|Year : 2019 | Volume
| Issue : 1 | Page : 24-31
Evaluation of technical efficiency of county referral hospitals in kenya and its determinants
Gilbert Koome Rithaa1, George Kosimbei2, Andrew Yitambe3, Peter Kithuka3
1 Department of Health Records and Information Management, College of Health Sciences, Mount Kenya University, Thika, Kenya
2 Department of Applied Economics, Kenyatta University, Nairobi, Kenya
3 Departments of Health Management and Informatics, Kenyatta University, Nairobi, Kenya
|Date of Submission||10-Nov-2018|
|Date of Acceptance||18-Apr-2019|
|Date of Web Publication||10-Jul-2019|
Mr. Gilbert Koome Rithaa
College of Health Sciences, Mount Kenya University, Thika
Source of Support: None, Conflict of Interest: None
Background: Kenya's gross national income per capita of $ 2250 is significantly lower than the global average of $ 6977. In addition, Kenya Government's health expenditure as a percent of the total government budget is approximately 7% which falls below the target of 15% recommended by the World Health Organization. It is, therefore, important that the country's health-care resources, specifically those allocated to the health sector, are optimally used. Methods: An one-stage data envelopment analysis (DEA) method was used to estimate the technical efficiency of county referral hospitals. A total of 34 county referral hospitals were randomly sampled and studied. Data analysis was performed in two stages as follows: first, input and output data were entered into MS Excel sheet after which DEA version 2.1 was used to determine the technical efficiency scores for the hospitals. In the second stage, interval regression analysis using censored interval regression model was used to identify determinants of technical efficiency for the sampled hospitals. Results: Results indicated that the mean constant return to scale technically efficient score was 82.4%, the mean variable return to scale (VRS) technically efficient score was 94.1%, and the mean scale efficiency technical score was 87.4%. The mean level of VRS technical inefficiency was 17.2%. The total inputs slacks in the inefficient hospitals were 104 beds and 840 staff which represented an input slack of 4% for the beds and 28% for the staff. Conclusions: Inefficient hospitals could have attained efficient frontiers using fewer resources, specifically 4% beds and 28% staff. The technical inefficiencies in county referral hospitals are occasioned by the use of inappropriate production functions characterized by the existence of excess production inputs and suboptimal outputs.
Keywords: County referral hospitals, data envelopment analysis, scale efficiency, technical efficiency
|How to cite this article:|
Rithaa GK, Kosimbei G, Yitambe A, Kithuka P. Evaluation of technical efficiency of county referral hospitals in kenya and its determinants. Int J Adv Med Health Res 2019;6:24-31
|How to cite this URL:|
Rithaa GK, Kosimbei G, Yitambe A, Kithuka P. Evaluation of technical efficiency of county referral hospitals in kenya and its determinants. Int J Adv Med Health Res [serial online] 2019 [cited 2020 Oct 30];6:24-31. Available from: https://www.ijamhrjournal.org/text.asp?2019/6/1/24/262496
| Introduction|| |
Kenya is located in the East Coast of Africa. In 2012, Kenya had a population of approximately 44 million with an average annual population growth of 2.7%. About 23% of the Kenyan population resides in urban areas compared to 51% for the African region., Kenya is ranked 145th among 186 countries in the poverty index with almost half (21 million) of the population living below poverty line. In 2012, the country had a gross domestic income per capita of $840. The average total expenditure on health as a percentage of its gross domestic product (GDP) is approximately $4.7.,
Tightening budget and increasing pressures on the efficiency of public spending represent major challenges for the Kenyan Government., The achievement of the sustainable development goals and related initiatives as reaffirmed by the UN state members requires not only availability of adequate resources but also their efficient use to improve access and quality of care., However, the resources required to meet the costs of achieving the development goals are far beyond the reach of many. In Kenya and other sub-Saharan African countries, hospitals absorb the greatest proportion of the total health expenditure which requires efficient use for maximum benefit to the population.
