1
Selecting Financial Health Ratios and reference values for
the Dutch Caribbean Hospitals based on benchmarking of
Dutch Hospitals 2012-2014
Master Thesis by Jan-Fedde Bouma
1JEL Classification: G30; I18; I11; M41
Keywords: Financial ratio analysis, Benchmarks, Healthcare, General Hospitals
MSc International Financial Management
University of Groningen, the Netherlands Faculty of Economics and Business
2
Abstract
The rise of healthcare costs in proportion to personal GDP in developed countries is deemed unsus-tainable. The transition to the financing of healthcare driven by market forces is expected to counter this trend. Fostering of competition and financial benchmarking play are in an advanced stage in the Netherlands. However, financial ratios used in benchmarking can also provide guidance to financial management where the scope for competition is limited, but the government continues to bear respon-sibility for health care provision, conditions present on the islands of the former Netherlands Antilles. Identifying a comparable group of hospitals that operate under market forces that can provide these benchmarks (in the shape of reference percentiles of relevant financial ratios) is an academic chal-lenge. In this dissertation I attempt to identify relevant ratios and reference values for the hospital on Aruba, Bonaire and Curacao based on benchmarking data from 76 Dutch hospitals between 2012 and 2014. Hospital size, remoteness and type (General vs Top Clinical) are potentially relevant parameters in this matching process. I refute the hypotheses that size and remoteness are relevant parameters in the Dutch context, but find support for the hypothesis that the type of hospital matters (in the “domain of profitability”). Because the Antillean hospitals send patients for specialized treatment abroad, I do not include the Top Clinical Hospitals in the peer-group, and calculate median, 10th and 90th percen-tiles for the remaining 48 Hospitals. Principal Component Analysis shows a considerable overlap of ratios used by Dutch accountancy firms in the “profitability” and “repayment capacity” domains, and I therefore select a single ratio for each domains. I use only the peer group data of 2014, because only in this year, when additional government subsidies cease, factor analysis successfully separates the results of the seven calculated ratios into the three domains recognized in financial industry. The per-centiles applied to the ratios I calculated from Antillean hospital data between 2010 and 2014 do show striking anomalies that reflect e.g. historical financial deficits, and highlight areas that deserve further research. Based on financial benchmark experience for non-commercial peripheral hospitals in the USA, I recommend ratios (additional to those particularly of interest to bankers and investors) that appear more relevant for the day to day financial management capabilities of hospital administrations. The assertion by one of the Dutch firms that hospital size strongly determines financial performance (and ranking) appears only to apply to Top Clinical Hospitals (I excluded for an Antillean peer-group). Their recommendation for scale-enlargement is not likely to apply to general hospital services and challenges the belief in mergers to achieve cost-efficiency.
3
Table of Contents
Abstract Page 2
1. Introduction 4
2. Research framework & Hypotheses 6
3. Benchmarking in management, finance and in the healthcare sector 8
3.1 Benchmarks in management and finance 8
3.2 Financial benchmarks in the healthcare sector 9
4. Ratios used for hospital benchmarking in the Netherlands 10
4.1 Ratios used by BDO and EY 10
4.2 The “stress-test” of Ernst & Young 12
4.3 Ratios used by BDO 13
5. Data and Methods 14
5.1 Data description 14
5.2 Methods used 14
6. Results 16
6.1
Differences of ratio scores between years and firms
166.1.1 Sources of variability between years 16
6.1.2 Sum-scores. Differences between BDO and EY 17
6.1.3 Principal Component Analysis of Dutch hospital ratios 2012-2014 18
6.2 Descriptive Statistics 20
6.3 Selection of suitable hospitals as “peer group” for hospitals on the ABC islands 21 6.3.1 The effect of hospital remoteness on performance in the Netherlands 21 6.3.2 The effect of hospital size and type on performance in the Netherlands 23 7. Application of benchmark ratios and reference percentiles to hospitals on ABC islands 27 7.1 Financial performance of ABC hospitals between 2010 and 2014 27
8. Discussion , Limitations & recommendations 28
9. References 29
10. Appendices 31
I. Dutch General Hospitals included in analysis 31
II. EY Descriptive statistics General Hospitals in the Netherlands 2010-2014 32
III. Results Ratios ABC hospitals 2010-2014 33
IV. Social, economic and demographic back ground ABC islands 34
V. Critical Access Hospitals in the USA 37
4
1. Introduction
Medical progress, ageing and associated rise of chronic diseases have pushed up healthcare costs considerably over the last decades. In 2014 the healthcare sector already accounted for 10% of global GDP (EY re-imagined, 2015), and as fraction of the personal GDP affordability have become particularly stained in Western countries. “Across the globe, governments, health care delivery systems, insurers, and consumers are engaged in a persistent tug-of-war between two competing priorities: meeting the increasing demand for health care services and reducing the rising cost of those services” (Deloitte, 2015). As the explosive rise in healthcare costs is considered unsustainable in the long run, attempts have been made in Western countries to improve cost-efficiency by subjecting the healthcare system to the forces of the market. In the Netherlands the government decided to move towards a system of regulated competition in the healthcare where the tariffs for healthcare are no longer set by the Ministry of Health, but negotiated between the health insurer and care providers. The change in regulation resulted in the transformation in healthcare and initiated a range of changes across the industry. One of these changes is that hospitals in the healthcare industry will have to behave like private companies in a free market. Therefore, the healthcare sector, with a strong non-profit background, must conform to financial performance requirements and practices maintained by banks and investors. They must adapt to sell and deliver their “health products”, have a capital base, borrow on the capital markets, attract investors and submit annual financial reports. Their “financial health” is likely to reflect both the quality and efficiency of the care provided by health staff, general and financial management. In the market, financial health is assessed through benchmarking using various “ratios” to reflect, e.g. profitability, solvency (Chapter 4).
5
Figure 1.1 Comparison of 2 years Hospital rankings from Algemeen Dagblad (AD). Each diamond representing a single hospital (Botje, Berenschot, 2014)The variability has been attributed to the lack of reliable performance indicators (on which ratios and rankings are based) as hospitals so far do not collect and manage (yet) the relevant data in a systematic and uniform fashion (rapport Beperkt Zicht van het AMC/UvA, 2012). Botje (2014) does not openly challenge the validity of the financial health judgements, but the “scatter” in figure 1.1 published does raise this concern. EY & BDO scores, as I will show below, display indeed considerable differences between firm and between years.
6
A financial crisis in the hospital of one of the islands in recent years has threatened the provision of health services. At present hospitals on the ABC islands have no mandatory publication of annual financial statements, and comparisons between these hospitals are not possible. However, I have managed to obtain the relevant data (2010-2014) for the hospitals on Aruba, Bonaire and Curacao to calculate the same ratios selected by BDO and EY for the assessment of Dutch hospitals (Methods). The analytical part of this thesis concerns the selection of relevant financial ratios and calculation of the reference values for a representative sample of Dutch hospitals (See Research framework, Chapter 2). As a preliminary exploration, I applied the percentiles (10, 50 and 90th) of the peer group of Dutch hospitals to the ratios I calculated for the annual data of the ABC hospitals between 2010 and 2014, and highlight some striking anomalies. The Dutch reference values I select in this thesis can serve as the basis for a more in debt investigation of the strength and weaknesses of the financial management of the hospitals on the islands.
