Approaching the management of hospital units with an operation research technique: The case of 32 Greek obstetric and gynaecology public units

Transcription

1 Available online at Health Policy 85 (2008) Approaching the management of hospital units with an operation research technique: The case of 32 Greek obstetric and gynaecology public units Maria Katharaki National and Kapodistrian University of Athens, Egnatias 25, Athens, Greece Abstract Controlling healthcare costs is a multifaceted problem for governments all over the world, as they have the difficult task of ensuring that patients receive high quality care, and that this is delivered as efficiently as possible. Through the use of quantitative analysis, an attempt is made to determine the areas of activity of 32 Greek Public Obstetrical and Gynaecological Units which present problems with regard to their performance. Based on the results that emerge from the application of Data Envelopment Analysis in the 32 hospital units of the sample, information is provided to their managers, which refer to: (i) the degree of utilization of their production factors, (ii) the particular weight of each factor of production in the formation of the relative efficiency score, (iii) the utilization level of each factor of production, and (iv) those hospital units that utilize their factors of production in an optimal way and constitute models for the exercising of effective management. The derived information assists in the formulation of an appropriate policy mix per hospital unit which should be applied by their management teams along with a set of administrative measures that need to be undertaken in order to promote efficiency Elsevier Ireland Ltd. All rights reserved. Keywords: Hospital performance; Hospital management; Quantitative analysis; Data Envelopment Analysis; Efficiency 1. Introduction The Greek health care system is characterized by the coexistence of the National Health Service (NHS), a compulsory social insurance and a voluntary private health insurance system [1]. The NHS provides universal coverage to the population operating on the M. Katharaki owns Ph.D in OR (National & Kapodistrian University of Athens, Greece). Tel.: ; fax: address: M.Katharaki@gmail.com. principles of equity, social cohesion and equal access to health services for all. Under this context, citizens are not directly dependent on a specific healthcare institution. On the other hand, they are free to choose amongst a variety of healthcare units depending on the type of treatment they wish to follow. It should be pointed out that the Greek Ministry of Health decides on the overall national health strategy and the relative health policy issues within Greek healthcare organizations. Its main responsibilities, amongst all, are the definition of priorities, the approval and extension of funding for proposed /$ see front matter 2007 Elsevier Ireland Ltd. All rights reserved. doi: /j.healthpol

2 20 M. Katharaki / Health Policy 85 (2008) activities and the resource allocation at a national level. With the latest reforms, the main objectives were the decentralization of the system with the establishment of 17 Regional Health Authorities. Decentralization efforts devolved political and operational authority to Regional Health Authorities but were only partially fulfilled [2]. Decision making and all administrative procedures continued to depend on a very centralized and bureaucratic Ministry of Health [2]. The consequences of this fragmentation, combined with the lack of a monitoring system had an impact on the extent and quality of services provided to beneficiaries of different funds, leading to over-consumption of services and serious socioeconomic health inequalities [3]. Moreover, Greek public hospital units operate within a framework characterized by limited economic resources, a restricted number of beds and a geographically unequal distribution of both personnel and patients [2,4]. There are wide discrepancies between the number of hospitals and number of hospital beds allocated in different regions [2] and a wide variation between the distribution of resources in urban and rural areas [5]. For example, in the greater Athens area in 2000 there were 6.4 hospital beds per 1000 population while the corresponding ratio in Central Greece was 1.2 beds per 1000 population [2]. These characteristics are more vivid in the provision of health care services in obstetrical and gynaecological (O&G) cases than in others [6,7] and are attributable to demographic and national factors. The lack of gynaecologists/obstetricians together with the limited experience of the serving staff on obstetric issues in rural areas, pose difficulties in the routine follow-up of pregnant women [7,8]. As a consequence, women choose to seek health services and even followup examinations at tertiary hospitals in Athens [8]. Under this context, hospital O&G management staff is expected to perform an optimal utilization of resources in terms of quantity and quality of offered services. In other words, managers are expected to achieve efficiency despite the fact that reality imposes certain, well-known limitations. Under these circumstances, the use of quantitative methods can provide the management staff useful information concerning: (1) the evaluation of the efficiency score regarding the utilization of the available production factors; (2) the contribution of every productive factor used and of the health services provided (outputs) in the formation of the efficiency score; (3) the policy mixture, that is, the combination of inputs and outputs, which must be applied to improve the degree of utilization of the production factors; (4) the prototype models of best practice hospital units which constitute models to be emulated in managing the other hospital units. This information, when complete, assures the improvement of efficiency and quality when combined with the managers ability to make analogous with reality, the needs of the receivers