Introduction
This chapter uses material from an Open Access article previously published by the research team [see Kyle SD, Siriwardena AN, Espie CA, Yang Y, Petrou S, Ogburn E, et al. Clinical and cost-effectiveness of nurse-delivered sleep restriction therapy for insomnia in primary care (HABIT): a pragmatic, superiority, open-label, randomised controlled trial. Lancet 2023;402(10406):975–87. https://doi.org/10.1016/S0140-6736(23)00683-9. Epub 10 Aug 2023. PMID: 37573859]. This article is published under licence to The Lancet. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/.
This chapter presents the health economic evaluation conducted as part of the HABIT trial. The base-case analysis took an NHS and PSS perspective and assessed the cost-effectiveness of nurse-delivered SRT relative to SH alone for insomnia in primary care. Health care and PSS resource utilisation data and EQ-5D-3L utility data were collected alongside clinical data. These data were used to conduct a cost–utility analysis, calculating the incremental cost per QALY gained during the 12-month trial period as the primary outcome of the economic analysis. Sensitivity analyses were conducted to explore how the result was affected by altering several key features of the economic evaluation, including (1) complete-case analysis without data imputation, (2) adopting a societal perspective and (3) adjusting the nurse training cost for SRT. Pre-specified secondary exploratory analyses were also conducted: (1) using two other utility measures (SF-6D and EQ-5D + Sleep) to calculate QALYs, (2) only including participants who attended at least one SRT session in the intervention arm for the per-protocol analysis, (3) using only NHS and PSS costs and EQ-5D-3L data over the first 6 months of follow-up and (4) using improvement of ISI scores and treatment response measured by ISI change score as health outcomes.
We followed current guidelines59,60 for conducting and reporting economic evaluations within clinical trials, including in relation to the design, conduct, data analysis and reporting.
Methods
Aim
The primary aim of the health economics component of the HABIT study was to address the question of whether nurse-delivered SRT is cost-effective compared with SH among patients with insomnia in primary care.
The within-trial economic analysis was performed using individual patient-level data collected from the HABIT trial. The analysis used data from the HABIT trial only and did not combine this with any external data or evidence. The primary analytical approach took the form of a cost–utility analysis, which uses QALYs as the main measure of health outcome. The economic analysis compared the costs and outcomes of each intervention group over the 12-month period following randomisation, with no extrapolation beyond the study period as pre-specified by the study protocol. The time horizon of the evaluation was 1 year and so no discounting of costs and QALYs was applied.
Measurement of resource use and costs
Resource use and costs during the trial follow-up period were estimated using an adapted version of the self-reported CSRI, which all participants were asked to complete. Data were collected at four time points: baseline, 3 and 6 months after randomisation (with a recall period of ‘in the last 3 months’) and 12 months after randomisation (with a recall period of ‘in the last 6 months’). The CSRI collected individual patient use of NHS and PSS services due to (1) their insomnia and (2) health-related reasons other than their insomnia. We followed current guidelines61 and included insomnia-related NHS and PSS resource use and costs in our analysis because these were deemed to be important and relevant to the intervention and the underlying condition.
The adapted version of the CSRI captured both NHS- and PSS-related resource use and costs, and insomnia-related resource use and costs borne by trial participants. This included: frequency of use of hospital care (including accident and emergency visits, hospital outpatient appointments, overnight hospital admissions), community-based health and social care (including consultations with GPs, consultations with PNs), mental health services (including consultations with psychiatrists, psychologists, mental health nurses, and counsellors) and prescribed insomnia-related medications. These items were identified as relevant and important to reflect the clinical care that patients with insomnia are provided in the NHS based on discussion with clinical experts in the team. For inpatient admissions, the trial participants were also asked to record how many nights of hospital stays they experienced. For contacts with various community-based health and social care professionals, in addition to frequency, the trial participants were also asked to record how many minutes, on average, each contact lasted. The trial participants were additionally asked to record the name and dose of their prescribed insomnia-related medications.
The trial participants also documented their purchases of over-the-counter remedies (name) and frequency of use of complementary therapies (such as homeopathy and acupuncture) at each of the four time points. These were treated as participants’ out-of-pocket spending on their insomnia because they are not normally provided by the NHS, and the associated costs were included in the sensitivity analysis with the societal perspective.
We made some assumptions when cleaning, analysing, or costing the health and social care resource use data. The assumptions included: (1) if a patient answered ‘No’ to a prompt question about resource utilisation, ‘have you used any of the services below for help with your insomnia?’, then we assumed that the frequency of service use for that particular item was equal to zero; (2) if a patient answered ‘Yes’ to a prompt question about resource utilisation, ‘have you used any of the services below for help with your insomnia?’ but did not report the frequency of service use for that particular item, then we assumed the data were missing; (3) where patients were asked to record the names and doses for prescribed insomnia medications and over-the-counter (OTC) remedies, we assumed one monthly pack (28-tablet pack) per 3-month period, so two monthly packs per 6-month period for short-acting hypnotics and sedative antihistamines; and nightly use (so three monthly pack per 3-month period and six monthly pack per 6 months) for other medications and (4) we assumed that patients purchased one item of over-the-counter remedies for their insomnia over the recalled period. The assumptions of prescribed and over-the-counter remedies use were based on clinical expert opinion.
Costing of the sleep restriction therapy intervention and sleep hygiene
The SRT intervention introduced and tested by the HABIT trial includes two main components: (1) SRT training and (2) nurse-delivered SRT sessions. Both arms provided SH advice, which is typically NHS usual care for patients who seek help in community care.
Training
We included the cost of training the community nurses in how to deliver the SRT intervention as it is not part of standard NHS practice. We assumed SH to be implicitly known without any training as it was delivered as standard NHS practice.
Two members of the research team (SK and NS) delivered a total of 17 training sessions (14 by SK and 3 by NS) to a total of 56 community nurses. Each nurse spent approximately 4 hours on their session. SK spent 4 hours to deliver each session and NS spent 3 hours to deliver each session. We assumed that SK and NS spent 5 minutes for the preparation of each training session. SK also spent 8 hours developing training materials. Among the 56 community nurses trained, only 40 delivered SRT sessions to participants. The other 16 trained nurses did not see any patients, or the practices they were based at did not open for the HABIT trial, and therefore their costs were not included in the analysis after discussions with the clinical team.
The total cost of SRT training within the HABIT trial was calculated mainly based on the time the 2 trainers and 40 community nurses spent on training sessions, multiplied by the trainers and community nurse cost per hour, which were obtained from the staff salaries from the project budget (SDK and ANS) and from the Unit Costs of Health and Social Care [Personal Social Services Research Unit (PSSRU)] 2019 compendium. This reflected the NHS cost of training as it was delivered in the HABIT trial and included the time costs for the trainers and the community nurses. The cost of venue hire, trainer’s and trainees’ travel costs to training sites and NHS parking charges were not included in the analysis. We applied and averaged the total cost of SRT training for the SRT intervention group across all participants randomised to obtain the mean training cost per patient, regardless of how many SRT sessions they attended.
In a sensitivity analysis we adjusted the per-patient training cost to reflect how many patients a nurse in the NHS may see over a 1-year period, if SRT were introduced into primary care. That is, in practice, the trained nurses would see many more patients than those involved in the trial and therefore the cost of training would be averaged across a larger number of patients. We assumed that a PN would hold a weekly sleep clinic lasting for 3 hours (i.e. 12 hours per month). We calculated how many patients each nurse would be able to see each month using the mean time to deliver four SRT sessions in the trial, and calculated how many patients each nurse would be able to see for 1 year. We divided the total SRT training cost in the trial by the number of patients the 40 nurses would be able to see.
