Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Determinants of community health workers' performance in India

Sunder Gujral, Rita Abbi, Rajni Mujoo, and Tara Gopaldas



Forty-three anganwadi workers (community health workers) in Gujarat state, India, were interviewed to record their education level, evaluate their nutrition knowledge, and collect information on the number of visits made by the auxiliary nurse midwife (ANM) in the preceding three months and the activities she performed for the anganwadi. The coverage of five services delivered or assisted by the anganwadi worker- supplementary feeding, growth monitoring, vitamin A prophylaxis, health check-ups, and immunization- was estimated by interviewing the mothers of 3,987 children 0-6 years old. The anganwadi worker's having at least a high school education, a nutrition knowledge score of more than 4 out of 7, more than one visit by the ANM in three months, and an ANM activity score of more than 2 out of 9 were significant determinants, individually or in combination, for the anganwadi worker's performance. Multiple regression analysis indicated that nutrition knowledge was the most powerful determinant of performance, followed by guidance from the ANM and education level. It is therefore concluded that anganwadi workers should receive nutrition health education and regular guidance from the ANMs, and their education level should be high school or above.


Editor's note

It is often difficult for the editor and reviewers to judge whether a paper is of interest mainly to the nutrition and health professionals of the country from which it originates or is of potential value to persons in other countries and regions. The Bulletin regularly returns articles, no matter how worthy, judged to be of mostly local interest. The following article at first reading appeared to be a description of a highly specific programme in India. However, the term "anganwadi worker" is simply the local name for the lowest level of community health worker-with equivalents in other countries, whether classified as volunteers (e.g. the kaders of Indonesia) or receiving some small allowance as in India. This paper describes three factors that most influence the performance of community health workers at this level (1) the adequacy of their training, (2) the adequacy of the supervisor, and (3) the level of their formal education. Rapid assessment surveys using anthropological methodologies* have revealed these same three principles in country after country. Too often the first and second are so poor that the third becomes almost irrelevant. This paper has been accepted, therefore, as a reminder of these principles in any nutrition and primary health activity.



The government of India launched the Integrated Child Development Services Scheme (ICDS) [1-3] on an experimental basis in various states of the country in October 1975. Since the impact of the programme was encouraging, it expanded from 33 projects initially to 1,952 by 1989 [4].

The anganwadi constitutes the basic institutional infrastructure through which the ICDS operates at the village level. Each anganwadi caters to a population of 1,000 in rural and urban areas and 700 in tribal areas.

The key functionary in the anganwadi is the anganwadi worker, the front-line individual who is expected to implement the programme at the village level. The success of the programme depends mainly on her performance. As the anganwadi worker is the key person for the development of the community through her performance at the anganwadi, the present study attempts to explore the determinants of her performance.

Performance was evaluated in terms of the proportion of children (0-72 months old) covered for five services: supplementary nutrition, growth monitoring, vitamin A prophylaxis, health check-ups, and immunization. The determinants studied were the worker's education level, her knowledge of nutrition, and the guidance or assistance she received from the auxiliary nurse midwife.

TABLE 1. Anganwadi workers performance related to their education level

Services < high school ³ high school Relative coverage
Children (no. ) Coverage (% ) Children (no. ) Coverage (% )
Supplementary Nutrition 1,636 56.5 1,965 59.6 1.1
Growth Monitoring 1,782 22.8 2,161 22.3 1.0
Vitamin A Prophylaxis 1,482 11.3 1,787 13.6 1.2
Health check-ups 1,796 13.1 2,166 17.4 1.4*
Immunization 751 13.7 832 19.0 1.5*

*P < .01.

Methods and materials

Three to seven villages were sampled at random from each of 11 blocks (subdistrict administrative units) in the Panchmahals district of Gujarat state. All the houses in a village were covered in census fashion. Data were collected with the help of a pretested questionnaire by teams of investigators working in pairs, consisting of a medical intern and a nutritionist-social scientist.

Data on performance

The coverage for each of the five services was recorded by interviewing the mothers of the children in each anganwadi area.

Data on determinants of performance

The anganwadi worker's education level was categorized as either less than high school, or high school or above.

Her nutrition knowledge was assessed on the basis of her responses to a set of seven questions relating to nutrition and health (covering weaning age, feeding frequency, diet during diarrhoea, causes of malnutrition, treatment of malnutrition, growth-chart knowledge, and preparation of oral-rehydration solution), with one point given for each correct answer, and was categorized as scoring either 4 or less, or more than 4.

The worker was asked the number of visits made by the auxiliary nurse midwife (ANM) to her anganwadi in the past three months, and the response was categorized as either one visit or less, or more than one visit.

