Economic Evaluation of an Investment in Medical Websites and Medical Web-Based Services

Valuing Human Life

Brent (2003) highlights the point that proper decision making and resource allocation is needed because even if all the expenditures of healthcare systems are vital, there is no way so many resources are found to meet the demand but only through evaluation process can anyone be sure that the system works effectively and is as safe as possible for the society. Also Brent (2003) pinpoints the paradox that economists and in general analysts who apply the valuation method have to think in and use finite values while human life has infinite value. So significantly Brent states that the purpose of every valuation in healthcare method is to find which method, process or action is more socially desirable.

Card and Moorey (1977) emphasize also the fact that although human life has infinite value and despite the fact that any attempt to attach a monetary finite value to life seems to be cynical and unethical, this is the only way to apply any valuation method and improve the quality and rationality of decision making in healthcare. Making the infinite value of human life finite secures coherence and stability in the decision making process and in services delivered by any healthcare unit or system in general. This opinion is supported also by the logical argument that if all resources were allocated to save a single life this would result in a great loss of healthcare benefits for billion people and the potential loss of the ability to save many more lives. Card and Moorey also propose three methods to evaluate human life: Productive Capacity, General Implicit values and Individual values.

According to Card and Moorey (1977) productive capacity is based on the idea that everyone offers to society through his productivity and he is compensated for this with an amount of monetary units/earnings. If these earnings discounted to their present value and other factors such as, expected productivity or expected working life, are taken into account, then the result is a value of human life. Although this approach sounds possible, this method is criticized for its drawbacks, that it does not take into consideration the view of the patients whose life is in danger and also that it ignores that the major objective of a healthcare unit is to improve the quality of human life and not only cynically enhance production for the society. The second approach involves the adoption of a mean value from previous trials in order to evaluate human life, but here different conditions and implications of every past value must be thoroughly examined. The third method, which is based on the individual’s opinion about what is the value of the human life, is the most insecure and subjective but combined with the other two can result in a quite accurate measure.

While attempting to compare Cost-Efficiency Analysis and CBA, Garber (1999) states also that CBA requires to measure human life in finite values which raises many ethical issues. He emphasizes the fact that information in the healthcare sector is imperfect and often asymmetric, and this asymmetry is due to the fact that anyone involved with the healthcare sector (producer, patients, working force, tax-payers) have different access to information about the characteristics of healthcare and this can cause issues difficult to overcome while trying to apply CBA. He also strongly claims that CE is preferred in measuring the value of healthcare interventions because of the fact that it avoids to measure healthcare effects in monetary units which is something that CBA does. Moreover, Sugden and Williams (1978) state that one might argue that the value of life in terms of the Pareto improvement criterion is the amount of monetary units an individual is willing to pay to prevent a premature death. On the other hand it can be argued that this is the compensation an individual would accept for his own death. Other suggestions presented are that the value of life equals to the costs imposed to others from one death. They propose that a more obvious approach to value human life is to identify situations where individuals have a choice between safety and extra monetary gain and the value they put ton their own safety, this way researchers/decision makers can suggest that the minimum value of life is equal to the monetary gain an individual needs in order to live/work/interact in a completely unsafe environment.

Quality Adjusted Life Years s, Health’s Years Equivalent and Disability Adjusted Life Years

The QALY approach is the most widespread approach used concerning the evaluation of healthcare intervention outcomes. According to the National Institute for Clinical Excellence’s (NICE) (2010) the quality-adjusted life year (QALY) is a measure of disease burden, based on the number of years of life that would be added by the intervention and it is used to rationalize the benefit from different medical treatments or procedures, so as to calculate relative cost-benefit. NICE (2008) refers to QALY as the most appropriate generic measure of health, recognizing at the same time that alternative measures exist (for example, the healthy-year equivalent), although few economic evaluations have used these methods, and their strengths and weaknesses are not fully established. Culyer and Wagstaff (1993), support the idea that QALY have become a valuable instrument for measuring a healthcare intervention outcome, serving well medics and decision makers. According to them, although QALY constitute is a very useful approach, which however suffers from its own drawbacks. One main criticism about the use of QALY is that QALY do not present always consumer preferences, and this is an issue that might result in wrong, not according to public interest, decisions.

