As estimated by the World Health Organization, depressive disorders will be the leading contributor to the Global Burden of Disease by 2030. In light of this fact, we designed a study whose aim was to investigate whether the value placed on health-related quality of life (HRQoL) for a depressive disorder is higher in patients diagnosed with a major depressive disorder (MDD) compared to non-patients in a matched sample.
We collected data on willingness to pay (WTP) for a total of four health-gain scenarios, which were presented to 18 outpatients diagnosed with a MDD versus 18 matched non-patient respondents with no symptoms of depression. Matching characteristics included age, income, level of education, and type of health insurance. Respondents were presented with different HRQoL scenarios in which they could choose to pay money to regain their initial health state through various treatment options (e.g., inpatient treatment, electroconvulsive therapy). To test whether the probability of stating a positive WTP differed significantly between the two samples, Fisher’s exact test was used. Differences regarding stated WTP between the samples were investigated using the Mann-Whitney U-test.
For most of the health scenarios, the probability of stating a positive WTP did not differ between the two samples. However, patient respondents declared WTP values up to 7.4 times higher than those stated by matched non-patient respondents.
Although the perceived necessity to pay for mental-HRQoL gains did not differ between respondents with MDD and respondents with no symptoms of depression, patient respondents stated higher values.
The probability of stating a positive WTP did not differ between samples. However, patient respondents stated WTP values as much as 7.4 times higher than non-patients.
The global burden of disease is shifting from premature death to years lived with disability (
Due to limited resources in the health-care sector, cost-effectiveness analyses are used as guidelines in priority setting, resource allocation, and reimbursement decisions. The preferred metric of health benefits in cost-effectiveness analyses is commonly the measurement of quality-adjusted life years (QALYs), combining the impact of health benefits on both health-related quality of life and quantity of life years (
Various studies have tried to estimate the value of a QALY through the WTP method (e.g.,
WTP per QALY seems to be related to several Unlike other European countries, Germany has a universal health-care system with two types of health insurance: Germans can choose between public (statutory) insurance and private health insurance, which is co-financed by employer and employee.
To the best of our knowledge, no study has ever investigated the effects of the individual relevance of the presented health-gain scenario on the respondent’s WTP per QALY. Additionally, several studies argued that the plurality of different perspectives should be acknowledged, and that values for health benefits (i.e., QALYs) should be based on preferences from both patients and the general public (
With an eye toward this need for more specific information on patient and non-patient preferences, the aim of our study was to assess whether WTP preferences for mental health gains differ between outpatients with a diagnosed major depressive disorder (the patient sample) and respondents from the general public with no symptoms of depression (the non-patient sample). To control for the effects of the above-mentioned individual characteristics on WTP, we matched respondents from the patient sample with respondents from the non-patient sample based on income, level of education, age, and type of health insurance (see Section ‘Participants and Procedures’). The above-mentioned meta-analytical comparison of patient and non-patient health-state assessments found that patients give higher valuations than non-patients (
The probability of indicating a positive WTP (WTP > 0) is higher throughout all the scenarios in the patient sample compared to its likelihood among respondents with no self-reported symptoms of depression (the non-patient sample).
Respondents from the patient sample are willing to pay significantly higher amounts for the health gains presented than respondents with no self-reported symptoms of depression (the non-patient sample).
This study was performed in accordance with the principles of the Declaration of Helsinki. The Ethical Review Committee of the University of Hildesheim, Germany, approved the study (Application number: 107).
Individuals with a suspected depressive disorder were screened at a German university outpatient clinic between May 2019 and March 2020. Possible participants were informed as to the objective of the study both verbally and in writing, and were required to provide their written consent. Participants were eligible for inclusion if they were more than 18 years of age and met the
For each respondent in the patient sample, we compared one matched respondent from the German general population who reported no symptoms of depression. Computer-based matching was conducted using the following characteristics: age at index rate (± 8 years), income category (see
On the first page of the online questionnaire, respondents were informed about the objective of the study and were asked to give their consent. The hypothetical scenario that was introduced assumed that no sickness funds exist in Germany, and therefore, respondents would not have to pay premiums or contributions toward health insurance, increasing their monthly net income by that amount. Respondents were asked to imagine that instead, they would need to pay for every medical service out of their own pocket.
