Utilities Associated with the Treatment of Growth Hormone Deficiency (GHD): A Time Trade-off (TTO) Study in the UK and Canada

Introduction

Growth hormone deficiency (GHD) is a rare condition caused by an insufficient production of growth hormone (GH) in the pituitary gland.¹ In studies conducted in the US and Europe, the prevalence of childhood GHD has been reported to be between 1.8 and 2.9 per 10,000.^2–4 Additionally, a Danish study reported the incidence of childhood-onset GHD to be 1.7 for females and 3.6 for males per 100,000.⁵

GHD is one of the most important endocrine-related causes of short stature.⁶ In childhood, GHD is associated with symptoms such as absence of strength, lower energy levels, unproductive sleep and poor muscle development as well as reduced appetite.⁷ In addition to the physical symptoms, GHD has an impact on multiple aspects of daily life, including the child’s social and emotional well-being.⁷ Children with GHD also face a greater comorbid burden compared with demographically matched controls.⁸ Adults with GHD have increased cardiovascular mortality, lower muscle tone, increased central abnormal adiposity, hyperlipidaemia, low bone mineral density, psychological effects and poorer health-related quality of life (HRQoL) compared with the general population.^9,10

GH therapies have been given as daily subcutaneous injections with GH since the 1980s. However, newer GH formulations given as weekly subcutaneous injections have been available in the US since the 2020s.^10–15 Despite the recognised benefits of GH therapy, several aspects of GHD treatment have shown to be burdensome.^16–18 Factors such as storage, injection device, injection frequency, injection pain and preparation are some of the aspects with a high impact.¹⁷ Previous findings indicate that 43% of people with GHD miss injections if the medication must be stored in the fridge, whilst the figure is 24% if the storage requirements are more flexible.¹⁹ Furthermore, treatments associated with complex administration could lead to non-adherence, and relatively easy-to-use treatment options are more desirable amongst people living with GHD.^18,20 The treatment burden of GHD is also substantial for potential caregivers.^21–23 For instance, 62% of parents worry about treatment administration when giving GH therapy to their child, 38% worry about causing pain to their child and 15% are frustrated with the injection administration.¹⁶

To measure the impact on HRQoL amongst people living with a specific disease and to inform economic evaluations, several health technology assessment (HTA) agencies require health state utility values (HSUVs). A number of HTA agencies also recommend inclusion of HSUVs for potential caregivers in relevant cases²⁴ such as GHD due to the disease occurrence in childhood. Generic preference-based measures, including, eg EuroQol-5 Dimension (EQ-5D) and Short-form 6-dimension (SF-6D), are often the preferred method when generating HSUVs.²⁴ However, there are situations where these are insufficient, eg in evaluations of rare diseases, where patient populations are small. In these situations, vignettes are a preferred tool. In addition to being an alternative for generic measures, vignettes offer the opportunity for making more condition-specific evaluations and measuring even small differences in HSUVs related to a specific disease or treatment.^24,25 Time trade-off (TTO) methodology is an accepted preference elicitation technique when estimating HSUVs through the evaluation of vignettes. The TTO methodology is acknowledged by several HTA agencies, including the National Institute for Health and Care Excellence (NICE) in the UK and the Canadian Agency for Drugs and Technologies in Health (CADTH) in Canada.^24–27

Whilst the methodology for evaluating health in an adult population is well established, robust guidelines for the most appropriate approach for children and adolescents have not yet been developed.²⁸ Earlier research has identified multiple challenges when valuing health amongst children, even though several methods have been applied and discussed.^28–30 One option is to ask a population of children and adolescents, since they can relate to how the condition could impact their lives. However, several contrary arguments emphasise that asking children and adolescents is inadequate, perhaps because they might not be able to understand the TTO questions.³¹ Another approach could be to ask adults to imagine being a child. However, earlier studies have found that the willingness to trade life years is strongly impacted by the perspective used in the survey as well as how the perspective is described.^29,32 Additionally, imagining being a child is difficult, which results in variations in what is imagined by the respondents.²⁹ These issues when valuing health amongst children make it challenging to generate HSUVs related to GHD treatment, and different approaches can be applied.

The aim of this study was to estimate utility gains or disutilities associated with GHD treatment for both people living with the disease, including children, and potential caregivers to inform future economic evaluations. This was investigated by evaluating several treatment aspects using TTO surveys in the UK and Canada.

