Results of nutritional status surveillance in El Salvador, 1975 - 77*

BuLlatin of the World Helth Organizwtion, ES (3): 433 -440 (1982) Results of nutritional status surveillance in El Salvador, 1975 - 77* FREDERICK L. TROWBRIDGE1 & HARRISON C. STETLER2 Nutritional status surveillance data based on the clinical diagnosis ofmalnutrition and on weight-for-age, as well as diarrhoeal disease data for preschool age children attending government health clinics in El Salvador are presented for a 3-year period (1975- 77). Surveillance results indicated consistently higher rates of clinical malnutrition and weight- for-age deficit in rural children as compared with urban children, and higher malnutrition rates in children I - 4 years ofage as compared with infants less than 1 year old. Consistent seasonal increases in malnutrition were observed that were most pronounced in older pre- school children (I - 4 years) in rural areas. Seasonal peaks in malnutrition consistently followed 1- 2 months after the major seasonal peak in diarrhoea at the onset of the rainy season, suggesting that diarrhoea may play a role in the etiology of malnutrition. A secon- dary seasonalpeak in diarrhoea in the cooler, dry season was mostprominent in infants but was not related to increased malnutrition. These results indicate that nutritional and health status indicators formedfrom data collected regularly at health centres can demonstrate consistent agegroup, urban/rural, andseasonal differences in nutritional status that may be useful in identifying risk groups and in monitoring nutritional changes for planning and evaluation purposes. Nutritional surveillance has been proposed as a means of monitoring malnutrition for the purposes of planning and evaluation (1). However, little experi- ence has been reported from established surveillance systems. In El Salvador, a nutritional status surveil- lance system has been monitoring indicators of nutritional and health status based on data from Min- istry of Health outpatient clinics since 1975 (2). In order to keep the system as simple and practical as possible, the surveillance system has collected only nutritional and health status data from the health sector rather than undertaking multisectoral data collection. Evaluation has been made of the performance of surveillance indicators by com- parison with field survey measurements (3). Three years of surveillance data are now tabulated and form the basis of this report. * This study was carried out in coliaboration with the Divisions of Maternal-Child Health, Statistics, Epidemiology, and Planning, of the Ministry of Health ofEl Salvador. Funding was provided, in part, by the Agency for International Development and the Biomedical Research Support Grant Fund of the National Institutes of Health. Requests for reprints should be addressed to Dr F. L. Trowbridge at Johns Hopkins University. ' Associate 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, Center for Infectious Diseases, Centers for Disease Control, Atlanta, GA 30333, USA. Data from the surveillance system provide a basis for the further evaluation of the performance of surveillance indicators in identifying age group, urban/rural, seasonal, and regional differences in the patterns of malnutrition and diarrhoeal disease. The surveillance data can also be useful for planning purposes both by identifying high-risk populations and by monitoring changes in nutrition and health status over time. MATERIALS AND METHODS The system for collecting and tabulating data for nutritional status and disease surveillance in El Salvador has been described previously (2, 3). Surveillance is based on both data from the routine reporting of weight-for-age in preschool children and diagnosis by physicians of malnutrition, diarrhoea, respiratory disease, or other illnesses in one of the approximately 275 outpatient clinics run by the Ministry of Health in El Salvador. About 2.5 million medical consultations occur each year, of which approximately two-thirds are first consultations for a given diagnosis, while the remainder are visits for follow-up examinations. Data collected for all consul- 413 -433- F. L. TROWBRIDGE & H. C. STETLER tations include the patient's age, sex, department of residence, urban or rural residence, primary diag- nosis, secondary diagnosis, and, since 1977, weight in pounds or kilograms and age in months for children less than 5 years of age. These data are