Delivery of sleep restriction therapy sessions and sleep hygiene
Patients randomised to the SRT intervention arm were provided a total of four SRT sessions with the community nurse. We recorded whether individual patients attended each of the four SRT sessions, whether the SRT sessions were delivered in person or via telephone, and associated start and end time of each SRT session attended. The duration of each SRT session delivered was calculated and used to indicate associated community nurse time spent on the delivery of the session. If the participant did not attend the planned SRT session or withdrew from the study, the time duration of that session was assumed to be 0 with 0 costs incurred.
The total and mean nurse time spent on delivering each of the four SRT sessions was calculated. The total and mean cost of community nurse time spent on delivering each of the four SRT sessions was calculated by multiplying the cost per hour [£41.5 per hour, or £0.7 per minute, Unit Costs of Health and Social Care (PSSRU) 2019] by the duration, regardless of whether they were in person or via telephone. The total time and associated cost of the 40 community nurses’ time spent on delivering the four SRT sessions to all the patients in the SRT arm were calculated as the sum of each session delivered. We calculated the mean intervention cost per participant by dividing the total SRT costs by the total number of the patients in the SRT arm regardless of whether they attended no session or only some of the four sessions. The SH leaflet was contained within the SRT patient guide and given at the end of the first SRT session, so no extra nurse time was calculated or included for patients in the SRT intervention arm.
In the SH arm, we consulted the clinical team and estimated staff time spent on e-mailing or posting the SH booklet, and postage cost for posting paper copies of the SH booklet. We estimated the mean cost of the SH per patient by multiplying average staff time on SH with their cost per minute.
Measuring and valuing productivity loss due to insomnia
Time off work and productivity loss due to insomnia for those in employment were captured and quantified using the WPAI44 questionnaire at baseline and 3, 6 and 12 months. Questions 1, 2, 4 and 5 of the WPAI were used to quantify productivity loss due to insomnia. Patients were asked whether they were currently employed (Q1). If they were employed, the participants were asked about hours of missed work due to insomnia (Q2) and hours actually worked (Q4). The degree to which insomnia affected productivity while working was recorded on a scale from 0 to 10, with a higher score indicating worse impairment (Q5).
Participants’ productivity losses had two components: absenteeism (work time missed) and presenteeism for employment (impairment at work). Participants’ time absence from work due to insomnia was taken directly from question 2 of the WPAI for those who were working at the time of data collection (captured by Q1). Presenteeism was calculated as the total working time (Q4) multiplied by the extent (converted from Q5) to which insomnia affected productivity while working. The 0–10 scale in question 5 was converted into a percentage score from 0% to 100%.9 The total productivity loss was calculated as the sum of time lost due to both absenteeism and presenteeism.
Work productivity and activity impairment questions asked about patients’ productivity loss over the past 7 days. In order to keep a consistent recall time period with other cost categories, the productivity loss over the previous 3-month period (i.e. baseline and the time points of 3 and 6 months after randomisation) and 6-month period (at the time point of 12 months after randomisation) were extrapolated by multiplying values by 12 for 3-month periods and by 24 for the 6-month period, assuming each month has 4 weeks.
Economic values associated with productivity losses were estimated by multiplying total working hours lost by average hourly salaries based on gender and age groups obtained from Office of National Statistics 2019. Productivity losses were valued for participants who were in employment. For the participants who were not in employment, we assumed 0 productivity loss. We then obtained average values of productivity losses due to insomnia across all participants randomised into the SRT and SH arms, regardless of whether they were in employment or not. Economic values associated with productivity losses are regarded as falling outside of the perspective of NHS and PSS, so these values were only included in the sensitivity analysis that adopted a societal perspective.
Valuation of resource use
The unit costs for clinical staff time to develop training materials, deliver and receive the SRT training, and deliver the SRT sessions were obtained from national standard sources [Unit Costs of Health and Social Care (PSSRU)62].
Unit costs of community health and social service inputs were based on PSSRU national cost compendia. The costs of medications were estimated from the British National Formulary.63 NHS references costs were assigned to use of alternative categories of hospital services.61 Complementary services were assigned an average cost according to clinical opinion. The costs of OTC remedies were obtained from the PAGB OTC directory (www.pagb.co.uk/product/pagbs-otc-directory/) and from a search of the websites of large pharmacies in the UK, including Boots and Lloyds pharmacies.
The cost of each resource item was calculated by multiplying the number of resource units used by the relevant unit cost. The total cost for each individual trial participant was estimated as the sum of the costs of resource use items consumed during the specific time period. For example, total cost at 12-month follow-up was the sum of total costs at 3-, 6- and 12-month follow-ups, as the data were collected for the previous 3 months at the 3- and 6-month follow-up time points, and for the previous 6 at the 12-month follow-up point.
All costs were reported in 2018–9 Great British pounds. Given that the trial follow-up period was 12 months, no discounting was applied to cost estimates.
Calculation of utilities and quality-adjusted life-years
Trial participants completed the EQ-5D-3L questionnaire at baseline and 3, 6 and 12 months post randomisation. The EQ-5D-3L questionnaire facilitates the generation of a utility score from the measure’s health status classification system. A utility score reflects the preference of the general population for any particular set of health states. The EQ-5D-3L has been recommended by NICE64 for the measurement and valuation of health outcomes in economic evaluations. Effectiveness was estimated in terms of QALYs, calculated as the baseline-adjusted utility curve of EQ-5D-3L utility scores across the baseline and 3-, 6-, and 12-month intervals, using the trapezoidal rule.
We understand that follow-ups may not have fallen exactly at the expected time points (e.g. 3, 6 and 12 months post randomisation). However, we made the assumption that the time points were exact to simplify the calculation of QALYs.
In addition to the EQ-5D-3L, participants completed two additional utility measures, the SF-6D (derived from the SF-36) and the EQ-5D-3L + sleep, at baseline, 3, 6 and 12 months. EQ-5D-3L + Sleep contains the same five dimensions of the original EQ-5D-3L questionnaire plus an extra dimension on sleep. A value set has been developed for EQ-5D-3L + Sleep. Utility values derived from the SF-6D and EQ-5D-3L + Sleep were used to estimate QALYs over the 12-month trial period using the same method described for the EQ-5D-3L above, and the QALYs derived were used for further secondary exploratory analysis. Given that the trial follow-up period was 12 months, no discounting was applied to QALY estimates.
Missing data
Many sources of information on patient characteristics, treatments, utilities and resource use are used to conduct the economic evaluation within a clinical trial. Therefore, missing data are a frequent and particularly challenging issue that requires careful consideration. Costs and outcomes for individuals with missing data may differ systematically from those individuals with observed data. We followed current method guidance65,66 on handling missing data in cost-effectiveness analysis conducted alongside clinical trials. We examined missing data status at the trial time points, and estimated logistic regressions to investigate association between missingness of NHS and PSS costs and QALYs with key baseline covariates including age, sex, region, EQ-5D utility score, PHQ-9 score, ISI score, and use of prescribed sleep medication, and treatment group.
