Managing Demand for Health Care

Presentation to the Wellington Health Policy Seminar, 10 March 2000

The purpose of this paper is to consider how the demand for health care can best be managed in a world of constrained resources, and the implications this has for the complementarity between the public and private sectors. It also reports on some recent research on the New Zealand health insurance markets.

The search for sustainable financing options for the health sector is driven by the view that sufficient funding can not be made available from general taxation. This view, however, is not obvious to all. Various opinion surveys in New Zealand suggest that the general population would be willing to pay more tax (or to give up tax cuts) in order to put more resources into health funding. In theory, a comprehensive health system could be fully funded from within the existing tax system.

The key defining feature of traditional tax-based financing options is that individuals’ access to health care is not directly linked to their tax payments, nor do their tax payments vary with their level of health system use. Why should the fiscal funding be so constrained that the national budget is fundamentally unable to finance the level of health care desired by the population?

If the people who make decisions about health care (either individuals themselves or health care providers) do not face the full costs of their choices, their demand for services is likely to be much higher than what is socially optimal. This view is often disputed because people tend to think that health care is either needed or not. In fact, health care decisions are all a matter of degree. The main reason for this is the risk associated with each individual case of health care. Three main types of risk are relevant here:

• The risk that the symptoms may signal a more serious condition than is really the case;
• The risk that a more advanced intervention would produce better results; and
• The risk that an intervention may not work.

For example, if an individual develops common influenza symptoms, there is always a risk that the symptoms are actually signalling the presence of meningitis. Generally, meningitis is relatively rare and, from society’s point of view, it would be inefficient to test every individual with flu symptoms for the more serious condition. In other words, from society’s point of view, the risk needs to be managed. From an individual’s point of view, particularly when there is no cost involved, it may be more appropriate to eliminate that risk.

The same issue arises with respect to interventions. From society’s point of view, it is more efficient to try simpler, cheaper interventions first, as long as they have some likelihood of success. By contrast, individuals who wish to eliminate risks for themselves, may opt for more advanced, costlier interventions "too soon" – that is before the perceived social benefits of interventions outweigh their social costs. (This ordering of tests and interventions is based on the information logic of medicine – the results of each treatment give an improved set of health data to the doctors, who are in every case trying to solve the problem of what is wrong with the patient based on limited and often misleading information.)

To repeat the key point: individuals (and the medical providers on their behalf) who are not constrained in some way are likely to seek to eliminate, or come close to eliminating, the risks which, from the point of view of the society as a whole, should merely be managed. As a result, individual demand for health services is likely to exceed the socially optimal level of these services.

This mismatch between the socially optimal and individually optimal treatment of risk is likely to exist in any system of health financing. Only if individuals were to pay for all of their health care directly out of their own pockets (assuming could afford it), would individual decisions be relatively more likely to match the optimal level of risk management from a societal perspective. With any insurance-type arrangement, where the signals about the cost of risk management are muted ex post the payment of premiums, individuals are likely to demand more than is socially optimal. This is precisely why countries with managed and financially constrained health systems – such as New Zealand or the UK – are able to achieve the same broad level of health outcomes as countries with insurance-based health system – such as the US – at significantly lower costs. Insurance-based systems, which give individuals greater control over their perceived health risks, are likely to lead to more aggressive health interventions, and possibly to a greater level of satisfaction from the consumers who feel that "everything possible has been done".

One problem with solely fiscally funded system is that the mismatch it creates between individuals’ demand and the socially optimal level of supply is likely to be especially strong.

It is important to reflect a bit more on the distinction between the systems of risk management implicit in managed care and in reimbursement insurance. This is, arguably, a more significant distinction than between public and private care. Managed health care can be either financed and run by the government, such as in the public hospital system, or by the market, such as the Health Maintenance Organisations (HMOs) in the USA. Equally, an insurance system can be publicly funded, such as Medicaid in the US, or can be entirely private.