Materials and Methods

The TTO Methodology

When using the TTO methodology, respondents are asked to choose between two health states (HSs): An impaired HS for an amount of life, t, and an HS in full health but for a shorter amount of life, x. By choosing the preferred HS, respondents “trade” a part of their life in order to live in full health instead of living in an impaired HS.

To identify the point where the two presented HSs are equally attractive to the respondents (point of indifference), they are repeatedly asked to choose between the impaired HS versus full health. Respondents who choose the impaired health state will be asked to trade a smaller number of life years in the next question, while respondents who choose to trade life years to live in full health are asked to trade a larger number of life years in the next question. Thus, for each trade, the number of years lived in full health, x, is varied until the point of indifference is identified, x*. An HSUV between 0 and 1 is calculated based on this point of indifference. The HSUV is calculated as x*/t (0=death, 1=full health).^33,34

Study Population

When using the TTO methodology to elicit HSUVs for health economic evaluations, it is generally recommended by HTA agencies to use HSUVs elicited from the general population, ie not the population living with the condition.^24–27 This recommendation is often based on the idea of social decision-making, which reflects that healthcare decision makers are acting on behalf of the general population. Therefore, the population in this study consisted of the adult general population in the UK and Canada.

To manage the challenges when valuing health amongst children, two surveys were developed based on discussions in the literature on the best approaches for measuring child health.^28,31 In the first survey, HSUVs of the adult general population were used as a proxy for children. In the second survey, utility values of the adult general population were elicited by asking the respondents to imagine that they were diagnosed with GHD as a child. To value health amongst caregivers, a third survey was developed in which an adult general population was asked to imagine themselves being a caregiver of a child living with GHD. For all three surveys, the inclusion criteria were (1) consent to participate and (2) at least age 18 years. As recommended by Powell et al and to increase relatability, only respondents with at least one child under the age of 15 were included in the third survey.²⁹

Respondents who had previously agreed to participate in internet-based surveys were recruited through existing Email panels. Respondents were rewarded (cash) points equivalent to EUR/CAD 1–2 for participation. The surveys were answered anonymously, and no sensitive information was collected or revealed. Additionally, the surveys were conducted according to the codes of conduct of the European Society for Opinion and Market Research (ESOMAR) and the Canadian Research Insights Council (CRIC) and followed the applicable guidelines. Since this study was not a clinical trial, it did not include patients, it did not gather biological or human samples or identifiable personal information, it was not carried out or funded by Health Canada or Public Health Agency of Canada, and it was not taking place in a National Health Service (NHS) setting; ethical review board approval was not required in Canada and the UK (as described by NHS Health Research Authority in the UK and in the Canadian TCPS2 2022).

Description of Disease and Definition of Health States

In the beginning of each of the three surveys, respondents were presented with an introduction to GHD and a description of the perspective (Figure 1). In the first survey, respondents were asked to imagine having a hormone deficiency disease, and in the second survey, respondents were asked to imagine that they were diagnosed with GHD at the age of 5. In the third survey, respondents were asked to imagine having a 5-year-old child diagnosed with GHD.

Figure 1 Description of GHD applied in the three surveys.

Based on input from clinical experts and findings from a previous qualitative study of the treatment burden for children living with GHD and their caregivers,¹⁶ the following five GHD treatment aspects were evaluated in each of the three surveys: injection frequency, device complexity, needle visibility, injection pain and storage possibilities. To make the results as accurate as possible, the treatment aspects were defined using product information from existing treatment options and a definition of injection pain from a previous Phase 3 clinical trial.³⁵ Table 1 presents the final definition of the five treatment aspects, and Figure S1 provides an example of how the treatment aspects were described to the respondents.

Table 1 Overview of Evaluated Aspects in the Treatment of GHD

To elicit the utility gain or disutility associated with each of the five GHD treatment aspects, eight HSs were designed and included in each of the three surveys. Table 2 presents an overview of how the GHD treatment aspects were described in the eight HSs.

Table 2 Overview of Health States Evaluated in the Three Surveys

Survey Design

The surveys were programmed in a commercial survey software package (SurveyXact). To make the trade-offs as realistic as possible, the time horizons applied in the surveys varied depending on each respondent’s individual life expectancy, calculated using the respondent’s age and gender as well as the most recent lifespan tables from the World Health Organization.³⁶

The point of indifference was identified by presenting each trade-off between an impaired HS and a full health HS four to six times but with different lifetimes for full health (always shorter than the impaired HS). For each choice made by the respondents, the difference in the lifetime for the full health option was narrowed until the point of indifference was identified. This procedure followed a standard bisection methodology, using a utility of 0.6 as a starting point to reduce the utility value to an interval of 0.025.