Achieving a high level of technical efficiency in the health sector is an important task in the face of significant budgetary allocation in the public health-care system. For instance, the Kenyan Government health expenditure as a percent of the total government budget was 6.5% for the financial year 2010/2011, which falls below the target of 15% recommended by the Abuja Declaration. Despite Kenyan health sector budget increasing overtime, there has not been a corresponding significant change in her key health indicators.,,, For instance, comparison from the UNICEF has pointed out that weak health-care management systems and poor quality of services are derailing health sector reform gains as they are associated with increasing inefficiency and resource wastages.
Kenya through a number of economic development programs such as Vision 2030, National Health Sector Strategic Plan III 2013–2018, and the Health Sector Policy Framework 2012–2017 places emphasis on improved efficiency in health-care delivery at all levels. This is important for the health sector since hospitals in Kenya are reported to account for over 75% of the health sector budget and employ over 80% of these key health professionals. Despite this increasing advocacy for efficiency, there has been limited effort in assessing the status of technical efficiency in health sector in Kenya. At present, studies on technical efficiency in county referral hospitals in Kenya are limited. The main objective of this study was to evaluate technical efficiency of county referral hospitals in Kenya and its determinants.
| Methods|| |
Efficiency refers to the use of available economic resources in a manner that results in maximum possible output. For a hospital to be economically efficient, it has to be also technically efficient. This requires the firm to use the right mix of inputs in light of the relative price of each input (input allocative efficiency) and produce the right mix of outputs given the set of prices (output allocative efficiency) which results into economic efficiency. Therefore, economic efficiency refers to the process of obtaining maximum benefits from a given cost or minimizing the cost of obtaining a given benefit. Technical efficiency of a hospital can be categorized into pure technical efficiency and scale efficiency. Pure technical efficiency refers to technical efficiency score which cannot be attributed to changes in economics of scale in the production process. It assumes no deviations from the optimal scale. It is depicted by regions in which there exists a constant return to scale (CRS) relationship between the outputs and inputs., On the other hand, scale efficiency refers to extent to which health-care outputs change due to changes in health-care inputs. In other words, according to Salvatore, scale efficiency assumes economics of scale in the production process which is mainly depicted in regions with variable returns to scale regions. Scale efficiency is equal to CRS technical score divided by variable returns to scale technical score.
Technical efficiency is evaluated relative to other decision-making units (DMUs). A DMU is said to be fully efficient (100%) when the performance of other DMUs indicates that altering the mix of its inputs and outputs could worsen the production mix. Therefore, technical efficiency can be measured in terms of the optimal combination of inputs to achieve a given level of outputs (an input-oriented model). The input-oriented model of technical efficiency measures how many fewer resources a hospital could employ and still produce the same level of output., In this context, efficiency is interpreted as a hospital's resource intensity relative to best practice. This is a better approach for estimating technical efficiency in public hospitals which have less flexibility to change their output but can change their use of inputs since they operate using a capped/fixed budget.
In this analysis, hospital efficiency is assessed using an input-oriented model. A two-stage data envelopment analysis (DEA) was used in estimating technical efficiency scores of the County Referral Hospitals in Kenya and its determinants. DEA which is defined as a nonparametric mathematical programming approach to frontier estimation which uses linear programming to sketch a boundary function (efficient frontier) to observed data for relatively homogeneous firms. DEA and econometric methods have been greatly used in analyzing technical efficiency of health-care sector in both developed and developing countries. In this study, DEA is used to estimate the hospital-level efficiency scores and the magnitude of technical inefficiencies in county referral hospitals.
Data and data management
In Kenya Health System, the hospital medical record officers collate daily summaries of the numbers of visits, discharges, and operations. Every month, hospitals upload a summary of this data on selected output and input indicators to the district health information system (DHIS) housed at the Ministry of Health headquarters. However, the staff data and capital data such as bed capacity are updated annually in the DHIS. Therefore, the input, output, and other capital data used in the study were obtained from the DHIS report 2012/2013 financial year. To complement this data, personal visits to the county health records and information offices were done to facilitate records review and validation of collected data. In this study, two main county referral hospitals were selected from each of the 47 counties in Kenya to form a sampling frame of 84 county referral hospitals. An excel sheet was used to enlist the hospitals using unique numerical codes for each of the sampled hospitals. The choice of the two main county referral hospitals selected was based on the scale of operations to enhance the comparability of results. In the second stage, random sampling was used to randomly sample 34 county referral hospitals for analysis using the excel sheet random selection function. This sampling technique was justified by its ability to ensure representation and control for selection bias. In this study, data analysis was performed in two stages: first, input and output data were entered into Excel sheet after which DEA version 2.1 was used to determine the technical efficiency scores for the hospitals. In the second stage, Stata, an analytic software, was used to perform interval regression analysis using censored interval regression model. This model was used to identify determinants of technical efficiency for the sampled hospitals.
