Validation of the Kiswahili version of the SF-36 Health Survey in a representative sample of an urban population in Tanzania.

The objective of this study was to assess the validity of a Kiswahili translation of the SF-36 Health Survey (SF-36) among an urban population in Tanzania, using the method of known-groups validation. People were randomly selected from a demographic surveillance system in Dar es Salaam. The representative sample consisted of 3,802 adults (15 years and older). Health status differences were hypothesized among groups, who differed in sex, age, socioeconomic status and self-reported morbidity. Mean SF-36 scale scores were calculated and compared using t-test and ANOVA. Women had significantly lower mean SF-36 scale scores (indicating worse health status) than men on all scales and scores were lower for older people than younger on all domains, as hypothesized. On five of the eight SF-36 scales, means were higher for people of higher socioeconomic status compared to those of lower socioeconomic status. People who reported an illness within the previous 2 weeks scored significantly lower on all scales compared to those who were healthy, as did people who said they had a disability or a chronic condition.

A Kiswahili version of the SF-36 Health Survey for use in Tanzania: translation and tests of scaling assumptions.

The objective of the study was to translate and adapt the SF-36 Health Survey for use in Tanzania and to test the psychometric properties of the Kiswahili SF-36. A cross-sectional study was conducted as part of a household survey of a representative sample of the adult population of Dar es Salaam, Tanzania. The IQOLA method of forward and backward translation was used to translate the SF-36 into Kiswahili. The translated questionnaire was administered by trained interviewers to 3,802 adults (50% women, mean (SD) age 31 (13) years, 50% married and 60% with primary education). Data quality and psychometric assumptions underlying the scoring of the eight SF-36 scales were evaluated for the entire sample and separately for the least educated subgroup (n = 402), using multitrait scaling analysis. Forward and backward translation procedures resulted in a Kiswahili SF-36 that was considered conceptually equivalent to the US English SF-36. Data quality was excellent: only 1.2% of respondents were excluded because they answered less than half of the items for one or more scales; ninety percent of respondents answered mutually exclusive items consistently. Median item-scale correlations across the eight scales ranged from 0.47 to 0.81 for the entire sample. Median scaling success rates were 100% (range 87.5-100.0). The median internal consistency reliability of the eight scales for the entire sample was 0.81 (range 0.70-0.92). Floor effects were low and ceiling effects were high on five of the eight scales. Results for n = 402 people without formal education did not differ substantially from those of the entire sample. The results of data quality and psychometric tests support the scoring of the eight scales using standard scoring algorithms. The Kiswahili translation of the SF-36 may be useful in estimating the health of people in Dar es Salaam. Evidence for the validity of the SF-36 for use in Tanzania needs to be accumulated.

The right objectives in health care planning.

In Dar es Salaam, United Republic of Tanzania, the traditional epidemiological approach to health service planning has been superseded by a process-oriented approach. The implications for managers are discussed below.

PIP: Health planning in Tanzania has been based upon epidemiological considerations, with population needs for health services expressed in terms of health status and disease profiles. In 1989, however, a project was launched with Tanzanian and Swiss government funding to improve the physical and functional status of the public health system in Dar es Salaam. The aims were to strengthen management capacities at the city and district levels, to rehabilitate the health service infrastructure, which had deteriorated since the early 1980s because of reduced resources and population growth, and to develop a health care strategy for the city. A process-oriented approach to health service planning is now in place in Dar es Salaam. Selected indicators are not epidemiological, but are pragmatic process markers intended to allow targets to be set which can easily be monitored. These changes are seen as a step toward improving service delivery.

Perinatal mortality in rural Tanzania.

In 1987, as part of the World Health Organization's Inter-Health Programme, we carried out a noncommunicable diseases survey in six rural villages in Tanzania. Each women completed a questionnaire, part of which was concerned with obstetric history, and underwent a physical examination. Our findings of unacceptably high perinatal mortality and abortion rates are described below.

