Contrary to the perception that undernutrition affects individuals only, it has affected families and communities for many decades and has emerged as a silent emergency1. Undernutrition contributes to half of the 10.8 million deaths that occur annually in the under-5s in developing countries; 90% of these deaths are caused by neonatal disorders, diarrhoea, pneumonia, malaria and HIV/AIDSReference Black, Morris and Bryce2. Based on the conceptual model posited by the United Nations Children's Fund, the underlying causes of undernutrition are multifaceted and include problems relating to food security, care-giving practices, health service provision and the environment. Thus it remains a major challenge in most countries to address undernutrition comprehensively and effectively in children from communities where these underlying factors are prevalent1, 3.
South Africa has made progress in developing a national nutrition policy, the Integrated Nutrition Programme (INP), which aims to reduce the problems of undernutrition in children based on a comprehensive and integrated approach4, 5. In addition, funding has been made available for nutrition-related provincial case management guidelines, as well as for implementation of the Integrated Management of Childhood Illness (IMCI) strategy, which aims to address the nutritional needs of children < 6 years of age at primary health care (PHC) facilities4, 5. Despite the implementation of these and other nutrition- and health-related measures, the prevalence of underweight (10%) and stunting (25%) of South African children remained static over a 5-year periodReference Labadarios, Steyn, Maunder, MacIntyre, Swart, Gericke and Huskisson6, 7. Previous studies that were conducted prior to implementation of the INP revealed poor growth monitoring and promotion (GMP) practices, problems with the targeting of vulnerable groups, inadequate food supplementation and poor functioning of the former Protein–Energy Malnutrition SchemeReference Chopra and Sanders8–Reference Schoeman, Dhansay, Fincham, Kunneke and Benade17.
The aim of the present study was to determine the knowledge and practices of PHC nurses in targeting nutritionally at-risk children for intervention at a PHC facility.
Methodology and study population
A cross-sectional, descriptive study was conducted at a PHC facility in a low socio-economic peri-urban area of the Western Cape in South Africa, between 15 November 2001 and 15 January 2002. Based on the population census of 1996, the estimated population size of this area was 34 306, with the majority being colouredFootnote * (98%) and a few black (0.5%)18. The census data further revealed that 43% of the population was under the age of 15 years, 78% under the age of 35 years, while 13% comprised children under the age of 5 years18. Unemployment, followed by low income, a high population growth rate and a high housing density were identified as the major problems in the community. These conditions, together with relatively easy access to the health facility and the absence of previous research undertaken in the area, contributed to the selection of this community for the study.
The participants consisted of infants and children aged 0–72 months, their caregivers and PHC nurses. The nursing staff had to work in the paediatric section of a health facility for at least 6 months to be selected for the study. For caregivers to be recruited, they had to accompany a child to the PHC facility and to reside in the area. These criteria were used to exclude nurses with inadequate working experience at a PHC facility or caregivers utilising the facility while visiting the area only.
A convenience sampling method was used to recruit caregivers and children. The nursing staff consulted caregivers immediately after registration and weighing of the children. Thereafter, the researcher recruited caregivers who had finished their consultation with the nurses and had no objection to participate in the study. This procedure was followed to ensure minimal disruption of the daily routine. Only if consultations were delayed did the researcher select a caregiver prior to consultation with a nurse. An average of three to five caregivers were recruited daily; at the end of the 2-month period, a total of 124 caregivers and 134 children were recruited and participated in the study.
The caregivers' and children's data were collected on a daily basis, while the data relating to the nurses were obtained towards the end of the study. Questionnaires were used to determine sociodemographic, health and nutrition information, and anthropometry of the children. The knowledge and practices of nurses regarding nutrition protocols were assessed. This was done by means of structured interviews and observation of nurses' weighing methods and practices.
The researcher used standard methods to obtain the weight and length/height of a sample of children and caregivers who attended the PHC facilityReference Jeliffe and Jeliffe19. Children were weighed in light clothing on a digital electronic load cell scale to the nearest 0.05 kg. The scale was calibrated with a 1 kg standard weight and zeroed prior to each weighing session. Recumbent length of children < 2 years was measured with a wooden board with a fixed headboard and a moveable footboard. The caregivers were asked to support the infant's head against the headboard while the researcher measured the infant's length. The height of children ≥ 2 years was measured in the standing position with a stature meter that was fixed to the wall. Both length and height were measured to the nearest 0.1 cmReference Jeliffe and Jeliffe19.
