The Promotion of Environmental Management in the South Korean Health Sector—Case Study

: Because of the comprehensiveness and urgency of environmental challenges, every stakeholder needs to be engaged in reducing environmental impacts. The healthcare sector has rarely been studied, despite its intense effects on the environment, particularly through generating various forms of hazardous waste and intensively consuming energy and water. Many healthcare facilities exist in South Korea, and every citizen frequently visits hospitals thanks to the convenient system. To reduce the environmental impacts of the healthcare sector, the South Korean government has implemented various policy measures aimed at promoting environmental management in that sector. This study evaluated the eco-efﬁciencies of 21 hospitals from 2012 to 2015 using data envelopment assessment (DEA), used the analytical hierarchy process (AHP) to analyze hospital staff members’ answers to a questionnaire asking about the relative importance and performance of individual environmental management tasks, and also identiﬁed environmental management tasks that should be prioritized by building an importance-performance analysis (IPA) matrix using those questionnaire responses. This study found that eco-efﬁciencies have improved during the period, and that mandatory policy measures were more effective than voluntary agreements for improving eco-efﬁciency. This implies that rigorous reporting and monitoring should be implemented along with any voluntary agreement. In addition, this study found that the top priorities are “establishment of vision and strategy for environmental management” and “organization of task team for environmental management and task assignment”. This shows the necessity of additional policy measures, such as training or consulting to promote the priorities. In addition to policy recommendations for diffusing environmental management in the South Korean healthcare sector, the methodological approach sheds light for researchers interested in environmental management in the healthcare sector because previous studies depended on qualitative approaches, particularly case studies. Management of hazardous chemical use and reduction activity; ⑪ Compliance with domestic and international environmental regulations; ⑫ Publication of environmental report and environmental information disclosure; and ⑬ Response to stakeholders requests for environmental information. Investment in new and renewable energy and introduction of technology; ⑦ Management of GHG (Greenhouse Gas) emissions and reduction activity; ⑧ Management of emissions of water pollutants and reduction activity; ⑨ Management of waste generation and reduction activity; ⑩ Management of hazardous chemical use and reduction activity; ⑪ Compliance with domestic and international environmental regulations; ⑫ Publication of environmental report and environmental information disclosure; and ⑬ Response to stakeholders requests for environmental information. a vision and strategy for environmental management; ② Organization of task team for environmental management and task assignment; ③ Guideline and compliance with green purchasing; ④ Management of water use and reduction activity; ⑤ Management of energy use and reduction activity; ⑥ Investment in new and renewable energy and introduction of technology; ⑦ Management of GHG (Greenhouse Gas) emissions and reduction activity; ⑧ Management of emissions of water pollutants and reduction activity; ⑨ Management of waste generation and reduction activity; ⑩ Management of hazardous chemical use and reduction activity; ⑪ Compliance with domestic and international environmental regulations; ⑫ Publication of environmental report and environmental information disclosure; and ⑬ Response to stakeholders requests for environmental information.


Introduction
The 2015 Decision of the Conference of Parties (COP) in Paris emphasizes the participation of non-Party stakeholders, including cities and other subnational authorities. In addition, the decision encourages Parties to cooperate closely with non-Party stakeholders in an effort to strengthen and expand mitigation actions [1]. The transnational actions of non-Party stakeholders become "the heart of the new climate regime" [2] because every sector's participation is paramount for coping with environmental problems and attaining sustainable development goals. However, environmental their implications are discussed in the Results and Discussion section. Finally, major findings and limitations of this study are presented in the Conclusions section.

