Is the distance a patient lives from hospital a risk factor for death from tuberculosis in rural South Africa?

SETTING: Tuberculosis programmes that rely on district hospitals for diagnosis and initiation of treatment may disadvantage those living furthest away. We present an analysis of such a programme in rural South Africa to see whether those living furthest from the hospital were at greatest risk of dying from tuberculosis.
METHODS: All patients diagnosed and treated for tuberculosis in three health districts in 1997 and 1998 were included. An estimate of the distance each patient travelled to get to the hospital was obtained. The distances the patients travelled were categorised into four groups. The furthest distance patients could reasonably be expected to travel to get to their nearest hospital was estimated as 60 km. Outcomes of treatment were recorded using standard definitions. The mortality of patients in each of the four groups was compared.
RESULTS: Of 1187 patients started on treatment for tuberculosis in the hospitals, 877 (74%) were known to be alive at the end of treatment, whereas 158 (13%) had died. Distance travelled was a risk factor for death, but only amongst those travelling more than 60 km to get to the hospital (0-20 km: n = 313, odds ratio [OR] 1; >20-40 km: n = 436, OR 1.09, 95% confidence interval [CI] 0.71-1.67; >40-60 km: n = 205, OR 0.97, 95%CI 0.57-1.65; >60 km: n = 79, OR 2.87, 95%CI 1.59-5.17).
CONCLUSION: The mortality from tuberculosis was high, even amongst those living closest to the hospital, and did not rise significantly within 60 km. The situation may be different for the relatively small number of patients who come from further away. The distance travelled to hospital for initial diagnosis does not account for the relatively high mortality amongst tuberculosis patients in this area.