She was also asked about the kind of guidance the ANM provided through helping to perform various activities during her visits. This guidance was assessed on the basis of nine possible activities (immunization, weighing children and nutrition gradation, health check-ups, vitamin A distribution, iron-folic acid distribution, home visits, nutrition, family planning and health education, referrals, and treatment of primary ailments), with one point given for each, and was categorized as scoring either 2 or less, or more than 2.

Statistical analysis

Multiple regression analysis, using the Statistical Package for Social Sciences (SPSS/PC) [5], was done to find out which determinants had the greatest influence on the anganwadi workers" performance.



Where the anganwadi worker had at least a high-school-level education, a significantly larger proportion of children received health check-ups and immunizations than in areas where the worker had less than a high school education (table 1).

A positive association was observed between the anganwadi worker's nutrition knowledge and her performance (table 2). Significantly more children participated in supplementary nutrition and received vitamin A, health check-ups, and immunization in areas where the worker had adequate nutrition knowledge (score >4) than where the worker's knowledge was inadequate. Similarly, the growth of a larger proportion of children was monitored when the worker's nutrition knowledge was adequate.

Table 2. Anganwadi workers' performance related to their nutrition-knowledge scores

Services Score £ 4 Score > 4 Relative coverage*
Children (no.) Coverage (%) Children (no.) Coverage (%)
Supplementary nutrition 1,460 51.9 2,141 62.4 1.5
Growth monitoring 1,592 16.8 2,351 26.4 2.0
Vitamin A prophylaxis 1,337 8.2 1,932 15.6 2.1
Health check-ups 1,601 10.4 2,361 18.8 2.0
Immunization 729 8.4 855 23.4 3.3

* P < .001

TABLE 3. Anganwadi workers" performance related to visits by the auxiliary nurse midwife in the preceding three months

Services £ 1 visit > 1 visit  
Children (no.) Coverage (%) Children (no.) Coverage (%) Relative coverage*
Supplementary nutrition 1,746 53.4 1,855 62.7 1.5
Growth monitoring 1,901 17.5 2,042 27.2 2.0
Vitamin A prophylaxis 1,585 9.6 1,684 15.4 2.0
Health check-ups 1,903 12.0 2,059 18.6 2.0
Immunization 819 11.0 765 22.4 2.3

*P < .0001.

The auxiliary nurse midwife is expected to visit each anganwadi once a month in order to provide guidance to the anganwadi worker in performing various activities. Table 3 shows the highly significant difference in the coverage of children for the five services between areas where the ANM visited more than once and those where she visited only once.

Further, when the ANM gave help in performing more activities during her visits, the anganwadi worker was better able to provide services to a larger proportion of children (table 4). In anganwadis where the ANM's activity score was above 2, the proportion of children covered for the services was twice as great as where the score was 2 or below.

Since the anganwadi worker's education level and nutrition knowledge play an important role in her performance, we examined their combined effect on the coverage of children. Table 5 clearly indicates that workers with at least a high-school-level education and whose nutrition knowledge was adequate (>4) were better able to cover a larger proportion of children. These two variables appear to have their greatest influence on immunization, followed by health check-ups, vitamin A prophylaxis, growth monitoring, and supplementary nutrition.

We also examined the combined influence of education level, nutrition knowledge, and guidance received from the ANM during her visits on the performance of the anganwadi worker (table 6). A highly significant association was observed between these determinants collectively and the worker's performance.

Multiple regression analysis (table 7) indicates that nutrition knowledge was the most powerful determinant of the anganwadi worker's performance, followed by guidance from the ANM and the worker's education level.

TABLE 4. Anganwadi workers' performance related to the auxiliary nurse midwife's activity score

Services Score £ 2 Score > 2 Relative coverage*
Children (no ) Coverage (%) Children (no.) Coverage (%)
Supplementary nutrition 1,850 53.8 1,751 62.8 1.5
Growth monitoring 2,033 17.7 1,910 27.6 2.0
Vitamin A prophylaxis 1,677 9.5 1,592 15.8 2.0
Health check-ups 2,033 12.1 1.929 19.0 2.0
Immunization 897 9.6 687 25.5 3.2

*P< .0001.