Before data collection, a research permit to authorize the study and data collection was granted by the National Commission for Science, Technology, and Innovation which is a research regulatory body mandated to regulate research activities in Kenya. Because the study used existing data with no human subject involvement, ethical review exemption was granted.
| Results|| |
[Table 1] presents the descriptive statistics for inputs and outputs of randomly sampled Kenyan county referral hospitals in this study. In the financial year of April 1, 2012–March 31, 2013, the 34 hospitals [Appendix 1] [Additional file 1] received 3,499,531 outpatient visits, discharged 241,436 inpatients, and performed 32,401 operations. These outputs were produced using a total of seven inputs which were 234 medical consultants, 284 general practitioners (medical officers), 739 clinical officers (specialized and registered clinical officers), 4542 nurses, 435 therapy specialists, 434 laboratory technologists, and 6110 hospital beds (including cots). The outputs and inputs varied widely across different hospitals. The means and standard deviations of hospitals outputs varied widely across the hospitals. The mean of outputs was as follows: 102,927 outpatient visits, 7101 discharges, and 959 operations. On the other hand, the mean inputs were as follows: 180 beds, 7 medical consultants, 8 general practitioners, 22 clinical officers, 134 nurses, 10 therapy specialists, and 13 laboratory technologists.
[Table 2] shows the scores for constant returns to scale technical efficiency, variable returns to scale technical efficiency, scale efficiency, types of scale efficiency and the efficiency reference set for each hospital sampled in this study.
|Table 2: Technical efficiency, and Lambda weights for the sampled hospitals|
Click here to view
(a) Constant Return to Scale (CRS)
We found that 14 (41%) hospitals were constant return to scale technically efficient, scoring 100% while the remaining 20 (59%) hospitals were constant return to scale inefficient. The mean CRS technical efficiency was 82.4% with a standard deviation of 21.1%. The average CRS technical efficiency score varied from a minimum of 26.2% to a maximum of 100%.
(b) Variable Return to Scale
In terms of variable returns to scale (VRS), of the 34 hospitals studied, 22 (64.7%) hospitals, were VRS technically efficient while the remaining 12 hospitals were VRS technically inefficient. The mean VRS technical efficiency score was 94.1% with a standard deviation of 11.1%. Kapenguria District hospital had the lowest variable returns to scale technical efficiency score of 63.2%.
Results indicated that 14 (42%) hospitals had an optimal production function for their input -output mix i.e. they had a scale efficiency score of 100%. The remaining 20 (59%) hospitals had scale efficiency scores of less than 100% and were thus deemed scale inefficient. The mean scale efficiency score was 87.4%, with a standard deviation of 19.8%. The average scale efficiency score varied from a minimum of 26.2 % to a maximum of 100%. Increasing the quantity of all hospitals inputs by a given proportion would result in:
- Increasing returns to scale in 41.2% hospitals.
- Constant returns to scale in 41.2% hospitals. This means that their health service outputs would increase in the same proportion.
- Decreasing returns to scale in 17.6% hospitals.
Magnitude of inefficiencies using VRS
A total of 22 (64.7%) hospitals were VRS technically efficient while the remaining 12 (35.3%) hospitals were VRS inefficient. Technical inefficiency varied from a minimum of 4.1% (Kathiani) to a maximum of 36.8% (Kapenguria). Among the technically inefficient hospitals, the mean level of VRS inefficiencies was 17.2% with a standard deviation of 10.4% [Table 3].