Literature Review
Studies of EM in hospitals investigated the factors motivating EM in hospitals [22,23], the status of EM implementation, and the effects of policy measures such as the VAs and the ETS on EM in hospitals.
There are competitive drivers, regulatory drivers, and ethical drivers that motivate hospitals to participate in EM [22]. Hospitals introduce EM to reduce administrative costs, to improve competitiveness by differentiating themselves to the public using certificates earned through EM practices, and to attain accompanying benefits such as safety enhancement [24,25]. Hospitals also conduct EM due to external factors such as government regulations and increased pressure for social responsibility [26]. In addition to the external pressure regarding social responsibility, the willingness of top management is another ethical driver for EM. The influence of these drivers varies along with various characteristics of hospitals, such as their governance structure and ownership. For example, the greater importance of regulations at public hospitals is attributed to their governance structure [22]. In addition, size and location have been found to affect the influence of regulatory and competitive drivers on hospitals' EM adoption and practice: expenses are perceived as more important in small hospitals than in medium ones, and even more important in large hospitals [23].
Many studies have focused on waste management in hospitals, due to the unique profile of waste generated in hospitals [8,9,13,[19][20][21]27,28]. Almost all case studies have pointed out issues related to the inappropriate segregation of waste generated in hospitals in developing countries and argued the necessity of training or educating hospital staffs [8,9,13]. One study argued for an integrated system to dispose of waste generated in hospitals, using composting, incineration, and recycling based on an LCA (Life Cycle Assessment) analysis of various waste disposal scenarios [27]. Recognizing the huge amount of waste generated from disposable materials used in hospitals, Campion et al. (2015) conducted a life-cycle assessment of the environmental impacts of the consumption of disposable packages, which consisted of "a set of sterile, disposable products prepackaged for a specific procedure," at 15 hospitals (12 in the United States, 2 in Thailand, and 1 in Global Links (an NGO (Non-Governmental Organization)). They then suggested an alternative green or environmentally preferred custom pack that could reduce the impact by 80% over the average disposable pack [29]. Some studies have been conducted on EM issues beyond waste management. For example, Saad (2003) focused on indoor pollutants, which can harm the health of hospital staffs and patients [30]. Carraro et al. (2016) conducted a comparative study that surveyed risks and legislation regarding hospital wastewater, finding that there are regulatory loopholes for some substances, including antibiotic residues, even in industrialized countries [20]. In addition to the waste or water, Chiarini, Opoku, and Vagnoni (2017) compared sustainable procurement practices at public hospitals in Italy and the United Kingdom by analyzing the responses to seven questions regarding sustainable procurement. They found that the Italian hospitals tended to focus on compliance with laws or regulations, while the British organizations tended to ask suppliers to improve their performance over time [31]. However, that study did not further investigate the factors that led to those differences. Some studies have investigated or suggested a framework for EM in the healthcare sector [7,32] and have identified barriers to the spread of EM [33]. Blass et al. (2017) developed a framework that enables consistent and robust measurement and reporting of environmental performance in hospitals [7]. Rian-Fogarty et al. (2016) investigated the effectiveness and applicability of an existing EM program, the Green Campus Program, in a case study of a teaching hospital in Ireland, finding that it is a "systematic approach to environmental action and education" [32]. Seifert (2018) conducted 14 in-depth telephone interviews with people who were in charge of EM at German hospitals that participated in voluntary environmental management systems. Through this qualitative approach, they found major barriers to the system, such as the heavy initial documentation burden and lack of knowledge and awareness across the all of the stages, including 'before initial registration', 'implementation and maintenance', and 'revalidation' [33].
Several studies have assessed EM and investigated its potential in hospitals. Faezipour and Ferreira (2018) identified factors influencing the water sustainability of three hospitals in the United States, and they modeled the causal relationship between the factors, water footprints of the hospitals, the cost of services and resources, and patient wellbeing. Water saving or water reuse was found to reduce the water footprint and cost of services and resources, and also enhance the overall wellbeing of the population that is affected by external freshwater quality [34]. Romero and Carnero (2017) established a multi-criteria model for environmental assessment of the health care sector. Their model consists of three criteria, 12 first-level sub-criteria, 46 indicators, and five alternatives. The questionnaires related to these indicators, e.g., "a value between 0 and 0.45 kg/patient of potentially infectious waste is obtained and is not removed in compliance with regulations," could be assessed using the AHP method based on different weightings of criteria. The authors tested this model to assess the environmental management status of a Spanish hospital. With a score of 0.6286, they showed the hospital carried out its environmental management well. The hospital did particularly well on the criteria related to energy consumption, hazardous waste generation, and legal matters [35].
Similarly, environmental management in the South Korean healthcare sector has been qualitatively studied. Early studies, conducted before the introduction of various policy measures, focused on defining eco-friendly hospitals and recommending policy measures for EM in the health sector. These presented key attributes of eco-friendly hospitals based on cases in developed countries [36], provided recommendations to improve EM in hospitals based on surveys [37], and pointed out the necessity of additional policy measures to strengthen EM in hospitals in preparation for an emission trading scheme [37]. More recent studies have scrutinized a specific policy measure or assessed EM practices in hospitals using specific criteria. Kang (2013) investigated the EIDS as practiced and presented recommendations for that specific policy [38]. Kim and Kang (2014) analyzed the EM practices at 44 hospitals using green management evaluation criteria and presented recommendations for the health sector [39]. Few studies have quantitatively and comprehensively investigated the performance of EM in hospitals. Therefore, a methodology for assessing EM in the health care sector, quantitatively as well as qualitatively, needs to be established.