2. Research framework and Hypotheses
In order to generate appropriate ratios for hospital management on the ABC islands and reference values that do justice to the condition on the islands, the experience on benchmarking in the Netherlands can serve as a source of information to study the determinants of financial hospital achievement. The differences between the hospital evaluation of the two accountancy firms has to be comprehended, as well as the variability in scoring of the same hospital by the same firm between years. Most important is the understanding how parameters that may differ between The Netherlands and the Antilles can affect the outcome on the “financial health”. That in order to issue the most realistic guidance figures for financial ratios Antillean hospital management can aim for.
The hospitals on the ABC islands are in comparison to those in the Netherlands rather small (annual budgets of around 50 to 85 million Euro, as compared to the group of smallest hospitals the Netherlands with revenues below 150 million Euro; See chapter 7.2), they are used by the whole population (people have little choice and health care providers have little competition, and there may be less scope for economies of scale). For this reason, hospitals in more remote areas of the Netherlands, those that are less populated, have fewer hospitals, have less hospital choice, and have less competing service providers, could be a better match for the Antillean hospitals. Thirdly, patients that need more specialized services are referred to service providers abroad (e.g. Colombia, Cuba, USA and the Netherlands). To match the Antillean with the Dutch hospitals, general hospitals which also provide more specialist services (the so-called Top-clinical Hospitals), should perhaps be excluded from the peer-group. Confusingly, EY separates these 2 types of hospitals for benchmarking purposes, but BDO does not. EY reports the results for General hospitals, excluding the TCZ, and BDO reports the results for General hospitals including the TCZ.
7
The scores on financial ratios are more favorable
1) In more populated Dutch provinces with more choice and competition (Null-hypothesis: remoteness does not affect ratio scores)
2) In larger hospitals that can benefit from economies of scale (Null-hypothesis: hospital size does not affect ratio scores)
3) For TKZ with higher remuneration for specialized services and perhaps their oligopolistic position. (Null-hypothesis: Top Clinical Hospital status does not affect ratio scores)
In order to maximize the sample of Dutch hospitals to be included in the peer-group, I will have to take these variables into account if the Null-hypothesis (Ho) is rejected. Subsequently, I then have to match the sample of Dutch hospitals accordingly or statistically correct for its effect. If the hypothesis is not rejected and I cannot find support for the effect of the variable on ratio-scores, I retain the hospitals in my peer-group (correction is not required, or, I do not have to exclude Dutch hospitals even if they superficially appear to be a poor match).
Of these potential confounders, hospital size appears the most important. A matching exercise with Antillean hospitals would reduce the peer group to a small number of hospitals, and thus with produce reference values with undesirable large confidence intervals. There is indeed evidence from international studies supporting the importance of hospital size. For the relative small Critical Access Hospitals in the USA (See Appendix V), size was indeed shown to be a very significant (Pink et al., 2007). For the larger NHS hospitals in the UK, the effect of size could only be demonstrated in one of the 3 years studied (Monitor, 2014). One of the Dutch accountancy firms (BDO) draws conclusions from its benchmarking results (particularly from its ranking exercise), in favour of larger hospitals, with a strong conclusion and advocacy for scale enlargement and amalgamation of (smaller) hospitals in order to improve financial health of Dutch hospitals (BDO 2012,2015).
8
conform to the 3 separate domains expected from commercial enterprises. The PCA will further measure the overlap between ratios within these domains. This will aid selection and possible reduction of the number of ratios.
3. Benchmarking in management, finance and in the healthcare sector
3.1 Benchmarks in management and finance
Benchmarking “compares a company’s current performance against its own previous performance or that of its competitors. It provides a standard of comparison for measurement.” (Brooks, 2010). It helps to identify best performance, provides a method to set aggressive targets for improvement, and identifies potential strategies on how to improve performance, and is considered a key component of many organizational performance measurement systems (Spendolini, 2002). Conditions for successful benchmarking focus essentially on careful preparation of the monitoring process of the relevant indicators, staff involvement and inter-organizational visits” (Ettorchi-Tardy et al., 2012).
Benchmarking involves the application of financial ratios, well validated in the financial industry. Ratios are concise measures of financial performance that are calculated using data from published (usually annual) financial statements. “Their popularity is due to their simplicity in measurement that expresses valid and very important relations among the economic data of the companies involved” (Curtis, 2009). Inputs to this process include balance sheets, income statements, and other financial and performance data. Common applications of financial ratios within analysis arrangements include:
o Snapshot: Showing a "picture" of an organization's financial condition at a specific point in time
o Single time period: Showing the performance of an organization over a specific period of time, generally for a quarter, year-to-date, or full-year period
o Year-to-year: Showing performance compiled by year, displaying trends and changes occurring over several individual time periods
o Peer-to-peer: Comparing performance ratios of an organization or bank against the performance ratios of a like organization
Brooks (2010) ‘Core Concept of Financial Management’ states that “financial ratios are relationships between different accounts from financial statements that serve as performance indicators. We can look at specific performance areas of a company by selecting key pieces of information from financial statements and by analyzing this information at a point in time or over a specific time horizon. We can either look at trends over time of an individual company or compare firms of different companies at a specific point in time.” Brooks identified five different types of ratios:
1. Liquidity ratio: Can the company meet its obligations over a short term?
9
3. Asset management ratios: how efficiently is the company managing its assets to generate sales/revenue?
4. Profitability ratios: How well has the company performed overall?
5. Market value ratios: How does the market (investors) view the company’s financial prospect?
Of these 5 domains, liquidity, solvency and profitability ratios have so far been applied in the healthcare setting, liquidity mainly in relation to Interest and debt repayment capacity.
3.2 Financial benchmarks in healthcare
The use of financial benchmarks in healthcare goes back several decades. “Financial ratio analysis is an accepted approach to hospital performance evaluation” (Zeller et al. 1996, 161). The late adoption of ratios in the analysis of the financial soundness of hospitals at the end of 1970s was attributed to “the possibility that financial pressures to the hospital industry were not as pervasive as they were in other industries and the lack of availability of comparable financial statement information” (Watkins 2000, 75). Ratios are used in tracing specific aspects of financial performance and especially in esti-mating liquidity, valuating profitability, performing competitor analysis and forecasting corporate bankruptcy.
10
4. Ratios used for hospital benchmarking in the Netherlands
In this chapter I review the ratios used by two Dutch accountancy firms EY and BDO. Each has pro-posed their own selection of ratios and the methods used to calculate the aggregated score for each individual hospital. To investigate the origin of differences between judgments on hospitals between firms and between years the used ratios and methods of each firm are described below. I first describe the similarities and differences of ratios used by these firms.
4.1 Ratios used by BDO and EY
The ratios used by BDO and EY are shown in Table 4.1. For each domain (A-C) there is one ratio used by both BDO and EY (shown in blue). EY has added the Current ratio (B) in a later stage, but it is not included in their sum-scores.
Domain
Ratio
EY Critical
Val-ue for “health”
Used
by:
A ) Interest- and
Repayment Capacity
(“Re-payment capacity”)
DSCR =Debt Service Coverage
Ratio
>1.25
EY &
BDO
ICR= Interest Coverage Ratio
>2
EY
Net debt /EBITDA (measure of
leverage)
<3.5
EY
B) Financial Structure &
Position (“Solvency”)
Current ratio = Current assets/
Cur-rent liabilities
(>1 EY)
BDO
Solvency Ratio = Equity (own
as-sets)/ Total assets.