transmitted monthly to the central statistical unit of the Ministry of Health where the data for 1 week each month are tabulated by computer for surveillance purposes. Only new diagnoses are tabulated, thus excluding persons who return for a follow-up visit for the same diagnosis. The indicators of malnutrition used in the tabula- tions include the percentage of clinic attenders clini- cally diagnosed as malnourished, based on the clinical impression of moderate to severe muscle wasting, depleted fat stores, retarded growth for age, and/or vitamin deficiency signs. Since January 1977, with the beginning of routine recording of weight, an additional malnutrition indicator has been based on the percentage of children under 5 years of age attend- ing the clinic with weight-for-age less than 75% of the median of the National Center for Health Statistics (NCHS) reference population (4). The indicator of diarrhoeal disease used is the percentage of clinic attenders with clinically diagnosed dysentery or diarrhoeal disease. A similar indicator of respiratory disease was defined, but did not appear to have any relationship with reported malnutrition. The 45 40 LU = 35 ) 302-J > 25 -J ' 20 -J " 15 1-1 L-) = 10 5 Au indicators are expressed as a percentage of clinic attenders rather than using the actual numbers of cases in order to facilitate comparisons among calen- dar months, age groups, and regions with unequal numbers of clinic attenders, and between urban and rural residents. The data were edited to exclude from tabulation cases with missing age, sex, place of residence, or diagnostic code information. In the 1977 data set, which included weight data, weight-for-age results less than 30%7 or greater than 150%o of the NCHS median for age were considered biologically implaus- ible, and were excluded as probably representing errors in weight measurement or age assessment. This procedure excluded about 1.5%o of weight-for-age values. RESULTS Surveillance data were tabulated from a sub- sample of approximately 300 000 outpatient clinic visits in 1975, 351 000 visits in 1976, and 370 000 visits in 1977. This extensive data base included 83 000- 109 000 visits per year by children 0 - 4 years of age (approximately 30%o), 57 000- 72 000 visits by chil- dren 5- 14 years of age (approximately 19Go), 44 000 - 55 000 visits per year by adult males 15 years __/:__/--R--//~\/~/ o_/./\/ LEGUEND: URBAN RURAL Fig. 1. Percentage of infants under 1 year of age with clinically diagnosed malnutrition, by month in urban and rural districts, El Salvador, 1975-77. IJ F M A M J J A S O N D J F M A M J J A S O N D J F M A M J J A SO N D 1975 1976 1977 MONTH AND YEAR 434 NUTRITIONAL SURVEILLANCE IN EL SALVADOR LEGEND: URBAN -- RURAL A M J J A S 0 N D A M J J A S 0 N MONTH AND YEAR D J F 1977 M A M J J A S 0 N D Fig. 2. Percentage of children aged 1 - 4 years with clinically diagnosed malnutrition, by month in urban and rural districts, El Salvador, 1975-77. of age and older (approximately 15/o), and 114 000- 134 000 visits by adult females 15 years of age and older (approximately 36%). Visits were about evenly divided between persons living in urban (55%) and rural (45%) areas. Clinical malnutrition The diagnosis of malnutrition was slightly, but consistently, more frequent among rural than among urban infants less than 1 year of age (Fig. 1). A slight seasonal trend was also observed with malnutrition tending to be higher in the months of June-October and lowest in November and December. Both urban/ rural differences and seasonal changes were more prominent and consistent among preschool children 1- 4 years of age than among infants (Fig. 2). The percentage malnourished among rural children 1 - 4 years of age was consistently higher than among their urban counterparts. Distinct seasonal peaks in malnutrition were observed each year among rural, preschool children in the months of July-September. Seasonal peaks in malnutrition were also seen in urban preschool children but were less consistent. An overall increase in the frequency of clinically diagnosed malnutrition was observed in infants and older preschool children in 1977. Weight-for-age deficit The routine collection of weight and exact age data, which was begun in 1977, permitted the calculation of weight-for-age statys for preschool children. Tabu- lation of these data (Fig. 3) indicated that low weight- for-age was observed more frequently in rural than urban