The structure, function and implementation of an outcomes database at a Ugandan secondary hospital: the Mbarara Surgical Services Quality Assurance Database

Risk-Adjustment of Perioperative Mortality Rate Measurement in a Low-Income Country

BACKGROUND

The health care systems in low-income countries have extremely limited capacity to treat surgical diseases. The perioperative mortality rate has been suggested as a key quality metric to guide the expansion of care, but there is little information on how to risk-adjust this outcome measure.

METHODS

We did a 42-month observational cohort study of surgical operations at a Ugandan secondary referral hospital. We examined factors associated with in-hospital 30-day perioperative mortality outcomes. The aim of the study was to suggest a suitable indicator metric for comparative health service research in low-income countries.

RESULTS

The 30-day perioperative mortality rate was 5.3 % (n = 381/7170). The adjusted odds ratios (95% confidence interval) of variables associated with mortality were as follows: procedure (P < .001; laparotomy 2.6 [1.6, 4.3], P < .001; cranial surgery 2.8 [1.6, 4.9], P < .001); American Society of Anesthesiologists (ASA) rating 3.1 (2.6, 3.6), P < .001; HIV serostatus (P < .001; positive 2.7 [1.5, 4.8], P < .001); procedure urgency (urgent/emergent) 1.7 (1.2, 2.3), P = .003; home district location (P = .015; distant referral 1.4 [1.0, 1.9], P = .027); and age decile 1.1 (1.0,1.2, P = .001). Laparotomy was the commonest procedure performed (n = 2361) and was associated with 56.3% (n = 216/381) of deaths.

CONCLUSIONS

Laparotomy had a strong independent association with mortality at a Ugandan secondary hospital. The laparotomy perioperative mortality rate may be a suitable outcome measure for comparative health service research in low-income countries.