of health services and the producers of health services. The objective of the current study is two-fold. Primarily, to estimate the relative technical efficiency by using a sample from public hospital units that provide obstetrical and gynaecological services in Greece. Secondly, to emphasize the policy implications for health sector policy-makers. These implications can trigger the associated policy-makers in order to conduct a national efficiency study amongst all healthcare organizations in Greece. The present paper is organized as follows. Section 2 below describes the materials and methods used. This section includes a thorough discussion on the data sources, the Data Envelopment Analysis approach and the inputs and outputs of the study. In addition, the selection strategy for the appropriate sample is presented in conjunction with a discussion for the analysis plan of the study. Continuously, Section 3 provides an outline of the results obtained along with their interpretation. Section 4 is then presenting a discussion on the overall study and its outcomes whereas the last section, Section 5, provides a summary and conclusion remarks. 2. Materials and methods 2.1. Data sources Data availability and notification of hospital units managers is very important as it would facilitate decision making and optimize efficiency. The research aim of this study is to provide such a framework by using

3 M. Katharaki / Health Policy 85 (2008) a comparative analysis of 32 Greek public hospital units with obstetrical and gynaecological services. The research is based on data collected from official public sources [3,5,9 13] and on data published in Yearbook of Health 1994 by Greek Ministry of Health [9]. The well-timed of data was partially covered from data given by Greek National Statistical Service the decade and on OECD (Organization for Economic Co-operation and Development) health data The data were also combined with primarily collected data directly from hospital units of the sample Data Envelopment Analysis The method of quantitative analysis, in the context of the current research study, is Data Envelopment Analysis (DEA), which is a technique that can be used to evaluate the efficiency of a number of producers or decision making units [14 19]. DEA works by estimating a piece-wise linear envelopment surface, known as the best practice frontier [19]. Despite its limitations [18,20], the DEA model has several important advantages over parametric and econometric approaches. Two of the most important are, primarily it does not impose a particular functional form on the production frontier [16 18] and secondly, the ability to handle multiple-output, multiple-input technologies in a straightforward way, which is considered an important feature when assessing efficiency in public sector activities [16]. A feature which is especially important in the assessment of efficiency in public sector activities [18]. Many different approaches have been taken for the evaluation of hospital units performance assessment Data Envelopment Analysis approach (DEA) [17 34]. For a review of DEA health care studies see Hollingsworth et al. [21], Ozcan et al. [22] and Chilingerian and Sherman [23]. Several other areas of application include hospitals [17,19,22 27], perioperative services [19], surgical operating rooms [29], and physicians [30]. This study extends the use of DEA in health care to hospital O&G services. The analysis is extended to provide detailed information at the level of Greek units performance that provide health services on female population of the urban and rural areas of the country. This level of detail is necessary for policymakers to make decisions on which individual units should undergo changes. To estimate the efficiency of Greek public O&G units, the CCR (Charnes, Cooper and Rhodes) inputoriented model was used [14,16]. With the application of DEA, information is acquired about the inadequate provision of services, that is, information on the concept of the slackness of the productive forces concerning the way in which the health units function and the proposed changes at the level of the utilization of the factors of production that can improve the efficiency of every O&G unit [8]. Low efficiency is usually due to excess resources and the model can be used to explore the effect on efficiency of decreasing input resources. The DEA model can be used to explore some of the underlying reasons for inefficiency Defining Inputs and outputs To carry out a DEA assessment for a group of units it is necessary to construct an input set to reflect the resources used by the units and an output set of the results obtained [16,20,23]. With the application of DEA in the quantitative investigation of the relative technical efficiency of the 32 O&G hospital units, an attempt is made to evaluate the degree of utilization of the following production factors (inputs): number of O&G beds; number of O&G medical personnel; total expenditure for the provision of care. Regarding the selected inputs, hospital size and capacity were measured by the number of beds. Most studies exclude the number of physicians because there exist independent contractors who may admit patients. For the purpose of the current study, it is important to include them as an input since there exist wide discrepancies between the number of specialized physicians in different regions of the country which they largely determine the volume of the O-G services that a hospital can perform. Moreover, the shortage of O&G physicians in rural units influences the female flow to Attica units [7,8]. The importance of more evenly distributed finances throughout the healthcare regions was the primary reason for performing a DEA analysis with the input being total expenditure. The focus of the current study is on grand total expenditure and not on the individual resource component costs (doctors salaries, nurses salaries, etc.). Therefore, O&G expenditures do not include medical personnel expenses.

