Objective To determine whether hospital controls could be used in case-control studies where resource constraints limit recruitment of community controls.

Methods Hospital controls and community controls were compared in terms of socio-demographic and risk factor variables in a study of smokeless tobacco (SLT) and coronary heart disease (CHD) in Bangladesh in 2010. Incident cases of CHD were selected from two cardiac hospitals. Hospital controls were selected from outpatient departments of the same hospitals. Community controls were selected from neighbourhoods matched to CHD cases. Four community controls and one hospital control were matched to each case on age and gender.

Results The study enrolled 302 cases, 1208 community controls and 302 hospital controls. There were no differences between hospital controls and community controls with respect to age, gender, marital status, occupation, economic status and risk factors for CHD. Hospital controls were more educated but less active physically than community controls. Current use of SLT was similar among community controls (33%) and hospital controls (32%), which was also not significant statistically (adjusted OR 0.81, 95% CI 0.58 to 1.12, p>0.05). Current use of SLT was not associated with increased risk of CHD when data from community controls were used (adjusted OR 0.87, 95% CI 0.63 to 1.19, p>0.05), nor when data from hospital controls were used (adjusted OR 1.00, 95% CI 0.63 to 1.60, p>0.05).

Conclusion For comparable future studies in situations of resource scarcity or difficult socio-political context, it is possible to enrol hospital controls with careful planning which are similar to potential community controls minimising bias.