The impact of no-fault compensation on health care expenditures: an empirical study of OECD countries.

G Model HEAP-3284; No. of Pages 8

ARTICLE IN PRESS Health Policy xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Health Policy journal homepage: www.elsevier.com/locate/healthpol

The impact of no-fault compensation on health care expenditures: An empirical study of OECD countries Tom Vandersteegen ∗ , Wim Marneffe, Irina Cleemput, Lode Vereeck Hasselt University, Faculty of Applied Economics, Martelarenlaan 42, BE-3500 Hasselt, Belgium

a r t i c l e

i n f o

Article history: Received 16 April 2014 Received in revised form 15 September 2014 Accepted 18 September 2014 Keywords: Health care expenditures Medical liability system No-fault compensation Defensive medicine OECD countries

a b s t r a c t Around the world, governments are faced with spiralling health care expenditures. This raises the need for further insight in the determinants of these expenditures. Existing literature focuses primarily on income, ageing, health care financing and supply variables. This paper includes medical malpractice system characteristics as determinants of health spending in OECD countries. Estimates from our regression models suggest that no-fault schemes for medical injuries with decoupling of deterrence and compensation reduce health expenditures per capita by 0.11%. Furthermore, countries that introduced a nofault system without decoupling of deterrence and compensation are found to have higher (+0.06%) health care spending. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Over the last decades, policy makers have increasingly focused on controlling spiralling health care expenditures.1 On average, health care expenditures as a proportion of Gross Domestic Product (GDP) have risen from 7.8% in 2000 to 9.3% in 2011 in OECD countries [1]. This increase was the highest in the United States (+4.0%) and the Netherlands (+4.0%). On the contrary, in Luxembourg (−0.8%) and Iceland (−0.5%) health spending has slightly declined over the past decade. These developments have incited many scholars to examine the determinants of health care expenditures. Some scholars state that a country’s medical malpractice system could be a determinant of health care spending.

∗ Corresponding author. Tel.: +32 11 26 87 57. E-mail addresses: [email protected] (T. Vandersteegen), [email protected] (W. Marneffe), [email protected] (I. Cleemput), [email protected] (L. Vereeck). 1 In this paper health (care) expenditures and health (care) spending are used as synonyms.

In 2002, the US Department of Health and Human Services stated that the medical liability system imposes large costs on the US health care system [2]. Kessler and McClellan [3] estimated that health care costs could be reduced by 5–9% by limiting unreasonable awards for non-economic damages, such as pain and suffering, without substantially affecting the quality of care. Hellinger and Encinosa [4] found that health care spending was statistically 3–4% less in US states capping non-economic damage awards in malpractice cases. Therefore, it is possible that differences in medical malpractice systems across countries do significantly affect health spending. The objective of this article is to assess the impact of a no-fault compensation system on health care expenditures. In addition to the problem of increasing health care costs, another policy issue in many OECD countries is the reform of their medical liability system. A medical malpractice system has two main purposes: compensating patients suffering damages due to a health care provider’s negligence and inciting health care providers to take appropriate precautions during medical treatments [5]. Until a few decades ago, injuries resulting from health interventions

http://dx.doi.org/10.1016/j.healthpol.2014.09.010 0168-8510/© 2014 Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Vandersteegen T, et al. The impact of no-fault compensation on health care expenditures: An empirical study of OECD countries. Health Policy (2014), http://dx.doi.org/10.1016/j.healthpol.2014.09.010


2

ARTICLE IN PRESS T. Vandersteegen et al. / Health Policy xxx (2014) xxx–xxx

in OECD countries were settled in court applying the conventional tort rules. Under these negligence-based rules, a patient carries the burden of proving a physician’s fault, harm suffered and causation between both to receive compensation. However, the lack of a specific compensation system for medical malpractice posed problems for patients as well as for physicians. Physicians practiced defensive medicine, which the Office of Technology Assessment describes as the ordering of additional tests, extra procedures and visits, or the avoidance of certain procedures or patients, due to concern about malpractice liability risk [6]. Mello et al. [7], moreover, estimated these health care costs at approximately $45.59 billion in 2008 dollars, or nearly 2% of total health care spending. Patients struggled with the burden of proving fault, damage and causation, making the outcome of medical malpractice trials unpredictable and uncertain. Some insurers restricted their coverage package, others no longer offered liability insurances to the most severe risk categories or exited the medical malpractice insurance market altogether. Specific regulation for medical malpractice cases was therefore implemented in several countries. The United States, for instance, established statutory tort reforms on the state level. Several states adopted caps on economic and non-economic damages, limited joint and several liability, put caps on attorneys’ fees and limited collateral source rules to limit malpractice premiums and awards [8]. New Zealand and Sweden introduced a publicly financed scheme for compensating medical injuries, respectively in 1974 and 1975 [9]. The main driver behind this shift was that the conventional tort system’s cumbersome nature obstructed patients’ access to due compensation in Sweden [10]. New Zealand’s compensation system originally arose as a consequence of the workers’ compensation reforms, not in response to concerns about medical malpractice [11]. An important feature in both countries was the abandonment of their negligence-based compensation system. Instead, an outof-court procedure was introduced, applying a no-fault rule. In Sweden, compensation is awarded to patients if the harm suffered could have been avoided under optimal circumstances (avoidability rule). Eligibility criteria in New Zealand have been revised several times. From 1974 to 1992, ‘personal injuries’ included medical, dental, surgical and first aid misadventures, and were compensated on a no-fault basis. From 1992 to 2005, ‘medical misadventures’, i.e. injuries resulting from a medical error or a medical mishap, were eligible for compensation [12]. Since 2005, eligibility is extended to all ‘treatment injuries’ regardless of injury severity or rarity, or error [13]. Damages are paid by a pool of insurers, while the discipline of medical providers is handled by another independent institution. As a result, deterrence and compensation are decoupled, possibly lowering the practice of defensive medicine. It took a decade before the other Nordic countries implemented a modified medical malpractice system. Finland (1987) and Denmark (1992) established a privately financed insurance-based compensation system [9]. Eligibility of compensation is determined by the avoidability

