Authors Helen Moestue, Public Health Intervention Research Unit, London School of Hygiene and Tropical Medicine, 50 Bedford Square, London WC1B 3DP, UK. Tel.: 0207 2994724; E-mail: firstname.lastname@example.org (corresponding author). Bonifacio Mahumane, Save the Children, USA, P.O. Box 1854, Maputo, Mozambique. Tel.: 258 493140; Fax: 258 493121; E-mail: email@example.com Anna Zacher, 4972 Del Monte Ave, San Diego, CA 92107, USA. Charles M. Kihamia and Wahab Issae, Tanzania Partnership for Child Development, P.O. Box 9383, Dar es Salaam, Tanzania. Tel.: 255 51 30131; E-mail: firstname.lastname@example.org Su-Tung Wen and Sam Adjei, Ghana Partnership for Child Development, Health research Unit, P.O. Box 184, Accra, Ghana. Tel.: +233 21 226739; E-mail: email@example.com Donald A. P. Bundy, World Bank, 1818 H Street NW, Washington DC 20443, USA. Tel.: +1 202 473 3636; E-mail: firstname.lastname@example.org Andrew Hall, School of Integrated Health, University of Westminster, London, NW1 3ET, UK. E-mail: email@example.com
Background Insufficient attention has been paid to the health problems of school-age children in sub-Saharan Africa. A questionnaire administered to schoolchildren about their ill-health has been developed to identify schools in which urinary schistosomiasis occurs. The data collected during the interviews can also be used to assess other common health problems.
Objectives To analyse data collected during health questionnaires in schools to assess how schoolchildren perceive their own health, and to compare the findings between three countries in sub-Saharan Africa.
Methods Questionnaires asking about recent health problems were administered by teachers to schoolchildren in 120 primary schools in Mozambique, 52 primary schools in Tanzania and 298 primary schools in Ghana. A total of 67 002 children aged 8–15 years took part.
Results Of the 10 health problems asked about in all questionnaires, the average number reported by each child was 3.9 in Ghana, 3.4 in Mozambique and 3.1 in Tanzania. The distributions of the prevalence of each condition among schools were similar and the prevalence of all conditions showed a similar ranking. For most conditions a greater percentage of girls than boys reported each health problem.
Conclusions Schoolchildren in Ghana, Mozambique and Tanzania do not perceive themselves to be healthy. The pattern of reported health problems was similar in each country. School health questionnaires are worthy of further study and validation.
There are few policies and programmes to combat morbidity in school-age children in developing countries, especially when compared with programmes for infants and pre-school children, and they could be considered a neglected group (Goodburn & Ross 2000). This is mainly because there is little information about the extent and causes of ill-health among school-age children and, what information does exist, tends to have been collected during small surveys on a few children. There is also little or no published information about how school-age children perceive their own health.
Questionnaires have been recognized for some time as a means to find out about people's health problems and to provide information which can help to plan programmes (Kroeger 1983; Ross & Vaughan 1986; Barreto 1998; Mafe et al. 2000). A study in Bangladesh, for example showed that children aged between 2 and 9 could reliably identify their own severe or moderately severe disabilities (Zaman et al. 1990) while interviews with mothers in Ghana and the Philippines have been used to diagnose specific health problems in their children (Belcher et al. 1976; Kalter et al. 1991). Questionnaires have also been used to explore perceptions and attitudes to health and ill-health (Ostberg et al. 2001).
If the information given by children in a questionnaire is diagnostically reliable, so long as any under- or over-reporting is systematic and consistent, it could be used to identify schools where a health problem is common and where curative or preventive measures could be effectively directed. This notion has been studied and validated for urinary schistosomiasis, a focal disease caused by the blood fluke Schistosoma haematobium which is responsible for some distinctive symptoms including pain when urinating and blood visible in urine. These self-reported symptoms can be validated easily, either by testing urine for blood using a reagent strip, or by examining microscopically a filtrate of urine or some sediment. In this way the relationship between the reported prevalence of a symptom and the prevalence of infection can be established. Research in Tanzania has shown that the prevalence of symptoms of urinary schistosomiasis in schools under-reports the actual prevalence by a consistent percentage (Lengeler et al. 1991; Ansell et al. 1997) and that, in some circumstances at least, children can quite reliably diagnose their own infections when they have moderate to heavy infections (Partnership for Child Development 1999a). The ‘red urine’ questionnaire as it has been called, which is administered by teachers, has been validated in several African countries (Red Urine Study Group 1995; Lengeler et al. 2002) and has become a useful tool to identify schools where urinary schistosomiasis occurs and in which mass treatment with praziquantel is warranted (UKUMTA 1997; Partnership for Child Development 1999a). The questionnaire has also been applied in areas where intestinal schistosomiasis caused by Schistosoma mansoni occurs, but it has generally been found to be less sensitive and specific than for Schistosomiasis haematobium, and according to a review by Lengeler et al. (2002), still requires validation.