Garber (1999) argues that typical measures of healthcare intervention outcomes are “years of life saved” and QALYS. Each additional year an individual lives raises the life expectancy measure, but in case of QALYS the amount with which each year of life raises life expectancy is calculated as a weight that can take values between 0 and 1 depending on the health status of the individuals during that year. QALY are measured by first calculating life expectancy, which is considered to be the sum of probabilities that an individual will be alive at each age of his life, in the future until his/her maximum life span. Then to calculate QALY from the life span, weights are applied for possible health states and a discount factor. The main problem for an analyst is to calculate the probabilities and to adjust the weights for the different health states. The first step in calculating QALY is the description of the health state the second step is the assessment of weights for the health state, and the estimation of the probability of that health state, which usually involves elements of modelling. He also emphasizes that the most difficult practical issue concerning QALY is the variation that can be observed in preferences, and he suggests that even the “theoretically best” treatment can vary when the preferences of the individuals vary. To solve this problem, it is suggested to use statistical weights rather than weights based on preferences.

On the subject of QALY, Culyer (1989) suggests that the QALY approach rejects “wellfarism”, which suggests that social welfare

e depends on the utility of the individuals, focusing more on the individual’s characteristics. Finally, Nord (1992) highlights that this approach has some drawbacks; e.g. such as it expresses quality of life per year as a number, which is difficult to be explained to potential users and is only having a significant meaning. Also the way the various health states are evaluated raises many ethical issues. Moreover this approach emphasizes only the healthcare effects and improvements, ignoring the condition of the patient before or in the long term after the intervention, and in general the QALY approach focuses on quality of life in life years and does not focus on the quality of life for the people. Nord (1992) further attempts to propose an alternative procedure that it seems to be more a complementary tool to QALY approach. His suggestion is to try to measure one particular effect by letting people compare it with a benchmark effect of healthcare, which in this particular case is the effect of saving a young life (SAVE). The most vital aspect of this alternative procedure is to find out what the number of beneficial outcomes of a potential program or healthcare intervention should be to be considered enough to get a SAVE unit. Although this proposed method, has the advantage that it takes into consideration distributional or ethical issues, measures social value and can help especially in decision making concerning resources allocation, its reliability should be studied further. Because of this drawback and the practical problems of the application of this process, also due to the large number and the complexity of the outcomes from healthcare interventions it is suggested that this method is better not to be used as alternative to the QALY approach, as Nord proposes, but as complementary procedure, one more support tool.

Health’s Years Equivalent (HYEs) can be considered as an alternative to the QALY, but their use has generated a lively debate between academics. Bleighrodt (1995) argues that HYEs as well as QALY, assume that health and non-health attributes are independent in the value function of an individual. In comparing QALY to Health Years Equivalent (HYEs), Culyer and Wagstaff (1993) state that HYEs engulf the time preference and risk attitude of the individuals and thus they are not to be adjusted as QALY,. In addition to this, if QALY are calculated under the assumptions of HYEs, analysis held equal results. Also they suggest that QALY calculated by using the time-trade-off approach are completely equal to HYEs.

Another useful approach in measurement of healthcare outcomes is Disability Adjusted Life Years (DALYs) approach, Murray and Acharya (1997) state that DALYs are used to measure the burden to a patient from a disease and the main aim of this measure is to enhance the ethical dimensions in healthcare evaluations and improve the attempts to quantify many other measures in healthcare. They highlight the fact that DALYs are not an indicator of utility lost because of the current health state and it cannot be used to assign a social preference according to individual preferences over different healthcare states. Their primary use is to help analysts quantify effects and express them in the same units. They also argue that the main motivation for the creation of a measure like the DALYs, was to deal with the problem of identification of inefficiencies in the allocation of resources and healthcare services through cost benefit/effectiveness methods.

According to Murray and Acharya (1997), allocation inefficiencies are caused due to:

• 
  Inequalities disparities in earning of individuals so that in systems which demand a percentage to be paid by the individuals, richer people will have access to more expensive services.
  