When using an online survey to estimate HSUVs, it is crucial that the respondents understand and accept the premises of the research and that they continue to be motivated to provide considered answers throughout the survey. Thus, to increase the validity of results, the following three features were built into the survey design. First, to familiarise the respondents with the TTO concept and to test their understanding of the methodology, a “warmup” exercise was implemented as the first HS in all surveys. The exercise included a TTO test question with a choice of (1) full health and a long remaining lifetime or (2) impaired health and a reduced lifetime. Respondents who chose the second option were excluded from the TTO analysis. Second, respondents who chose not to trade any lifetime or who chose to trade a very large proportion of their remaining lifespan were screened carefully before inclusion in TTO analyses. Respondents were excluded from the analysis of the specific HS if they reported that their choice was due to ethical or religious beliefs or if they stated that they did not understand the question. Respondents who reported that their choice was due to a manageable HS or a desire to live as long as possible, eg due to obligations in their life, were kept in the further analysis. Finally, to avoid fatigue and thereby potentially less-considered trade-offs, respondents only received trade-off scenarios for five of the eight HSs. The order in which HSs were presented to respondents was randomised to ensure that the results for each HS were not affected by the order.

The functionality of the three surveys was tested in a pilot study in the UK. In the second survey, an unusually high dropout rate of respondents was observed (28% vs 8% in the first survey and 4% in the third survey) due to their not understanding the test question. Based on these results as well as recommendations from the literature, it was decided to elicit HSUVs for people with GHD using only the first survey. Thus, the main data collection was completed for the first and third surveys only (hereafter, Survey A and Survey B). Survey A was conducted to elicit utilities for people with GHD, including children, and Survey B was conducted to elicit utilities for caregivers of children with GHD. The main data collection was conducted, in both the UK and Canada, from October to December 2022.

Statistical Analysis

The aim was to estimate the average utility gain or disutility associated with different aspects of GHD treatment (Table 1). Based on the respondents’ trade-offs, an HSUV was calculated for each of the eight HSs. The value represents the midpoint of the chosen range of indifference. Using these HSUVs for all respondents, an average HSUV for each HS was estimated. Thereafter, utility gains/disutilities were calculated as the difference between the average HSUVs of two HSs. For instance, the only difference between HS 1 and HS 2 was the presence of injection pain (Table 2). Thus, by calculating the difference between the average HSUV elicited for the two separate HSs, a utility gain/disutility for avoiding injection pain could be estimated.

To mitigate the risk of including people who do not provide considered responses, the 5% most extreme values were excluded (2.5% at each end). This ensures better confidence in the results and makes the analyses less sensitive to outliers. Non-parametric bootstrapping with 10,000 iterations was used to simulate standard errors and confidence intervals (CIs) and to test differences between the parameters.

Statistical analyses were conducted using SAS version 9.4 statistical software.

Results

Study Population

In total, 2026 and 2028 respondents completed Survey A and Survey B, respectively. Of the respondents completing Survey A, 1025 were from the UK and 1001 were from Canada. Of the respondents completing Survey B, 1017 were from the UK and 1011 were from Canada. 244 and 350 respondents failed the TTO test question in Survey A and Survey B, respectively. Thus, 1782 and 1678 respondents were included in the further analyses (Figure 2).

Figure 2 Respondent flowchart.

Baseline characteristics are presented in Table 3. In Survey A, the study population was almost equally distributed by age group and gender. The majority were in full-time employment (45%) and had a bachelor’s degree or higher (43%). In Survey B, the majority of the population was aged 30 to 50 years, most likely due to the inclusion criteria of having at least one child under the age of 15. Additionally, most respondents were female (55%), were in full-time employment (63%) and had a bachelor’s degree or higher (53%).

Table 3 Characteristics of Study Population

TTO Results for Survey A

Table 4 presents utility gains/disutilities associated with the five aspects of GHD treatment when the population evaluated HSs as if they took injections themselves.