rule, though damages can also be paid for unavoidable medical injuries if they are unusual or serious (endurability rule). Meanwhile, the Norwegian government initiated the public financing of a non-statutory no-fault scheme for medical injuries in 1988 [14], though negligence remained their main criterion for compensation. Eventually, also Iceland abandoned their tort system for medical malpractice in 2001 and instead implemented an insurancebased no-fault scheme, also applying the avoidability rule.2 France (2002) and Belgium (2010) initiated a more restricted no-fault scheme than the Nordic countries [15,16]: only in case of the absence of negligence, a nofault rule is applied to determine a patient’s eligibility for compensation. Some countries, such as New Zealand and Iceland, moreover abolished the option of going directly to court in case of a medical injury. Already in the 1970s, the English government acknowledged the shortcomings of their clinical negligence litigation system, though its replacement by a no-fault scheme has been rejected ever since. To date, also the NHS Redress Act 2006, adopting a compensation scheme without recourse to civil proceedings, has not been issued yet. Nonetheless, so-called pre-action protocols have been introduced to resolve clinical disputes without resort to legal action.3 On the other hand, the Welsh NHS Redress scheme has been issued in 2011,4 while the Scottish government currently explores the implementation of a no-fault system for medical injuries [17]. Using cross-country OECD data for the period 1970–2011, the effect of no-fault compensation on health care expenditures is estimated. In addition to the common determinants of health care expenditures, we added a variable accounting for the presence of a no-fault compensation system. This is in line with Gerdtham and Jönsson [18], who pointed out the need for testing new variables of health spending. The paper is organized as follows. The following section presents a brief overview of the determinants of health care expenditures in existing literature and discusses the data and the regression model. Section 3 provides the empirical analysis and the results, while Section 4 discusses the impact of malpractice systems on defensive medicine. Finally, Section 5 draws some conclusions. 2. Methods 2.1. Determinants of health care expenditure in literature The inclusion of other determinants of health care expenditures besides the presence of a no-fault compensation system was based on a review of the literature. We selected papers using the search terms “health care

2 Act on Patient Insurance, No. 111/2000, Iceland. Available at: http:// eng.velferdarraduneyti.is/media/acrobat-enskar sidur/Act on Patient Insurance as amended.pdf. 3 Pre-Action Protocol for the Resolution of Clinical Disputes. Available at: http://www.justice.gov.uk. 4 The NHS Concerns, Complaints and Redress Arrangements Wales Regulations 2011. Available at: http://www.wales.nhs.uk.




expenditures”, “health care costs”, “health care spending” and “determinants” in the “Thomson Reuters Web of Knowledge” database. We limited our scope to studies comparing health spending in OECD countries or regions. Appendix A [53–60] provides an overview of selected multivariate regression analyses of health care expenditure levels in OECD countries or regions. Generally, per capita health care expenditure has been adopted as the dependent variable. Income has been identified as one of the key determinants of health care expenditures and is mostly represented by per capita gross domestic product (GDP). Theoretically, a higher income will yield higher health care spending, which is also empirically confirmed. Twenty out of the 25 studies examined included a determinant accounting for the age structure of a population. According to Gerdtham et al. [19], it is generally accepted that a country’s or region’s age distribution affects the demand for health care services. An older population has a relatively higher propensity to consume health care [20]. In most of these studies, the proportion of the population over 65 or 75 years is applied and indeed found to have a positive influence on health expenditures. Rather than assuming only aged people affect health care expenditures, some authors also add the share of young people to their regression model [21–25]. Nonetheless, previous research revealed a relatively modest impact of age distribution compared to the income effect. Structural characteristics of the health care supply, like the number of physicians and hospital beds per capita, are included in 15 out of the 25 studies reviewed. According to the supply-induced demand for health care hypothesis (SID), health expenditures will increase as the ratio of the number of physicians to the population expands [26]. Most regression analyses for OECD countries establish this relationship to be statistically significant. Leu [21] presumed the share of public health financing to positively affect the level of total health care expenditures. Though, Culyer [27] attributed a more important role to the degree of open-endedness of financing, i.e. the lack of budget restrictions, than to financing distribution. Gerdtham et al. [28] concluded that the impact of the share of public health care financing cannot be determined a priori. Ten out of 25 studies accounted for a public health care financing variable in their analysis, indeed finding both (significant) positive as negative effects on total health spending. Although relatively few studies (7 out of 25) have explicitly accounted for medical progress, technological change is pointed out to account for the bulk of the increase in health care expenditures [29]. The absence of a similar variable in many regression analyses may be due by the fact that there is still a debate on what is an appropriate indicator of medical progress. Some authors suggest R&D spending in health care [30], the provision of high-technology services such as magnetic resonance imaging (MRI) [31], patent applications [32], life expectancy or infant mortality [33] as a proxy for technological change in health care. In the absence of a suitable measure and given that technological change occurs over time, OECD studies accounting for medical technology generally included a time trend or a set of time fixed effects. However, time effects represent