>20%
EY &
BDO
C) Results and
Perfor-mance (“Profitability”)
Profit Margin = Net Income/Total
Revenues
>1.5%
EY &
BDO
ROIC = Return on invested capital
EBIT/(equity + Net debt)
>7.5%
EY
Profitability = Operating results /
Total assets
BDO
Table 4.1: Ratios in 3 domains (A-C) and ratios used by EY (grey), BDO (black) and those used by both (blue).The ratios are calculated using reported financial figures from the balance sheet and income statements of year reports. For some ratios pre-calculations need to be performed. These are: the EBIT, EBITDA and Net Debt.
11
Debt Service Coverage Ratio (DSCR): This ratio shows if the cash flow from operations is sufficient to meet the regular and debt payment obligations of the hospitals and calculated in this thesis as:DSCR = EBITDA / (Interest Expense + Repayments) (1)
Interest Coverage Ratio (ICR): This ratio calculates the extent to which the company can meet the interest expense. MacGrayHill (2012) refers to this ratio as Times interest Earned and defines it as a measure of long term solvency calculated as:
ICR = EBIT / Interest Expense (2)
Net debt /EBITDA: Measure of how the business is financed. This ratio shows in how many years the company can repay its debt, given that the operating result stays the same.
Net debt / EBITDA = Interest bearing debt / EBITDA (3)
Current ratio: This ratio is a liquidity ratio and measures the ability to meet the short term debt obli-gation.
Current Ratio = Current assets / Current Liabilities (4)
Solvency ratio: This ratio is an indication of the capital structure of a company and show the capabil-ity to cover potential losses in the future. Also a high solvabilcapabil-ity ratio means a relative low amount of debt.
Solvency ratio = Equity / Total Assets (5)
Profit Margin: This ratio shows the relative performance margin of the company and measures how the firm manages its operations and how effectively it uses its assets.
Profit Margin = Net income / Total revenues (6)
Return on Invested Capital (ROIC): This ratio shows how efficient a company transforms its assets into profit. The ROIC shows how well the hospital is capable of investing its money to generate re-turns. Literature shows different ways of calculating this ratio, all calculations show more or less the same result. In thesis I follow the calculation used by EY in order to compare the ratios of the ABC islands with the Dutch results:
ROIC = EBIT / (Equity + Net debt) (7)
Profitability: This ratio shows the relationship between the performance and total assets.
12
4.2 The “stress test” of Ernst & Young
From 2012 onwards EY has issued an annual financial barometer of the Dutch healthcare system. These are divided into 6 sectors (EY, 2012). Three curative (The General hospitals, the academic hospitals and the top clinical hospitals), and a sample (n =30) of each of the 3 care sectors (Geestelijkegezonheidszorg, Gehandicaptenzorg and Verpleging, Verzorging en Thuiszorg). The justification for the division into 3 curative care sectors was not clearly motivated at the time, but appears rational because of the delivery of a different range of health “products”. Academic hospitals and general hospitals have different mission and remuneration structures that are likely to impede comparisons of financial health (Flex 2005, 26). Top clinical hospitals (TCH) (28 in the Netherlands) are hospitals that besides ordinary hospital care provision, also serve as a referral hospital for more complex care, based on their specialization in one or more care areas. As a result they receive patients from a wider region than the general hospitals. These TCH hospitals carry out innovative research and take part in the Dutch training of specialist (Wegwijs, 2014).
EY calculates for each of the subgroups the averages, providing an estimation of the financial health by sector. Of a large array of financial ratios used in the financial world, 6 were selected within 3 domains; Interest and Repayment capacity (3x), Results and Performance (2x) and Financial Structure & Position (1x) (See table 4.1). The score for each ratio is consists of a pass (1 point) or fail (0 point) in relation to an (in the sector accepted) “critical score”. The sum of these scores (between 1 and 6) is therefore a semi quantitative score based on 6 good/bad verdicts. The ratios and critical values used in the test are unchanged over the years, to allow comparisons between years,
In a period of the post 2008 financial crisis EY has referred to their assessments as a “stress tests”, presumably because ratios beyond their desirable range are associated with failure. However, strictly, “stress test” is a misnomer. The International Actuarial Association (IAA) defines a stress test as; “a projection of the financial condition of a firm or economy under a specific set of severely adverse conditions that may be the result of several risk factors over several time periods with severe consequences that can extend over months or years” (IAA, 2013). The EY assessment does not involve scenario testing under simulated adverse conditions. The “stress test” is a sum of binary valued financial benchmarks (ratios). There appears to be no inherent reasons not to preserve the full quantitative information of the ratios, and this simplifying procedure has important implications.
13
4.3 The BDO “Benchmark”
BDO has annually published their financial analysis of 82 Dutch hospitals since 2009. Contrary to EY, BDO includes a group of hospitals with distinct features, the Top Clinical Hospitals (TCH) as part of the General hospitals, as these hospitals do also provide general services besides more specialized care products.
Between 2009 and 2012 the emphasis of the analysis has been on the overall performance in the health care sector rather than on individual hospitals. From 2013, similar to the initiative of EY (see above), the results for individual hospitals are reported. BDO choose to allocate hospitals, for each ratio score, between 0 and 2. Instead of the 2 point scale (0 or 1) used by EY, a 5 point scale is used (0, 0.5, 1, 1.5 and 2. See table 4.2). The sum score is used by BDO to issue a yearly ranking of the Dutch hospitals. Due to delays in annual reporting of the hospitals for the year 2014, an annual report for this year was issued in 2015.
Awarded
Points
DSCR ratio
(A)
Current ratio
(B)
Solvency
ratio (B)
Profit margin
(C)
Profitability
(C)
0
<0.5
< 0.5
<5 %
<0.0 %
<0.0
0.5
0.5 - < 1
0.5 - < 0.9
5 - < 10%
0.0 - < 1.0%
0.0 - < 1.0
1
1 - < 1.5
0.9 - < 1.0
10 - < 15%
1.0 - < 2.0%
1.0 - < 2.0
1.5
1.5 - < 2.0
1.0 - < 1.1
15 - < 20%
2.0 - < 3.0%
2.0 - < 3.0
2.0
> 2.0
> 1.1
> 20%
> 3.0%
> 3.0
EY “pass”=1>1.25
> 1.0 (added)
>20%
>1.5%
NA
Table 4.2 BDO scoring on 5 ratios in 3 categories (A-C see Table 4.1) Ratios also used by EY (in blue) and EY “pass” values
14
5. Data and Methods
5.1 Data description
The data to calculate the financial ratios are derived from annual financial statements of the general hospitals. These statements can be accessed by the public for a fee. For hospitals in the Netherlands I have used the financial parameters extracted by the EY healthcare department in Amsterdam, and agreed not to share these data with other parties in view of the extensive amount of cost and effort that was incurred by the firm in processing the information. This dataset contains the parameters from which the EY, but also the BDO rations could be re-constructed and the sum-scores (re)calculated. The dataset includes the years 2012, 2013 and a large fraction of data for 2014. There are 3 kinds of hospitals listed in the Netherlands; General hospitals, Top clinical hospitals and Academic hospitals. The Academic hospitals are excluded in the analysis by both BDO and EY. However, BDO and EY use a different definition of “general hospitals”. BDO has included the “Top clinical hospitals” (n=28) with the General hospitals, whereas EY has only used the “General hospitals” (n=48) in the strict sense.