TABLE 5. Anganwadi workers' performance related to their education level and nutrition knowledge scores

Services Lowera Higherb Relative coverage*
Children (no.) Coverage (%) Children (no.) Coverage (%)
Supplementary nutrition 877 54.2 1,382 64.3 1.5
Growth monitoring 951 14.8 1,520 23.4 2.0
Vitamin A prophylaxis 798 6.4 1,248 14.7 2.5
Health check-ups 962 6.9 1,527 18.0 3.0
Immunization 436 4.8 539 21.9 5.5

a. Education < high school and nutrition-knowledge score £ 4
b. Education ³ high school and nutrition-knowledge score > 4
*P < .0001

TABLE 6. Anganwadi workers' performance related to their education level and nutrition knowledge scores and the auxiliary nurse midwife's visits and activity scores

Services Lowera Higherb Relative coverage*
Children (no.) Coverage (%) Children (no.) Coverage (%)
Supplementary nutrition 220 62.3 433 72.3 1.6
Growth monitoring 235 12.8 471 28.0 3.0
Vitamin A prophylaxis 195 2.1 399 20.1 12.0
Health check-ups 237 7.6 479 19.8 3.0
Immunization 110 5.5 150 34.7 9.2

a. Education < high school. nutrition-knowledge score £ 4 ANM's visits £ 1 . and ANM's activity score £ 2.
b. Education ³ high school, nutrition-knowledge score >4. ANM's visits > 1. and ANM's activity score >2.
*P< .0001.

TABLE 7. Regression model of determinants of anganwadi workers' performance

Dependent variable Independent variable
Worker's nutrition knowledge Visits by ANM ANM's activities level Worker's education level
b t test b t test b t test b t test
Health check-ups 0.068 5.64** 0.051 4.34** 0.054 4.60** 0.027 2.27* .0263
Supplementary nutrition 0.092 5.30** 0.075 4.52** 0.067 3.95** 0.009 0.52 .0232
Vitamin A prophylaxis 0.064 5.23** 0.045 3.80* 0.049 4.10** 0. 008 0.67 .0224
Growth monitoring 0.090 6.45** 0.082 6.07** 0.071 5.23** -0.026 -1.88 .0336
Immunization 0.137 7.39** 0.081 4.44** 0.137 7.47** 0.020 1.09 .0930

b = regression coefficient.
r² gives the percentage of variance in the dependent variable that is accounted for by the independent variables.
*P < .0522.
*P < .0001.



As the anganwadi worker is the key person in the programme [6], her education level and knowledge of nutrition and the guidance she received from the ANM individually or synergistically related to her performance in the anganwadi.

The relative coverage of children for the services provided by the programme was higher where the anganwadi worker had a high school education or more than where her education was below that level. Sharma [7], in findings similar to those of the present study, reported that education was positively related to performance. Perhaps relatively better educated anganwadi workers are better able to convince parents to have their children immunized against the six killer diseases and more confident in persuading children to come to the anganwadi for supplementary nutrition. It has also been reported that, in addition to education level, training anganwadi workers about growth monitoring plays a beneficial role in improving their performance [8].

Knowledge and understanding of some aspects of basic nutrition and health care is of great importance for the anganwadi worker's performance [6]. The present study shows that workers with adequate nutrition knowledge reached more children with various services than those whose knowledge was inadequate. We therefore conclude that anganwadi workers should receive nutrition health education periodically, perhaps through mobile in-service training or refresher courses.

Regular visits by the auxiliary nurse midwife or health functionaries to the anganwadi for guiding and helping the anganwadi worker provide repeated on-the-job training, and frequent and regular interaction between ANMs and anganwadi workers is beneficial for the latters' performance [7, 9].



1. Integrated Child Development Services. Soc Welfare 1975;22:9.

2. Integrated Child Development Scheme. Swasth Hind 1976;20: 16.

3. Integrated Child Development Scheme. Swasth Hind 1976;20:361.

4. Tandon BN. ICDS: evaluation and research 1975-1988. Central Technical Committee. New Delhi: Department of Gastroenterology, All India Institute of Medical Sciences, Ansari Nagar, 1990:1.

5. Norusis NJ. SPSS/PC for the IBM PC/XT/AT. Chicago: SPSS Inc., 1986.

6. Udani RH, Chothani S, Arora S. Kulkarni CS. Evaluation of knowledge and efficiency of anganwadi workers. Ind J Pediatr 1980; 47:289-92.

7. Sharma A. Monitoring social components of Integrated Child Development Services-a pilot project. New Delhi: NIPCCD 1987:28-106.

8. Gopaldas T, Christian PS, Abbi RD, Gujral S. Does growth monitoring work as it ought in countries of low literacy. J Trop Pediatr. 1990;36:322-27.

9. Walia BNS, Bambhir SK. Narant A, Gupta KB. Evaluation of knowledge and competence of anganwadi workers as agents for health care in a rural population. Ind Pediatr J 1978;15(10):797-801.


