1 Introduction

Osteoarthritis is a major cause of hip or knee pain. The main treatment available is joint replacement, an effective and widely used but expensive intervention for which there are no evidence-based criteria for the indications or timing of these procedures [1]. Some smaller studies suggest no significant differences in outcome between age groups, obese and non-obese people, pre-operative functional status or underlying disease [2, 3]. In contrast, larger observational, registry databases and economic evaluations indicate that these factors plus level of social support and education are significant predictors of outcomes and of cost [4, 5]. Economic evaluation can provide a useful framework to analyse and interpret data to inform decisions about whether joint replacement should be available and who should be treated. The methods of economic evaluation are described in detail elsewhere [6]. This paper outlines some of the key issues that should be addressed by an economic evaluation.

2 How Does Economic Evaluation Define Whether Joint Replacement Is Worthwhile?

In general terms, a health care intervention is worthwhile if the value of the outcome, in terms of gains in health and social well-being, is greater than the value of the resources and services used to produce or provide that intervention. More specifically, if the decision has been made that treatment should be provided for a health condition, such as hip or knee pain, the question an economist asks is: What is the most efficient method of producing a gain in health and social well being for this condition? In this context, efficiency is defined as maximising the gain in health and social well-being associated with an intervention, such as a hip replacement, for a given budget.

3 What Should Joint Replacement Be Compared to?

Any economic evaluation to inform questions of technical efficiency requires that the intervention of interest, in this case joint replacement, is compared to an alternative. This is so that the additional or incremental costs and benefits of the joint replacement compared to the next best alternative can be assessed. The alternatives for comparison may include no joint replacement and usual care for the management of pain, reduced mobility, etc. An alternative may include more active intervention with physical therapy, manipulation or psychological therapy to manage or improve mobility and reduce the impact of pain, anxiety or distress.

4 Framework of Analysis

The main frameworks of analysis for economic evaluation are cost-effectiveness analysis, cost–utility analysis and cost–benefit analysis. All these frameworks require that there is a comparison of costs and outcomes between two or more interventions. The key difference is in the method of measuring and valuing outcomes. Cost-effectiveness analysis measures health gain using clinical outcome measures, such as life years gained or improvements in pain or in mobility. However, cost-effectiveness analyses do not combine different measures that are potentially important to patients and health care decision makers. For example, in the decision to provide a hip or knee replacement, surgeons and patients will need to trade the potential benefits of surgery against the risks of serious adverse events, such as surgical mortality, deep wound infection or failure of the prosthesis and the costs of the procedure. These aspects also need to be compared to the benefits, risks and costs of no surgery, or alternative interventions. The trade-offs of benefits and risks will be implicit if cost-effectiveness analysis is used to provide information about the relative value of joint replacement compared to available alternative interventions.

In contrast, cost–utility and cost–benefit analyses combine preferences for the matrix of benefits and risks that affect and health and social well-being into a single measure. Preferences for length of life and consequences, such as pain, reduced mobility, or infection that affect quality of life are combined by explicit trade-offs between each of these attributes. This allows the derivation of value weights that can be used to produce a summary measure of outcome. For example, consider (a) a year spent in full health with no problems. Compare this year of life to (b) a year spent with moderate pain or discomfort, no problems with walking about, self-care or usual activities and not feeling anxious or depressed, and to (c) a year of life spent in extreme pain or discomfort, with some problems in walking about, some problems with self-care and performing usual activities, and not being anxious or depressed. A typical ranking of these three health states would be state (a) full health as most preferred state, (b) as the next preferred option and state (c) as least preferred.

Now consider three additional health states: (d) a year of life spent with moderate pain or discomfort, some problems with walking about, self-care and usual activities and feeling extremely anxious or depressed whilst waiting for a hip or knee replacement; (e) a year of life spent in extreme pain or discomfort, with some problems in walking about, no problems with self-care and performing usual activities and feeling moderately anxious or depressed whilst waiting for a hip or knee replacement; and (f) immediate death due to peri-operative complications during hip or knee surgery. The ranking of health states (a)–(f) is not clear cut. For some people, full health (a) will be the most preferred and death (f) the least preferred. For others, health states which include extreme pain or discomfort or extreme anxiety or depression may be worse than death. The ranking and valuation of health states that may occur also needs to consider the probability that they will or will not occur. Consider an intervention where the probability of death (f) was 2% and the probability of full health was 98%. Compare this to an intervention where the probability of death was 0% and the probability of health state (a) was 100%. Which would you prefer? Cost–utility analysis and cost–benefit analysis seek to evaluate the trade-offs between complex sets of benefits and risks by addressing the following questions. First, on average, what is the order of preferences for health states (a)–(f) and second, by how much is one health state preferred to another?

Cost–utility analysis uses a non-monetary measure of value, known as utility, to address these questions. The utility weights are multiplied by length of life to produce a measure of quality-adjusted life-years (QALYs). Cost–benefit analysis seeks to provide a monetary measure of the relative value of the different benefits and costs. The methods used to derive these utility weights and the relative merits of the methods are described in detail elsewhere [6].

5 Whose Costs, Outcomes and Values Are Important in Choosing Joint Replacement?

From the economic perspective, an evaluation of the efficiency of an intervention should include the costs of all resources used to provide care and all the outputs or consequences of the intervention. This is termed the societal perspective and encompasses more limited viewpoints, such as that of the hospital or secondary care sector responsible for providing the joint replacement services, or the third party insurance system (public or private) responsible for payment for the services used. Resources or services from a variety of sectors will be used as inputs to provide care and support prior to, during and after treatment. Care after the intervention will include longer term follow-up and management and treatment for any adverse consequences as a result of the intervention.