Production changes (input and output slacks)
35.3% technically inefficient hospitals required to change their input-output mix in different proportions as shown in [Table 4] to attain technically efficiency frontiers. Since the VRS model was an input oriented model, the inefficient hospitals were required to reduce their inputs and augment any resulting gaps with their outputs in different proportions to achieve optimal efficiency scores. In all the inefficient hospitals, the following total input reductions were required: 104 beds; 24 medical specialists; 33 general practitioners; 126 clinical officers, 513 nurses and 91 laboratory technologists.
|Table 4: Production slacks required to make inefficient hospitals efficient|
Click here to view
While beds slack varied from a minimum of 0 beds to a maximum of 41 beds (Kathiani), the total staff slacks ranged from a minimum of 4 staff (Kangundo) to maximum of 271 staff (Kapenguria). This translated to a total of 104 beds (including cots) and 840 staff in all the inefficient hospitals. This was an equivalent of 4% reduction in total beds and 28% reduction in total staff in all the inefficient hospitals. Reductions in input levels are not sufficient to achieve efficient levels in the inefficient hospitals. This required inefficient hospitals to augment their input reductions (slacks) with output additions (slacks). All the inefficient hospitals required a 23% increase in total outpatient visits which translated to an additional 266,376 outpatient visits.
Increase in outpatient visits varied between a minimum of 0 to a maximum of 89, 547 visits (Kisumu) with a mean of 22, 198 outpatient visits. Additions in outpatient discharges varied from a minimum of 0 to a maximum of 2, 649 (Taveta) with a mean of 571 discharges in all the hospitals. This translated to 9% increase in total discharges. A mean of 385 operations was required for all the hospitals to reach efficient production frontier. However, the additional operations varied from a minimum of 0 to a maximum of 877 operations in Kapkatet. All the hospitals required to augment their operations capacity with a sum of 4,616 operations which represented a 55% increase in total operations.
| Discussion|| |
The mean CRS technical efficiency for the hospitals was 82.4%. About a third of the hospitals were technically inefficient similar to findings of a study conducted in Kenya by Kirigia et al. which found that about 44% of the public health centers were inefficient with an average efficiency score of 84%. This means that the hospitals can attain efficiency levels in the efficiency frontiers by reducing their input mix by an average of 17.6%. However, the high variation in output and input production mix among the different hospitals can be explained by variations in the sizes and operations of the hospitals sampled for the study.
The variable return to scale (VRS) model indicated that the hospitals had a VRS efficiency mean of about 94.1% which meant that they could achieve efficiency levels by reducing their input levels by 5.9%. However, this differs from results by Kibe who found that the average technical efficiency score for level four hospitals in Kenya in the period 2008–2011 was 97.72%. However, the reduction in efficiency levels may point to increase in inefficiency in public hospitals amid the many health-care reforms targeting optimal resource utilization. This was further supported by the scale efficiency findings that about 13.6% of the total inefficiency scores can be attributed to scale inefficiency. Optimal operational mix in their production functions is critical for their optimal resource utilization. This finding supports the available evidence that hospitals in developing countries operate at unacceptably high levels of technical inefficiency.
It has been found in this study that two-fifths of the county referral hospitals have increasing return to scale production functions. This means that an increase in their health service inputs would increase their outputs by a greater proportion. These hospitals should increase their production sizes to achieve optimal scale efficiency and reduce unit costs of production. This is consistent with observations from Eretria and Zambia that the hospitals can perform better with the level of resources that are allocated to the sector without scaling them up., Similarly, about 40% of the county referral hospitals exhibited CRS in their production function. This meant that increasing their health service inputs would result in their health service outputs increasing in similar proportion; they are scale efficient. However, the close to a fifth of the hospitals with decreasing return to scale (DRS) meant that increasing their inputs in production mix would increase by a smaller proportion. These hospitals are not operating at their optimal scale sizes. This indicates the existence of excess inputs in these hospitals when other hospitals are grappling with resource shortages.
