Medical liability under limited resources: A law and economics perspective

Abstract

Should a lower standard of due care apply to healthcare providers facing scarce resources? We analyse tort law in its capacity to provide deterrent incentives to potential tortfeasors and to spread accident losses that were not avoided. The fact that resources in healthcare are limited, among others due to budget restraints, affects the necessary weighing of costs and benefits of precautionary measures. Limited resources may force healthcare providers to choose between treating more people with less care or fewer people with more care. We argue that limited resources provide a preference for a negligence rule over strict liability, because otherwise, part of the (limited) budget would be spent on tort damages for losses which were too expensive for society to prevent. A reversed burden of proof addresses information problems of the victims. Furthermore, regulation and medical guidelines are important in providing desirable behavioural incentives to avoid accidents.

Introduction

Since the beginning of this century, there has been an increasing debate on how the scarcity of resources for public health affects legal liability. This debate is relevant around the globe, because the question whether and how scarcity of resources allocated to healthcare should impact possible liability of healthcare providers (in establishing liability and/or in the magnitude of this liability) occurs in many countries. Questions concerning the allocation of scarce resources during a medical emergency arose especially in the United States after 11 September 2001 ¹ and more particularly – not only in the United States but worldwide – also after the H1N1 pandemic. ² The general question arose whether during disasters, referred to as public health emergencies ³ a different standard of ‘due care’ should apply than under normal circumstances. ⁴ The different level of due care is referred to as a crisis standard of care (CSC). ⁵ In the framework of the healthcare emergency, the introduction of the notion of CSCs was especially meant to immunise or shield healthcare providers from tort liability. ⁶ After the COVID-19 pandemic, government officials and emergency planners were equally confronted with difficult choices whereby many states in the US introduced liability shields to protect healthcare professionals from lawsuits based on allocation decisions. ⁷ The application of those statutes providing liability shields to healthcare professionals in situations of emergency has, however, not been undisputed. Some warn that the application of CSCs could lead to wrong decisions and even discriminatory impacts. ⁸ Others have criticised the application of CSCs as unethical and not effective in providing protection to healthcare professionals. ⁹

In fact, the problem that comes to the surface in full force in a medical emergency situation (like COVID-19) is of a much more general nature and comes down to the question of how limited resources in healthcare affect medical liability. The problems of limited resources have played a huge role during the COVID-19 pandemic where healthcare providers needed to decide how to allocate scarce healthcare resources (like ventilators, ICU beds, and other medical supplies) under the risk of tort liability. ¹⁰ But the problem can obviously arise outside the situation of medical emergencies as well, since resources available for effective healthcare may as a matter of principle be limited, and not only in emergency situations. The problem has come to the fore, not only in the United States, but in many other countries as well. ¹¹ Given the debates to which lacking resources have led in relation to medical liability, we will in this contribution examine the effects of limited resources for medical liability from an economic perspective, in order to contribute to the debate. The economic approach to law, also referred to as ‘law and economics’, applies economic concepts and methods to study a wide variety of legal topics. Human choice, and the impact of legal rules on such choices, is analysed from an economic point of view, applying i.a. rational choice theory and behavioural economics. In addition, law is studied in its capacity to improve social welfare, for example by providing incentives to avoid social losses and by trading off various types of social costs. ¹² Law and economics has paid extensive attention to the question of how legal rules (more particularly liability rules, but also regulation) can contribute to a minimisation of the social costs of accidents. ¹³ We will argue that this methodology can also be usefully applied to analyse the effects of limited resources on medical liability. We will thereby analyse this problem at a more general level (not only in medical emergencies), assuming that healthcare is by definition often confronted with limited resources, as a result of which choices and implicit prioritisation are to some extent unavoidable.

We therefore focus on a variety of scenarios, such as for example:

• The case where limited resources available in a hospital make it not possible to treat a patient with the most expensive treatment;

• The case where a hospital needs to make a choice between a variety of different treatments and tries to make an ‘optimal’ choice, taking into account the marginal costs of various treatments versus the marginal benefits each treatment would have.

To be clear: the case we are not envisaging in this contribution is the more classic medical liability case, such as the symbolic case of a surgeon who forgot to remove particular equipment (such as scissors) from the patient after an operation or typical cases where sub-standard medical care was provided. Those are the ‘easy’ cases where all would agree that they would give rise to liability, irrespective of the available resources for the simple reason that the treatment was sub-standard.

The general research question we therefore would like to address in this contribution is the following: should the limited availability of resources affect the required level of medical care or, more generally, a finding of medical liability?

As the economic approach to law uses a normative framework based on efficiency, the question is indeed not only posed in a positive fashion (‘does’), but also in a normative fashion (‘should’). That makes it clear that we do indeed take the economic framework including efficiency and cost-benefit analysis as the starting points for answering the research question. We do obviously fully recognise that there are many other non-economic (e.g. justice) considerations involved in answering our research question. One could easily imagine difficult ethical choices when there are indeed limited resources and choices have to be made concerning which patient will receive a particular treatment and which patient will not. ¹⁴ Those are undoubtedly important and relevant ethical choices. We will, although recognising their importance, not discuss those questions for the simple reason that it goes beyond the domain of economics and requires expertise from other fields to be able to address those ethical issues.

A final introductory remark: we will mostly focus on the medical liability of hospitals and assume that the hospital is the one making the trade-off on available resources on the one hand and the healthcare to be provided on the other hand. One reason to focus on hospitals rather than on individual physicians is that in many legal systems medical liability is de facto channelled to the hospital. Another reason is that the choices influenced by the available resources are often made at the level of the institution (hospital) rather than at the level of the individual (physician). ¹⁵ In making this trade-off, ‘priority setting’ obviously is very important. However, our article concerns the question whether scarcity of resources should affect the level of due care applied in healthcare situations. We hence do not focus on the priority setting by the hospitals, but on the possible impact of scarcity of resources on how the choices of healthcare providers are assessed in a setting of possible liability. The topic of priority setting as such therefore is not included in our article.