3

an upper bound estimation of technological impact, as in reality it will also account for other non-observed trended effects like policy shifts or changes in preferences [34]. Nonetheless, the effects of medical progress on health care expenditures are rather ambiguous. On the one hand, new techniques may generate less expensive ways of health outcomes resulting in lower health expenditures. On the other hand, expensive innovative treatments may boost the cost of health services [35]. Less frequently adopted determinants of health care expenditures are urbanization [19,21,23,28], gatekeeping [36–38], population density [4,22], the price of health care [34,39] and fee-for-service payment systems [19,28]. Furthermore, Gerdtham et al. [37] also included a number of institutional determinants in their analysis. 2.2. Data and model specification Our key exogenous variable of interest is whether a country has a no-fault compensation system for medical malpractice in a given year or not (NO FAULT). As endogenous variable, per capita health care expenditure (PCHCE) is used. Taking into account the availability and quality of data for the OECD countries and the findings from our literature overview, we controlled for per capita Gross Domestic Product (PCGDP), population age structure (proportion of the population 65 years of age (POP65)), number of physicians per 1000 inhabitants (PHYS) and proportion of total health care expenditures that is publicly financed (PUBFIN), which also accounts for (a part of) a country’s socio-political environment. Considering the lack of an appropriate measure for technological change, we include a nonlinear time trend (TREND). This proxy also accounts for other time-related macroeconomic effects. We use unbalanced panel data for all 34 OECD countries5 for the period 1970–2011, which yields 759 observations. The list of variables, definitions, and data sources are shown in Table 1 and the basic descriptive statistics are presented in Table 2. Our empirical models are specified in the following way: lnPCHCEit = ˛ + ˇ1 lnPCGDPit + ˇ2 POP65it + ˇ3 POP15it + ˇ4 PHYSit + ˇ5 PUBFINit + ˇ6 TREND + ˇ7 NO FAULTit + εi + uit

(1)

lnPCHCEit = ˛ + ˇ1 lnPCGDPit + ˇ2 POP65it + ˇ3 POP15it + ˇ4 PHYSit + ˇ5 PUBFINit + ˇ6 TREND + ˇ7 NO FAULTit + ˇ8 DECOUPit + εi + uit

(2)

5 Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea Republic, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, United Kingdom and United States.


G Model

ARTICLE IN PRESS

HEAP-3284; No. of Pages 8

T. Vandersteegen et al. / Health Policy xxx (2014) xxx–xxx

4 Table 1 Variable definitions and data sources. Variable

Definition

Data source

PCHCE PCGDP POP65 POP15 PHYS PUBFIN TREND NO FAULT

Per capita health care expenditure; in US$ (2005 PPP) Per capita Gross Domestic Product (GDP); in US$ (2005 PPP) Percentage of population above 65 years Percentage of population below 15 years Number of physicians per 1000 population Proportion of total health spending financed by public expenditures Nonlinear time trend Dummy variable; is coded 1 for observations corresponding to countries with a no-fault medical malpractice system (Belgium, Denmark, Finland, France, Iceland, New Zealand, Norway and Sweden) and 0 otherwise Dummy variable; is coded 1 for observations corresponding to countries with a no-fault medical malpractice system and no decoupling of compensation and deterrence (Belgium and France) and 0 otherwise Dummy variable; is coded 1 for observations corresponding to countries with a no-fault medical malpractice system and decoupling of compensation and deterrence (Denmark, Finland, Iceland, New Zealand, Norway and Sweden) and 0 otherwise Dummy variable; is coded 1 for observations corresponding to countries with a mainly privately financed no-fault medical malpractice system and decoupling of compensation and deterrence (Denmark and Finland) and 0 otherwise Dummy variable; is coded 1 for observations corresponding to countries with a mainly publicly financed no-fault medical malpractice system and decoupling of compensation and deterrence (Iceland, New Zealand, Norway and Sweden) and 0 otherwise

OECD.stat OECD.stat OECD.stat OECD.stat World Bank OECD.stat

NO DECOUP

DECOUP

DECOUP PRIV

DECOUP PUB

Own research

Own research

Own research

Own research

Own research

Table 2 Descriptive statistics (N = 759). Variable

Mean

Std. Dev.

