|
Economic evaluation of internet-based interventions for harmful alcohol use alongside a pragmatic randomized controlled trial |
Blankers M, Nabitz U, Smit F, Koeter MW, Schippers GM |
|
|
Record Status This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn. CRD summary This study evaluated the cost-effectiveness of internet-based interventions to reduce harmful alcohol use. The authors concluded that internet-based therapy offered value for money, compared with internet-based self-help, and could be considered as a first intervention or as a second step. The reporting was generally clear, but limited reporting on trial methods, significant missing data, inappropriate exclusion of the waiting-list control, and an inadequate time horizon, mean that the findings should be interpreted with caution. Type of economic evaluation Cost-effectiveness analysis, cost-utility analysis Study objective This study evaluated the cost-effectiveness of internet-based interventions to reduce the harmful use of alcohol. Interventions Internet-based self-help was compared with internet-based therapy. A waiting-list control was part of the trial, but not included in the economic evaluation. Both interventions were based on cognitive-behavioural therapy and motivational interviewing. Self-help introduced participants to treatment exercises designed to help them cope with craving, drinking lapses and peer pressure, without the assistance of a therapist. The therapy intervention had seven individually themed 40-minute written chat sessions with a therapist that were accompanied by homework assignments. Location/setting Netherlands/secondary care. Methods Analytical approach:The cost-effectiveness and cost-utility analyses were conducted alongside a small randomised controlled trial; 205 participants were recruited and 105 received an intervention, between 2008 and 2009. The authors stated that the analysis was conducted from the societal perspective. The evaluation had a six-month time horizon. Effectiveness data:The study recruited patients from a highly visited, substance abuse treatment centre website (Jellinek). Participants were required to be between 18 and 65 years old, resident in the Netherlands, have health coverage, have internet access at home, score above eight on the Alcohol Use Disorders Identification Test (AUDIT), and consume 14 or more alcohol units (10g ethanol per unit) weekly. They could not have had previous substance abuse treatment, and had to have a limited risk of severe problems during the intervention due to drug abuse, and/or physical and mental illness. The outcome data were collected at the start, and three months and six months after randomisation. The primary clinical outcome was the treatment response, based on reduction of alcohol consumption to no more than 14 standard alcohol units for women and 21 for men, during the last seven days, without any deteriorations of more than 10% in AUDIT score, the Quality of Life Scale, or the Global Severity Index. Further details were available in another publication by the authors. Analyses were conducted using intention-to-treat. Missing values were handled using multiple imputation. Monetary benefit and utility valuations:EQ-5D utility scores were used and were based on UK tariffs. Measure of benefit:The primary measure of benefit for the cost-effectiveness analysis was the proportion of treatment response. For the cost-utility analysis, the primary measure of benefit was quality-adjusted life-years (QALYs). QALYs were calculated for the six-month time horizon of the study. Cost data:All costs for the interventions and health care uptake, and the opportunity costs of the participants time, and productivity losses, were included. Intervention costs consisted of software and technology services, overheads, and the therapist for the therapy intervention. Therapist costs were based on actual work time, using national profession-specific wages. Participant costs consisted only of their time spent on the interventions, valued as leisure time using standard Dutch methods. Productivity losses were calculated from Dutch sources, using the human capital approach. Presenteeism (the cost of being less productive at work due to illness) and absenteeism (the cost of being absent from work) were included. The health care costs were from a global burden of disease study of alcohol-attributable societal costs. Costs were indexed to 2010 Euros (EUR) using a consumer price index. Analysis of uncertainty:Non-parametric bootstrapping was used on each of the imputed sets of data to produce probabilistic incremental costs, effects, and incremental cost-effectiveness ratios. These simulations were displayed on a cost-effectiveness plane and as a cost-effectiveness acceptability curve. Sensitivity analyses were conducted by varying the most important cost drivers and using only the health care costs. Results Total societal costs were EUR 2,010 for therapy and EUR 1,120 for self-help; an incremental