Level and Determinants of Economic Efficiency in Primary Public Health Care Facilities: A Study in Selected Areas of Bangladesh

Abstract

During the last three decades, the health sector of Bangladesh has made notable progress in respect of many health indicators. Contraceptive Prevalence Rate has rapidly increased; antenatal care, postnatal care, modern delivery care and immunization coverage have notably increased. Total Fertility Rate considerably declined, Crude Death Rate has declined; population growth rate has declined, life expectancy has increased. The sector, however, faces some challenges, viz., low quality of care, lack of trained manpower, absenteeism, inadequate coverage, insufficient amount of resources, and more importantly low economic efficiency. The problem of scarcity of resources and inefficiency of resource use has led to accomplishment of low performances of the sector, using the given resources the sector can achieve much higher level of outcome then it does at present. Added to this is the lack of sufficient allocation of fund for the sector. Efficient resource use basically refers to the best spending of resources by all agents involved. As such, it includes maximization of output by producers and maximization of consumer satisfaction. If resource use is inefficient, maximization of at least one party will not be achieved, implying that there will be waste of resources. As a result, utilization of resources will be lower than the amount invested- the sector will produce much less than it can using the allocated resources. Lower utilization will also cause lower quality of services and low level of client satisfaction. Furthermore, waste of resources can be a major factor to create the financing gap of the sector. Therefore, one major objective of the health sector is to accomplish highest possible efficiency in resource use in order to significantly increase coverage, improve quality of services and reduce financing gap so as to achieve the targeted goals. Measuring efficiency is important for health policy and planning. Studies suggest that as health care expenditure is large in size and grow rapidly, a small increases in the level of efficiency can lead to substantial savings of resources and increase coverage (Peacock et al. 2001). Economic loss due to inefficiency is estimated to be 20-40 percent of total health spending according to the World Health Report 2010. This means that health sectors can achieve much more with their existing resources. Inefficiencies are there in all health systems with different degrees which vary by country (Heredia-Ortiz, 2013). High efficiency can be achieved at low level of utilization. But the objective of health sector is to achieve high efficiency at high level of utilization and equity because of the externality and public good character of many health services and of the market failures to provide this to all citizens. Inefficient and underutilization of resources cause low performances of the health care facilities. Inefficient and underutilization of resources stems from both demand and supply side factors. On the supply side the factors are: absenteeism of the providers, lack of availability of complementary inputs (including supporting staff), lack of the necessary amount of fund, undue intervention from different quarters, lack of sufficient number of competent and committed manpower/provider, and allegedly insufficient attention and behavior of providers,. The main reasons on the demand side are: lack of sufficient knowledge and health consciousness, traditional values and customs, cost of time and transport, informal payment, presence of informal provider. To achieve the desired goals of the health sector of Bangladesh and in the face of the declining trends of resources, these issues of utilization of services and efficient use of available resources to maximize output need to be investigated. Against this backdrop, the present research aims at analyzing the level of efficiency of the primary level facilities in the Bangladesh health sector. The specific issue of the research is: what is the level of economic efficiency of resource use in the facilities and what are the causes of low efficiency, if any.

The major hypothesis is that available resources are not optimally utilized and the level of efficiency is much lower than what can be attained. The broad hypotheses of this study are: despite appreciable improvement in the performances of the Bangladesh health sector in the last few decades, levels of utilization of resource and efficiency are very low. Although abundant literature now exists on various aspects and issues of the health sector, and a large number of researchers hold that there are high inefficiency and underutilization of the resources in the sector, no rigorous study has so far been undertaken in the context of the Bangladesh health sector to estimate the level of inefficiency and underutilization and identify their causes. The findings of the study are expected to be highly useful to the policy makers and health economists. The magnitudes of inefficiency and underutilization will unravel the extent to which the sector can further improve the performances with the allocated resources and the determinants identified will enable the policy makers and managers of the sector to devise strategies and adopt effective measures to enhance economic efficiency and utilization of resources. Two types of surveys were conducted: facility survey and exit client survey. A primary survey was conducted in all types of primary public health care facilities of some selected upazilas of Bangladesh. The facilities are: Upazila Health Complex (UHC), Union Health and Family Welfare Centre (UH&FWC), and Community Clinic (CC). A total of 16 upazilas were selected for the survey. The survey includes one UH&FWC and one CC from each of the selected UHC. Therefore a total of 48 facilities were surveyed. The exit client survey covered a total of 448 exit clients’ interview in the study areas. The study used Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA) as a method of analysis. DEA