The concept of measuring health on a visual analog scale was introduced: Based on the European Quality of Life 5-Dimensions 3-Level Version (EQ-5D-3L;
Next, a description of typical symptoms of depressive disorders and their impact on everyday life, including mortality rates by suicide, was presented (see Online Resource 1 in the This treatment method was used because its efficacy is recognized by the German Association for Psychiatry, Psychotherapy, and Psychosomatics (DGPPN), and because it is a highly standardized procedure with rapid response rates (
Scenario | Health gain | Duration | Time | Initial health state achieved? | Treatment |
---|---|---|---|---|---|
A | 25 points | 4 years | In 1 year | 100% | pain-free treatment |
B | 10 points | 10 years | In 1 year | 100% | pain-free treatment |
C | 25 points | 4 years | In 1 year | 90% | 8-week inpatient treatment |
D | 25 points | 4 years | In 1 year | 90% | 8-week inpatient treatment plus electroconvulsive therapy |
To ensure that the questions were relevant to the individual respondents, and in accordance with the EuroVaQ report (
Respondents who indicated that “the government should pay” from the set of pre-coded responses as the reason for zero WTP (so-called “protest respondents”), were excluded due to their not having understood the hypothetical nature of the scenario (as is standard for WTP studies; see
Additionally, respondents were excluded from data analysis regarding scenarios A, C, and D if they rated their health state at less than 35 points (indicating poor health), and if they expected to live for less than 6 years as of that day. Respondents were excluded from data analysis regarding scenario B if they rated their health state at less than 20 points, and if their life expectancy was assumed to be below 12 years. The intention was to ensure that no health loss reduced the respondent’s health to below 10 points, and that all health gains were complete at least one year before the respondent expected to die.
All analysis was undertaken with IBM SPSS Statistics 26. The collection of open-ended responses allowed us to determine the mean and median values reported for each scenario, which were collected in Euros. The current study does not report trimmed means because
To test Hypothesis 1 — whether the likelihood of expressing a positive WTP differed across both samples — WTP responses were dichotomized as zero and non-zero values. Because of the small sample size, Fisher’s exact test and odds ratios were calculated. To assess Hypothesis 2 — whether WTP values for the described health gains differed between the patient and the non-patient sample — the nonparametric Mann–Whitney
From an initial sample of 20 screened outpatients,
The matching process based on income, level of education, type of health insurance, and age resulted in a sample of
Characteristic | Patient Sample |
Non-patient sample |
||
---|---|---|---|---|
Min/Max | Min/Max | |||
Age (in years) | 48.33 (15.22) | 22/77 | 47.89 (14.97) | 22/70 |
Life expectancy (age) | 82.28 (9.49) | 65/99 | 83.78 (8.45) | 70/110 |
Health status (0-100) | 61.67 (18.31) | 20/95 | 89.94 (9.17) | 70/100 |
% | % | |||
20 to 69 (poor) | 11 | 61.1 | 0 | 0.0 |
70 to 79 (rather poor) | 2 | 11.1 | 2 | 11.1 |
80 to 89 (rather good) | 4 | 22.2 | 2 | 11.1 |
90 to 100 (very good) | 1 | 5.6 | 14 | 77.8 |
Low remaining lifetime (< 16 years) | 4 | 22.2 | 1 | 5.6 |
Females (rather than males) | 16 | 88.9 | 11 | 61.1 |
Educational level | ||||
Basic (nine years) | 0 | 0.0 | 0 | 0.0 |
Secondary (ten years) | 8 | 44.4 | 7 | 38.9 |
Tertiary (> ten years) | 10 | 55.6 | 11 | 61.1 |
Monthly household income | ||||
No answer | 1 | 5.6 | 1 | 5.6 |
Below 500 € | 0 | 0.0 | 0 | 0.0 |
500 to below 1.000 € | 1 | 5.6 | 1 | 5.6 |
1.000 € to below 1.500€ | 1 | 5.6 | 1 | 5.6 |
1.500€ to below 2.000€ | 4 | 22.2 | 4 | 22.2 |
2.000€ to below 3.000€ | 4 | 22.2 | 4 | 22.2 |
3.000€ to below 4.000€ | 6 | 33.3 | 6 | 33.3 |
4.000€ and more | 1 | 5.6 | 1 | 5.6 |
Health Insurance | ||||
Social insurance | 17 | 94.4 | 17 | 94.4 |
Private insurance | 1 | 5.6 | 1 | 5.6 |
ICD-10 Diagnosis | ||||
Depressive episode | 8 | 44.4 | ||
Recurrent MDD | 10 | 55.6 |
Results from Fisher’s exact test indicate no association between the sample (patient vs. non-patient sample) and the probability of stating a positive WTP (WTP > 0) in three of four scenarios (Scenarios B, C, and D). Only in scenario A was the probability of expressing a positive WTP higher in the patient sample compared to the non-patient sample (χ2 = 6.84,
In the patient sample, the number-one reason for being unwilling to pay for the presented health gains across all scenarios was: “The effects of treatment are too small.” In the non-patient sample, the number-one reason stated was: “It would not be so bad/I could live with it.”
