To the best of our knowledge, this exploratory study is the first to estimate the value of QALYs by considering the perspectives of people with diabetes mellitus. Patients’ WTP per QALY ranged from $US1191 to 5043, which is equal to 0.23–0.95 times the Iranian GDP per capita in 2015. Although the QALY values changed in accordance with health preferences and discount rates, all the values were lower than one times the GDP per capita. Thus, in contrast to the WHO’s recommendation to use one to three times the GDP per capita as a cost-effectiveness threshold, our findings indicated a value close to the lower bound of the GDP per capita and did not support the higher bound of three times GDP per capita. The results are also consistent with findings from our recent study [
41] wherein we estimated WTP for QALYs from the perspectives of patients with cardiovascular disease. All in all, the findings indicated that the WHO’s recommendation (i.e., < 3 times the GDP per capita) should be used with considerable caution and care in decision making and resource allocation for cost-effective interventions. Moreover, our empirical results revealed an important methodological issue in relation to estimating WTP for QALYs through a chained approach (Eq.
1) for patients who rated/stated their current states as corresponding to full health in their health state valuations while having a positive WTP for obtaining treatment. A zero value in the denominator of Eq.
1, which would result in an undefined WTP/QALY, and exclusion of these patients or inclusion them with an arbitrary value for QALY gained in the denominator, may produce bias in the monetary estimation of QALYs. Byrne et al. [
42] asserted that eliminating respondents with very low QALYs can lead to a downward estimation of QALY values, whereas Zhao et al. [
9] found that excluding extreme values from the EQ-5D scale did not produce a significant bias in estimated WTP for a QALY. Therefore, we believe more investigation is required to identify and manage bias when direct or indirect health measures are used to elicit QALY gains. We also acknowledge that bias issues are a potential problem with the TTO technique. In our study, few respondents expressed full health valuations in the EQ-5D-3L scale and the VAS, but 22–38% of patients had zero QALYs gained in the adjusted and conventional scenarios of the TTO method. These respondents constitute a significant sample proportion that cannot be ignored. The interests of the patients’ families and children, the non-severity of their disease, and ongoing treatment for increased longevity were the three main drivers of a non-trading orientation among some of the respondents. All except one non-trader had a positive WTP for a permanent cure and restoration to full health; that is, they were willing to pay for zero gain, suggesting that, although patients were aware of their health states, they were unwilling to trade any duration of lifespan in exchange for restoration to full health because of the three reasons mentioned above. Attema et al. [
43] pointed out that excluding people with an infinite value for life from analyses is inappropriate, and Iezzi et al. [
44] recommended that non-traders be retained in analyses by permitting them to trade in small units of time (< 1 year). We believe this idea may be useful in evaluating patients’ health-related quality of life but is inappropriate when used to estimate WTP for QALYs using a chained method because very small units of QALYs gained in the denominator of Eq.
1 pose the risk of overestimation. Let us take the TTO-adjusted scenario as an example. Under the absence of non-traders, the monetary value of a QALY is $US4215, and assuming a 0.999 utility value instead of 1 causes the WTP for QALYs to be overestimated at $US109,872 or 21 GDP per capita. Such a value exceeds the resources of the healthcare system in any country with limited resources and is therefore unacceptable. On the other hand, eliminating respondents with very small QALYs gained could result in a lower estimation of WTP per QALYs [
42]. The apparent unsuitability of trading in small units of time and the continued existence of non-trader issues constitutes an analytical challenge. An alternative would be to employ other health state valuation techniques that have fewer problems of full health valuation.
Despite our results, nothing should be taken at face value, as by no means do we claim that our findings represent the absolute value of a QALY for patients with diabetes. More research is needed to further investigate the effects of including or excluding non-traders or respondents with a full health valuation and positive WTP for the monetary value of QALYs. Some of the main limitations of this study are the sample size and sampling technique. Our sample is not representative of people with diabetes in our society because we interviewed only patients who sought treatment from a publicly funded hospital. Patients from other geographical regions and those admitted into private hospitals or clinics were excluded from the research, as were patients with serious health problems, such as foot amputation, blindness, and severe kidney failure, given the nature of TTO questions. The study is also limited in terms of the remaining lifetime that we used to implement the TTO scenarios and calculate QALYs gained. It was not a life expectancy specific to patients with diabetes and was instead based on national tables for different age groups; such patients may have a shortened lifespan. In this case, the QALYs gained may be underestimated if respondents assume a gain in longevity as well as quality of life after they purchase a “cure.” Generalizing the findings, especially when it comes to decision making, requires additional surveys with a broader and more representative sample. For informed decision making and efficient resource allocation, more empirical research should be conducted to elicit disease-specific QALY values from the patient’s perspective, with particular emphasis on the nature of a disease, that is, whether it is life threatening, chronic, or disabling [
9].