Brief Survey of Healthcare Sector in South Korea
The number of South Korean health facilities totaled 89,919 in 2016 (see Table 1). Most of these were clinics, dental hospitals and clinics, oriental medical hospitals and clinics, and drug stores. Hospitals are classified according to specific requirements articulated in the Medical Service Act (Act No. 8366). Whereas clinics provide medical services primarily to outpatients, hospitals should have more than 30 patient beds. General hospitals should be equipped with at least 100 patient beds. Superior general hospitals are hospitals for treating high-level diseases; they are designated by the Ministry of Health and Welfare from general hospitals that meet the special requirement of having more than 20 specialized departments. As of 2016, there were 43 superior general hospitals, 298 general hospitals, and 2942 hospitals. institutes, mortuaries and crematoria, and prisons. Almost all of this medical waste (204,752.7 tons) was incinerated; the rest was treated with sterilization shredding devices [15]. The incineration cost for medical waste was about 600-800 thousand KRW per ton (about 560 to 750 USD per ton), which was 4-5 times more expensive than the cost of incinerating municipal wastes [16].

Policies and Instruments for Environmental Management in the South Korean Healthcare Sector
Four policy measures are practiced in South Korea to promote EM in the healthcare sector: Voluntary Agreements (VAs), the Environmental Information Disclosure System (EIDS), the Greenhouse Gas and Energy Target Management System (GETMS) and the Emission Trading System (ETS).
Since 2011, the Ministry of the Environment (MOE) has taken several initiatives, including establishing guidelines for EM in the public and service sector, including the healthcare sector; holding seminars and workshops for building EM capacities; sharing best practices in the healthcare sector; supporting the establishment of domestic eco-friendly hospital networks; and organizing expert meetings [41].
The MOE has implemented VAs with large general hospitals, university hospitals, and superior general hospitals since 2013 to promote EM in the healthcare sector. Ten hospitals signed a memorandum of understanding (MOU) for a VA in 2013; this signing ceremony between the Minister of the MOE and the hospital CEOs drew much attention from the press and public. In 2014, 11 more hospitals signed a VA, followed by nine more in 2015, six more in 2016, and five more in 2017; by the end of 2017, 41 hospitals participated in a VA for EM [17]. The hospitals with a VA can use the title of "Green Hospital" in their publicity. The MOE provides them up to 10 million KRW (about 9300 USD) in financial support for EM consulting, the establishment of an eco-friendly hospital network at the domestic and global level, and sharing of best practices and solutions regarding EM.
In July 2015, the South Korean government announced that it aimed to reduce national greenhouse emissions (GHG) by 37% compared to business-as-usual (BAU) emissions by 2030 [42]. To achieve this emission reduction target, two primary policy measures are in place: the GETMS and the ETS. Companies or business sites with GHG emissions or energy consumption exceeding specific criteria (Companies with emissions exceeding 50,000 tCO 2 eq/year or energy consumption exceeding 200 TJ/year, or business sites with emissions exceeding 15,000 tCO 2 eq/year or energy consumption exceeding 80 TJ/year, shall participate in the GETMS.) are required to participate in these measures. Entities with more substantial emissions or energy consumption are controlled under the ETS; those with emissions or energy consumption smaller than the ETS criteria are controlled through the GETMS. South Korea began the ETS on 1 January 2015. Tradable permits are allocated to controlled entities under the ETS according to a grandfathering principle (The whole allowance is distributed for free to the controlled entities for Phase I, from 2015 to 2017. From 2018 to 2020, 3% of the allowances are distributed through auction for Phase II). The entities must surrender permits equivalent to their emissions at the end of the year. Two hospitals, Kangnam St. Mary's Hospital and Seoul National University Hospital, have been subject to the ETS since 2015 [43]. Four general hospitals, Yangsan Busan University Hospital, Gacheon University Gil Hospital, Jeonbuk University Hospital, and Chungnam University Hospital, are controlled under the GETMS [44]. These controlled entities must make efforts to reduce their emissions or energy consumption to meet their annual target, which is established through a bottom-up process: the individual target is determined in consultation with the MOE [45]. At the end of 2016, 346 companies and facilities were controlled under the GETMS [45].
In addition, the MOE has operated the Environmental Information Disclosure System (EIDS) to promote the implementation of VAs and communication with people, to strengthen monitoring of environmental impacts, and to increase investment in eco-friendly companies by financial institutions and investors on the basis of the environmental information submitted to the system [46,47]. MOE initiated the EIDS as a pilot project in 2010; the MOE provides participating companies with benefits such as training programs and consultations. After 2012, the EIDS was expanded to include central government agencies, local governments, universities, public corporations, and large general hospitals and companies that have significant environmental effects [47]. These entities are required to disclose environmental information such as their business overview; plan and activity for EM; resource savings and pollution reductions; and EM attainments (e.g., EM-related awards, EM systems, or the status of the EM division) by Article 16.8 and 16.9 of the "Environmental Technology and Environmental Industry Support Act." As of 2015, 1216 companies and organizations participated in the EIDS, including 78 hospitals and public healthcare institutions. Details on the items of environmental information to be disclosed by the healthcare sector are shown in Table A1.
Thanks to government policy measures, some achievements have been observed. The hospitals participating in VAs saved costs through reducing energy and water consumption (See Table 2). In addition to these quantitative achievements, a Green Hospital in Seoul introduced the idea of green service, which treats elderly volunteers who clean up the neighborhood for free and helps low-income elderly people using revenue collected from an annual bazaar held with environmental NGOs (interview on 11 June 2017).