In the ‘red urine’ questionnaire children are also asked about several other common health problems with the intention of masking the main interest in urinary schistosomiasis. The information collected on the other health problems is usually not analysed or used. The present analysis uses data about the other health problems collected during questionnaire surveys for urinary schistosomiasis in primary schools in Ghana, Mozambique and Tanzania. The aim was to assess how schoolchildren perceive their own health and to compare the findings between three countries in Africa.
The questionnaire surveys were undertaken by Save the Children, USA, in Gaza Province in Mozambique and by the Partnership for Child Development in Tanga Region in Tanzania and in Volta Region in Ghana, between 1994 and 1999. The questionnaires were administered in order to identify schools where urinary schistosomiasis occurred and were validated using reagent strips to detect blood in urine in both Tanzania and Mozambique (results not reported here). Mass treatment with praziquantel was then given according to WHO recommendations (Montresor et al. 1998). The questionnaires used in each country were based on the one developed by Lengeler et al. (1991) in Tanzania but did not use a separate list of ‘diseases’ and ‘symptoms’, simply a list of common ‘health problems’, which included urinary schistosomiasis and blood in urine. The questionnaires were adapted in each country to list common health problems that were known by local school- children in the area where the survey was to be done. Seventeen questions were listed in the Mozambique questionnaire, 19 in Ghana, and 15 in Tanzania. Ten health problems were common to the questionnaires used in all three countries: malaria, headache, cough, stomach ache, vomiting, diarrhoea, toothache, eye infection, ear infection and blood in urine. Four health problems were common to the questionnaires used in Tanzania and Mozambique only: schistosomiasis, lice, jiggers and a wound or cut. These health problems were chosen because they were well known to children, such as malaria, cough and stomach ache; because they had clear symptoms, such as diarrhoea, vomiting and blood in urine; because they affected specific parts of the body such as the eyes, ears, or teeth; or because they were easily identified, such as jiggers and a cut or wound.
The questionnaire was provided to teachers in each country with written instructions about how it should be administered and how questions should be asked. In Mozambique the questionnaire was administered in Portuguese and in Ghana in English, both the national languages; if these languages were not well understood, especially by younger children, then the teacher translated them into the local language or dialect. In Tanzania the questionnaire was administered in Kiswahili, the national language. Each teacher was instructed to interview each child individually and in private, and to ask whether the child had experienced each of the listed health problems within the last 4 weeks in Ghana and Tanzania, and within the last 2 weeks in Mozambique. The answer to each question was recorded as a tick (✓) for ‘Yes’, a circle (○) for ‘No’, and a dash (–) for ‘Don't know’. A circle was used to denote ‘No’’ rather than a cross (×) because of the need to allow teachers to correct answers clearly.
All the data were single entered into a computer using EpiInfo versions 5 and 6 (Dean et al. 1994) for all children interviewed in Mozambique, for all children in a randomly selected sample of 15% schools in three Districts in Tanzania, and for children in all primary schools in three Districts in Ghana who returned questionnaires. The prevalence of ‘Yes’ answers for each health problem was calculated for each school and for all children in each country, and stratified by age and sex. The arbitrary minimum sample size for the smallest category for analysis was 50 children. Data were analysed using EpiInfo version 6 and SPSS for Windows version 9.0 (SPSS, Inc., Chicago, IL, USA). Chi-square tests were used to test the statistical significance of differences in proportions.
Data were analysed for a total of 67 002 children aged 8–15 years enrolled in 120 primary schools in Mozambique (n = 37 713 children), 52 primary schools in Tanzania (n = 9184) and 298 primary schools in Ghana (n = 20 105).
In Ghana, Mozambique and Tanzania 97.3%, 93.7% and 91.4% of children, respectively, had reported having at least one health problem in the previous 2–4 weeks. Of the 10 health problems that were common to all three questionnaires the percentages of children who reported at least one problem were 94.2%, 91.1% and 88.8%, respectively. Figure 1 shows the percentage of children in each country who reported between none and all of the 10 health problems that were common to all three countries. Many children reported having more than one health problem. Of the 10 health problems asked about in all questionnaires, the average number reported by each child was 3.9 in Ghana, 3.4 in Mozambique and 3.1 in Tanzania.
Figure 2 shows the distributions of the prevalence in schools of six reported health problems. The distributions were largely similar in shape for each problem. The distributions of the prevalence of blood in urine and diarrhoea were both skewed to the right, patterns which are consistent with either focal or epidemic infections. The distribution in Tanzania and Mozambique of jiggers, an infection with the flea Tunga penetrans, was similar (not shown), and ranged from 1 to 67% in Tanzania (mean 20%) and from 0 to 97% in Mozambique (mean 36%). The prevalence of reported headache and vomiting tended towards normality with vomiting showing a very wide range, especially in Ghana. The distribution of the prevalence of problems with eyes and teeth was again very similar between countries: a normal distribution with a mostly small proportion of children reporting each problem in most schools.