Anthropometric data obtained by the nurses and the researcher were analysed using Epi-Info version 6.0 (Centers for Disease Control and Prevention, Atlanta, GA, USA). The researcher used a growth chart based on the weight-for-age chart on the Road to Health Chart (RTHC) to plot the children's current and four previous weights obtained at the facility in order to interpret their growth curve. The advantage of obtaining four previous weights instead of two made it easier to identify weight errors or inconsistencies. For children, weight below the 3rd centile, birth weight less than 2500 g, and flattening or dropping off of the growth curve for at least two consecutive monthly visits were used to classify the child as being nutritionally at-risk4. The prevalence of underweight, stunting and wasting was based on the proportion of children with Z-scores less than minus two standard deviations ( < –2SD) below the median of the National Center for Health Statistics/World Health Organization reference values20. The researcher determined whether children identified as being nutritionally at-risk were referred to the Nutrition Supplementation Programme of the health facility-based programme and entered onto the register at the PHC facility. She also determined other nutrition interventions provided to those children who were nutritionally at-risk, e.g. vitamin A and iron supplementation and deworming.
Results
Sociodemographic characteristics of caregivers, infants and children
The study population comprised 124 caregivers and 134 pre-school children. Of the 124 caregivers, 105 (85%) were the biological parents, 15 (12%) were relatives and four (3%) were non-family members. The mean age of the caregivers was 29.5 (SD 7.5) years. The mean age of the children was 19.0 (SD 17.8) months. Only 47 (38%) caregivers were married, 18 (15%) lived with a partner, 55 (44%) were single and four (3%) were divorced. Most caregivers, 96 (77%), were unemployed; 72 (58%) had one or two children, 38 (31%) had three or four children and 14 (11%) had five or more children. Household heads were indicated as follows: 45 (36%) were the caregiver's spouse, 35 (28%) the caregiver's parents, 25 (20%) were the caregiver, 13 (10%) were a relative of the caregiver, and six (5%) a non-family member. Twelve (10%) consumed alcohol regularly, six (5%) consumed alcohol occasionally, 91 (73%) never consumed alcohol, and 15 (12%) had previously consumed alcohol but had stopped. Of the total sample of women 42 (34%) were smoking. Of those who smoked, the majority were coloured 40 (95%) compared with two (5%) black. Forty-one (33%) women also smoked during pregnancy.
Significantly more caregivers had children who were nutritionally at-risk if they were unemployed, 52 (54%) (P = 0.04), compared with those who were employed, nine (32%); and if there was household food insecurity (36; 63%) compared with household food security (25; 37%) (P = 0.004). It was also found that significantly more caregivers who were financially self-supporting or supported by their partners (46; 61%) had children who were not nutritionally at-risk (P = 0.003) compared with those who were not financially self-supporting or supported by their partners or a relative (17; 35%) (Table 1).
Nutritional risk status of children based on sociodemographic status of the caregiver (n=124)
Values are given as n (%).
Weighing practices
Differences were observed between the weighing practices of the researcher and the nurses. The nurses weighed the infants ( < 12 months) without clothing on an infant scale and the children ( ≥ 12 months) in light clothing on a bathroom scale. To ensure accuracy of weight readings, the nurses adjusted the infant scale's needle to zero at the start of the weighing session. The needle of the children's scale was not adjusted regularly and registered beyond zero at times. This observation was only made at the beginning of the weighing session and towards the end of the day. The researcher calibrated the scales with a 1 kg weight in the morning prior to the weighing and weighed all children in light clothing. The nurses weighed most children (95%) either without or in light clothes and 5% of the children fully clothed without their shoes.
Completion of the RTHC
The RTHC was available for 123 (92%) children. When assessing completion of the weight-for-age chart, it was found that dates of birth were entered on 50%, calendars were completed on 8%, weights were plotted on 46%, while only 4% had information on infant and or child feeding (Table 2). The researcher was able to assess children with underweight and growth faltering by obtaining four recent weights from the health facility records. These weights were plotted on a growth chart similar to the weight-for-age chart of the RTHC. This enabled the researcher to assess 67 (50%) children as being underweight or faltering in growth. Since the nurses did not plot the weights of 55% of the children, they failed to detect a large number of children who were underweight or faltering in growth.
Growth monitoring and promotion (GMP) practices of nurses and their utilisation of the Road to Health Chart (RTHC)
Values are given as n (%).
Comparison of weight measurements
Weights were available for comparison between the researcher and the nurses for 133 children. The mean weight of all children obtained by the nurses and the researcher was 8.94 kg and 9.18 kg, the mean weight for infants was 5.26 kg and 5.56 kg, and the mean weight for children 11.44 kg and 11.66 kg, respectively (P < 0.00). The researcher's measurements were consistently higher than the nurses' with the exception of the nurses' maximum weight measurement of 19.0 kg against the researcher's 18.3 kg for the same child (Table 3).
