Patient and Caregiver Health State Utilities in Lennox-Gastaut Syndrome and Dravet Syndrome

Patient and Caregiver Health State Utilities in Lennox-Gastaut Syndrome and Dravet SyndromeOriginal Research
Lennox-Gastaut Syndrome and Dravet Syndrome
Siu Hing Lo, PhD
2 ; and
2 GW Pharma Ltd, London, United Kingdom
ABSTRACT
Dravet syndrome (DS) are rare, treatment-resistant forms of epilepsy characterized by childhood onset of seizures. Cost-effectiveness analysis for new antiseizure medications typically requires health state utilities (HSUs) that reflect the patient and caregiver burden of the relevant conditions. This study aimed to estimate HSUs for LGS and DS. Focus was placed on valuing
the impact of seizure frequency and seizure-free days on health-related quality of life (HRQL) for patients and caregivers.
Methods: Health state vignettes that described the experience of living with and caring for a child with
LGS or DS were developed based on a targeted
literature review and feedback from interviews with
LGS and DS clinical experts and DS caregivers. Vignettes varied by the number of seizures and
seizure-free days per month. Twenty-four vignettes were developed that described patient HRQL (6 per condition) and caregiver HRQL (6 per condition) for LGS or DS. Vignettes were evaluated via interviews from the general population in the United Kingdom
and Sweden using a visual analog scale and time trade-off (TTO) method. Participants were recruited
by interviewers based in different regions of the United Kingdom and Sweden using convenience sampling.
Findings: Two hundred interviews were conducted
by video call from March to April 2020. One hundred
participants evaluated each of the 6 patient and 6
caregiver vignettes for LGS; a different 100 participants evaluated each of the DS vignettes. The estimated utility
values vary consistently according to seizure frequency
and number of seizure-free days. Patient TTO utility
values range from −0.186 (highest seizure frequency
and fewest seizure-free days) to 0.754 (seizure-free state) for LGS and from 0.171 to 0.778 for DS. The caregiver TTO utility values range from 0.032 to 0.810
and 0.510 to 0.881 for LGS and DS, respectively. Fewer seizures and additional seizure-free days are associated
with better patient and caregiver HRQL. Implications: This study estimated utility values
for patients with LGS or DS and their caregivers using visual analog scale and TTO valuation among a
general public sample. These utility values can be used
in cost-effectiveness assessments of new antiseizure medications. The findings indicate the importance of seizure-free days as well as seizure frequency in patient and caregiver HRQL, which may inform the design
of future clinical trials. ( Clin Ther. 2021;43:1861– 1876.) © 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC- ND license ( http://creativecommons.org/licenses/by- nc-nd/4.0/ )
Key words: cost-effectiveness analysis, Dravet syn- drome, health state utilities, health technology assess- ment, Lennox-Gastaut syndrome.
November 2021
INTRODUCTION
Lennox-Gastaut syndrome (LGS) is a rare, severe form of epilepsy that typically becomes apparent in childhood, is associated with an age at onset between
Accepted for publication September 24, 2021 https://doi.org/10.1016/j.clinthera.2021.09.017 0149-2918/$ - see front matter
© 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Clinical Therapeutics
1 and 8 years,1 , 2 and is characterized by frequent seizures.2 Some types of seizure are particularly hazardous because they may cause patients to fall if unsupported because of a temporary loss of muscle tone (atonic seizures) or stiffening of muscles (tonic seizures).1 , 2 Seizures that result in sudden falls are collectively referred to as drop seizures and can result in severe injuries.1 , 2 Dravet syndrome (DS) is another rare, severe form of epilepsy that typically becomes apparent in childhood and is characterized by frequent, prolonged epileptic seizures.3 , 4 General mobility issues have been reported in both LGS and DS.5–7 Most patients use a wheelchair or adaptive stroller as their primary means of mobility.6 , 7 Most people with LGS or DS experience behavioral difficulties 2 , 8 , 9 and have cognitive impairments, which may affect their ability to communicate.5 , 10 Reduced appetite and eating difficulties have also been reported in patients with DS.10 , 11 Patient mobility issues, cognitive impairments, behavioral difficulties, and feeding issues can all negatively affect caregiver burden.5 , 11 , 12
Both LGS and DS are associated with treatment- resistant epilepsy, and there is a need to develop new treatment options for patients with these condi- tions.13 , 14 The emergence of new treatments will trigger health technology assessments (HTAs) or comparative effectiveness research to determine their benefits for patients and health care professionals. HTA reviews can include a cost-effectiveness analysis for each treatment, reflecting the gain in health measured by both the length and the quality of life. Health state utilities (HSUs) are a measure of health-related quality of life (HRQL) and reflect the value that society places on gains in patients’ health status.15 Health gains from an effective treatment are not restricted to patients alone and may extend to caregivers. This issue is included in the National Institute for Health and Care Excellence (NICE) reference case and is the subject of a recent report from the NICE Decision Support Unit.16 There is substantial evidence that caring for a child with refractory epilepsy may significantly impair the health of parents or caregivers.5 , 8 , 17 Importantly, efficacious treatments may reduce the caregiver burden, which is an important consideration when conducting cost-effectiveness analyses.