As it is not mandatory for healthcare institutions on the Antilles to deposit annual statements, it has been difficult to obtain the annual reports for recent years. Eventually, the financial directors of all three hospitals have given me access to the annual statements from 2010 to 2014 under the condition that the data will kept confidential. The advisory board of the St. Elisabeth Hospital (SEHOS) stipu-lated that none of the information regarding the hospital can be made public. The precarious financial position of this hospital is likely to have contributed to this requirement. The other two general hospi-tals are: Dr. Horacio E. Oduber Hospital on Aruba (HOH) and Funadashon Mariadal on Bonaire (MD). The descriptive statistics of the Dutch General hospitals and the ABC hospitals can be found in Ap-pendix II and ApAp-pendix III respectively. For Appendix IV I used data for the construction of demo-graphic and mortality profiles from WHO, World Bank and CBS sources. (Life expectancy at birth, childhood mortality (<5yr), infant mortality (<1yr) and maternal mortality: mortality up to 1 year after childbirth). For per capita income I used the Gross National Income (GNI) per capita, and for interna-tional caparisons I used the Purchasing Power Parity (PPP) corrected figures for the health expenditure per capita.
5.2 Methods used
15
coefficient squared (R2) to assess the variance explained. Because some ratio variability exists between years of individual hospitals, for some regressions I have combined results for 2 years (2012 and 2013). Where means of 2 samples are compared, I used the T-test to determine significance levels.
Preliminary analysis has shown that most financial health ratios for the hospitals are linearly correlat-ed. The financial health ratios have been divided into 3 domains by BDO and EY, Liquidity (credit worthiness), Profitability and Solvency that may reflect “underlying factors” of the financial health ratios. However, the ratios selected to represent these domains are multiple, and are different for the two firms. To test the rationale of the division between domains, and make a judicial selection of os within the domains for the ABC hospitals, Principal Component Analysis (PCA) is used on all rati-os used by both BDO and EY. This is done with the programIBM SPSS Statistics 23. My hypothesis is that the hypothetical “factors” will be differentiated representing the 3 “domains” (Debt repayment capacity, solvency and profitability). In order to remove redundant or duplicating ratio’s within the domains, and select a few representative ratios for each domain (rather than recommend all ratio’s used by both BDO and EY), I calculate the factor loading of each ratio. Those with the highest load-ings will be selected for the defined purposes of the thesis, that is, suitable for recommendation to the health sector on the ABC islands.
I further used a multivariate analysis using the IBM SPSS Statistics 23 package. I used four independ-ent variables namely: revenue, number of beds, distance to hospital and the type of hospital. The first three variables are continuous variables and the last one is a categorical variable with two categories, general hospitals and top clinical hospitals. Dummies are made for these two categories and the cate-gory general hospitals will be used as the reference group. The dependent variables are the chosen ratios that follow from the principle component analysis. First, the effect of revenue, number of beds, distance to hospital and type of hospital on the chosen ratios is analyzed. Secondly, the effect of reve-nue, number of beds, distance to hospital and type of hospital on sum scores of BDO is analyzed. Thirdly, separate analyses are performed for general hospitals and top clinical hospitals to see if the effect of the independent variables is different. Lastly, the interaction effect of the type of hospital and the number of beds on the sum scores of BDO is analyzed.
16
and transformed this cut-off score to the value of 5. Scores between 0 and 5, and 5 and 10 were calcu-lated subsequently proportional to the chosen range. See table 5.1 below.
Chosen range Critical values EY and scoring BDO
DSCR 0 – 2.5 (EY Crit. Value: 1.25; BDO scores: <0.5, 0.5-<.09, 1 – 1.1, 1.5-2, >2) ICR 0 – 4 (EY 2)
N debt/EBITA 1.5 – 5.5 (EY 3.5)
Solvency 0 - 40 (EY 20; BDO scores: < 5%, 5-<10%, 10-<15%, 15%-<20%, >20%). Profit Margin 0 – 3 (EY 1.5; BDO scores: <0%, 0-<1%, 1-<2%, 2-<3%, >3%)
ROIC 2.5 – 12.5 (EY 7.5)
Table 5.1: Chosen range and Critical values
6. Results
6.1 Differences of ratio scores between years and firms and selection of ratios for Antillean
benchmarking
Understanding the sensitivities of ratio, domain and sum-scores between companies and years can provide guidance in selecting appropriate ratio for the ABC islands. As BDO and EY have different baskets of ratio spread over different domains, the domain emphasis of each firm is apparent. BDO has chosen 2 out of 5 ratios in the “profitability” and 1 out of 5 in the “interest/debt repayment capacity” domain, whereas EY has 2 out of 6 for “profitability” and 3 out of 6 in the last domain. There is a striking emphasis of EY on “interest/debt repayment capacity” (50%) compared to BDO (16.7%) which may reflect the focus of EY’s client base on the credit-worthiness of hospitals. This should explain part of variability between the individual hospital sum-sores of these firms. In 6.1.2 I will show that the method of (sub) scoring chosen by the firms also contributes the differences between firms, and for the same reason between years too. Botje (2014) has highlighted the striking changes in hospital scores between years (See introduction). In 6.1.1 I will explore the contribution of different domains to changes in scores between 2012 and 2013. In 6.1.3 I investigate whether the underlying vectors identified by PCA correspond to the 3 domains, and assess the loading of the individual ratios on these domains to guide a choice of ratios for the ABC islands.
6.1.1 Sources of variability between years
17
Fig 6.1A. Scatter plot of Solvency in 2012 and 2013 (Linear regression line stippled).Fig 6.1B, C. Scatterplot of DSCR in 2012 and 2013 (B) and Profit margin in 2012 and 2013 (C).
There is a larger degree of variability in the DSCR (interest repayment capacity) although still signifi-cant (for the linear regression, and on the polynomial regression shown in the figure (6.1B). The larg-est degree of variability appears to originate from the “Profit margin” (Fig 6.1C). There is no hint of any relationship, the origin of which may stem from the supplementing subsidies (“transitie” pay-ments) in 2012 and 2013. These subsidies from the government in 2012 and 2013 accounted for ap-proximately 70% of the profitability in the sector (Nederlandse Zorgautoriteit, 2014). BDO published (2013) the differences of the hospital rankings with and without “transitie” payments, and they are indeed considerable (not shown). These results favor the year 2014, without government subsidies, as the most suitable to infer reference values for the ABC islands. This choice will be further supported by PCA (See 6.1.3).
6.1.2 Sum-scores. Differences between EY and BDO
Rankings of BDO are based on the sum-score of equally weighted ratios as explained in Chapter 4. Scores for BDO range from 1 to 10 (max. 2 points for each of the 5 ratios), and 1 to 6 for EY (1 point for each of the ratio’s above (or below for the Net debt/EBIDTA ratio), the cut-off threshold- values. As explained in 6.1 the EY sum-scores has a larger weighting on “Interest and debt repayment capacity”. Here I show the sum-scores of both accountancy firms for 2012, using the truly general hospitals (thus excluding the top clinical ones).