Consequently, we decided to impute missing data for use in our base-case analysis. Both chained equations and predictive mean matching (PMM with knn = 8) were used for multiple imputation using the Stata command ‘mi impute chained’. The imputed variables included EQ-5D-3L, EQ-5D-3L + Sleep and SF-6D utility values, and ISI scores at 3-, 6- and 12-month follow-ups, and NHS and PSS costs at 3-, 6- and 12-month follow-ups, and non-NHS and PSS costs at 3-, 6- and 12-month follow-ups, and the value of productivity loss at 12 months. We used the same set of baseline covariates as predictor variables for multiple imputation and regression models to estimate incremental costs and incremental QALYs, which included age, sex, site, baseline ISI scores, baseline EQ-5D-3L utility scores, baseline PHQ-9 scores, prescribed sleep medication at baseline, and NHS and PSS costs at baseline. The imputation was conducted for cost and utilities for the two treatment arms separately within a single command. The chained equation method means that the costs and EQ-5D-3L utility scores at each time point contributed to the multiple imputation as both predictors and imputed variables, which made efficient use of the data.
We used multiple imputation to generate 50 data sets using PMM, which provides plausible values when costs and utility values are not normally distributed. The number of imputations was run following the rule of thumb. The imputation models were validated by comparing the distributions of the imputed data with the observed data.
Cost-effectiveness analysis
The base-case analysis was conducted using the full data set with missing data imputed as described above, comparing the two arms as randomised and including all patients in the analysis where practical. The purpose of the economic analysis was not to test statistical hypotheses such as whether there are significant differences between costs or/and health outcome such as QALYs. The fundamental aim of the economic analysis was to estimate the ICER associated with SRT, to quantify the uncertainty surrounding the ICER estimate, and to examine whether and to what extent the intervention is cost-effective by comparing the ICER with conventional cost-effectiveness thresholds for an extra unit of health outcome (i.e. QALY).
Reporting the cost and health outcomes
We report means and SDs (or standard errors) and medians and IQRs for EQ-5D-3L utilities and associated QALYs for the two arms at the different follow-up time points based on individual patient data. We similarly report means and SDs (or standard errors) and medians and IQRs for EQ-5D-3L + Sleep and SF-6D utilities and associated QALYs. We report intervention costs, including training and delivery of SRT sessions to participants in the SRT arm, and SH to participants in the SH arm. We report mean costs of key NHS and PSS services in relation to insomnia, including hospital services, community health and social care, and prescribed medications, as well as non-NHS out-of-pocket healthcare costs, costs due to productivity losses, and total societal costs for both arms of the trial at the different follow-up time points. We performed parametric t-tests (bootstrapped 95% CIs, 1000 samples) to compare mean costs of cost categories in relation to insomnia, and QALYs based on the EQ-5D-3L by treatment group at each assessment time point.
Regression analysis and bootstrapping
In the base case, bivariate regression using seemingly unrelated regression was used to estimate incremental NHS and PSS costs and incremental QALYs between the SRT and SH arms over the 12-month follow-up period on each of the imputed samples controlling for baseline covariates [ISI score, region, age, prescribed sleep medication use, sex, and PHQ-9 score, and either baseline EQ-5D-3L utility scores (for incremental QALYs) or baseline NHS and PSS costs (for incremental costs)]. The mean estimate of the ICER was calculated by dividing incremental costs by incremental QALYs.
Non-parametric bootstrapping was used to quantify uncertainty surrounding the mean ICER estimate by resampling 1000 times from incremental costs and incremental QALYs obtained from the seemingly unrelated regression. This method addressed the effects of missing data and sampling uncertainty using the MI Boot approach suggested by Schomaker and Heumann.67 This approach is simpler to implement and less demanding of computing capacity, and it has been shown to produce valid inference and to be equivalent to nesting bootstraps within imputations and combining results using Rubin’s rule. The outputs were displayed graphically on a cost-effectiveness plane to determine the uncertainty surrounding cost-effectiveness, enabling investigation of the joint distribution of both incremental costs and incremental QALYs by scatter-plotting the incremental cost-QALY pairs in the plane and exploring the joint density of the plots. NMBs were estimated from the incremental costs and incremental QALYs at alternative cost-effectiveness thresholds of £15,000, £20,000 and £30,000 per QALY gained in order to reflect the overall resource gain or loss associated with SRT. By calculating NMBs for each of these 1000 simulated ICER values at alternative levels of the cost-effectiveness threshold, the probability of cost-effectiveness of SRT (defined as the proportion of positive NMBs at a given threshold level) was calculated, and plotted as a cost-effectiveness acceptability curve (CEAC).
Sensitivity analysis and exploratory analysis
Sensitivity analyses were performed to explore how the ICER was affected by altering several key features of the economic evaluation. The sensitivity analyses were conducted with the intention of providing evidence on whether the results from the base-case analysis remained robust. Our sensitivity analyses included: (1) using complete-case analysis rather than imputed data to explore any potential effects due to data imputation; (2) adopting a societal perspective where extra costs beyond NHS and PSS costs were included in the analysis (these included non-NHS out-of-pocket spending on complementary therapies, OTC remedies, and the value of productivity losses due to insomnia); and (3) adjusting costs associated with SRT training from the overall cost of SRT intervention. We also conducted several pre-defined secondary exploratory analyses, including (1) using two other utility measures (SF-6D and EQ-5D + Sleep) to calculate QALYs. We compared cost-effectiveness results for these measures with those obtained from the EQ-5D-3L. (2) Including participants in the SRT arm who attended at least one SRT session for a per-protocol analysis. (3) Restricting the analysis period to the first 6 months post randomisation using NHS and PSS costs and using QALYs estimated from utilities obtained from the EQ-5D-3L up to the 6-month follow-up. The rationale was to explore short-term cost-effectiveness of the SRT given the primary outcome of the HABIT trial was insomnia severity at 6 months. (4) Performing a cost-effectiveness analysis using improvement of ISI scores between the baseline and 12-month follow-up as the health outcome, and expressed in terms of incremental cost per unit reduction in ISI score. In this analysis, hypothetical cost-effectiveness thresholds were used to estimate the probability of cost-effectiveness and net economic benefit of the SRT intervention. Other hypothetical cost-effectiveness thresholds were also used for a further exploration. (5) Defining treatment responders as those exhibiting a reduction of ISI score ≥ 8 points between baseline and 12 months, we estimated the incremental cost per additional treatment responder.
Results
Descriptive analysis and quality-of-life measures
Six hundred and forty-two participants were randomised in the HABIT trial, half of them (321) to the SRT arm and the other half (321) to the SH arm. Table 24 presents mean (SD) and median IQR values for the key health outcomes between the two arms at different time points using the available cases without imputation.
