Nutritional status surveillance in El Salvador.

Bulletin of the World Health Organization, SS (2): 327-332 (1980) Nutritional status surveillance in El Salvador* FREDERICK L. TROWBRIDGE1 & HARRISON C. STETLER 2 There has been increasing recognition of the potential usefulness of continuous nutritional surveillance based on routinely collected data to provide nutrition information for evaluation and planning. Such an activity has been developed within the Ministry of Health of El Salvador. Surveillance indicators have been developed based on routinely reported datafrom the countrywide system ofoutpatient clinics operated by the Ministry of Health. These indicators have been systematically evaluated by comparison with ongoing field survey measurements. An indicator based on the weight-for-age status ofpreschool children attending clinics has been shown to be the most reliable ofthosepresently available for monitoring nutritional status. Surveillance data are being tabulated every 6 months to show seasonal and long-term changes in malnutrition and to define important regional and urban/rural differences in nutritionalstatus. It may bepossible in thefuture to monitor data from outside the health sector such as food prices, agricultural production, and family income; such data may give insights into the causes of the malnutrition documented by the nutritional status surveillance system. Similar tabulation ofroutinely reported data toform indicators for nutritional status surveillance may prove useful for health planning and evaluation in other countries. In the past few decades many nutritional surveys have been carried out in countries in the developing world in an effort to gain baseline data on the magni- tude and causes of malnutrition. More recently a simplified methodology for the field assessment of nutritional status based on anthropometry has been developed by the Center for Disease Control (1). This has been applied in a number of countries (2-5). Such studies form an important first step towards the defi- nition and quantification of nutritional problems. However, in order to define changes in the nutritional status of a population with time, this type of survey must be repeated at intervals. Such surveys, even using simplified methodology, are expensive, require considerable staff and logistical support, and the data from the survey may not be available for up to a year or more after planning for the survey is begun. For these reasons the concept of an ongoing nutritional surveillance system is more appealing. At least in theory, surveillance can provide information less expensively and the data could be available on a * From the Bureau of Tropical Diseases, Center for Disease Control, Department of Health, Education and Welfare, Atlanta, GA 30333, USA, in collaboration with the Divisions of Maternal- Child Health, Statistics, Epidemiology, and Planning of the Ministry of Health of El Salvador. Funding was provided, in part, by the Agency for International Development. Requests for reprints should be addressed to Dr Trowbridge at Johns Hopkins University. I Assistant Professor, Department of International Health, Johns Hopkins University, School of Hygiene and Public Health, 615 N. Wolfe Street, Baltimore, MD 21205, USA. 2 Medical Epidemiologist, Bureau of Tropical Diseases, Central America Research Station, El Salvador. continuous, routine basis within a few weeks or months after collection. The need for developing a methodology for nutri- tional surveillance has been emphasized in the 1976 report of the Joint FAO/UNICEF/WHO Expert Committee (6). This report recommends that such systems be multisectorial in design, be organized within the framework of existing information chan- nels, be based on a standard set of indicators selected after careful assessment of their validity, and be aimed towards vulnerable population groups. Despite the recognized need and the definition of the basic re- quirements for a nutritional surveillance system, implementation has proceeded slowly, mainly because of the need to develop and test practical method- ologies and to adapt the theoretical design of surveil- lance to local economic, political, and organizational realities. The experience gained in the development of a nutritional status surveillance activity within the Ministry of Health of El Salvador is described here. This activity was developed as a simple, practical system for the routine collection of health and nutri- tional status information for use in health planning and evaluation. Although a multisectoral design for conducting nutritional surveillance has been recom- mended (6), the surveillance activity in El Salvador has been developed solely within the health sector. This approach was taken for practical reasons. It was felt that nutritional surveillance could best be devel- oped in a step-by-step manner and that the develop- 