Different methods exist for capturing HSUs.18
Utilities can be estimated using data from the EQ-5D or similar standardized measures. However,
1862
in the context of rare pediatric epileptic syndromes, the use of standardized measures is problematic. Self-completion of the youth version of the EQ- 5D (EQ-5D-Y) would not be appropriate for most pediatric patients with LGS and DS, who typically have cognitive impairments. Furthermore, the EQ-5D has been criticized for not capturing epilepsy-specific outcomes and potentially having poor responsiveness and sensitivity in seizure disorders.19 , 20 The very low prevalence of LGS and DS makes it difficult practically to recruit sufficiently large samples of patients or caregivers for (proxy) HRQL estimation across a large number of health states. The classification of health states in LGS and DS requires data from people who experience specific numbers of seizures of each type. The few available utility analyses in LGS and DS that exist only report on a limited number of health states.21–23 For these reasons, a vignette study was conducted rather than a standardized questionnaire. This approach used bespoke vignettes to describe the specific impact of the condition (LGS or DS) in narrative form and devise preferences for a specific health state.24
Previous work by Radu et al 25 described LGS and DS states in terms of seizures and seizure-free days. These measures represent 2 dimensions of seizure outcomes that are deemed to be patient relevant by clinical, patient, and caregiver experts and that have been included in previous cost-effectiveness analyses in refractory epilepsy.25–27 However, states were only assessed using a visual analog scale (VAS), which was completed by caregivers and patients. Many HTA bodies prefer a societal perspective and the use of a choice-based preference task, such as time trade-off (TTO). The aim of the present study was to elicit HSUs or utility weights for patient and caregiver health states related to LGS or DS through TTO interviews with the general public. This vignette research was designed to evaluate the specific impact of seizures and seizure-free days on patient and caregiver HRQL in LGS or DS.
PARTICIPANTS AND METHODS
Health state vignettes were developed to describe children living with LGS or DS and their caregivers. Patient vignettes consisted of an account that provided the following details on a hypothetical patient with LGS or DS: number of drop (LGS) or convulsive seizures (DS) per month and number of seizure-free
Volume 43 Number 11
S.H. Lo et al.
days per month. Seizure frequency is a common end point in clinical trials, but there is an increasing body of evidence that patients and their families place a high value on seizure-free days.28–30 The inclusion of both variables allows for flexibility in the application of these utilities. The composition and variation in attributes across patient and caregiver health states are described in Supplemental Figures 1 to 4 for different seizure frequencies and severities. Categories were defined to broadly cover the distribution of seizure frequencies and seizure-free days, while maximizing differentiation among states, observed in patients who participated in clinical trials conducted by the sponsor.
Caregiver vignettes corresponded with the patient vignettes to provide relevant context. The caregiver vignettes described the perspective of a person who is 1 of 2 primary caregivers/parents of a 10-year-old child. The vignettes included content regarding the caregiver’s own health and daily life alongside the patient for whom they provided care. Vignettes were initially developed using data from a literature review. They were then validated and refined using feedback from interviews with caregivers and clinical specialists. Participants from the general population in the United Kingdom and Sweden then assessed the vignettes using the TTO interview method.
Targeted Literature Review
A PubMed literature search was conducted to identify publications that reported the experiences of patients with LGS or DS and their caregivers. The review aimed to identify studies that reported research on the HRQL impact/burden of LGS or DS. The impact of seizures on HRQL was emphasized during the search. The following search terms were used: severe myoclonic epilepsy of infancy OR SMEI OR Dravet syndrome or Lennox-Gastaut syndrome OR LGS AND impact OR symptom
∗ OR burden OR quality of life .