PIP: Within the framework of the World Health Organization's [WHO] Inter- Health Program, a noncommunicable diseases survey was carried out in 6 villages located in 2 regions of Tanzania in 1987. A total of 3565 women were queried about their viable pregnancies, still-births, perinatal deaths (death within the first week of life), and abortions (loss of conceptus within 28 weeks of gestation). The overall abortion rate was 119/1000 pregnancies (range of 89/1000-170/1000 pregnancies). In the Kilimanjaro region, the abortion rate was 97/1000 compared to 145/1000 in the Morogoro region. The overall perinatal mortality rate was 73/1000 births (range of 49/1000-124/1000 births). In the Kilimanjaro region, the perinatal mortality rate was 58/1000 births vs. 91/1000 in the Morogoro region. The overall abortion rate for the 6 villages was 12% of pregnancies, while the range culled from the literature was 10-25%. The rate of miscarriage before 20 weeks of gestation has been estimated at 36-43%. In the Morogoro region, miscarriage rates were highest in the older age groups, which is indicative of improved health care. The difference in perinatal mortality rates between the 2 regions may be attributable to the relatively higher level of development of the Kilimanjaro region. This difference in socioeconomic development was also reflected in the infant mortality rate: 70/1000 births in the Kilimanjaro compared to 140/1000 in the Morogoro region. Other studies in east Africa showed perinatal mortality rates of 46/1000 births in Kenya's Machakos region during 1975-78, and a rate of 124/1000 births in Tanzania near the Morogoro region during 1984-85. Community-based studies focusing on women's health and care during pregnancy and childbirth are needed to investigate the causes of perinatal deaths, because hospital-based studies give a disjointed picture of both rates and the causes.

Coronary heart disease risk factors in sub-Saharan Africa: studies in Tanzanian adolescents.

STUDY OBJECTIVE: To assess the level of cardiovascular risk factors in young people in sub-Saharan Africa living in rural and urban settings. DESIGN: Cross sectional survey of the population aged 15 to 19 years. SETTING: Eight rural Tanzanian villages in three regions, and two districts in Dar es Salaam. PARTICIPANTS: 664 males and 803 females in rural villages and 85 males and 121 females in the city. Response rates for total population were 74% to 94% in the rural areas and 60% in the city. MEASUREMENTS AND RESULTS: Measurements included blood pressure, body mass index, serum lipids, and blood glucose concentrations (fasting and two hours after 75 g glucose). Blood pressure was slightly but significantly higher in young women than in young men (115/67 mmHg versus 113/65 mmHg) and increased significantly with age. Only 0.4% subjects had blood pressure greater than 140 and/or 90 mmHg. There were no urban-rural differences. Body mass index was higher in females (mean (SD) 20.3 (2.8) kg/m2) than males (18.5 (2.1)). Overweight was found in only 0.6% at age 15 years but 5.4% at age 19 years. Serum cholesterol concentrations were low at 3.5 mmol/l in males and 3.7 mmol/l in females. Only 7% had values above 5.2 mmol/l. The highest concentrations were found in the city and in Kilimanjaro, the most prosperous rural region. Serum triglycerides were 1.0 (0.5) mmol/l in males and 1.1 (0.5) mmol/l in females, and were highest in the city dwellers. Diabetes was rare (0.28% males, 0.12% females) but impaired glucose tolerance was present in 4.7% and 4.1% respectively. Drinking alcohol was equally prevalent in males and females, reaching 30% at age 19 years. Only 0.4% of females smoked compared with 7.3% of males. Smoking was commoner in rural areas that in the city. CONCLUSIONS: Several risk factors for cardiovascular disease were found in Tanzanian adolescents, but levels were much lower than in studies reported from developed nations. The challenge is to maintain these low levels as the population becomes more urbanised and more affluent.

Low prevalence of risk factors for coronary heart disease in rural Tanzania.