Starting point: Coase theorem

The basic idea of the so-called Coase theorem is that when parties are fully informed, an agreement between the parties can be reached irrespective of the legal rule. This is especially the case when a producer and a purchaser (assuming that this is the potential victim) are bound to each other via the price mechanism. ¹⁶ The price can provide an indication concerning the division of risk between the parties. In the context of medical liability, one would argue that the Coase theorem teaches that the price agreed between the hospital and the patient would be the result of an efficient bargaining process. Theoretically, the patient and the hospital could agree on a variety of different treatments and on a corresponding price. This would in theory imply that when a patient selects a low level of healthcare, he or she also pays a corresponding low price and the reverse would be the case when the patient chooses a high level of healthcare.

Of course, this may sound very theoretical in the context of medical malpractice. One problem is that a patient is (fortunately) often a so-called one-shotter. ¹⁷ As a result, the patient will often lack appropriate information. The Coase theorem assumes that both parties have full information (which is a part of the zero-transaction cost assumption underlying the Coase theorem). That assumption is of course unrealistic in the context of medical malpractice. Given the information asymmetry between the physician and the patient, the patient will undoubtedly rely (almost) completely on the expert advice provided by the physician. ¹⁸ Moreover, very often it is not the patient himself or herself paying for the healthcare. The system is rather one of third-party payment: it will be a social security provider or a first-party insurer who provides the payment and who might therefore also decide upon the price to be paid for a particular level of healthcare. In other words, it is not the patient him/herself deciding either upon the healthcare level or upon the price. As a result, the Coasean (contractual) model that assumes free negotiations between the hospital and the patient, resulting in an appropriate agreement on the appropriate care level and the corresponding price, seems in most cases unrealistic.

However, this finding that ‘Coasean bargaining’ often is unrealistic is mostly true as far as the majority of the population is concerned and for the public healthcare system. There is, however, also a segment of the population that could (and does) afford higher levels of healthcare (e.g. in private hospitals), often financed via private insurance (rather than through social security) purchased by the prospective patients themselves. These arrangements where higher (and more expensive) healthcare levels are provided to a particular segment of the population that can afford this is a perfect example of the application of the Coase theorem as these patients pay indeed a higher price (directly or via insurance) for a higher healthcare level. Those arrangements will, however, mostly only be available for ‘the happy few’.

Third-party liability

Given that we have just argued that the contractual model is not very realistic in the case of medical malpractice, we assume that third-party liability – tort liability – of hospitals is possible, even though there may be a contractual relationship between the hospital and the patient as well. Assuming that the applicable regime, in a case where unexpected harm occurs to the patient, will then be third-party liability, the first question that should be addressed is what the goal would be of a third-party liability regime in this particular context. Let us recall that it would be the context (see ‘Introduction’) where a patient would for example blame the healthcare provider for not having chosen a more expensive treatment simply because of limited available resources within the hospital concerned. Generally, two goals of a liability regime are distinguished. ¹⁹ The first goal (often stressed by lawyers) would be compensation. The liability regime would then be seen as a means to provide adequate compensation for the harm suffered by the patient. The second goal could be prevention. The idea would then be that a finding of liability could provide deterrence and would steer the behaviour of a healthcare provider to incentivise the latter for taking an optimal level of care. For both goals of liability, the question that could be asked would be why tort law would be chosen, given that there are alternatives. Let us address both goals in turn and address the potential of tort law.

First, regarding compensation, law and economics scholarship has been extremely critical of the ability of tort law to provide adequate compensation to accident victims. ²⁰ The tort law system is ultimately very uncertain in the sense that only a few (of the many) victims may be compensated via tort law. ²¹ It has to do with the high barriers for victims to access justice and to receive compensation via tort law. Estimates have indicated that literally only a few victims (not more than a few percent) effectively receive compensation via the tort litigation system. ²² Also as far as medical malpractice is concerned, studies indicate that only 1 in 10 cases of negligent injuries resulted in a claim. ²³ Also recent research confirmed a very low likelihood of claim and compensation payment ²⁴ and that the majority of the court rulings are dismissive of patient claims. ²⁵ Studies indicate that 1% of hospital patients are victims of medical negligence, but fewer than 2% of these injured patients file claims as many attorneys will reject legitimate cases under contingency fees. ²⁶ In other words, the tort system is extremely selective in its compensatory function and not able to provide systemic compensation for a large number of victims. There are, moreover, obvious alternatives that do much better as far as providing compensation is concerned than the tort system. One is social security; the other is (voluntary or mandatory) first-party patient insurance.

Second, similar questions can be asked regarding deterrence. In theory, an exposure to liability of the healthcare provider (hospital) could provide incentives for appropriate decision-making, more particularly in the weighing of various alternatives versus the available resources. However, there is an obvious alternative, which is (self-)regulation. There may, as will be shown below, ²⁷ even be a strong argument in favour of (self-)regulation, for example via guidelines. The studies devoted to the effects of medical malpractice on healthcare quality also shed doubts on the deterrence hypothesis: greater tort liability has, at least in the United States, not been systematically associated with improved quality of healthcare. ²⁸

It is important to reflect in further detail on what compensation in this particular case exactly entails. Losses that have been suffered must be borne by someone, somewhere in society. This might be the individual victim or the individual injurer, or the losses might be spread over larger groups (via third-party liability insurance, via first-party loss insurance, or because the government spreads them over all taxpayers). If one focuses on compensation, liability primarily deals with the issue of whether the victim bears their own risk or whether the risk can be transferred to the injurer. In the case of patients and hospitals:

• If the hospital is liable for the losses suffered by the patient, the costs of providing healthcare by the hospital increase with the amount of this liability. Hence, healthcare costs increase in comparison with the situation without liability. This may result in an increase in the price for health treatments and/or an increase in the premium for health insurance (if available). This implies that the liability is ultimately borne by the patients (via higher healthcare costs), by the insured (via a more expensive health insurance), and/or by ‘the taxpayers’ (in so far as healthcare is paid out of tax revenues).