3951 -327- F. L. TROWBRIDGE & H. C. STETLER ment of a means of monitoring nutritional status within the health sector was a logical first step. Surveillance of causal factors in malnutrition based on data from other sectors such as agricultural produc- tion, market prices for staple foods, and income may be added later as resources and opportunities permit, to form a more comprehensive surveillance system. STEPS IN PROJECT DEVELOPMENT Recognition of need Interest in the surveillance of nutritional status in El Salvador began in 1975 in the context of formulating a new five-year health plan. It was recognized at that time that there was a need for better definition and monitoring of the location and severity of nutritional problems. The data available included the nutritional survey carried out by the Institute of Nutrition of Central America and Panama (INCAP) in 1965 (7) and data from the outpatient services of the clinics run by the Ministry of Health. However, these data did not define current nutritional status. The INCAP study was 10 years old and the sample was not ad- equate to define regional differences. The Ministry of Health data was delayed several years in data proces- sing. Moreover, the relationship between the inci- dence of reported cases of malnutrition among clinic attenders and the actual nutritional status of the general population was uncertain. These factors indicated the need for new sources of nutritional status information. The advantages of developing a surveillance mechanism to provide a continuing flow of information in the future were recognized. Evaluation of indicators The first step in the development of surveillance was the definition of possible indicators of nutritional status from the data routinely collected in the out- patient clinic system of the Ministry of Health. Two indicators were proposed for use in the preschool age group: the percentage of children attending a clinic who were clinically diagnosed as malnourished by the clinic physicians (clinical diagnosis indicator), and the percentage of children in grades II and III of weight- for-age deficit as defined by Gomez et al. (8) (weight- for-age indicator). The choice of these indicators was based on the availability of clinical diagnosis and weight data. Height data were not routinely available so that it was not possible to form height-for-age or weight-for-height indicators. These indicators would have offered advantages for distinguishing chronic malnutrition as reflected in stunting of height growth from more acute malnutrition as reflected in weight- for-height deficit (9). It was decided to evaluate the clinical diagnosis and weight-for-age indicators that could be derived from available data by comparing them with field measure- ments of malnutrition from anthropometric surveys in various regions of El Salvador. The evaluation was undertaken by the Central America Research Station of the Center for Disease Control (CDC) in collabor- ation with the Ministry of Health of El Salvador. The study was carried out in January and February 1976 in the Department of La Paz (population 200 000). This Department included three distinct geographical sub- regions: mountains, hills, and coastal plains. Records from the clinics within each of these three subregions were reviewed to calculate the clinical diagnosis and weight-for-age indicators. Field surveys were carried out using standard methods (1) which measured height, weight, and arm circumference in approxi- mately 1400 children in each subregion and simul- taneously assessed fertility and mortality (10) as well as dietary intake using a 24-hour recall method in a subsample of approximately 100 preschool children per subregion. Anthropometric data analysis was made using the Harvard standard (11). The results of this study are summarized in Table 1. It may be seen that the anthropometric indicators, weight-for-age, weight-for-height (P<0.001), and arm circumference (P<0.05), as well as dietary and mortality data, suggest a higher level of malnutrition and mortality among children in the coastal region. The weight-for-age indicator derived from clinic data also indicated a higher level of malnutrition in the coastal region (P<0.01). However, the indicator based on clinical diagnosis appeared to be far more variable and did not conform with the pattern of increased malnutrition in the coastal region suggested by the weight-for-age indicator and the field measure- ments. The results of the study suggested that the weight- for-age indicator derived from clinic data might prove more reliable for use in surveillance than the indicator based on clinical diagnosis. However, it was felt that further evaluation of both indicators in additional regions was required. Therefore, a more extensive study was undertaken in collaboration with INCAP in which the same indicators from a sample of approxi- mately 10 