4 22 M. Katharaki / Health Policy 85 (2008) Table 1 The 32 obstetrical gynaecological hospital units in the region being studied Hospitals in the study area (sample) Number of obstetrical and gynaecological units in the hospitals in the sample Total obstetrical and gynaecological units of the hospitals in the five geographical districts Attica (central region) Remaining Continental Greece Aegean Islands Thessaly Peloponnese Total Total number Of O&G units in country 92 Percent (%) Country total 119 Percent (%) Source: Ministry of Health and Social Solidarity. Percent (%) The corresponding group of outputs that describe the health care services offered are: The number of O&G hospitalization days/bed-days. The number of female patients treated. The number of O&G examinations in outpatient clinics. The number of O&G lab tests. The use of number of O&G lab tests and patient days as outputs of the study was selected in order to become criteria for efficiency assessment of units as proxy factors of the degree of resources utilization. These criteria have been utilized in a plethora of related studies [4,23,35] Sample selection and analysis plan The study was based on 32 Obstetrical and Gynaecological (O&G) units located in 5 of the 10 geographical Greek NHS regions. With regard to the sample of the public O&G Units, it must be noted that they are located in the following geographical districts of Greece: Attica (central region), remaining Continental Greece, the Peloponnese, Thessaly, and the Aegean Islands. Eleven of the 32 hospitals are located in the central region (which includes Athens and Piraeus) and the remaining 21 units are in the districts outside of these major metropolitan areas (Table 1). On the basis of the criterion of number of beds in these units, 40% of the beds offered by these 32 units of the sample belong to 2 units of the central region the Alexandra General Hospital of Attica and the Elena Venizelou General Hospital of Attica. Based on the size of the above, in terms of how representative the sample is for the entire country, figures above 50% are considered satisfactory (Table 2). The separation of the sample into the O&G units of Attica (central region) and of the outlying regions is essential. This is because the geographically unequal distribution is real, both in terms of resources and of patients who demand and receive health care services outside their region of residence, that is, in the large urban centers [8]. Maternity services are of major importance for the rural areas and Aegean Islands of Greece, mainly due to their isolation, especially during winter. Traditionally, maternity and gynaecology services in those areas are offered by the Healthcare Centres, as well as by obstetricians working in private practices [7]. Healthcare Centres are manned by Internists, General Practitioners and young, non-specialized physicians [7]. The gynaecologists/obstetricians shortage together with the limited experience of the serving staff on obstetric gynaecology issues, pose difficulties in the health services delivery and in the handling of emergency cases. Private practices, on the other hand, are

5 M. Katharaki / Health Policy 85 (2008) Table 2 The representativeness of the sample of 32 O&G hospital units Percentage of the 32 O&G units in the 5 districts (%) Percentage of the 32 O&G units in the O&G field of the country (%) Inputs Beds Medical personnel Total expenditures Outputs Bed-days Patients hospitalized Patients in outpatient clinics Lab tests Source: Ministry of Health and Social Solidarity. Percentage of the O&G units of the 5 districts in the O&G field of the country (%) usually not adequately equipped to handle difficult or emergency cases. Under these circumstances, the primary reason that many female patients travel from the peripheral regions to Athens is because they are seeking for higher quality healthcare services [8]. These are inevitably available in city hospitals which have the facilities and expertise to deal with a large variety of cases. Birth rate at place of permanent residence of mother compared to birth rate at place of childbirth happen rise as a key indicator of women internal flow to urban hospital units [12]. Furthermore emergency cases are evacuated to tertiary hospitals of Attica either by boat or aeroplane, depending on the severity of the case and on weather conditions, resulting in intensity of available resources utilization and therefore in expenditures and lab tests increase [7,8]. As a consequence, it can be deduced that different types of problems are presented in the administrative practice of the hospital units of the central region than those encountered by the administrators of the units of the outlying regions. More specifically, in the hospitals of the Attica, a build-up of patients is observed, resulting in the administration having to confront functional and performance problems in their effort to satisfy the existing demand both qualitatively and quantitatively. The hospitals operating in the outlying regions are characterized by a limited utilization of the production factors [2,3]. As a result, their administrators are interested in providing services of a higher quality and in increasing demand. That is to say, their administrations confront problems in the utilization of the available resources. The composition of the sample, based on geographic location and the figures used as inputs and outputs, is presented in Table 3. From the data in the Table, the figures of the sample related to the activities of the hospital units of the central region are more than Table 3 Composition of the sample of 32 O&G hospital units between the central and outlying regions Total sample Central region (Attica) % Outlying regions % Inputs Beds 1, Medical personnel Total expenditures (D ) 49,353,333 28,295, ,057, Outputs Bed-days 226, , , Patents hospitalized 50,209 30, , Patients examined in outpatient clinics 215, , , Lab tests 1,900,462 1,098, , Source: Ministry of Health and Social Solidarity.

