Dropouts in a long-term follow-up study involving voluntary counseling and HIV testing: experience from a cohort of police officers in Dar Es Salaam, Tanzania.

BACKGROUND: Results of most population-based studies primarily are derived from people who responded positively and thereby continued to participate in such studies. It is, however, equally important to know the characteristics of study subjects who drop out to learn the reasons that kept them from continuing to participate in the study, especially because they had initially agreed to participate in such a study. In studies with long-term follow-up, reasons for nonresponse may provide invaluable information that may be gathered through continued contact with study subjects who have withdrawn from the study. OBJECTIVES: To determine characteristics of study participants who withdrew from an ongoing study of police officers, which involved counseling and HIV testing, and to determine reasons for their discontinued participation. METHODS: Demographic characteristics of a cohort of police officers who had been participating in a study to determine their suitability for HIV vaccine trials were analyzed. Characteristics of those who did not return for the second survey of appointments for HIV testing were compared with those who continued their participation. A randomly selected sample of 132 police officers who did not participate in the second survey of HIV testing were asked why they did not return. Answers were obtained from 84 people who had discontinued their participation. RESULTS: Of eligible police officers, 2087 (72.1%) responded to the call for follow-up appointments, whereas 807 (27.9%) did not return. Those who did not return to participate in the second survey had significantly higher rates of HIV seropositivity (17.2%) than those who did return (13.5%) (p <.05). The rate of return in unmarried participants was worse (p <.05) than the rate among married participants. Rates of sexual contacts with partners other than their spouses and levels of alcohol consumption did not differ between the two groups. Reasons for dropping out of the study included fear of knowing results of HIV testing in 54.6%, lack of time to continue in 34.5%, and fears about job security in 3.6%. CONCLUSION: Fears of finding out that one might be seropositive need to be answered at recruitment, and practical arrangements must be made to facilitate further follow-up. A bias for lower incidence might be introduced in vaccine trials if participants thought to be at highest risk for HIV infection discontinue participation.

The prevalence and incidence of HIV-1 infection and syphilis in a cohort of police officers in Dar es Salaam, Tanzania: a potential population for HIV vaccine trials.

OBJECTIVES: To assess the suitability of a cohort of police officers in Dar es Salaam for HIV vaccine trials by determining the prevalence and incidence of HIV-1 infection, active syphilis and their associated factors. DESIGN AND SETTING: An open cohort study of police officers in Dar es Salaam, Tanzania. METHODS: Recruitment of police officers began in 1994. A standardized questionnaire was completed at enrolment and subsequent visits. HIV antibodies were determined using two consecutive enzyme-linked immunosorbent assays. Samples repeatedly discordant on the two tests were tested by a Western blot assay. Treponema pallidum antibodies were first determined by Venereal Disease Research Laboratory (VDRL) test and reactive sera were confirmed by Treponema pallidum hemagglutination test. RESULTS: At the end of 1996 a total of 2850 police officers had been recruited of whom 2733 (96%) consented to be tested for HIV. The overall HIV-1 seroprevalence at recruitment was 13.8% (378 of 2733). Females had a significantly higher HIV-1 seroprevalence, 18.0% (55 of 306), as compared to males, 13.3% (323 of 2427), P< 0.05. From a total of 2215 married police officers, 585 (26.4%) responded to a question on extramarital sex within the previous 3 months of whom 36.2% (212 of 585) admitted to have had at least one extramarital sexual intercourse. Condoms were not used during these encounters by 178 of 212 (84.0%). As of 31st December 1998, among the 1524 males observed for 2553 person-years (PYAR), 50 had seroconverted and among 200 females observed for 357 PYAR, eight had seroconverted. The overall crude HIV-1 incidence was thus 19.9/1000 PYAR; 19.6 and 22.4/1000 PYAR for males and females, respectively. The overall prevalence and incidence of active syphilis were 3.1% (88 of 2850) and 8.6/1000 PYAR (26 of 3149), respectively. Males had a higher prevalence of active syphilis, 84 of 2525 (3.3%) than females, five of 325 (1.5%), P = 0.09. CONCLUSIONS: There was high risk sexual practice including low condom use in this cohort of police officers. The incidence and prevalence of HIV infection were high. Police officers in Dar es Salaam are therefore a potential population group for HIV vaccine evaluation.

Isoniazid prophylaxis for tuberculosis prevention among HIV infected police officers in Dar es Salaam.