Technical inefficiency can be attributed to inappropriate production function. Hollingsworth and Parkin indicated that one cause of inefficiency is excess use of resources such as labor and supplies and inefficient organizational processes. This agrees with the study finding that excess inputs such as overstaffing are causing inefficiencies. This has been shown by the existence of input and output slacks in the production functions. Previous research findings from developing countries such as South Africa, Eritrea, and Namibia have recommended reallocation and rationalization of the excess inputs to achieve maximum production frontiers in other facilities.,,
Inefficient hospitals with slacks in their production can make appropriate production changes which may require reducing the excess inputs to achieve optimal resource utilization. However, the input reduction may not be sufficient to make the hospitals inefficient hence the need for augmenting gaps with outputs additions. The World Health Organization (WHO) guidelines recommend proper staffing and resource resources for equality, equity, and improved health outcomes. In addition, public hospitals can benchmark their facilities with best practices that entrench efficiency principles. However, technical inefficiency has been associated with poor policy framework, governance, and leadership setbacks as supported by past studies done in Kenya and Botswana., This calls for a facilitative and transformative leadership committed to principles of efficiency at all levels of service delivery. This is consistent with the WHO (2000) which advocated for efficiency concerns in all functions of a health system.
| Conclusions|| |
The inefficient hospitals could achieve efficiency levels by reducing their input levels by 17.6% under CRS and 5.9% under variable returns to scale. A substantial number of county referral hospitals operated inefficiently. This is mainly because of inappropriate production function which has led to inefficient production frontiers in the allocation and utilization of the limited resources in county referral hospitals. The production slacks mean that the inefficient hospitals could produce more outputs with the same or even less inputs which an indicator of resource wastages.
For optimal resource production function to be attained, the excess staff needs to be reassigned to health facilities faced with staff shortage. This will require clear policy guidelines for effective, rational, equitable, and evidence-based staff deployment. This can be achieved by developing demand-based model as the basis for reassigning excess staff. Further, to augment for output slacks and reduce unit cost of production, a pool of excess staffs needs to be created for providing essential, and demand-based health services to the population through outreach programs within and outside the hospital's catchment area. Further, the excess beds should be transferred to other public health facilities with bed shortages to ensure that the extra supply of beds does not result in excess/unnecessary admissions and longer stays, an effect known as Roemer's Law, which has been shown to cause inefficiency in hospitals.
Financial support and sponsorship
We acknowledge National Commission for Science, Technology and Innovation, Kenya, [NACOSTI] for partial financing of this research.
Conflict of interest
There is no conflict of interest.
| References|| |
World Health Organization. World Health Statistics. Geneva: World Health Organization; 2011.
Health Sector Working Group Report. Medium Term Expenditure Framework (MTEF) 2012/13-2014/15. Health Sector Working Group Report; 2012.
Ministry of Health. Health Sector Budgetary Allocation and Spending: Health Sector Analysis Report. Nairobi: Government Printers; 2011.
Coelli T. A guide to DEAP Version 2.1: A Data Envelopment Analysis Programme. CEPA Working Paper 96/8, Department of Econometrics. University of New England, UK; 1996.
Constitution of Kenya. Constitution of the Republic of Kenya. Government of the Republic of Kenya; 2010.
World Health Organization. Health Statistics: Global Health Observatory. World Health Organization, Kenya Office; 2012. Available from: http://www.who.int/countries/ken/en/
. [Last accessed on 2019 Aug 20].
World Health Organization. The World Health Report 2006: Working Together for Health. Geneva: World Health Organization; 2006.
Nancy C, Florian H. Estimated Government Spending 2009/2010: Kenyan Health Sector Budget Analysis. Mars Group; 2010.
Kirigia JM, Lambo E, Sambo L. Are public hospitals in Kwazulu-Natal province of South Africa technically efficient? Afr J Health Sci 2000;7:25-32.
Schmacker ER, McKay NL. Factors affecting productive efficiency in primary care clinics. Health Serv Manage Res 2008;21:60-70.
Kenya Demographic and Health Survey. Kenya Demographic and Health Survey 2003-04. Nairobi: Government Press; 2003.
Kenya Demographic and Health Survey. Kenya Demographic and Health Survey 2008-09. Nairobi: Government Press; 2009.
