Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Sentinel surveillance in health and nutrition: Experience in Guatemala
Hernán L. Delgado and Mireya Palmieri
Health information systems in most Central American countries fail to produce reliable, timely, representative, and useful information to define target groups and interventions. This situation originated the need to develop new methodologies for the collection, analysis, and dissemination of information regarding health and nutrition, as well as health services. For this purpose, between 1985 and 1987, the Institute of Nutrition of Central America and Panama (INCAP) and the Ministry of Health and Social We/fare of Guatemala jointly implemented a sentinel surveillance subsystem. During this period, 119 rural communities of 1,000 + 250 inhabitants were longitudinally studied in four rounds of surveys. The Institute transferred the methodology and results to the Ministry of Health through INCAP's communication channels: dissemination of information, direct technical assistance, training, and research. The methodology has not been widely used by national health officialis; therefore, it has become necessary to develop operational research to assess the relationship between decision-making and information availability.
Health services require timely and high-quality information on the nature, magnitude, and distribution of food, nutrition, and health problems for national, ecodemographic, and socioeconomic groups, to plan regional or national programmes aimed at specific targeted groups. Information should distinguish between more efficient activities that have to be strengthened and those not requiring support. Finally, due to the political and economical crisis affecting specific social groups in the countries, it is necessary to have updated information on the food, nutrition, and health status of these groups [1, 2].
Most of the information managed by the health sector is derived from the interaction between care providers and users, and to a lesser extent from specific studies such as surveys on health and nutrition status and the performance of the programmes. Currently, regular information systems have limitations associated with the amount of information (more information than necessary is collected for decisionmaking), its quality (information is representative only of the health-care users and is usually unreliable), and the opportunity (type of information flow and limitations in data processing delaying its use). On the other hand, special studies may overcome these restrictions, but they are limited by high costs and low use by decision makers [3].
These restrictions have forced Central American technical groups to develop information subsystems that, by supporting the regular system, would provide valid, reliable, and timely information to plan, monitor, evaluate, and identify high-risk groups. In response to this situation, INCAP, in cooperation with the Ministry of Public Health and Social Welfare, implemented a sentinel surveillance subsystem that was developed in Guatemalan rural areas during 1985-1987.
In 1985 the Department of Maternal and Child Health requested INCAP's collaboration to develop an information subsystem to support planning activities in mother-child health and to monitor and evaluate continuing programmes. The programme would be directed toward rural communities with fewer than 2,000 inhabitants, who were considered priorities in the coverage expansion programme. The department also requested that, to strengthen local programming activities, the information should be representative of all Guatemalan health areas.
Based on agreements reached with national technicians and cooperation agencies in several working meetings to monitor and evaluate the national and subregional child survival plans, a sentinel surveillance methodology was proposed [4]; INCAP and UNICEF have been supporting evaluation activities in the region [3, 5]. Also, the characteristics of sentinel surveillance were outlined, defined in interagency coordination meetings as "the methodology that makes it possible to collect detailed, timely, and reliable information on specific aspects of the health situation and services, as a complement of the data produced by regular information systems" [6], supporting national planning and evaluation of specific interventions. It also included the timely identification of emergency situations and the sensitization of senior national officials regarding the type, magnitude, and distribution of the most striking nutritional and health problems. Data would be collected from a sentinel facility, which is the "basic observation element using the data collection methodology by sentinel facilities such as health centres, drugstores, schools, communities, families, etc." [6].
On the basis of these definitions, INCAP identified two types of sentinel surveillance: in areas or communities that would provide information on the health and nutrition status of the population; and in sites or facilities that would support the management of programmes and services. The latter would be focused on monitoring and evaluating processes involved in the delivery of health-care, as well as the interaction between health-care providers and users.
With the aid of the 1981 population census from the National Institute of Statistics, five communities of 1,000 ( + 250) inhabitants from the universe of communities of this size were randomly selected for each of the 24 health areas. The reason for choosing five communities from each health area was to obtain representative data on the health and nutrition states of these types of communities.