Scenario | It would not be so bad/ I could live with it | Effects of treatment are too small | I want my family to have the money | I would get better without treatment | I value the treatment but cannot afford it | Other reasons | |
---|---|---|---|---|---|---|---|
Patient sample | |||||||
A | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
B | 1 | 0 | 1 (6.3) | 0 | 0 | 0 | 0 |
C | 2 | 0 | 0 | 0 | 0 | 0 | 2 (11.8) |
D | 8 | 0 | 2 (11.8) | 0 | 0 | 0 | 6 (35.4) |
Non-patient sample | |||||||
A | 6 | 1 (5.6) | 0 | 1 (5.6) | 1 (5.6) | 1 (5.6) | 2 (11.2) |
B | 7 | 4 (22.2) | 0 | 1 (5.6) | 0 | 1 (5.6) | 1 (5.6) |
C | 6 | 1 (5.6) | 0 | 0 | 2 (11.2) | 2 (11.2) | 1 (5.6) |
D | 8 | 2 (11.2) | 0 | 0 | 2 (11.2) | 2 (11.2) | 2 (11.2) |
Mean, median, and maximum WTP values, as well as bias-corrected accelerated 95% confidence intervals around means, are displayed in
Scenario | Bootstrapped 95% CI | Maximum WTP | ||||
---|---|---|---|---|---|---|
Patient sample | ||||||
A | 17 | 17 | 54,794 | 14,646-116,424 | 15,000 | 350,000 |
B | 16 | 15 | 52,667 | 6,956-121,249 | 10,000 | 350,000 |
C | 17 | 15 | 23,867 | 10,714-45,548 | 10,000 | 150,000 |
D | 17 | 9 | 15,778 | 7,667-25,762 | 13,000 | 50,000 |
Non-patient sample | ||||||
A | 18 | 12 | 4,650 | 2,322-7,686 | 2,500 | 15,000 |
B | 18 | 11 | 2,277 | 1,000-4,126 | 1,500 | 10,000 |
C | 18 | 12 | 3,433 | 2,245-4,737 | 2,750 | 10,000 |
D | 18 | 10 | 2,415 | 1,183-3,567 | 1,750 | 5,000 |
As currently discussed (e.g.,
Results indicate that the probability of stating a positive WTP does not differ between patients and non-patient respondents. However, when assessing the number-one reasons indicated for zero WTP (patient sample: “Effects of treatment are too small,” vs. non-patient sample: “It would not be too bad/I could live with it”), it seems that respondents with no prior experience of depression underestimate the burden of depressive symptoms. As discussed by
In this study, we assessed respondents’ WTP for one specified treatment (electroconvulsive therapy) in detail due to its high standardization when compared to other psychotherapeutic interventions. Thus, when assessing results for this specified scenario, it seems unexpected that only 53% of the patient sample and 55% of the non-patient sample were willing to pay money for ECT. One possible explanation might be that 83% of the patient sample stated that they knew nothing or little about ECT, compared to 72% of the non-patient sample. The present findings accord with the conclusion of a recent study, which found that ECT is still largely underutilized due to persisting stigma and lack of knowledge about modern ECT techniques (
The cost-effectiveness of primary care for depressive disorders has been investigated by, for example,
Matching the respondents from the patient sample to respondents from the non-patient sample allowed us to control for the effects of individual characteristics (e.g., income, level of education) on WTP. Presenting the scenarios in a randomized order let us control for ordering effects. However, some limitations should be also mentioned.
First, the size of both samples (
Second, the broad majority (88.9%) of the recruited patient sample was female. Results from the EuroVaQ study indicate that men stated a higher WTP (
Additionally, presenting scenarios that emphasize the certainty of successful treatment — which may be especially unlikely with respect to mental health — may have led to the overestimation of estimated WTP values. More scenarios with uncertainty characteristics should be evaluated in further research, as well as other specified treatment options, such as psychotherapeutic treatment approaches or antidepressant medication (
Fourth, the assessment of the variable “knowledge about ECT” consisted of one item only, and did not objectively specify how much respondents know or how and where they became informed (e.g., movies, media, medical services). During administration of the present survey, a measure to assess perceptions and knowledge of ECT was published (
Additionally, we only recruited people who were being seen at an outpatient clinic. It is possible that patients of an inpatient clinic with more severe depressive symptoms would place higher values on mental-health-related quality of life, and might also be better informed about their treatment options — ECT in particular. Generalization of results may therefore be limited to patients from an outpatient setting with no co-occurring mental disorders.
Finally, the health-care system (including psychiatric and psychological care) in Germany is unique compared to that of other European systems (see
This study investigated the effect of the personal relevance of a presented health-gain scenario on the respondent’s WTP per QALY, and produced findings that add valuable information toward estimating the effects that individual characteristics have on the value that respondents place on a QALY. Additionally, our findings emphasize the need to assess hypothetical population preferences alongside actual patients’ preferences for health benefits.
We would like to thank Lars Paternoster and Robert Szczepanski for their help in implementing the questionnaire versions online. In addition, we would like to thank Sina Haider, Marieke Hansmann, Laura Lefarth and Kira Schamke, who conducted the screening interviews together with one of the authors.
The Supplementary Materials contain the following items (for access see
Supplementary Material 1: Translation of the health state description
Supplementary Material 2: Sample scenario
The authors have no funding to report.
The authors have declared that no competing interests exist.