• 
  Moral hazard and adverse selection mainly appeared in the health insurance market.
  

Discounting of costs and benefits

Sugden and Williams (1978) highlight the fact that sometimes important decisions involve not only the reserve of valuable resources but also costs that incur during the present period or in the future and result not in present instant benefits but in benefits in the future. This creates a problem related to the time dimension in decision making and evaluation process. According to these authors, in order to deal with this time dimension issue, an analyst should take into consideration the time-consumption-preferences of the individuals, which is the preference of individuals between consumption at different time periods. This raises the question among the scientific community whether benefits should be discounted as costs and questions about the appropriate discount rate both for costs and the effects of healthcare interventions.

One approach they propose, in order to solve the problem of how to select a social discount rate for public projects, is to seek for consistency between the criteria used to evaluate projects in the public sector and the private sector of the economy. Another approach is the time preference approach. They suggest the best solution is a synthesis of these two methods. Other problems are shot-slightness and the transferability of the effects to future generations.

Arrow and Lind (1994) question whether it is appropriate to treat public investment, concerning discounting in the same way as private investments. If risk is treated differently in the public than the private sector, “this can lead to overinvestment in public sector, in expense of private investment yielding higher returns”. On the other hand, they state the opinion government can spread and diversify risk in a more successful way, since it has a pool of large number of projects with different risk grades than the private sector, so government investments should be treated differently than investments in private markets. One third opinion, presented by Arrow and Lind (1994), states that the discount rate, and in general risk attitude, should be specified by an appropriate authority which would take into consideration risk and time preferences rather than individual preferences. According to Hirsheiler (1965), in the case the government diversification of risk is effective, government should take into consideration the more productive private investments. He also argues that the cases in which risk is treated differently between public and private sector are only justified, if it is very difficult to transfer the advantages of government diversification to private investors.

Brent (2003) tries also to deal with the difficult of CBA in miscalculating discount rates. Brent presents the argument that both costs and benefits must be discounted in order to end up with the interpretation of reasonable results and supports the idea that the discount rate for costs should be derived as a concept that will indicates the amount of money the society prefers to cut from a future time, in order to have the use of that amount of money today. This is called social time preference rate (STPR). Equally Brent (2003) supports that discount rate for benefits should be interpreted like life expectancy discount rate (LEDR) which derives from how society values an extra year of life in the future to a year in the present.

Based on a sample of studies from National Health Service Economic Evaluation Database (NHSEED) Smith and Gravelle (2000a) observed the practices related to discount rates of costs and benefits. According to, these results only one source, the English Department of Health, proposes a different discount rate for costs and benefits i.e. around 6% for costs and 1.5%-2% for benefits, and they found also the remarkable result that more than 90% of their sample researches were discounting costs and health intervention effects, using the same rate, whose value was between 0% and 7% with 0%, 3%, 5% being the most used values. Amazing was also the fact that almost 30% of their sample did not discount costs at all, which appears somewhat strange practice with unjustifiable results, because costs are almost always expressed in monetary values and value of money changes over time, so discounting is crucial in order to have realistic and reasonable results.

Moreover, CHE technical series 20 (2000b), it is reasonable stated that healthcare effects grow over time, arguing that value of healthcare effects is affected from an individual’s age and the time period when these effects occurred. The researchers in this report applied two models one individualistic and one socialistic; the socialistic model being the one more relevant to this research. The conclusion is that if any valuation is to be used, there must be a disclaimer about whether the researcher took into account the assumption that healthcare effects change in value aspects over time or not. Also it is argued that for CBA or CEA, if the above assumption is made, healthcare intervention effects must be adjusted either by adjusting the effect itself or by adjusting the discounting rate. Specifically for CBA, it is proposed that all the effects from a healthcare intervention must be valued in the income of the period during which they occurred, and then, through the application of a discount rate, to be discounted to their present value.