clinics in each of 5 geographical regions of El Salvador were compared with field survey measure- ments in the same regions. Anthropometric data were analysed using the Iowa standard employed by INCAP. The results of this study will be reported in detail elsewhere (F. L. Trowbridge and V. Valverde, un- published data). However, the most pertinent findings are summarized in Table 2. It may be seen that the field survey results indicated relatively high levels of malnutrition, as defined by weight-for-age, in three of 328 NUTRITIONAL SURVEILLANCE IN EL SALVADOR Table 1. Evaluation of surveillance indicators in children aged 1-4 years: La Paz Department, El Salvador, 1976 Surveillance indicators Field measurements Clinical diagnosis Weight-for-age % % % % < 75% < 90% <12.5 cm adequacy of Infant No. % No. % of weight- of weight- arm cir- dietary intake mortality Geographical exam- mal- exam- <75% of for-age for-height cumference (per 1000 region ined nourished ined standard standard standard Energy Protein live births) Mountain 2972 24.2 596 28.8 15.9 9.1 2.7 61.9 111.3 78 Hills 5851 8.2 1557 25.4 15.6 10.4 3.3 65.6 119.8 79 Coast 887 11.2 460 33.7b 16.3 14.6c 4.48 54.1 95.5 98 Significance levels by chi-square analysis: a P<0.05 b p< 0.01 C p< 0.001 Table 2. Evaluation of surveillance indicators in children aged 1-4 years. El Salvador, 1976 Surveillance indicators Field survey Clinical diagnosis Weight-for-age Weight-for-age No. % No. % No. % Region exam- mal- exam- <75% of exam- <75% of ined nourished ined standarda ined standarda Central 478 12.8 403 41.1 1047 26.4 Coffee 717 26.6 295 42.7 1043 23.2 Northern 197 38.1 362 47.5 1447 23.1 Coastal 374 26.5 237 28.4 1501 17.3 Urban 264 20.8 272 27.6 1369 15.5 8 P < 0.001 by chi-square analysis. the regions and significantly lower levels in the other two. The difference between the high and low preva- lence regions was highly significant (P<0.001). The weight-for-age surveillance indicator correlated well with field measurements, identifying the same three regions as having relatively high levels of malnutrition and the same two regions as having relatively low levels (P<0.001). The clinical diagnosis indicator failed to reveal the same pattern. Based on the results of these field evaluations it was concluded that weight-for-age would be the more reliable indicator of preschool child nutritional status. Consequently, in January 1977, supplementary columns were added to the clinical data reporting form of the Ministry of Health for recording the weight of all preschool children and the age in years and months, so that weight and age data would be routinely collected and available for calculation of the weight-for-age indicator. The indicator based on the clinical diagnosis of malnutrition would continue to be observed, but with a recognition of its apparent limitations. Finally, the need was recognized for a continuing, systematic evaluation of the surveillance indicators by comparison with additional field surveys in the same regions once routine data tabulation from the surveil- lance activity was under way. This continuing field evaluation was initiated in January 1978, with 2 additional surveys planned in the same regions. A further study is also planned to evaluate the perform- ance of weight-for-height and height-for-age indi- cators in pilot clinics where height measurement will be added to the routine collection of weight and clinical diagnosis data. Design of the surveillance data collection system The data base utilized for the surveillance activity in El Salvador is the routine monthly reports of all out- patient physicians' consultations in the approximately 275 government health facilities. Over two million consultations occur each year, of which approxi- mately 25% or 500.000 involve children less than 5 years of age. About 350 000 of these visits by children under 5 years of age are first visits for a given diagnosis, the remainder being follow-up visits for previously diagnosed disease. This extensive data base is collected by means of a standard form which is completed by the physician and submitted monthly to the statistical unit of the Ministry of Health for analysis. The form provides one line for the reporting of each individual consul- tation, including data on weight (for children 0-5 329 F. L. TROWBRIDGE & H. C. STETLER years old only), age, sex, urban or rural residence, date of consultation, primary diagnosis, secondary diagnosis, and whether the consultation is the first or a follow-up visit for the indicated diagnosis. Only "first visit" data are used for surveillance. Once received by the statistical unit of the Ministry of Health, a sample of one week of data is selected from each month. In this one-week sample the diag- noses are coded using the 1965 WHO International Classification of Diseases (12) and all data are punched on to IBM cards for computer analysis. Data