Assessing value using the QALY
Priority setting in healthcare has long been recognized as: “an intrinsically complex and value-laden process” [
63]. Society, including relevant stakeholders, such as patients, providers, insurers, and citizens, has a wide range of social values and interests that result in different perceptions of what makes health interventions valuable [
64]. A recent review has also identified heterogeneity in value assessment systems across Europe, which results in significant differences in coverage recommendations across settings based on how HTA agencies perceive or interpret evidence and the associated uncertainties [
65]. Current HTA value assessment frameworks do not adequately capture the range and diversity of stakeholder values. Frameworks used by HTA agencies are typically based on comparative clinical benefit assessment and economic evaluation (cost utility analysis or cost-effectiveness analysis) as the main method of determining the value of new technologies. Sometimes, this involves reflecting health gain as quality-adjusted life years (QALY), or an alternative patient-relevant outcome; for example, life-years gained [
65].
Typically, individuals with ill health want improved health; i.e., length of life and/or quality of life (patient experience in respect of, for example, delivery of care, pain or other factors, impact on the family, time off work). QALYs combine quality and quantity of life into a single composite measure: value per health state multiplied by the length of time in that state [
66]. QALYs can be used to compare the benefits of interventions across therapy areas and determine priorities by means of cost per QALY ratios [
66] against a threshold value of a QALY. In theory, this approach ensures that new treatments do not displace more health gain than they provide and do not diminish the overall value of benefits gained from the healthcare budget. Thresholds are in place in England, Poland and an academic proposal for a threshold was identified for Spain in a recent review [
65]. In England, the threshold ranges between £20,000 to £30,000 [
50] although evidence indicates that this threshold can be modified in practice with some products with an ICER below £20,000 receiving a negative recommendation and other products with an ICER above £30,000 receiving a positive recommendation [
65].
While the threshold provides a framework for decision-making, it does not take into account the fact that decision-makers may wish to account for other factors not captured in the cost-effectiveness ratio when assessing a medicine pricing/reimbursement decision [
67]. One way of managing this has been to vary the threshold ICER according to the type of medicine, the type of disease, and the decision-making context, such as the age of the patient or the severity of the disease [
37]. However, what constitutes value varies between jurisdictions. For example, in England, NICE uses an end of life criterion for which the underlying assumption is that society places a higher priority on treating those patients near the end of their life [
50]: an additional QALY under end of life conditions is worth more than a QALY at another point time although this is widely debated [
68]. The threshold is also higher for “ultra-orphan” health technologies [
69].
In Norway, patient benefit, resource use and severity of the disease are proposed as the three main sources of value for priorities in the healthcare system [
70]. Rarity is not mentioned as a specific criteria for value. In Sweden, society should be able to pay for more health gain and accept lower standards for scientific evidence when prioritising drugs for treating rare conditions if the following conditions are met: (1) the treatment has a high cost per health gain as a consequence only treating a few patients; (2) it involves a health condition with a very high level of severity; (3) that the treatment option being considered is assumed, based on firm grounds, to have a substantial effect, and (4) that no alternative treatment having a substantial effect is available [
71]. The Norwegian and Swedish proposals have broad public and political support, but should only been seen as general guidelines. It is up to Legemiddelsverket in Norway and TLV in Sweden to translate these guidelines into specific decisions for individual products applying for reimbursement. That includes making the trade-off between different dimensions of value.
Guidelines for reimbursement are silent about the role of empirical studies for assessing the value of specific products or classes of products from patients or the general public. Such studies did not find consistent support for a preference for health gains to cancer patients in the context of health maximization [
72,
73]. However, they observed a higer willingness to pay for severe diseases, while the empirical support for additional value for “end-of-life” treatments are weak [
74].
There is also evidence that with short life expectancy, few will trade off less survival vs. improved quality of life [
75]. Cost per life year-gained may, therefore, be a more adequate metric for assessing the value of new treatments for persons with short life expectancy. Presenting a calculation of cost per LYG in addition to cost per QALY, would provide decision-makers with additional information for decisions of value from a patient perspective.
Additional elements of value (beyond health gain)
While assessments using the cost per QALY approach include many important and relevant aspects of value such as impact on survival, quality of life, and potential cost savings, they do not include assessment of other elements of value beyond the health gain for the patient and costs to the health system which may also be relevant for payers, patients, and society [
76]. Examples of these wider benefits include disease severity, age of onset, lifetime burden of illness, socioeconomic impact, and possible spillovers from the initial innovation [
65], or improvements in the quality of or process of care that may also not be captured by measures of improvements in outcome (e.g., home vs. hospital treatment or oral vs. intravenous treatment). Some ATMPs may not be deemed cost-effective according to conventional decision rules, and in some cases this will be because the therapy simply does not represent good value for money. However, in other cases, the therapy may offer high value that is just not reflected in the HTA evaluation, and/or the therapy may be restricted by the set of aforementioned challenges (e.g., difficulties in establishing robust estimates of clinical effectiveness). Potential additional sources of value are explored in the following paragraphs, each of which could potentially justify some premium on top of what is deemed acceptable to pay for pure health gains.