6 24 M. Katharaki / Health Policy 85 (2008) twice those of the outlying regions. Thus, the sample composition between the central region and that of the outlying regions illustrates the geographic inequalities at the level of the available production factors and the health care services provided. 3. Results: model interpretation The implementation of DEA, with the use of the inputs and outputs described, led to the results presented in Table 4 that concern the evaluation of the O&G unit s performance of the sample. DEA identified 18 O&G units as efficient and 14 as inefficient. Attica s hospital units A, B, C and H are tertiary units with the biggest market share amongst the volume of patient treated from all over the country. Thus an evaluation of the relative technical efficiency of a unit of less than 100% demonstrates the degree to which the unit in question lags behind relative to the best practice unit of the sub-category of reference, with which it is compared. One of the by-products of DEA is that for O&G units it deems inefficient it produces a set of efficient peers with which the apparently inefficient unit is being compared [17,18,33]. The comparison is formed by taking a weighted average of each of inputs and outputs of the efficient units. The performance of the composite unit formed by this weighting procedure gives achievable targets for the inefficient unit. For instance, the efficiency reference set for the peripheral unit L is a combination of the actual outputs and inputs of the reference subset of hospitals and results in a composite hospital that produces as much or more outputs as unit L, but uses as much or less inputs than this unit. The composite unit L is formed by multiplying the weights (0.03, 0.26, and 0.16) of individual hospitals with the actual inputs and outputs of the 100% efficient O&G units (B, C and H). The results of the multiplication for the three hospitals are then combined to arrive at a hypothetical best practice hospital. Using Table 4 as a basis and with the use of the statistical measurement-the arithmetic mean of the efficiency score of the units of the sample, Table 5 emerges. Thus, the utilization of production factors (technical efficiency) in the total of the 32 units of the sample is 89.33%. The differentiation in the score between the hospital units of the central region and the outlying regions must however be noted. It becomes clear that the efficiency score of hospital units in the central region surpasses that of the hospital units of the outlying region. This is confirmed by reality, since in the current state of functioning of the health care units; the patients demand and receive services in Attica [7,8]. The logical consequence of this is the inadequate utilization of the available production factors in the O&G units in the outlying region (83.23%) relative to the O&G units of Attica (91.90%). From the quantitative analysis, the results of Table 6 emerge with regard to the significance coefficient of the individual production factors, or the percentage of participation of every factor of production used in the configuration of the efficiency score. Table 6 demonstrates that the technical efficiency score of the O&G units of Attica depends primarily on bed utilization and economic management, while the utilization of the potential of the medical personnel contributes to a lesser degree. On the contrary, for the hospital units of the outlying regions, it is the utilization of the potential of the medical personnel (52.5%) that contributes primarily to the shaping of the technical efficiency score, while bed utilization and economic management make a relatively limited contribution. Moreover, the differences of the population means of bed and medical personnel variables between Attica s and outlying units were statistically significant for the Mann Whitney test with p = and (p < 0.05), respectively. These results point to the areas of activity of the hospital units in which it is necessary for actions to be developed that will improve efficiency. The differences in the contribution of the inputs between hospital units in the central region and in the outlying regions indicate how different the problems and the conditions under which these units operate are. With regard to the contribution of the outputs to the formation of the technical efficiency score, the results presented in Table 7 emerge. From Table 7 it is clear that bed-days constitute the basic determining factor in the efficiency score of the hospital units of the central region, while on the contrary, the contribution of the number of patients hospitalized is not the basic factor. In contrast with the practices in the hospital units of Attica, there is a difference in the contribution of the outputs to the determination of the efficiency score in the respective units

7 M. Katharaki / Health Policy 85 (2008) Table 4 The relative technical efficiency of the 32 units of the sample Hospitals with O&G units by region Technical efficiency score (%) Benchmarks Attica N1 A (GH Attica Alexandra) N2 B (Obstetrical E. Venizelou) N3 C (GH Nea Ionia Agia Olga) N4 D (GH Athina) (0.0004), 3 (0.0245), 8 (0.4530), 21 (0.0469) N5 E (GH Laiko) (0.3223), 8 (0.3610), 11 (0.0025), 21 (0.1354) N6 F (GH Attica Evvagelismos (0.0230), 3 (0.2394), 8 (0.1927) N7 G (GH Attica elpis (0.2275), 8 (0.0475) N8 H (GH Agios Savvas) N9 I (GH Metaxas) N10 J (GH Janneio Piraeus) (0.0728), 2 (0.0300), 3 (0.4357), 8 (0.1247), 11 (0.0042), 21 (0.1159) N11 K (GH Nikaia Piraeus) Remaining continental Greece N12 L (GH Agrinio) (0.0044), 3 (0.1183), 8 (0.1125), 31 (0.3541) N13 M (GH Patra) N14 N (GH University of Patra) (0.0984), 8 (0.0879) N15 O (GH Amaliada) N16 P (GH Leivadia) (0.0510), 29 (0.1584) N17 Q (GH Halkida) (0.0013), 3 (0.0659), 11 (0.0507), 13 (0.0304), 28 (0.2669) N18 R (GH Lamia) (0.0837), 30 (0.3329), 31 (0.0537) N19 S (GH Amfissa) (0.0220), 21 (0.2342), 24 (0.0100) Thessaly N20 T (GH Larisa) (0.0128), 21 (0.2697), 28 (0.3121) N21 U (GH Volos) N22 V (GH Trikala) (0.2470), 31 (0.0672) Peloponnese N23 W (GH Argos) N24 X (GH Navplio) N25 Y (GH Trikala) N26 Z (GH Korinthos) N27 AA ((GH Sparti) N28 AB (GH Kalamata) N29 AC (GH-HC Kyparissia) Aegean Islands N30 AD (GH Mytilini) N31 AE (GH Rhodos) N32 AF (GH Spyros Varvakios) (0.006), 8 (0.670), 11 (0.366), 28 (0.773) Source: DEA results.

