Health state utilities associated with hyperphagia: Data for use in cost‐utility models

Abstract

Objective

Rare genetic diseases of obesity typically present with hyperphagia, a pathologic desire to consume food. Cost‐utility models assessing the value of treatments for these rare diseases will require health state utilities representing hyperphagia. This study estimated utilities associated with various hyperphagia severity levels.

Methods

Four health state vignettes were developed using published literature and clinician input to represent various severity levels of hyperphagia. Utilities were estimated for these health states in a time trade‐off elicitation study in a UK general population sample.

Results

In total, 215 participants completed interviews (39.5% male; mean age 39.1 years). Mean (SD) utilities were 0.98 (0.02) for no hyperphagia, 0.91 (0.10) for mild hyperphagia, 0.70 (0.30) for moderate hyperphagia, and 0.22 (0.59) for severe hyperphagia. Mean (SD) disutilities were −0.08 (0.10) for mild, −0.28 (0.30) for moderate, and −0.77 (0.58) for severe hyperphagia.

Conclusions

These data show increasing severity of hyperphagia is associated with decreased utility. Utilities associated with severe hyperphagia are similar to those of other health conditions severely impacting quality of life (QoL). These findings highlight that treatments addressing substantial QoL impacts of severe hyperphagia are needed. Utilities estimated here may be useful in cost‐utility models of treatments for rare genetic diseases of obesity.

1. INTRODUCTION

Rare genetic diseases of obesity are often caused by impaired function of genes involved in hypothalamic signaling that regulates appetite and energy homeostasis. ¹ , ² Some rare genetic diseases of obesity arise from variants in an expanding number of genes in the leptin‐melanocortin 4 receptor (MC4R) pathway, including LEPR, POMC, MC4R, and PCSK1. ² Additionally, several syndromic rare genetic diseases of obesity have also been identified, including the ciliary disorder Bardet‐Biedl syndrome, Alström syndrome, and Prader‐Willi syndrome (PWS). ² These diseases can present with a high comorbidity burden, including hyperphagia, which is characterized by an extreme, persistent desire to consume food, preoccupation with food, and lack of satiety following food consumption. ¹ , ² Hyperphagia is a hallmark feature of rare genetic diseases of obesity. Hyperphagia is a result of deficient signaling within the MC4R pathway and quickly leads to severe obesity in individuals with these conditions. Hyperphagia‐associated behaviors have been described as a “relentless, overwhelming, life‐threatening force” accompanied by an extreme burden negatively impacting the lives of affected patients and their caregivers. ¹ Currently, there is a gap in the understanding of the contribution of hyperphagia to the burden of rare genetic diseases of obesity. ³ The extreme burden of hyperphagia in patients with these diseases and their families and caregivers highlights the need to quantify the effects of hyperphagia to better inform further assessments related to this condition, such as cost‐utility analyses.

Decision‐making for healthcare reimbursement and resource allocation is often informed by cost‐utility analyses. These analyses are a type of economic modeling designed to determine the value of medical treatments based on the calculation of quality‐adjusted life years (QALYs). ⁴ The calculation of QALYs requires health state utilities for a given condition. Utilities are values representing strength of preference for various health states on a scale ranging from 0 to 1, where 0 = dead and 1 = full health. While several studies have assessed health state utilities associated with obesity, ⁵ , ⁶ , ⁷ , ⁸ there is no published evidence on the independent impact of hyperphagia on utilities. In this study, the health state utilities associated with various levels of hyperphagia severity were estimated using a vignette‐based time trade‐off (TTO) approach. ⁹ Specifically, the aim was to identify health state utilities associated with hyperphagia, independent of weight, obesity, and any additional comorbidities. Health technology assessment authorities generally prefer utilities derived from generic preference‐based measures such as the EQ‐5D. However, this method may not be feasible with rare conditions such as hyperphagia because it can be difficult to obtain a sufficient sample size to value each relevant health state. ⁹ , ¹⁰ Therefore, in situations where it is necessary to estimate utilities associated with rare conditions such as hyperphagia, vignette‐based methods are often a more feasible approach.