Min

Max

PCHCE PCGDP POP65 POP15 PHYS PUBFIN

1865.023 23591.84 12.92316 21.12547 2.399039 71.95644

1066.228 9634.025 3.501311 4.929513 0.8999911 14.52512

108.2694 5129.571 3.868939 13.05635 0.5 21.9391

7430.349 73912.59 23.02412 39.49369 6.167 98.2861

lnPCHCEit = ˛ + ˇ1 lnPCGDPit + ˇ2 POP65it + ˇ3 POP15it + ˇ4 PHYSit + ˇ5 PUBFINit + ˇ6 TREND + ˇ7 NO DECOUPit + ˇ8 DECOUP PRIVit + ˇ9 DECOUP PUBit + εi + uit

(3)

where i refers to the OECD countries (n = 1,. . .,34) and t refers to time (t = 1970,. . .,2011). Different characteristics of the no-fault compensation systems are used in each estimate to account for differences between these schemes. Decoupling of deterrence and compensation (DECOUP) refers to a medical malpractice system in which physicians are not personally burdened with the payment of damages. Since the remaining variables are expressed in percentages, logarithms are only used in the case of health expenditure and income. 3. Results To estimate Eq. (1) for all 34 OECD countries from 1970 to 2011, we performed Generalized Least Squares (GLS) regressions. Since our key variable of interest, NO FAULT, does not vary over time for most of the OECD countries, we used random effects to estimate our model.

We also ran fixed effects analyses including only the eight countries that have adopted a no-fault system for medical malpractice, though these models yielded insufficient evidence of differences in health care spending levels. Autocorrelation was detected by the Wooldridge test and corrected for by the cluster-option.6 According to the Breusch–Pagan/Cook–Weisberg test, heteroskedasticity was not an issue.7 Multicollinearity also proved not to be an issue. Furthermore, we tested for reverse causality as the variable NO FAULT might be endogenous. High levels of health care expenditures might induce countries to adapt their medical liability system. It appears that mainly countries with a Nordic legal origin adopted a no-fault compensation scheme for medical injuries. Moreover, the kind of legal origin does not affect health care spending levels. Therefore, a country’s legal origin might be a suitable instrument to test for the presence of endogeneity. We used the legal origin classification of La Porta et al. [40]. The results of the 2SLS method indicate legal origin is a suitable instrumental

6 7

H0 : No autocorrelation; Prob > F = 0.0000. H0 : Constant variance; Prob > chi2 = 0.1177.




variable.8 Moreover, the Durbin–Wu–Hausman test does not support the presence of reverse causality.9 Table 3 shows that almost all coefficients of our control variables have the expected sign and are statistically significant. However, both age structure variables are not significant, indicating that in our model a country’s population age structure does not affect health care expenditures. The proportion of the population over 80 years of age was also tested, though yielded similar results. People mainly use health care services at the end of their lives, regardless of a person’s stage of life. Elderly, moreover, live longer in good health. Hence, no significant effect on health spending can be determined. A similar result was found by Murthy and Ukpolo [20], Barros [36] and Roberts [34]. The effect of the share of public health expenditures is rather limited and only increases health care spending by 0.003%. On the other hand, an increase in the number of physicians raises health expenditures by 0.7%. As opposed to Gerdtham et al. [28], we measure an income elasticity of 1.1, which is rather close to unity. This finding is in line with Hitiris and Posnett [41] who reported similar results. Life expectancy and infant mortality were both tested as a proxy for medical progress, as suggested by Dreger and Reimers [33]. However, these variables did not yield any significant effect or improvement of the regression model. Instead, we included a nonlinear time trend as recommended by Di Matteo [35], which also accounts for other unobserved macroeconomic effects. Estimates from our models indicate that these time effects significantly raise health care expenditures. Furthermore, an important result is the seemingly insignificant effect of the variable NO FAULT in our basic model in column 1. This would imply that countries with a no-fault compensation system for medical injuries do not have statistically different levels of health care spending compared to countries without a no-fault scheme in a given year. However, this would wrongfully assume that the liability systems in those countries are identical. Therefore, in the second column we differentiate between the nofault countries. On the one hand, in Belgium and France the compensation systems still focus on blaming the physician in case of a medical injury. This is even aggravated by the implementation of a free administrative procedure to apply for compensation. On the other hand, the decoupling of deterrence and compensation is an important feature of the schemes in the Nordic countries and New Zealand. This means that physicians involved in a medical malpractice case are not personally burdened with the payment of damages. Column 2 shows that countries with decoupling of deterrence and compensation in their no-fault system have lower (−0.06%) health care spending than all other countries. This is probably the result of a decrease in the practice of defensive medicine, which is in line with Kessler and McClellan [3]. They found that malpractice reforms

8 F test of excluded instruments: F(1,33) = 15.74; underidentification test (Kleibergen–Paap rk LM statistic): Chi-sq(1) P-val = 0.0382; Weak identification test (Kleibergen–Paap rk Wald statistic): 15.738. 9 H0 : Regressor is exogenous; Prob > chi2 = 0.07464.