8 26 M. Katharaki / Health Policy 85 (2008) Table 5 Efficiency score based on the comparison of the arithmetic mean of the results of DEA Total sample 89.33% O&G units Attica 91.90% O&G units outlying regions 83.23% of the outlying regions. However, these differences of the population means are not statistically significant for the Mann Whitney tests. The comparatively greater contribution of the number of patient days and lab tests to the formulation of the efficiency score of the Attica units can be explained by the fact that the Attica units, by definition, serve as a welcome point of all acute cases. On the other hand, it should be noted that patients accommodated in Attica units usually spend more time within these units. Several reasons include the timeconsuming process of such patients for returning back home or the high level of patient succession and high level of bed coverage within these units. In addition, the fact that the Attica units are often academic units as well cannot be neglected since this also contributes to the increased lab tests utilization. Regarding the greater contribution of outpatient exams in formulating the efficiency score of the outlying regions units, it can be explained from the fact that patients tend to seek their initial treatment in a unit which is close to their location and, as a result, to become admitted to a clinic depending on the bed availability. For inefficient units, DEA provided information on the sources of inefficiency as given by the slack values (increased output, decreased input) [19,34]. DEA methodology however allows the possibility of information provision at the level of each unit. In addition to the identification of inefficient O&G units and their efficiency reference set, DEA provides additional insight into the magnitude of inefficiency of the inefficient units [4]. The magnitude of inefficiency is obtained from the magnitude of slack inputs and/or deficient outputs produced by inefficient hospitals [16,22]. Slack inputs and/or deficient output production must be eliminated before a given unit is said to be relatively efficient compared with its composite reference set of units. Thus, from the analysis, the policy mix emerges, that is to say, the target group of inputs and outputs proposed for the managers of the units in order to improve and optimize the effective utilization of the production factors (Table 8). The slack values of inputs indicate the way personnel and beds are adequately used in rural units, because of patient crowd in Attica s units (Table 8). The difference between the proposed policy mix for the hospital units of the central region and those of the outlying regions should be emphasized since the significant changes that must be made in the utilization of personnel serving in the units of the outlying regions are clear. In addition, the almost non-existent reduction in beds of the units of the central region must be emphasized in contrast with what is taking place in the outlying regions. This reduction of beds relative to the health services provided proves their inadequate utilization. The alternative proposal to the limitation of the inputs is the development of initiative and administrative measures which could contribute to the utilization of the production factors available especially when the Table 6 Percentage contribution of individual production factors in the determination of the efficiency score based on the arithmetic mean Production factors Total sample (%) O&G units Attica (%) O&G units outlying regions (%) O&G medical personnel O&G beds O&G expenditures Table 7 Percentage contribution of the outputs to the determination of the efficiency score of the production factors based on the arithmetic mean Outputs Total sample (%) O&G units Attica (%) O&G units outlying regions (%) O&G patients hospitalized O&G lab tests O&G bed-days O&G examinations in outpatient clinics

9 M. Katharaki / Health Policy 85 (2008) Table 8 Proposed policy mix (percentage reduction of the production factors based on the arithmetic mean) Production factors Total sample (%) O&G units Attica (%) O&G units outlying regions (%) O&G medical personnel O&G beds O&G expenditures intention to reduce these is, in practical terms, difficult to implement. The results that have been presented refer to the total sample and to its two sub-categories. Due to limitations of space, results for each unit cannot be given in this paper. Indicatively, the estimate which emerged from the quantitative analysis for four hospital units are presented. Two of these are in Attica (Obstetric Units of the Central Region [OUCR] A and F ) and the other two are from the sub-category of the outlying regions (Obstetric Units of the Outlying Regions [OUOR] O and Q ). According to Table 9, it can be demonstrated that in the hospital units in the central region, the primary factors determining the efficiency score are: (i) economic management and (ii) the number of bed-days. Differences however are observed in the hospital units of the outlying regions where the basic determining factors are: (i) the examinations in outpatient clinics and (ii) the number of patients hospitalized and secon- Table 9 Efficiency score and the inputs and outputs weights OUCR A OUCR F OUOR O OUOR Q Relative efficiency (score) Contribution of the production factors (%) O-Beds Total O&G expenditures O&G medical personnel Contribution of outputs (%) Outputs Number of O&G patients hospitalized O&G lab tests O&G bed-days O&G examinations in outpatient clinics Table 10 Proposed policy mix expressed as a percentage of the change in inputs and outputs, of which the efficiency score falls short of optimal OUCR F OUOR Q Relative efficiency (score) Percentage reduction of inputs O&G beds Total O&G expenditures O&G medical personnel Possible percentage increase in services Offered Number of O&G patients hospitalized O&G lab tests 0 0 O&G bed-days 0 0 O&G examinations in outpatient clinics