children, particularly in the 1 - 4-year age group. A seasonal pattern in the occurrence of weight- for-age deficit was also observed with a peak in the months of June and July. Seasonal variation was more prominent in 1- 4-year-old children than in infants, and in rural than in urban areas. The peak in weight-for-age deficit generally coincided with or slightly preceded the seasonal peak in clinically diagnosed malnutrition. Preliminary tabulation of weight-for-age data for 1978 (not illustrated) indicated a repetition of the age group, urban/rural, and seasonal trends observed in 1977. The relatively high frequency of weight-for-age deficit observed in the early months of 1977, particu- larly among rural children, did not correspond with the pattern of reported clinical malnutrition. Diarrhoeal disease Distinct and consistent seasonal patterns of diar- rhoea incidence were observed in all three years and, 451 40 Lu X 35 0-4 3 30 >- 25 -J (-.4z 20 -J ,_ 15 Lii L.) , 10LU 5 0 J F M 1975 J F M 1976 435 F. L. TROWBRIDGE & H. C. STETLER INFANTS <1 YEAR CHILDREN 1-4 YEARS / \ 5. .I ._ .-6 .'/\! \ *N,". .%5% % / * *55. .,, LEGEND: URBAN * RURAL u J F M A M J J A SO N D J F M A M J J A S O N D 1977 1977 MONTH AND YEAR MONTH AND YEAR Fig. 3. Percentage of infants under 1 year and children 1 - 4 years of age with weight-for-age less than 75% of the NCHS median, by month in urban and rural districts, El Salvador, 1977. 45 40 ~35/ \ =20- - 15 LLi lo10 LEGEND: URBAN RURAL J F M A M J J A S O N D J F M A M J J A S O N 1975 1976 DJ F M A M 1977 J J A S O N D MONTH AND YEAR Fig. 4. Percentage of infants under 1 year of age with diarrhoeal disease, by month in urban and rural districts, El Salvador, 1975-77. 40 35 30 25 20 15- 10 Z: U- CD V LL P.D 0 2 Li- := 3c I-. 4-) LIi al- L- 5 n 436 v NUTRITIONAL SURVEILLANCE IN EL SALVADOR I'I LEGEND: lJRBAN * RURAL J F M A M J J A 1975 S O N D J F M A M 1976 J J A S 0 N D J F M 1977 A M J J A SO N D Fig. 5. Percentage of children aged Salvador, 1975-77. MONTH AND YEAR 1 - 4 years with diarrhoeal disease, by month in urban and rural districts, El in different degrees, in all age groups, including adults. The most prominent seasonal patterns of diarrhoeal disease were observed in infants and in children 1 - 4 years age (Fig. 4 and 5). Both of these age groups showed sharp seasonal peaks in June of each year for both urban and rural children. Particularly among infants a distinct second and smaller seasonal peak occurred in the months of December-February. The secondary peak was seen in both urban and rural infants (Fig. 4), but in the I - 4 years of age group (Fig. 5) it was more promi- nent among rural children. The proportions with diarrhoeal disease were higher in infants than in older preschool children, and rural children had more diar- rhoea than those from urban areas. The peak in diar- rhoeal disease generally preceded by 1 - 2 months the peak in clinically diagnosed malnutrition, consistent with the possibility that diarrhoea is a major factor in the evolution of clinical malnutrition in these children. Regional differences Data were tabulated to examine regional differ- ences in clinical malnutrition and weight-for-age deficit (Table 1). Each of the regions showed the urban/rural differences that were observed for the data as a whole. Also the overall tendency towards an increase in reported clinical malnutrition in 1977 was observed in each of the regions, with the exception of health region 1. The level of weight-for-age deficit in regions 1 - 4 in 1977 was similar and suggests that the apparent interregional differences in clinically diag- nosed malnutrition may be based on differences in clinical judgement rather than on real differences in nutritional status. Clinically diagnosed malnutrition was relatively high in 1977 in region 5, the metro- politan region consisting principally of the capital city of San Salvador plus some adjacent rural areas. However, weight-for-age deficits were relatively low suggesting that the high level of clinically diagnosed malnutrition was a problem of overdiagnosis. Minor year-to-year variations were observed in the regional reporting of diarrhoea (not illustrated) but no consis- tent regional pattern was observed in the occurrence of this diagnosis. DISCUSSION Reliability of indicators The indicators formed from surveillance data provided a consistent definition of seasonal, urban/ rural, and age group patterns of malnutrition and 45 40 cn a 350-, _. 