Methodology and Data
DEA measures relative efficiency by comparing the efficiencies of individual decision-making units (DMU) in targets such as governments, hospitals, companies, programs, and policy measures. It was developed by Charnes, Cooper, and Rhodes (1978) [48] to measure the concept of efficiency, as defined by Farrel in 1957. DEA finds the "efficient frontier," which is the efficiency of the best practices; measures the "distance" of efficiency values of other DMUs from the "efficient frontier"; and provides rankings of efficiencies of individual DMUs [48,49]. As a result, potential improvements are identified for inefficient DMUs. DEA has advantages in providing information on the reference set and potential improvements that should be set as an example for inefficient units to become more efficient. Therefore, it has been widely used not only as a measure of efficiency, but also as a tool for setting goals for analysis and improvement.
Among the variants of the DEA model, this study uses an input-oriented model that aims to improve efficiency by minimizing the inputs required to achieve the targeted output [50]. The input-oriented model is more appropriate to environmental management in hospitals because it pursues enhancements in eco-efficiency by reducing inputs/environmental burdens given the desired output rather than increasing output while maintaining inputs. This study uses the input-oriented Banker, Charnes, and Cooper (BCC) model (1984) [51], allowing variable returns to scale (VRS). VRS assumes the increased outputs may or may not be proportional to the increased inputs. This differs from constant returns to scale (CRS), which assumes that the increased outputs should be proportional to the increased inputs [50]. The BCC allows measurement of technical efficiency by separating the effects of scale [51]; this is more appropriate for measuring pure technical efficiency while taking into consideration the different sizes of hospitals.
This study collected the data from all 21 hospitals that participated in the EIDS in South Korea over the period from 2012 to 2015 (see brief information on the 21 hospitals in Table 3) (The value of sales is adjusted to the 2012 value, after accounting for inflation.). First, this study analyzed four mandatory reporting items of the 20 reporting items that included quantitative information under the EIDS as input variables (see Table A1): annual water consumption (ton of water), annual energy consumption (TOE), annual waste generation (ton of waste), and annual hazardous chemical use in each hospital (kg of chemical). It used two other mandatory reporting times, the total annual sales (million Korean Won) and the number of patients (persons), as output variables. The variables can be seen in Table 4, and detailed values for each variable over the period can be found in Table A2. Only quantitative data that must be reported to the EIDS was used in order to obtain data from all of the EIDS-participating hospitals. Other qualitative information that is attainable from the EIDS was used for the additional IPA analysis. Eco-efficiency is defined "as being achieved by the delivery of competitively priced goods and services that satisfy human needs and bring quality of life, while progressively reducing ecological impacts and resource intensity throughout the life-cycle, to a level at least in line with the Earth's estimated carrying capacity." Furthermore, "progress in eco-efficiency can be achieved by providing more value per unit of environmental influence or unit of resource consumed" [52]. Eco-efficiencies are calculated and evaluated using Frontier Analyst (version 4, Banxia Software Ltd, Kendal, Cumbria, United Kingdom), a DEA software application. For example, the absolute value of eco-efficiency of H01 is estimated (see the equation below). Among the estimated eco-efficiencies of the 21 hospitals, the benchmarked or highest value is identified, and its relative eco-efficiency is set as one. As of 2015, H01 is one of the most eco-efficient hospitals, along with eight others (See Table 5). Then, the eco-efficiency values of the other hospitals are compared to the benchmarked value to establish their eco-efficiency values relative to the highest eco-efficiency value. A smaller value means relatively lower eco-efficiency: In addition, this study surveyed 51 hospital staffs about EM (the 21 hospitals whose eco-efficiencies were assessed by DEA in this study and an additional 30 hospitals participating in the eco-friendly hospital network). The questionnaire used a symmetrical 5-point Likert scale to ask about the relative importance of individual reporting items of the EIDS as major management indicators with respect to EM diffusion in hospitals (see Table A3 for the detailed questionnaire). The reporting items are grouped into four categories: environmental management (or procurement) system establishment; resource/energy management and reduction activities; greenhouse gas emissions and environmental pollution management and reduction activities; and social/ethical responsibility compliance. There are specific items under each individual category (e.g., within the category of social/ethical responsibility compliance, there are three specific questions on the relative importance of domestic and international environmental regulation compliance vs. sustainability report publication, domestic and international environmental regulation compliance vs. responses to stakeholder requests for environmental information; and sustainability report publication vs. responses to stakeholder requests for environmental information). The questionnaire also asked about the performance of EM tasks for 13 items (Of 20 information disclosure items for the South Korean healthcare sector, items analysed in the DEA analysis were excluded when constructing the questionnaires.) that should be reported to the EIDS in individual hospitals. The questionnaire was emailed to the 51 hospital staffs on 8 May 2017; and 33 responses were received by 15 May 2017. These responses to the multi-criteria decisions were assessed using AHP to solve multi-criteria complex problems. The AHP allows the respondents, the staff responsible for environmental management in these hospitals, to evaluate the relative importance of items through pairwise comparison [53]. Among 33 responses, 29 responses with a consistency ratio smaller than 0.1 were analyzed. Saaty (1980) proposed a consistency index to evaluate the level of consistency of final decisions among respondents [54]. Generally, if the consistency index is less than 0.1, the consistency of the pairwise comparison matrix is considered to be good and reliable. The responses to the questionnaires were analyzed using the importance-performance analysis (IPA) matrix. This study draws this matrix using SPSS 18 (IBM, Armonk, NY, USA). The IPA is a technique for diagnosing the performance of services or products, comparing performance to their importance and suggesting insights into priorities [55]. Specifically, services or products located in quadrant 4 of the IPA matrix-the service or product is highly important, but performance is low-should receive more focus [56] (see Figure 1).   Table 5 presents the changes in eco-efficiency from 2012 to 2015. The variance in the average differences of the eco-efficiency scores was statistically significant at the 5% level, as seen in Table 6.