Table 4 Utility Gain or Disutility Associated with Different Aspects of GHD Treatment Elicited with Survey A

Injection pain was considered the most important treatment aspect amongst the general population in both the UK and Canada. Avoiding weekly injection pain was associated with a significant utility gain of 0.030 in both countries (95% CI UK: 0.025–0.036, p<0.001, Canada: 0.023–0.036, p<0.001, total: 0.026–0.035, p<0.001).

A less complex injection device and a lower injection frequency were also preferred in both countries. In total, a less complex injection device was associated with a utility gain of 0.020 (95% CI 0.016–0.025, p<0.001) (UK: 0.019 (95% CI 0.013–0.026, p<0.001), Canada: 0.023 (95% CI 0.016–0.030, p<0.001)), and weekly instead of daily injections were associated with a significant utility gain of 0.009 (95% CI 0.005–0.014, p<0.001) (UK: 0.009 (95% CI 0.004–0.015, p=0.001), Canada: 0.010 (95% CI 0.003–0.017, p=0.008)).

In Survey A, there were no significant results for avoiding a visible needle at the time of injection or for different combinations of storage possibilities (Table 4).

All results for Survey A are illustrated in Figure 3.

Figure 3 Illustration of utility gain or disutility associated with different aspects of GHD treatment (Survey A). Abbreviation: RT, Room temperature.

TTO Results for Survey B

Table 5 presents utility gains/disutilities associated with the five treatment aspects when the general population evaluated HSs as if they gave injections to a child.

Table 5 Utility Gain or Disutility Associated with Different Aspects of GHD Treatment Elicited with Survey B

In the analysis, injection pain was considered the most important aspect of GHD treatment in both the UK and Canada. In total, avoiding a weekly injection that causes injection pain to one’s child was associated with a utility gain of 0.044 (95% CI 0.038–0.051, p<0.001). In the UK, the utility gain was 0.039 (CI 95% 0.030–0.048, p<0.001), and in Canada the utility gain was 0.050 (CI 95% 0.040–0.060, p<0.001).

When using aggregated data, a less complex injection device and a lower injection frequency were also preferred for caregivers. In total, a less complex injection device was associated with a utility gain of 0.008 (95% CI 0.002–0.014, p=0.006), and weekly instead of daily injections were associated with a significant utility gain of 0.009 (CI 95% 0.003–0.014, p=0.003). When using country-specific data, the utility gains were statistically significant in Canada but not in the UK (Table 5).

Survey B found no significant results for avoiding a visible needle at the time of injection. For different combinations of storage possibilities, the findings in Canada showed that the option of storing the medication at room temperature for up to 72 hours was associated with a significant disutility of −0.011 (95% CI −0.021– −0.002, p=0.019) compared with storage in the fridge. Storage factors were not associated with a utility gain/disutility in the UK or when data from both countries were combined (Table 5).

All results for Survey B are illustrated in Figure 4.

Figure 4 Illustration of utility gain or disutility associated with different aspects of GHD treatment (Survey B). Abbreviation: RT, Room temperature.

Discussion

This study applied a TTO methodology to estimate utility gains or disutilities associated with GHD treatment for both people living with the disease and potential caregivers. Avoiding weekly injection pain was the aspect associated with the highest utility gain amongst the five evaluated treatment aspects for both people living with GHD and caregivers. Additionally, a less complex device and injection frequency were associated with significant utility gains for both people with GHD and caregivers.

The results emphasise that injection pain is a fundamental aspect of GHD treatment. The importance of injection pain is in line with earlier evidence reporting that 54% of children with GHD describe injection pain as a burden, and 38% of caregivers report that they worry about causing pain to the child.¹⁶ Additionally, the findings suggest that device complexity and injection frequency also have an impact on especially people with GHD. These findings are in agreement with other studies showing that injection devices should be easy and comfortable to use and that “ready to use” devices are preferred amongst people living with GHD.^{18,20,21,23,37} Previous studies have also shown that less frequent injections are highly preferred due to greater convenience and reduced life interference.^16,38

The derived utility for living with or treating a child with GHD is similar to that of living with circulatory system or muscular diseases.³⁹ More recent studies have also investigated the disutilities for events within other endocrine disorders, including diabetes. In comparison, the findings of this study suggest giving a child weekly painful injections is associated with approximately the same disutility as experiencing monthly non-severe daytime hypoglycaemic events and a disutility comparable to experiencing a severe hypoglycaemic event once a year when living with diabetes.⁴⁰ Avoiding the weekly injection pain when self-injecting is associated with a disutility comparable to experiencing a severe hypoglycaemic event every two years.⁴⁰ Additionally, the identified disutility associated with the treatment of GHD is comparable to previous reported decreases associated with the treatment of type 2 diabetes. The study findings suggest that avoiding a complex injection device for the treatment of GHD is associated with a utility gain similar to that of avoiding injections that require reconstitution, waiting time and needle handling for the treatment of type 2 diabetes.^41,42 However, comparisons of utility gains/disutilities between diseases and across different populations should be treated with caution.

