2 Hospital corporatisation and its consequences across countries

Over the past two decades policymakers and researchers have taken an increased interest in hospital performance as the hospital is the main provider of both basic and advanced health care services, especially critical care services. This interest has become even more intense because of the widely reported problems faced by public hospitals in delivering these services, such as inefficiency, user dissatisfaction and failure to reach poor people (Jakab et al., Reference Jakab, Preker, Harding and Hawkins2002). Furthermore, the hospital sector absorbs the largest proportion of health care expenditure in both developed (Organization for Economic Cooperation and Development: OECD, 2010) and developing countries (Mills, Reference Mills1990a, 1990Reference Millsb). Hospitals thus often become the main target of reforms, which aim to improve their efficiency, equity and quality of care. One particular reform widely applied in the hospital sector is corporatisation, which intends to increase hospital performance specifically through the adoption of practices found in the private sector (Preker and Harding, Reference Preker and Harding2003; Vienazindiene and Ciarniene, Reference Vienazindiene and Ciarniene2007).

However, the consequences of corporatisation vary across countries. A number of studies have found these to be positive. For example, the successful implementation of corporatisation was carried out in the state of Victoria in Australia. In 1995, the government of Victoria restructured its 32 autonomous public hospitals into seven health care networks: six based in the regions and one specialist network. This reform was expected to increase the efficiency of hospitals by reducing duplication of services and sharing infrastructure costs. Corden (Reference Corden2003) found that the existence of these health care networks reduced unit costs. Real unit costs per case mix-adjusted inpatient in the network hospitals fell about 5% in the two years after the network was introduced. Furthermore, the introduction of networks contributed significantly to the improvement of quality and access to services as more hospitals were accredited and nearly 40,000 more patients were treated per quarter (a 20% increase) by September 1999.

Similar findings emerged in Asian countries, where Singapore, Malaysia and Hong Kong presented positive consequences of corporatisation. In 1985, Singapore applied corporatisation to its public hospitals after carrying out a successful pilot project at the new hospital at Kent Ridge. This reform achieved many positive results, including improvements in cost recovery ratios, services standards and responsiveness to patient’s needs (Phua, Reference Phua2003). In Malaysia in 1992, the government corporatised the National Heart Institute, following the successful corporatisation of the state-owned enterprise sector. This differs from previous cases, in that the reform was applied to a single hospital. Hussein et al. (Reference Hussein, Al-Junid, Nyunt-U, Baba and Geyndt2003) reported that the National Heart Institute gained higher revenue from private patients than other public hospitals while still maintaining its social function. At the same time, the stable amount of government subsidy earmarked for poor patients resulted in the Institute being able to successfully balance its social function with its role as a corporatised hospital. Turning to Hong Kong, in 1991 policymakers changed the public hospital network, integrating all public hospitals into a new, corporatised hospital authority. This reform is considered to be successful as it has resulted in increased bed availability, which in turn has reduced overcrowding and improved quality of care. Higher staff retention and unit cost reduction complement the success of the reform (Yip and Hsiao, Reference Yip and Hsiao2003).

Contrasting results of corporatisation in the hospital sector are found in New Zealand, the second developed country to implement hospital corporatisation under New Zealand company law in 1993. Here however, Scott et al. (Reference Scott, McKenzie and Webster2003) suggested that hospital efficiency remains unchanged. Given these mixed results, we now turn to the experience of Indonesia.

3 Corporatisation of public hospital in Indonesia

Indonesia is the fourth most populous country in the world, with over 230 million people inhabiting an extended archipelago between the Pacific and Indian oceans. Serving this population, Indonesia has a very low supply of hospital beds. In 2011, the ratio of hospital beds to population was 6.3:10,000, far below the global average of 30:10,000 (Awofeso et al., Reference Awofeso, Rammohan and Asmaripa2013). At the same time, utilisation of these beds was very low (55–60%) compared with that of the South East Asian region (over 80%) (Awofeso et al., Reference Awofeso, Rammohan and Asmaripa2013), implying high unmet needs for inpatient care.

The health sector in Indonesia includes public and private hospitals. These hospitals are grouped into four classes (A, B, C, D) according to the medical specialty, technological competencies and the number of beds (Rokx et al., Reference Rokx, Schieber, Harimurti, Tandon and Somanathan2009). Class A hospitals have highly specialised services (each is obliged to provide every specialist and most subspecialist services available) and a minimum of 400 beds, while Class B hospitals provide a minimum of 18 specialty services and 200 beds. Class C have four basic specialist services and a minimum of 100 beds. Class D, the lowest level of hospital, provides at least any two of the four basic specialist services and a minimum of 50 beds (Ministry of Health, 2010). Class A hospitals are majority-owned by central government, while class B hospitals are owned by provincial government and district government owns class C and D hospitals. Based on the services provided, each hospitals is categorised into two types: general and specialist hospitals. General hospitals provide services for all diseases, while specialist hospitals provide services for certain conditions, such as lung disease, leprosy and mental health.