• If the hospital is not liable for the losses, the patient-victim bears the losses. Who ultimately bears these losses depends on their nature. If it regards additional healthcare costs (e.g. more and/or more serious injuries or illnesses due to the initial medical treatment), they are born by whoever finances healthcare (patients, insured, taxpayers), just as was the case when hospitals are liable. If it regards loss of income, either the victim bears those losses, or it is spread over a larger group via private or public insurances. Only a small part of the losses (more particularly non-pecuniary losses) will be borne by the victim.

Therefore, irrespective of whether the hospital is held liable, the losses ultimately end up with similar groups of people (insured or taxpayers). For uninsured patients, liability of the hospital does result in spreading the losses, via the price of the health treatment. Without this liability, they would have to bear the full losses individually. Viewed like this, liability is more about spreading the losses than about compensating them as such.

Determining the optimal care level

From the famous paper of Shavell, ²⁹ it is known that the goal of liability law is not to provide incentives to avoid all accidents for the simple reason that the social costs of reaching a zero accident level would simply be too high. ³⁰ The question is rather how the point can be found where the marginal costs of prevention equal the marginal benefit. ³¹ Again, such an analysis could in principle also be made to select an optimal treatment whereby, given limited resources, the marginal costs of various treatments could be compared to the marginal benefits in reaching a particular result. However, there are several specific features for determining the optimal healthcare level under limited resources.

A first issue is that, probably in contrast to other tort situations, the choice can hardly be made with respect to just one patient, because the choice of a particular (expensive) treatment for one, may also have repercussions for others. The ‘tragic choices’ ³² that have to be made in this respect, have an important collective element. Indeed: choosing a very high care level for one particular patient may have as a consequence the result that the limited resources would no longer be available (of course depending upon the financing system). As a consequence, a decision in one case (of one patient) may have consequences at the collective level (for other patients). To put it differently: in this particular trade-off, the due care level cannot simply be determined at an individual level. That raises the question of who should make this particular trade-off. One thing is clear: these complicated resource allocation choices should preferably not be made at the bed-side level by the individual healthcare provider, ³³ but rather at a collective level.

In addition to this first point that decisions regarding due care levels should be taken at a collective level, a second issue is whether the limited availability of resources generally should result in a lower due care level. In theory, there are certainly arguments in that respect. Optimal care levels are the result of a balancing of marginal costs of prevention and marginal benefits. When resources are scarce, it can be argued that the marginal costs of care increase. The reasoning is as follows: in the case of limited resources, the so-called opportunity costs of care increase. ³⁴ The reason is that particular (costly) care provided will lead to less care for other patients for the simple reason that the available budget has been exhausted. In other words: the lower the resources available, the more serious opportunity costs become (e.g. because other seriously ill patients cannot be treated anymore). As a result of this reasoning, it can be argued that it is almost unavoidable to take the available resources into account when determining optimal care levels, ergo: limited resources will unavoidably reduce optimal care levels and should therefore (from an economic perspective) also reduce due care.

This could also be argued from a different angle: suppose that limited resources would not be considered as an argument when determining due care levels. In theory there are certainly arguments in that respect. They relate to two economic notions: scarcity and opportunity costs. In practice, that may imply that for example a hospital could not deny a patient a particular treatment simply because resources are not available. That would imply that the legal norm for the hospital would be determined as if there were no scarcity. As a result, the hospital may provide the expensive treatment, but that would imply that other patients could (given scarcity) no longer be treated as the budget has been exhausted. If the care costs per patient cannot be lowered, the result will be that the hospital will have to reduce the number of patients and waiting lists will be the result. If one rejects the argument that limited resources lead to lower due care levels, the unavoidable consequence is that, in that case, fewer patients can be treated which may, from society’s perspective, not be an optimal outcome. From that perspective, there could be a strong argument indeed to lower due care levels as a result of limited resources in order to be able to treat more patients, albeit with less care.

This result still merits a few comments:

1. Here one clearly notices the difference between the economic and the legal approach. (Tort) lawyers are often confronted with the (painful) situation of one individual victim and address liability from an ex post perspective, resulting in the desire to provide justice ex post to that individual victim. Economists, to the contrary, would address liability rules from an ex ante perspective and would therefore ask the question: what would the consequences of awarding high care to one individual victim be for society at large (i.e. for other patients)?

2. We do realise that our argument that limited resources can be leading to a lower optimal (and hence also due) care level, has to be applied with caution. One would have to be aware of (and avoid) the (moral hazard) risk that the healthcare provider would either pretend to be insolvent or organise its own insolvency (or at least lacking resources) in order to benefit from a lower due care standard.

3. The previous discussion shows that limited resources lead to an important societal trade-off: should the hospital provide less care to more people or more care to fewer people? From this the question arises of whether that is in fact a matter that tort law (and thus with the judge deciding in an individual case of medical liability) should be confronted with, or should it rather be the regulator.

Liability or (self-)regulation?