difference of EUR 845. The proportion of patients who responded was 0.53 (36/68) for therapy and 0.29 (20/68) for self-help; a difference of 0.24 in favour of therapy. The EQ-5D scores were 0.89 for therapy and 0.78 for self-help; a difference of 0.12 in favour of therapy. No patient died, so incremental QALYs were reported as 0.6 in favour of therapy, for the six-month study. The incremental cost-effectiveness ratio was EUR 3,683 per additional responder at six months, for therapy compared with self-help, or EUR 14,710 per QALY. The sensitivity analyses found that the ratios from the health services perspective were a third or less of those from the societal perspective. The societal costs were the primary driver of cost-effectiveness. The likelihood of cost-effectiveness was 60% at a willingness-to-pay of EUR 20,000 per QALY, and 85% at a willingness-to-pay of EUR 40,000 per QALY. Authors' conclusions The authors concluded that internet-based therapy offered value for money, compared with internet-based self-help, and could be considered as a first intervention or in a stepped-care approach. CRD commentary Interventions:The interventions were well described and appear to have been appropriate. The waiting-list control was inappropriately omitted from the economic evaluation; this control could have represented no intervention. Effectiveness/benefits:The study sample consisted of a specific subgroup of people who were seeking help for alcohol consumption problems. The time horizon of the trial and economic evaluation was insufficient to capture the downstream effects of an intervention designed to combat alcoholism, a life-long condition with many associated diseases and a high potential for future societal costs. Using a decision-analytic model to predict future benefits and costs would have been appropriate. No data were provided on who the utility scores were derived from; it was also unclear why UK tariffs were used instead of validated Dutch tariffs. It was not clear if the EQ-5D was an outcome measure in the trial. The trial had significant missing data; the reasons for this were not explored. Of 136 participants randomly allocated, only 92 completed data at three-month follow-up, and only 80 completed data at six months. It was unclear how many data were missing for each outcome and cost category at each follow-up. It was not possible to determine if the data were missing at random or due to participant and intervention characteristics; therefore, it was not clear whether the use of imputation was justified. Costs:The costs were clearly presented and appear to have been appropriately derived, except that it was not clear how presenteeism was calculated. The time horizon was insufficient to adequately capture the future benefits and costs of the interventions. In the trial sample there were large differences in absenteeism and presenteeism; over the six months preceding the intervention, therapy participants had costs of EUR 756 for absenteeism and EUR 1,137 for presenteeism, while self-help participants had absenteeism costs of EUR 1,863 and presenteeism costs of EUR 794. There were large differences in the baseline costs of absenteeism, with large variance. Analysis and results:While reporting was generally clear, problems with the trial data, inappropriate exclusion of the waiting-list control, and a far too short time horizon, make the validity of the results difficult to assess. The original trial and the economic evaluation proceeding from it appear to have been poorly conducted, leading to substantial loss to follow-up and inappropriate methodological choices. The sensitivity analyses for the health care provider perspective should be disregarded, as six months was not long enough for most health care costs associated with excessive alcohol consumption to occur in the relatively low-risk participants. Sensitivity analysis found that the primary driver of cost-effectiveness was the societal costs. The authors acknowledged the limitations of the generalisability of the cost data and the short time horizon. Concluding remarks:The study was clearly written, but limited reporting on the trial methods, significant missing data, inappropriate exclusion of the waiting-list control, and an inadequate time horizon for capturing all the benefits and costs, mean that the findings should be interpreted with caution. Funding Funded by the Netherlands ZonMw Addiction II Program. Bibliographic details Blankers M, Nabitz U, Smit F, Koeter MW, Schippers GM. Economic evaluation of internet-based interventions for harmful alcohol use alongside a pragmatic randomized controlled trial. Journal of Medical Internet Research 2012; 14(5): e134 Indexing Status Subject indexing assigned by NLM MeSH Adult; Alcoholism /therapy; Cost-Benefit Analysis; Female; Humans; Internet /economics; Male; Middle Aged; Netherlands AccessionNumber 22013004037 Date bibliographic record published 20/09/2013 Date abstract record published 20/01/2015 |
|
|
|