is a technique which measures relative efficiency of a set of production units which are termed as decision-making units (DMUs) using similar multiple inputs to produce similar multiple outputs. The relative efficiency of a unit is the ratio of the sum of weighted outputs to the sum of weighted inputs. The main objectives are to identify production units which are relatively inefficient and identify targets/peers for them based on practices of the efficient units. Pareto optimality is the foundation of DEA (Charnes et. al., 1985). “A DMU is considered relatively efficient if there is no other DMU or a combination of DMUs which can produce at least the same amount of all outputs with less of one input and not more of any other input” (Emrouznejad and Emrouznejad, 2001). The ratio gives a score that is expressed as 0-1 or 0-100 percent. A DMU with a score less than 1 or 100 percent is inefficient compared to its counterparts. Kontodimopoulos and Niakas (2005) revealed that the advantages of DEA are that it can include multiple inputs and outputs simultaneously. It identifies efficient units and performance targets for the inefficient units. DEA’s main limitations are the impacts of: omitting important variables, outliers, and missing observations. DEA is now widely used tool for health care efficiency measurement (Cowing and Holtman, 1983; Ning et al., 2007; Shetty and Pakkala, 2010; Osei et al., 2005). Malmquist Productivity Index has also been computed using DEA to measure the change in efficiency overtime and identify the cause of this change - technical efficiency change or technological change over time. Another method of measuring efficiency, Stochastic Frontier Analysis, has been used to complement the measures obtained by DEA. SFA is a parametric method. The parametric approach has the advantage of allowing noise in the measurement of inefficiency. The model assumes a traditional random error and a nonnegative error term representing the technical inefficiency. Sarafidis (2002) revealed that SFA is an extension of simple regression analysis in that it estimates the “frontier” of a set of functions with different levels of efficiency. SFA requires separate assumptions for the distributions of the “inefficiency” and “error” terms, leading to nearly accurate measures of efficiency. The findings on background characteristics of the areas indicate that in term of the major characteristics the upazilas under study by and large represent the upazilas of the country. In most upazilas the density of population is quite high, level of education is low everywhere, most of the households are engaged in agriculture and small business, and the majority of the households belong to the low income group. At the UHC level the average number of manpower employed was 64. More than 31 percent of the equipment in the facilities were out of order. At the UH&FWCs the average number of manpower employed was 4. Almost 20 percent of the equipment were found out of order there. At the CCs the average of number manpower employed was 3. Only 6 percent of the equipment were out of order. The findings clearly suggest that the level of utilization of health care is low in most facilities. The level of utilization varies widely among the facilities, although the capacities of the facilities of each type should be same, given the structure and budgetary provision of the government for the facilities. The question can arise as to why there is so much variation in output. Since on the supply side the factors are more or less same, the demand side factors may be mainly responsible for this. In any case, the level of utilization of health care in most of the UHCs as well as in the lower level facilities clearly indicates that economic efficiency is quite low there – volume of output is lower compared to the expenditure. However, utilization alone cannot show the level of economic efficiency. The Data Envelopment Analysis and Stochastic Frontier Analysis measure level of economic efficiency of resource using data on both inputs and outputs. The results to identify the level of efficiency of resource use at the Upazila Health Complexes (UHCs) under the survey showed that 6 facilities were efficient under input oriented CRS assumption. The average efficiency score of all the facilities was 76 percent. Seven UHCs scored lower than the average efficiency score. The lowest score was 20 percent. By Bangladesh standard, some facilities were found highly inefficient and some were inadequately efficient. Only a few UHCs were efficient to the acceptable limit. The implication is that there is waste of resource use, utilization rate is low and this creates financing gap. The average amount of waste per UHC was Tk. 3.72 crore. The total amount of waste of all UHCs in Bangladesh was estimated to be TK. 1539.87 crore and it is 12.13 percent of the total health budget of the country for the year 2015-2016. Most UHCs had slack in nurse manpower, followed by drug and equipment. The results of the econometric analysis showed that the relationship between the dependent variable and each independent variable was as expected. However, the estimates were statistically insignificant presumably due to small sample size. The results of sensitivity analysis showed that the performance rank of facility changes in different sensitivity scenario. Comparison between utilization rates and efficiency scores of the facilities clearly indicated that efficiency categories do not match with the utilization categories for most facilities. A number of facilities exist which show high efficiency at the low utilization rate. On the other hand, some facilities have achieved high or medium level utilization rate, but efficiency level of them is low. Only one UHC has achieved both objectives: Fulbari UHC. The time series data for three years were used to identify efficiency gain/loss over the years. There was a productivity loss from 2011 to 2012 but a productivity gain from2012 to 2013. The gain was due to technical efficiency increase, as the technological change is less than one. Only one facility was found both technical