DEA Results
The average eco-efficiency of the 21 hospitals was 0.940 in 2015. The eco-efficiency of 12 hospitals was 1.0, which means that those hospitals were the most efficient. Among these hospitals, four hospitals participated in government regulation through the GETMS. The other nine hospitals were less ecoefficient than the 12 benchmarked hospitals. The average eco-efficiency score rose from 2012 (0.830) to 2014 (0.933) but then dropped slightly in 2015 (0.902). The gradual increase in average eco-efficiency from 2012 to 2014 shows improvements at the relatively less eco-efficient hospitals. In contrast, the slight decrease in 2015 implies regression by the less eco-efficient hospitals. H11, H12, H14, and H17 participated in government regulation through the GETMS; these hospitals showed gradual improvement in their eco-efficiency scores from 2012 to 2015. The scores at H11, H12, and H17 had been particularly lower in 2012. The notable improvement of their eco-efficiency, up to 1.0 in 2015, can be attributed to the implementation of the GETMS.  Table 5 presents the changes in eco-efficiency from 2012 to 2015. The variance in the average differences of the eco-efficiency scores was statistically significant at the 5% level, as seen in Table 6.

DEA Results
The average eco-efficiency of the 21 hospitals was 0.940 in 2015. The eco-efficiency of 12 hospitals was 1.0, which means that those hospitals were the most efficient. Among these hospitals, four hospitals participated in government regulation through the GETMS. The other nine hospitals were less eco-efficient than the 12 benchmarked hospitals. The average eco-efficiency score rose from 2012  Although H02 and H05 participated in VAs, their eco-efficiencies not only did not improve, but became even worse. This means that these two hospitals preserved the status quo or regressed regarding EM while other hospitals improved their eco-efficiencies. This result makes the effectiveness of VAs questionable, shows that VAs cannot guarantee that hospitals will execute activities or practices for EM, and implies that VAs have limitations in their current form and need corrective measures such as strict monitoring, reporting, and a feedback system. The potential for improvement was evaluated for hospitals with lower eco-efficiency scores by comparing their actual production conditions and processes with hospitals that had the same production conditions and processes. Figure 2 shows the potential improvements that are achievable at nine inefficient hospitals. They could reduce water consumption by eight percentage points (ppt), energy consumption by 9 ppt, waste generation by 10 ppt, and hazardous chemical use by 24 ppt, to be in line with the most efficient hospitals. Notably, there is a significant room to reduce hazardous chemical use in less eco-efficient hospitals. Although H02 and H05 participated in VAs, their eco-efficiencies not only did not improve, but became even worse. This means that these two hospitals preserved the status quo or regressed regarding EM while other hospitals improved their eco-efficiencies. This result makes the effectiveness of VAs questionable, shows that VAs cannot guarantee that hospitals will execute activities or practices for EM, and implies that VAs have limitations in their current form and need corrective measures such as strict monitoring, reporting, and a feedback system. The potential for improvement was evaluated for hospitals with lower eco-efficiency scores by comparing their actual production conditions and processes with hospitals that had the same production conditions and processes. Figure 2 shows the potential improvements that are achievable at nine inefficient hospitals. They could reduce water consumption by eight percentage points (ppt), energy consumption by 9 ppt, waste generation by 10 ppt, and hazardous chemical use by 24 ppt, to be in line with the most efficient hospitals. Notably, there is a significant room to reduce hazardous chemical use in less eco-efficient hospitals. By analyzing the potential improvements at hospitals with low eco-efficiency scores, each hospital could identify which input variables needed to be improved. In addition, more substantial alternatives need to be reviewed and adjusted to achieve an eco-efficiency score of 1.0. By analyzing the potential improvements at hospitals with low eco-efficiency scores, each hospital could identify which input variables needed to be improved. In addition, more substantial alternatives need to be reviewed and adjusted to achieve an eco-efficiency score of 1.0. Table 7 presents the results of a combined weighting analysis of the responses to the questionnaires that asked about the relative importance of the individual EIDS reporting items for EM in hospitals.