Sentinel communities
The Guatemalan sentinel surveillance subsystem, implemented in 1985 in 119 rural communities, focused on the longitudinal prospective study of these populations. With the results of the initial community survey (92,665 inhabitants), three microsurveys or rounds were carried out. Data on health, food, and nutrition were collected, which are useful for policy makers and evaluators of health interventions of Guatemalan mother-child health and nutrition.
The first round, establishment of the baseline, was developed by the end of 1985, and obtained general information on demographic and socio-economic characteristics, as well as specific data on the health and nutrition of women of reproductive age and children under 36 months. This information provided the baseline for evaluating the effects of the child survival five-year plan, initiated in the second half of 1986. Previously, the pre-coded form for data collection was tested in communities of the Totonicapán health area, as was all material to be used in training activities in data collection and analysis. Collection and primary analysis of baseline data of the 119 communities were performed by the technical health personnel (nurse aides, rural health technicians, sanitary inspectors, etc.) from each health area, supervised by INCAP's trainers. The collection, processing, and primary analysis phase was developed in each health area by 10 technicians within a 10-day period.
Baseline data generated information regarding the nature, magnitude, and distribution of the most important mother-child health problems of the populations living in each health area. Specific information was obtained on the size of the population for each sentinel community by age group and sex, prevalence of infectious diseases such as diarrhoea and acute respiratory infections, global rates on protein-energy malnutrition (assessed through weight-forage index), indirect mortality rates for infants and children under two years of age using the Brass method modified by Trussel [7], fecundity and fertility rates, and direct mortality and cause in children under one year. Information on morbidity, malnutrition, and mortality differentials by socio-economic groups and environmental characteristics was gathered. Finally, information on health services made it possible to assess the use rate of health services, as well as the implementation of technologies such as oral rehydration therapy and management of acute respiratory infections.
Results of the baseline evaluation were useful for reviewing certain characteristics of infant and preschooler morbidity, malnutrition, and mortality. As seen in figure 1 (see FIG. 1. Percentage of children with stunting (Z score -2.0), acute diarrhoea, and acute respiratory infections, sentinel communities, Guatemala 1985- 1986), the stunting prevalence (in terms of percentage of children with weight for age below two standard deviations of the WHO reference population) [8] follows the diarrhoea and acute respiratory infections trend, finding an increase during the first year and reaching its peak at the second year. Also, the characteristics of infant and preschool mortality were reviewed, finding that the main causes of death were infectious and respiratory diseases, diarrhoea, and neonatal pathology, including neonatal tetanus.
After wide dissemination of findings, a training workshop for the technical teams in the health areas was organized, in a joint effort of the Ministry of Health and INCAP, to promote the use of data collected for local programming activities in health and nutrition. At the same time, the use of the methodology was transferred to non-government organizations working in the health area (projects Hope and Concern), and to communities in the health areas not covered in the first round, in order to have representative data at this level. Also, some health area coordinators used the data as primary sources to define programme priorities at the department level.
The second microsurvey was developed with the aim of contributing to the assessment of the national vaccination rounds carried out in 1986. For this purpose, and based on the initial information collected in 1985, the lot quality assurance sampling (LQAS) methodology was used [9]. To assess vaccination adequacy, a smaller number of children, 29, was randomly selected for each sentinel community. They were surveyed by health personnel on vaccines applied during the national vaccination round. The information was jointly analysed with health officials responsible for monitoring and evaluating this activity, allowing identification of health areas where coverage in communities of approximately 1,000 was inadequate. Recommendations were drained to strengthen immunizations for future similar activities in populations where coverage was low.
During the third round, in 1987, information was collected on health knowledge and practices of a household sample. The issues addressed were: diarrhoeal diseases, acute respiratory infections, immunizations, breast-feeding, infant feeding, physical growth, pregnancy, delivery and prenatal care, access to and use of health services, and sources of communication and information, including mass media and local and personal communications. This round was part of the national survey at the community level on knowledge, attitudes, and practices in mother-child health within the assessment of the EPI-ORT national project. This information was used to identify accurately the degree of knowledge and behaviours of the population on mother-child health, to design activities in community education and communication, and to train strategies of health personnel. As in the case of previous surveys, data were collected by personnel from the health areas who were supervised by INCAP staff.