2. METHODS

2.1. Hyperphagia health states

Health state vignettes representing varying severity levels of hyperphagia were developed using published studies ¹ , ¹¹ and iterative interviews with clinicians who have experience treating patients with hyperphagia (2 from the US; 1 from the UK). Clinicians were asked to define hyperphagia and its symptoms, impact, screening procedures, and concepts that may be included in screening tools. These interviews included open‐ended questions designed to elicit description of the typical experience of a patient with hyperphagia and continued until all clinicians agreed on clearly described health states and accurate hyperphagia descriptions.

Four health states were developed: A (no hyperphagia), B (mild hyperphagia), C (moderate hyperphagia), and D (severe hyperphagia). Health states used in this study are included in Tables S1−S4. Instead of labeling the health states A through D, they were labeled as T, K, F, and V to avoid suggestion of any severity order. The health states and utility assessment procedures were pilot tested with 21 individuals in April 2021 to ensure the health states and methods were clear to respondents before conducting the larger utility valuation study. All participants understood the health states and reported key factors in their health state preferences to be emotional distress; impact on daily life, family, and social relationships; and never feeling full.

2.2. Participants and interviews

The study was conducted with individuals from the UK general population in England and Scotland who were recruited through distributed fliers and through advertisements posted in newspapers and online (in classifieds and on social media). Interested individuals who contacted the study team were screened for eligibility. Participants were required to be 18 years or older and able to understand assessments and provide consent to participate in virtual interviews and questionnaires. Individuals with a cognitive or visual impairment, hearing difficulty, severe psychopathology, or insufficient knowledge of English were excluded from the study. Study documents including four hyperphagia health states and a demographic questionnaire were mailed to participants prior to their interview. Eligible respondents participated in a one‐time Zoom interview between April and May 2021. The study protocol and data collection materials were approved by an institutional review board (Ethical and Independent Review Services; Study 21028‐01), and all participants provided written informed consent.

The interviews consisted of two parts: a health state ranking task and a TTO assessment. For the introductory task, participants ranked the four health states in order of most preferable to least preferable. All participants valued all four health states. For each participant, the four health states were presented in random order before the ranking task.

These health state rankings were then assessed as a utility score from dead (0) to full health. ¹ For the TTO task, participants were given a choice of spending 10 years in a particular health state or spending time in full health. These choices were provided in 6‐month increments (10 years, 0 months [dead]; 9.5 years, 6 months; 9 years, 1 year; etc.) for each health state perceived as better than dead. Given persistent measuring and scaling problems in valuing “worse than dead” health states, ¹² , ¹³ a composite TTO (cTTO) ¹³ was used for this study, where classic TTO was used for health states considered to be better than dead and lead‐time TTO was used for health states perceived to be worse than dead. ¹² , ¹⁴ Disutility for each severity level of hyperphagia (health states B–D) was derived as the utility difference between a given health state and health state A. This method allows worse than dead and better than dead TTO to be similar tasks while still providing a clear distinction between them. The cTTO approach in this study has been used elsewhere and is the recommended protocol of the EuroQol group. ¹² , ¹⁴ Respondents were additionally asked to explain their rankings and TTO valuations, which were transcribed by the interviewers.

2.3. Statistical analysis

Statistical analyses were performed with SAS, version 9.4 (Statistical Analysis System; Cary, North Carolina). Descriptive statistics were used to summarize utility differences and general demographic and clinical characteristics. Assessments using t‐tests (2‐tailed, p < 0.05) were performed to examine comparisons between groups (independent t‐tests) in demographics and utility, as well as differences between health states (paired t‐tests). Continuous variables are summarized as means and standard deviations. Categorical variables are reported as frequencies and percentages.

3. RESULTS

3.1. Sample characteristics

For the full study, 220 potential participants were screened, of which 215 were eligible for the study and completed interviews. The study sample was 60.5% female (n = 130), with a mean age of 39.1 years (range, 18–76 years). Most participants reported their ethnicity as white (84.2%), and the majority reported being married/cohabitating/living with a partner (60.5%). Most participants reported being employed (52.1% full time, 20.9% part time). Over half of the participants had attained a university degree or higher education (57.3%). The most frequently reported health conditions were anxiety (23.7%), asthma (17.7%), and depression (15.3%). Five participants (2.3%) reported an eating disorder. No participant reported a diagnosis of hyperphagia, but 4 participants (1.9%) knew someone with hyperphagia. Mean body mass index was 25.3 kg/m² (range, 16.8–43.4 kg/m²). Demographic information is outlined in Table 1.

