The aim of this paper is to compare the Stochastic
Frontier Analysis (SFA) and Data Envelopment Analysis (DEA) methods as two
alternative approaches to evaluate technical efficiency and efficiency factors
for the nursing home sector in Ireland. DEA is a non-parametric linear programming
frontier technique, and is the dominant
approach used to measure efficiency in healthcare. This method does not require the explicit specification of the form of the
underlying production relationship, but it has been criticised for not
considering errors due to chance, measurement errors, or environmental
differences; hence all deviations are attributed to the measured inefficiency.
Consequently, the technical efficiency measure is sensitive to outliers and is
likely to be downward biased. On the other hand, SFA requires a parametric
functional form for the production frontier. This frontier is estimated using a maximum likelihood method which
assumes that any deviation from the frontier is composed of two parts, one
representing randomness or statistical noise and the other inefficiency. However, the parametric approach might suffer from the problem of
misspecification of the functional form, and possible multicollinearity. To our knowledge, there has been very
little research on the estimation of technical efficiency in the nursing home sector using the SFA approach. Moreover,
we believe that this is the first empirical study that compares the SFA and DEA methods using long-term care data. Applying the conventional DEA and the homogeneous
bootstrap DEA procedure we found that the efficiency levels for both public and
private nursing homes in Ireland do not exceed 50 per cent on average – a
finding which corroborates the DEA results obtained for the nursing home sector in
other countries. Applying the SFA method, we find that the average technical
efficiency is much higher and ranges above 70 per cent. We also compare the
impact of environmental variables (ownership, size, age, regulatory constraints, quality of care and the case mix) on the
technical efficiency scores using the two alternative approaches and we obtain
similar results. Overall, the results indicate that the main drawback of the DEA
method is that it does not take into account the random error which leads in
our case to overestimation of inefficiency, and hence underestimation of the
technical efficiency scores. Moreover, in terms of policy
implications, our findings suggest that considerable inefficiencies exist in
the nursing home sector in Ireland and therefore more engagement with performance
measurement and improvement is needed.