Finally, and also during 1987, the fourth round was carried out in 51 of the 119 sentinel communities. Information was collected on the nutrition status of children and women of reproductive age, as well as food availability at the household level. It is possible to compare data on the nutrition status of children under 36 months of age in the 51 sentinel communities with the baseline information collected in the same communities during 1985. Thus, as shown in figure 2 (see FIG. 2. Anthropometric and nutritional indicators in children under 36 months of age, sentinel communities, Guatemala 1985-1986 and 1987), the average of Z scores for weight for age and height for age, both of which reflect past nutritional problems, and weight for height, indicative of current nutritional problems, show that children in the second year of life have the highest prevalence of both. Furthermore, by comparing Z scores of weight for age estimated in 1985 with those of the 1987 average and the reference population (see FIG. 3. Nutrition status of children under three years of age, sentinel communities, Guatemala 1985-1986 and 1987), it is possible to assess nutritional deterioration between these two rounds. Similarly, it is possible to determine the changes that occurred during this period in the infant mortality rate, or the causes of death in children under three years. It must be stressed that, even though the infant mortality rate showed a slight decrease between 1985 and 1987, the malnutrition rate in children under three years increased significantly.
On the basis of these data, officials at the central and area level were trained, focusing on processing, analysing, and interpreting available information, and emphasizing its use in operative programmes to channel resources at the central and area levels.
Sentinel facilities
Sentinel facilities are any type of health centre where it is possible to collect information to measure structure, process, and impact variables, and assess quality of health care including user satisfaction and the relationship between user and health-care provider. The processes to be studied were resources, planning, organization and administration, personnel training, supervision, information system, supplies, and education of the population. Testing of this methodology is now in its final phase at the health care facilities, and a guide to be used in the healthcare delivery system is being prepared.
As in the case of sentinel communities, it is proposed that a random selection of facilities will be studied in an integral and prospective fashion. Data collection techniques are review of documents, observation, focus groups, guided surveys, and interviews of community and service personnel.
Recently, Thacker et al. [10] defined seven criteria for assessing the quality of health surveillance systems: sensitivity, specificity, representativity, opportunity, simplicity, flexibility, and acceptability. Cost and usefulness analysis should also be included. The sentinel surveillance subsystem implemented in Guatemala adheres to these characteristics. During the 1985-1987 period, this subsystem was the only source of representative data on the nature, magnitude, and distribution of health and nutrition problems in rural communities of approximately 1,000 inhabitants. Prospective data collection in these communities enabled us to assess changes in health and nutrition status that remained undetected through other information sources. Striking features of the subsystem are its simplicity and low cost. Therefore, the system can be organized and implemented at the local level, contributing to the development of primary data analysis at this level. The results would be readily available to programme, monitor, and evaluate activities, as well as to detect groups at risk.
INCAP has transferred the methodologies developed for the surveillance subsystem to national technical counterparts through information dissemination, direct technical assistance, training of personnel, and development of collaborative research. However, the transfer of methodologies does not necessarily lead to generalized use and application, and with few exceptions, they have not been implemented in Guatemala, especially by decision makers at the central, intermediate, and local levels. This is consistent with the fact that, for almost all information and surveillance systems, availability of information does not imply that it will necessarily be used for decision-making purposes. A relevant future line of operational research should be related to identifying information requirements and delivery to decision makers, to ensure use at the technical and political levels.
This study was financed by the Regional Office for Central America and Panama of the United States Agency for International Development, and by the Institute of Nutrition of Central America and Panama. The study was developed during 1985-1987 with the collaboration of counterparts from the general office of Health Services of Guatemala, Dr. Francisco Arenas, Dr. Juan Jose Arroyo, Dr. Jose Lima, and Dr. Ricardo Lopez Urzua.