TABLE 1.

Participant demographics and clinical characteristics

Parameter	N (%)
Age, mean (SD), y	39.1 (14.3)
Sex, n (%)
Male	85 (39.5)
Female	130 (60.5)
Ethnicity, n (%)
White	181 (84.2)
Black	6 (2.8)
Hispanic/Latino	3 (1.4)
Other a	25 (11.6)
Marital status, n (%)
Single	65 (30.2)
Married/cohabitating/living with partner	130 (60.5)
Other b	20 (9.3)
Employment status, n (%)
Full‐time work	112 (52.1)
Part‐time work	45 (20.9)
Other c	58 (27.0)
Education level, n (%)
University degree or higher	123 (57.2)
No university degree	92 (42.8)
Physical characteristics
Height, mean (SD), m	1.7 (0.1)
Weight, mean (SD), lbs	163.9 (36.5)
BMI, mean (SD), kg/m2	25.3 (4.9)
Health conditions
Anxiety	51 (23.7)
Arthritis	17 (7.9)
Asthma	38 (17.7)
Diabetes	2 (0.9)
Depression	33 (15.3)
Eating disorder	5 (2.3)
Hypertension	11 (5.1)
Other health conditions	33 (15.3)
None	89 (41.4)
Diagnosed with hyperphagia	0

Abbreviations: BMI, body mass index; SD, standard deviation.

^a Other ethnicity: Mixed Asian (n = 1), mixed race (n = 4), mixed race: Black Caribbean and White (n = 2), mixed race: Black and White (n = 1), South American/Middle Eastern (n = 1), Spain (n = 1), Turkish (n = 1), and not specified (n = 5).

^b Other marital status: In a relationship (n = 2), long‐term relationship (n = 1), and partner (n = 1).

^c Other employment status: Carer (n = 1), child minder (n = 1), medical pension (n = 1), self‐employed (n = 2), and visual artist (n = 1).

3.2. Health state rankings and preferences

All but two participants (n = 213) ranked health states from most to least preferable as A, B, C, then D (Table 2). One participant ranked health states as B, A, C, then D, and another ranked health states as A, B, D, then C. These participants' responses were queried. One participant confirmed satisfaction with their response, and the other participant cited D as more likely than C to be identified as a health condition for which a person would seek treatment.

TABLE 2.

Health state rankings

Health state, n (%)	1 (Most preferred)	2	3	4 (Least preferred)
A (no hyperphagia)	214 (99.5)	1 (0.5)	0 (0.0)	0 (0.0)
B (mild hyperphagia)	1 (0.5)	214 (99.5)	0 (0.0)	0 (0.0)
C (moderate hyperphagia)	0 (0.0)	0 (0.0)	214 (99.5)	1 (0.5)
D (severe hyperphagia)	0 (0.0)	0 (0.0)	1 (0.5)	214 (99.5)

3.3. TTO utilities

Mean utility scores were highest for health states A (0.985), followed by B (0.909), C (0.702), and D (0.218) (Table 3 and Figure 1). Disultilites were calculated for health states B, C, and D with respect to health state A. Health state D had the greatest disutility (−0.767), followed by health states C (−0.283) and B (−0.076). All differences between health states were statistically significant (p < 0.0001). Utilities were also compared by subgroups. No differences in utility were found by age (Table S5) or sex (Table S6), with the exception of disutility of severe hyperphagia, which was significantly greater for women than for men (−0.830 vs. −0.670; p = 0.048). Most participants (209) rated all health states as better than dead (utility>0): no hyperphagia, 100.0%; mild hyperphagia, 100.0%; moderate hyperphagia, 98.1%; and severe hyperphagia, 76.7%. Six participants rated severe hyperphagia as equal to dead (i.e., utility = 0). Among participants who had at least 1 negative utility (indicating a health state percieved as worse than dead), 26 had at least 1 utility of −1. Some participants provided justification for their negative utilty rankings, presented in Table 4.