18
Figure 6.2 A & B. Regression between sum scores of BDO (x-axis) and EY (y-axis) for Dutch general hospitals (n= 40) for 2012 (A). Note that not all hospitals are individually visible due to some identical scores. Largest anomalies in red, & the same ratios with re-vised sum-scores for EY in the same year, using a continuous scale for ratio scoring (B).There is a broad correspondence in marking the hospitals by both firms judging by the R2 values for 2012, but the unexplained variability around 40% is in excess of what could be expected based on the structural (domain) bias of both firms. Particularly with the scoring method of EY, hospitals with ratio values just below the critical value receive zero points in the EY ratings. As EY de facto “degraded” their ratios (on a ratio-scale) to a nominal scale one (good or bad), and BDO theirs to a 5 point scale, discrete sizeable chance fluctuations are introduced. More outliers are found on the X-axis (with 0 EY score on the Y axis) than on the Y-axis. Using the same 2012 ratio data, but scoring the ratio for EY on a linear scale (between a set range, see methods: 5.2) a much higher R2 value is obtained (See Fig 6.2B). The 15% improvement (78 – 63%) in the variance explained appears simply due to EY’s de-grading the ratio-scale to a nominal (binary) scale. This demonstrates the high price paid for the simplistic approach to producing sum score. Further improvements in reconciling the BDO and the EY score can be expected when the BDO score too is made fully continuous. There is no technical reason to degrade the ratio scales, and where the use of sum-scores and rankings is desirable, the use of a continuous scale can be recommended. However, applying sum-scores to the hospitals of the ABC islands appears arbitrary and without merit.
6.1.3 Principal Component Analysis for Dutch hospital ratios 2012-2014
The domains underlying the individual ratios relate to different aspects of financial assessments and interests of stakeholders. A banker considering a loan to an organisation is concerned with the ability of the recipient to repay his interest and his principal, whereas an investor looking for yield would be more interested in the profitability of a company. Ratio’s that reflect these different perspectives are conceptually but also empirically selected to measure these different domains. The financing of hospital care originates from non-commercial beginnings. If the transition of hospitals to commercial enterprises has been successful, one would expect, in reverse, that the results of the ratio scores will identify these core-domains of the financial industry. I test this hypothesis here using Principal Component analysis (See Chapter 5.2, methods). The loading of individual ratio’s on these domains
R² = 0,6324 0 1 2 3 4 5 6 0 1 2 3 4 5 6 7 8 9 10
A) Sum scores BDO (x-axis) & EY (y-axis) R² = 0,7816 0 2 4 6 8 10 0 1 2 3 4 5 6 7 8 9 10
19
(principal vectors) would further inform to what extend the individual ratio represents the domain, and ratios can be compared and perhaps reduced in number in case of overlap.
I have used 8 ratio values of 76 hospitals (2012 and 2013) and 72 hospitals (2014) for the PCA, thus including both General and Top Clinical hospitals. This analysis is based on linear regression, and exploratory analysis using scatterplots indeed supported this assumption/requirement. The results of the analysis for 2012, 2013 and 2014 show that only in 2014, when transition payments ceased, the 3 components (domains) are fully differentiated (Fig 6.3A). In 2012 and 2013, the ratios of the repayment domain were divided. The ICR being part of the profitability domain, and the DSCR a principal vector (“domain”) on its own (See Figure 6.3 B&C). In the years 2012 and 2013 the government made its last “transition payments” to bridge the ease the adaptation to market forces. In 2012, 267.6 million was disbursed, and in 2013, 189.2 million 2013 (Nederlandse Zorgautoriteit, 2014). These additional subsidies are likely to have distorted the income distribution of the Dutch hospitals, and this may explain the incomplete differentiation of the ratio results over their domains. This result may be interpreted as a success of the transition period, and strongly favors 2014 as the most suitable year to use as the basis for reference values for the ABC islands.
Figures 6.3 A-D. Vector loadings of ratios (Rotated components matrices) in 2014 (6.3B), 2012 (6,3C) and 2013 (6.3D), with the total variance explained by vector (6.3A).
With very similar vector loadings of the ratios in the 3 domains, I have chosen a single ratio to repre-sent each domain.2 These are DSCR (for the domain of Repayment capacity A), Profit margin,
20
vency (for Financial Position and structure, B) and Profit margin (for Profitability domain C). The main reason for this choice is that EY and BDO share these ratios (See Table 4.1). These three ratios that represent the three domains are the dependent variables for the multivariate regression analysis performed with SPSS 23.
6.2 Descriptive statistics
The following tables (Table 6.1 A-C) presents the means (M), standard deviations (SD), and the correlations of the independent variables (revenue, number of beds, hospital remoteness and type of hospital) and the dependent variables (DSCR, current ratio and profit margin).
Table 6.1 A. Descriptive Statistics and Variable Correlation (2012)
Variables M SD 1. 2. 3. 4. 5. 6. 1. Revenue (mln) 196,43 103,14 2. Beds 465 216 .821** 3. Distance to Hospital (km) 6,5 1,7 ,132 ,177 4. Type of Hospital ,37 ,49 .712** .699** -,006 5. A) DSCR 2,68 3,53 -,143 -,076 ,046 -,119 6. B) Current Ratio 0,97 0,32 ,068 ,194 .342** ,146 ,147 7. C) Profit Margin 1,96% 2,59% ,070 ,124 .307** ,037 .306** .376** Notes: Type of Hospital is coded as 0= GH and 1=TCH; *p<0.05; **p<0.01; N=76
Table 6.1 B. Descriptive Statistics and Variable Correlation (2013)
Variables M SD 1. 2. 3. 4. 5. 6. 1. Revenue (mln) 205,43 106,89 2. Beds 451 198 .855** 3. Distance to Hospital (km) 6,5 1,7 ,118 ,144 4. Type of Hospital ,37 ,49 .701** .745** -,006 5. A) DSCR 2,62 2,46 -,171 -,132 -,023 -,139 6. B) Current Ratio 1,05 0,37 ,047 ,129 .226* ,155 ,168 7. C) Profit Margin 2,45% 2,86% ,064 ,058 -,131 ,126 .300** .335** Notes: Type of Hospital is coded as 0= GH and 1=TCH; *p<0.05; **p<0.01; N=76
Table 6.1 C. Descriptive Statistics and Variable Correlation (2014)
21
6.3 Selection of suitable Dutch hospitals as “peer group” for hospitals on the ABC islands
6.3.1 The effect of hospital remoteness on performance in the Netherlands
There are striking regional differences in population density and hospitals density within the Nether-land (Figure 6.4A) with the lowest densities observed in the more remote provinces FriesNether-land, Gro-ningen, Drenthe and Zeeland (FGDZ provinces).
Figure 6.4A. Number of hospitals within 20 km, and Distance to hospital at neighbourhood level in 2007 (Deerenberg et al., 2009).
The Caribbean island general hospitals could perhaps be better matched with those in the FGDZ prov-inces. Inchapter 2 (research framework) I have explained that remoteness may serve as the reciprocal of competition level, a factor that may affect financial performance. I investigate if there are regional (population and/or hospital) density dependent differences in the financial performance.
Figure 6.4B. Density of population and General Hospitals in the Dutch provinces and Curacao (cross). The FGDZ provinces (green diamonds) are not individually marked.