5

that directly reduce provider liability lead to reductions of 5–9% in medical expenditures for all elderly Medicare beneficiaries treated for serious heart disease in the US. Also based on US data, Hellinger and Encinosa [4] concluded that laws limiting malpractice payments reduce state health care expenditures by approximately 3–4%. Baicker et al. [42], Thomas et al. [43] and Smith-Bindman et al. [44] found similar results, while Beider and Hagen [45], Baicker and Chandra [46], and Sloan and Shadle [47] found no evidence of tort reforms reducing medical spending. In our model, no-fault systems that reduce physician liability only lower health spending by 0.06%, which is rather limited compared to existing literature. In contrast, estimates from our model suggest that no-fault systems without decoupling of deterrence and compensation have statistically higher (+0.06%) levels of health care spending, compared to all other countries. Analogously, a possible explanation might be the increasing exposure to liability for physicians, leading to more defensive medicine and thus higher levels of health expenditures. An increase of 0.06% is, however, rather modest. Since Belgium and France only recently introduced their no-fault compensation system, these findings should be interpreted with caution. Finally, in the third column of Table 3 we separate the no-fault systems that decouple deterrence and compensation according to their financing structure. The schemes in Denmark and Finland are mainly privately financed, while those in Sweden, New Zealand, Norway and Iceland are mainly publicly financed. We find that no-fault countries with privately as well as publicly financed structures have lower health spending than all other countries. While the former relationship is significant, the latter is not. After the reform, compensation payments in Denmark and Finland are paid from an insurance pool and liability premium levels are set by the government. As a result, physicians are no longer personally accountable for the payment of damages. Liability exposure has decreased, leading to less defensive medicine and thus reduced health spending. These privately financed structures appear to lower health care spending by 0.11%. Thus, the overall impact of no-fault systems on health care costs is rather modest, which was also found by Thomas et al. [43] concerning the impact of tort reforms. Nonetheless, these outcomes raise the need for further research. 4. Discussion One should bear in mind that no-fault systems for medical malpractice are primarily introduced to facilitate access to compensation for patients. We believe that these systems might also affect physicians’ clinical decision making, although conclusive evidence is lacking. To our knowledge, the only existing study on a no-fault system affecting defensive practices is provided by Edwards [48], who found at most mild evidence that Virginia’s Birth Injury Program, i.e. a no-fault scheme for birth-related neurological injuries, reduces obstetricians’ defensive behaviour. The replacement of malpractice litigation for medical injuries by a no-fault compensation system, however,


G Model

ARTICLE IN PRESS

HEAP-3284; No. of Pages 8

T. Vandersteegen et al. / Health Policy xxx (2014) xxx–xxx

6

Table 3 GLS regression results for health care expenditures in 34 OECD countries for the period 1970–2011 (N = 759).

NO FAULT NO DECOUP DECOUP DECOUP PRIV DECOUP PUB lnPCGDP POP65 POP15 PHYS PUBFIN TREND Constant Within R2 Between R2 Overall R2

(1)

(2)

(3)

−0.0373923 (0.0324898) – – – – 1.092795*** (0.0840761) 0.0043915 (0.0105825) −0.0120519 (0.0136703) 0.0676811*** (0.0254997) 0.0025911* (0.0015675) 0.2156921*** (0.0685156) −4.495946*** (1.020332) 0.9311 0.8684 0.8926

– 0.0642403** (0.0304019) −0.0626132** (0.0318745) – – 1.095709*** (0.0849842) 0.0031646 (0.0103882) −0.0121685 (0.0136183) 0.0697046*** (0.0254054) 0.0025377 (0.0015734) 0.2175754*** (0.0680451) −4.512842*** (1.025715) 0.9319 0.8677 0.8924

– 0.0643195** (0.0305008) – −0.1052055*** (0.0341522) −0.0481781 (0.0295104) 1.096422*** (0.0853196) 0.0031945 (0.0103902) −0.0118451 (0.0138686) 0.0692794*** (0.0254501) 0.0026144* (0.0015804) 0.2185501*** (0.0688088) −4.533993*** (1.040512) 0.9321 0.8668 0.8920

Standard errors appear in parenthesis. (–) Variable not included. The bold values indicate the variables that are statistically significant (at any level). * Significant at 10%. ** Significant at 5%. *** Significant at 1%.

does not necessarily imply that physicians practice in a void of accountability. For example, in New Zealand the no-fault compensation system (outcome-based) is supplemented with a medical professional accountability system (process-based). Physicians may be referred to the Medical Council, the Health and Disability Commissioner (HDC) system and the Health Practitioners’ Disciplinary Tribunal for possible disciplinary actions [13]. Although New Zealand’s regulatory system was intended to be non-punitive, evidence exists that physicians fear the patient complaints system. Cunningham and Dovey [49] found that doctors altered their clinical behaviour in the direction of defensive medicine due to the patient complaint process. After the 2005 reforms, the same authors concluded that the New Zealand complaints process has the potential to improve health care quality, though it also may cause doctors to practice defensively [50]. On the other hand, Wallis [51] claims that disciplinary proceedings against New Zealand doctors or even a performance review, as a result of the patient complaints process, are very unlikely. Paterson [52] even stated that doctors practice unnecessarily defensive, given the statistical odds of disciplinary actions. Nonetheless, physicians’ defensive practices may not only be impacted by the compensation system, though also by the (co)existence of patient claims systems. Further research could focus on the impact of both compensation systems and medical professional accountability processes. Moreover, it cannot be excluded that the differences found are attributable to socio-political features, which implies more economical health care systems for New Zealand and the Nordic countries compared to other countries. However, a country’s socio-political environment is generally rather stable throughout time, while our dummy variable only takes into account the years in which a country had a no-fault system for medical injuries. Thus, unless the socio-political environment in each country would have changed in the same year of introducing a no-fault system for medical injuries, our no-fault