10 28 M. Katharaki / Health Policy 85 (2008) darily economic management and the number of beds. We must note the particularity displayed by the productive factor medical personnel between the two hospitals units of the outlying regions which makes it necessary for the administration of these units to take different measures in order for this factor to be utilized. With regard to the proposed policy mixture which should be used in order to optimize the efficiency of the two hospital units (OUCR F, OUOR Q ); see Table 10. Increased efficiency of the two hospital units will be achieved through the reduction of inputs at different levels for each input, a fact which also highlights the areas as well as the extent of the need for administrative measure to be taken. Particularly in OUOR2, the reduction in beds and of medical personnel is larger, which means not that they should be limited but rather that they must be utilized more productively to cover all hospital needs. 4. Discussion The existing homogeneous statutory framework Greek National Health System under which all public healthcare units operate along with the common public source of funding, form the basis for the creation of a comparative evaluation amongst them. This evaluation is based on the level of utilization of available resources and the determination of responsibilities areas for any possible pathogenesis within all Greek healthcare units, independent of their location. In this study, DEA was applied to 32 Greek public O&G units. O&G units operations were represented by means of an input output model whereby each unit uses quantities of inputs to generate outputs in the form of services. Specifically, clinics were considered to transform labour (physicians) and capital (approximated by the number of beds and expenditures) into services, which were assumed to be approximated by the number of female patients, in-patient days, lab tests and outpatient exams. The results of the quantitative analysis differ both with regard to the type of production factors used as well as to the geographical location of the hospital units. This necessitates the taking of different administrative measures both on a general level as well as at the level of individual units. Consequently, for the administration, the areas of activity are noted where problems exist while at the same time indications are provided on the breadth of and the needs for measures to be taken. More specifically, prioritizing the problem areas, the following are noted. For the hospital units in the central region, measures concerning the more rational utilization of economic resources should take priority, in addition to measures concerning personnel which are also important, but to a lesser degree. At the same time, a more genuine evaluation of bed-days and laboratories is essential. It is imperative that measures be taken for more rational utilization of patient hospitalization in the units of the central region in combination with the bed utilization. The results which emerge for the hospital unit OUCR F are a characteristic example. In order to optimize the efficiency score, the taking of measures to increase the effective use of inputs by more than 20%, in addition to the increase of some of its outputs to as much as double the number of lab tests in the outpatient clinics are assumed. The assumption to double the number of lab tests in outpatient clinics, is based on the ascertainment that the use of telemedicine will assist in delivering healthcare services in remote locations. Thus, the number of patients remaining for treatment at their home location will be increased which will consequently lead to more frequent patient monitoring and lab tests increase. Besides, the main objective of telemedicine is the geographical spreading of all the medical incidents and the proportional exploitation of all the productive resources [36] within healthcare units, both in urban and in regional level. With reference to the problem areas and to the breadth of the needs for the hospital units of the outlying regions to take measures, the following should be noted. It is imperative that measures be taken immediately to utilize the potential of the personnel which is underemployed by more than 50%. It is necessary for beds available in the units of the outlying regions to be utilized and secondly for the more rational utilization of expenditures to be take place. At the same time, it is considered crucial that policy be implemented that will result in the increase of (almost all) services provided by approximately 20%. In support of the above, the information which is provided for the administration of the hospital unit OUOR Q justifies the taking