| SRT | SH | |||||
|---|---|---|---|---|---|---|
| N | Mean (SD) | Median(IQR) | N | Mean (SD) | Median(IQR) | |
| EQ-5D-3L utility | ||||||
| Baseline | 321 | 0.70 (0.26) | 0.73 (0.66–0.85) | 321 | 0.72 (0.24) | 0.76 (0.69–0.85) |
| 3-month | 245 | 0.72 (0.29) | 0.80 (0.69–1) | 284 | 0.68 (0.28) | 0.73 (0.62–0.85) |
| 6-month | 233 | 0.72 (0.27) | 0.80 (0.69–0.85) | 281 | 0.72 (0.25) | 0.76 (0.69–0.85) |
| 12-month | 223 | 0.72 (0.27) | 0.80 (0.69–0.85) | 266 | 0.72 (0.23) | 0.75 (0.69–0.85) |
| QALYs (by EQ-5D-3L) | ||||||
| 3-month | 245 | 0.18 (0.06) | 0.19 (0.17–0.22) | 284 | 0.18 (0.06) | 0.19 (0.16–0.22) |
| 6-month | 218 | 0.36 (0.13) | 0.39 (0.34–0.45) | 267 | 0.36 (0.12) | 0.38 (0.32–0.43) |
| 12-month | 202 | 0.73 (0.24) | 0.79 (0.66–0.88) | 249 | 0.72 (0.21) | 0.77 (0.66–0.85) |
| SF-6D | ||||||
| Baseline | 321 | 0.63 (0.11) | 0.64 (0.57–0.70) | 321 | 0.63 (0.10) | 0.62 (0.56–0.69) |
| 3-month | 243 | 0.68 (0.13) | 0.66 (0.61–0.79) | 283 | 0.63 (0.11) | 0.64 (0.56–0.70) |
| 6-month | 230 | 0.67 (0.13) | 0.67 (0.58–0.76) | 282 | 0.65 (0.11) | 0.64 (0.58–0.7) |
| 12-month | 222 | 0.68 (0.13) | 0.67 (0.6–0.76) | 262 | 0.65 (0.10) | 0.64 (0.59–0.71) |
| QALYs (by SF-6D) | ||||||
| 3-month | 243 | 0.17 (0.03) | 0.16 (0.15–0.18) | 283 | 0.16 (0.02) | 0.16 (0.14–0.17) |
| 6-month | 215 | 0.34 (0.06) | 0.33 (0.3–0.38) | 267 | 0.32 (0.05) | 0.32 (0.29–0.35) |
| 12-month | 200 | 0.68 (0.11) | 0.67 (0.60–0.76) | 245 | 0.65 (0.09) | 0.64 (0.59–0.7) |
| EQ-5D-3L + Sleep | ||||||
| Baseline | 321 | 0.76 (0.15) | 0.79 (0.71–0.87) | 321 | 0.77 (0.14) | 0.79 (0.72–0.87) |
| 3-month | 245 | 0.79 (0.16) | 0.86 (0.73–0.9) | 284 | 0.76 (0.16) | 0.79 (0.7–0.87) |
| 6-month | 233 | 0.79 (0.16) | 0.86 (0.73–0.9) | 282 | 0.78 (0.14) | 0.80 (0.72–0.87) |
| 12-month | 221 | 0.79 (0.16) | 0.86 (0.73–0.9) | 266 | 0.78 (0.13) | 0.82 (0.72–0.87) |
| QALYs (by EQ-5D-3L + Sleep) | ||||||
| 3-month | 245 | 0.19 (0.03) | 0.21 (0.18–0.22) | 284 | 0.19 (0.03) | 0.2 (0.17–0.22) |
| 6-month | 218 | 0.39 (0.07) | 0.42 (0.38–0.44) | 268 | 0.39 (0.07) | 0.4 (0.35–0.43) |
| 12-month | 201 | 0.79 (0.14) | 0.85 (0.76–0.88) | 249 | 0.78 (0.12) | 0.81 (0.72–0.87) |
In general, more data were missing at follow-up for the SRT intervention group than for the SH group. The mean and median differences of EQ-5D-3L utilities and QALYs between the two arms at different time points were very small. The difference in mean QALYs derived from the EQ-5D-3L over 12 months post randomisation was 0.01. We also present utilities at baseline to assess for any imbalance of health states between the two arms. On average, the SH group had a slightly better HRQoL at baseline.
The EQ-5D-3L + Sleep utilities and QALYs were higher than those for the EQ-5D-3L but the mean and median differences were similarly small. The difference in mean QALYs derived from the EQ-5D-3L + Sleep over 12 months post randomisation was also 0.01.
The SF-6D utilities and QALYs were the lowest among the HRQoL measures. Interestingly, the mean and median differences for SF-6D utilities at different time points and QALYs between the two arms seem to be slightly larger than those for the EQ-5D-3L and EQ-5D-3L + Sleep, although they were also very small. The difference in mean QALYs derived from SF-6D over 12 months post randomisation was 0.03.
Missing data analysis
Table 25 summarises the proportion of individuals with missing health economic data by treatment group over time. There were very few missing data at baseline, and more data were missing at subsequent time points. We explored the patterns of missing data, which indicated that patients with missing data included those lost to follow-up or who withdrew from the trial, as well as those who had missing data at one time point but not at the next. We estimated logistic regressions to investigate the association between missingness of NHS and PSS costs and QALYs with key baseline covariates, including age, sex, region, EQ-5D utility score, PHQ-9 score, prescribed sleep medication use and ISI score. A significant association was found between missingness of NHS and PSS costs and QALYs over 12 months and baseline ISI scores (p < 0.05), with worse ISI scores (greater insomnia severity) being associated with missingness. Region and intervention group were also significantly associated with missingness of NHS costs and QALYs over 12 months (p < 0.05).
| Missing values | ||||||
|---|---|---|---|---|---|---|
| SRT (N = 321) | SH (N = 321) | Total (N = 642) | ||||
| N | % | N | % | N | % | |
| EQ-5D index at baseline | 0 | 0 | 0 | 0 | 0 | 0 |
| EQ-5D index at 3 months | 76 | 24 | 37 | 12 | 113 | 18 |
| EQ-5D index at 6 months | 88 | 27 | 40 | 12 | 128 | 20 |
| EQ-5D index at 12 months | 98 | 31 | 55 | 17 | 153 | 24 |
| QALYs at 12 months generated from EQ-5D utility scores | 119 | 37 | 72 | 22 | 191 | 30 |
| Total NHS cost at baseline (over previous 3 months) | 6 | 2 | 2 | 1 | 8 | 1 |
| Total NHS cost at 3-month follow-up | 80 | 25 | 49 | 15 | 129 | 20 |
| Total NHS cost at 6-month follow-up | 91 | 28 | 48 | 15 | 139 | 22 |
| Total NHS cost at 12-month follow-up | 103 | 32 | 65 | 20 | 168 | 26 |
| Total NHS cost over 12-month trial period | 129 | 40 | 92 | 29 | 221 | 34 |
| Hospital cost at baseline (over previous 3 months) | 0 | 0 | 0 | 0 | 0 | 0 |
| Hospital cost at 3-month follow-up | 73 | 23 | 34 | 11 | 107 | 17 |
| Hospital cost at 6-month follow-up | 69 | 21 | 28 | 9 | 97 | 15 |
| Hospital cost at 12-month follow-up | 92 | 29 | 51 | 16 | 143 | 22 |
| Total hospital cost over 12-month trial period | 111 | 35 | 66 | 21 | 177 | 28 |
| Primary care cost at baseline (over previous 3 months) | 0 | 0 | 0 | 0 | 0 | 0 |
| Primary care cost at 3-month follow-up | 75 | 23 | 39 | 12 | 114 | 18 |
| Primary care cost at 6-month follow-up | 81 | 25 | 39 | 12 | 120 | 19 |
| Primary care cost at 12-month follow-up | 96 | 30 | 55 | 17 | 151 | 24 |
| Total primary care cost over 12-month trial period | 119 | 37 | 77 | 24 | 196 | 31 |
| Mental Health service cost at baseline (over previous 3 months) | 0 | 0 | 0 | 0 | 0 | 0 |
| Mental health service cost at 3-month follow-up | 75 | 23 | 39 | 12 | 114 | 18 |
| Mental health service cost at 6-month follow-up | 83 | 26 | 38 | 12 | 121 | 19 |
| Mental health service cost at 12-month follow-up | 95 | 30 | 57 | 18 | 152 | 24 |
| Total mental health service cost over 12-month trial period | 117 | 36 | 77 | 24 | 194 | 30 |
| Prescribed insomnia medications cost at baseline (over previous 3 months) | 6 | 2 | 2 | 1 | 8 | 1 |
| Prescribed insomnia medications cost at 3-month follow-up | 73 | 23 | 42 | 13 | 115 | 18 |
| Prescribed insomnia medications cost at 6-month follow-up | 86 | 27 | 41 | 13 | 127 | 20 |
| Prescribed insomnia medications cost at 12-month follow-up | 98 | 31 | 55 | 17 | 153 | 24 |
| Total prescribed insomnia medication cost over 12-month trial period | 119 | 37 | 80 | 25 | 199 | 31 |
Intervention-related training costs
Table 26 summarises the time that the 2 trainers (SK and NS) and 40 community nurses spent on preparing, delivering and receiving the SRT training. A total cost of £10,182 and an average of £31.7 per participant was estimated and included the training cost of SRT. The training costs would be likely to reduce after scaling up.