It is not surprising that population density and (general) hospital are correlated as shown in figure 6.5, but it provides a basis to identify 3 regions within the population/hospital density range. North &
Flevoland Overijssel Gelderland Utrecht N.Holland Z. Holland N-Braband Limburg Curacao R² = 0,7292 0,0 1,0 2,0 3,0 4,0 5,0 0 200 400 600 800 1000 1200 1400 H o sp itals /1000 km 2 Persons/km2
22
South Holland and the FGDZ provinces form the ends of the spectrum. Densities of people and hospi-tals in Curacao match the group of Provinces in between the extremes. Utrecht, with a population den-sity of 905/km2, is the largest outlier and excluded from this exploratory regional analysis. In table 6.2 the financial performances on the EY ratios are shown for the 3 identified clusters.
North & South Holland FGDZ Other
Average Revenue/ Hospital (SD) 132.7 (56.3) 120.6 (30.2) 160 (80.9)
No. General Hospitals Reporting 17 6 25
DSCR Mean 3.56 3.70 2.40 SD 5.64 2.22 1.68 Solvency (%) Mean 16.12 22.60 17.84 SD 8.40 13.76 8.80 Profit Margin (%) Mean 1.39 2.33 2.30 SD 3.44 1.54 1.81 Table 6.2A
From this regional comparison, contrary to expectation, the FGDZ provinces score best on the 3 ratios. These regional differences in financial performance could be biased too by differences in the popula-tion structure (e.g. average age) and disease prevalence. To explore this for the 3 province clusters we averaged the values per province and for the disease prevalence for the GGDs in the 3 regions (not weighted for the size of the population in the parts that make up one region).
2012-2013 North & South Holland Other FGDZ Curacao
Average Age 41.5 41.9 43.4
Gen Hosp. /106people 3.1 2.9 4.9 6.5
% Staff costs 56.6 56.0 54.2
% Asthma 8.6 9.3 8.8
% Diabetes 6.3 6.0 5.5 9.2
% Card.& Vasc. disease 5.4 5.3 5.3
% Cancer 2.0 2.0 1.5
Table 6.2B. Statistics for 3 clusters of provinces (corrected for age and sex & level education).
The average age in the FGDZ provinces is higher than in the other clusters and can therefore not account for the more favourable ratio’s considering the higher health costs for ageing populations. The prevalence of cancer in the FGDZ provinces is below the values for the other clusters, as is the incidence of diabetes. Note that the prevalence of obesity and related diabetes is strikingly higher on the Antilles as shown for Curacao in Table 6.2B. The patterns of these Dutch regional differences provide little justification of selecting a cluster with a better match for the ABC islands.
23
data to obtain Table 6.2A&B), and 2 to out of 3 hospitals that have closed in recent years, 2 are from the FGDZ provinces (Ziekenhuis Zorgsaam, Terneuzen and Zionberg, Dokkum). These peripheral hospitals may also have received more transition payments in 2012 and 2013. This possibility is supported by a multivariate regression with remoteness, hospital size (number of beds) and type of hospital (TCH vs GH) as independent variables. With remoteness parameterized as the number of km from a hospital for the average person in a province (CBS, 2014), remoteness is significant (for the current ratio and profit margin) in 2012 and, only marginally for the current ratio in 2013. In 2014, the year that transition payments ceased, no significant relations are identified. Results remain the same when hospital (annual) revenue is used instead of the number of beds to quantify hospital size. The results of the multivariate analysis for the current ratio for 2012-2014 are given in appendix VII. Therefore I cannot reject the null-hypothesis that remoteness matters for 2014, the year with the least biased ratio scores. As no convincing evidence can be found that general hospitals in less populated and more remote parts of the Netherlands yield less favourable financial ratios (and for 2014 not significantly better scores), no matching of the Dutch peer-group appears required.
6.3.2 The effect of hospital size and type on performance in the Netherlands
24
Figure 6.5. The 6 (EY) Ratios expressed as fraction of critical (EY) values for 3 revenue domains (in Million Euro) of Dutch hospitals (General and Top Clinical).I explored the relationships between revenue (annual turnover) as approximation of hospital size for the General Hospitals (n= 48), and the 3 selected ratios (for each domain; Profit margin, DSCR and Solvency, See figure 6.1) with simple scatterplots using averaged data for 2012 and 2013. No hint of any relationship is apparent in any of the 3 plots (R2 < 0.01, results not shown).
I further examined the possible effect of hospital size on financial performance in a multivariate re-gression (See Methods chapter 5). The analysis mentioned in 6.3.1 examined the effect of 3 independ-ent variables, average distance (in a province) from hospital, size of hospital (number of beds) and type of hospital (GH vs TCH) on the ratios of the 3 domains. Nor size, nor hospital type showed sig-nificant contributions in any of the years (2012-2014) in any of the domains. However, distance to the hospital had significant on the Current Ratio in the years 2012 and 2013. This is shown in Table 6.3 Also in view of BDO’s strong conclusions regarding the importance of size for financial health, the possible contribution of hospital size was more thoroughly explored.
Current Ratio 2012 Current Ratio 2013 Current Ratio 2014
Distance to hospital .332** .234** .205
Number of beds .061 -.048 -.058
Type of hospital .106 .192 .224
R Square .141 .076 .071
Table 6.3 Model summary for the effect of “Distance to hospital”, “number of beds” and “type of hospital” on “cur-rent ratio” (Note: Type of Hospital is coded as 0= GH and 1=TCH; **p<0.01
When the general hospitals and the Top Clinical ones (TCH) are treated as separate groups, and I ex-clude, for reasons mentioned in 6.3.1 the distance to hospital as parameter, a different picture emerges. In the group of Top Clinical hospitals there is, indeed, a sizeable effect of hospital size on perfor-mance, for which I can identify the domain for profitability as the one responsible. In table 6.3, I show the results for the regression for both groups of hospitals for the year 2014, with “profit margin” as dependent variable. The un-adjusted R2 for the TCH is 0.246 (p=0.007), whereas no effect shows amongst the truly general hospitals. No significant correlations could be identified in the other
25
mains (results not shown). From the results in 2012 and 2013 (when subsidies were phased out) it appears that size-effect in TCH has become stronger (Results not shown). The un-adjusted R2 value for 2012 was 0.005 (not significant), and for 2013 it was 0.2, with no significant results for the truly general hospitals (R2 in 0.038 in 2012, and 0.003 in 2013).
It appears that this size dependent “profit-margin” results of TCH have a strong bearing on the sum-scores (and rankings) BDO has issued (BDO, 2013, 2015). I find similar effects of hospital size on BDO (sum) scores (Results not shown). This can explain that the published rankings of BDO are superior for the larger hospitals that are over represented by Top clinical ones. Indeed, the average ranking of reported TCH in 2012 is 33,1, as compared to 43.8 for the GH, and in 2013, 44 as compared to 32.8 respectively. For 2013 this difference is significant (t=2.09, p<0.05, df =77), and just below significance level (0.05) in 2012 (t=1.98, df = 77).
Profit margin 2014
0 = GH Number of beds 0.000**
1 = TCH Number of beds 0.000**
Rsquare 0.246
Table 6.3. Model summary of the effect of the number of beds on profit margin in 2014 of GH (0) and TCH (1) sepa-rately with number of beds as independent and the profit margin as dependent variable. Note: **p<0.01
As there appears to be an interaction between the variables “type of hospital” and number of beds, I perform a regression on all hospitals (TCH and GH) that includes the interaction parameter. Now I do find a significant contribution of both the type of hospital and the interaction variable (and not the number of beds), although the variable explained in this model is small as compared to the one shown in 7.2. Again, this is both visible with the profit margin and with the sum-scores as dependent variable. (Table 6.4). None of the other domains (as dependent variable) shows any significant associations.