Comparison of weights (kg) obtained by the nurses and the researcher (n=133)
*,**P < 0.00.
Identification of nutritionally at-risk children
The number of children with underweight identified by the researcher was 28 (21%) compared with 34 (25%) by the nurses. In addition, the researcher found that among the infants and children, 29 (22%) were stunted and four (3%) were wasted. The nurses had no information on stunting and wasting since height is not measured routinely at PHC facilities. Birth weights of 128 infants and children were obtained from clinic records and the children's RTHC, of whom 19 (15%) were low birth weight. The researcher and nurses identified underweight and growth faltering in 67 (50%) and 14 (10%) of the children, respectively.
Nutrition intervention for the nutritionally at-risk group
Of the sample of 67 (50%) nutritionally at-risk children identified by the researcher, the nurses' nutrition intervention was as follows: 14 (21%) were registered on the Nutrition Supplementation Programme, 11 (16%) received food supplements, three (4%) received vitamin A supplements, 23 (34%) received anthelmintics and 15 (22%) received iron supplements (Table 4).
Nutrition intervention by nurses for nutritionally at-risk children (n=67)
Values are given as n (%).
Follow-up visits
Follow-up visits were arranged for the following reasons: immunisations, 54 (40%); treatment of disease, 38 (28%); screening, five (4%); and growth monitoring, 10 (7%). Twenty-seven (20%) children were not given a date or asked to return for follow-up.
Training of nurses and their knowledge of nutrition case management guidelines
None of the nurses (n = 12) received full training in implementing the nutrition case management guidelines. One (8%) received formal training regarding the INP guidelines, which covered only the Nutrition Supplementation Programme, seven (58%) received training in the IMCI strategy and primary health care, and four (33%) received no training.
None of the nurses were knowledgeable about the recommended frequency of GMP visits for children aged 0–24 months, and only two (17%) knew about the frequency of GMP visits for children aged 24–60 months.
All of the nurses (100%) were knowledgeable about underweight and growth faltering as risk criteria for targeting children < 6 years, while two (17%) indicated that tuberculosis and HIV/AIDS were also risk criteria for this group. None of the nurses knew that underweight or body mass index < 18.5 kg m–2 was a criterion for supplementing lactating mothers, but five (42%) indicated that teenagers who have no income, lack household food security or who have other social problems while breast-feeding their infants, needed support.
Discussion
Although the spouse/partner, caregiver or a relative headed 95% of the households in this study, this did not necessarily translate into optimal nutrition for many of the children1. The large number of caregiver-headed households (20%) could influence the quality of life of mothers. It has been found that when mothers are exposed to long working hours, their child caring capacity can be diminished, while non-working mothers are often financially dependent on others for food and basic requirements. In either case, the children in these households are at risk of undernutrition1, Reference Allen and Gillespie21, 22.
Fifty per cent of children were found to be nutritionally at-risk and factors such as unemployment, food insecurity and dependency on non-family members for financial support significantly increased the risk of undernutrition. It was also revealed that 48% of the caregivers were single and unsupported by their partners, which coincides with previous studies of children with undernutrition in similar communitiesReference Hendricks, Le Roux, Fernandes and Irlam13. Since food availability at the household level is an important indicator of related social risks, the identification of food-insecure households needs to be strengthened and should be included with other criteria, e.g. household income, in targeting nutritionally at-risk children1, 3, 5, Reference Labadarios, Steyn, Maunder, MacIntyre, Swart, Gericke and Huskisson6. Once such households are identified and targeted, access to the child support and care dependency grants as well as food supplementation could be initiated and facilitated via the PHC system or community points (unpublished data). This could improve social security measures for the caregivers and families with undernourished children.
In this study 34% of the caregivers smoked, which is probably related to the high prevalence of low birth weight in this population22. Similarly, another study in a neighbouring community, Bishop Lavis, between 1993 and 1995 showed that 60% of the women smoked and a low birth weight rate of 22% was reported for the children in this community (unpublished data). Since smoking and low birth weight are predominantly rife among coloured women, these trends need to be observed with the emerging racial integration among the different groups in South African communities.
