Cover photo: Surveillance systems track the occurrence of common diseases, such as acute diarrhoea among children at the International Centre For Diarrhoeal Disease Research, Bangladesh Photo by Jean-Luc Ray for AKF The surveillance of undernutrition and other nutritional problems requires regular monitoring of the growth and development of children under age five years Photo by Pierre Claquin for AKF ISBN: 1-882839-15-3 Library of Congress Catalog Number : 92-75490 Dedicated to Dr. Duane L. Smith (1939-1992), Dr. William B. Steeler (1948-1992) and all other health leaders, managers and workers who follow their example in the effort to bring quality health care to all in need. Screening for and surveillance of anaemia is important among women of childbearing age, particularly wherever malaria and hookworm remain problems, such as in Bangladesh Photo by Jean Luc Ray for AKF An overview of PHC MAP The main purpose of the Primary Health Care Management Advancement Programme (PHC MAP) is to help PHC management teams collect, process and analyse useful management information. Initiated by the Aga Khan Foundation, PHC MAP is a collaborative programme of the Aga Khan Health Network1 and PRICOR. An experienced design team and equally experienced PHC practitioner teams in several countries, including Bangladesh, Chile, Colombia, the Dominican Republic, Guatemala, Haiti, India, Indonesia, Kenya, Pakistan, Senegal, Thailand and Zaire, have worked together to develop, test and refine the PHC MAP materials to make sure that they are understandable, easy to use and helpful. PHC MAP includes nine units called modules. These modules focus on essential information that is needed in the traditional management cycle of planning-doing-evaluating. The relationship between the modules and this cycle is illustrated below. PHC MAP modules and the planning-evaluation cycle PHC MAP MODULES 1. Information needs 2. Community needs 3. Work planning 4. Surveillance 5. Monitoring indicators 6. Service quality 7. Management quality 8. Cost analysis 9. Sustainability 1 The Aga Khan Health Network includes the Aga Khan Foundation, the Aga Khan Health Services, and the Aga Khan University, all of which are involved in the strengthening of primary health care. 2 Primary Health Care Operations Research is a worldwide project of the Center for Human Services, funded by the United States Agency for International Development. Managers can easily adapt these tools to fit local conditions. Both new and experienced programmers can use them. Government and NGO managers, management teams, and communities can all use the modules to gather information that fits their needs. Each module explains how to collect, process and interpret PHC-specific information that managers can use to improve planning and monitoring. The modules include user's guides, sample data collecting and data processing instruments, optional computer programs and facilitator's guides, for those who want to hold training workshops. The health and management services included in PHC MAP are listed below. Several manager's guides supplement these modules. These are: Better management: 100 tips, a helpful hints book describing effective ways to help managers improve what they do; Problem-solving, a guide to help managers deal with common problems; Computers, a guidebook providing useful hints on buying and operating computers, printers, other hardware and software; and The computerised PRICOR thesarus, a compendium of PHC indicators. OTHER HEALTH CARE Water supply, hygiene and sanitation School health Childhood disabilities Accidents and injuries Sexually transmitted diseases HIV/AIDS Malaria Tuberculosis Treatment of minor ailments Chronic, non-communicable diseases HEALTH SERVICES Planning Personnel management Training Supervision Financial management Logistics management Information management Community organisation MANAGEMENT SERVICES Health and management services GENERAL PHC household visits Health education MATERNAL CARE Antenatal care Safe delivery Postnatal care Family planning CHILD CARE Breast feeding Growth monitoring Nutrition education Immunization Acute respiratory infection Diarrhoeal disease control Oral rehydration therapy Antenatal care and the early detection and surveillance of high risk pregnancies are essential for lowering maternal mortality rates Photo by Jean Luc Ray for AKF The Primary Health Care Management Advancement Programme has been funded by the Aga Khan Foundation Canada, the Commission of the European Communities, the Aga Khan Foundation U.S.A., the Aga Khan Foundation's head office in Geneva, the Rockefeller Foundation, the Canadian International Development Agency, Alberta Aid, and the United States Agency for International Development under two matching grants to AKF USA. The first of these grants was "Strengthening the Management, Monitoring and Evaluation of PHC Programs in Selected Countries of Asia and Africa" (cooperative agreement no. OTR-0158-A-00-8161-00, 1988-1991); and the second was "Strengthening the Effectiveness, Management and Sustainability of PHC/Mother and Child Survival Programs in Asia and Africa<170> (cooperative agreement no. PCD-0158-A-00-1102-00, 1991-1994). The development of Modules 6 and 7 was partially funded through in-kind contributions from the Primary Health Care Operations Research project (PRICOR) of the Center for Human Services under its cooperative agreement with USAID (DSPE-6920-A-00-1048-00). This support is gratefully acknowledged. The views and opinions expressed in the PHC MAP materials are those of the authors and do not necessarily reflect those of the donors. All PHC MAP material (written and computer files) is in the public domain and may be freely copied and distributed to others. Contents QUICK START 1 INTRODUCTION What is surveillance? 3 Surveillance methods 8 How to use this module 13 SURVEILLANCE PROCEDURES Step 1: Specify the objectives 16 Step 2: Define the data to collect 20 Step 3: Select the methods and procedures 24 Step 4: Develop the data collection and reporting procedures 27 Step 5: Collect and report the data 34 Step 6: Analyse the data 37 Step 7: Investigate causation (optional) 45 Step 8:Develop an action plan 47 Step 9: Prepare and present reports 48 APPENDICES A. Blank worksheets and analysis templates 51 A.1 Planning, reporting worksheets 51 A.2 Templates for producing charts 53 B. Common diseases, definitions, and indicators 65 C. Case investigation: The example of neonatal tetanus 77 D. Determining the cause of death 81 E. Verbal autopsy protocols 91 E.1 Newborn and child death inquiry 91 E.2 Maternal death inquiry 108 REFERENCES AND BIBLIOGRAPHY 121 ACRONYMS AND ABBREVIATIONS 122 GLOSSARY 123 Acknowledgements In May 1992, based on recommendations from the International Conference on Management and Sustainability of PHC programmes and the PHC MAP Technical Advisory Committee, the PHC MAP Management Committee decided to add a module on surveillance to the PHC MAP series. An outline was agreed upon in August by the authors. Ms. Veronica Walker had the difficult task of liaison among the authors in Geneva, Jakarta and Karachi, and between the authors and the reviewers, in addition to typing several draft versions of the text. Her contribution is gratefully acknowledged.This module draws heavily from, and is patterned after, the excellent WHO/EPI manual, Training for mid-level managers: Disease surveillance. It also reflects, complements, and draws on material developed for Module 2, Assessing community health needs and coverage and Module 5, Monitoring and evaluating programmes, especially the instruments and indicators on mortality and morbidity. It also draws on material developed for Module 3, Planning and assessing health worker activities, particularly the CHW and clinic registers, which can easily be adapted for surveillance purposes. The verbal autopsy material is based, in part, on material prepared by Dr. Abhay Bang of the Society for Education, Action and Research in Community Health (SEARCH) Gadchiroli, Maharashtra, India. Thanks also to Dr. Ron Gray of the Johns Hopkins University for his contributions to this section. The text was prepared by Drs. Claquin and Reynolds. Dr. Marsh prepared the material on cause of death and "verbal autopsies". We wish to thank the participants of the Bangkok conference for their strong recommendation that this module be added to the series, and especially David Fraser, Peter Tugwell, Khatidja Husein, and Hugh Annett, who lobbied convincingly for its inclusion and who contributed their ideas and time to the development of the outline. We also want to thank Paul Richardson for his contributions to the internal debate on measuring mortality. The results are reflected in this module, as well as in Modules 2 and 5. The draft was tested in Nairobi, Kenya during an Aga Khan Health Service, Kenya, workshop in September on, "Introduction to the use of Epidemiology in the Surveillance of Morbidity and Mortality." Feedback from all participants and facilitators, especially Dr. Joseph Valadez and Dr. Ruth Chunge, led to simplification of the text in November and December. Reviewers: David Fraser Social Welfare Department, Secretariat of His Highness the Aga Khan, Aiglemont, France Khatidja Hussein Community Health Sciences, The Aga Khan University, Karachi, Pakistan Alain Roisin Brussels, Belgium George Stroh Center for Disease Control, Atlanta, Ga., USA Peter Tugwell Department of Medicine, University of Ottawa, Canada Quick start Setting up a basic surveillance system This is a simple program that you can use to set up your own surveillance system. The program is on the diskette that comes with this module, named MOD4_QS. Simply load it into a spreadsheet, such as Lotus 1-2-3 or Quattro Pro, then just follow these instructions. You can also do this manually. Just fill in the data table and make up your own graphs. The objective of this simple surveillance system is to monitor trends in morbidity and/or mortality of up to five diseases, or some other indicators that interest you. You can monitor anything you wish, as long as you have the data to do so. This program assumes that you have, or will be able to get, the data you need. First, decide which indicators you want to monitor. You should begin by identifying your major target groups, your health goals for them, and the indicators you will use to assess progress. You may have selected these indicators when you went through Module 1. If not, you can enter your summary in the following table. Here is an example: Target groups, Health goals, Indicators Children << 2 years, Reduce mortality Reduce morbidity Reduce disability, No. of deaths of children << 1 year No. of children << 2 years with 3rd degree malnutrition No. of cases of immunizable preventable diseases Married women 15-49 years, Reduce mortality Reduce fertility, No. of maternal deaths No. of live births Set up a table like the one below. Revise it to fit your own calendar. Enter up to five indicators in the left hand column. Each month tabulate the number of cases reported and enter them into the table. Here are some illustrative data. Table 1, J, F, M, A, M, J, J, A, S, O, N, D, Tot. A. Infant deaths B. 3rd degree malnutrition C. Cases imm. prev. dis., 0, 23, 3, 1, 32, 1, 2, 14, 2, 6, 18, 0, 2, 22, 0, 4, 26, 3, 3, 32, 0, 6, 38, 9, 2, 44, 8, 4, 56, 7, 5, 54, 8, 3, 52, 4, 38, 411, 45 D. Maternal deaths E. Live births, 0, 12, 0, 15, 0, 22, 0, 45, 1, 42, 0, 38, 0, 30, 1, 48, 2, 54, 0, 43, 0, 56, 0, 38, 4, 443 If you use the computer file that comes with this module, it will tabulate the totals for you and produce a graph for the first indicator (A). When you enter new data, the totals in the table and the graph will update themselves automatically. If you want to see a graph of another indicator, press /Graph, Name, Display, and select one of the names (A, B, C, D, E) and press Enter. This is a very basic graph (it is a "Quick start," after all). You can make more informative and sophisticated graphs. See the suggestions and templates in the appendices. Months Surveillance: A. Infant deaths Introduction What is surveillance? Surveillance of morbidity and mortality is the collection and analysis of selected health and vital events to: identify, investigate and control outbreaks or epidemics identify specific population groups at high-risk of illness and death from priority diseases confirm current priorities among disease control activities evaluate the impact of preventive and curative control activities on the incidence and prevalence of priority diseases in the community monitor disease trends so as to adjust plans to meet current needs A more formal definition comes from the Centers for Disease Control (CDC), which defines epidemiological surveillance as "the ongoing and systematic collection, analysis and interpretation of health data in the process of describing and monitoring a health event. This information is used for planning, implementing and evaluating public health interventions and programs. Surveillance data are used to both determine the need for public health action and to assess the effectiveness of programs."Guidelines for evaluation surveillance systems. CDC Morbidity and monitoring weekly report supplement. Vol 37, No. S-5, May 6, 1988> In the PHC MAP framework, morbidity and mortality surveillance data are used largely for monitoring and evaluation to determine whether the programme is having an impact on health. As such, it is an important tool that you can use to see if you are reaching your health goals. But surveillance data can also be used to assess health needs and to set health goals in planning. Whereas Module 2 concentrates on assessing health coverage, Module 4 is the one that concentrates on assessing health status. In this module, surveillance is presented largely as a technique for monitoring and evaluating morbidity and mortality. And two levels of analysis are included. The first is quantitative surveillance of the number of cases of a given disease or death. The second is qualitative investigation of the causes of a disease or death. Managers need to have accurate information on the number of cases and the percentage of their target groups that are affected by a specific disease, and that are dying from specific diseases. They need this information to be able to evaluate the impact of their PHC strategy on health and to determine if core PHC services are effective. Morbidity, Mortality Incidence, morbidity reports , mortality reports Cause , case,, outbreak investigation (diseases) , case,, outbreak investigation (autopsies) Sometimes they also need to know what has caused a particular death or the outbreak of a particular disease. This information may be essential to make sure that the assumed causes are confirmed, and to adjust the programme to prevent such events from happening again. Surveillance does not have to be complex to be useful. In fact, a common problem with many surveillance systems is that they are too complex and too large. By trying to collect comprehensive information on all diseases, there is little time left for analysis of the data and for taking action to reduce the number of cases of disease. For this reason, we strongly recommend that you collect only as much data as you can use and that you concentrate on the most important health problems. Principles of surveillance The following list will give you an idea of what makes up a good surveillance system. According to CDC guidelines, an effective surveillance system: addresses health events which are of considerable public importance, i.e., cause a substantial amount of morbidity and/or mortality, and are amenable to practical control or prevention; identifies and correctly classifies a large proportion of target health events; correctly reflects the distribution of events over time, place, and person; consists of components which include clear definitions of health events under surveillance, a clear and logical path for data flow, adequate knowledge of the population under surveillance and defined and appropriate methods for collection, analysis, interpretation and feedback of information gives rise to meaningful and effective public health action based on the data processed in the systemis uncomplicated is adaptable and responsive to new demands engenders a high level of participation provides information rapidly enough to allow effective action to be taken requires minimal resources appropriate to the circumstances.Ibid> Surveillance systems are often national or regional in scope. They are designed to collect data on specific diseases, such as AIDS, for use in national policy-making. The systems we are suggesting in this module use the same methods, but are designed to be used at local levels, to help PHC managers monitor morbidity and mortality in their own programme area. As such, they can reflect the health priorities and concerns of local communities. An example The University Medical Centre in Lome, Togo, has been monitoring data on paediatric deaths for nearly a decade. It is one of the few sites in Togo that has enough data to evaluate changes in infant mortality over time. Table 1 shows the ten leading causes of death among children less than five years of age in 1989. Malaria tops the list, accounting for one out of five deaths, followed by anaemia (13%) and malnutrition (12%). This type of data has been collected every year since 1985. Figure 1 shows the data on malaria and anaemia deaths for the five-year period 1985-1989. The data clearly show an upward trend in the proportion of deaths due to these two diseases. The anaemia deaths, were associated with malaria. Vernon, A. Cause of death from hospital data in developing countries: A review based on the CCCD experience with hospital-based mortality surveillance systems in Sub-Sarahan Africa. Draft discussion paper, 9 May, 1992.> Figure 1: Proportion of paediatric deaths from malaria and anaemia, Togo, 1985-1989 Table 1: Leading causes of death of children under age 5, Togo,1989 Cause No. deaths, Percentage Malaria, 199, 22 Anaemia, 117, 13 Malnutrition, 113, 12 ALRI, 85, 9 Meningitis, 66, 7 Diarrhoeal diesases, 58, 6 Coma,, NOS, 42, 5 Tetanus, 20, 2 FUO, 19, 2 Hepatitis, 18, 2% Limitations of surveillance Surveillance is a labour-intensive activity. It requires much effort to collect the needed data, especially if it is not already being collected as part of a routine recording and reporting system. Tabulation and analysis of the data is also time-consuming. For these reasons, most systems are limited to a few key indicators. It can take several years of data collection before trends can be identified, particularly for diseases, such as dengue, that break out every five years or so. You may be able to use retrospective data to look at past trends, but this is often difficult. This is because the indicators you want may not have been collected, are slightly different than what you need, or are not reported consistently. It can be difficult to assess impact if your target population is small, or if you cannot set up control groups to compare disease patterns between programme and non-programme areas. You need a large population to compute rates, especially mortality rates, and it may be difficult to identify all of the deaths that occur in your area. Reporting of surveillance data is often incomplete, especially from remote areas. Some programmes are fortunate to get 50% reporting, and even those reports may be incomplete or include errors which go undetected, if they are not checked carefully. Despite these limitations, surveillance can be an important tool for PHC managers. It can help identify changes in diseases and injuries that require immediate action. And by investigating causation, a manager can often identify programmatic changes that could help prevent a problem from recurring. Surveillance methods This module describes, and recommends, four surveillance methods. There are actually six principal methods, but two of them, vital registration systems and censuses, are not often feasible in developing countries. The six methods are summarised below. Surveillance methods, Description Routine reporting systems, Information is routinely collected and reported by PHC staff Sentinel reporting systems, A small number of reporting units (usually health centres or hospitals) carefully collect and report requested data Surveys and special studies, Sample surveys are conducted periodically to estimate the level of a disease or condition in a given area Case/outbreak investigations, Special investigations are undertaken to determine the cause of a disease or death and to recommend action to prevent its recurrence Vital registration systems, Public and private health providers report births,, deaths and other selected data to a central system Census, A count of all of a population,, often including questions about health habits,, diseases,, etc. The first three of these methods are used largely to gather quantitative information about the incidence of diseases and deaths. The fourth is used to investigate the causes of diseases and deaths. 1. Routine reporting systems In routine reporting, health staff, and sometimes, non-health people, collect information about the number of cases of reportable diseases and of selected deaths that occur in their area. Data are collected as a part of the routine screening and diagnosis process during home visits or during visits to health facilities, health centres and hospitals. Thus, the reports are based on direct contacts with the individual who is sick, dying, or has died. In some cases the information is provided second-hand from a relative, often the mother, friend or neighbour. Routine data is usually recorded in family or individual folders and then transferred to summary tally sheets for reporting of the aggregated data. At the end of each month the information is compiled and sent to a supervisor for further compiling and analysis. Most routine reporting revolves around health centres and hospitals. They obtain information from staff within the facility and from reports prepared by CHW's and other outreach workers. In some cases, community members help collect data from households. Advantages. Routine systems have the advantage of taking data from an ongoing system. Thus, they are inexpensive and efficient ways to collect information. A new system, with additional staff to operate it, does not have to be set up. The data cover all health activities from routine well-baby clinics to surgery. The recording and reporting systems have been standardised, which means that the same definitions are used throughout the system, the same type of data is collected, and the same reporting periods are used. This is a great advantage for making comparisons among areas. Disadvantages. These systems almost always provide an incomplete picture of the total number of cases that occur. Some of the reasons for incomplete reporting are: Not all cases come to the health facility for treatment. Some people go to private providers for treatment, others to facilities outside the area, still others do not seek treatment at all. Distance, transportation costs, hours of operation, loss of income, cultural taboos, and many other reasons tend to limit the use of health facilities. Not all cases are identified by outreach workers. CHW's may visit a household only once every 3-6 months, and are likely to miss many cases that are not picked up by the health facilities. When they do visit, they may neglect to ask about health problems that happened months ago, and the mother may not consider them important enough to report on. Some diseases, such as neonatal tetanus, are more commonly treated at hospitals than health centres, and therefore go undetected by the health centre and CHW. Complete and accurate reporting are always a problem in PHC. Some CHW's are illiterate, others concentrate on the most pressing concerns; some nurses and doctors skip over some items because they are too busy, they don't believe the item is important, or they didn't ask. Even when reporting is fairly regular and complete, there may be inconsistencies among health workers if they don't use the same procedures, definitions and guidelines. Nevertheless, routine reporting is the most common method used in surveillance. And because it is so inexpensive, it is likely to be attractive to many PHC managers. 