TABLE 3.

Health state utilities and disutilities

Health state	Utility score		Disutility score
	Mean	SD	Mean	SD
A (no hyperphagia)	0.985	0.022	–	–
B (mild hyperphagia)	0.909	0.103	−0.076	0.102
C (moderate hyperphagia)	0.702	0.302	−0.283	0.301
D (severe hyperphagia)	0.218	0.586	−0.767	0.582

Abbreviation: SD, standard deviation.

FIGURE 1.

Health state utilities and disutilities. Mean (95% confidence interval) utility and disutility values are reported for the various levels of hyperphagia. Severity levels are as follows: A, no hyperphagia; B, mild hyperphagia; C, moderate hyperphagia; D, severe hyperphagia. Disutility scores were derived as the utility difference between health state B, C or D and health state A.

TABLE 4.

Selected quotations from participants explaining reasons for negative utility scores

Selected quotations
“It would be really difficult to live like this. Worse than dead, as I knew someone with this condition, and it looked and felt horrible to see this person this way. It would really get in the way of daily life.”
“You get upset when denied food and your daily activities are severely impacted. I don't see the point. It also doesn't seem like you can have relationships which seems like a lonely life.”
“Worst one because you have severe problems doing daily activities and you're feeling discomfort every time you eat. It also gets in the way of your relationships. I would not like anything to do with this health state.”
“This has a very severe impact on mental health—constantly having that relationship with food, thinking about it all the time. I don't think I could live like that for 10 years.”

4. DISCUSSION

Results from this study generally follow expected patterns, with most participants ranking the most severe hyperphagia as least preferable with a utility that was lower than that of other health states. Several participants ranked severe hyperphagia as worse than dead and expressed concerns that hyperphagia would considerably affect daily activities and impair relationships. Together the data from this study show the profound impact that severe hyperphagia has on quality of life (QoL) for individuals with this condition and related disorders.

This study is the first to estimate the impact of hyperphagia on health state utilities independently of any specific underlying indication. Given that the health states only differed with regard to hyperphagia, and no differences were related to weight, obesity, or other comorbidities, any utility differences can be attributed to hyperphagia. Quantifying the independent effect of hyperphagia on health states allows for the evaluation of the impact this condition has on QoL in related disorders. The inclusion of comorbidities decreases health state utilities substantially. ³ , ¹⁵ An example of decreased utilities with comorbidities that includes hyperphagia can be observed in a recent study that sought to determine the effect of hyperphagia on disease burden in PWS, a rare genetic syndromic cause of obesity accompanied by cognitive impairment, behavioral problems, and hormone deficiencies. ² , ³ Health states of PWS were valued at 0.707 for those who did not include hyperphagia or obesity, 0.574 for those with just hyperphagia, and 0.384 for those with combined hyperphagia and obesity. ³ These results support the present study, which highlights the contribution of hyperphagia to the burden of rare genetic diseases of obesity and the additional impact of hyperphagia on QoL of afflicted patients.

The present study has several limitations. Utilities for severe hyperphagia are comparable to other severe health states, such as stroke, ¹⁶ progressive and metastatic cancers, ¹⁷ , ¹⁸ and severe chronic pain, ¹⁹ which similarly have a broad impact on many aspects of QoL. While this study was designed to isolate the unique utility impact of hyperphagia, it is important to remember that hyperphagia is generally accompanied by several comorbidities, such as diabetes and obesity. ² Comorbid conditions have been shown to have an additive or multiplicative effect with respect to utility values. ¹⁵ , ²⁰ Therefore, more research is needed to examine the utility impact of hyperphagia in the context of frequently associated comorbidities.

An additional factor not examined in this study is impact of hyperphagia on other family members and caregivers. Constant food seeking creates extraordinary stress on families, caregivers, and support systems. In studies of patients with PWS, family members showed poor QoL, difficulties in family functioning, communication problems, increased number of conflicts, high levels of depression and anxiety, and symptoms of post‐traumatic stress disorder. ²¹ , ²² Caregivers had an overall high level of burden that correlated with their depressed mood, anxiety, and sleep and work disruption. ²³ Inclusion of these effects would also contribute to a more complete assessment of hyperphagia utilities.