On the other hand, Keller and Cretin (1983) present a more comprehensive argument supporting the view of equally discounting costs and hard to monetize benefits. They argue that discounting cost, but not effects at all, boosts cost effectiveness ratios for all the issues under evaluation. The logical argument is that if also benefits are in monetary units, then the researcher should use the same discount rate as all kind of money should be treated in the same way. Keller and Cretin (1983) are using examples which yield some interesting results when different discount rations are used for costs and benefits/effects.

They concluded that discounting costs but not benefits can hinder any decision making or even block any decision making process, since for any attractive programme, there will be always a more cost effective delayed programme which should be funded instead (“postponing paradox” due to discounting). On the other hand, using a discount rate for benefits/effects, but lower than the discount rate used for costs will not necessarily result in the postponing paradox that will paralyze decision making process, but this practise can result in inefficiencies in budget allocation. A decision not to fund a programme today because it is not cost effective, may result in the allocation of these resources in the future on even less efficient programs. Moreover, they examined the case in which discount rates of benefits are higher than those for the costs. This case is based on the argument that there is much more future uncertainty for the benefits than the costs. The results of the implementation in this case where a defensive approach implying that the longer the delay the more efficient the programme should be in order to be funded. Keller and Cretin concluded that using equal discount rates for costs and benefits is the only way to justify any cost- effectiveness analysis.

According also to the National Institute for Clinical Excellence (NICE) guideline (2004, 2008) the discount rate for appraisal methods in general and more specific cost-effectiveness analysis should be 3.5% both for benefits and costs. So NICE proposes one discount rate and not two different (one lower for benefits and one higher for costs as other studies suggest) mainly because the guideline advises also healthcare intervention’s benefits to be expressed in QALY so not to be interpreted in monetary units. Brouwer et al. (2005) strongly criticize this practice because QALYs and in general healthcare effects, do not stay stable but are sensitive to changes in the future and the “postponing paradox” argument seems illogical in real-world practice. Thus, as healthcare intervention effects change over time, their price must be adjusted according to their rate in order to have a valid evaluation result. Brouwer et al. (2005) think that this decision and in general the adaptation of a single discount rate for both effects and costs, is anachronistic and cannot be considered as step forward and may negatively influence not only NHS, but also other health systems.

Cost type selection, estimation of social value, uncertainty and Willingness to Pay approach

Among the scientific community, there is also a debate going on what type of cost is most appropriate to be used for better results in CBA. Extant literature also highlights the dangers concerning the correct identification of costs and effects for the analysis.

Birch and Donaldson (1987) highlight one of the major dangers concerning the application of CBA, the risk to misidentify some costs as benefits so that the real benefits are ignored which can lead to decisions that are not according the social preferences. They pinpoint also the fact that during the selection of a programme, the analysts should take into consideration not only the efficiency of the programme, but also the associated budget constraints, otherwise the programme selection will be completely unrealistic. They proposed a procedure for finding the optimal budget allocation having calculated the budget costs and net benefits. This is based to the assumptions that there is a fixed budget and that the projects are indivisible and independent one from the other. However, the main problem with these assumptions is that in general programmes often interact, and are not completely independent one another, and so can have different benefit value if they applied together, than in the case they are applied independently. For example, programmes that encourage people to adopt activities that improve their health, programs about quitting smoking and programmes about cardiovascular improvements, if applied all together, can produce a greater total social value and have better outcomes. Also another point is that projects can always be reformed in order to fit a budget.

Garber (1999) discusses the choice of which measure of costs (average, marginal or total) should be used. Taking into consideration market imperfections and the cases where fixed costs are significant, we end up with the conclusion that these measures of cost can vary a lot. It is suggested that the measure of total costs whether born by patient or others involved in the healthcare sector should be chosen to represent costs. As in many cases, e.g. in NHS, government has both the role of provider and consumer some governments may take into consideration for the analysis only the costs they can handle for healthcare. But this may lead to decisions that are far from the main aim of government and healthcare systems which is to serve people. He also suggests that because a healthcare intervention can result in changes concerning the costs and the effects /consequences from its application, in the long run it is advisable that all future costs and effects should be taken into account when a Cost-Benefit or Cost-Effectiveness analysis is applied.