Draft vignettes that described the experience of patients with LGS or DS and their caregivers were developed based on findings from the literature review. The initial draft states were used as a basis for discussion with caregivers and clinical experts.
Interviews With Caregivers and Clinical Experts Six caregivers of patients with DS were recruited
via Dravet Syndrome UK
31 and the Dravet Syndrome Association Sweden.32 Caregivers reviewed the seizure-
November 2021
free DS health state and the health state that most closely matched their own experience. No caregivers of patients with LGS were recruited because no LGS- focused patient organizations were identified in the United Kingdom or Sweden. The LGS and DS clinical experts were recruited based on their substantial experience of clinical practice in managing patients with these diseases. Two clinicians experienced in treating patients with LGS or DS were interviewed: 1 clinician was an expert in the treatment of pediatric populations, whereas the other clinician specialized in the treatment of adult patients. Both clinicians reviewed all LGS and DS vignettes.
Interviews were conducted using a semistructured interview guide, and caregivers and clinical experts reviewed the content of the draft vignettes. Interviews were designed to validate and refine the content of the vignette and fill in evidence gaps that were not addressed by the literature review. Draft health states were iteratively revised based on notes from the caregiver and clinical expert interviews, ensuring that subsequent interviewees reviewed incremental changes made to the drafts. Interviewees were asked to provide feedback on the relevance and comprehensiveness of the vignettes in terms of the patients’ and caregivers’ HRQL. Key changes made to the LGS patient vignettes following clinical expert feedback are summarized in Supplemental Table I; key changes made to LGS and DS caregiver vignettes following clinical expert and DS caregiver feedback are summarized in Supplemental Table II. When similarities between LGS and DS were confirmed by clinical experts, the final DS vignettes were used to update the LGS vignettes. Similarly, differences between the LGS and DS vignettes were based on confirmed differences between these conditions. A final round of vignette revisions ensured that patient and caregiver vignettes only differed in aspects where the participants’ perspective justified variations in wording and content, such as an adult self as the patient versus self as 1 of 2 primary caregivers of a 10-year-old child with the condition.
Health State Valuation
TTO is a standardized interview method for valuing health states that was used to generate health utility weights for each vignette.33 , 34 The method is designed to determine the point at which participants consider 10 years in the target health state to be equivalent to the prospect of X years in full health. Time in full health
1863
Clinical Therapeutics
is varied until this point of indifference is reached, such that the participant views them as being equivalent. To minimize possible bias, we used the ping pong method, which varies the amount of time in full health alternating between high and low values, changing by 6-month intervals. If a participant indicates that he/she believe that zero years in full health is preferable to any time living in a health state, then this indicates that the participant considers the state to be worse than dead. At this point, the interviewer switches to a lead-time TTO exercise, which asks participants whether they would prefer to live for 10 years in full health followed by 10 years in a health state or to live for 10 years in full health.35 This lead-time procedure allows the participant to trade more years of life to determine how much worse than dead they consider the health state to be.
Members of the UK and Swedish general population were recruited by 5 interviewers trained in the TTO ping pong method based in different regions of the United Kingdom using convenience sampling, including snowballing using informal networks of earlier participants, in March to April 2020; quotas were set for age and sex. The aim was to recruit a broadly representative sample of the UK and Swedish general population in terms of age and sex. Participants were eligible if they were adults ( ≥18 years of age). To avoid upsetting the participants, caregivers and parents of patients with epilepsy were not eligible. A total of 200 TTO interviews were conducted. One hundred participants evaluated each of the 6 patient and 6 caregiver vignettes for LGS; a different 100 participants evaluated each of the DS vignettes. In response to the UK COVID-19 lockdown at the start of the data collection, all interviews were conducted using online video (Skype or Zoom), except for 5 face-to-face interviews. Participants were provided with information about the study and asked to complete a consent form along with a brief background questionnaire.
Interviewers requested participants to have vignettes printed or presented on a separate screen. The online video interview was presented on a second screen (if vignettes were not printed) so that the interviewer and participant were visible to one another at all times, enabling the VAS and TTO board to be shown alongside the interviewer on the screen. The online interview encouraged eye contact and was designed to mimic a standard face-to-face interview. The TTO-
1864
trained interviewers conducted all interviews and were instructed to complete all patient vignettes in one block and all caregiver vignettes in another block, each in a quasirandom order (paper cards were shuffled by the interviewer to determine the order).