2. Sentinel reporting systems In sentinel systems, a small number of health units in a programme area is selected to report cases of diseases and deaths that are seen and diagnosed at their facility. They may also be asked to report additional information, such as the age and immunization status of the children treated at the facility. Staff at sentinel sites are given special training and supervision to ensure that reporting is complete and accurate. The sites chosen are not necessarily chosen because they are representative of a given area. They may be chosen because they are likely to see cases of certain diseases, or because their caseloads are high enough that rare events are likely to be identified. Other criteria that are important are: reliability of data collection, timeliness of reporting, willingness of the staff to participate, high-quality laboratory or diagnostic capabilities. Hospitals are often included as sentinel sites, since they are likely to have much higher caseloads than health centres, they are more likely to see serious diseases, infant and maternal deaths are more likely to occur there, and they are more likely to have staff trained in diagnosis and data processing. Advantages. Sentinel sites may provide a more consistent picture of illness in a given area than routine reporting. Data collected from these sites may also show whether routine reporting is accurate or not. In addition, being chosen to participate in surveillance tends to motivate the staff to do the best they can to report accurately and on time. Disadvantages. A major disadvantage of sentinel systems is that they are not representative of the entire population at risk. The data they generate may not be of sufficient volume to generate rates and ratios, which are important for assessing changes in health status. Another disadvantage is that populations served by the sentinel facility may change, making the study of trends invalid. Nevertheless, the sentinel system has a major advantage in the quality of the data produced. They are also relatively inexpensive to set up and operate, especially in contrast to a "universal" system where all potential facilities would be required to report. 3. Surveys and special studies Sample surveys, such as those described in Module 2, are often used for surveillance. They usually provide a broad estimate of the incidence or prevalence of a disease. They can also be used to estimate mortality rates, although the sample sizes required to do this are very large. Surveys can also be used to evaluate the reliability of the routine or sentinel systems. For example, a morbidity and mortality survey in a high-risk, underserved population may be a more accurate and practical method for measuring disease patterns that tend not to be seen at health facilities. The surveys usually have to be repeated periodically, at least annually, to develop trend data. Although this can be expensive, the rapid surveys described in Module 2 are an affordable option. They also relieve the health staff of the burden of continual reporting. 4. Case and outbreak investigations These investigations are attempts to identify the causes of a death or disease. They are not alternatives to routine and sentinel systems, but are used as the next step in epidemiological investigation. In general, a "case" investigation is an investigation of a single case of a disease or death; an "outbreak" investigation is an investigation of many cases. However, when the occurrence of a particular disease is very low, polio for example, even one case can be considered an "outbreak." Purposes. The usual purposes of these investigations are to: confirm diagnoses and determine the causes (main/leading and associated) confirm the existence of an outbreak (an increase in the number of expected cases/deaths) identify the most appropriate control measures identify where and to whom to apply these measures determine why the outbreak occurred determine what can be done to prevent similar outbreaks in the future These investigations are conducted systematically, usually following a written protocol and a logical cause-effect chain of events to identify the underlying cause of the problem. An example of a protocol for a neonatal tetanus investigation is found in Appendix C. Advantages. These investigations have a programmatic as well as a medical objective. They can identify errors that should be corrected (e.g., using contaminated needles) and procedures that could be changed to prevent problems from occurring in the first place (e.g., increase home visits to monitor high-risk infants). The investigations can provide staff with a chance to learn more about the conditions and causes of diseases and deaths. They can use this information to improve internal procedures, and also to help community members understand what they can do to prevent unnecessary health problems. Verbal autopsy. A special technique for case investigations of deaths, the "verbal autopsy," is featured in this module. A PHC health worker, who needs to be trained in the technique, conducts an in-depth investigation of the death through interviews with the mother and anyone else who was a witness to the death and the circumstances leading up to it. See Appendix E.1 for the instrument that is used. The decision to conduct a case/outbreak investigation may be triggered by a standard protocol, such as investigate every case of polio, neonatal tetanus, and hepatitis; or, investigate every neonatal and maternal death. In many cases the decision may depend on the circumstances. If a strange or unexpected event occurs, this might be enough reason to investigate it, especially if it could be a threat to health. In summary, each of these methods has its advantages and limitations. You may want to try a combination of systems, relying on your existing reporting system for as much data as is reasonable, setting up a small sentinel system to continuously gather additional data on a few important health problems, and adding mortality and morbidity questions to an annual community survey to assess changes in health needs as well as coverage. Also, don't expect too much of your surveillance system. Surveillance usually monitors outcome data - impacts and effects. A good system might identify changes in outcomes, but it will not necessarily explain what caused the changes. You may need to use other modules in the PHC MAP series to look more closely at programme inputs, processes, and immediate outputs to find the cause of a problem - or of a success. How to use this module Who and when This module is designed to provide PHC managers, whether working in government or private organisations, with simple and inexpensive tools for setting up and operating a local surveillance system. PHC managers are not likely to design or operate the system themselves. Staff with a background in planning, evaluation, MIS, as well as epidemiology and any similar discipline, can easily learn how to use the tools in this module to set up a system. PHC consultants, especially those who are interested in data for decision-making, should also find the module helpful. How the module is organised The module can be easily adapted to meet local needs. The procedures that are described in the next section are general and flexible. Those programmes that already have a surveillance system will find that the module can help them simplify their system. Those who do not have a system already should find that they can start small, designing a highly focused system of a few indicators at first, and expanding it later if it proves to be worthwhile. This User's guide has two main sections, in addition to the Introduction and Quick start. Procedures (or steps) come next, describing a simple, but systematic process you can go through to determine the kind of surveillance you need and how to set up a system that meets your needs. The appendices contain useful guidelines, worksheets, and tools that you can use in designing your surveillance system. There is a computer file that comes with this guide that includes a number of "templates" in Lotus 1-2-3 and Quattro Pro, or pre-designed tables that you can use to tabulate your data. Each template includes pre-designed graphs that you can display just by pressing one or more keys on your computer's keyboard. A Facilitator's guide has also been developed to help you set up a workshop to train staff to use the module. How to begin If you haven't tried the Quick start section, you may want to do so, just to see how easy it is to set up a simple, but useful surveillance system. Then go to the next section and read through the Procedures. This should give you a better idea of what a surveillance system involves and the kind of system you want for your programme. Surveillance procedures This section describes how you can design and operate your own surveillance system using one or more of the methods described in the Introduction. There are eight steps in these procedures. The first three describe how to design a system that will meet your needs. The next two describe how to develop the appropriate data collection procedures and then to collect and tabulate the surveillance data. The last three steps describe how to analyse and interpret the data so that you can take action quickly, if indicated, and report your findings to others who might benefit from your information. These eight steps are summarised below. Steps in designing and operating a surveillance system Step 1:, Specify the objectives of surveillance Step 2:, Define the surveillance data to collect Step 3:, Select the surveillance methods Step 4:, Develop the data collection procedures Step 5:, Collect and tabulate the data Step 6:, Analyse the data Step 7:, Take action Step 8:, Prepare and present reports The major decisions you will have to make can be recorded on the worksheets illustrated in this section. There are blank worksheets in the Appendices that you can copy. Step 1: Specify the objectives Before you begin to design a system it is very important that you are absolutely clear on what the system is supposed to do. This step guides you through a simple process to help you specify: 1)purpose of the surveillance 2)users of the information that the system produces 3)scope of the surveillance (which geographic area and which PHC services it will cover) 4)target groups that will be monitored 5)whether you just want to identify cases of mortality and/or morbidity only, or, also the causes of mortality and/or morbidity 6)time period the surveillance will cover - one year, five years, three months and how frequently you will collect and process your surveillance data. WORKSHEET 1: SPECIFYING THE OBJECTIVES OF SURVEILLANCE Purpose(s) Assess needs X Monitor trends X Explain causes Identify risk factors Evaluate impact on infant & maternal mortality Identify outbreaks Other:________________________ Identify unusual events Other:________________________ User(s) Board of directors PHC Manager Community Government officials PHC Staff CDC* Supervisors Other:___________________ ___ Donors Other:_______________________ Scope * local Communicable Disease Centre Geographic area(s): Entire project area Programme service(s): M/M preventable by PHC services Target group(s) X Children 1 month Women 15-49 yrs X Children 12-23 months Married women 15-49 yrs Children 1-4 yrs X Pregnant women Children <<5 yrs Other: ________________ _______ Cases Causes Other X Mortality X Mortality Specify:____ X Morbidity Morbidity Specify:____ Time period: 12 months Frequency: Monthly & quarterly If you completed Module 1, it can be very helpful to refer to Worksheets A and B. They will help you remember your overall information needs and how surveillance fits in. When you define what you want to monitor, keep your programme's goals clearly in mind. If a goal is to reduce mortality or malnutrition, then this is probably what you should design your surveillance system to monitor. It may be useful to go through this step with a group, so that all significant points of view are heard. Involve the potential users of the information if at all possible. Make sure someone is in the group who understands what can and cannot be collected. This will help you avoid designing an unrealistic system. Also, make sure that the objectives are limited to something reasonable. Don't try to measure everything. It is probably a good idea to limit it to 3-5 important items, especially if your programme is small. 1. Purpose: First you need to specify the purpose of your system. Most systems can be used to meet several purposes, especially if they are designed to do so from the beginning. Among the principal purposes are the following: Assess needs. You may want to confirm that your current disease control activities are correct, or you may want to update an assessment of needs in your area. Identify risk factors. You may want to identify specific population groups that are at high-risk of illness and death so that you can develop interventions to protect them. Identify outbreaks. A major purpose of surveillance is to identify quickly any outbreak or epidemic so that it can be brought under control. Outbreaks of communicable diseases, such as measles and AIDS, are especially serious and should be identified and acted upon immediately. Identify unusual events. Equally important is to identify quickly any event that is unexpected but a serious threat to health. Examples would be a number of deaths from drowning and food poisoning among school-age children. Monitor trends. Many diseases are seasonal and you may want to watch trends to identify unusual deviations from expected patterns. You may also want to track improvements in disease control over time. Evaluate impact. Most managers want to know if their programme is having any impact on health. Surveillance can help you watch key trends, for example, in infant mortality and the incidence of diarrhoeal diseases. Explain causes. In some situations you may want to investigate the cause of an illness or death to determine whether your programme can do anything in the future to prevent such an occurrence. Some programmes investigate every maternal death for this reason. Others look at reasons for complications during childbirth to identify changes in procedures that might be warranted. Don't expect your surveillance system to tell you everything; it will mainly alert you to a problem. Then you can use some of the other modules in the PHC MAP series to get more detailed information on the cause of the problem. 2. Users: The person or persons who will use the results of the surveillance system should decide what the purpose is, and specify the scope, target groups and other elements of the surveillance objectives. Otherwise the information will be of little use. There may be one primary user, often the PHC manager, and one or more secondary users, superiors, donors, and so forth. If your system will be part of a larger regional or national system, then you will probably be expected to report your findings to your communicable disease centre. If you have multiple users, take care to make sure that each of them is consulted. You may have to negotiate compromises to avoid designing an enormous system to meet everyone's needs. 3. Scope: Next, make sure to define the scope of the geographic and programmatic areas to be monitored. Most programme managers would like to monitor their entire catchment area and all major diseases, but that may be too much to be feasible. You may have to settle for setting up a sentinel system that is representative, or that watches for specific diseases. Set priorities. Identify what is most important in your situation. Table 2: Common priority diseases and health problems for surveillance in developing countries Vaccine-preventable diseases Enteric diseases Parasitic diseases Measles Cholera Malaria Neonatal tetanus Dysentery Onchocerciasis Tuberculosis Watery diarrhoea Schistosomiasis Poliomyelitis Lymphatic filariasis Diphtheria Leprosy Pertussis Ascariasis Mumps Guinea worm Rubeola Trypanosomiasis Pregnancy-related Other infectious and Other important diseases problems communicable diseases Obstructed labour Chicken pox* STD/HIV/AIDS Eclampsia Yellow fever Malnutrition Prematurity Meningitis Anaemia Post-partum infection Haemorrhagic fever Heart disease Diabetes *Several epidemiologists have recommended monitoring chickenpox as a measure of the quality of surveillance. 4. Target groups: Most PHC programmes have a limited number of target groups, usually children under age five or under age three years, or two to four years old, etc., and married women in the reproductive age range. Don't try to monitor all age groups for all diseases, unless you plan to set up a comprehensive system. Again, set priorities. Which are your priority target groups, and what aspect of their health are you most concerned about? If you completed Worksheet A (in Module 1), you have probably already identified your key target groups and the impact goals you want to measure. 5. Cases and/or causes: You will need to decide if you just want to monitor quantitative data on the number of cases that occur, or if you also want to know the causes of a death or disease. The latter requires an investigation, which can be time-consuming and expensive. You may want to be selective and only investigate causes of major outbreaks, infant deaths, and certain high-priority concerns, e.g., a decline in nutritional status. 6. Time period and frequency: You will probably want to monitor events for several years, and process your data every one to three months. This will enable you to develop trend patterns. In some cases, if you have the data, you may want to go back in time to examine previous trends. If the data are available, that could be a useful and not particularly expensive undertaking. Other considerations: Although this module concentrates on mortality and morbidity, keep in mind that you can use surveillance systems to monitor a number of other things, as well. For example, you can monitor coverage; numbers of children fully immunized; numbers of new acceptors of family planning. You can also monitor disabilities and fertility. If any of these are important to you, make sure you include them in your design. Don't be afraid to leave some of these substeps incomplete at first, and to come back to this step later to revise your objectives. You may find out as you go along that some objectives are just not feasible, or that some are more important than you originally thought. Step 2: Define the data to collect This step guides you through a series of substeps to determine precisely which data your system will collect, how frequently, from which source, and using which procedure. You can use Worksheet 2 to keep track of your decisions. This step is a bit more technical than Step 1, so you may not want to have too large a group working on it. It would help to have at least one representative of management, one or two staff who understand the current record-keeping system, and one person, perhaps a consultant, who understands the requirements for and constraints on collecting data on mortality and morbidity. Again, remember to keep your programme's goals in mind when deciding what to monitor. Target group mortality. Enter the target groups you identified in Step 1 in the left column of Worksheet 2. In the next column enter whether you are going to monitor the mortality of each target group. It is generally unrealistic to try to monitor all deaths. The more practical, and programmatic priority, would be to monitor; maternal deaths: These are deaths that occur during pregnancy, delivery, and up to 42 days after delivery. infant deaths: These are deaths that occur between birth and the first birthday. You may want to distinguish between perinatal (died at birth), neonatal (died within the first 28 days of life), and post-neonatal (died between the 29th and 365th day of life). The total of these, of course, is infant mortality. WORKSHEET 2: SPECIFY THE SURVEILLANCE PROCEDURES Step 2: Define the data to collect Step 3: Select the methods and procedures, +, +, +, +, + Target group, Mortality/ Morbidity/ Other, Indicator, Data collection, +, + ^, ^, ^, Frequency, Source, Method* Children << 24 mos., Mortality,, all causes, No. of deaths children <<24 months., Monthly, CHW HH visit reports, Routine,, all health centres , Morbidity and mortality of NNT, No. of cases NNT Mortality: cause of death, Quarterly, Clinic records, Routine,, all health centres Verbal autopsy for cause Children <<5 yrs., Morbidity,, all immunizable diseases, No. of cases by disease,, age,, sex,, location, Monthly, Clinic records, Sentinel,, 6 health centres Pregnant women, Mortality, No. of deaths,, all causes Cause of death, Quarterly, Clinic,, CHW & TBA reports, Routine,, all centres Verbal autopsy for cause Routine, Sentinel, Sample survey/special study, Case/outbreak investigation, Verbal autopsy Investigating cause of death. If you want to determine the cause of death, in addition to keeping track of the number of deaths, then you will probably need to conduct in-depth interviews to gather the needed information. If you are able to get reliable autopsy reports, by all means use those. If not, you can try the "verbal autopsy" approach. Appendix E describes how to do this and includes prototype questionnaires that you can use. Before you make your final decision on this question, however, look at the following chart. It summarises the causes of death that you are likely to be able to determine through verbal autopsies. In general, the causes of some deaths are fairly easy to assess, others are more difficult. For example, if a child dies in an automobile accident, from a fall, or is electrocuted, the cause is usually obvious to witnesses. On the other hand, all but the most obvious causes of neonatal deaths are very difficult to determine. The majority <%-3>of neonatal deaths are classified as "unknown." See Appendix D for a more detailed discussion of the feasibility of ascertaining the cause of childhood and maternal death. Table 3: Causes of death that can be determined through interview Childhood mortality Maternal mortality Relatively easy Injury Post-partum haemorrhage Neonatal tetanus Obstructed labour Measles Eclampsia Diarrhoea Post-partum infection Acute respiratory infection Abortion-related Malnutrition Relatively difficult Neonatal sepsis Morbidity. Repeat these steps to list the diseases that you will monitor for each target group. Again, some are likely to be more important than others to you. Keep your surveillance objectives (Worksheet 1) in mind as you select your priorities. Use Table 2 checklist of common diseases to identify the ones you want to include in your surveillance system. See Appendix B for more information on several of these, including standard case definitions, lay definitions, and indicators. Investigating cause of morbidity. As with mortality, if you plan to investigate the cause of one or more of these diseases, you may have to conduct in-depth interviews to get the information you need. In most cases it is unrealistic to investigate more than a few cases each year. You may want to take a selective approach to this issue and determine on a case-by-case basis if an investigation is warranted. Obviously, if most of your children have been fully immunized and there is an outbreak of polio, you would want to investigate that immediately. Thus, you might limit investigations of causation to significant and unusual events. That could come as a result of the outcomes of steps 6 and 7, when you analyse your data and decide