11 M. Katharaki / Health Policy 85 (2008) of measures which will increase the utilization of the available factors of production by more than 30%. 5. Conclusions The research aim of this paper was to primarily estimate the relative technical efficiency by using a sample from public hospital units that provide obstetrical and gynaecological services in Greece and secondly, to emphasize the policy implications for health sector policy-makers. In order to effectively address the above goals, a comparative analysis of 32 Greek Public Hospital Units with obstetrical and gynaecological services was conducted. The research was based on data collected from official public sources [3,5,9 13]. From an analysis of the evaluation of the utilization of the production factors of these 32 Greek Public Hospital Units of the Central and Outlying regions, using the DEA method, the following emerged: (i) The areas were noted which must be reorganized in order to increase the efficiency of the functioning of the hospital units. (ii) Estimates were made of the breadth of the interventions which must be made by the administrations of the hospitals. (iii) It was ascertained that both the problems and their breadth differ between the hospital units of the central region and those of the outlying regions and as a consequence, the administrative measures to be taken also differ. (iv) The model hospital units that should be emulated as well as the production factors which contribute to the creation of these models were identified. These results give an overall picture of the benefits. Inevitably to implement changes it is necessary for policy and decision makers to know exactly which units need changes and the magnitude of these changes. With respect to this DEA is a powerful tool as it provides information about the efficiency of individual units taking into consideration multiple-inputs and multiple-outputs. It can thus enable decision makers to know exactly which inputs (beds or staff), need to be increased or decreased in an individual unit to maximise efficiency and cost savings. Inevitably, the reliability and validity of the DEA results are vital if decisions are to be made based on these, and further work may be required to establish their accuracy. However, even if the predictions from DEA are not completely accurate, DEA will still almost certainly indicate which areas could be targeted to improve use of resources and to maximise efficiency. Consequently, the following benefits should be achieved to a large extent by implementing changes: Prevention of unnecessary journeys by female patients. Better geographical distribution of O&G cases. More efficient use of resources in rural areas. Increased provision of services in rural areas. Information indicating which resources should be targeted to improve efficiency maximise resources and provide more equitable care. DEA results can help administrators by providing new insights on the distribution of health resources to individual hospital units. The present study provides valuable information regarding deployment of medical staff and beds, the utilization of financial resources and the deployment of medical supplies and equipment. Resource planning and the identification of the needs of O&G units are greatly aided by the availability of up-to-date DEA results. Briefly, all of the above are believed to constitute useful information for the managers of the hospital units, which will assist them in making decision that will lead to the more effective operation of the units. Over and above these measures, which fall into the competencies and responsibilities of the managers of the hospital units, the possibility for comparison between these units facilitates monitoring by those in charge while at the same time it contributes to the creation of a spirit of rivalry and competition with each other. Finally, the differences in the evaluations between the hospitals of the central region and of those of the outlying regions underline the necessity for the geographic redistribution of the cases. Underline, also, the necessity to investigate the reasons that patients prefer the central Attica hospitals, considering that are reputed to be better. Marketing campaigns and regulation to create barriers to movement of non-local patients into the centrally located hospitals are issues for further research. Social marketing is an approach to

12 30 M. Katharaki / Health Policy 85 (2008) changing behaviour and thus improving public health. It could help to facilitate this critical review, the object of which would be to isolate those approaches that really do enable individuals and communities to gain greater control over their health and the quality of their lives [37]. Nevertheless, the redistribution is assumed to be achieved through the implementation of Information and Communication Technologies (ICT). This can be used not only to store and transfer patient information but also to improve decision making, to improve institutional efficiency, to promote better health behaviour, and to enhance more rational management of resources [38 41]. The essence of telemedicine lies in transferring expertise and not the patient [39]. This enables the needs and demands of healthcare to be met across large distances and is an important means for achieving geographical redistribution of resources and thus this would facilitate the taking of more effective measures at the management level of the units. Telemedicine allows local services to be provided to patients wherever and whenever it is possible, eliminating unnecessary journeys for patients. The feasibility of the introduction of such a national system of telemedicine and its impact on the efficiency of hospital units of the central and outlying regions is an issue which requires further research. References [1] Allin S, Bankauskaite V, Dubois H, Figueras J, Golna C, Grosse-Tebbe S, et al. In: Grosse-Tebbe S, Figueras J, editors. Snapshots of health systems. WHO, European Observatory on Health Systems and Policies; Available from: 1. [2] Tountas Y, Karnaki P, Pavi E. Reforming the reform: the Greek national health system in transition. Health Policy 2002;62: [3] Center for Health Services Research. The state of health in Greece. Athens: Ministry of Health and Welfare; 2000 [in Greek]. [4] Giokas DI. Greek Hospitals: how well their resources are used. Omega 2001;29(1): [5] Center for Health Services Research. Health services in Greece. Athens: Ministry of Health and Welfare; 2001 [in Greek]. [6] Desai J. The cost of emergency obstetric care: concepts