(dummy) variable should not be interpreted as reflecting a country’s socio-political system. 5. Conclusions Spiralling health care expenditures have led to numerous studies on the determinants of health care expenditures in OECD countries. Mainly focusing on income, ageing, health care financing and supply variables, they concluded income to have the largest impact on health spending. In this paper, we included the medical liability system as an explanatory variable. The rationale behind this is the effect of malpractice system reforms on defensive medicine and thus on health expenditure levels. Estimates from our regression models provide evidence that a country’s medical liability system significantly affects national health care expenditures. Countries that do not decouple deterrence and compensation in their no-fault malpractice system, like Belgium and France, have significantly higher (+0.06%) levels of health expenditures. However, since they only recently introduced their no-fault compensation system, these findings should be interpreted with caution. No-fault systems with decoupling of deterrence and compensation, like in the Nordic countries and New Zealand, lead to significantly lower (−0.06%) health spending. These outcomes might be the result of changing levels of defensive medicine due to the type of medical malpractice system being adopted. Moreover, no-fault systems with decoupling of deterrence and compensation that are mainly privately financed appear to yield lower (−0.11%) health care expenditures. Notwithstanding these findings only represent a small fraction of differences in health spending per capita, for society the amount is not trivial. The medical malpractice system as determinant of health care expenditures should, therefore, receive more attention from scholars and policy makers. In addition to the compensation system, also the role and the impact of medical professional accountability processes should be considered.




Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10. 1016/j.healthpol.2014.09.010.

[21]

[22]

References [1] OECD Health Statistics, 2013. Available at: http://www.oecd. org/els/health-systems/oecdhealthdata.htm [accessed 20.01.14]. [2] Department of Health and Human Services (HHS). Confronting the new health care crisis: improving health care quality and lowering costs by fixing our medical liability system; 2002. Available at: http://aspe.hhs.gov/daltcp/reports/litrefm.htm [accessed 16.01. 14]. [3] Kessler D, McClellan M. Do doctors practice defensive medicine? The Quarterly Journal of Economics 1996;111(2):353–90, http://dx.doi.org/10.2307/2946682. [4] Hellinger FJ, Encinosa WE. The impact of state laws limiting malpractice damage awards on health care expenditures. American Journal of Public Health 2006;96(8):1375–81, http://dx.doi.org/10.2105/AJPH.2005.077883. [5] Danzon PM. Medical malpractice: theory, evidence and public policy. Cambridge, Mass: Harvard University Press; 1985, http://dx.doi.org/10.4159/harvard.9780674331143. [6] Office of Technology Assessment (OTA), U.S Congress. Defensive medicine and medical malpractice. OTA-H-602. Washington, DC: U.S. Government Printing Office; 1994. [7] Mello MM, Chandra A, Gawande AA, Studdert DM. National costs of the medical liability system. Health Affairs 2010;29(9):1569–77, http://dx.doi.org/10.1377/hlthaff.2009.0807. [8] Bal BS. An introduction to medical malpractice in the United States. Clinical Orthopaedics and Related Research 2009;467(2):339–47, http://dx.doi.org/10.1007/s11999-008-0636-2. [9] OECD, 82 p. Policy issues in insurance no 11. Medical malpractice: prevention, insurance and coverage options. Paris: OECD Publishing; 2006, http://dx.doi.org/10.1787/9789264029057-en. [10] Kachalia AB, Mello MM, Brennan TA, Studdert DM. Beyond negligence: avoidability and medical injury compensation. Social Science & Medicine 2008;66(2):387–402, http://dx.doi.org/10.1016/j. socscimed.2007.08.020. [11] Bismark M, Paterson R. No-fault compensation in New Zealand: harmonizing injury compensation, provider accountability, and patient safety. Health Affairs 2006;25(1):278–83, http://dx.doi.org/10.1377/hlthaff.25.1.278. [12] Farrell A-M, Devaney S, Dar A. No-fault compensation schemes for medical injury: a review. In: Scottish Government Social Research interim report; 2010. [13] Wallis K. New Zealand’s 2005 ‘no-fault’ compensation reforms and medical professional accountability for harm. The New Zealand Medical Journal 2013;126(1371):33–44. [14] Jørstad RG. The Norwegian system of compensation to patients. Medicine and Law 2002;21(4):681–6. [15] G’Sell-Macrez F. Medical malpractice and compensation in France: Part I. The French Rules of Medical Liability since the Patients’ Rights Law of March 4 2002. Chicago-Kent Law Review 2011;86(3):1093–123. [16] Boone I, Lierman S, editors. Reeks Gezondheidsrecht 1. Vergoeding van slachtoffers van medische ongevallen. Praktijkgerichte analyse van de wet van 31 maart 2010. Antwerpen: Intersentia; 2011. [17] Goldberg R. Medical malpractice and compensation in the UK. Chicago-Kent Law Review 2012;87(1):131–61. [18] Gerdtham U-G, Jönsson B. International comparisons of health expenditure: theory, data and econometric analysis. In: Culyer AJ, Newhouse JP, editors. Handbook of health economics. Amsterdam: Elsevier; 2000. [19] Gerdtham U-G, Søgaard J, Jönsson B, Andersson F. A pooled cross-section analysis of the health care expenditures of the OECD countries. In: Zweifel P, Frech HE, editors. Health economics worldwide. Dordrecht: Kluwer Academic Publishers; 1992., http://dx.doi.org/10.1007/978-94-011-2392-1 14. [20] Murthy NRV, Ukpolo V. Aggregate health-care expenditure in the United States: evidence from cointegration tests. Applied