Human nutrition

Evaluation of weight for height: Validation of a visual scale

Esther Casanueva, Laura Magaña and Tania Cárdenas



The design and validation of an instrument for use in programmes for the detection of alterations in the nutrition status of non-pregnant, non-lactating women of fertile age is described. A group of 150 non-pregnant, non-lactating women was studied to classify them in one of Sheldon's five somatotypes according to the visual perception of a trained observer, as well as a self-classification by each woman. Also, the percentage of weight for height for each somatotype was estimated. The sensitivity and specificity of the cut-off points for the percentage of weight for height were calculated in an independent group of 105 women and were found acceptable.



Most programmes to detect alterations in the nutrition status of adults begin with the registration of weight and height to construct the various indicators that permit an indirect estimate of body composition [1]. This anthropometric strategy requires adequately calibrated weighing scales and stadiometers, which are not always available for transportation to the field. It is also necessary to have personnel who are experienced in making the measurements, but they too are not always available [2]. It is therefore important to have instruments to make the measurements with a sufficient degree of precision that can be transported easily and can be used by auxiliary personnel [3].

The objective of this work was to adapt Sheldon's somatotype scales [4] as a screening test to evaluate the nutrition status of adult women, establishing the percentage-of-weight-for-height interval for each somatotype. The objective of Sheldon's scheme is to describe the different female and male human biotypes.


Material and methods

A group of 150 non-pregnant, non-lactating women between the ages of 14 and 40 years was studied, of whom 66% had completed at least basic education. After a careful examination of a chart with pictures of Sheldon's five somatotypes, each woman was asked to classify herself in one of them. At the same time, one of two trained female observers classified the woman according to her own visual perception. The same observers evaluated each woman independently with anthropometric measurements. Weight was measured on a clinical scale for up to 160 kg with 100-g precision; the confidence limit was 200 g. The women were always weighed with a robe of known weight, which was subtracted from total body weight. Height was measured with 1-mm precision and a confidence limit of 2 mm, using a 2-m stadiometer.

The percentage of expected body weight for height* was calculated based on the values proposed by Casillas and Vargas [5] for the Mexican adult population.

Weight-for-height intervals for each somatotype were estimated using the classification from standardized observations assigned to the weight for height of the 150 woman. Feinstein's stratification method was used to establish cut-off points corresponding to each somatotype where the proportion of cases with a defined characteristic changes significantly.

To evaluate whether the weight-for-height intervals assigned to each somatotype were reproducible, the same procedure was applied in a different population of 105 women who were classified by weight for height from actual measurements and then by somatotype by the standardized observer. The classification was assumed to put the woman in the "true" classification, and the visual rating was then compared against this. Sensitivity and specificity values were calculated for each of the cutoff points previously identified.

FIG. 1. Sheldon's somatotypes, modified


Results and discussion

The distribution of weight for height in the population studied was highly dispersed, with an interval of 76%-202%. There was a marked tendency toward overweight.

In the analysis of the distribution of percentage of weight for height, it was not possible to distinguish between Sheldon's somatotypes 2 and 3, so they were combined into one group, leaving a scale of only four categories (fig. l). The percentage-of-weight-for-height intervals for each category are shown in figure 2. In general, no overlapping in the cut-off points was observed.

To validate the calculated intervals, the sensitivity and specificity values were calculated (table 1). The values were highest for categories l and 4 and lowest for categories 2 and 3. The self-classifications and the classifications assigned by the observer showed a high correlation (Spearman's r = .99, p < .001). This indicates that it is feasible to use the woman herself as an observer in programmes for population detection provided she has sufficient basic education. It would be necessary to validate the instrument in a population with a lower distribution of fatness.

FIG. 2. Percentages of expected weight for height by somatotype (N = 150)

TABLE 1. Sensitivity and specificity of cut-off points identified in a control population

Weight/height (%) Somato-
Sensitivity Specificity
<90 1 6/6 (1.0) 90/99 (0.91)
91-110 2 30/45 (0.67) 42/60 (0.70)
111-140 3 29/47 (0.62) 54/68 (0.79)
>=141 4 7/11 (0.64) 93/94 (0 99)

These results indicate that the proposed instrument may be used as a screening test in community programmes where the evaluation of nutrition status is important; and in urban populations the woman herself may even do the classification [7].

The identification of a subject as being in category 1 (thin) or 4 (obese) indicates the need for a more direct evaluation of nutrition status, including at least some anthropometry (weight, height, skinfold thickness). An exploration of food consumption and the possible influence of acute and chronic infections must be taken into account in determining appropriate interventions.



l. Jelliffe DB. Evaluación del estado de nutrición de la comunidad. Geneva: World Health Organization, 1968:7677.