The choice is basically whether this trade-off has to be made via the liability regime or instead through (self-)regulation. This choice between regulation and liability rules was thoroughly examined by Steven Shavell in 1984 in a paper in which he advances several criteria that influence the choice between safety regulation and medical malpractice. ³⁵ The first criterion is information asymmetry. There may be a strong argument in favour of safety regulation: strict liability would assume that healthcare providers themselves would (under the deterrent effect of liability rules) determine the appropriate standard of healthcare. For individual healthcare providers, the costs of determining the optimal standard may be high. This is exactly what they would have to do under a strict liability regime. Under a fault/negligence regime it is the judge who does the balancing and determines the due care standard. Also, a judge may lack appropriate information to determine this optimal care level. As the optimal level of healthcare would probably not be determined just for one specific patient, economies of scale could be used in regulation. That would mean that at a central level (e.g. the national government) the optimal standard would be determined which would subsequently be passed on via (self-)regulation.

The second criterion advanced by Shavell is the insolvency risk. Liability rules no longer have a deterrent effect in case there might be insolvency. Obviously, the likelihood of insolvency of hospitals is not that large, but there is always a risk (certainly in case of multiple losses), that harm could be caused which could substantially outweigh the assets of the healthcare provider. In that respect one should also take into account that hospitals could be organised as limited liability corporations, which would also allow them to externalise harm to society. ³⁶

The third criterion advanced by Shavell is the danger that a liability suit would not be brought, for example because of proof-of-causation problems or latency. Latency is the problem that a long timelapse may have occurred between the moment that a wrongful treatment took place on the one hand and the moment that the damage occurred on the other hand. The chances of a liability suit for medical malpractice are, as already indicated, usually very low. ³⁷ It is often hard for patients to prove a causal link between the treatment and their damage. In some cases, victims will also not recognise that their harm has been caused by a tort but might think that the particular ailment is simply the result of their general bad health condition. Empirical evidence indeed indicates that many cases of medical malpractice are never recognised as such and do not give rise to liability. ³⁸ For all those reasons a liability suit might not be brought and hence safety regulation is necessary to ensure that an optimal care level is implemented.

These criteria point in the direction of regulation. Given that especially medical malpractice claims can be very complicated, time-consuming, and (therefore) expensive, ³⁹ in our view the application of this Shavellian framework results in the conclusion that the optimal trade-off between various treatments and the limited resources available should in principle be determined via regulation. The next question then, of course, is how this should exactly be implemented.

Self-regulation and guidelines

The determination of an optimal trade-off between the most effective treatment, its costs and available resources is obviously a very complicated one which should probably be determined not just by the government via ex ante regulation. An alternative is to work with guidelines within the framework of self-regulation. In many countries, the level of healthcare to be followed by a physician is primarily determined through disciplinary rules, set by the professional organisation. Non-compliance with the standard set means that disciplinary sanctions will be imposed on the particular healthcare provider. Moreover, in many legal systems, self-regulation and government regulation are combined in the sense that the non-respect of disciplinary rules by the healthcare provider could also be sanctioned via public enforcement. As a result, throughout this article we refer to guidelines, but these are not purely self- or private regulation. They are rather the result of a collaboration between the government and professional groups. The guidelines therefore have a rather hybrid character, holding the middle ground between pure private regulation (rules only drafted by the professional group) and pure government regulation.

The literature has indicated that self-regulation has the advantage that healthcare providers may have better information to determine the appropriate care level than the regulator. ⁴⁰ Compliance may also be better if the professionals had been involved in the drafting of the rules. However, that also entails the risk that the self-regulation may not always reflect the public interest, but rather the private interest of the healthcare providers concerned. ⁴¹ It is for that reason that it should not only be the healthcare providers who should determine the guidelines, as there is a public interest that plays a role as well (access to affordable healthcare for all). Therefore, it may make sense to develop guidelines by means of mutual agreements in which the healthcare providers (hospitals) are involved, but also the government (the health inspection services) and those providing the finances (such as healthcare insurers or social security carriers). In the ideal case, negotiations between those stakeholders could lead to an agreement which could be translated into guidelines. Therefore, as just mentioned, those guidelines have a rather hybrid character. The advantage of those guidelines is that they would make the life of healthcare providers and hospitals in practice much easier. During the COVID-19 pandemic, one of the problems was that healthcare providers sometimes needed to make critical life-or–death decisions given the scarce resources at the bedside, without any precise guidance ⁴² or triage protocols. ⁴³ The use of guidelines implies that the trade-off would not have to take place in an individual case by the healthcare provider at the bedside, but that it could rely on the guidelines providing an answer to the trade-off.

A supplementary role for liability rules?

If the just advocated system of guidelines worked perfectly, there would obviously be no need for a complementary role of liability rules. However, in the law and economics literature with respect to the optimal combination between regulation and liability rules (a so-called smart mix), ⁴⁴ it has been advocated that there are many reasons why safety regulation may not be able to deal with all externalities: ⁴⁵

• Safety regulation cannot deal with all possible situations that could emerge in practice;

• Safety regulation is often static, not dynamic and may be outdated fast;

• Sometimes lobby groups (in the case of this article, of healthcare providers) could affect the level of safety regulation as a result of which that may be suboptimal;

• Safety regulation only works if there is a full compliance as a result of effective enforcement, which is hardly ever the case.

That can also play a role in this particular case: even though guidelines can be the primary instrument to determine the optimal trade-off between a particular treatment and limited resources, there may be situations not foreseen by the guidelines and thus the question arises as to the potential complementary role of liability rules in addition to regulatory guidelines. A complementary role of liability rules would be important for providing deterrence where the guidelines would fail (e.g. because the guidelines dictate a too low level of healthcare or because there is a situation not provided for by the guidelines). ⁴⁶ That then obviously also raises the question as to the optimal relationship between (compliance with) the guidelines and the (complementary) liability regime. In that respect the nature of the liability regime is of course of importance.