[23]

[24]

[25]

[26] [27] [28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

7

Economics 1994;26(8):797–802, http://dx.doi.org/10.1080/ 00036849400000094. Leu RE. The public–private mix and international health care costs. In: Culyer AJ, Jönsson B, editors. Public and private health services: complementarities and conflicts. Oxford: Basil Blackwell; 1986. Crivelli L, Filippini M, Mosca I. Federalism and regional health care expenditures: an empirical analysis for the Swiss cantons. Health Economics 2006;15(5):535–41, http://dx.doi.org/10.1002/hec.1072. Wang Z. The determinants of health expenditures: evidence from US state-level data. Applied Economics 2009;41(4):429–35, http://dx.doi.org/10.1080/00036840701704527. Baltagi BH, Moscone F. Health care expenditure and income in the OECD reconsidered: evidence from panel data. Economic Modelling 2010;27(4):804–11, http://dx.doi.org/10.1016/j. econmod.2009.12.001. Ang JB. The determinants of health care expenditure in Applied Economics Letters 2010;17(7):639–44, Australia. http://dx.doi.org/10.1080/00036840802297868. Newhouse JP. A model of physician pricing. Southern Economic Journal 1970;37(2):174–83, http://dx.doi.org/10.2307/1056127. Culyer AJ. Cost supplement in Europe. Health Care Financing Review (Annual Supplement) 1989;11:21–32. Gerdtham U-G, Søgaard J, Andersson F, Jönsson B. An econometric analysis of health care expenditure: a cross-section study of the OECD countries. Journal of Health Economics 1992;11(1):63–84, http://dx.doi.org/10.1016/0167-6296(92)90025-V. Newhouse JP. Medical care costs: how much welfare loss? Journal of Economic Perspectives 1992;6(3):3–21, http://dx.doi.org/10.1257/jep.6.3.3. Okunade AA, Murthy NRV. Technology as a ‘major driver’ of health care costs: a cointegration analysis of the Newhouse conjecture. Journal of Health Economics 2002;21(1):147–59, http://dx.doi.org/10.1016/S0167-6296(01)00122-9. Sen A. Is health care a luxury? New evidence from OECD data. International Journal of Health Care Finance and Economics 2005;5(2):147–64, http://dx.doi.org/10.1007/s10754-005-1866-4. Kleinknecht A, Van Montfort K, Brouwer E. The nontrivial choice between innovation indicators. Economics of Innovation and New Technology 2002;11(2):109–21, http://dx.doi.org/10.1080/10438590210899. Dreger C, Reimers HE. Health care expenditures in OECD countries: a panel unit root and cointegration analysis. International Journal of Applied Econometrics and Quantitative Studies 2005;2:5-20 [IZA Discussion Paper No. 1469]. Roberts J. Sensitivity of elasticity estimates for OECD health care spending: analysis of a dynamic heterogeneous data field. Health Economics 1999;8(5):459–72, http://dx.doi.org/10. 1002/(SICI)1099-1050(199908)8:53.0.CO;2-U. Di Matteo L. The macro determinants of health expenditure in the United States and Canada: assessing the impact of income, age distribution and time. Health Policy 2005;71(1):23–42, http://dx.doi.org/10.1016/j.healthpol.2004.05.007. Barros PP. The black box of health care expenditure growth determinants. Health Economics 1998;7(6):533–44, http://dx.doi.org/10. 1002/(SICI)1099-1050(199809)7:63.0.CO;2-B. Gerdtham U-G, Jönsson B, MacFarlan M, Oxley H. The determinants of health expenditure in the OECD countries: a pooled data analysis. Developments in Health Economics and Public Policy 1998;6:113–34, http://dx.doi.org/10.1007/978-1-4615-5681-7 6. Jönsson B, Eckerlund I. Why do different countries spend different amounts on health care? In: A disease-based comparison of health systems: what is best and at what cost? Paris: OECD; 2003, http://dx.doi.org/10.1787/9789264100053-en. Gerdtham U-G, Jönsson B. Price and quantity in international comparisons of health care expenditure. Applied 1991;23(9):1519–28, http://dx.doi.org/10.1080/ Economics 00036849100000204. La Porta R, Lopez-de-Silanes F, Shleifer A, Vishny R. The quality of government. The Journal of Law, Economics and Organization 1999;15(1):222–79, http://dx.doi.org/10.1093/jleo/15.1.222. Hitiris T, Posnett J. The determinants and effects of health expenditure in developed countries. Journal of Health 1992;11(2):173–81, http://dx.doi.org/10.1016/ Economics 0167-6296(92)90033-W.