OBJECTIVE: To determine the acceptability, compliance and side effects of isoniazid (INH) prophylaxis against tuberculosis among HIV infected police officers (PO) in Dar es Salaam. DESIGN: A nested study from a prospective follow up of a cohort of police officers. SETTING: Dar es Salaam, Tanzania. SUBJECTS: One hundred and forty three HIV-1 infected police officers. MAIN OUTCOME MEASURES: Acceptance and compliance to INH prophylaxis. RESULTS: Of the 400 HIV-1 infected officers, 143 (35.7%) came forward for post-test counselling and HIV test results. Sixty per cent (87/143) of them accepted to be on INH prophylaxis. However only 42.5% (37/87) came forward for evaluation regarding their suitability for INH prophylaxis. During the evaluation, eight (21.6%) of 37 otherwise asymptomatic PO were found to have active pulmonary tuberculosis (TB). Eventually only 29 PO were actually started on INH, and only 16 (55.2%) of them completed the six months course. No serious side effects were observed. One PO developed TB two months after loss to follow up before completing the six months. CONCLUSIONS: There was low acceptability of and poor compliance with INH prophylaxis among the HIV-1 infected PO despite being educated on the benefits of prophylaxis. The prevalence of PTB among asymptomatic HIV-1 infected PO was high, and therefore persons with HIV infection should be examined for TB even in the absence of symptoms.

Clostridium perfringens septicemia with massive hemolysis in a patient with Hodgkin's lymphoma.

A 55-year-old woman with stage IV-B nodular sclerosing Hodgkin's lymphoma presented to the emergency department with fever and lethargy of 12 hours' duration. The patient developed massive intravascular hemolysis secondary to Clostridium perfringens sepsis and cardiac arrest unresponsive to transfusions and cardiac pulmonary resuscitation, and died within 4 hours of presentation. The differential diagnosis of massive intravascular hemolysis, as well as the pathogenesis and treatment of C perfringens-induced hemolysis, are discussed.

Blood pressure changes in diabetes in urban Tanzania.

Little is known of the natural history of blood pressure (BP) levels in diabetic patients from sub-Saharan Africa. BP levels were therefore recorded in such patients in Dar es Salaam, Tanzania, over 2, 5, and 7 years. Hypertension was found in 5% of insulin-treated diabetes mellitus (IDDM) and 29.2% of non-insulin-dependent diabetes mellitus (NIDDM) patients at presentation with diabetes. Hypertension developed in a further 2 IDDM (3.7%) and 27 NIDDM (15.6%) patients at 2 years, and in 3 IDDM (13.0%) and 9 NIDDM (9.8%) patients at 5 years. Seven NIDDM (18.4%) patients had developed hypertension by 7 years. In NIDDM patients with normal BP initially, the mean systolic BP rose from 131 to 141 mmHg (P < 0.001) 2 years later (n = 146); from 131 to 138 mmHg (P < 0.001) for those followed for 5 years (n = 82); and from 131 to 138 mmHg (P < 0.05) for those followed for 7 years (n = 31). The mean diastolic BP was 83 mmHg initially and 84 mmHg (NS) for those followed for 2 years (n = 146). There was no observed rise in mean diastolic BP at 5 or 7 years of follow-up. In IDDM patients without hypertension, only the systolic BP rose significantly by 5 years, from 124 to 132 mmHg (P < 0.001; n = 20). These changes were independent of age, sex, body mass index, and proteinuria.(ABSTRACT TRUNCATED AT 250 WORDS)

Must diabetes be a fatal disease in Africa? Study of costs of treatment.

OBJECTIVE: To estimate the costs of diagnosis and treatment of diabetes in Tanzania. DESIGN: Costs estimated from the reported and recorded experience of patients with newly presenting diabetes in 1989-90 and of diabetic patients first seen in 1981-2. SETTING: Muhimbili Medical Centre, Dar es Salaam. SUBJECTS: 464 patients (315 men and 149 women). 262 patients diagnosed during 1 September 1989-31 August 1990 (group 1) and 202 during 1 June 1981-31 August 1982 (group 2). RESULTS: The average annual direct cost of diabetes care in 1989-90 was $287 for a patient requiring insulin and $103 for a patient not requiring insulin. Purchase of insulin accounted for 68.2% ($156) of the average annual outpatient costs for patients requiring insulin. For patients not requiring insulin the cost of oral hypoglycaemic drugs and treatment of chronic complications and infections accounted for 42.5% ($29.3) and 48.8% ($33.7) of costs respectively. Cost of outpatient care of diabetic patients for the whole of Tanzania was estimated at $2.7m, *75,128 (32.2%) of which was for insulin. Doctors' and nurses' costs accounted for 0.2% of total costs of outpatient care. The annual direct inpatient care costs were estimated at $1.25m. Around 0.2% of the Tanzanian population aged 15 years and over used the equivalent of 8% of the total government health expenditure, which was $47,4088,382. CONCLUSION: Diabetes places a severe strain on the limited resources of developing countries. If African patients with diabetes have to pay for their treatment most will be unable to do so and will die.