Strict or fault liability?

A classic question in the law and economics literature is whether an injurer should be exposed to strict liability or to fault liability. The law and economics literature has developed several criteria which determine whether strict liability or fault liability is better suited to address the particular situation. ⁴⁷ We will now review those criteria in turn to verify how they affect the choice between strict liability and negligence in the particular case where tough choices in healthcare need to be made with limited resources.

Determining optimal care

Determining the optimal care level with limited resources involves weighing the marginal costs of care against the marginal benefits thereof. Assessing this optimal care level is by no means an easy exercise and in fact involves high information costs. ⁴⁸ Depending upon the liability regime, the stakeholder who incurs the information costs may be a different one. This can therefore already be an important indication for the choice between negligence and strict liability. There are roughly three options in that respect:

If the potential injurer has the best information about the possible care measures and their effect on the expected accident losses, strict liability is likely to provide the best care incentives. After all, we then leave it to the potential injurer to decide which care measures to take. Strict liability is a regime whereby the injurer (healthcare provider) will be confronted with all the social costs of accidents, that is, cost of prevention and the expected accident costs. Being confronted with all costs, under this strict liability regime, the injurer will use his superior information to determine the optimal care level.

If the court has the best information (e.g. because the injurer is a layman and because there are so many relevant characteristics of the case that general regulation is not feasible), then negligence is the best option. Here, the court formulates a behavioural norm based on the exact circumstances and determines whether the tortfeasor took enough care.

If the legislator has better information than the injurer and the court (e.g. due to economies of scale and the possibility of undertaking specialistic research), then ‘breach of a statutory duty’ may be preferrable. The legislator uses its superior information to formulate (via legislation) a behavioural norm and communicates this to the relevant legal subjects. They merely need to know the rule, so that they know which care measures to take. This resembles a negligence rule in the sense that there will be liability of the injurer in case specific, previously determined rules have been violated.

In medical cases, if the legislator has the best information about costs and benefits of care measures or can undertake large-scale research, breach of a statutory duty seems the best option. This fits in with the finding in Section ‘Liability or (self)regulation?’ that regulation may be preferred to liability in healthcare cases. The same would apply if regulation was not government regulation but instead private regulation (e.g. guidelines set by the medical association). Also, in that case it could be argued that these professionals may have better information and would therefore be better equipped to determine the standard via guidelines. Tort law then would have the function of private enforcement of this regulation. This might be relevant if public enforcement were to fail, for example, due to the influence of an interest group lobbying for lax enforcement. ⁴⁹ If the regulation itself posed restrictions which were too lenient, then negligence could play a good role, because courts are less susceptible to the influence of interest groups than the legislator. However, informational problems on the side of the court (and probably also on the side of the victim; see the next subsection) make it unlikely that negligence would function perfectly. Then, strict liability could be desirable. But this assumes that the hospitals have enough information about costs and benefits of care measures to be able to take correct decisions.

Possibilities to sue for the victim

Even though there might be arguments for a negligence rule, this may be problematic from the perspective of the victim. It requires the victim to prove a violation of the norm and that this violation caused their losses. If the victim cannot prove negligence and/or causation, the injurer will not be liable.

In medical cases it is highly doubtful whether the patient can prove the negligence of the healthcare provider. Therefore, in order to give correct incentives to the healthcare provider to take optimal care, strict liability may be more appropriate. In this case, the victim does not have to prove the negligence of the healthcare provider, but only that its losses were caused by the healthcare provider.

An alternative is a negligence liability with a reversal of the burden of proof. ⁵⁰ Just as with strict liability, the victim does not have to prove that the hospital has violated a norm. The victim can simply pose that this is the case, and it is then up to the hospital to prove that they did not violate the norm. This will provide incentives to the hospital to carefully register the various steps in the treatment in order to show that they complied with the regulation. ⁵¹ This may strike a balance between, on the one hand, the desire to use regulation (if the legislator has superior information) and, on the other hand, to avoid a problematic burden of proof of the patient (which would make the threat of liability illusory). There are some other advantages as well, which we will touch upon in the subsections below.

Note that if the healthcare provider could not show compliance with the regulation, this would amount to negligence. As a result, the healthcare provider would be held liable to compensate the victims’ losses. But the reverse is not necessarily the case. In other words: even if the healthcare provider could show compliance with the regulation, this would not necessarily mean that there is no negligence as the optimal care level could be higher than the care level fixed by the regulation. The healthcare provider could still be liable if the victim could show that an optimal care level (higher than that fixed in the regulation) should be applied and that it was violated by the healthcare provider. However, in the latter case, the burden of proof (of both the higher optimal care level and the violation) then lies with the victim, which obviously may not be easy. The reversal of the burden of proof would only apply as far as the compliance with the regulation is concerned.

Activity level

Parties can not only prevent losses by taking more care, but also by lowering their activity level. A party chooses the optimal activity level if he is confronted with all the costs that are caused by the activity (both to this party and to others). Hence, only the party that ultimately bears the expected accident losses will choose the correct activity level. ⁵² The conventional wisdom in the law and economics literature therefore holds that strict liability provides incentives to choose the correct activity level for the injurer, while fault liability gives the correct activity incentives to the victim. It is not possible to provide correct incentives to both parties simultaneously, because only one party will bear the accident losses.

In medical accidents, it is in principle more important to control the activity level of the healthcare provider. This would plead for strict liability. However, three comments should be made in that respect.