IPA Results
Of 33 collected responses, four responses with a consistency ratio larger than 0.1 were excluded [54]: thus, a total of 29 responses were analyzed. As a result of the combined weighting analysis, it was found that the hospital staffs recognized "management of energy use and reduction activity" as the most important EIDS reporting item for diffusing EM. In addition, the scores for "the establishment of a vision and strategy of EM", "organization of task team for EM and task assignment", "management of water use and reduction activity", and "management of GHG emissions and reduction activity" are greater than the average score of 0.077. This implies that these items are recognized as the highest priorities for promoting EM in the healthcare sector. The results of the relative performance of individual EIDS reporting items are shown in Table 8. It was found that "compliance with domestic and international environmental laws and regulations," "management of waste generation and reduction activity," "management of energy use and reduction activity," and "management of water use and reduction activity" scored the highest: 4.20, 4.10, 4.10, and 4.00, respectively. Scores were lower for "investment in new and renewable energy and introduction of technology," "publication of environmental report and disclosure," "organization of task team for EM and task assignment," "response to stakeholder requests for environmental information," and "guideline and compliance with green purchasing." Table 8. Relative performance of EIDS reporting items.

EIDS Reporting Items
Score Ranking 1. Establishment of a vision and strategy for environmental management 3.60 9 2. Organization of task team for environmental management and task assignment 3.50 Using the results for importance and performance, the IPA matrix was constructed (see Figure 3). "Management of water use and reduction activity" and "management of energy use and reduction activity" were in the first quadrant, which means that the current state needs to be maintained because both the importance and the performance of the task are high. Tasks located in the second quadrant should be priorities because the activity is important, but its performance is low. "Establishment of a vision and strategy for EM," "organization of task team for EM and task assignment," and "management of GHG emissions and reduction activity" fall in this area of "concentrate here." In contrast, "guideline and compliance with green purchasing," "investment in new and renewable energy and introduction of technology," "publication of environmental report and environmental information disclosure," and "response to stakeholder requests for environmental information" have low priorities, which means that both the performance and importance of tasks are low: these tasks are less emphasized. In the fourth quadrant ("possible overkill" area), tasks are relatively unimportant, but a great deal of effort is concentrated on them; unnecessary work should be discarded. "Management of emissions of water pollutants and reduction activity," "management of waste generation and reduction activity," "management of hazardous chemicals use and reduction activity," and "compliance with domestic and international environmental regulations" were included in this quadrant.
Comparison of the relative importance of environmental input variables for 2015 enables us to deepen the analysis of eco-efficiency. Water use (0.762) and energy consumption (0.754) are placed in the fourth quadrant because the eco-efficiencies of these two environmental input variables were higher than the average. On the other hand, the eco-efficiency scores of waste generation (0.532) and hazardous chemicals use (0.323) were lower than average, as were their importance levels. This result reveals the gap between actual performance (eco-efficiency measured by the individual environmental input variable) and the recognized performance (responses to the relative performance of corresponding activities). More specifically, hospital staff did not recognize the management of waste generation and hazardous chemical use as priorities; instead, they believed that their performance on those items was excellent because waste and hazardous chemical measures practiced in hospitals are straightforward and easy to mobilize staff for participation. In contrast, the actual eco-efficiency or performance of these items is lower, due to the large gap with the superior performance of hospitals regarding management of waste and hazardous chemical use. This result implies that it is necessary to adjust or raise the level of hospitals' reduction targets so they can be more eco-efficient. excellent because waste and hazardous chemical measures practiced in hospitals are straightforward and easy to mobilize staff for participation. In contrast, the actual eco-efficiency or performance of these items is lower, due to the large gap with the superior performance of hospitals regarding management of waste and hazardous chemical use. This result implies that it is necessary to adjust or raise the level of hospitals' reduction targets so they can be more eco-efficient.