30 =- 25 0- = 20 ,- 15 , "I-' 10 0- 5 ff a - 437 ._ 6, 'S." ,0-110/ F. L. TROWBRIDGE & H. C. STETLER Table 1. Percentage of children 1 - 4 years of age with clinically diagnosed malnutrition and low weight-for-age, by health region, El Salvador, 1975 - 77 Clinically diagnosed malnutrition Low weight-for-age 1975 1976 1977 1977 Health region urban rural urban rural urban rural urban rural 1 10.3 22.8 15.0 26.2 14.0 21.2 24.6 38.1 2 17.9 24.1 20.3 23.9 23.3 32.7 26.4 38.4 3 12.9 20.6 10.5 19.2 19.0 23.3 27.6 39.1 4 9.4 12.7 13.0 18.3 19.4 29.8 26.9 38.1 5 15.4 22.0 14.4 28.6 24.3 43.6 15.7 35.5 diarrhoeal disease. Malnutrition indicators showed consistent seasonal peaks which corresponded with observations of seasonally increased malnutrition in field surveys (5, 6). Although the surveillance indi- cators for malnutrition are based on data from self- selected clinic attenders, field evaluation studies have shown that the levels of weight-for-age deficit observed in the surveillance data from a given region are indicative of the relative level of weight-for-age deficit in the general population of the same region (2, 3). Thus, both previous field studies and the overall consistency of surveillance data over the three years reported here provide a basis for confidence that the surveillance data are reflecting real malnu- trition patterns. Definition by the surveillance data of seasonal patterns and urban/rural differences was particularly sharp for diarrhoeal disease. The same seasonal pattern has been observed in other studies in El Salvador (5) and in other countries in relation to the onset of the rainy season (7, 8). Moreover, the secondary peak in diarrhoea incidence that occurred primarily in infants in the months of December- February may correspond with similar observations in other Central American countries such as Hon- duras, and could represent diarrhoea caused by path- ogens different from those responsible for the main summer diarrhoea peak, for example, by rotavirus (M. Levine, personal communication, 1980). These observations of similar diarrhoea patterns in neigh- bouring countries lend support to the validity of observed seasonal changes in diarrhoea defined by the surveillance mechanism. Although the surveillance data have provided generally consistent results, there has been some vari- ability in the data and this merits comment. The clinical malnutrition indicator showed reasonably consistent seasonal peaks but also indicated a tendency to rise in 1977. This rise may be an artefact possibly related to an increased awareness of malnutrition among physicians stimulated by the collection of weight-for-age data that started in Janu- ary of 1977. The overall level of weight-for-age deficit in the 1977 surveillance data is similar to the levels observed in 1975 and 1976 (3), indicating that the apparent rise in clinically diagnosed malnutrition in 1977 is not supported by any corresponding rise in weight-for-age deficit among clinic attenders. The relatively high weight-for-age deficits observed in the early months of 1977 may also be an artefact. In the first two months of data collection the percentage of children weighed was relatively low (30- 40(0/) and there may have been a bias towards weighing the chil- dren who appeared to be clinically malnourished. By March 1977,70 - 800o of children were being weighed so that any selection bias would have had a relatively smaller effect on the data. The similar levels of malnutrition in health regions 1 - 4 does not mean that there are no geographically based differences in nutritional status in El Salvador. On the contrary, field studies in ecologically distinct regions of the country have shown significant differ- ences in nutritional status related to geographical features, life styles, and patterns of economic activity (9). However, the boundaries of the administrative regions defined by the Ministry of Health cut across the ecological regions and so tend to obscure the measurable differences in nutritional status and disease patterns. The relatively low level of weight- for-age deficit in the urban population of region 5, the metropolitan region consisting of the capital city and adjacent areas, is probably a real difference, consis- tent with the relatively low level of weight deficit found in field surveys in urban San Salvador (5, 9). 