and production efficient. Two UHCs had high technical efficiency scores but low production efficiency and cost efficiency scores in terms of both the technical efficiency measures. The finding indicates that although the existing inputs were used at maximum in those facilities other conditions such as appropriate mixture of inputs, and procurement of inputs at market prices have not been fulfilled by those upazilas. As a result of which they have higher production efficiency scores. The results of SFA showed that none of the facilities was fully efficient. Increase in absenteeism reduced efficiency of the facilities. Due to absenteeism efficiency on average was reduced by 0.03 percent. Two UHCs ranked high in both DEA and SFA. Three ranked medium and only one ranked low in both methods. The results of the level of efficiency of resource use at the Union Health and Family Welfare Centres (UH&FWCs) and Community Clinics (CCs) showed that only one UH&FWC and one CC were found efficient and the average efficiency score was 53 and 46 percent, respectively. The time series data for three years revealed that there was a productivity loss from 2011 to 2012 but a productivity gain from 2012 to 2013 at the UH&FWCs and productivity gain in both periods for CCs. None of the facilities, of any type whatsoever, was found fully efficient according to the SFA scores. Comparison between the results of DEA and SFA for both UH&FWCs and CCs showed that for most facilities SFA score moderately improves except for one UH&FWC. The major finding of the research are – Average level of efficiency of the primary public health care facilities is lower than expected. The efficiency level of the technically low efficient facilities is very low- as low as 20 percent. There is, therefore, a huge gap between the efficiency level of the efficient facilities and the inefficient facilities. It was found that high efficiency is achieved by some facilities with low utilization rate. The objective of the health sector is to achieve high efficiency at high level of utilization so that higher equity is not attained. Only one facility achieved both the objectives: high efficiency and high utilization. Results of economic efficiency show that the mean production efficiency is 0.70. Only one facility was found both technically and production efficient. There is slack in all inputs in the inefficient facilities. The survey facilities are using more inputs than is required to produce current level of output. Either these inputs should be redistributed or they should be utilized in a manner so that these facilities attain efficiency. Absenteeism of providers was found at all levels except in CCs. Moreover, a significant number of the posts remain vacant in the UHCs and UH&FWCs. The vacancy is low at the CCs. There are equipments out of order in all facilities. A major portion of the equipments remain out of order at the facilities, adversely affecting the efficiency level of the facilities. However, the out of order equipment is low at the CCs. The sensitivity analysis reveals that manpower, especially doctor and nurse, is an important factor determining efficiency. One encouraging finding of the time series analysis is that there is a productivity gain from 2012 to 2013 at UHCs. Although there was a productivity loss from 2011 to 2012. The results show that only two facilities, Nabiganj and Sundarganj UHCs, experienced both technical efficiency improvement and technological improvement. At the UH&FWCs and CCs, all facilities experienced productivity gains during the periods. The regression results show that a larger size of the population of the area(via its effect on the number of clients), increase in female education, high distance of the upazila from Dhaka, and low level of poverty increase have their impact on the efficiency level of the facilities.

The major conclusion stemming from the thesis is that a high proportion of UHCs under survey have been found inefficient in that the resources employed therein are not fully and properly utilized. Underutilization and improper utilization of resources have caused huge waste of resources in those facilities. It can be assumed that the condition of the sample UHCs under survey greatly represents that of all UHCs of the country. Then it appears that the vast majority of the UHCs are inefficient and the entire health sector is incurring huge financial loss in the form of waste of resources – the total national loss is estimated to TK. 1539.87 crore constituting 12.13 percent of the total health budget of the country for the year 2015-2016. The findings strongly indicate that if efficiency of UHCs is enhanced, it will save huge amount of resources, which can be spent for increasing coverage and quality of services. Thus, the financing gap of the sector will considerably decline and performances of the sector will improve. The possible reasons of low efficiency could be absenteeism of providers, inappropriate skill mix, leakages and misuse of medicines and supplies. On the basis of the findings of the research it can be recommended that measures should be taken to reduce absenteeism, inputs present at the facility should be enabled to work properly, excess manpower could be transferred to other under-staffed facilities, merger of two facilities in close geographic proximity can be an option to improve efficiency, transport and communication in the remote areas should be improved, efforts should also be made to increase demand and utilization of services, and efficiency measurement should be institutionalized to improve system and management. The study has a few limitations. The results obtained using DEA assumed that all inputs were fully employed. The inclusion of absenteeism in the DEA would change the level of efficiency of some facilities. The tool does not yield absolute scores rather it shows relative scores. The DEA requires large number of DMUs to obtain appropriate result. In this study, the number of facilities at each tier was small. The sample size could not be increased due to the constraints of time and fund available for the study.