processing, from initial recording in the clinic to avail- ability on tape for analysis, could, in theory, take as little as 2-3 months: I month for collecting the data from the field and 1-2 months for coding, key-punch- ing, and passing on to tape. Further economy of time could be gained by passing the data for the sample week directly from the clinic forms on to tape or cas- sette. However, in practice, delays of six months or more have occurred due to the limited resources for data processing. Although it is planned that all data will eventually be analysed in El Salvador, the devel- opment phase of the surveillance activity has required the use of the computer facilities of the Center for Disease Control in Atlanta, Georgia for data tabulation. The reporting of the surveillance data has been initially planned as a six-monthly report produced by Table 3. Percentage of children aged 1-4 years examined with Gomez Grade II or IlIl malnutrition by month and urban/ rural residence-El Salvador, 1977 % malnourished No. Month examined Urban Rural Total January 2 314 18.2 37.2 25.2 February 4 429 20.1 37.5 28.2 March 4 163 18.9 34.9 26.0 April 4 202 21.2 33.5 27.2 May 5 460 18.5 33.8 25.0 June 6 296 23.9 40.5 31.2 January-June 26 864 20.5 36.4 27.5 July 7 737 23.2 41.7 32.2 August 6 118 22.1 38.7 30.1 September 4 739 19.8 35.8 27.1 October 4 521 18.3 32.8 24.8 November 5 134 16.1 33.0 23.6 December 4 359 16.3 30.0 22.4 July-December 32 608 19.7 36.3 27.4 Annual total 59 472 20.0 36.3 27.4 the Nutrition Section of the Division of Maternal and Child Health of the Ministry of Health. As it becomes possible to process the data more rapidly, more fre- quent reports may be possible. Data for the reports are presented in a standard format to permit easy com- parison with previous data. The tables present the nutritional status indicators by region and depart- ment, by month, and by urban/rural residence within specific age groups. A summary of the weight-for-age indicator by month for children of 1-4 years for 1977 is presented by way of illustration (Table 3). This table permits the definition of seasonal and urban/rural differences in malnutrition. Other tabulations of the same data are used to define regional and age-group differences. Similar tabulations of data based on the clinical diag- nosis indicator are also prepared for comparison with the weight-for-age data. In addition, data on the clinical diagnosis of diarrhoeal disease are routinely tabulated for comparison with nutritional indicators, because of the important association between diar- rhoeal disease and nutritional status (13). APPLICATION TO PLANNING The nutritional status surveillance activity has been developed step-by-step within the Ministry of Health with the objective of providing reliable information on nutritional and health status in El Salvador for use in health planning. The intention has been to provide more continuous and updated information on nu- tritional status in order to stimulate an awareness of nutritional problems and to better define changes and trends in nutritional status in specific regions and population groups. Although the surveillance activity has been developed within the health sector, the infor- mation collected is being actively shared with the Nutrition Office of the Ministry of Planning where it can be appropriately used for defining multisectorial approaches to reduction of the severity of nutritional problems. The surveillance activity provides information that is potentially useful in showing trends in nutritional and health status by regions, seasons, and specific population groups. However, the data that are rou- tinely collected within the health sector do not provide information on factors such as food prices, agri- cultural production, and family income which might give insights into the causes of the malnutrition observed. It is hoped that, in the future, information from other sectors can be combined with the nu- tritional status data to provide a more complete, routinely collected information resource that could justifiably be termed a "nutritional surveillance system". 330 NUTRITIONAL SURVEILLANCE IN EL SALVADOR 331 ACKNOWLEDGEMENTS The evaluation of surveillance indicators and the development of surveillance activities have depended upon the full support and collaboration of the professional personnel of the Divisions of Maternal-Child Health, Statistics, Epidemi- ology, and Planning of the Ministry of Health of El Salvador. RESUME SURVEILLANCE DE L'ETAT NUTRITIONNEL EN EL SALVADOR Les methodes d'evaluation de 1'etat nutritionnel sur le terrain ont certes et grandement simplifiees ces dernieres ann&es, mais il n'en faut pas moins consacrer beaucoup de temps et d'argent a la collecte de donn&s qui peuvent perdre rapidement leur actualite. Ce fait a conduit a mieux apprecier l'int&et pour les planificateurs, qui doivent disposer