and issues. International Journal of Gynecology and Obstetrics 2003;81: [7] Gatzonis M, Deftereos S, Vasiliou P, Dimitriou F, Creatsas G, Sotiriou D, et al. Maternity Telemedicine Services in the Aegean Islands. In: Proceedings of 2nd International conference on telemedicine p [8] Katharaki M. The efficiency impact of Telemedicine on Obstetric and Gynaecology services: effects on Hospitals units management. Dissertation. Greece: University of Athens; 2006 [in Greek]. [9] Central Council on Health. Yearbook of health. Athens: Ministry of Health and Welfare; 1994 [in Greek]. [10] National Statistical Service of Greece. Statistical yearbook of Greece Athens: National Statistical Service of Greece; 2001 [in Greek]. [11] National Statistical Service of Greece. Statistics for social care and health. Athens: National Statistical Service of Greece; [in Greek]. [12] National Statistical Service of Greece. Vital statistics of Greece. Athens: National Statistical Service of Greece; [in Greek]. [13] OECD. Health data software; 2003, Version 06/15/2003. [14] Charnes A, Cooper WW, Rhodes E. Measuring the efficiency of decision making units. European Journal of Operational Research 1978;2: [15] Charnes A, Cooper WW. Preface to topics in data envelopment analysis. Annals of Operations Research 1985;2: [16] Cooper WW, Seiford LM, Zhu J. Data envelopment analysis: history, models and interpretations. In: Cooper WW, Seiford LM, Zhu J, editors. Handbook on data envelopment analysis. Boston/London: Kluwer Academic Publisher; p [Chapter 1]. [17] Al-Shammari M. A multi-criteria data envelopment analysis model for measuring the productive efficiency of Hospitals. International Journal of Operations and Production Management 1999;19: [18] Salinas-Jimenez J, Smith P. Data envelopment analysis applied to quality in primary health care. Annals of Operations Research 1996;67: [19] O Neill L, Dexter F. Evaluating the efficiency of Hospitals perioperative services using DEA. In: Brandeau ML, Sainfort F, Pierskalla WP, editors. Operations research and health care. Norwell, MA: Kluwer Academic Publishers; [20] Anderson T. A data envelopment analysis (DEA) home page; Available from: homedea.html. [21] Hollingsworth B, Dawson PJ, Maniadakis N. Efficiency measurement of health care: a review of non-parametric methods and applications. Health Care Management Science 1999;2: [22] Ozcan YA, Merwin E, Lee K, Morrisey JP. Benchmarking using DEA: the case of Mental Health Organizations. In: Brandeau ML, Sainfort F, Pierskalla WP, editors. Operations research and health care. Norwell, MA: Kluwer Academic Publishers; [23] Chilingerian JA, Sherman HD. Health care applications. From Hospitals to Physicians, from productive efficiency to quality frontiers. In: Cooper WW, Seiford LM, Zhu J, editors. Handbook on data envelopment analysis. Boston/London: Kluwer Academic Publisher; 2004.

13 M. Katharaki / Health Policy 85 (2008) [24] Butler TW, Li L. The utility of returns to scale in DEA programming: An analysis of Michigan rural hospitals. European Journal of Operational Research 2005;161(2): [25] Ballestero E, Maldonado JA. Objective measurement of efficiency: applying single price model to rank hospital activities. Computers and Operations Research 2004;31(4): [26] Chang H. Determinants of hospital efficiency: the case of central government-owned hospitals in Taiwan. Omega 1998;26(2): [27] Maniadakis N, Thanassoulis E. Assessing productivity changes in UK hospitals reflecting technology and input prices. Applied Economics 2000;32: [28] Sarkis J, Talluri S. Efficiency measurement of hospitals: issues and extensions. International Journal of Operations and Production Management 2002;22(3): [29] Basson MD, Butler T. Evaluation of operating room suite efficiency in the Veterans Health Administration system by using data envelopment analysis. American Journal of Surgery 2006;192: [30] Chilingerian JA. Evaluating physician efficiency in hospitals: a multivariate analysis of best practices. European Journal of Operational Research 1995;80: [31] Goni S. An analysis of the efficiency of Spanish primary health care teams. Health Policy 1999;48: [32] Banker RD, Conrad RF, Strauss RP. A comparative application of Data Envelopment Analysis and tran slog methods: an illustrative study of hospital production. Management Science 1986;32(1): [33] Thanassoulis E, Boussofiane A, Dyson RG. Exploring output quality targets in the provision of perinatal care in England using data envelopment analysis. European Journal of Operational Research 1995;80: [34] Miller JL, Adam EE. Slack and performance in health care delivery. International Journal of Quality and Reliability Management 1996;13(8): [35] Chilingerian JA, Sherman HD. Benchmarking physician practice patterns with DEA: a multi-stage approach for cost containment. Annals of Operations Research 1996;67(1): [36] Hovenga EJS, Hovel J, Klotz J, Robins P. Infrastructure for reaching disadvantaged consumers. Telecommunications in rural and remote nursing in Australia. Journal of the American Medical Informatics Association 1998;5: [37] Walsh DC, Rudd RE, Moeykens BA, Moloney TW. Social marketing for public health. Health Affairs 1993: [38] Economic Commission for Africa (ECA). Information and Communication Technology for Health Sector. The African Development Forum 99. United Nations Economic Commission for Africa, Available from: depts/eca/adf/pforum.htm. [39] Rao SS. Integrated health care and telemedicine. Work Study 2001;50: [40] Charles BL. Telemedicine can lower costs and improve access. Healthcare Financial Management 2000;4(4):66 9. [41] Håkansson S, Gavelin C. What do we really know about the cost-effectiveness of telemedicine? Journal of Telemedicine and Telecare 2000;6(Suppl. 1):133 6.