whether you need to take action. How much information is needed? In addition to a count of each case, you may want to collect other information, such as age, gender, immunization status, whether prescribed PHC treatment was provided, and so forth. We strongly recommend that you only collect as much information as you will use. That means: Routine reporting should be limited to: a) the total number of cases of death or disease in each priority category; and b) their distribution by gender, age group, residence, date of occurrence, and, depending on local conditions, the cause(s) of infant and maternal deaths. Sentinel sites should report all of the above in addition to information on immunization status for immunizable diseases, on index cases, and on contributing factors. Maternal and infant deaths should be systematically investigated by detailed verbal autopsies. Routine reporting of age and gender is not superfluous information. There is mounting evidence of mortality differences between boys and girls, and this should be monitored. Age at death is important for maternal as well as infant mortality. Indicators. At this point you can specify the indicator for each item that you want to monitor. If you went through Module 1, you may already have a tentative list that you can examine. Or turn to Appendix B for a list of common indicators for several of the diseases and health problems mentioned above. Finally, you can consult Module 5: Monitoring and evaluating programmes, which contains a complete set of morbidity, mortality, disability and fertility indicators. An important decision that you need to make is whether or not to calculate mortality and morbidity rates and ratios. Unless you have a large population base, you probably will not have enough cases to do this. A rough rule of thumb is that you will need to have a population of at least 50,000 to identify enough infant deaths to compute a reasonable infant mortality rate. Since maternal mortality is far less common, you would need a population about ten times larger. You also need to collect data from the entire population. You cannot use a sentinel system. Thus, you must have a very good recording and reporting system that identifies every infant and maternal death. That is difficult to find in most developing countries. The other approach is to collect the required data through sample surveys. Module 2 describes how to do this, and the sample size requirements, which are still large. In general, you would need a sample of about 2,000 respondents (not population) for infant mortality, and 7,000 for maternal mortality. The data collection procedures are relatively simple, and Module 2 includes sample questionnaires that you can use. However, the computation and analysis procedures are difficult, especially for maternal mortality estimates. The advice from Module 2 is to consult a trained demographer if you want to measure mortality rates or ratios. See Appendix G.5 in Module 2 for a fuller discussion. Finally, don't forget that you can also monitor coverage, fertility, disability and other non-mortality/morbidity items. If you plan to do that, consult Module 5 for extensive lists of indicators that you can adapt to your system. Step 3: Select the methods and procedures In this step you will select the data collection procedures for each of the indicators. Basically, that means that you will decide whether to gather the data from routine reports, set up a sentinel system, or conduct special surveys to collect the data. The advantages and disadvantages of these approaches were described in the Introduction. If you are not sure which approach to select, you may want to read this section again. Keep in mind that you can select more than one procedure. You could collect most of your data through routine records and set up a sentinel system for a few additional indicators. Also, keep in mind that you can change the procedures at any time. You can add indicators, drop some, switch to rapid surveys to collect data on an annual basis, and so forth. Fill in the second part of the last worksheet (Worksheet 2), which asks you to indicate: 1) the frequency of reporting; 2) the source of the data for the indicator; and 3) the data collection method for each indicator. As the example shows, most of the data will be reported monthly, and some quarterly. The sources are largely routine reports from CHWs and the health centres. And the methods will rely largely on routine reporting, supplemented by a surveillance system in six health centres and verbal autopsies to determine the cause of each maternal and infant death. Frequency of data collection/reporting. For the majority of indicators of a routine system, the frequency of data tabulation and reporting is monthly. However, for high priority indicators, such as a case of polio in a fully covered area, the reporting of cases should be weekly, or instantly so that action can be taken immediately. Sentinel sites usually report weekly so that outbreaks can be detected quickly. For unusual or significant events, the reports should be immediate. Sources. Data are either already available or not. If they are available, they are called secondary sources, which include all kinds or reports and records. If they are not available, they have to be collected especially for surveillance, and are called primary sources, which include observations and interviews. The most common sources for surveillance systems are listed below, using a CHW as an example of the data collection agent: Direct observation, Direct reports, Records and reports Unstructured: CHW happens to see event, Free association: CHW overhears discussion, Statistical: CHW keeps track of cases Structured: CHW uses checklist to observe, Unstructured interview: CHW talks to mother, Documents: CHW prepares monthly report , Semi-structured interview: CHW uses checklist to talk to mother, Secondary reports: CHW reads laboratory reports , Structured interviews: CHW conducts formal interview, , Tests: CHW examines child, , Inventories: CHW counts EPI supplies , The sources of data are always people, originally. Thus, you can substitute "laboratory technician," or "doctor" for CHW and you will quickly see that there are many potential sources of data for surveillance. In PHC programmes these people might be clients, health workers, or non-health workers þ school teachers, community leaders, university researchers. The main sources of surveillance data will usually be health workers: outreach, health centre staff, and health providers in other parts of the community, hospitals, maternity clinics, private midwives and doctors, etc. Typically, the health workers observe, interview, and then record data about morbidity and mortality. They do not always report this information. If you are setting up your own system, don't rely exclusively on available reporting systems, such as health centre and hospital activity reports. There may be a better way to get information quickly and easily, even if it means setting up a new systems for a while. See Module 5, Monitoring and evaluating programmes for some suggestions. Methods. The introduction described six of the most common methods for collecting surveillance data. Although this module emphasises two of them, routine and sentinel reporting systems, the others are also listed to remind you that there are other options. You will need to consider your selections carefully, especially if you are going to set up a new sentinel reporting system. This can be expensive and time-consuming, especially if the reporting requirements are large. However, if you can keep it small and focused on the data you need most, it can be invaluable to you. You should complete the last part of Worksheet 2, then review your overall design, consolidate or revise it as you think necessary, and then move on to designing data collection procedures. Surveillance methods, Description Routine reporting systems, Information is routinely collected and reported by PHC staff Sentinel reporting systems, A small number of reporting units (usually health centres or hospitals) carefully collect and report requested data Surveys & special studies, Usually sample surveys to estimate the level of a disease or condition in a given area Case/outbreak <%-3>investigations<%0>, Special investigations of one or more cases of a disease or death to determine its cause and recommend action to prevent its recurrence Vital registration systems, Public and private health providers report births,, deaths and other selected data to a central system. Census, A count of all of a population,, often including questions about health habits,, diseases,, etc. Step 4: Develop the data collection and reporting procedures Once the surveillance procedures have been selected, you can begin to design your data collection procedures. This consists of three sub-steps: Develop operational definitions of cases Develop or revise the data collection/recording instruments Pretest the instruments Develop an operational definition of a case A "case" is defined as an individual situation or occurrence. In health, a case is usually an individual person who has a particular disease. Thus, a person who comes to a clinic complaining of severe diarrhoea is a case. A child suffering from 3rd degree malnutrition is a case. An important surveillance requirement is to be able to define each case that is identified. A mother may bring a child in for examination, not knowing what is wrong. The health worker must diagnose that child's condition so that it can be treated properly, but also so that the health problem can be accurately identified, recorded and reported. Appendix B contains a listing of common diseases together with their standard and "lay" (non-clinical) definitions. Two examples are shown below. Disease, Standard case definition, Lay definition Measles, History of a generalized maculo-papular rash lasting three or more days and history of any one of the following: cough,, coryza,, conjunctivitis., History of fever and rash and any one of the following: cough,, running nose,, red eyes Diphtheria, Acute pharyngitis,, acute nasopharyngitis,, or acute laryngitis,, with a pseudo membrane, Sore throat,, with grey patch or patches in the throat You need to have a case definition for every disease that you plan to monitor. This is necessary to ensure that all health workers use the same definitions and criteria to diagnose a specific disease. Otherwise, the surveillance system will be of no use. You can use the WHO definitions in Appendix B as a start. There may be variations in your area that require a slightly different definition. Contact your local epidemiology or communicable disease control centre to get the standard definitions used in your country. There are a few important rules to convey to your staff about identifying and recording cases: Avoid double-counting. If a child makes two visits to a health centre for the same disease episode, count it as one case only. Only count those cases that have been diagnosed by a health worker. Count separately, but do not record or report cases that have been reported by community members, unless they have been diagnosed by health staff. This is necessary to avoid misdiagnosis and double counting. Count current cases only. You need to set a time frame for including cases. That could be the number of cases that occurred during the past seven days. Then, do not include any cases that occurred eight or more days ago. Develop/revise data collection/recording instruments There are three types of instruments that you might use in routine and sentinel surveillance: registers, survey questionnaires, and case investigation protocols. Registers. The most likely instruments that you will use for routine surveillance are your clinic and outreach registers. If you set up a sentinel system you may want to expand them to include additional data, or you may want to develop a special form to collect additional data on a limited number of diseases. Clinic registers. All ongoing PHC programmes will already have registers, which can probably be used for surveillance as well as basic record-keeping. If they aren't exactly what you need, you can probably adapt them fairly easily. Module 3, Planning and assessing health worker activities, suggests two clinic registers that can easily be used for surveillance. Individual clinic treatment record. This form records details of visits made by an individual for routine care as well as for treatment of various diseases. The form, a portion of which is shown in Exhibit 1, can be adapted easily to include whatever disease you are interested in monitoring. A separate record can be used to tabulate entries from individual forms. This form is useful if you want to include special characteristics of each case, such as age, sex, marital status, etc. Daily clinical treatment record. This form is used to record all visits made to a facility each day (Exhibit 2). In addition to identifying all cases seen during the day, you can also record the individual's sex and age group. Daily totals can easily be summarised each week or month. Outreach registers. Module 3 also describes and illustrates an outreach activity register that can also be used for surveillance (see Exhibit 3). This two-page form is used by each CHW to record the results of monthly visits that are made to each household in the CHW's work area. This form is designed to collect data on births, deaths, immunization, diarrhoea and nutritional status of all children under age five, and the pregnancy and immunization status of all married women in the household. The form can be adapted easily to collect other information needed for surveillance, including morbidity history, current health problems, and even cause of death. Exhibit 1: Excerpt from Module 3 <197> Individual clinical treatment record Household #: 1146 Individual #: 38 Medical record #A-65 For: unregistered only, +, +, +, +, +, +, +, +, +, +, + Patient name: Rosa Sanchez Date of birth 28/11/53, +, +, +, +, +, +, +, , , , Father/Husband name: Pedro Sex (M/F) F, +, +, +, +, +, +, +, , , , , VISIT DATE, 18/9, 17/10, 15/11, , , , , , , , Routine antenatal care, 4, 4, 4, , , , , , , , Routine well-baby care, , , , , , , , , , 1, Tuberculosis, , , , , , , , , , 2, Polio, , , , , , , , , , 3, Diphth/Pertus/Tetanus, , , , , , , , , , 4, Measles, , , , , , , , , , 5, Mumps, , , , , , , , , , , Other diseases, , , , , , , , , , 6, Malnutrition, , , , , , , , , , 7, Diarrhoea/Dysentery, , , , , , , , , , 8, Intestinal parasites, , , , , , , , , , Exhibit 2: Excerpt from Module 3 Daily clinical treatment record Name of clinic Mt. Vernon Date: 28,, Nov,, 1992 MO/CHN: Pamela, +, +, +, +, +, +, +, +, +, +, +, +, + Registered, +, +, +, +, +, +, +, +, +, +, +, +, + , , 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, , Total , Household no./Reg. no., 146, 128, 62, 243, 416, 24, 518, 661, 84, , , , ID #, 238, 216, 45, 306, 610, 47, 618, 721, 96, , , M=4 , Sex: M/F, F, M, M, M, F, F, M, F, F, , , F=5 , Age: Years <<1, , , 4, , , , , , , , , 2 , 1-5, , 4, , , 4, , , 4, , , , 3 , 6-14, , , , , , , , , , , , , 15-49, 4, , , , , 4 , , 4, , , 3 , Over 49, , , , 4, , , , , , , , 1 , Routine ANC, 4, , , , , 4, , , 4, , , 3 , Routine well-baby, , , 4, , 4, , 4, , , , , 3 1, Tuberculosis, , , , , , , , , , , , 2, Polio, , , , , , , , , , , , 3, Diphth/Pertus/Tetanus, , , , , , , , , , , , 4, Measles, , 4, , , 4, , , 4, , , , 3 5, Mumps, , , , , , , , , , , , , Other diseases , , , , 4, , , , , , , , 1 6, Malnutrition, , , , , , , , , , , , 7, Diarrhoea/Dysentery, , , , , , , , , , , , 8, Intestinal parasites, , , 4, , 4, , , , , , , 2 Exhibit 3: CHW Activities register (left side) , , , +, +, Children <<3 years, +, +, +, +, , Sr. No., House-hold No., I.D. No., Age, Newly identified birth (Jan-Mar), +, +, Jan., +, Feb., March, +, + , , ^, ^, Live DT BW, +, Still DT, Wt., Wt.* Ch., Wt.* Ch., Wt.* Ch., Nut. St.**, Imm. St.***(<<1) , , , (1), (2), (3), (4), (5), (6), (7), (8), (9), (10) 01, 242, C5, 2.1, , , , 10.5, <196>, O, O, N, C 02, 243, C2 C3, 2.7, NB, 12/3, 2.4, , 8.3, ++, , O, I, I, C, IC 03, 244, C6, 2.1, , , , 11.2, ++, ++, NW, N, IC 04, 245, C2, C3, 1.8, 0.8, , , , 10.4, 8.2, +, 0, , N, N, IC, A 05, 246, C4, C5, 2.9, 1.8, , , , 10.1, 8.3, +, +, 0, +, I, I, C,C 06, 247, CI, 2.6, , , , 8.5, ++, , , II, C, 07, , 248, C4, C5, C6, 2.1, 1.1, 0.2, , , , 7.6, 8.0, 3.4, +, +, 0, 0, +, 0, 0, 0, II, I, I, IC, IC, N Total , x, x, 12, 1, LBW=1,,, +=7, O=0, O=4, NW=0, +=3, O=4, -=3, NW=0, +=1, O=6, -=3, NW=1, N=4, I=6, II=2, III=0, C=5, A=1, IC=5, N=1 * WEIGHT CHANGE, +, +, ** NUTRITION STATUS, +, +, ++, =, Increase in weight, N, =, Normal, O, =, Same weight, I, =, First degree malnourished, , =, Decrease in weight, II, =, Second degree malnourished, BW, =, Birth weight, III, =, Third degree malnourished, LBWNB, ==, Low birth weight, Newborn, NW, =, Not weighed, Exhibit 3: CHW Activities register (right side) Date of visits, +, +, Deaths, Married women, +, +, +, +, +, , Jan., Feb., Mar., Jan-Mar., Jan., Feb., March, +, ***Imm. sta., +, ID#, Number of family members,,,ID#,PR, PR, FP, PR, All, DLVD., , (11) 12/1, (12) 15/2, (13) 18/3, (14), (15), (16), (17), (18) , (19) C, (20), (21) M, (22) 6 12/1, 14/2, 18/3, D 25/12, 7, 8, , , C, C, M, 8 13/1, 14/2, 12/3, , , , , , IC, , M, 3 13/1, 14/2, 12/3, , 4, 5, , 6, C, , M, 5 10/1, 12/2, 13/3, , , , , , IC, , M, 6 10/1, , 13/3, , , , , , IC, , M, 3 10/1, 13/2, 13/3, , , 4, , , ICIC, , M, C1, 5 7, 6, 7, 1, 2, 3, , 1, C=3, IC=5, N=0, C=1, IC=0, N=0, 8, 36 IMMUNIZATION STATUS, +, +, PR, = , Pregnancy month of pregnant woman C, =, Complete, DLVD, = , Women who have delivered during the quarter IC, =, Incomplete for age, FP, = , Family planning A, =, Appropriate for age, DT, = , Date N, = , No immunization, , , , , , , , Death report. If you don't already have one, you can supplement the clinic and outreach register with a simple "cause of death" form. Exhibit 4 is an example. One of these forms would be filled out for each death identified during the reporting period. The results can then be tabulated at the end of the reporting period. Exhibit 4: Death report Cause of death, + Name: Conchita Juarez Sex Address: House # 146,, Mt. Vernon F, Date of birth: 12 Oct. 1991 Date of death: 14 Nov. 1992 Age at death: 13 months Cause(s) of death: = main 1. Measles* 2. 3. 4., Supporting CHW observed rash * Comments: Child had not been immunized against measles, + Signed: Pamela Jones , Today's date: 16 Nov,, 1992 Survey questionnaires. Module 2, Assessing community health needs, includes three instruments that you can use if you plan to conduct surveillance through sample surveys. The first is a Vital events and health status questionnaire. It allows you to measure the recent morbidity and mortality status of every person in a selected household. The second is a Child mortality questionnaire, designed specifically to collect data on mortality of children under age one, between ages two to four, and under age five. The third is a Maternal mortality questionnaire. You can adapt each of these questionnaires to suit your programme's needs and the diseases you want to monitor. Exhibit 5 is an excerpt from the morbidity section of the vital events questionnaire. Case investigation. The current module includes two prototype instruments that you can use to do in-depth investigations of the causes of a disease, health problem or death. Case investigation form. Appendix C contains a protocol for investigating a case of neonatal tetanus. Verbal autopsy form. Appendix E contains two detailed protocols, one for investigating the cause of a child death, the other for a maternal death. Pretest the instruments After you have selected (or developed) your instruments, you should pretest them under real conditions. That is, try them out in your programme to see if they are understandable to your staff, easy to use, and produce the type of data you need. Exhibit 5: Excerpt from Module 2 <197> Vital events and health status Morbidity 14. Is there anyone in your household who has been sick this week? Yes__4___(1) No_____(0) Go to Q.60 DK/NR____(9) Go to Q.60 Who is/are sick (probe and fill out following table, using SI. No., e.g., 7.3, 9.2): 15, , 16, 17, 18, 19, 20, 21 SI, Name, Age m/y, Sex M/F, Disease (Code) 01,, 02,, etc, Treated Y/N, Where Treated 1,2,3,, etc., Outcome 1,2,3,, etc 2, Pasquel, 8/0, M, 01, Y, 5, 2 , , , , , , , 18., Disease Code:, , , 20., Where Treated, 21., Outcome 01 Diarrhoea/Dysentery, 08, Polio, 1, G<%-9>overnment clinic/hospital, 1, Cured/Recovered 02, Anaemia, 09, Tuberculosis, 2, Mobile clinic, 2, Still Recovering 03, Scabies, 10, Acute respiratory inf., 3, Private clinic/hospital, 3, Permanent disable 04, Diphtheria, 11, Fever, 4, Private doctor, 4, Died 05, Whooping cough, 12, Malaria, 5, Private midwife/nurse, 5, Other: 06, Tetanus, 13, Other:_________, 6, Traditional practitioner, 9, DK/NR 07, Measles, 99, DK/NR, 7, Pharmacy/drug store, , , , , , 8, Other:___________, , , , , , 9, DK/NR, , Step 5: Collect and report the data Training, supervision and quality control Once you have your procedures and instruments, you can start your surveillance. Good training and supervision of your staff in data collection, tabulation, and reporting will be crucial to the success of the system. Anyone who is involved in any of these functions should be trained. That includes community volunteers, TBAs, school teachers, as well as CHWs and health centre staff. Training should cover the following topics: the purpose and utility of surveillance how to recognize and classify specified diseases using standard or lay case definitions how to record data on the clinic and outreach registers how to summarise and report the data on a weekly or monthly basis how to determine if further investigation is needed. Case investigations, including verbal autopsies, require special attention. One consideration is the number of investigators you need. If there will be few investigations, it may be best to train a few people and let them handle all of the investigations. But if there are likely to be many investigations, and if they are likely to be spread out over a wide geographic area, it may be better to train a larger number of people who are assigned to different areas. Investigations are not easy to conduct, and the people chosen to do this must have a certain amount of persistence and dedication to their task. It is natural for PHC staff to dislike asking intimate questions, and a common problem is that they avoid some or all of the required questions. Some ways to deal with this are: 1) make sure the interviewers are well trained to begin with; 2) watch them during practice sessions to see if they have the tenacity to go after the needed information; 3) make sure that there is enough time allocated to allow the interviewer to get to and from the site, establish rapport, and probe for answers; 4) encourage the investigators to express their concerns and feelings, and help them to deal with them; 5) emphasise the importance of their task for the health and well-being of others; 6) teach them culturally appropriate ways to obtain the information; and 7) accompany them periodically on investigations to observe their technique and provide constructive feedback. Supervision is important. Make sure that the interviewers follow up on all designated cases and deaths. Don't let them skip some because they are inconvenient. Make sure that they follow up on answers that are not complete, or questions which aren't answered. Encourage them to probe, to go beyond the questionnaire to find out what really happened. Discipline them to write responses down, especially explanations that cannot be recorded easily on the questionnaire. Tell them to use the local language or dialect, and to record local words and phrases that are used. These sometimes have subtle meanings, and the differences can be important. Before they terminate an interview, encourage them to summarise what they have recorded and ask for