2. Lohman T, Roche A, Martorell R Anthropometric standardization reference manual. Human Kinetics Hooks USA, 1988:3-9.

3. Cravioto J. El significado del crecimiento físico en pediatría. In: Problemas en pediatría: IV. Mexico. Ed. Hospital Infantil de Mexico, 1966:3-13.

4. Sheldon WH, Stevens SS, Tucker W. The varieties of the human physique: an introduction to the constitutional physiology. New York: Harper, 1970:290-99.

5. Casillas L, Vargas L. Cuadros de peso y talla pare adultos mexicanos. Arch Invest Med (Méx) 1980;11:57-74.

6. Feinstein AR. Clinical biostatistics. St. Louis, Mo, USA: CV Mosby, 1977:385-97.

8. Gonzalez-Bernal G, Avila-Rosas H, Breña-Flores H. Utilización de la curve de tolerancia a la glucose oral como criterio pare validar algunos indicadores de riesgo pare presenter intolerancia a la glucose. Rev Invest Clin (Méx) 1988:40:259-64.


The iodine deficiency disorders (IDD): Current status

John B. Stanbury

Deficiency of iodine was firmly established as the cause of endemic goitre in the United States more than half a century ago. The spectrum of disorders known to arise from iodine deficiency has been considerably broadened in recent years and now includes endemic cretinism, endemic deaf-mutism, short stature, reduced fertility, foetal wastage, increased neonatal mortality, and, most importantly, reduction of intellectual and neuromotor performance. These are referred to collectively as "iodine deficiency disorders" (IDD).

With the iodization of salt and increased iodine intake from other sources, IDD has virtually disappeared from North America. Not so in much of the developing world: a fifth of the world's population remains at significant risk of the disorders that flow from lack of iodine. IDD continues to be an important public health problem in much of the Andean world, in North and Central Africa, in much of India and Pakistan, especially the Gangetic plain and the Himalayan region, in Indonesia, and in the People's Republic of China. It is still found in regions of Central Europe and even in Germany, Italy, and Spain. The situation in the Soviet Union is unclear. Iodine deficiency is the most frequent cause of preventable mental retardation in the world today.

How can it be that so important a set of disorders that are so readily and inexpensively prevented continue in the modern world? There are several answers, among them lack of political will, indifference or ignorance, difficulties in the distribution of iodized salt to remote or impoverished communities, and the expense of salt iodization when national budgets are tight and other health issues seem more immediate.

An alternative to the iodization of salt is the distribution of iodinated poppy-seed oil (lipoidol). This product has a high content of iodine that is released so slowly that, if it is given orally, a single dose can meet requirements for a year or more and, if given intramuscularly, for three or more years. It has been widely and successfully used in Papua New Guinea, Indonesia, China, Nepal, Ecuador, Peru, and elsewhere. It is the agent of choice if need is urgent or if the penetration of iodized salt is to be long delayed.

Several techniques are available for appraising the presence of IDD. The traditional method is palpation for the presence of goitre, but the recent introduction of ultrasonography has demonstrated the inaccuracy and inadequacy of palpation. In addition, palpation only gives an integral over time of the subject's exposure to iodine; it does not necessarily indicate current needs. Another measure of IDD is analysis of the concentration of iodine in the urine. This measures the current intake of iodine. The new ultrasensitive methods for measurement of the serum concentration of thyrotropin in the blood give excellent assessment of the current status of the thyroid system and can be done on dried blood spots. The contemporary IDD epidemiologist should be armed with a portable ultrasonograph and collect blood spots for thyrotropin analysis at an established regional control laboratory, and must also determine the current level of iodine intake by analysing urine samples.

Assessment alone leaves the job unfinished. The government and the ministry of health must be made aware of the extent and severity of the problem. The salt industry and the ministry of transport must be brought into the campaign. Legislation favouring the iodization of salt is helpful. Education must proceed through a variety of channels-the schools, radio, television, and the press. An appropriate programme of prevention must be developed and implemented. This may be done by a free-standing programme, or through health posts, or in conjunction with other programmes such as the expanded programme of immunization. Most important is to institutionalize IDD control; otherwise lapse will occur and the disorders will stage a stealthy return.

Many international agencies, such as UNICEF and WHO, and the foreign-aid programmes of Australia, Canada, and the Netherlands have become concerned and involved in the prevention of IDD. An independent organization, the International Council for the Control of the iodine Deficiency Disorders (ICCIDD), is actively promoting awareness, education, and investigation in this field and is seeking to energize preventive programmes in many countries. The goal is to eliminate IDD by the end of the century. Considerable progress has already been made.

Contents - Previous - Next