In the case of hip or knee pain and joint replacement, these will include: first, primary care services, such as GP visits for referral to secondary care or pain management; second, community or social care to provide support, equipment and aids for reduced mobility or problems in self-care and usual activities prior to joint replacement and rehabilitation following the joint replacement; third, hospital inpatient and outpatient services to assess the need for surgery and the patients suitability in terms of the risks and potential benefit of a joint replacement, inpatient care for the joint replacement and post-surgical care, treatment of adverse consequences of surgery, longer term follow-up and medication; and fourth, the costs to the patient and family or friends in terms of time spent in care related activities and additional expenditure. This may include time spent by the patient and carers to participate in exercise programmes to regain mobility or lose weight before or after the operation and time spent attending hospital services or other rehabilitation programmes. Both the patients and carers may incur additional expenditure, for travel costs, aids and adaptations to the home, social support or residential care and medication.

As noted above, the outputs will include benefits in terms of improved health and social well-being and also the impact of reduction health and social well-being associated with any adverse consequences of care. As with the inputs to care, the outputs to a number of viewpoints are potentially important. These include the impact on the health status and well-being of patients and of family or friends whose health or social well-being may be affected by providing a joint replacement. In addition, the provision of a joint replacement may have a less direct impact on the health and social well-being of health and social care providers or society in general. This may include feelings of satisfaction or well-being that care is provided to those in poor health, or distress over the impact of any adverse consequences of care (termed as externalities). Finally, joint replacement surgery may have a value to the economy of patients or carers who are able to participate in activities that result in the provision of additional goods or services (termed as indirect costs or benefits).

Whose valuations should be used to derive a summary measure of outcome will depend in part on the decision to be taken. If the decision is whether joint replacement surgery should be available to groups of patients or in a community, then the values should reflect a societal perspective. In this case, utility or monetary values of preferences for health states should be derived from a representative sample of the overall society or community in which the decision to provide joint replacement or care for hip or knee pain is to be taken. This will include a proportion of patients, family or health care professionals concerned with the management of hip or knee pain as well as those who are not directly involved. However, if the decision is whether a specific patient should receive joint replacement surgery then a more limited perspective is appropriate. This would include the patient, their family/friends whose health and social well-being are directly affected and the health professionals responsible for providing care to that patient.

6 Who Should Receive Joint Replacement Surgery?

Economic evaluations do not currently incorporate distributional concerns [7]. In economic evaluation, measures of the value of health gain are aggregated to provide a population or societal estimate of value. A key assumption underlying the aggregation is that the social value of a unit of health gain is equal between individuals and across all population groups and that the value is constant over time. In addition, economic evaluation favours those with the greatest capacity to benefit, such as patients more severely affected by osteoarthritis. Using this method of evaluation may systematically discriminate to the advantage of certain groups against others. Inequalities in health between individuals and/or groups of individuals may increase as a result of seeking to maximise health gain.

It has been argued that if measures of health gain such as QALYs represent close approximation of utilities they will incorporate a measure of value for the distribution of health and social well-being and support the maximisation of health gain [7, 8]. However, concern for others’ well-being and concepts of justice mean that there are societal preferences for the distribution of that health gain. These differences in valuations may be due to preferences for equity in the distribution of life, life years gained, quality adjusted life expectancy or “just desserts”. The latter may be based on a sense of individuals taking some responsibility for engaging in health-promoting behaviour or that individuals who for whatever reason are already disadvantaged should not be made worse off. For example, recent research indicates that the value of health gain varies by patient characteristics such as age and severity of illness prior to care rather than quantity of health or health gain alone [9, 10].

In the case of joint replacement, current utilisation data and predictors of outcome for indicate that decision makers may value equal health gains differently between eligible population groups. Hospital statistics for England and Wales show wide regional variations in operative rates, implying geographical inequalities in access, and American data suggest the elderly, the obese and blacks, are less likely to get a total joint replacement than middle-aged, middle class whites [1]. Some reports also suggest that there is a large unmet need, particularly for knee replacements [1]. This appears to contradict the principle of equal access to or equal treatment of equals, where the determinant is equal need (horizontal equity). However, it could be consistent with horizontal equity if the definition of need includes socio-demographic or economic characteristics as well as the more traditional clinical, functional and disease severity measures. Alternatively it could be consistent with an equal distribution of health or health gain across population groups and/or generations (vertical equity), which would require unequal treatment of unequals.

A recent review suggests that to take distributional concerns into account both individual and societal utilities for health care should be reflected more fully in economic evaluations [1]. It is possible to incorporate concern for equality into cost-effectiveness analysis by using equity weights that represent values for preferences for the distribution of gains in health and social well-being. The greater the equity weight the more society is willing to sacrifice a unit of health gain in pursuit of fairness.

7 Summary

In summary, the economics perspective indicates that decisions about whether joint replacement surgery should be available for a population or used for an individual will require evidence about the relative costs and outcomes to the society in which the decisions are taken. This will include those sectors and groups that are responsible for the provision or funding of time, staff, equipment and facilities used as inputs to produce care and those whose health and social well-being is affected by the provision and outcomes of care. Outcomes should include the direct impact on health status and well-being of the patients and others, and measurement and valuation of preferences for those outcomes. Ideally, the evidence should also incorporate preferences for the distribution of the costs and outcomes between sectors and groups of individuals.