| Items | Number | Number of hours | Time (hours) | Unit cost (£ per hour) | Total cost (£) |
|---|---|---|---|---|---|
| SRT training | |||||
| Nurse SRT training | 40 | 4 | 160 | 41.5 | 6,640 |
| Trainer 1 delivery | 14 | 4 | 56 | 44.16 | 2473 |
| Trainer 1 preparation | 14 | 0.08 | 1.2 | 44.16 | 52 |
| Trainer 2 delivery | 3 | 3 | 9 | 71.78 | 646 |
| Trainer 2 preparation | 3 | 0.08 | 0.25 | 71.78 | 18 |
| Others | |||||
| Time for trainer 1 to generate SRT training materials | 1 | 8 | 8 | 44.16 | 353 |
| Total training cost | – | 226.45 | – | 10,182 | |
An adjusted average of £2.52 per participant for the training cost of SRT was estimated and used in sensitivity analysis, reflecting that the trained nurse would likely see more patients than the 321 patients in the SRT arm in the trial. This assumed a PN will hold a weekly sleep clinic that lasted for 3 hours each week (12 hours per month). The mean time to complete four treatment sessions was 85.5 minutes in the trial, and so assuming 8.4 patients a month, a PN would see 101 (12 × 8.4) patients a year in routine practice. There were 40 nurses in the trial so 4040 patients would be seen, generating an average SRT training cost of £2.52.
Nurse-led sleep restriction therapy sessions, average duration and associated costs
Details of the nurse-led SRT sessions, including number attending, duration and associated costs, are presented in Table 27. The unit cost of community nurse time was identified as £0.7 per minute (£41.5 per hour).
| Session 1 | Session 2 | Session 3 | Session 4 | |
|---|---|---|---|---|
| Attendance | 296 | 250 | 217 | 209 |
| Mean (SD) duration (minutes) | 39.63 (12.70) | 15.07 (7.14) | 17.82 (9.36) | 13.76 (6.57) |
| Range duration (minutes) | 8–98 | 1–55 | 5–75 | 4–49 |
| Mean cost (£) | 27.74 (8.89) | 8.87 (6.03) | 9.14 (7.87) | 6.80 (5.85) |
The first session lasted the longest and hence cost the most. The average cost of the nurse-led SRT sessions was £52.6 for participants in the intervention group. Adding training cost and delivery cost together, the average SRT cost was £84.3 (31.7 + 52.6).
For the SH group, the cost of sending out the leaflet was estimated as £1.7 per participant.
Insomnia-related healthcare utilisation and associated costs
Table 28 shows the number of participants (n) who had data for various insomnia-related healthcare contacts at 3, 6 and 12 months post randomisation by group. It also summarises mean (SD) frequencies of those services at the three time points between the two groups. Very few participants used hospital-based services for their insomnia. They tended to go to their GP, request repeat prescriptions for medications, and purchase OTC remedies for their insomnia. It is worth noting that whether the health resource utilisation was associated with insomnia relied on participant’s judgement and attribution. A few participants also used alcohol and internet apps for their insomnia, although we do not have sufficiently detailed data to convert these into costs.
| SRT (N = 321) | SH (N = 321) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | Mean (SD) 3 months | n | Mean (SD) 6 months | n | Mean (SD) 12 months | n | Mean (SD) 3 months | n | Mean (SD) 6 months | n | Mean (SD) 12 months | |
| NHS services | ||||||||||||
| Hospital service | ||||||||||||
| Accident and Emergency visits | 251 | 0.01 (0.11) | 255 | 0.008 (0.09) | 230 | 0 | 287 | 0.02 (0.13) | 293 | 0.007 (0.08) | 273 | 0.01 (0.15) |
| Hospital admission | 251 | 0.02 (1.26) | 255 | 0.008 (0.13) | 230 | 0.004 (0.07) | 287 | 0.01 (0.13) | 293 | 0.003 (0.06) | 274 | 0.05 (0.79) |
| Outpatient | 250 | 0.21 (1.04) | 252 | 0.10 (0.65) | 229 | 0.14 (0.51) | 287 | 0.19 (0.83) | 293 | 0.08 (0.37) | 272 | 0.36 (2.37) |
| Primary care service | ||||||||||||
| GP | 249 | 0.25 (0.89) | 245 | 0.20 (0.62) | 228 | 0.35 (1.06) | 284 | 0.38 (0.90) | 288 | 0.20 (0.60) | 273 | 0.37 (0.96) |
| PN | 247 | 0.38 (1.01) | 247 | 0.04 (0.26) | 228 | 0.10 (0.43) | 287 | 0.11 (0.42) | 291 | 0.02 (0.17) | 273 | 0.10 (0.39) |
| Repeat prescription | 248 | 0.62 (1.71) | 241 | 0.69 (1.36) | 225 | 1.41 (3.18) | 283 | 0.78 (1.35) | 282 | 0.70 (1.54) | 266 | 1.42 (2.66) |
| Mental health service | ||||||||||||
| Psychiatrist | 248 | 0.008 (0.09) | 240 | 0.008 (0.09) | 227 | 0.04 (0.30) | 284 | 0.007 (0.08) | 284 | 0.007 (0.08) | 265 | 0.02 (0.21) |
| Psychologist | 247 | 0.008 (0.13) | 240 | 0 | 227 | 0.02 (0.21) | 284 | 0.10 (0.83) | 284 | 0.05 (0.63) | 264 | 0.01 (0.11) |
| Mental health nurse | 248 | 0.008 (0.13) | 240 | 0.04 (0.43) | 227 | 0.12 (1.37) | 284 | 0.08 (1.19) | 284 | 0.02 (0.24) | 265 | 0.003 (0.06) |
| Counsellor | 247 | 0.07 (0.54) | 239 | 0.08 (0.63) | 227 | 0.16 (1.11) | 284 | 0.11 (0.74) | 283 | 0.14 (1.05) | 264 | 0.06 (0.75) |
| Other mental health professional | 247 | 0.02 (0.18) | 240 | 0.004 (0.06) | 227 | 0.11 (1.35) | 282 | 0.01 (0.19) | 284 | 0.02 (0.26) | 265 | 0.03 (0.23) |
| % participants prescribed insomnia medication | 249 | 50 (20.1) | 236 | 37 (15.6) | 226 | 55 (24.3) | 285 | 66 (23.2) | 282 | 70 (24.8) | 267 | 67 (25.1) |
| Non-NHS services | ||||||||||||
| Homeopathy | 246 | 0.05 (0.47) | 237 | 0.03 (0.24) | 227 | 0.01 (0.15) | 285 | 0.05 (0.39) | 280 | 0.03 (0.26) | 265 | 0.07 (0.67) |
| Other complementary therapies | 247 | 0.13 (0.52) | 237 | 0.10 (0.54) | 227 | 0.10 (0.59) | 283 | 0.13 (0.98) | 278 | 0.13 (0.80) | 265 | 0.14 (0.80) |
| Alcohol use (%) | ||||||||||||
| Not at all | 249 | 201 (80.7) | 235 | 184 (78.3) | 226 | 182 (80.5) | 285 | 211 (74.0) | 281 | 205 (73.0) | 266 | 198 (74.4) |
| Less than once a week | 249 | 17 (6.8) | 235 | 21 (8.9) | 226 | 24 (10.6) | 285 | 37 (13.0) | 281 | 35 (12.5) | 266 | 34 (12.8) |
| Once or twice a week | 249 | 17 (6.8) | 235 | 20 (8.5) | 226 | 9 (4.0) | 285 | 18 (6.3) | 281 | 26 (9.3) | 266 | 25 (9.4) |
| Three or more times a week | 249 | 14 (5.6) | 235 | 10 (4.3) | 226 | 11 (4.9) | 285 | 19 (6.7) | 281 | 15 (5.3) | 266 | 9 (3.4) |
| Proportion who used internet/apps (%) | ||||||||||||
| 249 | 22 (8.8) | 235 | 22 (9.4) | 226 | 27 (12.0) | 285 | 29 (10.2) | 281 | 31 (11.0) | 266 | 30 (11.3) | |
Table 29 summarises unit cost estimates for calculation of NHS and PSS services and broader categories of service use and costs, obtained from various national sources.