verification. Quality control is important, also. The best way to ensure quality is to instil the desire for it in the interviewers. Encourage them to examine their own procedures and to identify ways to improve them. Bring the interviewers together as a group to share experiences and to seek solutions to common problems. Mistakes will still be made, despite all good intentions. In sentinel systems and case investigations, in particular, it is useful to have a second person verify the data. For example, have someone check the entries on the registers and the tabulations. Have a supervisor re-interview five to ten percent of the case investigations, or have the supervisor observe five to ten percent of the case interviews. A sure indicator of problems is a high "unknown" ratio. If more than ten to 20% of the causes of a disease o<%-6>r death are unknown, then it is likely that something is wrong. It could be that the instrument, interview technique, or timing of the interview is causing the bias. Data collection, tabulation and reporting If you plan to collect data through a survey, see Module 2 for specific instructions. Routine and sentinel system data collection will be done daily. You decided in Step 1 the time period and frequency of data collection. You may want to try out your system for a few weeks and then assess whether you need to make any changes. You may also have changed your mind about frequency as you have clarified your procedures. For example, if your CHWs make a routine monthly call to every household, or if they do an annual remapping update, that may be a good time to collect your surveillance data. Normally your staff would tally data daily on a register and then add up the totals at the end of the day or week. Those summaries can then be compiled at the end of the recording period (usually a month). If there are several health centres (or CHWs) reporting the data, you will probably enter the totals from each one on a separate register. Exhibit 6 illustrates one way to do that on a routine register. Sentinel data might be more complex, if additional data on such items as immunization status and other variables are required. But the idea is the same. Just tally the data, compute the totals, and compile the reports from the various units. Case investigation data is likely to be different. In the first place, there are likely to be few investigations in a given reporting period. Second, the amount of data collected is likely to be much greater. Nevertheless, summary reports can be prepared in the same way. Exhibit 6 could be modified, for example, to report the total number of deaths in the period, by sex and age group. A list of causes of death could be substituted for the diseases listed in the left column. Step 6: Analyse the data Analysis should be encouraged at each level of the surveillance system. If CHWs learn to interpret the data they are collecting they will have a better understanding of the needs of their communities. Health centre staff should also be encouraged to analyse their data. Surveillance data can be easily tabulated in three ways: summary tables, disease charts, and maps. Summary tables Most tables will be made up of simple counts of the numbers of cases. Some tables may include percentage distributions, and a few might include averages. Table 4 illustrates a table of counts. This is a summary of the data from Exhibit 6 on measles and diarrhoea. Disease charts You can also construct charts from these tables. Figure 2 is a chart drawn from the data in the table above. Note how the high number of measles cases in two health centres is immediately noticeable in a bar chart. One of the values of charts is that patterns and trends can be seen quickly. Appendix A contains a number of "templates," which are pre-formatted tables on computer files. If you use these, you simply enter the basic data (counts of cases) and the computer automatically computes the totals, percentages (where appropriate), and averages (also, where appropriate). The computer will also construct a graph of your data for you. If you change any figures, the calculations and graphs will change automatically. Summary disease charts are also useful. They are usually made for a 12-month period, as shown in Figure 3. This chart presents data from one health centre. Similar charts can be prepared for all health centres combined. That is, the total number of cases for all ten health centres would be entered each month. Long-term trend data are easier to visualise in a line chart, as illustrated in Figure 4. As will be discussed later, this type of chart is especially useful for identifying seasonal and epidemic patterns. Exhibit 6: Summary of data from ten health centres Excerpt from a modified daily clinical treatment record Programme: Mt. Vernon PHC Date: 30 Nov.,, 1992, +, +, +, +, +, +, +, +, +, +, +, +, + Health centres, +, +, +, +, +, +, +, +, +, +, +, +, + 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, , <%-3>Tot<%0>. , Sex:Male, 36, 25, 42, 65, 33, 43, 53, 32, 44, 47, , 420 , Female, 35, 41, 71, 40, 35, 62, 43, 54, 61, 61, , 494 , Age: Years <<1, 8, 10, 5, 12, 8, 6, 9, 16, 16, 12, , 99 , , 1-4, 64, 39, 73, <%-3>108<%0>, 57, 57, 62, 33, 58, 21, , 572 , 5-14, <%-7>3, 2, 5, 3, 5, 7, 8, 4, 2, 3, , 42 , 15-49, 4, 6, 8, 7, 8, 4, 6, 16, 21, 12, , 91 , Over 49, 2, 4, 3, 6, 5, 4, , 6, 2, 3, , 35 , Routine ANC, 3, 4, 6, 5, 8, 2, 4, 5, 5, 4, , 46 , Routine well-baby, 6, 8, 5, 12, 6, 4, 8, 14, 16, 10, , 89 1, Tuberculosis, , , , , , , , , , , , 2, Polio, , , , , , , , , , 1, , 1 3, Diphtheria/ Pertussis/TT, , , , , , , , , , , , 4, Measles, 1, , 2, , , 1, , , 18, 12, , 34 5, Mumps, , 1, , , , , , , , , , 1 , Other diseases, 2, 6, 4, 3, 8, 10, 7, 9, 2, 4, , 55 6, Malnutrition, , , , , , , , , , , , 7, Diarrhoea/ Dysentery, 2, 6, 14, 2, , 3, 8, 1, 10, 8, , 54 8, Intestinal parasites, , , , , , , , , , , , Table 4: Measles and diarrhoea cases reported, December, 1992 Health centre:, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, Total Measles, 1, , 2, , , 1, , , 18, 12, 34 Diarrhoea, 2, 6, 14, 2, , 3, 8, 1, 10, 8, 54 Total, 3, 6, 16, 2, 0, 4, 8, 4, 28, 20, 88 Maps Mapping is described in detail in Module 3: Work planning. The advantage of a visual presentation of the geographical distribution of cases is that it is easier to identify the location of cases. That can often give you a quick picture of how communicable diseases are spreading. You may see diarrhoea cases clustered in one area, STDs spread along transportation routes and other diseases limited to urban areas. Many PHC programmes now use maps for planning and those same maps can be used to identify households with a disease or health problem. If you plan to monitor several diseases, you can use different coloured pins or symbols to indicate each disease. Figure 2: Measles and diarrhoea cases, December, 1992 Maps are also useful at higher administrative levels. For example, you might show the number of cases of malaria in each village, town, or district. Figure 5, malaria in this example, illustrates how data on the progression of a disease, malaria in this example, might be monitored. Charts and maps are powerful visual aids. But you can usually show only a few diseases on each chart, otherwise they become too cluttered and difficult to understand. Thus, you should be selective in developing charts and maps. You could easily develop 50 charts for just five diseases reported from ten health centres. Figure 3: Measles cases, Mt. Vernon Health Centre, 1992 Figure 4: Measles cases, Mt. Vernon Health Centre, 1990-1992 Exhibit 5: Map of malaria cases Computing rates and ratios Most programme managers will not (or should not) calculate rates and ratios unless their target population is large enough to produce reasonably accurate figures. You should be especially cautious about computing mortality rates. As we mentioned in Step 2, a rough rule of thumb is that you will need to have a population of at least 50,000 to identify enough infant deaths to compute a reasonable infant mortality rate. You would need a population ten times that size to compute a maternal mortality rate. Module 2 shows how to estimate the population size needed to identify a particular "attribute" (such as the proportion of children under age two who are malnourished). The recommendation from Module 2 is to consult an expert for advice on whether you can (or should) calculate rates and ratios. If your population is large enough, then you should compute these rates and ratios. The most typical of these are defined and illustrated below: Incidence rate. The number of new events that occur in a population in a given period of time divided by the total number of persons exposed to risk during that same period; No. new TB cases last month x 10n = 13 x 100,000 = 1.53 Total population 850,000 This is usually expressed as 1.53 per 100,000 population; Attack rate. A cumulative incidence rate that is observed for a limited period, like an epidemic. The formula is the same as the incidence rate formula; Prevalence rate. The total number of all individuals who have an attribute or disease at a particular point (or period) in time divided by the population at risk. The following example is for May, 1992. The rate would be expressed as 156 cases per 100,000 population in May, 1992; Total no. TB cases now x 10n = 1325 x 100,000 = 155.9 Total population 850,000 Appendix B includes appropriate rates for each common disease that you are likely to include in your surveillance system. You can also consult Module 5, which includes lists of morbidity and mortality indicators that you could use. Analysis The purpose of analysis is to: 1) identify patterns, and 2) if possible, causes of diseases or deaths. Look for the following patterns in your data: Spikes: Unusual or sharp increases or decreases in the number of cases. This can indicate an outbreak of a disease, perhaps due to an unexpected general problem, such as the introduction of a new strain of flu. Clusters: Groupings of cases by time period, area, age group, etc. This can indicate an outbreak that is limited to a certain part of the population, such as an outbreak of cholera in a specific part of a city. Trends: Gradual increases or decreases in cases over time. You need to watch trends carefully, since they may occur so gradually that change is not obvious. Examples could be increases in tuberculosis cases, or decreases in anaemia. Systematic variations: Regular changes, such as seasonal variations in diseases. Measles, for example, varies seasonally. A sudden spike might be natural. If you are running a sentinel system, you are likely to have more data, which will enable you to do more detailed analyses. The most typical is a correlation. For example, you might examine disease patterns by age group, by sex, by age and sex, by parity, by site, and so forth. This is the same type of analysis described in Module 2 as "cross-tabulations." If you are planning to use a computer to analyse your data, try the Epi Info programme that comes with Module 2. It was designed for surveillance. Changes in disease patterns may indicate an improvement in health, if the trend is down, or a deterioration, if the trend is up. But the changes could also reflect better or worse surveillance, rather than a real change in health status. Surveillance systems are very sensitive to the number of reported cases. Ideally, all cases would be identified and reported. But if they are not, then your figures will underestimate the actual level of disease. On the other hand, if more cases are reported than actually occurred, then you will overestimate the threat. There are four major factors that can influence the number of cases that are reported. completeness of reporting seasonal variation epidemics or outbreaks coverage Completeness of reporting: Surveillance reports may change for two major reasons. First, people may use the health centre more, and as a result, more diseases would be recorded. Second, health workers may improve their case-finding skills and identify more diseases than before. Thus, an apparent increase in diseases may actually just be an increase in reporting. The opposite can occur as well, of course. If neither of these factors has changed between reporting periods, you can be more confident that a change in the data will reflect a real change in cases. Seasonal variation: Some diseases vary with seasons. Measles, malaria and Guinea worm, for example, vary with the seasons. When coverage, e.g., malaria prevention and control, increases, the seasonal variation will be less noticeable. Seasonal outbreaks vary from one area to another, however, and you may need to chart your own experience for several years to identify seasonal variations. Epidemics: Some diseases reach epidemic levels every few years. Incidence may be relatively steady for several years and then increase dramatically. Measles can be like this. Epidemic patterns also vary from one area to another, and they also have to be charted to identify the patterns. Coverage: As your target population becomes protected from a particular disease or health problem, the incidence should decline. As more women get enrolled in ANC and have safe deliveries, the incidence of childbearing problems should decline. This is the one change that you would hope to see. Other factors: If your target population is changing due to in or out-migration, you could see significant changes in disease patterns. In some central cities slum populations increase 10-15% annually, which can have a significant negative effect on overall health status. Even if coverage is high, diseases can still spread if the health intervention is ineffective. For example, a breakdown in the cold chain can destroy the efficacy of vaccines. You may also see changes in health status due to a general improvement (or deterioration) in socio-economic status, due to an increase (or decrease) in the availability of other health services, or of food, clothing and shelter. Natural disasters, floods, earthquakes, volcanic eruptions, fires can bring about tremendous changes in health. Analysis of differences due to age, sex, education, income and other factors: Although you are not likely to turn your surveillance activities into a research programme, you may want to examine some factors every once in awhile to see if there are any subgroups that are more affected by a particular disease. For example, you might compare disease and mortality patterns of infant boys and girls. If mortality is higher among one group, there is an obvious need to direct more attention to that group. You may also want to examine changes over time. For example, if you compare ARI cases for several years, you may notice a shift in the age groups affected by the disease. Step 7: Investigate causation (optional) You may want to determine the cause of certain diseases, health problems or deaths. As noted previously, this requires special training and in-depth investigation. For most PHC programmes, including those with sentinel surveillance systems, investigation of causation will probably be done on a selective basis. Case and outbreak investigations Appendix C contains a prototype form for investigating a case of neonatal tetanus. You can modify this form to apply to other diseases. You will need one form for each case selected for investigation. Then follow these general guidelines for conducting the investigation: Observe or examine the patient, if possible Interview the mother or other relative who has first-hand knowledge of the case Interview the health worker who examined, observed or treated the case Collect basic descriptive information about the case Confirm the diagnosis Identify symptoms the patient had before and during the illness Determine whether the patient had received appropriate preventive care Identify the treatment the patient received during the illness Identify the outcome of the treatment and illness Analyse the information that is collected to identify: 1) the main cause of the disease or problem; and 2) action that could be taken in the future to prevent a recurrence of the disease. Verbal autopsies Appendix D provides important information and guidelines on ascertaining the cause of death. Appendix E contains two verbal autopsy protocols, one for child, and the other for maternal death investigations. The procedures for using the verbal autopsies are similar, but not identical, to those for case and outbreak investigations. Observe or examine the body, if possible Interview the mother or other relative who has first-hand knowledge of the death Interview the health worker who examined, observed or treated the deceased (prior to, during, or after the death) Collect basic descriptive information about the death circumstances of the death, symptoms and condition of the person at time of death suspected causes (diarrhoea, ARI, injury, etc.) Complete the coding chart Identify the main (most probable) cause of death Identify associated causes Determine whether the patient had received appropriate preventive care Identify the treatment the patient received during the illness As with case/outbreak investigations, the analysis should identify: 1) the main cause of the death; and 2) action that could be taken in the future to prevent this type of death. Verbal autopsy data can be cumulated, as well. That is, the results of several investigations can be summarised, as in Table 5. This example shows the results of verbal autopsies of 61 children. Diarrhoea was the main cause of death for 40, and ARI for 21. Of those who died from diarrhoea, 20 had non-associated causes, 12 had ARI, eight were malnourished, three had measles and ten had a low birthweight. For more information about ascertaining cause of death and analysing the results, see Appendices D and E. Step 8: Develop an action plan Obviously, the whole purpose of surveillance is to find ways to prevent unnecessary disease and death. The data you collect through your system will help you do that. Once you have identified problems, and then determined their causes, the next step is to do something to eliminate those problems. In many cases the actions that are needed will be obvious. For example, if many women are having pregnancy complications because they are not getting adequate care, then they need to be enrolled in ANC. In other cases the solutions aren't going to be obvious. For instance, diarrhoea may continue to be a problem even though most mothers know how to prevent and treat it. In this case you may need to gather more specific information on the cause of the problem before you can recommend a solution. Table 5: Main and associated causes of death of 61 infants Main cause, No. cases, none, ARI, diarrhoea, malnutrition, measles, low birthweight Diarrhoea, 40, 20, 12, , 8, 3, 10 ARI, 21, 11, , 8, 3, 4, 5 Total, 61, 31, 12, 8, 11, 7, 15 In either case, you will need to take action. In addition to the suggestions in this module, Module 6: Service quality, can help you do a more in-depth assessment of problem PHC services. The Problem-solving guide will also provide you with some ideas. If you are still uncertain, this may be a good time to conduct a brainstorming session with your staff, community representatives, and others to identify the root cause of the problem and to search for solutions. If you have a viable solution, or even if you don't, but want to develop one, then outline a plan of action. The plan does not have to be detailed, but it should include: What: The action(s) to be taken should be specified, e.g., provide ANC training to TBAs, or brainstorm what can be done to enrol high-risk women in ANC. Who: The specific people who will be responsible for each action should be identified, by name or position. When: The dates for starting and/or completing the actions. In some cases it may be important to include where, to specify the sites or locations where the actions will take place, how, to outline the procedures that will be followed, and the resources that will be made available to carry out the actions. Specific details may need to be worked out later, and even some of the above elements may have to wait until the formal report is ready and can be studied more carefully. If so, then they should be incorporated into the preliminary action plan. Step 9: Prepare and present reports Most surveillance systems are designed at the central level and require reports to be sent to that level for analysis and interpretation. This module has proposed a local surveillance system, where the reports should address the concerns of local PHC managers. However, the data generated by this system should be useful to other levels as well. If you are going to make formal reports to other levels, then the following guidelines may be of some help to you. Review your original objectives (see Worksheet 1) to make sure that you know the user's most important objectives, questions, issues, and indicators. Make sure that your report addresses them. Review your tables, charts, and maps. Decide which ones will provide the users with the information that most interest them. WORKSHEET 3: DEVELOP ACTION PLANS, +, +, + ACTION TO TAKE (What), RESPONSIBLE (Who), DATES (When), OTHER (Where,, How,, Resources) Identify all 3rd degree malnourished children <<3 years of age, All CHWs and HC staff, June, Via HH visits and routine clinic services Add a short narrative to explain each one. Order them in some logical sequence, such as the following; The number of cases of each disease The number of deaths, by cause An analysis of trends, what they mean Major issues that need to be addressed A summary of the actions that you have (or plan to) take. It is also a good idea to keep a record of your reports, and to supplement your official reports with more frequent (perhaps monthly) internal reports, actions taken, and results observed. Documenting the actions that you have taken to deal with a problem is invaluable information. It can help you þ and others þ avoid mistakes in the future, and most important, replicate those actions that really work. Appendix A: Blank worksheets and analysis templates A.1: PLANNING, REPORTING WORKSHEETS WORKSHEET 1: SPECIFY THE OBJECTIVES OF SURVEILLANCE Purpose(s) Assess needs Monitor trends Explain causes Identify risk factors Evaluate impact on Identify outbreaks Other:___________________________ Identify unusual events Other:___________________________ User(s) Board of directors PHC Manager Community Government officials PHC Staff CDC* Supervisors Other: _______________________ Donors Other: _______________________ * local Communicable Disease Centre Scope Geographic area(s): Programme service(s): Target group(s) Children 1 month Women 15-49 yrs Children 12-23 months Married women 15-49 yrs Children 1-4 yrs Pregnant women Children 5 yrs Other: _________________________ Cases Causes Other Mortality Mortality Specify:________ Morbidity Morbidity Specify:________ Time period: Frequency: WORKSHEET 2: SPECIFY THE SURVEILLANCE PROCEDURES+, +, +, +, + Step 2: Define the data to collect, +, +, Step 3: Select the methods, +, + Target group, Mortality/ Morbidty/ Other, Indicator, Data collection, +, + ^, ^, ^, Frequency, Source, Method* * Routine, Sentinel, Sample survey/special study, Case/outbreak investigation, Verbal autopsy WORKSHEET 3: DEVELOPING ACTION PLANS, +, +, + ACTION TO TAKE (What), RESPONSIBLE (Who), DATES (When), OTHER (Where,, How,, Resources) A.2: TEMPLATES FOR PRODUCING CHARTS This section is made up of a number of pre-formatted charts of various kinds. These charts have been designed to illustrate various ways that you can display different sets of data. In most cases the same data are used for a number of charts so that you can see the differences immediately. Just select the pattern you prefer and draw your own charts. The charts are also on the computer disk that comes with this module, both in Lotus 1-2-3 and Quattro Pro. We call these "templates." You can substitute your own data, titles, legends, etc. into the computer