8


[42] Baicker K, Fisher ES, Chandra A. Malpractice liability costs and the practice of medicine in the Medicare program. Health Affairs 2007;26(3):841–52, http://dx.doi.org/10.1377/hlthaff.26.3.841. [43] Thomas JW, Ziller EC, Thayer DA. Low costs of defensive medicine, small savings from tort reform. Health Affairs 2010;29(9):1578–84, http://dx.doi.org/10.1377/hlthaff.2010.0146. [44] Smith-Bindman R, McCulloch CE, Ding A, Quale C, Chu PW. Diagnostic imaging rates for head injury in the ED and states’ medical malpractice tort reforms. The American Journal of Emergency Medicine 2011;29(6):656–64, http://dx.doi.org/10.1016/j.ajem.2010.01.038. [45] Beider P, Hagen S. Limiting tort liability for medical malpractice. CBO; 2004. [46] Baicker K, Chandra A. The effect of malpractice liability on the delivery of health care. Forum for Health Economics & Policy 2005;8(1), http://dx.doi.org/10.2202/1558-9544.1010. [47] Sloan FA, Shadle JH. Is there empirical evidence for “defensive medicine”? A reassessment. Journal of Health Economics 2009;28(2):481–91, http://dx.doi.org/10.1016/j.jhealeco.2008.12. 006. [48] Edwards CT. The impact of a no-fault tort reform on physician decision-making: a look at Virginia’s Birth Injury Program. Revista Juridica Universidad de Puerto Rico 2010;80(1): 285–310. [49] Cunningham W, Dovey S. The effect on medical practice of disciplinary complaints: potentially negative for patient care. The New Zealand Medical Journal 2000;113(1121): 464–7. [50] Cunningham W, Dovey S. Defensive changes in medical practice and the complaints process: a qualitative study of New Zealand doctors. The New Zealand Medical Journal 2006;119(1244).

[51] Wallis K. ‘Woe is me!’: New Zealand’s non-punitive regulatory environment. Journal of Primary Health Care 2012;4(1):73–6. [52] Paterson R. The bogeyman of defensive medicine. New Zealand Doctor 2006:14–7. [53] Di Matteo L, Di Matteo R. Evidence on the determinants of Canadian provincial government health expenditures: 1965–1991. Journal of Health Economics 1998;17(2):211–28, http://dx.doi.org/10.1016/S0167-6296(97)00020-9. [54] Giannoni M, Hitiris T. The regional impact of health care expenditure: the case of Italy. Applied Economics 2002;34(14):1829–36, http://dx.doi.org/10.1080/00036840210126809. [55] Di Matteo L. The income elasticity of health care spending. The European Journal of Health Economics 2003;4(1):20–9, http://dx.doi.org/10.1007/s10198-002-0141-6. [56] Cantarero D. Decentralization and health care expenditure: the Spanish case. Applied Economics Letters 2005;12(15):963–6, http://dx.doi.org/10.1080/13504850500377926. [57] Costa-Fonti J, Pons-Novell J. Public health expenditure and spatial interactions in a decentralized national health system. Health Economics 2007;16(3):291–306, http://dx.doi.org/10.1002/hec.1154. [58] Moscone F, Tosetti E. Health expenditure and income in the United States. Health Economics 2010;19(12):1385–403, http://dx.doi.org/10.1002/hec.1552. ˜ S. The determinants of health care expendi[59] Cantarero D, Lago-Penas ture: a reexamination. Applied Economics Letters 2010;17(7):723–6, http://dx.doi.org/10.1080/13504850802314437. ˜ [60] Prieto DC, Lago-Penas S. Decomposing the determinants of health care expenditure: the case of Spain. The EuroJournal of Health Economics 2012;13(1):19–27, pean http://dx.doi.org/10.1007/s10198-010-0276-9.


Convergence and determinants of health expenditures in OECD countries.

Measuring and comparing health care waiting times in OECD countries.

The Impact of Certificate-of-Need Laws on Nursing Home and Home Health Care Expenditures.

The Impact of Obesity on Health Care Utilization and Expenditures in a Medicare Supplement Population.

Determinants of health-system efficiency: evidence from OECD countries.

Financial impact of the GFC: health care spending across the OECD.

Current Trends in Health Insurance Systems: OECD Countries vs. Japan.

Impact of random assignment on study outcome: an empirical examination.

Current trends in health insurance systems: OECD countries vs. Japan.

How relevant is environmental quality to per capita health expenditures? Empirical evidence from panel of developing countries.

Public Health Adaptation to Climate Change in OECD Countries.

The impact of changes in county public health expenditures on general health in the population.

The concentration of health expenditures: an update.

The impact of health systems on diabetes care in low and lower middle income countries.

Expenditures on health research in sub-Saharan African countries: results of a questionnaire-based survey.

The socioeconomic determinants of health: economic growth and health in the OECD countries during the last three decades.

The Association between Peace and Life Expectancy: An Empirical Study of the World Countries.

An Empirical Analysis of Rural-Urban Differences in Out-Of-Pocket Health Expenditures in a Low-Income Society of China.

The Impact of Visualizing Nested Sets. An Empirical Study on Tree Diagrams and Unit Squares.

Utilization rates of knee-arthroplasty in OECD countries.

Utilization rates of hip arthroplasty in OECD countries.

Impact of Noncommunicable Disease Multimorbidity on Healthcare Utilisation and Out-Of-Pocket Expenditures in Middle-Income Countries: Cross Sectional Analysis.

Systemic opioid elimination after implantation of an intrathecal drug delivery system significantly reduced health-care expenditures.

Health care expenditures attributable to smoking in military veterans.