A community-based survey was used to assess the prevalence of risk factors for coronary heart disease (CHD) in rural Tanzanians. In all, 8581 subjects (3705 men, 4876 women) aged > or = 15 years in eight villages in three regions in rural Tanzania representing a range of socioeconomic deprivation were studied. The main outcome measures were serum cholesterol and triglyceride level, blood pressure and prevalence of dyslipidaemia, hypertension, smoking, overweight, impaired glucose tolerance (IGT) and diabetes; as well as ECG changes. Mean serum cholesterol levels in men were 4.2, 3.4 and 3.7 mmol/l, and in women 4.4 3.6 and 3.9 mmol/l in Kilimanjaro, Morogoro and Mara regions respectively. In Kilimanjaro region 17.4% of men and 19.0% of women had values above 5.2 mmol/l compared with only 5.0% and 6.7% in Morogoro region and 4.8% and 6.9% respectively in Mara region. Systolic and diastolic blood pressures increased with age in both men and women in all three regions with the most marked increase in Kilimanjaro region and the smallest rise in Mara region. Mean age-adjusted values were highest in Kilimanjaro region (124/75 mm Hg, and 125/76 mm Hg in men and women respectively) and lowest in Mara region (120/70 mm Hg in men and 118/68 mm Hg in women). Hypertension was found in 6.6% of men and 7.5% of women in Kilimanjaro region, 3.3% and 4.7% in Morogoro, and 2.6% and 3.4% in Mara region. Cigarette smoking was found in 42.6% of men in Kilimanjaro region, 28.2% of Morogoro region and 8.6% in Mara region. Less than 4% of women smoked in all three regions.(ABSTRACT TRUNCATED AT 250 WORDS)

Prevalence and incidence of epilepsy in Ulanga, a rural Tanzanian district: a community-based study.

A random cluster sample survey of approximately 18,000 people in 11 villages was performed in Ulanga, a Tanzanian district with a population of approximately 139,000 people. Well-instructed fourth-year medical students and neurologic and psychiatry nurses identified persons with epilepsy using a screening questionnaire and sent them to a neurologist for detailed evaluation. Identified were 207 subjects (88 male, 119 female) with epilepsy; of these, 185 (89.4%) (80 male, 105 female) had active epilepsy. The prevalence of active epilepsy was 10.2 in 1,000. Prevalence among villages varied, ranging from 5.1 to 37.1 in 1,000 (age-adjusted 5.8-37.0). In a 10-year period (1979-1988) 122 subjects living in the 11 villages developed epilepsy, with an annual incidence of 73.3 in 100,000. Generalized tonic-clonic seizures (GTCS) accounted for 58% and partial seizures accounted for 31.9%, whereas in 10.1% seizures were unclassifiable. Of the partial seizures, secondarily generalized seizures were the most common. Possible etiologic or associated factors were identifiable in only 25.3% of cases. Febrile convulsions were associated in 13.4 of cases. Other associated factors included unspecified encephalitis (4.7%), cerebral malaria (1.9%), birth injury (1.4%), and other (3%). In 38% of the cases, there was a positive family history of epilepsy.

Is diabetes mellitus related to undernutrition in rural Tanzania?

OBJECTIVE: To investigate the relation between undernutrition and diabetes. DESIGN: Survey of glucose tolerance in rural Tanzania. SETTING: Eight villages in three widely separated regions of Tanzania. SUBJECTS: 8581 people aged 15 and above: 3705 men and 4876 women. MAIN OUTCOME MEASURES: Oral glucose tolerance, body mass index, height, and low haemoglobin and cholesterol concentrations. RESULTS: In the eight villages 42.7-56.9% of all men and 30.0-45.2% of all women had a body mass index below 20 kg/m2; the lowest quintile was 18.2 kg/m2 in men and 18.6 kg/m2 in women. The prevalence of diabetes did not change significantly from the lowest to the highest fifths of body mass index in men (lowest 1.6% (95% confidence interval 0.8% to 2.9%) v highest 1.3% (0.7% to 2.5%)) or women (1.1% (0.6% to 2.1%) v 0.5% (0.2% to 1.2%)). In men and in women prevalence of impaired glucose tolerance was greater in the lowest fifths of height (8.2% (6.3% to 10.6%), and 11.1% (9.2% to 13.3%)) respectively and body mass index (9.6% (7.5% to 12.1%), and 8.4% (6.7% to 10.5%)) than in the highest fifths (impaired glucose tolerance 4.7% (3.4% to 6.5%); and 5.1% (3.9% to 6.7%); body mass index 5.1% (3.7% to 7.0%), and 7.7% (6.2% to 9.6%). CONCLUSION: Rates of diabetes were not significantly associated with low body mass index or height, but overall rates were much lower than those in well nourished Western populations. Increased impaired glucose tolerance in the most malnourished people may reflect the larger glucose load per kilogram weight. The role of undernutrition in the aetiology of diabetes must be questioned.