file and then choose the type of display you wish. The name of each chart and computer file are shown in the title for easy reference. To use a template, simply load the appropriate computer file, substitute your own data, titles, and any other information you want to add, and then select the view mode. You can print these charts out, of course, make transparencies of them to use on overhead projectors and make copies for handouts. This is a list of the charts included in this appendix: Template A: One variable counts - one series of data (GRAPH_A.WQ1) A-1BAR A-2LINE A-3ROTATED BAR A-4AREA Template B: One variable counts - two or more series of data (GRAPH_B.WQ1) B-1STACKED BAR B-23D-BAR B-3RIBBON B-4BAR Template C: Frequency distributions - one series of data (GRAPH_C.WQ1) C-1PIE C-2EXPLODED PIE C-3BAR Template D: Frequency distributions - two or more series of data (GRAPH_D.WQ1) D-1COLUMN D-23-D BAR Template E: Correlations - two variables (GRAPH__E.WQ1) E-1XY E-2BUBBLE Template A: One variable counts - one series of data [File: GRAPH_A.WQ1] Months Cases 1 6 2 8 3 11 4 4 5 12 6 8 7 8 9 10 11 12 The two columns on the right are all you need to make a basic graph. The first column is the list of months (you can substitute names: JAN, FEB; or letters: J, F, M). The second column is the data for each month. In this example, these are reported cases of VD each month. Data for the first six months have been entered. Just add data for each month as they become available. You don't have to change anything else. Substitute your own data and press the F10 key to see the graph. Press <> to return to this screen. You can display the same data using different types of graphs. This template has been set up to display ONE VARIABLE using four alternative graphs: LINE, BAR, ROTATED BAR, and AREA. To switch from one graph to another, press /, Graph, Type, and select one of these four types of graph. Then press View to see it. To change the titles, press /, Graph, Text and select the line you want to edit. See your computer manual for more information about making graphs. Reported cases of VD: 1991 Bimanu Health Centre A-1: Bar A-2: Line A-3: Rotated bar A-4: Area Template B: One variable counts - two or more series of data [File: GRAPH_B.WQ1] Very often you will want to display two or more series of data. The following examples show how to do that. The example also illustrates how the þX-axisþ can be changed from a time dimension (months, weeks, years) to places (health centres, provinces, villages, etc.). You can also substitute people (CHWs, children by immunization status, etc.). Thus, these graphs can be used to display data on TIME, PLACES, and PEOPLE, merely by substituting the data and titles. Also note that you can enter your data in the spreadsheet in a horizontal or vertical direction. It doesn't matter. These illustrative data are of reported cases of measles (one variable) from several health centres over a two-year period. Data for 1991 are the FIRST series, those for 1992 are the SECOND series. You can enter six or more series in most graph programmes, however, they can become cluttered and difficult to read if there are more than three or four series. The two new types of graphs on the left are called STACKED BAR and RIBBON. The graph on the upper right is called a 3D-BAR. Also notice the addition of LEGENDS in these graphs. Often, when you have more than one series of data, you need to add legends to distinguish one series from the other. In these examples, the legends distinguish 1991 from 1992. Health centres , North, Tustin, Baro, Selin, Bagio, 1991, 45, 67, 87, 65, 56, Series 2 1992, 82, 98, 123, 88, 77, Series 1 Reported cases of measles: 1991-1992 Routine surveillance: Under age five B-1: Stacked bar B-2: 3D Bar B-3: Ribbon B-4: Bar Template C: Frequency distributions - one series of data [File: GRAPH_C.WQ1] Templates A and B are used to display display counts of variables. They show HOW MANY cases occurred, HOW MANY people died, etc. Frequency distributions show the PERCENTAGE of cases in various categories. For example, the percentage of children who died of malaria, ARI, injuries, and so forth. You can use bar graphs to show percentage distributions as well as counts. And you can also use PIE charts. The following example shows how to use pie graphs to display frequency distributions. It also shows one way to display data about PEOPLE, in this case, the percentage of neonatal deaths by cause. Data for C-1: Pie graph and C-2: Exploded pie graph; Main causes of neonatal deaths Main cause Number Prematurity/trauma 432 Tetanus 255 ARI/Pertussis 123 Diarrhoeal diseases 67 Others 221 Total 877 Pregnancy complications Age Number Percent 15-19 14 10.3% 20-24 26 19.1% 25-29 28 20.6% 30-34 34 25.0% 35-39 22 16.2% 40-44 12 8.8% Total 136 100.0% These percentages would be entered as Series 1 of your graph. You can then select other types of graphs, as explained in Template A. You don't need to compute the percentages yourself with these kinds of graphs, the computer program does it for you. Pie graphs have a unique feature that you may want to use. They can be "exploded." One or more pieces of the pie can be partially moved for emphasis, as shown in C-2. The graphs on the next page show several other ways to display frequency distributions. One way is to compute the frequency percentages yourself and then enter them into one of the graphs shown in Templates A or B (line, area, bar, stacked bar, rotated bar, ribbon). If you want the percentages to be displayed, you can do this by selecting the Interior labels option (press /, Graph, Customize series, Interior labels, select the series to be displayed, <>, then select where you want to place the labels (top, right, etc.). Main causes of neonatal deaths 877 infants, 1986 C-1: Pie C-2: Exploded Pie Pregnancy complications by age group: 136 women, 1992 C-3: Bar Template D: Frequency distributions - two or more series of data [File: GRAPH_D.WQ1] You can also use a COLUMN graph to display frequency distributions. Note that this graph displays the name of the disease as well as its percentage, just as the PIE graphs do. You can compare data from different sites, or different periods of time, or different types of people by entering each one as a separate series. You need to save each series as a separately NAMED graph. In this example the first graph is named HC#1, and the second is HC#2. To display (or print) one of the graphs, press /, Graph, Name, and select the graph you want, then press View to see it. Data for D1-2 for column graphs Sentinel centres, +, + , HC#1, HC#2 Malaria, 154, 231 STD, 198, 256 ARI, 221, 316 Diarrhoea, 422, 366 Other, 677, 743 Data for D3 column graph Sentinel centres, +, +, +, +, +, + , HC #1, , HC #2, , HC #3, Number, Percentage, Number, Percentage, Number, Percentage Malaria, 154, 9.2%, 231, 12.1%, 186, 10.6% STD, 198, 11.8%, 256, 13.4%, 245, 13.9% ARI, 221, 13.2%, 316, 16.5%, 267, 15.1% Diarrhoea, 422, 25.2%, 366, 19.1%, 321, 18.2% Other, 677, 40.5%, 743, 38.9%, 744, 42.2% Total, 1672, 100.0%, 1912, 100.0%, 1763, 100.0% You can also compute the percentages yourself and enter these into a graph, just as in Template D. The following example shows data from three sentinel surveillance centres. Disease report 1991: Sentinel centres Sentinel centre 1 D-1: Column Sentinel centre 2 D-2: Column Sentinel centres 1-3 D-3: Columns Template E: Correlations - Two Variables[File: GRAPH_E.WQ1] You can show the relationship between two series of data, such as age and a disease. You do this by listing two blocks of data and then identifying the spot in the graph where the values of the two variables intersect. The type of graph that is produced is often called an XY graph, because you plot one value (say age) on the X axis and the other value (say the number of ill people) on the Y axis. You can either plot a large number of data points, which produces a SCATTER diagram, or group the data (say into five-year age groups) and connect the data points with lines. If you have enough data, you can also do statistical tests (usually regressions) to see if there is a statistical relationship between the two variables. In general, does one increase or decrease as the other changes? One such graph is shown below. It is based on the first two variables and shows the number of children in three-month age groups who are malnourished. The slope of the line shows that malnutrition increased with age up to 18 months and then declined sharply. Age (months) 6 9 12 15 18 21 24 ARI 18 23 36 33 42 28 21 Severity 1.30 1.80 2.20 2.30 2.40 1.20 0.80 Malnourished cases by age of child, June-July, 1990 ARI cases by age of child, June-July, 1990 The next graph, a BUBBLE graph, lets you add a third variable. In this example, that is SEVERITY. Malnutrition is usually measured in degrees (1, 2, 3), with 3 being severly malnourished. By taking the average severity of malnourishment in each age group, we can show where the problem was greatest, since the size of the bubble indicates the severity. The graph shows that the severity of malnutrition increased as the number of cases increased. Then it began to decline. Consult your computer manual if you are interested in learning more about displaying your data. The computer programs are very versatile now, and very "user-friendly." With a little effort, you, or one of your staff, can learn how to make effective presentations fairly rapidly and easily. Appendix B: Common diseases,definitions and indicators This appendix is a listing of some common diseases that many PHC programmes deal with regularly. Some of these are specific to certain regions of the world and may or may not be problems in your area. Many more could be added, and you are encouraged to identify those that are important to you but which are not listed. You may find it helpful to develop a summary like those listed in this appendix. That way everyone will have a common understanding of definitions and indicators to use. The diseases included in this appendix are listed below. Common priority diseases and health problems for surveillance in developing countries Acute respiratory infections HIV/AIDS Tetanus Acute watery diarrhoea Measles Trachoma Cholera Neonatal tetanus Tuberculosis Diphtheria Pertussis Urinary schistosomiasis Dysentery Poliomyelitis The terminology used in this appendix is defined below: Standard case definition: Clinical definition of the disease. Lay definition: Common, non-clinical definition, using clear terms to identify symptoms. Incubation: The time interval between initial contact with an infectious agent and the first symptom of the disease. Indicator(s): An indirect measure of a phenomenon that cannot be easily measured directly (for example, weight-for-age as an indirect measure of health/nutritional status). Background indicator: Information that provides a context against which morbidity and mortality will be interpreted. Estimation: The type of measure used to estimate the incidence or prevalence of a disease. Rate: The frequency of occurrence of some event, such as kilometres/hour, cases/month. Data sources: Typical sources of data for these indicators. Used for: Typical management uses, such as for needs assessment, baseline and follow-up assessments, etc. Mortality ascertainment via interview: Degree to which the indicator can be used to determine the cause of death. Value and limitations: Utility of the indicator with respect to monitoring, evaluation, etc. Limitations, such as cost, validity, etc. Comments: Relevant comments on the effect of the diseases characteristics on measurement; other factors that affect the indicator. ACUTE RESPIRATORY INFECTIONS (ARI) Standard case definition: Fast breathing (60 per minute or more if child less than two months; 50 per minute or more if the child is two months up to 12 months; 40 per minute or more if child 12 months up to five years) is a sign of pneumonia. The severity of the pneumonia is judged on the presence of chest indrawing (the lower chest wall draws in when the child breathes in). Look for non-specific signs of pneumonia, sepsis or meningitis: the child stops feeding well, is abnormally sleepy or difficult to wake, has fever or low body temperature (35.5 C) or has convulsions. Lay definition: þPneumoniaþ (defined as fast breathing) and þchangeþ in normal behaviour. Incubation: Varies according to the agent. Often unknown. Indicator(s): Number of children diagnosed with pneumonia. Background indicator: Currently not identified. Estimation: Incidence by age groups; Case fatality rate. Data sources: Hospital and health centres registers; health care facilities and household surveys. Used for: Needs assessment and monitoring of ARI control activities. Mortality ascertainment via interview: High. Value and limitations: Not all cases are brought to curative facilities. However, after an initial training of the health staff and of community health workers in identifying the symptoms of pneumonia and of severe pneumonia, the quality of reporting should improve significantly. Comments: Currently, ARI is probably the leading cause of death among young children in less developed countries. Surveillance of pneumonia is an essential step in its recognition as a priority and in monitoring progress in its control. ACUTE WATERY DIARRHOEA Standard case definition: Frequent loose and watery stools often associated with fever and vomiting. Lay definition: Watery diarrhoea as defined by the mother. Incubation: Variable depending upon the agent responsible. Generally between eight to 18 hours. Indicator(s): Number of reported cases by age groups. Background indicator: Percentage of households: a) using a latrine; b) having access to safe water. Estimation: Incidence of acute watery diarrhoea; number of episodes per child per year. Data sources: Hospital and health centre registers. Health centres and household surveys. Used for: Needs assessment. Monitoring of diarrhoeal diseases control activities (as a denominator of the estimation of case management). Outbreak investigation. Mortality ascertainment via interview: High. Value and limitations: Most clinical cases of diarrhoea go unreported, except in settings having extensive monitoring of decentralized ORS packets delivery. Incidence has seasonal variations which has implications for the timing of surveys. Comments: Although case identification is necessary, WHO stresses the importance of adequate case management. CHOLERA Standard case definition: Sudden and severe watery diarrhoea with rapid dehydration. Lay definition: Many local terms often meaning þdeadly diarrhoea.þ Incubation: From a few hours to five days; usually two-three days. Indicator(s): 1. Number of suspected cases and deaths; 2. Number of laboratory-confirmed cases. Background indicator: Occurrence of outbreaks in the past. Estimation: Attack rate by age group and gender. Data sources: Outbreak investigation in non-endemic area; hospital and health centres registers in endemic areas. Used for: Outbreak investigation. Needs assessment. Mortality ascertainment via interview: High. Value and limitations: Assessment of the quality of food-handling and water and sanitation in non-endemic areas; assessment of health education (ORT) and case management in endemic areas. Comments: Reporting of cholera cases is often considered as sensitive information by MOH, because of the implication for exportation of edibles and on tourism. DIPHTHERIA Standard case definition: Acute pharyngitis, acute nasopharyngitis, or acute laryngitis, with a pseudo membrane. Lay definition: Sore throat, with grey patch or patches in the throat. Incubation: Usually 2-5 days, occasionally longer. Indicator(s): Number of cases. Background indicator: Vaccination coverage DPT3 by 12 months of age. Estimation: Incidence not predictable. Data sources: Health centre registers and, occasionally, from outbreak investigation. Used for: Not used. Mortality ascertainment via interview: Low. Value and limitations: Low. Comments: Often occur as outbreaks when social or natural conditions lead to crowding of susceptible children. DYSENTERY Standard case definition: Diarrhoea with blood, mucus and pus accompanied with fever, nausea and sometimes vomiting. Lay definition: Diarrhoea with blood and mucus. Incubation: 1-3 days on average. Indicator(s): Number of cases by month and by age groups. Background indicator: Water and sanitation status. Estimation: Incidence. Data sources: Hospital and health centre registers. Health centres and household surveys. Used for: Needs assessment. Monitoring of diarrhoeal disease control activities (as a denominator of the estimation of case management). Outbreak investigation. Mortality ascertainment via interview: High. Some types of Shigella (the agent) have a case fatality rate among hospitalized patients as high as 20%. Value and limitations: Generally under reported. Seasonal variations. Comments: In addition to continued feeding and prevention of dehydration, dysentery cases must receive appropriate antibiotic treatment. HIV/AIDS Standard case definition: For epidemiological surveillance an adult (above 12 years) is considered to have AIDS if: A test for HIV antibody gives positive results AND one or more of the following are present: >>10% body weight loss or cachexia, with diarrhoea or fever, or both, intermittent or constant, for at least one month, not known to be due to a condition unrelated to HIV infection. Tuberculosis with the combination of weight loss plus fever or plus diarrhoea as described above. Tuberculosis that is disseminated (involving at least two different organs) or miliary; or extra-pulmonary tuberculosis (which may be presumptively diagnosed). Kaposi's sarcoma. Neurological impairment sufficient to prevent independent daily activities, not known to be due to a condition unrelated to HIV infection (for example, trauma). Candidiasis of the oesophagus. Lay definition: Local terminology. In many countries of Africa, the term þslimþ is used. Incubation: One to three months between exposure and sero-positivity; up to ten years between HIV-1 infection and the development of AIDS. Indicator(s): Number of cases by age group. Background indicator: Prevalence of sero-positivity. Estimation: Incidence and Prevalence. Data sources: Hospital and health centre registers; TB control unit surveillance reports; laboratory registers. Used for: Need assessment. Monitoring of control activities. Mortality ascertainment via interview: High. Value and limitation: The case definition: is simple to use; includes extrapulmonary and pulmonary tuberculosis; and has high specificity. However it has a low sensitivity. Comments: Screening of blood donors in hospitals is often an initial step in surveillance to document the existence of sero-positive individuals in the community and discuss further steps. MEASLES Standard case definition: History of a generalized maculo-papular rash lasting three or more days and history of any one of the following: cough, coryza, conjunctivitis. Lay definition: History of fever and rash and any one of the following: cough, running nose, red eyes. Incubation: About ten days, varying from seven to 17 days from exposure to onset of fever, usually 14 days until rash appears. Indicator(s): 1. Number of cases of measles a) total b) by age: less than nine months between 9-23 months more than two years 2. Percentage of cases of measles with a documented valid vaccination against measles, by age group. Background indicator: Percentage of infants having a valid measles vaccination by 12 months of age. Estimation: Depending upon the time of the year, the vaccination coverage and the population density. Each country and district differs. As an example, between 1891-1991, the rate for the countries of the WHO South East Region ranged between 29.2 and 10.54 per 100,000 population. The expected number of cases among children in your area can be estimated, using the formula: a x b x .85 x .90 a= number of surviving children b= coverage in the age group c= vaccine efficacy: 85% d= incidence rate: 90% Data sources: Routine and sentinel reporting; outbreak investigation. Used for: Assessing effectiveness EPI and quality of MCH. Must be included in needs assessment, baseline and periodic evaluation. Mortality ascertainment via interview: High. Value and limitations: Useful to assess the quality of protection to high-risk groups, particularly among refugees. Comments: Measles is a seasonal disease each year with an epidemic occurring every two to three years depending upon local conditions. Increased vaccination coverage rates lowers the magnitude of yearly seasonal peaks and widen the epidemic intervals. NEONATAL TETANUS Standard case definition: History of normal suck and cry for the first two days of life, AND history of onset of illness between three and 28 days of age AND history of inability to suck followed by stiffness and/or þconvulsionsþ and death (80% of cases). Lay definition: History of normal suck and cry first two days of life, AND history of inability to suck, convulsions and/or stiffness between three and 28 days of age. In several cultures, history of the baby turning þblue.þ Incubation: Between three to 21 days with a range between one day and several months. In the majority of cases incubation is about ten days. Indicator(s): 1) total number of cases of NNT: 2) number of NNT cases according to the immunization status of their mothers 3) percentage of NNT cases delivered by a trained attendant. Background indicator: a) percentage of infants found protected against NNT at the time of receiving DPT 1; b) TT vaccination coverage of women of child bearing age. Estimation: To calculate the expected number of cases: a x b x .8 x .01 a = number of newborns b = coverage (at the time of DPT 1) c = vaccine efficacy: 80% d = attack rate: 10 per 1000 live births (1%) Data sources: Vital registration and/or lay TBA reporting;Vaccination records; MCH records. Used for: Needs assessment and PHC overall monitoring. Mortality ascertainment via interview: High by the use of verbal autopsy technique. Value and limitations: NNT rate is a good predictive indicator of IMR and of MMR. Comments: Neonatal tetanus cases have a tendency to cluster. Each case should be investigated. PERTUSSIS Standard case definition: History of severe cough and history of any one of the following: cough persistent two or more weeks, fits of coughing, cough followed by vomiting. Lay definition: History or observation of repeated and violent coughing and history or observation of any one of the following: cough persisting two or more weeks, fits of coughing, cough followed by vomiting, typical þwhoopþ in older infants and children. Incubation: Commonly seven to 10 days, rarely exceeding 14 days. Indicator(s): Number of cases by age-groups. Background indicator: Vaccination coverage DPT 3 by 12 months of age. Estimation: The expected number of cases can be estimated by the formula: a x b x .8 x .8 a = number of surviving children b = coverage in the age group c = vaccine efficacy: 80% d = incidence rate: 80% Data sources: Hospital and clinic records. Used for: Pertussis is generally not considered a high priority for surveillance. Mortality ascertainment via interview: Medium; case fatality rate is usually 1%. Value and limitations: Unless the staff has been trained in the differential diagnosis of pertussis (ARI), case ascertainment can be difficult. However pertussis surveillance might be of interest in the monitoring of active ARI control programmes and because of its importance (like measles) in worsening malnutrition. Comments: Should be included in the list of reportable diseases in refugee situations. POLIOMYELITIS Standard case definition: Any patient with acute flaccid paralysis (including any child less than 15 years of age diagnosed to have Guillain Barr‚ Syndrome) for which no other cause can be identified. Lay definition: History of sudden onset of weakness and paralysis of the leg(s), and/or arm(s) and/or trunk, AND documentation that paralysis was not present at birth or associated with serious injury or mental retardation. Incubation: Commonly between seven to 14 days for paralytic cases. Indicator(s): 1. Number of cases of acute flaccid paralysis with date of onset and age at onset, vaccination status. 