Indonesia’s public hospitals have faced major challenges in recent decades, arising from increasing demand for good quality health services and increasing numbers of private hospitals dedicated to serve the same population. Moreover, public hospitals experience rigid bureaucracy (especially in managing finances, human resources and procurement), which in many cases leads to inefficiency. In 1991, in an effort to improve performance and reduce dependency on government subsidy, Indonesia began the process of reforming its public hospitals (Lieberman and Alkatiri, Reference Lieberman and Alkatiri2003). The objectives were the establishment of strict budget constraints to control costs and provide a better quality service for patients at a lower cost.

To put the current reform in context, a brief historical overview of reforms over the last four decades follows. The first step on the road to reform in 1991 was the government launch of the autonomisation of public (or ‘swadana’) hospitals. Autonomised hospitals have greater authority over their management and are encouraged to operate a commercial section in addition to the standard non-commercial section (Government of Indonesia, 1991). The maximum allotment of the commercial section is 25% of hospital beds (Ministry of Health, 1995). Here patients are charged a higher rate for enhanced room facilities such as private rooms, an en-suite bathroom, television and air conditioning. The generation of revenue is intended to lessen the hospital’s need for government subsidy.

Several studies examine the effect of the hospital autonomisation in Indonesia on patient satisfaction and financial performance. Gani (Reference Gani1996) looked at two public hospitals in West Java and concluded that hospital revenue and patient satisfaction increased under autonomisation, while at the same time the use of facilities by the poor was sustained. However, other studies present different results. Lieberman and Alkatiri (Reference Lieberman and Alkatiri2003) found that autonomisation failed to reduce government subsidy, while Bossert et al. (Reference Bossert, Kosen, Harsono and Gani1997) showed that subsidies to two autonomised hospitals had actually increased five years into autonomisation. In terms of equity, Bossert et al. (Reference Bossert, Kosen, Harsono and Gani1997) found that autonomous hospitals failed to protect the poor, allocating them fewer beds and charging higher fees than the non-autonomous hospitals.

Lieberman and Alkatiri (Reference Lieberman and Alkatiri2003) suggested that the possible explanation of these disappointing results was the increased role of private hospitals. In the mid-1990s the number of private hospital beds in Indonesia has increased significantly, equalling more than a third of public hospital capacity, up from a quarter 10 years before. This circumstance combined with the relative stability of hospital care demand, resulted in lower utilisation of public hospitals. An additional explanation was supplied by Bossert et al. (Reference Bossert, Kosen, Harsono and Gani1997), who used the concept of decision space to identify the range of choice available to hospital managers under autonomy. Examining management functions such as human resources, finance and service delivery, they concluded that autonomisation had conveyed a limited increase in decision space available to the managers, which together with institutional capacities and accountability is an important factor in the improvement of services when authority and responsibility are shifted to the local executives (Bossert and Mitchell, Reference Bossert and Beauvais2011).

After 10 years of the reform, Suwandono et al. (Reference Suwandono, Gani, Purwani, Blas and Brugha2001) analysed the cost recovery rate of inpatient units in three autonomous hospitals in East Java. They found that commercial sections in these hospitals covered only 60% of their costs. One reason for this was the failure of hospital managers to set realistic bed fees (partly the result of the cash-based accounting systems used in public hospitals at the time not being set up to record sufficient financial data to perform comprehensive and routine financial analysis). Accrual-based accounting systems are suggested to provide managers with regular data which compares revenue with costs. Similar empirical evidence is provided by Maharani et al. (Reference Maharani, Femina and Tampubolon2015), who extended the previous study by Suwandono et al. (Reference Suwandono, Gani, Purwani, Blas and Brugha2001) by examining the cost recovery rates in five service units (inpatient, outpatient, operating room, laboratory and radiology) and showed that autonomous hospitals failed to cover their cost in these units. Maharani et al. (Reference Maharani, Femina and Tampubolon2015) concluded that 20 years after autonomisation, hospital financial performance remains weak.