Sugden and Williams (1978) strongly support the idea that there are many factors that make the use of markets values in general inappropriate to be used in a CBA for measuring the marginal social costs or marginal social values. To deal with this problem, they suggest that CBA should use some values called shadow values or accounting, prices which are not taken directly from the market. In a competitive economy, if a buyer can buy as much as he wants of a good at its market price, this price can be used to measure the marginal value of the good or service to the consumer. Moreover, if producers or service providers can supply whatever quantity of a good or amount of services they offer for sale at its market value, this price can be used to estimate the marginal cost of producing the good. Only in these cases or under these assumptions, an analyst can consider market values significant to measure effectively social value. But in the real world, according to Sugden and Williams (1978) this is not the case due to taxation that adds to the market price of a good or a service.

Moreover Sugden and Williams (1978) argue that finding the social value of public goods can be a very challenging task. One way is to try to measure it through the effects that they may have on market value of a ”property”, or in general the effects they have on markets that trade access to this public goods. Knowledge of what time/ cost/ travel or other inconveniences an individual is willing to bear in order to receive a service (opportunity cost) is helpful in estimating the value of an extra unit of the receiving service. Usually service prices can be objective and vary depending on the point of view of different groups of people and forces the analysts to observe simultaneously a large number of individual points, but on the same demand curve, assuming that all individuals have the same demand curve, which may be convenient and facilitate the analysis, but also is completely unrealistic.

In addition to this Sugden and Williams (1978), they state that healthcare activities are not the true outputs of healthcare, but means to produce the real outputs that have social value, which are improvements in people’s health condition and quality of life. These outputs can be measured in two units: the number of cases of a condition cured q₁ and the number of patients with a second condition cured q₂. So the social value of medical activities will be V= a₁q₁+a₂q_2, where a₁ and a₂ are constants and express the social value of each activity to cure the two conditions. So to maximize the social value of output, we have to maximize the quantity of the output. The distribution of income implied by the Pareto criterion is based on the idea that there are costless direst transfers of income. But in the real world and especially in a well organised tax system, in order to transfer a monetary unit to one person, an equal monetary unit must be deducted from another individual, so if direct transfers of income are not costless the distribution of income may not increase social welfare. In CBA, this can be solved by the use of weights to determine distributional judgments. They also argue that a decision or a project can not only have direct effects, but also indirect effects, which are more widespread, and are very difficult for analysts to predict, identify and measure them. This usually causes the common problem of double counting an effect in the analysis process.

To deal with the problem of uncertainty in many aspects of CBA and CE analysis, Garber (1999) the use of Sensitivity Analysis to present the effects of variability of uncertain variables on the effect of the analysis. In one–way sensitivity analysis one parameter changes over time while in a two way sensitivity analysis, two parameters can vary at the same time. This method has a limitation that it cannot show the result that many variables that change at the same time have on the final result. This can be solved by using more statistical approaches, such as calculating confidence intervals around the outcome variable. Arrow and Lind (1994) also state that individuals will be unwilling to value programs with uncertain results since they are not indifferent to uncertainty.

Finally, Sugden and Williams (1978), argued that the expected value criterion which states that in cases of uncertainty a project should be undertaken, if the expected value of its returns is greater than the expected value of its costs, can be reformed referring to social returns and costs and thus to be adopted in CBA . It can be assumed that financial risks are small because individuals will not risk major wealth changes and the tax system can be a very effective way to spread risk, in case a project or a decision affect individuals only through tax. But for direct risk, such an assumption cannot be made.

A further aspect of CBA is the use of willingness to pay approach (WTP) in benefits/value monetisation. Weinstein (1996) argues that an issue that raises questions about the willingness to pay approach is the validity of the estimates for this approach. The analysts should identify what are the main aspects under consideration in individual’s assigned value to health programs. This implies whether the individual will value a healthcare programmes from a holistic perspective or he will value the consequences of the medical intervention. For holistic healthcare programs, misallocation of information to the consumer about the consequences of medical interventions makes estimates of WTP less reliable.

Download 1.95 Mb.

Share with your friends: