Skip to main content

Table 1 Application of Bradford-Hill criteria: Katana, Democratic Republic of Congo

From: Assessing the impact of humanitarian assistance in the health sector

Starting in December of 2000, the International Rescue Committee (IRC) began a general health programme to support existing government services in Katana Health Zone, Democratic Republic of Congo (DRC). The IRC conducted population-based mortality surveys in this area with 345,000 mostly rural residents. The programme consisted of the provision of drugs, supplies, training and medical oversight in the clinics, a water provision and hygiene education programme in villages with the highest rates of cholera in 2000, a measles immunisation and vitamin A provision campaign, and support to the local health committees which included the donation of vouchers for the most indigent community members. Figure 1 below shows the crude mortality rate (CMR) over the period covered by 5 surveys conducted between 1999 and 2002. IRC claims to have reduced the excess CMR by 60% (from 4.9 to 2.8 deaths per 1000 per month where the baseline is assumed to be 1.5) during the period from 6 to 12 months after implementation and by 70% (from 2.8 to 1.9 deaths per 1000 per month) over the period from 12 to 24 months after implementation. In support of the results in figure 1 being a consequence of the health programme, IRC reported that:
• attendance at the clinic rose by 147% between 1999 (~7400 visits per month) and 2001 (~18,300 visits per month average)
• 70% of treatments were for malaria and diarrhoea, the main reported causes of death in the 1999 and 2000 surveys, and decreased as a cause of death in 2001 & 2002
• CMR in the five eastern provinces of DRC was estimated by IRC to have increased slightly in 2001 compared to 2000
• A survey in November of 2001 found that 60% of residents that had experienced fever in the preceding two weeks had sought treatment at a clinic
Employing Bradford-Hill's criteria, this example shows that: 1) there was a considerable drop in CMR associated with the establishment of the intervention, 2) there was no dose-response effect, 3) the fact that IRC's two other areas of health programmes had similar (but somewhat less dramatic) reductions implies repeatability, 4) the benefit occurred after implementation, 5) the findings are biologically plausible (although 1 visit per resident per year seems low), 6) alternative explanations for the reductions cannot be ruled out given the variance over time and the dramatic changes in violent conflict, although IRC reports that the violence did not dramatically subside until 2002, 7) these are not experimental data. Finally, the fact that the CMR was measured by an apparently valid survey method implies that IRC probably did contribute to a reduction in mortality in Katana [26].