Sealy S, Rosbach K. Kenya Health Sector Budget Analysis. Estimated Government Spending. Gtz; 2010.
United Nations Children's Fund. Social Budgeting Investments in Kenya's Future. Government of Kenya/UNICEF Programme of Cooperation 2004-2008. Working Paper No. 2; 2007.
Vision 2030: Second Medium Term Plan 2013-2017. Government of the Republic of Kenya; 2013.
Ministry of Health. National Health Sector Strategic Plan (NHSSP) III 2013-2018. Government of Republic of Kenya; 2014.
Kirigia J, Mensah O, Mwikisa C, Asbu Z, Emrouznejad A, Makoudode P, et al
. Technical efficiency of zone hospitals in Benin. Afr Health Monit 2010;12:30-9.
Banker D, Charnes A, Cooper W. Some models for estimating technical and scale inefficiencies in data envelopment analysis. Manage Sci 1984;30:1078-92.
Carlson SN. Microeconomics. London: Blackwell Publishers; 1996.
Coelli T, Rao P, O'Donnell J, Battese E. An Introduction to Efficiency and Productivity Analysis. London: Springer; 2005.
Fried H, Lovell C, Schmidt S. The Measurement of Productive Efficiency: Techniques and Applications. New York: Oxford University Press; 1993.
Salvatore D. Managerial Economics: Principles and Worldwide Applications. New York: Oxford University Press; 2008.
Charnes A, Cooper W, Rhodes E. Measuring the efficiency of decision-making units. Eur J Oper Res 1978;2:429-44.
Cooper W, Seiford M, Zhu J. Data envelopment analysis: History, models and interpretations. In: Handbook on Data Envelopment Analysis. Vol. 1. New York: Springer; 2011. p. 1-40.
Farrell J. The measurement of productive efficiency. J R Stat Soc 1997;120:253-90.
StataCorp. STATA 10 Data Analysis and Statistical Software. Texas: StataCorp; 2010.
McDonald F, Moffitt R. The uses of tobit analysis. Rev Econ Stat 1980;62:318-21.
Kirigia JM, Asbu EZ. Technical and scale efficiency of public community hospitals in Eritrea: An exploratory study. Health Econ Rev 2013;3:6.
Kibe N. Analysis of Efficiency in Public Hospitals in Kenya 2008-2011, unpublished Research Paper. University of Nairobi; 2010.
O'Neilla L, Raunerb M, Heidenbergerb K, Kraus M. A cross-national comparison and taxonomy of DEA-based hospital efficiency studies. Soc Econ Plann Sci 2008;42:158-89.
Masiye F. Investigating health system performance: An application of data envelopment analysis to Zambian hospitals. BMC Health Serv Res 2007;7:58.
Ruggiero J. Performance evaluation in education: Modelling educational production. In: Handbook on Data Envelopment Analysis. Boston: Kluwer Academic Publishers; 2004.
Hollingsworth B, Parkin D. The efficiency of Scottish acute hospitals: An application of data envelopment analysis. IMA J Math Appl Med Biol 1995;12:161-73.
Osei D, d'Almeida S, George MO, Kirigia JM, Mensah AO, Kainyu LH, et al
. Technical efficiency of public district hospitals and health centres in Ghana: A pilot study. Cost Eff Resour Alloc 2005;3:9.
Zere E, Mbeeli T, Shangula K, Mandlhate C, Mutirua K, Tjivambi B, et al.
Technical efficiency of district hospitals: Evidence from Namibia using data envelopment analysis. Cost Eff Resour Alloc 2006;4:5.
World Health Organization. The World Health Organization: Health Systems-Improving Performance. Geneva: World Health Organization; 2000.
Tlotlego N, Nonvignon J, Sambo LG, Asbu EZ, Kirigia JM. Assessment of productivity of hospitals in Botswana: A DEA application. Int Arch Med 2010;3:27.
Roemer MI. Bed supply and hospital utilization: A natural experiment. Hospitals 1961;35:36-42.
[Table 1], [Table 2], [Table 3], [Table 4]