No deterioration of oral glucose tolerance during pregnancy in rural Tanzania.

There is still controversy concerning the reference ranges for glucose tolerance tests in pregnancy. The WHO has recommended the universal use of the 75 g oral glucose load with 2-h post-load values of greater than 6.7 mmol l-1 to be considered impaired glucose tolerance (IGT) in the non-pregnant, and equivalent to gestational diabetes in the pregnant. Some data are available for pregnant Caucasians but little information is available for other ethnic groups. Oral glucose tolerance tests (75 g) have therefore been performed in 189 pregnant women in rural Tanzania. Mean fasting blood glucose values were 4.0 mmol l-1 in non-pregnant women, and 3.7, 3.5, and 3.3 mmol l-1 in pregnant women in the first, second, and third trimesters, respectively. Two-hour OGTT values were 4.7 mmol l-1, and 4.6, 4.5, and 4.2 mmol l-1 while the upper limit of normal values (mean + 2SD) were 7.1 mmol l-1, and 6.8, 6.8, and 6.1 mmol l-1. The 2-h glucose levels are therefore close to WHO recommendations but lower than those reported for Caucasians. By contrast with reports for Caucasians, glucose tolerance did not deteriorate during pregnancy. The prevalence of diabetes and IGT was zero in the pregnant group.

Study in Tanzania of impaired glucose tolerance. Methodological myth?

During a study of diabetes prevalence in six rural Tanzanian communities, a repeat oral glucose tolerance test (OGTT) was carried out in 514 subjects greater than or equal to 15 yr of age within 1 wk of an initial 75-g OGTT. In 498 subjects, blood glucose was measured 2 h after the glucose load on both occasions, and in 175 subjects, fasting blood glucose measurement was also repeated. Of the 498 subjects, 245 had normal glucose tolerance in the first test and were selected at random for further testing; 223 subjects had impaired glucose tolerance (IGT), and 30 had diabetic values. Diabetes and IGT were diagnosed on the basis of the 2-h blood glucose values. In the second test, 241 (98.4%) of the 245 subjects with normal tolerance continued in this category and 4 (1.6%) showed IGT. Of the 223 with IGT in the first test, 171 (76.2%) reverted to normal on the second test, 7 (3.1%) had diabetic values, and 45 (20.2%) persisted with IGT. Of the 30 subjects diagnosed as diabetic in the first test, 8 (26.7%) remained with diabetic values, 11 (36.7%) had IGT, and 11 (36.7%) were normal. Based on the second test, the population-prevalence rates of diabetes and IGT would have been 0.5 and 3.3% vs. 1 and 7.6% based on the first test. There was a significant downward trend in the mean 2-h blood glucose values in all three diagnostic groups. Regression toward the mean could not account for the downward shift in blood glucose values observed on retesting.(ABSTRACT TRUNCATED AT 250 WORDS)

Prevalence of diabetes and impaired glucose tolerance in rural Tanzania.

The prevalence of diabetes mellitus and impaired glucose tolerance (IGT) was assessed by use of WHO diagnostic criteria in 6299 Africans aged 15 years and above living in six villages in Tanzania. 0.87% (1.1% male, 0.68% female) had diabetes and 7.8% (6.9% male, 7.7% female) had IGT. Prevalence rates were 1.1% and 8.4%, respectively, when age-adjusted to the USA population. Only 7 (13.5%) of the 53 individuals with diabetes had been known to have the disorder; 34 (74%) of the other 46 were symptom-free. Mean age was 54 (SD 20) for diabetic subjects and 37 (17) years for the whole population. Diabetes and IGT rates did not differ significantly between villages despite geographical, socioeconomic, and dietary differences. Diabetes rates increased modestly with age and body mass index (BMI). Fasting blood glucose (FBG) levels did not rise significantly with age but correlated positively with systolic blood pressure (BP) and negatively with haemoglobin concentration (Hb) and BMI. The 2 hour post-glucose load blood glucose values correlated positively with age, sex, and systolic BP and negatively with Hb. Diabetes is less prevalent in rural Africa than in developed countries, even when age has been corrected for. This difference is probably related to body weight, diet, and exercise.