Figure 3 shows for boys and girls in each country the percentage who reported each of the 10 common health problems ranked according to the frequency of problems reported by children in Ghana. Figure 3 shows that the rank of health problems was similar in each country: headache, cough and stomach ache were the three most commonly reported conditions and were all ranked in the same order. Vomiting and malaria were reported by about 30–50% of children in all countries but were ranked fourth and fifth in Ghana compared with fifth and fourth in Mozambique and Tanzania. All other conditions were reported by about 20% of children.
Figure 3 also shows that except for blood in urine and diarrhoea in all countries, and for toothache in Ghana, a greater percentage of girls than boys reported all other health problems. Blood in urine, a symptom of urinary schistosomiasis, was reported by 20–21% of boys compared with 9–13% of girls.
Four health problems showed a statistically significant decline in prevalence with age in all three countries: headache, cough, vomiting and diarrhoea. Figure 4 shows as an example, the prevalence by age of reported vomiting, which showed the steepest decline and statistically strongest association.
Schoolchildren in Ghana, Mozambique and Tanzania were asked whether they had experienced 10 common health problems within the last 2–4 weeks. On average children reported between three and four health problems, and 89–94% reported at least one. It seems that children in sub-Saharan Africa do not perceive themselves to be healthy, which is unsurprising given that a disproportionately large burden of all disease occurs in this region (World Bank 1993). The pattern of reported health problems was remarkably similar between three countries in both East and West Africa: the distributions in all countries of the prevalence of conditions in schools were similar (Figure 2); the prevalence of all conditions showed a similar ranking (Figure 3); for all conditions except vomiting, the difference between the highest and lowest percentage was on average 20% of the maximum, indicating a fairly narrow range in prevalence for a particular condition between countries (Figure 3); and finally, for most conditions a greater percentage of girls than boys reported each health problem (Figure 3).
The number of problems reported by children in Mozambique might have been underestimated because a 2-week recall period was used compared with 4 weeks in Ghana and Tanzania. As most of the health problems enquired about were acute and short-lived rather than chronic and long term, this could also add to the underestimation of children's health problems in Mozambique. There is a need for studies to evaluate the reliability of recall over different periods, perhaps by comparing recall in the last 24 h with recall over the previous weeks to see if the prevalence of conditions is proportional. Studies have shown that reporting error may be related to the length of the recall period and the severity of the disease (Alam et al. 1989; Ramakrishnan et al. 1999).
The reported health problems need to be validated because children in each country may have different perceptions and knowledge of illnesses, which may affect how they report it. A comparison of problems reported by people in Ghana during a health questionnaire followed by a clinical examination found that malaria, diarrhoea and chronic conditions were under-reported (Belcher et al. 1976). Yet even if children were correctly reporting health problems that they had experienced recently, they may over-report because of the ‘concertina effect’ in which people report ill-health experienced over a longer recall period than asked about. The only reported health problem in the questionnaire that has been validated by an objective diagnosis is ‘blood in urine’, a condition mostly caused by infections with S. haematobium, although other causes may become apparent when the prevalence is low, especially among girls (Hall & Fentiman 1999).
Although validation of reported illnesses would shed light on the scale and severity of the health problems, the very fact that children have such poor perceptions of their health implies a need to assess the notion of the healthy schoolchild, to examine in more depth the health of this age-group, and develop school-based health services.
The fact that children assemble in schools provides an obvious opportunity to develop simple screening methods, such as for eye problems (Wedner et al. 2000), and then deliver health services within an existing infrastructure. There is growing experience of school-based health services: teachers in Ghana and Tanzania have been shown to be able to treat their pupils for intestinal worms and urinary schistosomiasis (Partnership for Child Development 1999b; Guyatt et al. 2001); teachers in Mali have effectively given weekly iron tablets to schoolchildren (Hall et al. 2002); and in Tanzania and Malawi teachers have been shown to be able to make a presumptive diagnosis of malaria and provide treatment (Magnussen et al. 2001; Pasha et al. 2003). A questionnaire is a quick, easy and inexpensive method to assess schoolchildren's health, while the involvement of school staff and pupils is likely to improve community awareness of ill-health, an essential first step in the development of school health programmes.
The data presented here show that schoolchildren in Ghana, Mozambique and Tanzania do not perceive themselves to be healthy, and that the pattern of reported health problems was similar in each country. The findings stress the need for greater efforts to improve the health and sense of well-being of older children and adolescents. Questionnaires administered by teachers could also provide more than just information about the distribution of urinary schistosomiasis and are worthy of further study and validation for other common self-reported health problems.
The Partnerships for Child Development in Ghana and Tanzania were funded by the Edna McConnell Clark Foundation.