Conclusions
Through the analysis of changes in the eco-efficiency of 21 hospitals from 2012 to 2015, this study investigated the status and effectiveness of environmental management in the South Korean healthcare sector. It is found that eco-efficiencies improved during the four years that were studied, and that the best 12 hospitals were equally eco-efficient in 2015. Obligatory policy measures, such as GETMS, were found to be more effective for improving the eco-efficiency of hospitals, while it was found that the effectiveness of voluntary agreements was not guaranteed. Therefore, rigorous reporting and monitoring should accompany VAs. It is worth noting that the United Kingdom regularly estimates and issues reports on the carbon footprint of the healthcare sector, which achieved an 11% GHG emission reduction from 2007 to 2015 [14]. Although the average value of eco-efficiency improved from 0.830 in 2012 to 0.902 in 2015, a significant potential for more improvement of eco-efficiency exists by reducing hazardous chemical use. The existing approach to hazardous chemical use in healthcare sector is somewhat reactive, focusing on how to effectively separate and safely dispose of hazardous wastes. This room for improvement implies the necessity of proactive approaches such as the introduction of innovative technologies or the development of alternatives. es practiced in hospitals are straightforward actual eco-efficiency or performance of these ormance of hospitals regarding management that it is necessary to adjust or raise the level icient.

Concentrate here
Each number refers to: ① Establishment of a rganization of task team for environmental compliance with green purchasing; ④ gement of energy use and reduction activity; tion of technology; ⑦ Management of GHG anagement of emissions of water pollutants n and reduction activity; ⑩ Management of mpliance with domestic and international ntal report and environmental information nvironmental information. of 21 hospitals from 2012 to 2015, this study management in the South Korean healthcare he four years that were studied, and that the atory policy measures, such as GETMS, were cy of hospitals, while it was found that the nteed. Therefore, rigorous reporting and hat the United Kingdom regularly estimates hcare sector, which achieved an 11% GHG verage value of eco-efficiency improved from ore improvement of eco-efficiency exists by o hazardous chemical use in healthcare sector rate and safely dispose of hazardous wastes. ctive approaches such as the introduction of s.
Management of hazardous chemical use and reduction activity; excellent because waste and hazardous chemical measures practiced in hospitals are straightforward and easy to mobilize staff for participation. In contrast, the actual eco-efficiency or performance of these items is lower, due to the large gap with the superior performance of hospitals regarding management of waste and hazardous chemical use. This result implies that it is necessary to adjust or raise the level of hospitals' reduction targets so they can be more eco-efficient.

Keep up the good work Possible overkill
Low priority Concentrate here

Conclusions
Through the analysis of changes in the eco-efficiency of 21 hospitals from 2012 to 2015, this study investigated the status and effectiveness of environmental management in the South Korean healthcare sector. It is found that eco-efficiencies improved during the four years that were studied, and that the best 12 hospitals were equally eco-efficient in 2015. Obligatory policy measures, such as GETMS, were found to be more effective for improving the eco-efficiency of hospitals, while it was found that the effectiveness of voluntary agreements was not guaranteed. Therefore, rigorous reporting and monitoring should accompany VAs. It is worth noting that the United Kingdom regularly estimates and issues reports on the carbon footprint of the healthcare sector, which achieved an 11% GHG emission reduction from 2007 to 2015 [14]. Although the average value of eco-efficiency improved from 0.830 in 2012 to 0.902 in 2015, a significant potential for more improvement of eco-efficiency exists by reducing hazardous chemical use. The existing approach to hazardous chemical use in healthcare sector is somewhat reactive, focusing on how to effectively separate and safely dispose of hazardous wastes. This room for improvement implies the necessity of proactive approaches such as the introduction of innovative technologies or the development of alternatives.
Compliance with domestic and international environmental regulations; excellent because waste and hazardous chemical measures practiced in hospitals are straightforward and easy to mobilize staff for participation. In contrast, the actual eco-efficiency or performance of these items is lower, due to the large gap with the superior performance of hospitals regarding management of waste and hazardous chemical use. This result implies that it is necessary to adjust or raise the level of hospitals' reduction targets so they can be more eco-efficient.