First, particular activities (such as healthcare) may also create substantial positive effects (sometimes referred to as positive externalities). ⁵³ In some cases, an increase in the activity level can be socially desirable. ⁵⁴ That may well apply to the situation of healthcare providers whose activities create huge social benefits.

Second, in situations where the losses of the victim may also be caused by something other than the behaviour of the injurer, strict liability triggers the risk of so-called ‘crushing liability’. ⁵⁵ For example, if certain health conditions could be caused by the medical treatment but also by the underlying health condition of the patient, there is a risk that under strict liability the healthcare provider will not only be liable for the losses it actually caused, but also for the losses which also would have happened without the medical treatment. The same holds for medical conditions which may be a side-effect of medication, but can also be caused by the illness which the medication is trying to tackle. Obviously, this could be avoided by strictly and correctly applying the causation requirement, that is, holding the hospital liable only for the additional losses that were caused by the wrongful medical treatment and not for losses that would have occurred anyway. However, precisely because it may in practice often be highly difficult to differentiate between those, there is a serious danger that the hospital would also be held liable for losses that were in fact not caused by a wrongful medical treatment.

Third, strict liability which is applied too harshly could lead to so-called ‘defensive medicine’. ‘Negative defensive medicine’ implies the withdrawal from certain (types of) treatments. It may result in an activity level that is lower than is desirable, because the positive effects of the treatment are now lost. ⁵⁶ Fault liability could have avoided this, because by taking the required level of care, the healthcare provider can escape liability altogether. Unfortunately (and connected to the first point we raised in the choice between strict liability and fault liability), if the courts can make mistakes in assessing the care level that the healthcare provider actually applied (and possible in establishing the due care level as well), out of fear for liability the healthcare provider may engage in ‘positive defensive medicine’. This implies that they may provide too much care to an individual patient, in order to reduce the probability of being found at fault if the patient should suffer a loss. And here as well, negative defensive medicine may occur for the types of treatments where the risk of liability is too high.

Unobservable care dimensions

Expected accident losses may be reduced in various ways. Some care measures are observable, while others are not (or difficult). Monitoring whether certain measures are taken is relatively easy for the observable dimensions of care, but not for the unobservable ones. It might be easily observable whether stitches are done correctly, whether the right medication was given and whether enough people were present during surgery. It may be much more difficult to monitor whether the healthcare providers were not too tired to perform the activities they performed, or whether they were not distracted while performing them.

A problem is that medical errors are sometimes caused by medical negligence (which is the situation where the patient did not receive treatment consistent with appropriate medical care). ⁵⁷ The traditional economic model regards negligence as a deliberate choice to take too little care. However, much medical negligence is accidental. This may for example be the case where a physician does not have enough information to correctly diagnose the patient or to select a correct treatment. The risk of those errors can be reduced through ex ante patient safety investments. Those investments, for example in expertise (via training), medical technology, adequate staffing, and support systems are costly, but may reduce the risk of accidental errors. From an economic perspective, these investments should be undertaken when their marginal cost is lower than the marginal benefit. According to Arlen, a problem that arises is that within the assessment of medical negligence those patient safety investments are often not taken into account. Medical negligence is often assessed on the basis of patient-specific care. As a result, the way in which medical negligence is assessed may only provide incentives to invest in these ex post care measures whereas the ex ante patient safety investments may be more cost effective. The problem is also that those ex ante patient safety investments are a collective good, that is, it is not just one, but all patients who benefit from those investments. As a result, incentives to do those essential investments cannot be provided via individual contracts nor via negligence rules. Tort liability can therefore not provide individual physicians with correct incentives to invest in ex ante patient safety and neither can contractual liability.

Incentives to also invest in ex ante patient safety can be provided via strict liability. The advantage of a strict liability regime is that it also provides incentives to invest in the non-observable care dimensions (like, for example, ex ante patient safety investment), because taking care in those dimensions affects the accident probability and therefore also the expected losses and hence the expected liability. Such dimensions would not be incorporated in a negligence regime. There the judge would only examine whether the healthcare provider has taken the observable, ex post required care measures.

Summarising, the more important the non-observable dimensions of care, the more important it may be to expose a healthcare provider to strict liability. For the particular example of ex ante patient safety investments, one can also argue that those constitute a strong argument in favour of hospital liability as those institutions can provide optimal investments in ex ante patient safety. Moreover, given the public good nature of ex ante patient safety investments, it may also be indicated to subject those to regulation (such as the types of guidelines discussed before).

Administrative costs

The negligence rule is a relatively costly system. Legal norms have to be formulated and in individual cases it has to be established whether the injurer violated the norm. Furthermore, it has to be established that this violation caused the losses. Strict liability does not involve an examination of those issues which makes a lawsuit easier. A potential problem under strict liability is that it could lead to more lawsuits (as in principle every accident gives raise to compensation). However, because the outcome of a case under strict liability is more predictable, more cases will be settled, which reduces the system costs even further. ⁵⁸

The judgement proof problem

The preventive function of tort law presupposes that the tortfeasor is able to pay damages. In the extreme case where an injurer has no funds at all, he is judgement proof and cannot pay damages. Tort law then cannot fulfil its deterrent role.

