Avoidable deaths in developing coutries: a study of factors contributing to death in patients admitted to the medical wards of Muhimbili Medical Centre Dar es Salaam.

Abstract

medical wards of Muhimbili Medical Centre could be regarded as "avoidable" and to identify possible contributing factors. Design: Prospective study of patients admitted to the medical wards of Muhimbili Medical Centre,Dar es Salaam over a 6- months period with information on each patients' history prior to admission being obtained from accompanying relatives. Setting: Muhimbili Medical Centre, Dar es Salaam. Subjects: 835 patients: 508 males and 327 females admitted to two medical wards between 1 July and 31 December, 1993. Results: There were 132 (15.8%) deaths. 80 (64.8%)' deaths could have been . considered "avoidable". 42 (32.1%) deaths were "unavoidable" but suffering could possibly have been alleviated and life prolonged. Only 4 deaths were considered completely "unavoidable". Of the 80 deaths which could have been considered "avoidable" 17 (21.25%) were due to all forms of tuberculosis (excluding those associated with RN infection), 10 (12.5%) were due to diabetes mel1itus; 9 (11.25%) were due to cerebral malaria, 8 (10%) were due to meningitis, 7 (8.75%) due to all forms of stroke; 6 (7.5%) to cardiomyopathy, 4 (5.0%) due to hypertension, ••

(ill) 4(5.0%) due to tetanus, 3 (3.75%) due to clinical malaria, 2 (2.5%) due to acute alcohol intoxication, 2 (2.5%) due to acute viral Hepatitis. 7 deaths grouped as others included acute gastroenteritis, renal failure, adult malnutrition, pneumonia, cellulitis, drug poisoning,and urinary outflow obstruction. Of the 42 deaths in which life could have been possibly prolonged 23 (54.8%) were due to tuberculosis associated with HIV infection 15 (37.7%) were due to AIDS, 3 (7.1 %) due to chronic liver disease and only 1 due to cancer. The 4 completely "unavoidable" deaths were due to hepatoma 2, rabies 1, noxious substance 1. The following factors related to care seeking behaviour were found to be significantly associated with death: Age (p<0.0001), male sex (p = 0.012), alcohol (p = 0.0132), Glasgow coma score (p = <0.012), Karnofsky Performance Scale (p = <0.0001), duration of illness (p = 0.0001) Duration of days in the hospital (p = 0.016), marital status for females when not married (p = 0.04), traditional healing (p = 0.0179) belief in witchcraft (p = 0.00029), delay in care seeking (p = 0.000992). For those who were ever referred to any care unit before admission they were significantly associated with death (p = 0.0062). Use of the following services was not significantly associated with death: government (p = 0.0982), private (p = 0.9852), . pharmacy or shops (p = 0.6668), past medical history (p = 0.14) current medication (p = 0.7). Neither was failure to pay for medical services significantly associated with death (p = 0.63).

(iv) Relatives' account of patients' account of care seeking behaviour matched that of the patients significantly. For services sought agreement was 92.4% for" No service" or assistance by family members, 95% for government services, 97.7% for private services, 76.6% for traditional healing and 99.5% for other services. For reasons for delay agreement was 94.7% for health service to be far, 93.0% when patient could not afford, 85.3% for belief in witchcraft and 93%' when patient was unwilling to seek care early. For outcome agreement was 96.4% for discharge on request/absconded, 84.6% for referral before admission and 70% for referral after admission. Distribution for age, sex and education by respondents was similar to those by patients. Conclusion: 64.8% of deaths are "avoidable" and in 32.1 % deaths life could be prolonged and suffering alleviated. Care seeking behaviour accounted significantly for most of these deaths. Efforts to motivate our people towards a "good" seeking behaviour is warranted. Health units should be equipped and health workers motivated. - - ---- ----