Most recently, in 2004 the Indonesian government launched corporatisation (Government of Indonesia, 2005). Corporatised hospitals have broader decision space on finance and inputs than autonomised hospitals have (see Table 1). Managers have significantly increased autonomy over the following functional areas: setting hospital budget, utilising revenue (both from subsidy and operational), initiating long-term investment programmes, contracting with private sector services and investors, procuring debt and accounts receivable, and having their own permanent (non-civil servant) staff.

Table 1

Decision space of autonomisation and corporatisation in Indonesia

Note:

^a Based on Bossert et al. (Reference Bossert, Kosen, Harsono and Gani1997).

^b Based on Regulation of Government of Indonesia number 23 and 24 (2005), Regulation of Ministry of Health number 12 (2003).

With this increased decision space, corporatised hospitals are expected by the stakeholder (central government) to depend less on government subsidies, to successfully compete with other public and private hospitals, and to better serve the community. As yet however, no study has examined the consequences of corporatisation in Indonesia, thus making an assessment of how it affects hospital performance of immediate importance. To do this, we use primary data from public hospitals in East Java province.

4 Data and methods

4.1 Data

In the summer and autumn 2013, we collected data from 54 public hospitals in East Java province (see Figure 2). This has 29 districts and nine cities, with a total area of 47,130 km². In 2007, East Java province was the second most populous province in Indonesia with total population of 36,895,571 inhabitants served by 55 public hospitals. This province was chosen for this study as it contains hospitals of every class, type and ownership, and is thus representative of all public hospitals in Indonesia. Of the 55 public hospitals, only one refused to participate, as it was dealing with the accreditation process at that time. The targeted respondents were hospital managers, who answered questions on the characteristics and performance of their hospital over a five-year period from 2008 to 2012.

Figure 2

Spatial distributions of the 54 public hospitals in East Java province.

The quality of data obtained (especially financial data) varied among hospitals: this study presents the best data available. It was collected over several visits; however, several hospitals did not provide sufficient data for our analysis (see Table 2). Out of all the sample hospitals, only 13 provided five-year financial data, while five-year data on BOR and LOS was available in 33 and 32 hospitals, respectively. All hospitals provided five-year data on the proportion of class 3 bed (serving mainly the poor), except one hospital which only started to run an inpatient unit in 2010.

Table 2

Data availability of hospital performance indicators

4.5 Methods

This study aims to measure the consequences of corporatisation on hospitals performance using a panel data regression model (Gujarati, Reference Gujarati2011). Including a binary indicator of corporatisation as an explanatory variable, the estimated model is as follows:

$$Y_{{it}} \,{\equals}\, \beta X_{{it}} {\plus}\delta Z_{{it}} {\plus}\gamma D_{{it}} {\plus}\alpha _{i} {\plus}v_{{it}} ,$$

where i=1, … , N indexes the hospital and t=1, … , T indexes year. X _it is a set of time-varying hospital characteristics, Z _it the time invariant hospital characteristics and D _it the binary variable for corporatisation. D the step variable, using the value 0 in all periods prior to corporatisation and 1 in all periods at and after the change.

Important is the presence of hospital heterogeneity, term α _i . This error component is often correlated with the explanatory variables, meaning that simple regression on the pooled data results in biased and inconsistent estimates of β, γ and δ. The fixed effect model remedies this problem by including N intercepts as parameters to be estimated, which for the estimates for β and γ are equivalent to within regression:

$$Y_{{it}} {\minus}\bar{Y}_{i} \,{\equals}\,\beta \left( {X_{{it}} {\minus}\bar{X}_{i} } \right){\plus}\gamma \left( {D_{{it}} {\minus}\bar{D}_{i} } \right){\plus}(v_{{it}} {\minus}v_{i} ),$$

where $$\bar{Y}_{{i\,}} \,,\,\bar{X}_{i} $$ and $\bar{D}_{i} $ are the mean of hospital characteristics. It is clear that fixed effect model will be consistent when $(X_{{it}} {\minus}\bar{X}_{i} )$ and $(D_{{it}} {\minus}\bar{D}_{i} )$ are not correlated with (v _it −v _i ), i.e. when X _it and D _it are strictly exogenous.

As an alternative to the fixed effect model, we use a random effect model. Assuming the unobserved effect α _i is not correlated with explanatory variables, the random effect model uses the following transformation to remove the fixed effect model and obtain efficient estimates in the presence of serial correlation:

$$Y_{{it}\,} {\equals}\,\beta X_{i} {\plus}\delta Z_{i} {\plus}\gamma D_{i} {\plus}\alpha _{i} {\plus}v_{{it}} .$$

We use the Hausman test to choose between the fixed effect and random effect model.