de renseignements a jour, des donn&es recueillies au moyen d'actes sanitaires de routine. Le Ministere de la Sante d'El Salvador a collabore a l'etablissement d'indicateurs fondes sur les donnees enregistr&s au jour le jour dans tous les centres de consulta- tion rattaches au systeme national de protection sanitaire place sous l'autorite du Ministere, et ces indicateurs ont e systematiquement evalues en les comparant aux donnees recueillies dans les enquetes en cours sur le terrain. La posi- tion sur l'echelle poids/age s'est revele constituer l'indica- teur le plus fiable pour appr&cier l'etat nutritionnel des enfants d'age prescolaire frequentant les dispensaires; les etudes men&es dans 5 grandes zones geographiques ont montre que cet indicateur permettait de definir des zones a prevalence forte ou faible de malnutrition tout a fait comparables a celles identifiees au moyen des enquetes de terrain portant sur des &chantillons aleatoires de la popu- lation. En revanche, le pourcentage d'enfants d'age pre- scolaire pour lesquels avait e pose un diagnostic clinique de malnutrition lors d'une consultation dans un dispensaire n'a pas presente une correlation satisfaisante avec les constata- tions faites a l'occasion des enquetes de terrain. L'indicateur reposant sur le diagnostic clinique presente cependant l'avantage de renseigner sur les variations saisonnieres et les differences entre zones urbaines et rurales. Les series de donnees qui viennent d'etre mentionn&es en tant qu'elements principaux de la surveillance (soit l'indica- teur poids/Age et l'indicateur fonde sur le diagnostic clinique) sontsoumises tous les six mois A une tabulation qui fait ressortir l'evolution saisonniere ou A long terme de l'etat nutritionnel, ainsi que les variations de celui-ci en fonction de la residence (urbaine ou rurale). Aux fins de cette analyse, on preleve dans les rapports mensuels des dispensaires un echantillon d'une semaine dont les donnees sont codees et reportees sur des cartes perforees. Le systeme de surveillance de l'etat nutritionnel applique en El Salvador repose donc uniquement, pour le moment, sur les donnees provenant du secteur de la sante. Cette approche a e dictee par des raisons pratiques. On a estime en effet que la surveillance nutritionnelle etait une activite A developper par &tapes et que la mise au point d'une methode de surveil- lance au sein dudit secteur constituait logiquement la premiere etape. Quant A l'identification des facteurs A l'origine de la malnutrition A partir de donnees provenant d'autres secteurs-production agricole, prix sur le marche des aliments essentiels, montant du revenu-elle viendra s'ajouter aux indicateurs sanitaires pour constituer un systeme de surveillance plus complet lorsque les ressources et les possibilites le permettront. REFERENCES 1. MILLER, D. C. ET AL. Simplified field assessment of nu- tritional status in early childhood: practical suggestions for developing countries. Bulletin of the World Health Organization, 55: 79 (1977). 2. BRINK, E. W. ET AL. Nutritional status of children in Nepal, 1975. Bulletin of the World Health Organ- ization, 54: 311 (1976). 3. BRINK, E. W. ET AL. Sri Lanka nutrition status survey, 1975. International journal of epidemiology, 7: 41 (1978). 4. HOGAN, R. C. ET AL. Sahel nutrition surveys, 1974 and 1975. Disasters, 1: 117 (1977). 5. STETLER, H. C. ET AL. Nutritional status of preschool children in Togo, 1976-77. Bulletin of the World Health Organization, (in press). 6. WHO Technical Report Series No. 593, Geneva, 1976 (Methodology of nutritional surveillance: report of a Joint FAO/UNICEF/WHO Expert Committee). 7. Institute of Nutrition of Central America and Panama/ Interdepartmental Committee on Nutrition for National Development. Nutritional evaluation of the population of Central America and Panama: regional summary. Washington, DC, US Department of Health, Education and Welfare, 1972 (DHEW Publication No. (HSM) 72-8120. 8. GOMEZ, F. ET AL. Mortality in second and third degree malnutrition. Journal of tropical Pediatrics, 2: 77 (1956). 332 F. L. TROWBRIDGE & H. C. STETLER 9. WATERLOW, J. C. Classification and definition of protein-calorie malnutrition. British medical journal, iii: 566 (1972). 10. UNITED NATIONS. Manual IV: Methods of estimating basic demographic measures from incomplete data. New York, United Nations, 1967, pp. 73-74. 11. STUART, H. C. & STEVENSON, S. S. Physical growth and development, In: Nelson, W., ed., Textbook of pediatrics, 7th ed., Philadelphia, Saunders, 1959, pp. 12-61. 12. Manual of the international statistical classification of diseases, injuries and causes of death, 1965 Revision, Vol. I, Geneva, World Health Organization, 1967. 13. TROWBRIDGE, F. L. & NEWTON, L. Seasonal changes in malnutrition and diarrheal disease among preschool children in El Salvador. American journal of tropical medicine and hygiene, 28: 136 (1979).