In the less extreme case where an injurer has some assets but not enough to cover all losses he has caused, tort law may provide fewer incentives than in situations without this problem. Strict liability is more affected by this insolvency problem than fault liability. The reason is that the ‘bonus’ of taking care is bigger under fault liability (namely, escaping liability altogether) than under strict liability (merely reducing the probability of being held liable, but not escaping it altogether). Smaller funds may therefore still be enough to provide adequate incentives under fault liability but not under strict liability. ⁵⁹

Therefore, in principle, the issue of insolvency provides an argument in favour of fault liability. There are, however, a few qualifications to be made. First, whether insolvency actually will be problematic depends on the size of the losses in comparison with the size of the assets. In the case of large hospitals with large budgets, the problem may not arise very often, especially now hospitals will (almost) always be insured against liability. This holds even more if ultimately the government would provide financial back-up of the hospital. ⁶⁰ Of course, if a medical issue arises which affects many patients in the hospital, for example, a bacterial infection which infects many patients, or a doctor that treats many patients incorrectly, the losses may be substantial. However, it is very doubtful whether, especially given the insurance of the hospitals, insolvency really will become an issue.

Summary

Looking at these six factors, some of them point in the direction of strict liability and some in the direction of fault liability, but there is no clear preference for one of the two approaches. It is most important, from the economic perspective, that accident losses are optimally prevented and that the non-prevented losses subsequently are spread over a larger group. ⁶¹

If both strict liability and negligence can provide adequate care incentives, against the background of limited resources an argument can be made for a negligence rule. With strict liability the hospital is the residual loss bearer while under fault liability it is the victim. So, if the hospital has taken the desirable care measures but losses still occur, these are the losses for which it is too expensive to avoid them. Liability then only entails a redistribution of welfare. Under strict liability, the hospitals would still have to bear these losses, which implies that part of the (already limited) budget will be spent on this expected liability or on liability insurance. ⁶² This, in turn, means that under strict liability, there is less budget remaining available for healthcare than under fault liability. This could result in the situation where certain health treatments can no longer be provided because there is no budget anymore. This would mean a ‘lower than optimal’ activity level because the ‘efficiently non-avoided losses’ are still allocated to the activity. This relates to the issue of ‘positive externalities’ we referred to above: if only negative externalities are internalised but not positive externalities, the activity becomes more expensive than is socially desirable and the activity level may become lower than optimal.

If strict liability would provide better care incentives than negligence, a trade-off has to be made: on the one hand, strict liability may result in fewer accidents (due to the better care incentives), which is desirable. On the other hand, the hospitals are then also liable for the losses which are too expensive to avoid so that the activity level may be too low. This is undesirable. The essential (and empirical) question then becomes which losses are larger: those that are caused by less available healthcare (under strict liability) or those that are caused by a possible larger number of medical incidents (under fault liability)?

Above, we indicated a preference for a negligence rule with a reversal of the burden of proof. This may equally solve the trade-off mentioned above. By reversing the burden of proof, the informational advantages of strict liability can still be kept, because the hospital can only escape liability if it shows that it has taken all the measures prescribed by the regulation. If the hospital proves compliance with the regulation, there is no liability. This avoids part of the available budget having to be spent on compensating losses that did not have to be avoided in the first place. As a result, more of the limited budget and resources remain available for providing healthcare.

An unbalanced application of strict liability risks leading to over-deterrence, specifically in this case where the specific standard of care depends on a careful weighing of different healthcare options and the available budgets. It is therefore wise, as we argued before (in Section ‘Self-regulation and guidelines’), to have this primary decision made by healthcare providers jointly with the government and those who finance healthcare (such as social security institutions). That would de facto be an argument to opt for a negligence rather than for a strict liability regime. The advantage of the negligence rule would be that it can largely rely on the guidelines that have been set ex ante and that will provide a standard concerning the optimal level of healthcare. The advantage of this model (combining ex ante guidelines with a negligence rule) is twofold: (1) it provides the judge with an indication of the due care level required under negligence; guidelines can substantially reduce information costs for the judge; (2) victims will have an incentive to show non-compliance with the guidelines. If that will automatically constitute negligence and they can prove a causal relationship with the harm they suffered (or causation is assumed, unless the hospital proves it was lacking), compensation can be obtained. In this way, victims de facto become (private) enforcers of the guidelines.

Linking a negligence regime (with reversal of the burden of proof) with (non)compliance of guidelines in our view makes a lot of sense: by having those guidelines drafted jointly by hospitals, healthcare authorities and patient organisations with participation and full transparency, the potential influence of special interests (inherent in government regulation) can be avoided. And by considering a breach of the guidelines as negligence, healthcare providers obviously get excellent incentives to prove compliance with the guidelines (certainly in combination with a reversal of the burden of proof).

Chilling effects?

We assumed a system whereby ex ante guidelines on the optimal level of healthcare would be fixed, taking into account limited resources and budgets. However, a problem may arise when the guidelines are not precise, but unclear and vague. It is known from public authority liability that when due care levels are vague, this may lead to so-called chilling effects. ⁶³ The chilling effect could either amount to operators adopting (inefficient) high levels of care in order to avoid liability or operators abstaining from the activity altogether (with as a consequence that positive externalities generated by the activity would be lost as well).

As we indicated above, in the literature a distinction is made between negative and positive defensive medicine. Negative defensive medicine would lead to lower activity levels (the chilling effect indicated here), whereas positive defensive medicine would lead to an increased care level in order to reduce the risk of liability. The latter may more particularly play a role as far as individual physicians are concerned, especially if physicians could be held liable for not providing a particular treatment or medication, whereas it would be the hospital that would bear the costs of that treatment or medication. In the latter case, individual healthcare providers obviously obtain an incentive to prescribe more treatments or medications, not because this would be necessary for medical reasons, but simply to avoid a liability claim. ⁶⁴ As it is someone else (in this case the hospital) picking up the bill, it is a clear example of moral hazard. This behaviour of the healthcare provider can be explained by the difference between his private incentives (avoiding liability) and the societal interest (in reducing overly expensive treatment given limited resources). This divergence of interests results in the healthcare provider following its own, rather than society’s, interests.