Profit Margin 2014 Total BDO sum score 2014 Type of hospital 1.207** 1.230**
Number of beds 0.275 0.408
Beds x type hospital (interaction) -1.207** -1.320**
Rsquare 0.133 0.147
Table 6.4 Model summary for regressions of “number of beds”, “type of hospital” and the interaction parameter explaining the Profit margin and the total BDO sum-score in 2014. (Note: Type of Hospital is coded as 0= GH and 1=TCH; **p<0.01)
26
7. Application of benchmark ratios and reference percentiles to hospitals on
Aruba, Bonaire and Curacao.
In appendix IV, I provide the social economic and healthcare background of Aruba, Bonaire and Cura-cao. I show that differences in demographic profiles and mortality patters between The Netherlands and the Caribbean islands, which may affect healthcare costs and possibly bias the financial perfor-mance, are negligible and should not impair any cross-continental comparisons. In this chapter I will present the results of calculated hospital ratios of the ABC islands, and provide some preliminary ob-servations. The Dutch percentiles are, as explained, based on all general hospitals (excluding the TCH) for the year 2014, the year the Dutch government did not carry out transition subsidies. The Dutch reference values are summarised in Appendix II.
7.1 Financial performance of ABC hospitals between 2010 and 2014
The figures in this chapter show the calculated ratios for Aruba in blue, Bonaire in red and Curacao in green. For the Dutch general hospitals (48) the 10th and 90th percentile (dotted lines) and the median (continuous line) are given in orange. The 10th and 90th percentile indicate the bottom and top performers of the general hospitals in the Netherlands. Below I show the results for the 3 selected ratios (for the 3 domains). The remaining ratios are shown in Appendix VI.
A) Interest- and Repayment Capacity: Profit Debt Service Coverage Ratio (DSCR)
Figure 7.1: DSCR for Aruba (blue), Bonaire (red) and Curacao (green) within the Dutch (orange) percentiles
The figure shows that, in line with the other ratios, the hospital in Bonaire has the best DSCR and that the cash flow from operations is sufficient to meet the repayment obligations in the near future. Aruba shows a marginal DSCR between the 10th percentile and the median of the Dutch general hospitals. Curacao has a bad DSCR and clearly has problems to pay its repayment obligations from the cash flow from operations.
27
B) Financial Structure: Solvency
Figure 7.2: Solvency for Aruba (blue), Bonaire (red) and Curacao (green) within the Dutch (orange) percentiles
The figure shows that Aruba and Bonaire have a similar capital structure and score between the 10th percentile and the median. In comparison with the hospitals in the Netherlands
The Solvency ratio for Curacao is omitted in the figure because the ratio shows a negative result over the years due to the negative equity.
C) Results & Performance:
Profit Margin (net income / Total revenues)
Figure 7.3: Profit Margin for Aruba (blue), Bonaire (red) and Curacao (green) within the Dutch (orange) percentiles
Bonaire’s profit margin is exceptionally high. Similarly, Aruba’s Profit margin is within Dutch (and USA, see Appendix V) bounds. In absence of the historical debt of the hospital in Curacao in this ratio, the profit margin is in normal territory, suggesting an acceptable performance in recent years.
0% 5% 10% 15% 20% 25% 30% 35% 2010 2011 2012 2013 2014
Solvency
-1% 1% 3% 5% 7% 9% 2010 2011 2012 2013 201428
8. Discussion and limitations of the research
In this thesis a first attempt has been made to generate quantitative recommendations for the financial management of hospitals on the ABC islands. The “reference percentiles” of financial ratios used in the financial industry are based on Dutch hospital data that have been processed in recent years to accompany the transition in the Netherlands to the financing of aided by market forces. The judicious selection of a representative sample of Dutch hospitals (a “peer-group”) to produce the percentiles has been my academic challenge. As a result this thesis is mainly concerned with the analysis of financial benchmarking in the Netherlands, and more specifically, the aim was to identify which parameters influence the performance measured by these financial ratios. The findings I here present, particularly those that challenge the common view that “size matters” and that scale enlargement and amalgamation of smaller hospitals should result in a more cost-efficient healthcare system, are obviously relevant in general discussions on future health care provision in developed and less developed countries. Further studies into the origin of differential financial performance of general vs top-clinical hospitals I identified seem warranted. I have also highlighted shortcomings in the Dutch ratings of individual hospitals, pointing out that sum-scores can be improved by using more continuous scale to measure assess the sub-scores. Unfortunately the data (financial statements) from the hospitals that failed in recent years is not available to me. This excluded the possibility to assess the predictive value (for success/failure) of the benchmarking.
The financial ratios I have studied here appear to serve financial institutions, banks and investors more than hospital managers in their day to day financial management. Strikingly, for the Critical access Hospitals in the USA, the peripheral smaller non-commercial hospitals have an extended list with ra-tios more tailored towards the hospital administrator (See appendix V. In these non-commercial hospi-tals cost-efficiency cannot be enforced by competition and market forces but only by well informed and motivated administrators.
29
References
BDO, Benchmark Ziekenhuizen, 2009-2015
Botje D. 2014 . Ziekenhuis Top 100 zegt weining over kwaliteit. Berenschot on-line publication,
www.consultancy.nl/nieuws/9343/ziekenhuis-top-100-lijst-zegt-weinig-over-kwaliteit
Caldwell, J., 2012, A framework for board oversight of enterprise risk, Chartered Accountants of Canada, Healthcare Management Forum. Fall:145-149
Casey, M., Klingner, J., 2004. CAH survey National Data. Flex publication No 2, University of North Carolina. http://www.flexmonitoring.org/publications/bp2/
Curtis, P., Roupas, T.A., 2009, Health Care Finance, the Performance of Public Hospitals and Financial Statement Analysis, European Research Studies, volume 7(4)
Ernst & Young, Barometer Nederlandse Gezondheidszorg, 2012-2015
Ernst & Young, Internal Learning Environment, Web Based Learning’s
Deerenberg, I., Melser, C., van Leeuwen, N., 2009. Most people in the Netherlands live within 5 km from a hospital facility. CBS publication. https://www.cbs.nl/en-gb/news/2009/33/most-people-in-the-netherlands-live-within-5-km-from-a-hospital-facility
Financial Statements Dutch Hospitals: https://www.jaarverslagenzorg.nl/
Flex Monitoring Team, 2005, Select Performance Dimensions for Critical Access Hospitals, Team Briefing Paper No.7, University of Minnesota North Caroline at Chapel Hill
HAH, Financial climate for hospitals in Hawaii, On-line publication of the Healthcare Association of Hawaii. http://hah.org/wp-content/uploads/2013/12/Financial-Climate-for-Hospitals-in-Hawaii.pdf
HRSA website: www.hrsa.gov/hospitals/criticalaccesshospitals/index.html
IAA, 2013, Stress testing and scenario analysis, Publication of the International actuarial Association, Ontario. www.actuaries.org/CTTEES_SOLV/Documents/StressTestingPaper.pdf
Integrated Risk Management for Healthcare Organizations, Risk Resource Guide October 2014, HIROC
UNICEF, 2013, The situation of children and adolescents in Curacao, UNICEF publication
Monitor publication, Facing the future: smaller acute providers, Monitor, June 2014. Publication code: IRRES 05/14. Wellington House, London, www.gov.uk/monitor
Nederlandse Zorgautoriteit, Voorlopige vaststelling transitiebedragen medisch specialistische zorg, April 2014
OECD, 2014, OECD statistics, Retrieved from database: http://stats.oecd.org/
Pink, G.H., Holmes, G.M., D'Alpe, C., Strunk, L.A., McGee, P., Slifkin, R.T., 2005, Financial indicators for Critical Acces Hospitals, Briefing paper No 7. University of North Carolina.