Participants were first asked to read the introductory text (Supplemental Figures 5 to 8) that asked them to imagine they had the condition themselves or they were a primary caregiver of a child who had the condition. The conditions were only referred to as epilepsy A (LGS) or epilepsy B (DS). The first exercise used a VAS that ranged from 0 (worst possible state) to 100 (full health). As a warm-up exercise, participants were asked to rank 2 practice vignettes with simple health states, unrelated to the LGS and DS vignettes. Six patient or caregiver vignettes and the health state dead were then presented one by one. After the VAS exercise, participants completed the TTO exercise for patient or caregiver vignettes.
Statistical Analysis Sociodemographic data were summarized descrip-
tively. The VAS ratings for each vignette were rescaled such that the value for the dead state was fixed at zero and all other values varied between 100 and the worse health state. The following formula was used to rescale the data
V
where V
′ is the rescaled VAS value, V is the original VAS value, and V_Dead is the value given to the dead state. After rescaling, the VAS data were summarized descriptively. The TTO data were scored according to the point of indifference. Correlation analysis was conducted between the mean of the 6 patient TTO values and the mean of the 6 caregiver TTO values of each condition (ie, LGS or DS) using pairwise correlations.
RESULTS
Literature Review Findings The literature search revealed 1467 articles. Only
the first 400 results, organized by relevance, were reviewed by title and abstract for the targeted literature review. A total of 12 publications were selected for inclusion (Supplemental Table III). Studies were included if they had collected data in an LGS or DS patient or caregiver population and reported findings from research on the HRQL impact/burden of these
Volume 43 Number 11
S.H. Lo et al.
conditions. Data on the overall LGS or DS burden, HRQL, and the specific impact of seizures in these conditions were extracted.
Patient HRQL
Overall, patient HRQL was low in both LGS and DS.8 , 36 , 37 Factors that contributed to low patient HRQL included both seizure-related factors and other factors related to the comorbidities of LGS and DS. Seizures in LGS and DS have a significant negative impact on patient HRQL.5 , 6 , 36 , 38 High seizure frequency and severity are associated with a greater number of comorbidities, emergency treatment events, and injuries.36 , 37 The use of wheelchairs and helmets to protect patients from injury when having a seizure has been reported in LGS.5 Seizures in LGS can affect self-care and continence and cause sleep disturbance attributable to nocturnal seizures.5 , 7 Seizures can limit social and leisure activities because of fear of injury from falls and fear of seizure triggers.5 , 10 Reducing seizure frequency in LGS may result in fewer injuries and fewer days missed at school, which in turn may improve social or family relationships.5
Caregiver HRQL
For both LGS and DS, caregivers of patients have reported considerable effects on their overall HRQL.5 , 6 , 8 , 12 The demands of caregiving can result in caregivers having little or no time for them- selves.5 , 6 , 10 , 38 Moreover, caregivers often experience stress and anxiety as a result of financial con- cerns.5 , 6 , 8 , 37 , 38
In the event of a seizure, caregivers may be required to intervene at any given moment to avoid injuries to their child, administer rescue medication, or make emergency hospital visits.5 , 6 , 38 Physical injuries related to seizures have been reported.5 The unpredictability of seizures can affect the caregivers’ and siblings’ ability to partake in recreational activities, resulting in a withdrawal from social life.5 , 38 Seizures have a profound impact on caregivers’ emotional well- being. Caregivers have reported stigma and social isolation.5 , 6 , 38 Caregivers also report poor sleep, fatigue, emotional exhaustion, depression, stress, and anxiety.5 , 6 , 37 , 38
Health State Valuation
Examples of the final vignettes for patients with LGS, LGS caregivers, patients with DS, and DS
November 2021
caregivers are shown in Figures 1 to 4 , respectively. Demographic characteristics of the TTO interview participants indicated that our samples were broadly representative of the general population, in terms of age and sex, in the United Kingdom and Swe- den ( Table I ). Mean age of the UK participants was 40.3 years compared with a median age of 39.4 years in the general population.39 For the population in Sweden, these values were 37.8 and 40.9 years.40 Sex distribution in each sample was close to that of the general population (male: UK…