The danger of chilling effects is, as indicated, also real and recognised as serious in the domain of medical malpractice. There, it is known under the heading of the above-mentioned defensive medicine. Defensive medicine generally is seen as healthcare providers not necessarily providing the most cost-effective treatment, but rather those treatments that could avoid them being held liable. The danger of defensive medicine has been well-documented and could certainly be an issue when a healthcare provider might fear that he would be blamed, for example for adopting the optimal (from a cost-effectiveness perspective) treatment rather than the most expensive one. It might theoretically lead healthcare providers to choose the more expensive treatment (just to avoid liability) although that could lead to social losses from a collective perspective (as the limited available resources would be used in a suboptimal manner). In an extreme case the fear for liability could also lead to an activity level change, that is, that physicians choose the type of activity with less liability risks or even leave the profession altogether. ⁶⁵ All this would obviously lead to a loss of the positive externalities generated by healthcare and thus to huge social losses.

As the source of chilling effects is the uncertainty concerning the guidelines, the answer is also relatively clear: guidelines should be formulated in a manner which is as clear as possible and, in cases where uncertainty exists, it would be important that healthcare providers are under an obligation to provide clear information on the choices made (e.g. why an optimal treatment rather than the most expensive one was chosen). This reporting in addition with the application of fault liability could keep the danger of chilling effects within reasonable limits.

Concluding remarks

We started from the problem which the legal system apparently has with the application of crisis standards of care in medical emergencies. We argued that the question of how to allocate resources (in economic terms how to determine optimal care) does not in fact only apply to crisis situations, but may emerge in any situation of limited resources. Given the limitations of public budgets, this may in fact be a general problem with (public) healthcare.

The solutions reached in (some) American states during COVID-19 of providing immunity to healthcare providers who were allocating resources in a healthcare emergency are, in a way, a binary approach as it leads to a total absence of liability of the healthcare provider (immunity) if particular conditions are met. That was one of the reasons why from a variety of legal perspectives, these so-called crisis standards of care were debated. We argue in this contribution that the problem of medical liability in a situation of limited resources is of a much more general nature and deserves therefore a more general answer to which, so we argued, a law and economics approach can help.

The advantage of the law and economics approach is that it points to the importance of identifying the goal of a (medical) liability regime. For the economic perspective, that goal should primarily be prevention, even though both from a theoretical perspective and given the limited empirical evidence available, other instruments, such as regulation, may be more suited to provide those incentives for prevention.

The crucial question, given limited resources, is how the optimal level of care (crucial in a liability setting) should be determined. One advantage of the economic approach is that it does not view care (and corresponding liability) as binary choices (either liability or immunity), but rather as a gradual decision-making process whereby, taking into account marginal costs and benefits of a variety of healthcare treatments, optimal care levels can be determined. Economic analysis indicates that in the determination of that optimal care level, limited resources do indeed play an important role, given scarcity of the resources and opportunity costs (care provided to one patient reducing the budget available for other patients).

The law and economics approach conducted in this article leads to some conclusions that can be useful for the current policy debate as well.

A first conclusion is that, as also indicated in the literature on medical liability, decisions on how to allocate scarce resources should in principle not be made by the healthcare provider at the bedside of the patient. ⁶⁶ These decisions involve complicated trade-offs that go beyond the information and responsibility of the individual healthcare provider and call for collective decision-making at multiple levels of society. ⁶⁷

Second, we argued that the best level to make those decisions is probably via multi-stakeholder agreements, including experts from healthcare, insurers, and governments, resulting in clear guidelines indicating how resource allocation and healthcare levels should be determined in a situation of scarce resources.

The major criticism on decision-making during COVID-19 was precisely that healthcare providers had to act without any protection or guidance in the absence of clear triage protocols. ⁶⁸ Clear guidelines not only have the advantage of providing indications to healthcare providers on how to act in cases of complicated ethical dilemmas; they may equally have the advantage of facilitating the applicability of the medical liability regime.

Also, as far as medical liability is concerned, several studies distinguish between either negligence or strict liability pointing to the relative advantages of both rules. ⁶⁹ We side with those who criticise an unbalanced application of strict liability for healthcare providers. That could indeed lead to both positive and negative defensive medicine, either increasing levels of care (e.g. unnecessary testing) with budgets which are already limited, or to a withdrawal from the market (in which case the positive externalities of healthcare would obviously be gone as well). As the chilling effects leading to defensive medicine are largely related to uncertainty concerning care standards, the guidelines mentioned above could, provided that they are sharply describing how to behave with limited resources, to an important extent provide certainty to healthcare providers and thus limit the dangers of defensive medicine. Moreover, guidelines could equally play an important role in the dichotomy between strict liability and negligence. Although strict liability indeed has serious risks, a pure negligence regime is not ideal either: the numbers indicate that it is virtually impossible for (the many) victims of medical malpractice to bring a successful lawsuit against a healthcare provider. The informational and procedural problems on the side of the victim could be alleviated by applying a negligence rule with a reversal of the burden of proof: healthcare providers who could show compliance with the protocols would thus escape liability. That once more underscores the role of the guidelines and also incentivizes hospitals to keep track of the way in which the guidelines have been followed in specific cases (thus reducing their liability risk). This regime holds the middle ground between the advantages of both strict liability and negligence and provides a smart mix between regulation (in the form of guidelines) and the liability regime. Through this smart combination, on the one hand ex ante better information can be provided to healthcare providers about the optimal care level under limited resources (which was exactly the problem under COVID-19) and the combination with the reversal of the burden of proof might make the liability regime as complementary to regulation more effective than it is today. In this manner, the guidelines combined with the medical liability regime might better serve the primary goals of the liability regime of both prevention and compensation.