30
Pink, G.H., Holmes, G.M., Thompson, R., Slifkin, R., 2007, Variations in Financial Performance Among Peer Groups of Critical Access Hospitals, National Rural Health Association 299 - 305
Rapport Beperkt Zicht van het AMC/ UvA, 2012, Onderzoek naar de betrouwbaarheid, validiteit en bruikbaarheid van prestatie-indicatoren over de kwaliteit van de Nederlandse ziekenhuiszorg,
http://www.nfu.nl/
Spendolini, M.J., 2002, The Benchmarking Book. Amacom, New York
Watkins, A., 2000, Hospital Financial ratio classification patterns revisited: Upon considering non financial information, Vol. 19, 73-95.
Wegwijs 2014. Wat betekent het als een ziekenhuis zich topklinisch noemt.
https://www.wegwijs.nl/artikel/2014/12/wat-betekent-het-als-een-ziekenhuis-zich-topklinisch-noemt
World Bank, 2013, Financial development and structure dataset November 2013
World Health Organization, https://www.mindbank.info/collection/country/netherlands/
31
Appendix I: Dutch General Hospitals included in analysis
Admiraal de Ruyter Ziekenhuis
Pantein
Algemeen Ziekenhuis Westfries Gasthuis
R.K. Ziekenhuis St. Franciscus
Antonius Ziekenhuis
Rijnland Zorggroep
Bernhoven
Rivas Zorggroep
Bovenij Ziekenhuis
Rode Kruis Ziekenhuis
Bronovo-NEBO
Saxenburgh
St. Jansdal
Slingeland Ziekenhuis
CuraMare
Slotervaart
Diaconessenhuis Leiden
St. Anna Zorggroep
Diakonessenhuis
Stichting Het Nederlands Kanker Instituut
Elkerliek Ziekenhuis
Protestants Christelijk Ziekenhuis Ikazia
Flevoziekenhuis
St. Jans Gasthuis
Gemini Ziekenhuis
Streekziekenhuis Koningin Beatrix
Groene Hart Ziekenhuis
t Lange Land Ziekenhuis
Havenziekenhuis
Tergooi
IJsselland Ziekenhuis
Tweesteden Ziekenhuis
Ijsselmeerziekenhuizen
Waterlandziekenhuis
Rivierenland
Wilhelmina Ziekenhuis Assen
Laurentius
Zaans Medisch Centrum
Lievensberg Ziekenhuis
Ziekenhuis Amstelland
Nij Smellinghe
Ziekenhuis Gelderse Vallei
Ommelander Ziekenhuis Groep
Ziekenhuisgroep Twente
Orbis Medisch en Zorgconcern
Zorgcombinatie Noordenboog
32
Appendix II:
EY Descriptive statistics General Hospitals in the Netherlands
2012 DSCR ICR Net Debt/EBITDA Current ratio Solvability Profit Margin ROIC Rentability Personnel %
Count 48 48 48 48 48 48 48 48 48 Minimum -0,06 -3,64 -38,39 0,09 -5,8% -12,1% -18,8% -9,7% 6% Maximum 29,94 24,04 11,70 2,07 46,0% 10,2% 46,1% 17,1% 70% Average 3,00 2,20 3,08 0,94 16,7% 1,9% 7,3% 3,6% 56% Median 2,07 1,81 3,86 0,90 17,0% 1,9% 6,1% 3,5% 56% ST dev 4,37 3,80 6,55 0,36 8,9% 3,1% 8,7% 3,4% 9% Percentile 10 1,23 -0,15 1,25 0,61 7,5% 0,27% 3% 1,7% 50% Percentile 25 1,57 1,34 2,55 0,73 12,1% 1,12% 4% 2,6% 54% Percentile 50 2,07 1,81 3,86 0,90 17,0% 1,95% 6% 3,5% 56% Percentile 75 2,40 2,24 5,20 1,04 20,2% 2,82% 8% 4,8% 58% Percentile 90 4,97 3,02 6,55 1,42 27,5% 4,24% 12% 6,3% 62%
Interest- and Repayment Capacity Financial Structure & Position Results & Performance
2013 DSCR ICR Net Debt/EBITDA Current ratio Solvability Profit Margin ROIC Rentability Personnel %
Count 48 48 48 48 48 48 48 48 48 Minimum 0,76 -8,55 -1,49 0,21 -5,2% -6,7% -13,9% -6,9% 48,9% Maximum 20,69 19,04 9,20 2,29 47,5% 14,9% 34,0% 22,3% 72,0% Average 2,88 2,16 3,76 1,01 19,2% 2,2% 7,5% 4,1% 55,3% Median 2,23 2,01 3,60 0,93 18,6% 2,5% 7,2% 4,2% 54,2% ST dev 2,99 3,75 2,46 0,42 9,9% 3,4% 8,3% 4,2% 4,4% Percentile 10 1,25 -1,04 1,28 0,56 6,6% -2,9% -3,0% -1,5% 51,8% Percentile 25 1,66 1,36 2,17 0,73 14,7% 0,8% 5,1% 2,6% 53,1% Percentile 50 2,23 2,01 3,60 0,93 18,6% 2,5% 7,2% 4,2% 54,2% Percentile 75 3,15 2,97 5,21 1,21 25,1% 3,8% 9,5% 5,8% 56,1% Percentile 90 4,01 4,48 7,63 1,59 30,4% 5,1% 15,3% 7,2% 60,2%
Interest- and Repayment Capacity Financial Structure & Position Results & Performance
2014 DSCR ICR Net Debt/EBITDA Current ratio Solvability Profit Margin ROIC Rentability Personnel %
Count 44 44 44 44 44 44 44 44 44 Minimum -1,02 -7,67 -10,05 0,24 -4,3% -8,1% -44,0% -9,9% 47,7% Maximum 8,73 20,19 7,76 2,23 45,6% 5,9% 23,4% 7,9% 69,7% Average 2,09 1,95 3,10 1,07 20,4% 1,6% 5,5% 3,4% 56,2% Median 2,06 1,67 3,20 1,06 21,3% 2,2% 6,8% 4,0% 55,9% ST dev 1,32 3,50 2,90 0,44 11,0% 2,4% 8,9% 3,0% 4,8% Percentile 10 1,14 0,17 0,02 0,55 5,0% -0,6% 1,6% 1,0% 51,3% Percentile 25 1,57 1,29 2,34 0,71 14,2% 0,7% 4,1% 2,4% 53,2% Percentile 50 2,06 1,67 3,20 1,06 21,3% 2,2% 6,8% 4,0% 55,9% Percentile 75 2,39 2,50 4,53 1,27 26,5% 2,7% 8,5% 4,9% 57,9% Percentile 90 2,74 3,12 5,97 1,66 31,7% 4,0% 10,4% 6,0% 63,0%