The main difference is that a reimbursement health insurance system provides individuals with a funding entitlement tied to their situation, which they can draw on to manage the risks as they see fit. By contrast, a managed health care system provides funding for a range of services, with individuals’ risks being managed by professionals who make decisions broadly on the basis of the social value of risk. Of course, in individual cases, professionals may be persuaded to adopt an individual patient’s perspective on risk, but overall, in a financially constrained managed care system, individual risk preferences are likely to be less significant.

Overall, managed care and reimbursement insurance provide two archetypes of health financing, with all health systems around the world falling into one or other category. Managed care is commonly financed through taxation, as typified by the British National Health Service, although HMOs in the US are an example of private funding. Reimbursement health insurance can either be privately purchased, as in the US, or can take the form of social insurance, typical of continental European countries. One important fact is that regardless of whether the health system is financed publicly or privately, countries which rely on managed care for their core policy tend to achieve approximately the same health outcomes by spending a significantly smaller proportion of their GDP on health than those which rely on reimbursement insurance. Similarly, medical professionals appear to be able to attain higher incomes under reimbursement systems, regardless of whether the professionals are public or private.

Countries with the managed care basic systems tend to have substantial add-on private reimbursement insurance. By contrast, it is difficult to find a country which relies primarily on reimbursement mechanisms where the market offers an additional managed care option. The US may be the only exception, in the sense that private HMOs are increasingly replacing the traditional private insurance option. However, the growth of the HMOs is largely linked to the employers’ (the main contributors) attempts to cut costs. The reasons for this contrasting trend are clear. Managed care systems, no matter how generous, leave pent up demand for individually initiated risk management. There will always be people who feel that a system built around social risks does not cater to their preferences, and who will be willing to pay to express those preferences. Reimbursement based systems already give expression to a wide variety of preferences, at a cost. Consequently, when people are locked into social insurance systems, they have no incentive to search for alternatives or add-ons. Alternatives arise only when people can opt out of the reimbursement system.

While it has been argued above that the essential features of managed care do not depend on whether it is funded by the taxpayers or privately, it is interesting to note that, apart from the HMOs in the US, such systems tend to be publicly financed. Is there a reason for this tendency? We think there is. Managed care does not rely on direct incentives for payers to create efficient risk management incentives. The key to managed care is the incentives for and the culture of health professionals. At the margin, direct payments by the patients may be of use from the point of view of efficiency. For example, prescription part-charges encourage the use of cheaper generic drugs and reduce waste. However, in the great scheme of things such direct incentives are almost insignificant compared to the efficiencies achieved through socially oriented risk management.

There are two key preconditions to achieving socially accepted risk management objectives: the budget available to providers must be capped, and the expenditure must be targeted at national level outcomes and priorities. Tax-based systems are uniquely suited to that. While funding the health care from general taxation removes any incentive on the taxpayers to minimise consumption of services, this is offset at the macro level by the stringent controls imposed through the fiscal constraint.

In general, to achieve efficient levels of risk management, it is important that the core of a national health care system be built around managed care principles. Equally, there will always need to be an add-on insurance system, which would allow people with particularly strong risk aversion, or particularly strong time preference, to manage risks as they wish. Again, so far in this argument, it is not important whether the core or the add-on are publicly or privately funded.

In fact, this is illustrated by the debate currently taking place in the US about the legal liability of HMOs. Under the present American law, HMOs are exempted from liability for medical malpractice (individual doctors working for HMOs are still liable). This allows such organisations to be relatively aggressive at managing risks – in effect, denying care in situations where an insurance firm would have authorised treatment. This has become an important political issue, with a recent opinion survey finding that 60% of voters rate it as a significant concern for the next congressional elections. Analysis would suggest that if the restriction on liability is removed, and the HMOs are forced to adopt the risk management standard of the insurance industry, then the whole value of having such organisations would be defeated.

The problem is that there is always going to be tension at the boundary between this core – where by definition individual preferences about the elimination of risk can not be satisfied – and the additional insurance which is primarily designed to allow individuals to manage risks according to their own perceptions and preferences. There are two problems that need to be addressed.

