by A. Somanathan, K. Hanson, T. Dorabawila and B. Perera
HPP Occasional Paper Series

Sri Lanka is well known for having achieved very good health outcomes at low cost. Analysis of a survey of health facilities in 1991 found that average costs of care in public sector health facilities were very low by international standards. Nonetheless, considerable variation was identified among facilities offering similar services, suggesting that there is potential for improving efficiency. The objectives of this study were to (i) to explore different methods for quantifying the magnitude of technical and economic inefficiency in service provision by public sector providers and (ii) to identify institutional and behavioural factors which explain difference in efficiency.

A variety of techniques were used to quantify the extent of inefficiency in service provision, including standard service indicators (length of stay, occupancy rate, turnover rate) average cost, and econometric cost and production functions. The results of the different methods were compared using rank correlation coefficients. Lasso diagrams were used to compare the relative performance of facilities. Other potential correlates of facility performance studied included a series of management indicators, which describe the characteristics of the facility manager, the system used for managing key inputs such as drugs and staff, and the characteristics of the environment.

The study found that average costs of care in 1997 continued to be below international norms, but that there remained an important degree of variation among similar facilities, with ratios of high: low cost facilities ranging from 4.3 (for cost per patient day in complex inpatient facilities) to almost 30 (for outpatient visits in basic inpatient facilities). Differences in average length of stay and occupancy rate explain only a small proportion of the variation in facility cost. Indicators of management characteristics do not seem to explain much of the variation in costs either.

The findings of this study led us to question the adequacy of microeconomic approaches to efficiency for understanding the way in which public hospitals in Sri Lanka operate. The neo-classical production model relies on several assumptions such as perfect information and choice over inputs and outputs that do not necessarily hold in the context of Sri Lankan public hospitals. In a situation where budgets are fixed and demand is exogenous, unit costs are mainly demand driven and are unlikely to be adequate measures of economic efficiency at the hospital level. A macroeconomic perspective of efficiency that takes into account the equity and efficiency objectives of health planners who are responsible for resource allocation would be more effective at explaining the huge variation in unit costs and performance indicators between the same types of facilities.