In this study, health state utilities for hyperphagia were derived from preferences of participants within the UK, the majority of whom were white, female, and on average middle aged. The health state with no hyperphagia in this study has similar utility value to those reported for normal healthy adults based on the EQ‐5D. ²⁴ Although the findings in healthy adults are similar to utilities found elsewhere, ²⁴ the present study may be limited in its generalizability, and subsequent studies should explore utilities associated with hyperphagia in other populations, including those with an equal sex distribution, across the age range to provide a more holistic understanding of these preferences.

Utility scores represent preferences for descriptions of various health states rather than the real‐world experience of actual patients with hyperphagia. While this vignette‐based study provides a reasonable estimate of utilities associated with hyperphagia, preferences of patients remain unknown at this time and may differ from preferences of individuals without hyperphagia. ¹⁵ It is also important to note that a consensus on the differentiation between mild, moderate, and severe hyperphagia has not yet been reached. Therefore, when using the utility values estimated in this study in a cost‐utility analysis, modelers will need to review the health state content to determine which health state utilities are the most appropriate for representing the populations being modeled.

Finally, another limitation may stem from the use of lead‐time TTO for health states perceived as worse than dead. Although the lead‐time approach does not increase the amount of negative scores, it does tend to result in a substantial amount of scores of −1, which may be lower than negative scores obtained via other methods. Therefore, the lead‐time approach could yield lower utility scores than other methods for health states that receive negative values (i.e., health states C and D in the current study). Despite this potential limitation, lead‐time TTO was used in this study because it is has been recommended by the EuroQoL group for health states perceived as worse than dead. ¹⁴

5. CONCLUSIONS

This research highlights the independent impact of hyperphagia on health state utilities. These data provide an estimate of hyperphagia utilities that can be incorporated into cost‐utility models conducted to inform decision‐making related to hyperphagia. The mean utility of the severe hyperphagia health state was substantially lower than the other hyperphagia health states, demonstrating its profound impact on QoL. These results aid in assessing the substantial impact of severe hyperphagia on QoL, underscoring the need for effective therapies to alleviate the burden to individuals with this condition.

AUTHOR CONTRIBUTIONS

Timothy A. Howell, Louis S. Matza, and Usha G. Mallya contributed to the study design and development, collected and analyzed data, made significant intellectual contributions and revisions to the manuscript, and approved the final version. Anthony P. Goldstone, W. Scott Butsch, and Ethan Lazarus contributed significant intellectual content and revision to the manuscript and approved the final version.

CONFLICT OF INTEREST

Louis S. Matza and Timothy A. Howell are employed by Evidera, a company that received funding from Rhythm Pharmaceuticals, Inc., for time spent on this research. Ethan Lazarus received payment or honoraria for lectures, speaker bureaus, advisory boards or educational events for Novo Nordisk, Currax Pharmaceuticals, Nestle Health Services, and the Obesity Medicine Association; and serves as president of the Obesity Medicine Association and a delegate of the American Medical Association. Usha G. Mallya is an employee of Rhythm Pharmaceuticals, Inc. As an employee, they receive stocks or stock options. Anthony P. Goldstone has been principal investigator for clinical trials sponsored by Rhythm Pharmaceuticals, Inc.; a member of the Data Safety and Monitoring Board for clinical trials for and received speaker honorarium from Novo Nordisk; and a consultant or member of medical advisory board for Millendo Therapeutics, Soleno Therapeutics, Helsinn Healthcare S.A., Evidera, Rhythm Pharmaceuticals, Inc., and Radius Health. W. Scott Butsch has served as a consultant or member of clinical and education advisory boards for Novo Nordisk; Rhythm Pharmaceuticals, Inc.; and Merck.

Supporting information

Supporting Information S1

Howell TA, Matza LS, Mallya UG, Goldstone AP, Butsch WS, Lazarus E. Health state utilities associated with hyperphagia: data for use in cost‐utility models. Obes Sci Pract. 2023;9(4):376‐382. 10.1002/osp4.652