First, there is no easy way to decide what the optimal size of the core should be. In particular, the optimal size of the managed care system probably cannot be determined from the level of individuals’ dissatisfaction or public complaints. In fact, some level of complaint is evidence that the system is working as intended. Decisions about the size of the system require macro level assessments about whether the risks are being handled in a socially optimal way. The sorts of questions that need to be asked are:

• What is happening to the health status of the population?
• What kind of risks are we willing to accept as a society? For example, what is the net social benefit from investing in facilities and technology that would reduce mortality from, say, heart attacks by 1%?
• The existence of risk means that bad things will happen. Are negative events (such as lack of treatment) consistent with the desired level of risk, or is the system mishandling risk?
• How quickly do we want health risks to be reduced at the social level, and how much are we prepared to pay for it?
• In the political context, how wide a gap between socially optimal risk management and the individual desire to eliminate risk can be sustained?

Secondly, individual transfers from the core to the add-on will always be fraught with tension. Assuming that there are some costs associated with transferring into an add-on insurance system – either financial or administrative – individuals will always have an incentive first to lobby for the managed care system to satisfy their preferences, and only make the transfer, if nothing else works from their point of view.

In a sense, an add-on insurance system is a safety valve for the managed care core. Its role is to deal with some of the gap between demand and supply that the core must generate if it is to be efficient. The problem is that the safety valve can not just flap open to every demand. Its value is in controlling some steam build-up, and in only letting some of it out. But how much?

There are two possible ways to answer this question. The first is through policy analysis, by answering the kinds of questions that were outlined above. The second is by finding a way for those whose pressure is going to be the greatest to self-select their way around the system.

One possibility is to think of a three tiered system:

• A publicly funded core: managed care system;
• An expanded managed care system that people can opt into by paying an annual service charge, which offers a bit more in terms of time management and facilities options; and
• A private insurance system which allows people to manage risks according to their own preferences.

The choices that people make in selecting into one of the three tiers would provide signals about the adequacy of funding for the core. Transition between tiers will be considered in discussion of entitlements later in the paper.

As argued before, once one accepts that the core of the health system must be managed care, it is easy to see the logic of having that core funded publicly. Public funding will be efficient precisely because it is severely constrained, and the focus of the public system is dealing with increasing demands from existing resources. The core of the health care system must also address complex equity issues and deal with social priorities, which is best done via a politically responsive arrangement. While much is wrong with politically determined resource allocation processes, a society which seeks both equitable and efficient outcomes may be unable to avoid them. Ideally, we may wish to find a way to fund the core health care that avoids the idiosyncrasies and the political tensions of the present battle for the health budget. However, such an ideal does not exist.

Once the core is set, private funding for any add-ons makes sense because it no longer needs to address the equity concerns. When people choose to pay insurance premiums or an annual service charge, such payments do not have the negative supply side effects of general taxation because there is a direct link between what people pay, and the services they receive.

Thinking of entitlements in terms of different approaches to risk management can help us think through various options for making transitions between the three tiers described above. The transition to the third tier – the insurance market – is relatively easy to understand. It would most likely work in the same way as it does now. Private insurance would attract individuals who have a very high rate of time preference (ie, those who are willing to pay additional amounts to ensure that procedures are scheduled at their convenience), who are extremely risk averse (ie, they are willing to pay for an opportunity to manage their health risks according to their own preferences), and who have a high preference for controlling their environment (eg, they may wish to choose their own doctor).