Conclusions
Through the analysis of changes in the eco-efficiency of 21 hospitals from 2012 to 2015, this study investigated the status and effectiveness of environmental management in the South Korean healthcare sector. It is found that eco-efficiencies improved during the four years that were studied, and that the best 12 hospitals were equally eco-efficient in 2015. Obligatory policy measures, such as GETMS, were found to be more effective for improving the eco-efficiency of hospitals, while it was found that the effectiveness of voluntary agreements was not guaranteed. Therefore, rigorous reporting and monitoring should accompany VAs. It is worth noting that the United Kingdom regularly estimates and issues reports on the carbon footprint of the healthcare sector, which achieved an 11% GHG emission reduction from 2007 to 2015 [14]. Although the average value of eco-efficiency improved from 0.830 in 2012 to 0.902 in 2015, a significant potential for more improvement of eco-efficiency exists by reducing hazardous chemical use. The existing approach to hazardous chemical use in healthcare sector is somewhat reactive, focusing on how to effectively separate and safely dispose of hazardous wastes. This room for improvement implies the necessity of proactive approaches such as the introduction of innovative technologies or the development of alternatives.
Publication of environmental report and environmental information disclosure; and excellent because waste and hazardous chemical measures practiced in hospitals are straightforward and easy to mobilize staff for participation. In contrast, the actual eco-efficiency or performance of these items is lower, due to the large gap with the superior performance of hospitals regarding management of waste and hazardous chemical use. This result implies that it is necessary to adjust or raise the level of hospitals' reduction targets so they can be more eco-efficient.

Conclusions
Through the analysis of changes in the eco-efficiency of 21 hospitals from 2012 to 2015, this study investigated the status and effectiveness of environmental management in the South Korean healthcare sector. It is found that eco-efficiencies improved during the four years that were studied, and that the best 12 hospitals were equally eco-efficient in 2015. Obligatory policy measures, such as GETMS, were found to be more effective for improving the eco-efficiency of hospitals, while it was found that the effectiveness of voluntary agreements was not guaranteed. Therefore, rigorous reporting and monitoring should accompany VAs. It is worth noting that the United Kingdom regularly estimates and issues reports on the carbon footprint of the healthcare sector, which achieved an 11% GHG emission reduction from 2007 to 2015 [14]. Although the average value of eco-efficiency improved from 0.830 in 2012 to 0.902 in 2015, a significant potential for more improvement of eco-efficiency exists by reducing hazardous chemical use. The existing approach to hazardous chemical use in healthcare sector is somewhat reactive, focusing on how to effectively separate and safely dispose of hazardous wastes. This room for improvement implies the necessity of proactive approaches such as the introduction of innovative technologies or the development of alternatives.
Response to stakeholders requests for environmental information.

Conclusions
Through the analysis of changes in the eco-efficiency of 21 hospitals from 2012 to 2015, this study investigated the status and effectiveness of environmental management in the South Korean healthcare sector. It is found that eco-efficiencies improved during the four years that were studied, and that the best 12 hospitals were equally eco-efficient in 2015. Obligatory policy measures, such as GETMS, were found to be more effective for improving the eco-efficiency of hospitals, while it was found that the effectiveness of voluntary agreements was not guaranteed. Therefore, rigorous reporting and monitoring should accompany VAs. It is worth noting that the United Kingdom regularly estimates and issues reports on the carbon footprint of the healthcare sector, which achieved an 11% GHG emission reduction from 2007 to 2015 [14]. Although the average value of eco-efficiency improved from 0.830 in 2012 to 0.902 in 2015, a significant potential for more improvement of eco-efficiency exists by reducing hazardous chemical use. The existing approach to hazardous chemical use in healthcare sector is somewhat reactive, focusing on how to effectively separate and safely dispose of hazardous wastes. This room for improvement implies the necessity of proactive approaches such as the introduction of innovative technologies or the development of alternatives.
Based on the IPA results, priorities for EM in the healthcare sector were identified, areas in which performance is low even though the tasks are essential. Interestingly, "establishment of a vision and strategy for environmental management" and "organization of task team for environmental management and task assignment" are included in these priorities. As García et al. (2015) argued, changes in management bring challenges such as time and cost [10], which are likely to slow down the spread of EM in the healthcare sector. Therefore, know-how and best practices need to be shared to enable followers to get past the challenges more smoothly. The government needs to support building a network or platform for sharing knowledge among hospitals. In addition, the government should focus on training and improving awareness of these priorities.
Based on quantitative analysis of reported data and qualitative analysis using responses to questionnaires, this study assessed EM in South Korean hospitals and provides implications for improving policy measures to diffuse EM in South Korean hospitals. The methodological approach sheds light for researchers interested in EM in the healthcare sector because previous studies depended more on qualitative approaches, particularly case studies. However, the number of analyzed hospitals is too small to represent all the hospitals in South Korea due to the limited availability of the data. Furthermore, because this study surveyed staffs at the 21 hospitals that participated in the EIDS program, the hospitals that did not participate in the EIDS were not represented. Additional analysis to investigate the status and performance of EM at those hospitals needs to be conducted shortly.

Acknowledgments:
The authors express sincere gratitude to the reviewers, whose comments enriched this paper.

Conflicts of Interest:
The authors declare no conflict of interest.