438 NUTRITIONAL SURVEILLANCE IN EL SALVADOR Application ofsurveillance information Data from the nutritional status surveillance system have been useful in defining the magnitude and distribution of malnutrition and related disease problems in El Salvador. The data identify age groups, particular areas, and seasons with the highest incidence of nutritional deficiency so that interven- tions such as the provision of supplementary food can be organized more effectively. The consistent associ- ation of diarrhoeal disease with malnutrition suggests that programmes aimed at decreasing diarrhoeal disease could well have important secondary benefits in reducing malnutrition. The epidemiological description of disease patterns is also useful in defin- ing fruitful areas for investigation to better under- stand cause/effect relationships. For example, the distinct summer and winter peaks in diarrhoea, especially in infants, may be related to different etio- logical agents and modes of transmission and may, therefore, require different preventive measures. Despite the potential usefulness of nutritional status surveillance data, the information collected has important limitations. The data are based on a self- selected population of clinic attenders and the surveil- lance data do not identify the causal interrelationships underlying differences or changes in malnutrition patterns. In addition, because of the delays involved in processing and tabulating the data the surveillance system in El Salvador does not provide data on current short-term changes in nutritional status, or an "early warning" of such changes. However, the problem of malnutrition in El Salvador, as in most developing countries, is fundamentally a chronic condition related to poverty, rather than an acute condition requiring or even benefiting from close monitoring. The nutritional status surveillance system provides not an early warning of nutritional crisis, but rather an ongoing information base from which the magnitude and location of nutrition and health problems can be identified and with which medium and long-term trends in nutritional status can be assessed. Nutritional surveillance cannot be expected to pro- vide all the information,necessary for nutrition and health planning. But surveillance data based on selected indicators of demonstrated reliability can serve as an information base to define the extent and location of nutritional problems and to monitor changes over time. The ability to identify and reliably monitor nutritional and health problems is an essen- tial first step in defining, implementing, and evalu- ating more effective policies and programmes. ;UMt RESULTATS DE LA SURVEILLANCE DE L'ETAT NUTRITIONNEL EN EL SALVADOR, 1975 - 77 On a estim6 que la surveillance nutritionnelle 6tait un moyen de d6tecter la malnutrition a des fins de planification et d'evaluation. Cependant, les systemes de surveillance 6tablis n'ont fourni que peu de donn6es. En El Salvador, un systeme de surveillance de l'6tat nutritionnel a 6tudie les indicateurs de la nutrition et de l'6tat de sant6, d'apres les donn6es des dispensaires du Ministere de la Sante, depuis 1975, a savoir le pourcentage d'enfants dont le poids est inf6rieur A 75% du poids-pour-l'Age pr6vu, ou pour qui le medecin du dispensaire a pos6 le diagnostic de malnutrition. Le pourcentage de ceux chez qui une diarrh6e a ete diagnos- tiqu6e cliniquement est de meme mis en tableaux, comme indicateur d'une maladie infectieuse en rapport 6troit avec la malnutrition. Les donn6es portent sur les trois ann6es 1975- 1977. Les resultats de la surveillance ont indique r6gulierement des taux plus eleves de malnutrition clinique et de deficit du poids-pour-l'age chez les enfants des milieux ruraux que chez ceux des milieux urbains, ainsi que des taux de malnu- trition plus elev6s chez les enfants Ag6s de 1 a 4 ans que chez les nourrissons de moins de 1 an. On a observ6 r6gulitrement des accroissements saisonniers de la malnutrition, particu- lierement prononcas chez les plus vieux des enfants d'&ge pr6scolaire (1 A 4 ans) dans les zones rurales. Ces pics saison- niers de la malnutrition survenaient toujours un ou deux mois apres le principal pic saisonnier des diarrh6es, qui se produisait au d6but de la saison des pluies. Un pic saisonnier secondaire, au cours de la saison fraiche et seche, s'observait principalement chez les nourrissons et ces diarrh6es peuvent correspondre a des infections par des rotavirus. Des carac- teres saisonniers