| Unit cost (£) | Source (2018–9) | |
|---|---|---|
| Accident and emergency (per visit) | 166 | PSSRU |
| Hospital admission (per visit) | 1311.2 | Reference costa |
| Hospital extra days beyond trim point (per day) | 276.6 | Reference costb |
| Outpatient/day case (per consultation) | 224.8 | Reference cost |
| GP (per consultation) | 33 | PSSRU |
| PN (per hour) | 41.5 | PSSRU |
| Repeat prescription (per service) | 6 | PSSRU |
| Psychiatrist (per hour) | 111 | PSSRU |
| Psychologist (per hour) | 56.3 | PSSRU |
| Mental health nurse (per hour) | 37 | PSSRU |
| Counsellor (per hour) | 44.3 | PSSRU |
| Other mental health professional (per hour) | 34 | PSSRU |
| Homeopathy (per visit) | 40 | Expert opinion |
| Acupuncture (per visit) | 40 | Expert opinion |
| Other complementary therapies (per visit) | 40 | Expert opinion |
| Average hourly salary | National Office of Statistics (2019) | |
| Age (18–21) and male | 8.6 | National Office of Statistics (2019) |
| Age (18–21) and female | 8.5 | National Office of Statistics (2019) |
| Age (22–29) and male | 12.3 | National Office of Statistics (2019) |
| Age (22–29) and female | 11.43 | National Office of Statistics (2019) |
| Age (30–39) and male | 15.7 | National Office of Statistics (2019) |
| Age (30–39) and female | 13.7 | National Office of Statistics (2019) |
| Age (40–49) and male | 17.5 | National Office of Statistics (2019) |
| Age (40–49) and female | 13.5 | National Office of Statistics (2019) |
| Age (50–59) and male | 16.4 | National Office of Statistics (2019) |
| Age (50–59) and female | 12.2 | National Office of Statistics (2019) |
| Age (60 and over) and male | 13.6 | National Office of Statistics (2019) |
| Age (60 and over) and female | 10.8 | National Office of Statistics (2019) |
Average costs for various cost categories, EQ-5D-3L utility values and associated QALYs by study time points are presented in Table 30 and compared across the study arms.
| Treatment group, cost (£) | Mean difference | Bootstrap 95% CI | p-value | ||||||
|---|---|---|---|---|---|---|---|---|---|
| SRT | SH | ||||||||
| N | Mean | SE | N | Mean | SE | ||||
| Cost categories by period | |||||||||
| Baseline to 3 months | |||||||||
| Primary care services | 246 | 17.80 | 2.76 | 282 | 17.88 | 2.07 | −0.08 | −7.04 to 6.89 | 0.982 |
| Hospital services | 248 | 48.22 | 15.96 | 287 | 50.54 | 12.52 | −2.32 | −41.96 to 37.32 | 0.909 |
| Mental health services | 246 | 7.30 | 3.85 | 282 | 8.09 | 2.67 | −0.79 | −10.04 to 8.46 | 0.866 |
| Prescribed insomnia medications | 248 | 0.85 | 0.19 | 279 | 1.25 | 0.37 | −0.40 | −1.23 to 0.43 | 0.336 |
| NHS and PSS | 241 | 69.53 | 16.37 | 272 | 76.93 | 14.46 | −7.40 | −50.40 to 35.61 | 0.735 |
| Non-NHS and PSS | 246 | 9.02 | 2.55 | 279 | 9.11 | 1.87 | −0.09 | −6.44 to 6.26 | 0.978 |
| Productivity losses | 237 | 619.74 | 76.74 | 268 | 823.49 | 83.77 | −203.75 | −434.17 to 26.68 | 0.073 |
| 3–6 months | |||||||||
| Primary care service | 240 | 10.93 | 1.55 | 282 | 11.07 | 1.47 | −0.14 | −4.16 to 3.88 | 0.948 |
| Hospital services | 252 | 27.98 | 9.89 | 293 | 21.04 | 5.11 | 6.94 | −14.52 to 28.41 | 0.533 |
| Mental health services | 238 | 4.03 | 1.74 | 283 | 20.14 | 12.57 | −16.11 | −41.30 to 9.09 | 0.205 |
| Prescribed insomnia medications | 235 | 0.99 | 0.41 | 280 | 1.02 | 0.23 | −0.03 | −0.95 to 0.90 | 0.956 |
| NHS and PSS | 230 | 46.55 | 11.22 | 273 | 41.05 | 6.47 | 5.50 | −19.90 to 30.91 | 0.671 |
| Non-NHS and PSS | 233 | 6.47 | 1.62 | 276 | 7.76 | 2.04 | −1.29 | −6.37 to 3.78 | 0.620 |
| Productivity losses | 217 | 553.64 | 83.81 | 265 | 639.51 | 70.06 | −85.87 | −292.26 to 120.53 | 0.432 |
| 6–12 months | |||||||||
| Primary care services | 225 | 20.74 | 3.25 | 266 | 20.80 | 2.56 | −0.06 | −7.93 to 7.81 | 0.988 |
| Hospital services | 229 | 33.32 | 9.89 | 270 | 56.48 | 5.12 | −23.16 | −61.63 to 15.30 | 0.230 |
| Mental health services | 226 | 13.55 | 5.32 | 264 | 7.52 | 3.76 | 6.03 | −6.74 to 18.80 | 0.355 |
| Prescribed insomnia medications | 223 | 3.21 | 0.94 | 266 | 1.73 | 0.36 | 1.48 | −0.45 to 3.41 | 0.136 |
| NHS and PSS | 218 | 69.69 | 11.45 | 256 | 88.95 | 19.70 | −19.26 | −64.18 to 25.66 | 0.398 |
| Non-NHS and PSS | 205 | 5.44 | 1.74 | 243 | 8.94 | 2.63 | −3.50 | −9.62 to 2.62 | 0.268 |
| Productivity losses | 212 | 970.81 | 136.01 | 250 | 1372.84 | 155.00 | −402.03 | −800.02 to −4.04 | 0.052 |
| Total cost over 12-month period (complete case for entire period) | |||||||||
| Total NHS and PSS cost | 192 | 182.13 | 26.56 | 229 | 189.16 | 28.90 | −7.04 | −82.29 to 68.22 | 0.858 |
| Total NHS and PSS cost and intervention cost | 186 | 268.26 | 27.17 | 229 | 190.86 | 28.90 | 77.39 | 2.91 to 151.87 | 0.047 |
| Total NHS and PSS cost and intervention cost with reduced SRT training cost to £2.52 | 186 | 236.56 | 27.17 | 229 | 190.86 | 28.90 | 45.69 | −30.53 to 121.92 | 0.250 |
| Total societal costa | 142 | 2176.35 | 298.83 | 180 | 2676.03 | 306.89 | −501.38 | −1312.76 to 313.40 | 0.244 |
| EQ-5D at follow-up points | |||||||||
| EQ-5D-3L value at baseline | 321 | 0.704 | 0.015 | 321 | 0.723 | 0.014 | − | − | - |
| EQ-5D-3L value at 3-month follow-up | 245 | 0.724 | 0.018 | 284 | 0.684 | 0.016 | 0.040 | −0.009 to 0.089 | 0.105 |
| EQ-5D-3L value at 6-month follow-up | 233 | 0.718 | 0.018 | 281 | 0.722 | 0.015 | −0.004 | −0.049 to 0.041 | 0.857 |
| EQ-5D-3L value at 12-month follow-up | 223 | 0.722 | 0.018 | 266 | 0.721 | 0.014 | 0.001 | −0.045 to 0.047 | 0.969 |
| QALYs for different periods | |||||||||
| QALYs (baseline to 3 months) | 245 | 0.180 | 0.004 | 284 | 0.176 | 0.004 | 0.004 | −0.007 to 0.015 | 0.471 |
| QALYs (3–6 months) | 218 | 0.181 | 0.008 | 267 | 0.177 | 0.007 | 0.004 | −0.007 to 0.015 | 0.487 |
| QALYs (6–12 months) | 209 | 0.363 | 0.007 | 258 | 0.361 | 0.009 | 0.007 | −0.020 to 0.024 | 0.85 |
