The results from the Tobit regression analysis performed in the second stage analysis are presented in Table 4.13 below. The dependent variables in the Tobit model are the constant returns to scale model result. The CRS model measures total efficiency with strong disposability of output (Valdmanis, Rosko and Mutter, 2008); that is, all outputs are considered desirable. Therefore, total efficiency = pure technical efficiency (VRS) *Scale efficiency* Congestion.
Table 4.13 Parameter Estimates of Tobit Regression Model
Variables Parameter
Coefficients Std Error Z-statistic Prob
a b A b a b a b
Constant 0.45 0.38 0.18 0.18 2.52 2.14 0.012 0.03
MarkCon β 0.002 0.002 0.002 0.002 1.10 1.24 0.21 0.214
Population β2 2.93E-07 4.59E-07 7.6E-07 7.51E-07 0.38 0.61 0.70 0.54
Servscope β3 -0.01 -0.003 0.023 0.021 -0.52 -0.14 0.60 0.89
Doctors β4 -0.001 -0.002 0.01 .002 -0.14 -0.86 0.89 0.39
BTR β5 0.004 0.004 0.001 0.001 3.44 3.59 0.00 0.00
Nurses Β6 - -0.001 - 0.001 - -.86 0.39
a. R2= 0.35; adj R2 =0.1764; log likelihood 0.4959; Avg log likelihood 0.0171
b. R= 0.37; adj R2 =0.198; log likelihood 0.8545; Avg log likelihood 0.0295
Generally, a positive sign of the coefficients β indicates a positive increase in efficiency while negative sign implies a reduction of efficiency. Put differently, positive coefficients are associated with efficiency increase and negative coefficients are related to decrease in efficiency. The result of the Tobit model for explaining the determinants of efficiency scores indicates that beds turnover ratio (BTR) β5, numbers of facilities offering health services in the environment proxied by MarkCon (β1) and population(β2), all have positive impact on hospital efficiency. However, only the coefficient for β5 beds turnover ratio (BTR) is statistically a significant determinant (p<.005)
The result of the Tobit model (Table 4.13) suggests that only 17.6% of the variations in the efficiency of Ogun state hospitals can be explained by the variables included in the study’s Tobit model adjusted R2 = 0.1764. This is indicative of the need to probe other variables for possible impact on hospital efficiency. However, the signs of the estimated coefficients of the explanatory variables suggest some interesting findings.
The scopes of health services offered have negative impact on efficiency (β3). This is somewhat unexpected. Expectations are that service scope or depths will affect output positively by way of increasing attendance at the facility in terms of more patients demanding the different and diverse health services offered. The β3 coefficient is not statistically significant (β3 = -0.012, p>.10), however, this suggests that most of these services are not actively demanded. Consequently, resources deployed for these services are generally idle and must be kept in expectations of future demand. This poinst to management that diversities of services do not necessarily lead to increased demand in order to ensure prudent resource use rather a new dimension to service planning for the state’s hospital might be required to ensure that varieties of services translate to intense use of health resources. An approach, for example, is for hospitals that are in close proximity to share resources or be
made to specialise along some service orientation. This will permit re-assignment of un- utilised staff and better utilisation of critical health resources in the health system
In addition, the negative impacts of service scope on efficiency scores indicate the need to consider the current method of health service planning for the state’s hospitals. There seems to be the need for more autonomy for hospital managers on changing or determination of scope of health services offered. Greater autonomy has the potential of making public hospitals to become similar to those in market system (Uslu and Linh, 2008). Furthermore, the more managerial decisions are under the control of hospital mangers, the more incentive they have to improve efficiency performance. The basis for this suggestion arises from other studies that found positive correlations between autonomy and organisational efficiency of public organisation (Perelman and Pestieau, 1988; Gathon and Perelman, 1992)
The impact of doctors on efficiency of these hospitals was unexpectedly negative on the efficiency scores of these hospitals (β4). The regression results, however, corroborate findings from our DEA models. The DEA results recommended that doctors in the state health system are in excess in some hospitals (Table 4.5). Though doctors are relatively scarce as indicated in table 1, it seems that poor resource allocation and management further limit the efficient utilisation of this critical health resource. There is however, another argument to this negative impact as it relates to the issue of dual practice which may possibly have divided the attention and loyalty of the physician. Also the states of equipments in these hospitals seem to have limited the use of doctors’ skills and intensive involvement in care procedure. Consequently, nurses seem to be more in use in care delivery. This line of thought have implications for health personnel management when costs and labour substitution in the state’s hospital system is considered
According to common wisdom in business and economic reasoning the numbers of health facilities available should increase the intensity of competition (β1). This rests on the logic that supply in the health facilities will exceed demand precipitating pressures to compete more vigorously for patients (consumers). However, from the regression analysis, its impact on the behaviour of hospital managers seem to be an incentive for public hospitals to
improve hospital efficiency, even though productivity and efficiency hardly constitute a major determinant of wage rate in the public sector. it is also possible that public hospitals enjoyed relatively better health resource advantage that positively influence health outputs in their favour. As Lambo (1989) pointed out as well patients tend to have more ‘faith’ in these secondary care level hospitals than others or the care costs are significantly lower than any other.
The impact of bed turnover ratio (BTR) is clearly significant in explaining hospital efficiency (β5 =0.005, p <0.005). Bed turnover ratio which measures the productivity of hospital beds represents the number of patients treated per hospital bed within a defined period. The impact of BTR on hospital efficiency here is, however, inconsistent with the findings of Rosko, et al, (1995). Research environment may have accounted for this since Rosko’s et al study was based on data from a developed nation with a more organized health sector. The findings in our study indicate the hospitals that produced more admissions or inpatients outputs from available inputs (beds and health personnel) have higher likelihood of being efficient. This finding implies that the higher the turnover ratio of hospital beds relative to other hospitals, the higher the efficiency of the hospital. Hospitals that admit less complicated cases which require inpatients’ less time of stay in the hospital and therefore are able to admit more patients, might seem to be more efficient.
Finally, the coefficient associated with population (β2) is positive on efficiency but statistically it is not significant. It is important to note that higher populations have more tendencies for higher demand for health care. The concentrations of health facilities to care for large population possibly foster competition and drive for prudent resource use. In any case, given the intensity of use of health services associated with large population, it is reasonable to suggest that intensive use of health resources in response to demand will have some impact. This gives empirical support to the argument for considerations of population data as major input in location of health facilities, particularly hospitals.