The transition between the two tiers of managed care is likely to be more difficult to define. A number of options can be considered:

• By type of risk: obviously, any system of managed care must be able to deal with the key health risks affecting the population, including any life threatening condition. It is also hard to imagine New Zealand society accepting different social targets, in the sense, for example, that the first tier would aim for one rate of mortality for a particular condition, while the second tier would add for a lower rate. However, there are other risks which may be addressed differently, including the risk of inconvenience and of personal discomfort. While it sounds callous, this kind of differentiation already takes place. Doctors make rationing decisions that affect such outcomes for people. Welfare may be improved by allowing people to move between tiers defined by explicit rationing criteria.
• By type of procedure: many medical procedures address the quality of life, rather than directly manage life risks. For example, an injured sports person may need an advanced operation to return to the sporting career, but is able to function otherwise. In this context, people may wish to make transition between the tiers on the basis of the life-style choice. People who place a higher value on particular life-style options would be willing to pay more for a managed care system that was willing to help them maintain those options.
• By how the risk is managed: the same health risks may be managed in different ways. For example, one option to deal with accidents and emergencies in remote locations is to build hospital facilities, another is to provide helicopter evacuation services to a big city hospital. People may have a strong preference for one or other of these options. The first, publicly funded, tier could be envisaged as aiming to manage risks in the least expensive way, while the second tier could give individuals the option to pay for the more expensive alternative.

Would the above mean that people on higher incomes will always opt into the second or third tier, leaving the first tier to provide an "inferior" service to the poorest members of society? Not necessarily. For example, people with children in isolated communities may have a preference for a local service, while single individuals may have no objection to travelling. There is no reason to expect that families with children would have higher incomes than single individuals, but they may be more willing to pay for the option of being treated locally. Similarly, people with less skill who perceive themselves as more at risk in the labour market may place a greater emphasis on the speed of treatment than more skilled people with stable jobs.

The same individuals are also likely to make transitions between the three tiers during different stages in their life times. For example, people may decide to place less value on time once they retire.

So how has the existing multi-tier system been evolving in New Zealand? Recent research on the New Zealand market for private health insurance using the Household Expenditure Survey has shown that the proportion of households reporting expenditure on health insurance has been fairly constant for the past 10 years. However, between 1995 and 1998 it declined. Four possible explanations for the decline have been considered:

• Access to private health care facilities – did those people who were at some distance from private health care facilities reduce their cover?
• Affordability of health insurance – have people tended to reduce the amount of cover they purchase, or have those in lower income groups tended to let their policies lapse?
• Ability to manage the risk internally – have people reduced their cover because they consider it to be cheaper to manage the risk of illness themselves, making out-of-pocket payments when necessary?
• Access to public services – have public services improved to obviate the need for private cover?

The first explanation has been largely discounted. Surpisingly, it was Aucklanders who were reducing their cover. Accessibility is unlikely to be an issue for those living in Auckland. Furthermore, the number of private hospitals has increased since 1995 suggesting that the availability of private facilities has, at the very least, not been reduced. At a more general level, accessibility may be an explanation for the continuation of above average levels of health insurance in Auckland.

Affordability concerns are more likely to prevent a household in the lower income groups from using their limited discretionary income to buy health insurance. Indeed lower income households do buy less cover, in general. Yet the decrease in insurance uptake in Auckland has been surprisingly concentrated in the top income bracket. This suggests that although affordability may be part of the story, something else is driving high-income households to let their cover lapse. If affordability were an issue, households may choose to reduce the amount of cover they purchase. Yet the average amount spent by households which still have insurance has remained relatively constant. One limitation of the analysis is that it is not possible to determine the extent of churn, or entry and exit from the market. If the length of time that particular households remained in the market was brief, this could suggest that they treat medical cover as highly discretionary, and probably strongly income dependent.

Another explanation for the decline in health insurance is that households have decided on the basis of balancing cost, risk and return that they would prefer to manage the risk of expenditure themselves. One possible cause of this could be employers moving away from subsidising health insurance as part of remuneration packages. This would increase the direct cost to the household, altering the balance of risk and return. This may explain why high-income households (who may have been more likely to have had such packages) have reduced their cover. This suggests that a more complex problem exists than straight cost concerns, involving the appreciation of the risk of a medical event, and the potential expense involved.

Another interesting fact is that, in recent years, there has been a slight shift of government funding toward the Auckland region. It is possible that this has encouraged Aucklanders to put their trust in the public system. If that were the case, it would suggest that public expenditure has been poorly targeted, with benefits disproportionately accruing to higher income households, which have discontinued private insurance.