| QALY (baseline – 12 months based on available EQ-5D-3L data at all time points) | 202 | 0.729 | 0.017 | 249 | 0.719 | 0.013 | 0.01 | −0.033 to 0.054 | 0.642 |
Mean NHS and PSS costs (excluding intervention costs) for the 12-month period were similar between the SRT and SH groups (£189.16 vs. £182.13). After including intervention costs, mean NHS and PSS costs for the SRT arm were significantly higher than the SH arm (£268.26 vs. £190.86).
Cost-effectiveness and cost–utility analysis
Table 31 reports results from the base-case cost–utility analysis, sensitivity analyses and secondary exploratory analyses. The base-case analysis was conducted from an NHS and PSS perspective and used NHS and PSS costs and QALYs (obtained from the EQ-5D-3L) over the 12-month follow-up, applying multiple imputation for missing data and controlling for baseline characteristics. It generated incremental costs of £43.59 (95% CI −18.41 to 105.59) and incremental QALYs of 0.021 (95% CI 0.0002 to 0.042) associated with SRT relative to SH. This resulted in a mean ICER of £2076 per QALY gained. The probability that the SRT is cost-effective at the NICE cost-effectiveness threshold of £20,000 per QALY gained was 95.3%, with a mean NMB of £377.84. The probabilities that SRT is cost-effective at the NICE cost-effectiveness threshold of £15,000 and £30,000 per QALY were 94.4% and 96.2%, with respective mean NMBs of £272.12 and £589.28. The cost-effectiveness plane (Figure 4) displays graphically the uncertainty surrounding the mean ICER estimate, while the CEAC (Figure 5) summarises the effects of uncertainty surrounding the value of the cost-effectiveness threshold.
FIGURE 4 Cost-effectiveness plane representing bootstrapped mean differences in costs and QALYs for SRT compared with SH (NHS and PSS costs and QALYs based on EQ-5D-3L). (Figure not reproduced here; see the original article.)
FIGURE 5 Cost-effectiveness acceptability curve for the base-case analysis (NHS and PSS costs and QALYs based on EQ-5D-3L). (Figure not reproduced here; see the original article.)
| Scenario | Mean cost (£) (SE) | Mean QALY (SE) | ICER (£) | Probability that SRT is cost-effective at NICE cost-effectiveness thresholds | Mean NMB (95% CI) at NICE cost-effectiveness threshold | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SRT | SH | Incremental cost (bootstrap 95% CI) | SRT | SH | Incremental QALYs (bootstrap 95% CI) | £15,000 (%) | £20,000 (%) | £30,000 (%) | £15,000 | £20,000 | £30,000 | ||
| Base-case analysis | |||||||||||||
| NHS and PSS cost and QALYs based on EQ-5D-3L [multiple imputation (n = 50); covariates adjusted]a | 266.00(25.55) | 222.41(26.24) | 43.59(−18.41 to 105.59) | 0.723(0.008) | 0.702(0.008) | 0.021(0.0002 to 0.042) | 2076 | 94.4 | 95.3 | 96.2 | 272.12(261.59 to 282.65) | 377.84(364.06 to 391.62) | 589.28(568.96 to 609.60) |
| Sensitivity analyses | |||||||||||||
| NHS and PSS cost and QALYs based on EQ-5D-3L (complete-case analysis; covariates adjusted) | 252.35(28.41) | 199.23(25.45) | 53.12(−20.00 to 126.23) | 0.742(0.009) | 0.726(0.008) | 0.017(−0.007 to 0.041) | 3125 | 85.8 | 88.0 | 89.7 | 207.96(195.93 to 219.98) | 295.02(279.23 to 310.81) | 469.15(445.77 to 492.52) |
| Societal cost and QALYs using EQ-5D-3L [multiple imputation (n = 50); covariates adjusted] | 2340.62(189.56) | 3426.75(186.21) | −1086.13(−1485.59 to−686.67) | 0.723(0.008) | 0.702(0.008) | 0.021(0.0003 to 0.042) | Dominates | 100 | 100 | 100 | 1404.30 (1387.82 to 1420.77) | 1510.54 (1491.78 to 1529.30) | 1723.03 (1698.99 to 1747.06) |
| NHS and PSS cost using £2.52 as SRT training cost and QALYs using EQ-5D-3L [multiple imputation (n = 50); covariates adjusted] | 236.82(25.55) | 222.41(26.24) | 14.41(−47.59 to 76.41) | 0.723(0.008) | 0.702(0.008) | 0.021(0.0002 to 0.042) | 686 | 96 | 96.30 | 97 | 301.30(290.77 to 311.83) | 407.02(393.24 to 420.80) | 618.46(598.14 to 638.78) |
| Secondary analyses | |||||||||||||
| NHS and PSS cost and QALYs based on EQ-5D + Sleep [multiple imputation (n = 50); covariates adjusted]b | 266.22(25.55) | 222.20(26.23) | 44.02(−17.93 to 105.97) | 0.789(0.005) | 0.767(0.005) | 0.022(0.010 to 0.034) | 2001 | 99.6 | 99.9 | 100 | 281.58(275.32 to 287.84) | 390.58(382.51 to 398.64) | 608.57(596.83 to 620.31) |
| NHS and PSS cost and QALYs based on SF-6D [multiple imputation (n = 50); covariates adjusted]c | 266.19(25.55) | 222.23(26.23) | 43.95(−17.88 to 105.78) | 0.666(0.004) | 0.642(0.004) | 0.025(0.015 to 0.034) | 1758 | 100 | 100 | 100 | 324.36(319.35 to 329.37) | 447.58(441.19 to 453.97) | 694.03(684.81 to 703.25) |
| NHS and PSS cost and QALYs based on EQ-5D-3L (per-protocol analysis) [multiple imputation (n = 50); covariates adjusted] | 257.25(25.67) | 220.86(26.17) | 36.40(−27.72 to 100.52) | 0.722(0.009) | 0.703(0.008) | 0.019(0.00009 to 0.038) | 1916 | 92.3 | 94.9 | 95.9 | 256.72(247.01 to 266.43) | 353.97(341.28 to 366.67) | 548.47(529.76 to 567.19) |
| NHS and PSS cost and QALYs based on EQ-5D-3L (6-month follow-up) [multiple imputation (n = 50); covariates adjusted] | 198.37(18.36) | 126.60(17.01) | 71.76(31.94 to 111.59) | 0.362(0.004) | 0.347(0.004) | 0.015(0.006 to 0.025) | 4784 | 97.6 | 98.7 | 99.8 | 158.38(153.47 to 163.28) | 235.13(228.74 to 241.53) | 388.65(379.24 to 398.05) |
All sensitivity analyses confirmed the robustness of the result that SRT is likely to be cost-effective at a cost-effectiveness threshold of £20,000 per QALY (see Table 31 and Figures 6–11). Indeed, when the societal perspective was used for the analysis, SRT cost less and generated more QALYs, on average, and so dominates SH in health economic terms. When the SRT training cost was reduced to £2.52 from £31.7, the ICER reduced to £686 per QALY gained (from £2076 in the base-case analysis).