Neither needle visibility nor varying storage possibilities were associated with significant utility gains/disutilities in this study. These findings are in opposition to other literature suggesting that needle-free devices are preferred by both people with GHD and caregivers²³ and that refrigeration requirements are burdensome.^16,19,22 The difference in results could be explained by differences in study designs. The current study examines quite different treatment aspects, and some of the HSs include complex and detailed information. In trade-offs where respondents were provided with a lot of information, there is a risk that each respondent will put more focus on specific treatment aspects and potentially neglect the impact of other aspects.

This study was challenged by the currently missing guidelines for measuring and valuing child health. Thus, two surveys taking different perspectives were tested. A very high dropout rate at the beginning of the second survey in the pilot study indicated difficulties amongst respondents in understanding the hypothetical exercise of imagining being diagnosed with GHD at the age of 5. Therefore, the second survey was excluded from the main data collection. Due to the varying findings in the research field and to maintain consistency in measures, it has been recommended to use the adult general population perspective when evaluating children’s health.³¹ Thus, the utility gains/disutilities elicited using Survey A can be considered a proxy for how different aspects of GHD treatment impact children. As a solution for obtaining more precise estimates when using an adult population as a proxy for child health, quality adjusted life year weighting or deliberation has been suggested.²⁸ Since there are no clear guidelines for applying this method and no major HTA agencies currently accept this approach as valid evidence, this approach was not applied in the present study.

Based on the findings of this study as well as those of previous studies, people and caregivers who live with GHD seem to face a treatment burden, which should be considered for inclusion in future health economic evaluations and HTA decisions. The findings underline the value of treatment options that are associated with a minimal amount of pain, are easy to use and are administered at a frequency less than daily. Choosing options associated with these features in GHD treatment has the potential to increase HRQoL amongst both people with GHD, including children, and caregivers.

This online TTO survey highlights both strengths and limitations. First, one limitation is that people with GHD and caregivers were not directly included in the development and validation of the surveys. To mitigate this limitation and to ensure that the wording used was representative of the investigated treatment aspects, both clinical experts and findings from a previous qualitative study of the treatment burden for children with GHD and their caregivers were included in the development of the surveys.¹⁶ In addition, the validity of the study findings is enhanced by similar results from the UK and Canada. Second, the approach implies a hypothetical exercise which is not targeted to people with GHD. However, this follows the recommendation from several HTA agencies, including NICE and CADTH, about eliciting HSUVs in a general population.^24–27 The hypothetical approach in the general population provides a large sample size, which provides more robust results. Third, the internet-based design of the study does not provide an opportunity to explain HSs further in case of misunderstandings, and it involves a risk of including respondents who do not accept or understand the premises of the methodology. On the other hand, the approach ensures that HSs are presented in the same way for all respondents, making the trade-offs more comparable. In addition, several features were built into the survey design, making it possible to exclude respondents who did not understand or accept the methodology. Finally, the study design might have caused bias in the results if internet users in general evaluate HSs differently from non-internet users. Since the UK and Canada have a 98% and a 94% internet accessibility rate, respectively, that poses a minor limitation.⁴³ At the same time, the internet-based approach increases the opportunity for obtaining a representative sample compared with an interview-based design.

Conclusion

This study finds that several aspects of GHD treatment are associated with a significant impact on the associated HSUVs amongst both people living with GHD, including children, and potential caregivers. Based on the results, obtained from a large sample of respondents from the UK and Canada using an online TTO survey, treatment options without (1) injection pain, (2) a time-consuming and complex injection process and (3) daily injections are associated with utility gains. Therefore, these treatment aspects could be expected to result in higher HRQoL.

These study results can be used to inform future economic evaluations of GHD treatment. In addition, the study may inform future treatment choices.