Short life expectancy at diagnosis or at the start of treatment is a common feature of many illnesses. In healthcare, a tension sometimes arises between the injunction to do as much good as possible with scarce resources (“cost-effectiveness”) and the injunction to rescue identifiable individuals at immediate risk, regardless of cost (the “Rule of Rescue”) [
77]. Culyer notes that the “rule of rescue” argument has two important caveats: first, the strength of the argument is weakened if the additional time is of poor quality or of a worse quality than would be the case under normal palliative care, and, second, that the end-of-life argument applies to all patients, and not only cancer patients, in which context it is commonly discussed [
76].
In an end of life context, patients themselves may also be willing to take risks or pay for options with greater uncertainty or immediate mortality risk if there is a significant chance of increased long-term survival [
78]; this has been referred to as the “value of hope” [
79]. Literature on prospect theory also suggests that patients with life-threatening conditions sometimes appear to make risky treatment decisions as their condition declines, contradicting the risk-averse behaviour predicted by expected utility theory [
80]. Rasiel et al. demonstrated that behaviour links to a reference point; for example, when the recently diagnosed patient has not yet adapted to their new prognosis, the prospective value of the investigational therapy exceeds that of the conventional therapy [
80]. Patients’ reference points may take time to adjust following a change in diagnosis, with implications for predicting under what circumstances a patient may select experimental or conventional therapies or select no treatment [
80].
While prospect theory captures individual valuations when faced with difficult choices under uncertainty, the application of the theory to inform public decisions has its challenges. Suppose, for example, a diagnostic was developed that can predict which patients will be cured by a therapy. This may reduce the value from an individual patient perspective, since the hope of cure for all patients is replaced by certain cure for a few. A public payer may still be interested in paying for this test, thus reducing the cost of treatment, but there may be perverse incentives against the development of companion diagnostics that may reduce the size of the market since this takes away the “value of hope”.
Curative therapies could eliminate the need for long-term management and provide longer term increases in quality of life. It is not known, however, whether curative therapies are valued more highly by society than treatments that offer the same “total” health gains through marginal gains over many years and/or patients. Little evidence is available to suggest whether or not this preference does exist, and no weighting is currently available to adjust for this [
10].
Additional elements of value identified in the literature in relation to complementary diagnostics may also be relevant for consideration in the assessment of ATMPs include: innovation and real option value, and scientific spillovers [
79]. These elements are briefly discussed below; however, there is currently limited research into how these elements should be formally incorporated into the decision-making framework.
Producing innovative drugs is increasing in cost. The role of cost-effectiveness analysis in incentivizing innovation is controversial; currently cost effectiveness analysis rewards gains in clinical benefit. Innovation has been defined as the: “successful introduction of something new and useful, for example, introducing new methods, techniques, or practices or new or altered products and services” [
81]. It is typically characterised in the economic literature as “a cumulative process of success” [
82]. In healthcare, the term innovation lacks specificity and differs by country [
83]. Italy has published criteria for identifying an innovative product. With this algorithm, pharmaceuticals are designated as an important, moderate, or modest therapeutic innovation based on: (i) the availability of existing products or (ii) the extent of the therapeutic benefit [
83]. In France, an improvement of medical benefit Amerlioration du Service Medical Rendu (ASMR) level (major innovation, important imIf ATMP improvement, significant improvement, minor improvement and no improvement) is assigned for each product vs. standard of care, but the criteria used to determine these levels is not defined [
84]. Innovations that result in benefits for society or facilitate benefits from future technologies might justify some reward [
81]. Culyer argues that innovation is “already rewarded (or at least encouraged) through the patent system and profit with special pricing and profit regulatory scheme in most countries” [
76]. To best assess the value of innovation, perhaps consideration of the overall health need alongside innovation goals and priorities is required [
81].
Real option value refers to the investment in healthcare that can lead to potential treatment pathways for patients in the future as other new technologies become available [
79]. Evidence suggests that patients perceive “option value” from treatment as getting one treatment despite its disadvantages (e.g., side effects) increases the likelihood of benefiting from a better treatment in the future [
10,
79].
Finally, scientific spillovers relate to knowledge spillover whereby one company’s achievements can lead to the success of another company developing a similar technology [
79,
85]. Sweeney and Goss suggest that the market authorisation of a new therapy leads to additional research and the generation of additional evidence to understand the benefits of a treatment in a given clinical context [
79,
85]. Combinations of approved therapies are used in successive clinical trials, further increasing scientific spillovers over time [
79,
85]. The advantages and disadvantages of the benefits of scientific spillovers are, however, unpredictable.
To summarise, traditional cost-effectiveness analysis conducted as part of HTA focuses on life-years gained, improvements in patient quality of life, and cost savings within healthcare. While the current framework may be appropriate for the assessment of ATMPs, it is also important that the full potential value of ATMPs is recognised. We suggest that this will involve incorporating an assessment—or at least consideration—of other aspects of value into the current evaluative framework for reimbursement. However, if such additional considerations are to be included in future assessment, then further work is required to ensure that these considerations are also applied to competing technologies and those that may potentially be displaced by new expenditure on ATMPs [
86].