2.Number of suspected cases of polio. 3.Number of confirmed cases of polio. Background indicator: Vaccination coverage OPV 3 by 12 months of age. Estimation: From zero polio in the Americas and 1814 cases reported in 1991 in the WHO Africa region to 6020 reported for India alone in 1991. Locally, the expected number of cases among children can be estimated: a x b x c x d with: a = number of surviving children b = coverage (use local figure if available) c = vaccine efficacy: 90% for OPV 3 d = attack rate: 0.6% (use local rate if available and recent). Data sources: Weekly or monthly surveillance reports; laboratory reports. Used for: Assessing the effectiveness of polio eradication activities and the quality of surveillance. Mortality ascertainment via interview: Low. The number of deaths can be guessed by assuming a case fatality rate of five percent. Value and limitations: Each new case should be considered on its own merits as a challenge to the health team. Comments: A must in the current context of polio eradication efforts. TETANUS (besides neonatal tetanus) Standard case definition: A case of tetanus occurs when a person has: (1) a stiff jaw and trouble opening the mouth or swallowing; (2) painful stiffness of the neck and abdominal muscles (often other body muscles get stiff too); (3) a clear mind; (4) a wound, often infected, or history of a wound within the past few weeks. In severe cases, the person may appear to be smiling (risus sardonicus) with raised eyebrows. His back and neck may be arched, his arms bent with fists clenched at his chest, and his legs extended. Noise, light or touching the person may trigger sudden, painfully tightening of the muscles (convulsions). Lay definition: History of injury or ear infection followed by difficulty in opening mouth (or jerking of the mouth) or stiffness of the neck or body. Incubation: Between three to 21 days with a range between one day and several months. In the majority of cases incubation is about ten days. Indicator(s): Number of cases by age group. Background indicator: None for male adults as vaccination status is generally unknown. For females of child-bearing age: TT 2 or TT 3 vaccination coverage. Estimation: Currently difficult as surveillance is neglected. Data sources: Hospital or health centre registers; in some places, lay reporting of vital events. Used for: Initial needs assessment. Measuring rate of complications of associated diseases (like Guinea worm, for example) or other hazards (contaminated street drugs in urban settings, for example). Mortality ascertainment via interview: High by verbal autopsy techniques. Among males, search for a history of an open wound; among females, tetanus is often a complication of septic abortion or delivery. Value and limitations: Useful for the monitoring of MMR and of adult health for occupational hazards (agriculture). Comments: Generally underestimated. TRACHOMA Standard case definition: Chronic inflammation of the eyes, leading to shrinkage and turning-in of lids and blindness. Lay definition: Blinding disease. Incubation: Five to 12 days. Indicator(s): Diagnosis of lymphoid follicles and hypervascularisation of the cornea, particularly on the tarsal conjonctiva lining the upper eyelid; Background indicator: Prevalence of corneal scars and blindness among adults. Estimation: Prevalence among children. Data sources: Periodic surveys; in areas where the health staff has been trained, from health registers. Used for: Needs assessment and monitoring of intervention. Mortality ascertainment via interview: N.A. Value and limitations: Good for a given ecological area. Comments: Trachoma is a disease of poverty, unhygienic surrounding (flies and scarce water resources) and dust (Sahel particularly). Trachoma is rarely a primary cause of consultation. TUBERCULOSIS Standard case definition: An ill child with a history of contact with a suspect or confirmed case of pulmonary tuberculosis. Any child who does not return to normal health after measles or whooping cough; with loss of weight, cough and wheeze who does not respond to antibiotic therapy for acute respiratory disease; with abdominal swelling with a hard painless mass and free fluid; with painful form or soft swelling in a group of superficial lymph nodes; with any bone or joint lesion of slow onset; with signs suggesting meningitis or disease in the central nervous system. In adult, confirmed isolation of mycobacterium bacilli (acid-fast bacilli or AFB) in sputum; and/ or typical pulmonary lesions in X-rays. Lay definition: An ill child with a history of contact with a suspected or confirmed case of pulmonary tuberculosis. Any ill child with one of the following: Who does not return to normal health after measles or whooping cough. With loss of weight, cough and wheeze who does not respond to antibiotic therapy for acute respiratory disease. An adult with history of persistent cough not responding to antibiotic therapy and a loss of weight during the previous months and low-grade fever and/or sweating at night. Incubation: From infection to demonstrable primary lesion or tuberculin reaction, about four to 12 weeks. While the risk of pulmonary or extrapulmonary reaction is greater in the following two years, the infection can remain latent throughout life. Indicator(s): 1. Number of suspected cases (on clinical basis) by age group. 2. Number of cases bacteriologically and/or X-ray confirmed by age group. Background indicator: Vaccination coverage for BCG. Estimation: Incidence rates not predictable. Data sources: Registers from health centres and/or from TB clinics; microbiology laboratory registers. Used for: Needs assessment. Measuring the quality of reporting systems. Mortality ascertainment via interview: Difficult, unless the deceased was a known patient who did not take his/her treatment regularly. Value and limitations: Recently, the incidence of tuberculosis has been increasing worldwide for two reasons: a) the HIV/AIDS pandemic which increases the rate of secondary infections; b) increased resistance to known drugs. The monitoring of the incidence of tuberculosis might be of increasing value as a marker of AIDS transmission. Comments: A disease often neglected by surveillance systems in the past, tuberculosis (particularly pulmonary tuberculosis) is be included in the future because of new interest in respiratory infections and because of its growing toll on health. URINARY SCHISTOSOMIASIS Standard case definition: Clinical symptoms (frequent and uncomfortable urination with blood appearing in urine at the end) associated with presence of eggs of Schistosoma Haematobium in the urine. Lay definition: Blood in urine. Incubation: Two to six weeks. Indicator(s): Number of cases of children six to 15 years having haematuria (blood in urine). Background indicator: Percentage of children six to 15 years bathing under safe conditions; percentage of households having access to safe water. Estimation: Incidence and prevalence. Data sources: Health centre registers; periodic school surveys. Used for: Needs assessment. Monitoring of control activities. Mortality ascertainment via interview: Low. Urinary schistosomiasis is a chronic disease of the adult with a differential mortality. Value and limitations: The presence of haematuria (assessed by dip-sticks) has been found an acceptable test to confirm the diagnosis of S. haematobium. Re-infection occurs after treatment. Comments: Surveillance is generally focused on school-attending children. Appendix C: Case investigation forms THE EXAMPLE OF NEONATAL TETANUS 1 WHO/EPI/MLM/9.14 Training for mid-level managers; disease surveillance, revised 1991> appendix C Neonatal tetanus has been targeted for control by most national health programmes. Neonatal tetanus incidence is a reliable indicator of the quality and effectiveness of two important PHC activities: EPI and MCH. Each reported case of neonatal tetanus becomes a challenge to the PHC manager since it points to a failure of the programme; insufficient TT coverage, cold chain breakdown, poor TBA training or unhygienic delivery practices or handling of the umbilical cord. PHC managers should keep in mind the fact that NNT cases are often clustered. Health centre supervisors should use a case investigation form to investigate all cases of neonatal tetanus. The purpose of investigating neonatal tetanus cases is to identify why the case occurred so that future cases can be prevented. When investigating a case of neonatal tetanus, ask the mother of the infant if she is willing to answer some questions about her infant's illness. Explain that the information she provides will help you prevent future case. With the case investigation form in front of you, ask her the questions listed on the form, and carefully record her responses. The questions are about: The immunization status of the mother. Whether the mother received antenatal care. Where the baby was born. Whether a trained attendant was present at the time of the actual delivery. How the cord was cut and treated (was any substance smeared on it). Whether the infant sucked normally at birth, then later developed problems with sucking, convulsions and stiffness. Whether the infant was treated in a hospital for the illness and where. If the mother is unable to answer, ask a close female relative who attended the birth or saw the infant during his first weeks of life. On the following pages you will find a sample of a neonatal tetanus case investigation form. PHC supervisors should monitor to make sure neonatal tetanus cases are investigated promptly and correctly and help with the investigations if needed. For example, a trained officer may conduct some case investigations himself or provide additional training if health workers are uncertain of the procedures to follow. Follow these steps when analysing neonatal tetanus caseinvestigation forms: 1. Confirm the diagnosis of each case. Review the form to make sure that each reported case of neonatal tetanus fulfilled the appropriate case definition of the disease. 2. Determine the immunization status of the mother of each neonatal tetanus case. Was the mother immunized? If not, why not? Was there a missed opportunity? 3. Look at the geographic distribution of cases by reviewing the disease map. This will tell you whether cases of neonatal tetanus are located in certain geographic areas or if they are widespread. 4. Look for similarities among cases. You may see a pattern that will help you identify solutions. For example, did all mothers of neonatal tetanus cases go for immunization to the same health worker who does not immunize women at every opportunity? Did they use a common birth attendant or did they deliver their child at the same hospital? Neonatal tetanus case investigation (continued) Determining the cause of death NEONATAL TETANUS CASE INVESTIGATION FORM Instructions: Either circle the appropriate response, or record the answer in the blank space following each item. Province: Village: Health centre: Date case reported: Reported by: Date of investigation: Household address of case: Name and job title of investigator: A. CASE INFORMATION Family name: Given name: Date of birth: Sex: Male/Female How long has the mother been resident in the area? Ethnic group (if applicable): B. IMMUNIZATION STATUS OF MOTHER AND ANTENATAL CARE Was the mother immunized against tetanus)? Yes/No If yes, number of doses: date of last dose Did you see an immunization record? Yes/No Did the mother receive antenatal care? Yes/No If yes, location and dates: Did the mother visit a health facility for reasons other than antenatal care during this pregnancy? Yes/No If yes, give the reason for visit and dates: C. BIRTH OF INFANT Where was the baby born? Hospital/home/other:________________ If the delivery was in an institution, give the name and address of the institution: Was the delivery attended by: doctor/traditional birth attendant/nurse/midwife/other: If attended by a traditional birth attendant, give the name and address: Describe how the cord was cut and with what type of equipment: Describe how the cord stump was treated or dressed: D. SYMPTOMS Date of onset of illness: Did the child suck and cry normally for the first two days of life? Yes/No If no, describe: Did the child later have a problem with sucking? Yes/No If yes, describe: Was the child stiff later? Yes/No Did the child later have convulsions? Yes/No/Unknown Other complications: E. TREATMENT/OUTCOME Was the patient cared for in health facility? Yes/No If yes, where? Was the patient seen by a health worker? Yes/No If yes, give diagnosis: Did the patient die? Yes/No If yes, give date and details: Comments: F. TO BE ANSWERED BY HEALTH FACILITY STAFF: Is this a case of neonatal tetanus? Yes/No If yes, could this case have been prevented? Yes/No If yes, describe: What actions should be taken to prevent similar cases in the future? Appendix D: Determining the cause of death Assessing the probable leading causes of death allows managers to document the effectiveness of health interventions, and to identify changes that they can make in procedures to reduce unnecessary mortality. However, causation is not always easy to determine. This appendix describes some of the common causes of death among young children and among women during pregnancy and childbirth. It discusses the ease or difficulty of measuring causation accurately. Identifiable causes of death The list of common causes of death in children is small, generally limited to diarrhoea, injury, neonatal tetanus, measles, ARI, malnutrition, and possibly such other neonatal conditions as sepsis. The following table lists these conditions. Those nearest the top are easier to verify through an interview with the mother than those near the bottom. Table D-1: Childhood causes of death that can be determined through interview Childhood mortality Relatively easy Injury Neonatal tetanus Measles Diarrhoea Acute respiratory infection Malnutrition Relatively difficultNeonatal sepsis Note that death due to injury or trauma tops the list. Drowning or falling off a cliff is seldom confused with other common causes of child death. Likewise the neonatal tetanus picture is quite distinct: an initially healthy baby develops stiffness/convulsions and an inability to suck dying on or after day three. Similarly the cough, fever, red eyes, and rash of measles is recognizable by family members. Diarrhoea and ARI are often more difficult to identify in different settings. Malnutrition is likely to be recognized if acute with marasmus, kwashiorkor, or marasmic kwashiorkor while chronic malnutrition, i.e., stunting, or milder forms will be missed. Many other conditions are important of which neonatal ones are among the most challenging. Indeed, all but the most obvious causes of neonatal death are notoriously difficult to establish. þUnknownþ rates for newborns are almost always higher than for older children. In neonates: Birth defects, if obvious, will be recallable if there is no social taboo. Major internal abnormalities, such as heart defects, of course will go undetected. Tiny babies will be reported, but in areas where small babies are common or even desired, they will go unnoted. Of great programmatic interest are newborns dying of birth trauma or birth asphyxia. While theoretically distinct, there are many hurdles to identifying these babies, especially the reluctance of mothers and birth attendants to report poor delivery practices. Newborns who die generally have many indistinguishable signs regardless of underlying cause: grey colour, irritability/sleepiness, poor feeding, abnormal breathing, vomiting and so on. In addition to expanding the scope of ascertainment to the newborn, researchers are attempting to break down syndromes into meaningful distinct subsets. For example, diarrhoea syndromes probably can be divided into acute watery diarrhoea, persistent diarrhoea, and dysentery. This is important because the programmatic response varies for each. Likewise, pertussis (an EPI preventable disease) may be characteristic enough to be distinguishable from other fatal ARI conditions. In older children: Validly identifying common childhood killers, notably malaria and AIDS, poses additional challenges. Table D-2 lists some of these conditions by age and the likelihood that they could be diagnosed by interview. Two points from Table D-2 need amplification. "Local ecology" (see malaria and AIDS) refers to the prevalence of the condition in the manager's setting. For example, if malaria is common, the general public, including the informant, is more likely to be able to recognize it. And malaria is statistically more likely to be the correct diagnosis than in areas where it is unusual. Related is the notion that specific programmes may have unusual killers that are not on this list. For example, visceral leishmaniasis is said to account for the largest PMR of child death in one north Pakistan valley. It is likely that it could be identified by interview although one will not find much expert guidance. Table D-2: Other causes of childhood death ascertainable by interview by age with likely validity* Condition, 0-1 Month, 1 Month-5 Yrs, Comments Failure to feed, +++, , e.g.,, after maternal death. Prematurity, ++, ++, LBW is a childhood risk for 1-2 years. Small for gestational age (SGA), ++, ++, Although most LBW babies in developing countries are probably SGA,, the distinction is important as programme response differs. Hypothermia, ++, , More likely with LBW baby or in extreme environmental cold. Milk aspiration, ++, , Not common. Sepsis, ++, ++, Non-specific presentation,, especially in neonates. Hyaline membrane disease, ++, , Non-specific even in prematures. Tetanus, +++, ++, Readily diagnosable in newborns; uncommon thereafter. Meconium aspiration, ++, , Difficult to distinguish from asphyxia ,, but programme response is similar for both. Dysentery, +++, ++, Important to distinguish as emphasis of programme response is different in each case: antibiotics vs. oral rehydration therapy vs. nutritional therapy. Acute watery diarrhoea (AWD), +++, ++, Different in each case: antibiotics vs. oral rehydration therapy vs. institutionaal therapy. Persistent diarrhoea, +++, ++, Pertussis, , ++, EPI significance; the þwhoopþ may be highly specific,, but prolonged coughing is not (asthma,, TB,, foreign body,, para-pertussis,, etc.). Cerebral/CNS infection, ++, ++, Very non-specific in newborns; children may have meningitis,, TB,, cerebral malaria,, encephalitis. Malaria, , ++, Local ecology imperative. AIDS, , +/++, Local ecology imperative. * + = low validity; ++ = medium validity +++ = high validity , +, +, + In women of child-bearing age: Women in the reproductive years make up the other main target group of PHC. Many, perhaps 25-50% of all deaths in this age, are considered þmaternal mortality,þ ie., related to pregnancy, labour, delivery, or the first six weeks after termination of pregnancy. Most maternal mortality falls into five syndromes which can be recognized through a survey interview instrument (Table D-3). Of course, due to the social implications, abortion-related deaths are severely under-recognized. Table D-3: Maternal causes of death that can be determined through interview Maternal mortality Relatively easy Post-partum haemorrhage Obstructed labour Eclampsia Post-partum infection Relatively difficult Abortion-related Reporting Deaths Step 4 in the main text described several ways that deaths can be identified, recorded and reported. The two major ones are: 1) a death report (see Exhibit 4), which extracts information on deaths from hospital, clinic, and provider records, and 2) an outreach register (see Exhibit 3). The usefulness of medical records is limited even when they exist and when they can be located. Busy medical staff appropriately spend more time attending to desperately ill patients than they do writing in charts. Outreach workers do not usually have first-hand knowledge of the death and can only get a rough idea of the cause of death. As a result, you probably will not find the detail that you would need to identify cause of death in these reports and records. In addition, you also have to be careful to ensure that the record consulted is, indeed, that of the deceased and that the deceased is from your catchment area. Verbal autopsies As noted before, in the abscence of a formal autopsy, the best way to establish cause of death is through a verbal autopsy (VA). This concept has been broadly introduced in previous sections and a step-by-step approach is proposed in Appendix E. The technique consists of þverbally,þ i.e., with an interview, conducting an þautopsy,þ i.e., ascertaining the causes of death. There is usually no better way because the death may have been undocumented and in the (possibly remote) past. How to process the data Assign a cause of death Once deaths have been reported and a VA conducted, a cause must be assigned and coded before the information can be tabulated. Assigning a diagnosis can be done in two different ways. Some use clinical judgement of one or more experts, usually paediatricians and obstetricians. Ideally, at least two should reach agreement by consensus. The other approach does not rely on the presence of an expert panel. It uses objective scoring systems for each condition, various signs are sought and are pre-classified as either major or minor, different combinations of which make the diagnosis. For example, a probable ARI death requires the presence of cough for more than two days, rapid breathing for more than one day plus one of the following four: chest wall indrawing, nostril flaring, grunting on breathing out or blue lips/tongue. An added complexity is the probabilistic weighting added to assigning the diagnosis. This can be omitted; however, it adds little work and is quickly grasped. The rationale is that some combinations of clinical information are likely to have a greater positive predictive value than others for the condition in question. In the ARI example, cough and rapid breathing lead to a possible diagnosis whereas cough and rapid breathing AND any one of the additional four signs of respiratory difficulty lead to a probable diagnosis. Note that in the field a certain diagnosis is never possible. Assigning cause of death, even using these scoring systems, requires additional skills. This should be the responsibility of the interviewer's supervisor. Occasionally the verbatim history will clarify þDKþ (don't know) responses enough to allow assigning a possible or even probable cause of death. You may decide to re-interview households if your unknown ratio is unacceptable, say, above ten to 20%. Coding the cause of death Coding the diagnosis is the next step, having made the diagnosis or diagnoses for the case. International convention uses terminology such as þimmediate, underlying, and antecedent cause,þ and þother significant conditions.þ These reflect a biomedical, Western bias where patients are more likely to die in a hospital fully documented. These terms have little relevance to the developing world where people die from multiple causes often without a clear primary condition and with no documentation. Thus, a different coding system using two types of diagnosis is proposed: main and associated. The main diagnosis is the one, in the opinion of the coder, which is likely to be the most important. All others are associated causes. Because sequence plays a role, a condition closer to the time of death is the main cause. So for a baby with diarrhoea for two weeks who develops cough and rapid breathing three days before death: Main = ARI, Associated = diarrhoea. It is often difficult or impossible to distinguish which is the main cause, as in the case of nearly simultaneous ARI and diarrhoea. Decision rules are required for a host of circumstances. In this example, we code ARI as the main and diarrhoea as the associated cause. Malnutrition can be confusing, too and should be listed as an associated cause unless there is evidence that the child died purely from insufficient food, a very rare event. Obviously, the vast majority of children dying with malnutrition die of infection, and this would be listed as the main cause. Unforeseen circumstances require specific rules as they arise. These should be recorded in a VA procedure manual and consistently followed in the future. Each PHC program can decide for itself its coding rules. It does not really matter what is decided. Bias will be present, but this can be dealt with in the interpretation. More important is that random error is minimized. Most important is that the coding of diagnosis will not lose information so long as all information is analyzed. Coding requires some technical skills, including using the procedure manual and making logical decision rules for special cases. We recommend that the assigner make only tentative recommendations. The coder, then, checks the assigner's scoring, reviews the case, and codes the diagnoses. How to analyze mortality data Dual analysis Following the coding methods described previously, analysis is a dual process: 1) cause of death by main cause and 2) cause of death by consolidated cause, that is, all causes combined. For comparison purposes over time and among programs, analysis by single (main) cause is standard. However, analysis by consolidated cause yields the PHC manager more information. In the consolidated analysis the numerator is the count of diagnoses, not individuals. These die with a condition rather than of it. Once this is grasped, one can see that consolidated analysis is a powerful way to measure the true burden of given conditions. Table D-4 shows hypothetical data by way of illustration. Consider diarrhoea. Of 61 deceased children, 40 died of diarrhoea as the main cause. Of these, 20 had no associated cause; 12 had ARI; eight had malnutrition; three had measles; ten were LBW. Look at the ARI row. Of the 21, 11 had ARI without any associated cause. However, eight had diarrhoea as an associated cause, and so on. The number of children dying with diarrhoea, then, is 48 (40 + 8). Similarly, the number of children dying with ARI is 21 + 12 = 33, not 21. Table D-4: Causes of death: main and associated (hypothetical) Main cause , +, Associated causes, +, +, +, +, + , Total, none, ARI, diarr., malnutr., measles, LBW Diarrhoea , 40, 20, 12, -, 8, 3, 10 ARI , 21, 11, -, 8, 3, 3, 3 Total , 61, 31, 12, 8, 11, 6, 13 Note also that the row totals do not equal the total number listed in the main cause column. In other words, in the diarrhoea row, 20 + 12 + 8 + 3 + 10 = 53, not 40. How can this be? It is because children dying with diarrhoea not only had one, but often two or more associated conditions. That is, the 20 (40 - 20) who did have associated conditions actually accounted for (53 - 20) 33 associated conditions. Play with the numbers for a few minutes and you will grasp this easily. Note that the only total figure which corresponds to the number of deceased children is 61. The totals of the numbers of children dying with various conditions is greater than the number of deaths because children die with more than one condition. This also means that there are strategies for controlling different diseases that will impact a single disease's mortality burden. Conversely, reducing the cause-specific mortality of a specific disease will also reduce mortality due to other killers. Indeed, ARI programs are remarkably effective in lowering both pneumonia mortality and measles mortality. Examples of analysis and response The following examples of types of analysis and data presentation may guide PHC managers. The figures come from AKU's experience in five Karachi PHC programs with a population base of about 45,000 in 1991 when the IMR was 63. The results were obtained using an earlier instrument which has since been improved. These types of analysis and computations are readily performed by hand with or without a pocket calculator. Table D-5 shows the breakdown by main cause of death among all children under five years of age. EPI is well in place, but controlling diarrhoea is still a major challenge. ARI is a moderate problem although we believe that we are under-recognizing it. Experience has shown that the interviewers þrelaxþ once they establish one diagnosis, usually diarrhoea which is not difficult to identify in our setting. Thus, co-infections with ARI and diarrhoea are certainly being missed. Improved training and supervision is needed. Note also the unacceptable unknown rate, another indication of problems either in the instrument or in the collecting/assigning/coding processes. Table D-5: Main cause of death: Age 0-5 years, 1991 MAIN CAUSE, No., Percent % Diarrhoea syndromes, 38, 31.7 ARI/pneumonia, 15, 12.5 Birth injury/asphyxia, 7, 5.8 Prematurity/low birth weight, 7, 5.8 Injury, 5, 4.2 Malnutrition, 1, 0.8 EPI-preventable, 0, 0.0 Others, 11, 9.2 Unknown,, ill-defined syndromes, 36, 30.0 TOTAL, 120, 100 Not shown, but a highly recommended next step, would be a set of tables (or columns) ranking causes of death by age. We usually look at deaths in newborns (0-28 days of age), post-neonatal infants (one through 11 months of age), and children one through four years of age. Table D-6 shows the same children after analysis by consolidated cause. Note that the numbers refer to diagnoses, not children. Note also some important differences with Table D-5. First, malnutrition is ranked as the second most common cause with a proportionate mortality ratio of almost 15% (compared to a PMR of 0.8% as a main cause). This is because of our convention of coding it as an associated cause rather than a main cause. Also note that LBW assumes greater importance. Again this has been because of the decision to assign LBW as an associated cause (because the more proximate þbiomedicalþ causes are infection, low temperature, low blood sugar, immature lungs, etc). Unfortunately these main causes are usually unknown. One could dispute this rule; we are probably trying to obtain an unrealistic degree of detail. Finally note that the unknown rate, while still high, is lower than in Table D-5 (which has many main causes listed as unknown). Again, a separate set of tables (or columns) for various ages at death is suggested. Table D-6: Consolidated* cause of death: Age 0-5 years, 1991 CONSOLIDATED CAUSE, No. Diagnosis, Percent Diarrhoea syndromes, 42, 21.5 Malnutrition, 29, 14.9 Prematurity/low birth weight, 27, 13.8 ARI/pneumonia, 19, 9.7 Birth injury/asphyxia, 7, 3.6 Injury, 5, 2.6 EPI-preventable, 1, 0.5 Other, 21, 10.8 Unknown, 45, 23.1 TOTAL, 196, 100.5 (*main plus associated causes combined; counts refer to diagnoses,, not patients), +, + Other types of descriptive analysis are rewarding and not difficult to perform by hand. For example, cumulative mortality curves (0-100% of mortality by age at death) depict the ages of death for various sub-sets of the data: by total, by sex, or by individual diagnosis. Figure D-1 is an example showing age of death by sex, two very similar curves (in that example). Figure D-1: Cumulative under five mortality by age and sex Urban PHC Programme, Karachi, 1990-1991 o Male + Female Cumulative % Male = 141 Female = 141 Missing values = 4 Age at death in months The table shows that almost 50% of infant mortality occurred in the first six weeks of life, a time when cultural practices make access to the baby difficult. Teaching trained birth attendants to recognize illness in young infants might be an answer. Likewise if a PHC manager has noticed that no ARI deaths occurred after 24 months of age, he should focus the target group of his ARI control programme more narrowly to the less than two years of age. A cautionary note in recommending policy: mortality trends need several years of data before they can be assessed and, by the time they have been documented, they are probably already evolving. As PHC programme managers know, they are in the business of þchanging the distribution of death.þ There are several reasons why the level and composition of mortality could change in a target group. For example, adding an ARI control programme would hopefully lower the ARI proportionate mortality ratio, the IMR, and the ARI ASCSMR. Comparable changes have been seen through EPI and CDD programmes. In addition to specific programmatic inputs, the distribution of deaths is changing due to larger forces related to the "health transition" (a term coined to enlarge the search for the causes of improved health beyond the usual economic development and health services explanations). One definition of the health transition is the þsocial, cultural, and behavioral factors which parallel the epidemiologic transition and may do much to propel it.þ This related term, þepidemiologic transition,þ refers to the change in disease patterns usually accompanying þdevelopment,þ i.e., a decrease in (primarily childhood) mortality secondary to infection, malnutrition, and contaminated environment with an increase in morbidity and mortality due to chronic disease (in the aged) as the life expectancy increases and birth rates fall. The terms are often used interchangeably. The bottom line is this: a þsuccessfulþ PHC manager will among other achievements like þcommunity empowermentþ cause a favourable change in the level and distribution of deaths in target groups. This will probably be due to several effects: 1) specific programme activities, 2) more intangible þsocialþ factors that encourage mothers to be more assertive in providing their children the care they need; and 3) cross-sectoral economic factors, such as income generation schemes. Appendix E: Verbal autopsy protocols E.1: NEWBORN AND CHILD DEATH INQUIRY INSTRUCTIONS INTERVIEWER: , + 1., Complete pages 1-3 2., Complete all the modules indicated on page 3 ASSIGNER:, + 1., Review the completed form. 2., For each module completed,, complete the score column,, and 3., Circle the most certain cause of death. CODER:, + 1., For each module completed,, review the assigned cause(s) of death. 2., Transfer all assigned causes of death to page 12,, and 3., Tick one main cause and other associated causes (including other and unknown). 4., Comment as needed. IDENTIFYING INFORMATION:, +, +, +, +, , , , 1), Village:, , 2), Division:, , 3), House #:, 4), Family no.:, , 5), I.D.#:, , , , 6), Name of deceased: ___________________, +, +, +, +, 7), Sex: M( ); F( ), + 8), Name of informant: __________________________________, +, +, +, +, +, +, + 9), Relationship with deceased: _____________________________, +, +, +, +, +, +, + 10), Date of birth: ___/___/___, +, +, +, , , , + 11), Date of death: ___/___/___, +, +, +, , , , + 12), Age at death: ___________, +, +, +, , , , + , , (___), 2., Private hospital/clinic , , (___), 3., Govt. hospital , , (___), 4., Other: specify:_____ 14), Did the child receive treatment outside home before death?, +, +, + , , (___), 1., Yes , , (___), 2., No þ GO TO QUESTION 17 , , (___), 3., Don't know þ GO TO QUESTION 17 , , +, +, + 15), If YES,, where did the child receive treatment?, +, +, + , , (___), 1., At PHC centre , , (___), 2., Private doctor , , (___), 3., Govt. hospital , , (___), 4., Private hospital , , (___), 5., Homeopath , , (___), 6., Unqualified person , , (___), 7., Other: Specify._______________ 16), What treatment did the child receive?, +, +, , +, +, +, + , +, +, +, + , +, +, +, + INFORMATION FROM FAMILY FOLDER OR OTHER SOURCES: 17) Registration date ___/___/___ 18) Total family monthly income ___________________ 19) Is the father employed? Yes (_), No (_), Don't know (_) 20) Is the mother working away from home? Yes (_), No (_), Don't know (_) 21) Is the mother divorced/widowed/separated? Yes (_), No (_), Don't know (_) 22) Ethnic group 23) Religion 24) Nutritional status (most recent prior to last illness): NL (_); 1st (_); 2nd (_); 3rd (_) Immunization dates:, +, +, , , 25), BCG, ___/___/___, 26), Measles, ___/___/___ 27), DPT 1, ___/___/___, 28), DPT 2, ___/___/___ 29), DPT 1, ___/___/___, 30), Polio 1, ___/___/___ 31), Polio 1, ___/___/___, 32), Polio 3, ___/___/___ 33) How did the child die? (Include the informant's exact words. Probe, but do not interpret the history. Focus on physical events) SCREENING QUESTIONS FOR CAUSES OF DEATH Interviewer: Tick each condition present within three days of death, and go to all modules indicated. (N = no; Y = yes; DK = don't know) No., Condition at time of death, Present within 3 days of death?, +, +, If Y or DK,, go to: ^, ^, ^, ^, ^, ^ ^, ^, N, Y, DK, ^ 34, Three or more loose stools in 24 hours?, , , , DIARRHOEA page 95 35, Cough? ORrapid breathing >> 12 hrs? ordifficult breathing >> 12 hrs? ornoisy breathing >> 12 hrs?, , , , ARI page 96 36, Cough with vomiting in baby over three months old?, , , , PERTUSSIS þ page 97 37, Extreme sleepiness & fever? ORirritability & fever? ORconvulsions?, , , , BRAIN INFECTION þ page 98 38, Stiff body? ORlockjaw? ORconvulsions?, , , , TETANUS þ page 99 39, Death in the 1st week of life?, , , , PERINATAL þ page 100 39a, Underweight ORmalnourished (local term), , , , MALNUTRITION þ page 101 40, Injury or accident?, , , , INJURY þ page 102 No., Condition at time of death or within three months, Present within 3 days of death?, +, +, If Y or DK,, go to: ^, ^, ^, ^, ^, ^ ^, ^, N, Y, DK, ^ 41, Rash with cough within three months prior to death?When: ____________________, N, Y, DK, MEASLES þ page 103 DIARRHOEA Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the weight letter in the score column.Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, N, Y, DK, Weight, Score 42, Were there three or more loose or liquid stools per day?, , , , E, 43a, Did the baby have: dry mouth? OR, , , any =S, 43b, sunken fontanel? OR, , , , ^, ^ 43c, extreme thirst? OR, , , , ^, ^ 43d, sunken eyes?, , , , ^, ^ 44a, Was there blood in the stools? OR, , , , any =D, 44b, Was there severe cramping? OR , , , , ^, ^ 44c, Was there fever?, , , , ^, ^ 45, Did the loose stools persist for more than two days?, , , , C, 46, Did the loose stools persist for more than 14 days?, , , , P, SCORING CAUSE OF DEATH: DIARRHOEA SYNDROMES Assigner: Circle all possible causes at the highest level of certainty Diagnosticlikelihood, ACUTE WATERY DIARRHOEA, DYSENTERY, PERSISTENT DIARRHOEA probable, 1E + 1C + 1S, 1E + 1D + 1C, 1E + 1P + 1C TABLESP-C, possible, 1E + 1 S, 1E + 1D, 1E + 1P ARI Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, N, Y, DK, Weight, Score 47, Was there a cough for more than two days?, , E, 48, Was there rapid breathing for more than one day?, , , , E, 49a, Was there,, for more than 1/2 day:chest wall indrawing? OR, , , , any =E, 49b, nostril flaring? OR, , , , ^, ^ 49c, grunting on breathing out? OR, , , , ^, ^ 49d, blue lips or tongue?, , , , ^, ^ SCORING CAUSE OF DEATH: ARI Assigner: Circle cause at the highest level of certainty Diagnostic likelihood, ARI probable, 3E possible, 2E PERTUSSIS Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þYþ or unshaded þDK,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, DK, N, Y, Weight, Score 50, Were there bouts of severe coughing for more than two weeks?, , , , C, 51, Was there choking or vomiting after coughing bouts?, , , , C, 52, Was there a þwhoopþ sound during breathing in?, , , , C, 53, Was there swelling of the eyelids?, , , , S, 54, Were there other cases of þwhooping coughþ in the village/neighbourhood?, , , , S, 55, Did the baby receive three or more DPT shots?, , , , X, SCORING CAUSE OF DEATH: PERTUSSIS Assigner: Circle cause at the highest level of certainty Diagnostic likelihood, PERTUSSIS probable, 3C + no X or2C + 2S + no X possible, 2C + no X BRAIN INFECTION Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þYþ or unshaded þDK,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, DK, N, Y, Weight, Score 56, Was there fever?, , , , E, 57, Was the child more irritable when picked up than when left alone?, , , , C, 58, Was there extreme sleepiness?, , C, 59, Was the neck stiff?, , C, 60, Were there any convulsions?, , C, 61, Was there a raised fontanel?, , S, 62, Was there vomiting?, , S, SCORING CAUSE OF DEATH: BRAIN INFECTION Assigner: Circle cause at the highest level of certainty Diagnostic likelihood, BRAIN INFECTION probable, 1E + 3C or1E + 2C + 2S possible, 1E + 2C TETANUS Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. Table A. NEWBORNS DYING BETWEEN AGE 3-28 DAYS OF LIFE, +, +, +, +, +, (younger newborns go to PERINATAL MODULE; older children: see Table B below), +, +, +, +, +, No., Was this symptom present?, DK N Y, Weight, Score 63, Was the death between 3-28 days of age?, , E, 64, Was the baby able to suck well after birth?, , E, 65, Was the jaw locked?, , C, 66a, Was there: back-arching/ rigidity? OR, , any = C, 66b, convulsions?, , ^, ^ Table B. OLDER CHILDREN (MORE THAN 28 DAYS OF AGE), +, +, +, +, +, No., Was this symptom present?, DK N Y, Weight, Score 67, Did the child's body become rigid?, , Ch, 68, Was the jaw locked?, , Ch, 69a, Were there convulsions? AND, , , 69b, Was the child awake between them?, , both =Ch, SCORING CAUSE OF DEATH: TETANUS SYNDROMESAssigner: Circle all possible causes at the highest level of certainty Diagnostic likelihood, NEWBORN TETANUS, CHILD TETANUS probable, 2E + 2C, 3Ch possible, 2E + 1C, 2Ch PERINATAL SYNDROMES Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, DK N Y, Weight, Score 70, Did the baby die before age 7 days?, , E, 71, Did the baby fail to shown signs of life? (crying,, breathing,, moving), , SB, 72a, Was the baby smaller than usual? AND, , both =PR, 72b, born of a pregnancy of << 8 months?, , ^, 73a, Was the baby smaller than usual? AND, , both =SGA, 73b, born of a pregnancy of >>or = 8 months?, , ^, 74, Did the baby not cry for at least five minutes after birth?, , C, 75, Did the baby have convulsions?, , C, 76, Was the delivery longer than 12 hours?, , C, 77, Did the baby show any abnormality of head,, chest/abdomen,, arms/legs?, , A, SCORING CAUSE OF DEATH: PERINATAL SYNDROMES Assigner: Circle all possible causes at the highest level of certainty Diagnostic likelihood, STILLBIRTH, PREMATURITY, SMALL FOR GESTATIONAL AGE, BIRTH INJURY/ASPHYXIA, CONGENITAL ANOMALY probable, 1E + 1SB, , , 1E + 3C, 1E + 1A possible, , 1E + 1 PR, 1E + 1SGA, 1E + 2C, MALNUTRITION Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, DK N Y, Weight, Score 78, In the month before death was the child:Listless? ORLacking energy? ORDisinterested in surroundings?, , C , 79, Did the child have malnutrition (local term)?, , C , 80, FOR CHILDREN >> six MONTHS: Was the child fed any milk besides breast milk before age six months?, , C, 81, In the three months before death,, did the child have diarrhea three or more times?, , S, 82, In the three months before death did any diarrhea episode last more than 14 days?, , S, ^ 83, In the three months before death did the child have diarrhea AND pneumonia (either at the same time or at different times)?, , S, 84, Was the child especially hungry compared to other children his/her age?, , S , SCORING CAUSE OF DEATH: MALNUTRITION Assigner: Circle cause at the highest level of certainty Diagnostic likelihood , MALNUTRITION probable, 2C1C - 2S Possible, 1C or 3S INJURY Interviewer: Answer each question. (DK = don't know; N = no; Y = yes)Make sure that the informant's history is complete (page 2). Additional space is provided below if needed. No., Were any of these situations present?, DK N Y 85, Was there a bite or sting from an animal or poisonous insect?, 86, Was there a burn?, 87, Was there drowning?, 88, Was there poisoning?, 89, Was there a traffic accident?, 90, Was there a fall?, 91, Was there choking?, 92, Was there cutting or piercing from a sharp object or bullet-like projectile?, SCORING CAUSE OF DEATH: INJURY Assigner: Each þYþ represents a probable injury diagnosis. Diagnostic likelihood, INJURY probable, 1Y MEASLES Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the weight letter in the score column. Refer to the SCORING CAUSE OF DEATH table to assign cause of death. No., Was this symptom present?, DK N Y, Weight, Score 93, Did the child die after age four months?, , E, 94, Was there rash for more than three days?, , E, 95, Was there fever for more than three days?, , E, 96, Was there cough with the rash?, , E, 97, IF the child survived for more than five days after the rash started,, did the skin peel all over? (not applicable), , C, 98, During the fever & rash,, were the eyes red?, , C, 99, Were there other cases of measles in the house or in the village/neighbourhood?, , S, 100, In your opinion,, did the child have measles within three months of death?, , S, 101, Did the rash ever look like little blisters filled with water?, , X, SCORING CAUSE OF DEATH: MEASLES Assigner: Circle cause at the highest level of certainty Diagnostic likelihood , MEASLES probable, 4E + 2C + no X or4E + 1C + 2S + no X possible, 4E + 1C + no X CAUSE OF DEATH: CODING CHART Coder: Follow decision rules to code cause of death by ONE main and any associated cause(s). Add comments as desired. Sign and date. Thank you. DXCodeno., CAUSE OF DEATH, probable, possible, Main(1 only), Associated , Acute watery diarrhoea, , , , , Dysentery, , , , , Persistent diarrhoea, , , , , ARI, , , , , Pertussis, , , , , Measles, , , , , Brain infection, , , , , Neonatal tetanus, , , , , Childhood tetanus, , , , , Stillbirth, , , , Prematurity, , , , Small for gestational age, , , , Birth injury/asphyxia, , , , , Congenital anomaly, , , , Injury: specify, , , , Other: specify, , , , , , Unknown, , , , CODER COMMENTS: CODER SIGNATURE: DATE: REVIEWER COMMENTS: REVIEWER SIGNATURE: DATE: NEWBORN AND CHILD DEATH INQUIRY INSTRUCTIONS TO TRAINERS AND INTERVIEWERS: EXPLANATION OF QUESTIONS No.Explanation 34. Three or more loose stools in 24 hours: separate episodes in a day 35. Cough: Rapid breathing: faster than normal Difficult breathing: requires more muscular effort; sucking in of skin <%-3>between/above/below ribs; <%-4>unable to drink<%-3> normally; chest<%0> wall caving in Noisy Breathing: normal breathing is quiet; noisy breathing has sounds on breathing in or out; squeaks, wheezes, grunts 36. Cough with vomiting: coughing so hard that it causes the baby to vomit 37. Extreme sleepiness and fever: does not wake up enough to recognize mother or eat properly Irritability and fever: cries more than usual 38. Stiff body: Lockjaw: jaw won't fully open 41. Rash with cough within three months of death: red spots on body 42. Same as 34 43a. Dry mouth: loss of wetness or saliva on inner lips, inner cheeks, tongue 43b. Extreme thirst: frantic for liquids 43c. Sunken eyes: eyes that look þowlishþ with darkish depressions around them due to loss of water 43d. Sunken fontanel: baby's soft spot may normally sink a bit when upright; this is much more than that 44a. Blood in stool: usually small reddish spots in the motions, but may be lots 44b. Severe cramping: cries just before or during passing motions 44c. Fever: temperature above 101 degrees F (38.5 C) or felt warm 45. Loose stools persist for more than two days: at least into the 3rd day 46. Loose stools persist for more than 14 days: passed loose stools daily for this period 47. Cough for more than 2 days: at least into the third day 48. Rapid breathing: see 35 49a. Chest wall in-drawing: lower ribs suck in with breathing in (the opposite of what normally happens) 49b. Nostril flaring: nostrils widen and open up to let more air in 49c. Grunting on breathing out: brief vocalizations during exhaling as if in pain; best to demonstrate: þhghþ (listen to your trainer) 49d. Blue lips or tongue: dusky blueness (can also be seen in nail beds) 50. Bouts of severe coughing: coughing spells for more than a minute separated by quiet episodes 51. Choking or vomiting after coughing bouts: like 36, but adds the idea of choking. Babies can choke either on mucus or their vomit 52. þWhoopþ sound: or þHoooþ sound as baby breathes in usually at the end of a long coughing bout 53. Swelling of eyelids: puffiness of upper and/or lower lids 55. Did the baby receive three or more DPT shots: by written record only. Most will not be able to prove this 56. Fever: see 44c 57. More irritable when picked up: usually sick children prefer to be picked up and become less irritable. Babies who hurt don't want to be touched 58. Extreme sleepiness: see 37 59. Neck stiff: hurts to bend the neck forward. The head may be bent back 60. Convulsions: episodes of jerking of arms and/or legs with or without loss of consciousness; seizures, fits 61. Raised fontanel: full or bulging soft spot, especially when baby held erect 63. Death between 3-28 days: dying on the 3rd through the 28th day of life 64. Suck well after birth: Normal sucking at least three different times 65. Jaw locked: can't open enough to suck 66a. Back arching: back bent backwards and stiff in that position 66b. Convulsions: see 60 67. Rigid body: arms and/or legs and/or back all stiff, possibly with back arching like 66b 68. Jaw locked: can't open enough to eat 69. Convulsions AND awake between them: although the child is very sick, he is conscious between the convulsions; he may move eyes to mother or express emotion 70. Die before age seven days: dies in the first week of life 71. Fail to show signs of life: was the baby dead at birth, i.e., without heartbeat, breathing, crying, movement or arms/legs/face muscles 72a. Smaller than normal: smaller than the usual newborn? or less than 2500 grams (if weighed) 72b. Pregnancy << 8 months: premature 73a. See 72a 73b. Pregnancy or = 8 months: at full term or nearly so 74. Not cry for at least five minutes after birth: almost all healthy babies cry after a few seconds or minutes of birth. Five minutes is a long time NOT to cry, and the mother should remember it 75. Convulsions: see 60; they are harder to see in a newborn 76. Delivery 12 hours: was there pushing for more than 12 hours 85. Bite or sting: animal (large or small): mammal, bird, insect, spider, reptile, fish, jellyfish, and so on 86. Burn: from flame, heater, stove. etc. 87. Drowning: in pond, river, sea, bucket of water, dish of milk, etc. 88. Poisoning: medicine, plant, cleaner, chemical, insecticide, petrol, rat poison, etc. 89. Traffic accident: an event involving any of the following: pedestrian and/or passengers on wheeled conveyance (car, rickshaw, bicycle, tonga, motorcycle, bus, etc.) 90. Fall: from standing onto a hard surface, from a tree, from a cliff, etc. 91. Choking on food, on a toy, on a stick; suffocating while sleeping, etc. 92. Cutting, piercing object: gunshot, knifeblade, glass, arrow, etc. 93. After age four months: after he turned four months 94. Rash for more than three days: at least into the fourth day 95. Fever for more than three days: see 44c 96. Cough with rash: cough for three days also 97. Peeling rash after the fifth day: skin flakes off 98. Red eyes: visible red lines (tiny blood vessels) or generalized redness of the whites of the eyes 101. Rash look like blisters; did it look like chickenpox E.2: MATERNAL DEATH INQUIRY INSTRUCTIONS INTERVIEWER: 1. Complete pages 1-3. 2. Complete all the modules indicated on page 3. ASSIGNER: 1. Review the completed form. 2. For each module completed, complete the score column, and 3. Circle the most certain cause of death. CODER: 1. For each module completed, review the assigned cause(s) of death. 2. Transfer all assigned causes of death to page 10, and 3. Tick one main cause and other associated causes (including other and unknown). 4. Comment as needed. IDENTIFYING INFORMATION 1)Name of deceased: 2)I.D.#: 3)Village:4) Division: 5)House #:6) Family no.: 7)Name of informant: 8)Relationship to deceased: 9)Date of deceased's birth: 10)Date of deceased's death 11)Deceased's age at death 12)Total family monthly income 13)Language group: 14)Religion: 15)What was her marital status at the time of death? ( ) Married ( ) Divorced ( ) Separated ( ) Widowed ( ) Single 16) If MARRIED, how long had she been married?(_____) 17) How many time had she been pregnant?(_____) 18) How many abortions had she had?(_____) 19) How many stillbirths had she had?(_____) 20) How many children were born alive to her?(_____) 21) How many of those children are still alive?(_____) 22) Was she pregnant in the last five years?(_____) 23) If YES, how many times?(_____) 24) Was she pregnant at the time of death?(_____) 25) If YES, for how many months?(_____) 26) Had she been pregnant within 42 days of death?(_____) 27) Where did the woman die?( ) 1. Residence ( ) 2. Private hospital/clinic ( ) 3. Govt. hospital ( ) 4. Other: specify:__________________ 28) Did she receive treatment outside home before death? ( ) 1. Yes ( ) 2. No ( ) 3. Don't know 29) If YES, where did the she receive treatment? ( ) 1. At PHC centre ( ) 2. Private Doctor (MBBS) ( ) 3. Govt. hospital ( ) 4. Private hospital ( ) 5. Homeopath ( ) 6. Unqualified person ( ) 7. Other: specify:__________________ 30) How did the woman die? Include all treatments received. (Include the informant's exact words. Probe, but do no interpret the history. Focus on physical events) SCREENING QUESTIONS FOR CAUSES OF DEATH Interviewer: Tick a response for each TIME and sign combination, and go to all modules indicated. (N = no; Y = yes; DK = don't know) No., Time and Sign, Present at death?, +, +, If Y or DK,, go to: ^, ^, N Y, DK, ^ 31, Post-delivery ANDfever ORfoul smelling vaginal discharge, , , PUERPERAL INFECTIONþ page 111 32, mid- OR late pregnancy ANDvaginal bleeding up to death, , , ANTEPARTUM HEMORRHAGE þ page 112 33, Post-delivery ANDvaginal bleeding up to death, , , POSTPARTUMHEMORRHAGE page 113 34, During labour ANDbaby not delivered in 24 hours, , , OBSTRUCTED LABOR þ page 114 35, During pregnancy OR labour OR delivery OR delivery day AND seizures, , , ECLAMPSIA MODULEþ page 115 36, Early OR mid-pregnancy AND fever OR, , , ABORTION (MISCARRIAGE) þ page 116 37, Abdominal pain OR foul smelling vaginal discharge, , , PUERPERAL INFECTION Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the Weight letter in the score column. Refer to the ASSIGNING CAUSE OF DEATH table to assign cause of death. No., Sign or symptom, No Don't know Yes, Weight, Score 38, Was the death after delivery?, , , , E, 39, Was the death within six weeks of delivery?, , , , E, 40, Was the death within one week of delivery?, , , , C, 41, Was there fever?, , , , S, 42, Were there chills and sweats?, , , , C, 43, Was there a foul-smelling vaginal discharge?, , , , C, 44, Was there abdominal pain (below the navel)?, , , , C, 45, Was there bleeding from several sites?, , , , C, RULES FOR ASSIGNING CAUSE OF DEATH: PUERPERAL SEPSIS Assigner: Determine the score (above), and circle the highest level of certainty. Probability, Score possible, 2E + 1C + 1S probable, 2E + 2C ANTEPARTUM HEMORRHAGE Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the Weight letter in the score column. Refer to the ASSIGNING CAUSE OF DEATH table to assign cause of death. No., Sign or symptom, No, Don't know, Yes, Weight, Score 46, Did the death occur in mid- or late pregnancy?, , , , E, 47, Was the woman not in labor?, , , , E, 48, Was the bleeding painless?, , , , P, 49, Had there been previous painless vaginal bleeding during this pregnancy?, , , , PS, 50, Was there sudden abdominal pain below the navel?, , , , A, 51, Did the abdomen swell below the navel?, , , , AS, RULES FOR ASSIGNING CAUSE OF DEATH: ANTEPARTUMHEMORRHAGE Assigner: Determine the score (above), and circle the highest level of certainty. Probability, Diagnosis Score, + , Placenta Previa, Placental Abruption possible, 2E + 1P, 2E + 1A probable, 2E + 1P + 1PS, 2E + 1A + 1AS POSTPARTUM HEMORRHAGE Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the Weight letter in the score column. Refer to the ASSIGNING CAUSE OF DEATH table to assign cause of death. No., Sign or symptom, No, Don't know, Yes, Weight, Score 52, Was there heavy vaginal bleeding up to the time of death,, such as:bright red blood? ORclots?ORpads could not keep up with it?, , , , E, 53, Was the afterbirth incompletely delivered?, , , , C, 54, Were shreds of membranes visible in the birth canal?, , , , C, RULES FOR ASSIGNING CAUSE OF DEATH: POSTPARTUM HEMORRHAGE Assigner: Determine the score (above), and circle the highest level of certainty. Probability, Score possible, 1E probable, 1E + 1C OBSTRUCTED LABOR Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the Weight letter in the score column. Refer to the ASSIGNING CAUSE OF DEATH table to assign cause of death. No., Sign or symptom, No, Don't know, Yes, Weight, Score 55, Was the woman in labour for more than 24 hours?, , , , E, 56, Was the woman in labour for more than 48 hours?, , , , C, 57, Was the abdominal pain worse than the usual pains of labour?, , , , C, 58, Was there abdominal swelling or change in abdominal shape?, , , , C, 59, Was there heavy vaginal bleeding up to the time of death?, , , , C, 60, Was there an abnormal presentation of the baby (feet,, bottom,, arm,, shoulder,, cord first; or twins)?, , , , S, 61, Was the mother short (less than five feet or less than 150 cm)?, , , , S, 62, Was there a history of large previous babies?, , , , S, RULES FOR ASSIGNING CAUSE OF DEATH: OBSTRUCTED LABOR Assigner: Determine the score (above), and circle the highest level of certainty. Probability, Score possible, 1E + 2C probable, 1E + 2C + 2S or1E + 3c ECLAMPSIA Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the Weight letter in the score column. Refer to the ASSIGNING CAUSE OF DEATH table to assign cause of death. No., Sign or symptom, No, Don't know, Yes, Weight, Score 63, Was the woman pregnant?, , , , E, 64, Was there at least one seizure?, , , , E, 65, Was there swelling of the:face? ORHands and ankles? OR, , , , C, 66, Was there high blood pressure during the pregnancy?, , , , C, 67, Was this the first pregnancy?, , , , S, 68, Was there:headache? ORvisual disturbance?, , , , S, 69, Was there abdominal pain?, , , , S, RULES FOR ASSIGNING CAUSE OF DEATH: ECLAMPSIA Assigner: Determine the score (above), and circle the highest level of certainty. Probability, Score possible, 2E + 1C probable, 2E + 2C or2E + 1C + 2S ABORTION (MISCARRIAGE) Interviewer: Answer each question. (DK = don't know; N = no; Y = yes) Assigner: For each þY,þ enter the Weight letter in the score column. Refer to the ASSIGNING CAUSE OF DEATH table to assign cause of death. No., Sign or symptom, No, Don't know, Yes, Weight, Score 70, Was the woman in early or mid-pregnancy?, , , , E, 71, Was there fever?, , , , C, 72, Was there foul smelling vaginal discharge?, , , , C, 73, Was there abdominal pain below the navel?, , , , C, 74, Was there:Depression ORUnwanted pregnancy OR, , , , C, 75, Was there vaginal bleeding up to time of death?, , , , C, 76, Was there passage of incompletely formed baby or tissue?, , , , E, RULES FOR ASSIGNING CAUSE OF DEATH: Abortion(miscarriage) Related Assigner: Determine the score (above), and circle the highest level of certainty. Probability, Score possible, 2E + 1C probable, 2E + 2C CAUSE OF DEATH: CODING CHART Coder: Follow decision rules to code cause(s) of death by ONE main and any associated cause(s). Add comments as desired. Sign and date. Thank you. DXcodeno., Cause of death, probable, possible, main(1 only), associated , Puerperal infection, , , , , Placenta previa, , , , , Placental abruption, , , , , Postpartum hemorrhage, , , , , Obstructed labor, , , , , Eclampsia, , , , , Abortion-related, , , , , Other: (specify), , , , , Unknown, , , , , , , , , , , , , , , , , , , , , , , , CODER COMMENTS: CODER SIGNATURE:DATE: REVIEWER COMMENTS: REVIEWER SIGNATURE:DATE: DEFINITIONS: TIME OF DEATH AND MAJOR SIGNS Time of death Early pregnancy; missed periods with prior regular periods; breast swelling; without noticeable abdominal swelling; may not be known by anyone, including the decedent Mid-pregnancy; visibly pregnant; no periods for 4-6 months; baby due in 3-5 months; baby may be kicking; usually noticed by family; almost always by the mother-to-be Late pregnancy; very visibly pregnant; definite baby kicking; no periods for 6-9 months; baby due in 0-3 months Labour; intermittent pains present with or without the urge to push; passage of water; bloody show Delivery; urge to push with the appearance and birth of the baby and passage of the afterbirth Post-delivery; within 42 days of delivery Signs Vaginal bleeding up to death: passage of copious bright red blood, dark blood, or clots from the vagina/birth canal Seizures: interrupted consciousness with stiffening and/or jerking of one or more limbs, +/- jaw clenching, +/- loss of urinary continence Abdominal pain/tenderness (not in labour): in this case, usual lower mid-line, anterior discomfort which is below the navel, i.e., where the womb is Difficult delivery with labour 24 hours: the normal discomfort of intermittent labor pains evolved into prolonged bouts of unrelenting pain, often with maternal physical and mental exhaustion without achieving delivery within a day's time Fever: temperature above 100 degrees F or þfeels warm to touchþ or intermittently clammy (cold) and flushed (sweaty, warm) skin; may have chills (mild shivers) or rigours (severe shivers) Foul-smelling vaginal discharge: bloody or purulent mucus discharge from vagina/birth canal that has an unusually bad odor References and bibliography General CDC. Guidelines for evaluation surveillance systems. Morbidity and monitoring weekly report supplement. Vol 37, No. S-5, May 6, 1988. Halperin W. and E. L. Baker. Public health surveillance. Van Nostrand Reinhold, New York, 1992. WHO/EPI/MLM/91.4 Training for mid-level managers: disease surveillance. Revised, 1991. Vaughan J. P. and R. M. Marrow. Manual of epidemiology for district health management. WHO, 1989. Vernon, A. Cause of death from hospital data in developing countries: A review based on the CCCD experience with hospital-based mortality surveillance systems in Sub-Saharan Africa. Draft discussion paper, 9 May 1992. Verbal Autopsy Bang, A.T. and R. A. Bang and the SEARCH team. Diagnosis of causes of childhood deaths in developing countries by verbal autopsy; suggested criteria. WHO Bull., pp. 499-507, 1992. Gray, R. H. "Verbal autopsy: using interviews to determine causes of death in children," IIP Occasional paper. no. 14. The Johns Hopkins University, Baltimore, USA, March, 1991. Gray, R. H., et al. "The use of verbal autopsy methods to determine selected causes of death in children." IIP Occasional paper no. 10. The Johns Hopkins University, Baltimore, USA, February, 1990. Kalter, H. D., et al. "Validation of postmortem interviews to ascertain selected causes of death in children." Int J Epi. 19:380-386, 1990. Snow, B. and K. Marsh. "How useful are verbal autopsies to estimate childhood causes of death?" Health policy and planning. 7:22-29, 1992. Snow, R. W., et al. "Childhood deaths in Africa: uses and limitations of verbal autopsies." Lancet, 340:351-355, 1992. Acronyms and abbreviations acronyms and abbreviations AKFAga Khan Foundation ALRIAcute lower respiratory infection AKUAga Khan University ANCAntenatal care ARIAcute respiratory infection AWDAcute watery diarrhoea BCGBacillus of Calmette and Guerin CCCDCombatting Childhood Communicable Diseases CDCCenter for Disease Control, U.S. Public Health Service (USA) CDDControl of diarrhoeal diseases CHNCommunity health nurse CHWCommunity health worker DPTDiphtheria, pertussis, tetanus EPIExpanded Programme for Immunization FUOFever of unknown origin GMGrowth monitoring HHHousehold HIVHuman immunodeficiency virus IECInformation, education, communication IMRInfant mortality rate LBWLow birth weight MCHMaternal and child health MISManagement information system MMRMaternal mortality rate MOHMinistry of health NNTNeonatal tetanus OPVOral poliovirus vaccine ORSOral rehydration salts ORTOral rehydration therapy PHCPrimary health care PHC MAPPrimary Health Care Management Advancement Programme PMRProportionate mortality rate SGASmall for gestational age STDSexually-transmitted diseases TBTuberculosis TBATraditional birth attendant TTTetanus toxoid WHOWorld Health Organization Glossary Attack rate: The percentage of individuals in a defined group who get a disease over a defined time period. Case: An individual situation or occurrence. In health this usually refers to an individual person with a disease, health problem or who has died. Case fatality rate: The proportion who die, of the number of persons diagnosed as having a specific disease. Catchment (area): The geographic area surrounding one or more health facilities. It refers to the population residing in that area, which includes the programme's target populations. Census: A count of all members of a population. Community health worker (CHW): A person indigenous to the community who provides basic health promotion, disease prevention and curative health services to members of the community. Includes "village health workers," "health guides," and other terms. Coverage: The proportion of a target group that has received a service or is protected from a disease or health problem. Disease surveillance: The collection of information about cases of diseases and the use of that information to evaluate the effectiveness of preventive activities to correct any problems which hinder disease-reduction objectives from being met. Disease trend: The pattern formed by increases and decreases in the number of reported cases of disease over time. Effectiveness: The degree to which desired outcomes are achieved. Efficiency: The degree to which desired outcomes are achieved without wasting resources. Endemic: The constant presence of a disease or infectious agent within a given geographic area. Epidemic: The occurrence in a community or region of more cases of a disease than usually occur in a specified period of time. Synonym: outbreak. Epidemic pattern: The occurrence of a disease in a pattern in which more cases occur during certain periods of time than in other periods. Incidence: The number of new cases of a disease in a defined population during a specified period of time. Incubation: The time interval between initial contact with an infectious agent and the appearance of the first sign or symptom of disease. Indicator: An indirect measure of an event or condition. For example, a baby's weight-for-age is an indicator of the baby's nutritional status. Inputs: Resources, e.g., personnel, equipment, information and money. Lay definition: Non-medical definition, using terminology easily understood and used by non-clinical individuals to describe a medical event or condition. Management: Th<%-2>e art and science of getting things done through pe<%0>ople. Objectives: The output and/or effect your PHC programme hopes to have. Outbreak: The occurrence in a community or region of more cases of a disease than usually occur in a specified period of time. Synonym: epidemic. Outbreak investigation: A study conducted for the purpose of collecting data about an outbreak, with the objective of controlling the outbreak and preventing similar outbreaks in the future. Outcomes: Results of your PHC programme, including outputs, effects and impacts. Outputs: Products and services provided by a PHC programme. Effects: Changes in knowledge, skills, attitude, and behaviour (including coverage) as a result of a PHC programme. Impacts: Changes in health status, (mortality, morbidity, disability, fertility) as a result of a PHC programme. Percentage: A proportion multiplied by 100. For example 3,500 children immunized out of 5,000 x 100. (3,250/5,000) * 100 = 65%. Prevalence: The total number of cases of a disease in a defined population at a specified point in of time. Also used with "coverage," as with the þcontraceptive prevalence rate,þ meaning the proportion of the target population that is currently practising family planning. Primary health care: Essential health care, accessible at affordable cost to the community and the country, based on practical, scientifically sound and socially acceptable methods. It includes at least eight components: health education, proper nutrition, basic sanitation, maternal and child health care, immunization, control of common diseases and injuries, prevention of local endemic diseases and provision of essential drugs. Processes: Activities or tasks carried out through the PHC programme. Proportion: A special type of ratio expressing a relationship between a part and the whole. For example, 3,250 children immunized out of 5,000 (3,2500/5,000 = .65). Rate: A measure of the frequency of occurrence of an event, such as cases per month. Ratio: Two numbers related to each other in a fraction or decimal, such as the number of cases of measles per 1,000 children. Any fraction, quotient, proportion, or percentage is a ratio. Routine reporting site: Compilation and reporting of selected epidemiological data by all health facilities in a designated surveillance area. Data are usually taken from routine records. Seasonal variation: The occurrence of a disease in a pattern where more cases occur in one (or more) seasons of the year. Sentinel reporting site: A health facility that is specially chosen to collect and report specific epidemiological data, often taken from special records and forms. Signs of disease: The evidence of disease found in a case by the examiner. Special surveys: Studies that collect data that cannot be obtained from routine and sentinel reporting systems. Symptoms of disease: The sensations of disease felt by the patient. System: A set of discrete, but interdependent, components designed to achieve one or more objectives. Target group: Specific groups of people designated to receive a PHC service, such as children under age three. Vaccine efficacy: The ability of a vaccine to prevent disease when used in routine immunization services. Verbal autopsy: An in-depth investigation by structured interviews of the cause of death and the circumstances leading up to it. Vital registration: Recording and reporting of births and deaths on a routine basis to a central authority by public and private health providers. PHC MAP MANAGEMENT COMMITTEE Dr. Ronald Wilson Aga Khan Foundation, Switzerland (Co-Chair) Dr. Jack Bryant Aga Khan University, Pakistan (Co-Chair) Dr. William Steeler Secretariat of His Highness the Aga Khan, France (Co-Chair) Dr. Jack Reynolds Center for Human Services, USA (PHC MAP Director) Dr. David Nicholas Center for Human Services, USA Dr. Duane Smith Aga Khan Foundation, Switzerland Dr. Pierre Claquin Aga Khan Foundation, Switzerland Mr. Aziz Currimbhoy Aga Khan Health Service, Pakistan Mr. Kabir Mitha Aga Khan Health Service, India Dr. Nizar Verjee Aga Khan Health Service, Kenya Ms. Khatidja Husein Aga Khan University, Pakistan Dr. Sadia Chowdhury Aga Khan Community Health Programme, Bangladesh Dr. Mizan Siddiqi Aga Khan Community Health Programme, Bangladesh Dr. Krasae Chanawongse ASEAN Institute for Health Development, Thailand Dr. Yawarat Porapakkham ASEAN Institute for Health Development, Thailand Dr. Jumroon Mikhanorn Somboon Vacharotai Foundation, Thailand Dr. Nirmala Murthy Foundation for Research in Health Systems, India PHC MAP TECHNICAL ADVISORY COMMITTEE Dr. Nirmala Murthy Foundation for Research in Health Systems, India (Chair) Dr. Krasae Chanawongse ASEAN Institute for Health Development, Thailand Dr. Al Henn African Medical and Research Foundation (AMREF), formerly of the Harvard Institute for International Development Dr. Siraj-ul Haque Mahmud Ministry of Planning, Pakistan Dr. Peter Tugwell Faculty of Medicine, University of Ottawa, Canada Dr. Dan Kaseje Christian Medical Commission, Switzerland, formerly of the University of Nairobi, Kenya KEY PHC MAP STAFF AT THE CENTER FOR HUMAN SERVICES Dr. Jack Reynolds (PHC MAP Director) Dr. Neeraj Kak Dr. Paul Richardson Ms. Lori DiPrete Brown Dr. David Nicholas Ms. Pam Homan Dr. Wayne Stinson Dr. Lynne Miller Franco Ms. Maria Francisco Ms. Mary Millar