Implications for school-based health programmes of age and gender patterns in the Tanzanian primary school


ProfessorD.A.P. Bundy, Partnership for Child Development, Wellcome Trust Centre for the Epidemiology of Infectious Disease, University of Oxford, South Parks Road, Oxford OX1 3PS, UK. E-mail:


This paper examines children's potential access to school health services by analysing data on the demographic structure, enrolment patterns and reported causes of early school-leaving in 347 schools in Tanga Region, Tanzania served by a school-based health programme. The analysis indicates that net enrolment ratios have risen over the previous 6 years, particularly among children under 10 years. However, in 1994 children were still much older than expected for a basic school population: 81% were adolescents ($10 years) and the mean age was 12 years. These data suggest that schools can provide equitable access to health education and school-based health services for a majority of children, even in a low-income country, and that the primary school population is predominantly adolescent and would benefit from health programmes targeted at that age group.


There is increasing recognition that many schoolchildren, particularly in the poorest countries of the world, suffer from health conditions that may constrain their ability to benefit from education (Pollitt 1990; Sternberg et al. 1997) and that the school setting itself offers one approach to providing improved health delivery to school age children (Halloran et al. 1989). The United Republic of Tanzania has been exploring a school-based approach to health services delivery whose preliminary findings have been summarized elsewhere (Hall et al. 1996; Partnership for Child Development 1997). This paper examines the demographic structure, enrolment patterns and reported causes of early drop-out in the schools covered by the Tanzanian programme, and examines the implications of these data for the coverage and content of the school-based approach to health delivery.

Sources of data and methods

Tanga Region is situated on the coast of northern Tanzania, on the border with Kenya. It had a population of just under 1.3 million according to the 1988 census. Three districts – Tanga, Muheza and Korogwe – are participating in a large-scale school-based health programme implemented by the Ministry of Education and Culture and the Ministry of Health as part of an international initiative to strengthen school health services. The Tanzania Partnership for Child Development Programme, called Ushirikiano Wa Kumwendeleza Mtoto Tanzania (UKUMTA), is providing treatments for intestinal worms and urinary schistosomiasis as well as health education to a population of about 110 000 primary schoolchildren in these three districts.

Two sources of data were used in the analysis. Firstly, enrolment forms submitted by schools to the District Education Offices in Tanga, Muheza and Korogwe in 1994, which were entered into a computer database using EpiInfo software (Dean et al. 1994) and analysed using EpiInfo and SPSS (Norusis 1993). Secondly, data from the 1988 Census (United Republic of Tanzania et al. 1992a) were used to estimate the percentage of the school-age population who should have been enrolled in 1994. Death rates were calculated for children aged 1–4 years according to the method described by Newell (1988) using the rates for infants and children under 5 years reported, by district, in the 1988 census (Tanzania Bureau of Statistics 1992b). Death rates for children aged 5–9 years and 10–14 years, were calculated using data on the number of deaths of boys and girls reported in rural Tanga Region during the 1988 census (United Republic of Tanzania, 1992a). Immigration and emigration were assumed to be the same.


Data were obtained from 347 of the 350 schools in the three districts of Tanga Region in 1994: 63 schools in Tanga District, 147 in Muheza, and 137 in Korogwe. Forms were not available from three schools in Muheza District, with an estimated 750 pupils. A total of 108349 pupils (55146 boys and 53682 girls) were enrolled in 2345 classes. The average number of pupils per school was 526 in Tanga District, 249 in Muheza and 282 in Korogwe.

Table 1 indicates that net enrolment ratios have increased by nearly 13% between 1988 and 1994, with most of the increase resulting from greater net enrolment among children aged 7, 8 and 9 years. The greater net enrolment of girls than boys in 1988 (χ2 = 161.7, P < 0.001) has fallen slightly but has been sustained, so that a greater proportion of girls than boys were enrolled still in 1994 (χ2 = 37.2, P < 0.001). The gross enrolment ratio in the three districts of Tanga Region in 1994 was estimated to be 80.3%.