FIGURE 6 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs and QALYs based on EQ-5D-3L, complete-case analysis). (Figure not reproduced here; see the original article.)
FIGURE 7 Cost-effectiveness acceptability curve (NHS and PSS costs and QALYs based on EQ-5D-3L, complete-case analysis). (Figure not reproduced here; see the original article.)
FIGURE 8 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (societal costs and QALYs based on EQ-5D-3L). (Figure not reproduced here; see the original article.)
FIGURE 9 Cost-effectiveness acceptability curve (societal costs and QALYs based on EQ-5D-3L). (Figure not reproduced here; see the original article.)
FIGURE 10 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs using £2.52 as SRT training costs, and QALYs based on EQ-5D-3L). (Figure not reproduced here; see the original article.)
FIGURE 11 Cost-effectiveness acceptability curve (NHS and PSS costs using £2.52 as SRT training costs, and QALYs based on EQ-5D-3L). (Figure not reproduced here; see the original article.)
Secondary analyses (see Table 31 and Figures 12–20) also demonstrated that SRT is likely to be cost-effective when other utility measures were used, applying per-protocol analysis (those attending at least one treatment session), and using data for 6 months follow-up only. When the cost–utility analysis was repeated for the SF-6D and EQ-5D-3L, the point estimates of cost-effectiveness were very similar to those using the EQ-5D-3L, although the 95% CIs were smaller. After taking account of uncertainties, the SF-6D produced a stronger conclusion; that SRT has a 100% probability of being cost-effective at the cost-effectiveness threshold of £20,000 per QALY. The probability of cost-effectiveness when using the EQ-5D-3L + Sleep was also higher than that of the EQ-5D-3L (99.9% probability of being cost-effective at a cost-effectiveness threshold of £20,000 per QALY).
FIGURE 12 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs and QALYs based on EQ-5D-3L + Sleep ‘bolt on’). (Figure not reproduced here; see the original article.)
FIGURE 13 Cost-effectiveness acceptability curve (NHS and PSS costs and QALYs based on EQ-5D-3L + Sleep ‘bolt on’). (Figure not reproduced here; see the original article.)
FIGURE 14 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs and QALYs based on SF-6D). (Figure not reproduced here; see the original article.)
FIGURE 15 Cost-effectiveness acceptability curve (NHS and PSS costs and QALYs based on SF-6D). (Figure not reproduced here; see the original article.)
FIGURE 16 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs and QALYs per-protocol analysis). (Figure not reproduced here; see the original article.)
FIGURE 17 Cost-effectiveness acceptability curve (NHS and PSS costs and QALYs per-protocol analysis). (Figure not reproduced here; see the original article.)
FIGURE 18 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs and QALYs at 6-month follow-up). (Figure not reproduced here; see the original article.)
FIGURE 19 Cost-effectiveness acceptability curve (NHS and PSS costs and QALYs at 6-month follow-up). (Figure not reproduced here; see the original article.)
FIGURE 20 Cost-effectiveness plane: bootstrapped mean differences in costs and QALYs (NHS and PSS costs and reduction of ISI scores between baseline and 12-month follow-up). (Figure not reproduced here; see the original article.)
Furthermore, restricting the time horizon of the economic evaluation suggested that SRT remains cost-effective over 6 months post randomisation (ICER of £4784), but not as cost-effective as over 12 months post randomisation.
The ICER was estimated at £14 per unit reduction in ISI when reduction of ISI score was used as the measure of effectiveness. At a hypothetical cost-effectiveness threshold of £30 per unit reduction in ISI, SRT had a probability of cost-effectiveness of 88.1%.
For treatment response (defined as ISI reduction of ≥ 8 points), the mean incremental cost was £44.23 and the mean incremental probability of a treatment response was 0.26 with an ICER of £170, indicating a mean incremental cost of SRT of £170 is required to achieve a clinically relevant treatment response.
Conclusion
The economic analysis within the HABIT trial evaluated the cost–utility of SRT compared with SH in the NHS primary care setting. The analysis quantified the mean cost of SRT as £84, although the initial training costs are likely to reduce following scaling-up of the intervention, or delivery via alternative methods. Further implementation research is needed to consider how nurse-delivered SRT would operate in practice. For example, nurses were trained by experienced members of the research team but in practice we envisage that training and ongoing support could be provided by a clinical psychologist, or mental health professional with experience in sleep disorders and cognitive–behavioural approaches.
The primary cost–utility analysis used EQ-5D-3L-derived QALYs and NHS and PSS costs over the 12-month follow-up. Based on available data, the SRT arm produced a mean QALY of 0.73 versus 0.72 for the SH arm. After data imputation and adjustment for baseline covariates, the mean difference in QALYs between the two arms over the 12-month follow-up period was estimated at 0.02. Although the mean QALY gain is not large, the ICER was estimated at £2076 per QALY gained, which suggests great potential for SRT to be cost-effective at the NICE £20,000 per QALY cost-effectiveness threshold. Further exploration of the decision uncertainty around the estimate of mean ICER showed that SRT has a 95.3% probability of being cost-effective at a £20,000 cost-effectiveness threshold, and indeed a 94.4% probability of being cost-effective at a £15,000 cost-effectiveness threshold. Sensitivity analysis using the available sample with no imputation also confirmed the cost-effectiveness of SRT, although the mean ICER was larger relative to the baseline analysis that applied multiple imputation. When adjusting the SRT training cost, the mean ICER decreased from £2076 to £686. SRT had the effect of reducing productivity-related losses, which was reflected in the sensitivity analysis that adopted a societal perspective where SRT dominates SH.
The exploratory analysis using QALYs derived from the EQ-5D-3L and the SF-6D confirmed the conclusion that SRT is highly likely to be cost-effective, and the probabilities of SRT being cost-effective are higher using the EQ-5D-3L + Sleep and the SF-6D than using the EQ-5D-3L. Restricting the economic evaluation to a 6-month time horizon also confirms that SRT is highly likely to be cost-effective, but not as cost-effective as when a longer time horizon is adopted.