A major limitation in GMP practices of nurses was under-utilisation of the RTHC with concomitant poor plotting and no entry of important nutrition-related information on the RTHC. Because of this, the nurses missed 84–95% of nutritionally at-risk children eligible for food and vitamin A supplements and nutrition monitoring recommended by the INP4. Other shortcomings in GMP related to poor maintenance, calibration and standardisation of scales as well as inaccuracies in weighing. Although the discrepancies between the weight measurements of the nurses and researcher were not extreme, the constant difference of 200–600 g could affect the nutritional assessment of individual children. This was demonstrated in the small number of nutritionally at-risk children who were targeted to receive food supplements and the few caregivers receiving follow-up dates. In the long term, these shortcomings undermine strategies aimed at addressing undernutrition and promoting child survivalReference Kuhn, Zwarenstein, Thomas, Yach, Conradie and Hoogendoorn14, Reference Morley and Elmore-Meegan16. The consequences of poor GMP at PHC facilities have been documented in many studies, yet little has been done to change the situationReference Coetzee and Ferrinho10, Reference Hendricks, Perez, Roberts and Weight12–Reference Schoeman, Dhansay, Fincham, Kunneke and Benade17.
Given the poor socio-economic status of the population, the underweight prevalence of 21% (researcher) and 25% (nurses) in this study, which is twice the national mean of 10.3% for South African pre-school children, is not surprisingReference Labadarios, Steyn, Maunder, MacIntyre, Swart, Gericke and Huskisson6, 7. Unlike studies targeting impoverished communities, national surveys include participants across the socio-economic spectrum, which could obscure a higher prevalence of underweight among vulnerable groups of children. Studies focusing on impoverished communities are therefore essential in targeting more vulnerable groups for nutrition intervention as suggested in the INP. It is also important that weighing equipment, methods and procedures are standardised to ensure reliable anthropometric outcomes and improved targeting of at-risk individuals.
The serious consequences of vitamin A deficiency and the benefits of vitamin A supplementation to mothers and children have been documented in numerous studies5, Reference Allen and Gillespie21, 22. A meta-analysis showed in six out of eight studies that vitamin A supplementation could reduce the all-cause mortality rate by 23% among children between the ages of 6 months and 5 years. It was found that improvement in vitamin A status was more likely to impact on mortality attributed to diarrhoeal disease and measles than on mortality attributed to respiratory infection or malariaReference Beaton, Martorell, Aronson, Edmonston, McCabe and Ross23. One trial showed that periodic vitamin A supplementation of HIV-positive children could reduce all-cause mortality by as much as 63%Reference Fawzi, Mibise, Hertzmark, Fataki, Herrera and Ndossi24. It was also found that HIV-infected infants born to mothers who were given vitamin A in the last trimester had better intestinal mucosal integrity when tested at 1–14 weeks, which was not found in non-infected infants22. Despite the proven benefits and cost-effectiveness of vitamin A supplementation and blanket deworming of children, these interventions were poorly implemented at the PHC facility1, Reference Allen and Gillespie21–Reference Beaton, Martorell, Aronson, Edmonston, McCabe and Ross23, Reference Fincham, Evans, Dhansay and Yach25.
The training of nurses to improve GMP and the targeting of nutritionally at-risk children needs strengthening through the human resources development sub-directorate within the province. The feasibility of community-based GMP and linking it to other community initiatives needs to be considered. For example, a community-based GMP system in the Langebaan community on the West Coast in the province has proved successful in facilitating 80–100% coverage and blanket deworming of pre-school children at 4-month intervals over a 3-year periodReference Schoeman, Dhansay, Fincham, Kunneke and Benade17. This model has also facilitated community nutrition surveillance and monitoring, and the feasibility of utilising it in other urban communities should be testedReference Schoeman, Dhansay, Fincham, Kunneke and Benade17. Findings from the research undertaken indicate it is important that the emphasis on what is wrong be moved to how to address the current problems which are experienced in nutritional programmesReference Berg26.
Conclusion
The research findings of this study highlight numerous shortcomings relating to implementation of the INP of the Department of Health at the primary level of care. There was poor targeting and intervention of nutritionally at-risk children by those given the task of implementing the INP. Factors contributing to this were the lack of training of PHC nurses, lack of supervision and nutrition monitoring, inconsistent GMP practices and the absence of community-based nutrition programmes to address underweight and poor food security in the long term. It is recommended that policy-makers urgently address these issues to ensure appropriate intervention for nutritionally at-risk children in similar communities.
Acknowledgements
We wish to acknowledge the assistance and support of Dr C de Villiers (City of Tygerberg); the nursing staff and the mothers of Delft Community Health Centre; Drs Sheik and Lawrence and Mrs Olivier (Provincial Administration of the Western Cape); and Dr C Lombard (Medical Research Council).