Table 1.   Net enrolment ratios of boys and girls by age for children aged 7–14 years (the official age range for primary education) in three districts of Tanga Region, Tanzania in 1988, calculated from the census (Tanzania Bureau of Statistics 1992b) and estimated for 1994 Thumbnail image of

This basic school population, which officially should range from 7 to 14 years of age with a mean of about 9.5 years, actually extends beyond 17 years of age and has a mean of about 12 years. Figure 1 shows the extraordinary range of ages found in each class. In Class 1, for example, it may range from 6 to 14 years. There is no obvious trend for the heterogeneity to differ amongst classes.

Figure 1.

 The distribution by age of all children in all classes of Primary Schools in Tanga, Muheza and Korogwe Districts, Tanga Region, Tanzania, in 1994. ▪ Class 1; □ Class 2; ▴ Class 3; ▵ Class 4; ● Class 5; ○ Class 6; ♦ Class 7.

There is a trend for girls to be of the order of 1 year younger on average than boys in any given class (data not shown), but to have a similarly broad age range. An examination of the summary data produced in the Ministry of Education and Culture each year (United Republic of Tanzania 1996) indicates that these age range patterns also occur in other parts of the country.

Some reasons for leaving school are shown in Table 2. The reported death rates show great variation by class, but this heterogeneity may reflect the wide and variable age ranges represented. Note that there is no apparent difference by sex, which is biologically improbable, and which may indicate that the gender-dependency component is probably overwhelmed by the age variation. Drop-out is initially higher for boys, but the rates for girls are higher from Class 6 onwards. The drop-out rate is, reportedly, different from that attributable to pregnancy, which is given as the reason for leaving by one in every 30 girls in Class 7. Note that the high pregnancy rate refers to the, perhaps better-off, minority of girls still in school at this age.

Table 2.   Reasons for pupils leaving school in 1993 in three districts of Tanga Region, Tanzania. The results are expressed per 1000 children enrolled in 1994 by class and sex. There were no significant differences between the sexes for all children combined Thumbnail image of


Although there may be underlying inconsistencies in these data, some basic trends are strongly represented, some of which suggest positive change is underway in Tanzania. There has been a very significant improvement in net enrolment ratios between 1988 and 1994, and girls continue to represent the majority of the younger age classes. But an important conclusion of these analyses is that schoolchildren are, on average, much older than the official class structure would suggest. In the population examined here, 81% of the children are older than 10 years, the WHO threshold for adolescence (WHO 1995), an age which marks the start of an important stage in both physical and behavioural development. There is evidence that life skills training and other methods of self-empowerment in health decision-making (including reproductive health and prevention of STDs/HIV) should be initiated at this early stage in life if they are to be effective (Kirby et al. 1994). The clear implication of our data is that such messages would be relevant to the majority in primary school, including some children in class 1. And for most children in Tanzania, primary school provides the only opportunity for education because of the few places available in secondary school: of the approximately 4 million children enrolled in basic education in 1995, only 2.3% were in secondary school (United Republic of Tanzania 1996).

These data suggest that primary schools can provide a means of reaching a significant proportion of the school-age population, even in low-income countries. It has been suggested that in the most impoverished countries enrolment may be so low as to cause major inequities in access by using school-based programmes. The data for Tanzania show that although significant numbers of children could not be reached in this way in 1988, about three quarters of the children could be reached by 1994. Since universalization of education is continuing, one would expect that an even larger fraction could be reached today. The same hopeful view might be adopted with regard to gender inequity, since it would appear that girls are now better represented in initial enrolment.

Drop-out remains a major problem, particularly for girls: the rates for girls substantially exceed those of boys in the later classes, particularly if one includes the pregnancy rates. These observations give further support for the need for early and effective health education, both to ensure that girls are adequately informed before their premature departure from school and to promote reproductive health at an age when the pregnancy statistics indicate that some children are sexually active.

In conclusion, these data suggest that delivering health services through the school can provide equitable access to children even in a low-income country, and that this situation is likely to improve. Whatever else these programmes may include, it is clear that much of the target population is adolescent and would benefit from adolescent health programmes.


UKUMTA receives its principal support from the Edna McConnell Clark Foundation. The Partnership for Child Development programmes and activities are supported by UNDP, WHO, the UK Department for International Development, UNICEF, the World Bank, the Edna McConnell Clark, Rockefeller and James S McDonnell Foundations, and the Wellcome Trust.