BACKGROUND: Estimates of the burden of disease in adults in sub-Saharan Africa largely rely on models of sparse data. We aimed to measure the burden of disease in adults living in a rural area of coastal Kenya with use of linked clinical and demographic surveillance data. METHODS: We used data from 18,712 adults admitted to Kilifi District Hospital (Kilifi, Kenya) between Jan 1, 2007, and Dec 31, 2012, linked to 790,635 person-years of observation within the Kilifi Health and Demographic Surveillance System, to establish the rates and major causes of admission to hospital. These data were also used to model disease-specific disability-adjusted life-years lost in the population. We used geographical mapping software to calculate admission rates stratified by distance from the hospital. FINDINGS: The main causes of admission to hospital in women living within 5 km of the hospital were infectious and parasitic diseases (303 per 100,000 person-years of observation), pregnancy-related disorders (239 per 100,000 person-years of observation), and circulatory illnesses (105 per 100,000 person-years of observation). Leading causes of hospital admission in men living within 5 km of the hospital were infectious and parasitic diseases (169 per 100,000 person-years of observation), injuries (135 per 100,000 person-years of observation), and digestive system disorders (112 per 100,000 person-years of observation). HIV-related diseases were the leading cause of disability-adjusted life-years lost (2050 per 100,000 person-years of observation), followed by non-communicable diseases (741 per 100,000 person-years of observation). For every 5 km increase in distance from the hospital, all-cause admission rates decreased by 11% (95% CI 7-14) in men and 20% (17-23) in women. The magnitude of this decline was highest for endocrine disorders in women (35%; 95% CI 22-46) and neoplasms in men (30%; 9-45). INTERPRETATION: Adults in rural Kenya face a combined burden of infectious diseases, pregnancy-related disorders, cardiovascular illnesses, and injuries. Disease burden estimates based on hospital data are affected by distance from the hospital, and the amount of underestimation of disease burden differs by both disease and sex. FUNDING: The Wellcome Trust, GAVI Alliance.

BACKGROUND: Despite the publication of several studies on the subject, there is significant uncertainty regarding the burden of disease among adults in sub-Saharan Africa (sSA). OBJECTIVES: To describe the breadth of available data regarding causes of admission to hospital, to systematically analyze the methodological quality of these studies, and to provide recommendations for future research. DESIGN: We performed a systematic online and hand-based search for articles describing patterns of medical illnesses in patients admitted to hospitals in sSA between 1950 and 2010. Diseases were grouped into bodily systems using International Classification of Disease (ICD) guidelines. We compared the proportions of admissions and deaths by diagnostic category using chi2. RESULTS: Thirty articles, describing 86,307 admissions and 9,695 deaths, met the inclusion criteria. The leading causes of admission were infectious and parasitic diseases (19.8%, 95% confidence interval [CI] 19.6-20.1), respiratory (16.2%, 95% CI 16.0-16.5) and circulatory (11.3%, 95% CI 11.1-11.5) illnesses. The leading causes of death were infectious and parasitic (17.1%, 95% CI 16.4-17.9), circulatory (16%, 95% CI 15.3-16.8) and digestive (16.2%, 95% CI 15.4-16.9). Circulatory diseases increased from 3.9% of all admissions in 1950-59 to 19.9% in 2000-2010 (RR 5.1, 95% CI 4.5-5.8, test for trend p<0.00005). The most prevalent methodological deficiencies, present in two-thirds of studies, were failures to use standardized case definitions and ICD guidelines for classifying illnesses. CONCLUSIONS: Cardiovascular and infectious diseases are currently the leading causes of admissions and in-hospital deaths in sSA. Methodological deficiencies have limited the usefulness of previous studies in defining national patterns of disease in adults. As African countries pass through demographic and health transition, they need to significantly invest in clinical research capacity to provide an accurate description of the disease burden among adults for public health policy.

BACKGROUND: The most common method for determining cause of death is certification by physicians based either on available medical records, or where such data are not available, through verbal autopsy (VA). The physician-certification approach is costly and inconvenient; however, recent work shows the potential of a computer-based probabilistic model (InterVA) to interpret verbal autopsy data in a more convenient, consistent, and rapid way. In this study we validate separately both physician-certified verbal autopsy (PCVA) and the InterVA probabilistic model against hospital cause of death (HCOD) in adults dying in a district hospital on the coast of Kenya. METHODS: Between March 2007 and June 2010, VA interviews were conducted for 145 adult deaths that occurred at Kilifi District Hospital. The VA data were reviewed by a physician and the cause of death established. A range of indicators (including age, gender, physical signs and symptoms, pregnancy status, medical history, and the circumstances of death) from the VA forms were included in the InterVA for interpretation. Cause-specific mortality fractions (CSMF), Cohen's kappa (kappa) statistic, receiver operating characteristic (ROC) curves, sensitivity, specificity, and positive predictive values were applied to compare agreement between PCVA, InterVA, and HCOD. RESULTS: HCOD, InterVA, and PCVA yielded the same top five underlying causes of adult deaths. The InterVA overestimated tuberculosis as a cause of death compared to the HCOD. On the other hand, PCVA overestimated diabetes. Overall, CSMF for the five major cause groups by the InterVA, PCVA, and HCOD were 70%, 65%, and 60%, respectively. PCVA versus HCOD yielded a higher kappa value (kappa = 0.52, 95% confidence interval [CI]: 0.48, 0.54) than the InterVA versus HCOD which yielded a kappa (kappa) value of 0.32 (95% CI: 0.30, 0.38). Overall, (kappa) agreement across the three methods was 0.41 (95% CI: 0.37, 0.48). The areas under the ROC curves were 0.82 for InterVA and 0.88 for PCVA. The observed sensitivities and specificities across the five major causes of death varied from 43% to 100% and 87% to 99%, respectively, for the InterVA/PCVA against the HCOD. CONCLUSION: Both the InterVA and PCVA compared well with the HCOD at a population level and determined the top five underlying causes of death in the rural community of Kilifi. We hope that our study, albeit small, provides new and useful data that will stimulate further definitive work on methods of interpreting VA data.