similaires ont ete observ6s pour la malnu- trition et la diarrh6e dans d'autres pays d'Am6rique centrale, ce qui confirme l'exactitude des tableaux de la malnutrition et des maladies r6vel6s par le mecanisme de surveillance. Les donn6es de surveillance n'ont pas constam- ment mis en evidence des diff6rences r6gionales dans l'6tat nutritionnel, mais les frontieres d6finissant les r6gions sanitaires, adopt6es par le Ministere de la Sante, passent au milieu de zones 6cologiques dans lesquelles des diff6rences notables de l'etat nutritionnel ont &6 mesur6es au cours d'autres 6tudes. Les donn6es fournies par le systeme de surveillance de l'etat nutritionnel ont aid6 A d6terminer l'amplitude et la localisation des problemes relatifs A la malnutrition et aux maladies connexes en El Salvador. Ces donn6es mettent en 6vidence les groupes d'iges, les secteurs particuliers et les saisons dans lesquels l'incidence du d6ficit nutritionnel est maximale, ce qui permet d'organiser plus efficacement les interventions, telles que la fourniture d'aliments compl6- mentaires. L'association constante des maladies diarrh6i- 439 440 F. L. TROWBRIDGE & H. C. STETLER ques et de la malnutrition donne a penser que les programmes visant a r6duire ces maladies pourraient avoir pour avantage secondaire important de r6duire la malnutrition. La description 6pid6miologique du tableau de la morbidit6 est 6galement utile, car elle permet de d6terminer les secteurs fructueux pour des recherches visant a mieux comprendre les relations de cause a effet. Par exemple, en ce qui concerne les diarrh6es, les pics distincts d'6t6 et d'hiver, particulierement chez les nourrissons, peuvent etre en rapport avec des agents 6tiologiques et des modes de transmission diff6rents et sont donc susceptibles d'exiger aussi des mesures pr6ventives diff6rentes. Malgr6 l'utilit6 potentielle des donn6es de la surveillance de l'6tat nutritionnel, l'information recueillie a d'impor- tants d6fauts. Ces donnees sont fond6es sur une population auto-s;lectionn6e de sujets fr6quentant des dispensaires. Elles ne r6velent pas les interrelations causales A la base des diff6rences ou des modifications dans les caracteres de la malnutrition. De plus, en raison des d6lais qu'exigent le traitement et la mise en tableaux des donn6es, le systeme de surveillance d'El Salvador ne fournit pas de renseignements sur des changements actuels A court terme dans l'6tat nutri- tionnel, ni d'<<avertissement pr6coce> de modifications de ce type. Toutefois, ce systeme apporte une masse d'informa- tions continues permettant de d6terminer l'amplitude et la localisation des problemes de nutrition et de sante, ainsi que d'6valuer les tendances de l'etat nutritionnel a moyen et a long terme. La capacit6 de cerner et de surveiller de maniere fiable les problemes de nutrition et de sant6 repr6sente une premiere 6tape essentielle sur la voie de la d6finition, de la mise en ceuvre \et de l'evaluation de politiques et de programmes plus efficaces. REFERENCES 1. WHO Technical Report Series No. 593, 1976 (Method- ology of nutritional surveillance: report of a Joint FAO/UNICEF/WHO Expert Committee). 2. TROWBRIDGE, F. L. & STETLER, H. C. Nutritional status surveillance in El Salvador. Bulletin ofthe World Health Organization, 58: 327 (1980). 3. TROWBRIDGE, F. L. ET AL. Evaluation of nutrition surveillance indicators. Bulletin of the Pan American Health Organization, 14: 238 (1980). 4. HAMILL, P. V. V. ET AL. Physical growth: National Center for Health Statistics percentiles. American jour- nal of clinical nutrition, 32: 607 (1979). 5. STETLER, H. C. ET AL. Anthropometric nutrition status and diarrhea prevalence in children in El Salvador. American journal of tropical medicine and hygiene (in press, 1982). 6. TROWBRIDGE, F. L. & NEWTON, L. H. Seasonal changes in malnutrition and diarrheal disease among preschool children in El Salvador. American journal of tropical medicine and hygiene, 28: 135 (1979). 7. POSKETr, E. M. Seasonal variation in infection and malnutrition in a rural paediatric clinic in Uganda. Transactions of the Royal Society of Tropical Medicine and Hygiene, 66: 931 (1972). 8. WALDMANN, E. Seasonal variation in malnutrition in Africa. Transactions of the Royal Society of Tropical Medicine and Hygiene, 67: 431 (1973). 9. VALVERDE, V. ET AL. Life styles and nutritional status of children from different ecological areas of El Salvador. Ecology offood and nutrition, 9: 167 (1980).