|
An economic evaluation of adaptive e-learning devices to promote weight loss via dietary change for people with obesity |
Miners A, Harris J, Felix L, Murray E, Michie S, Edwards P |
|
|
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 adaptive electronic learning devices, to promote weight loss by dietary change, in obese individuals. The authors concluded that e-learning devices were unlikely to be cost-effective, without much lower fixed costs or much more effective designs, but the value of further research was high. The study was generally well reported and reached appropriate conclusions. Type of economic evaluation Study objective This study evaluated the cost-effectiveness of adaptive electronic learning devices, to promote weight loss by dietary change, in obese individuals. Interventions A hypothetical e-learning programme was assessed, based on an internet intervention, evaluated in a published UK study. This was compared against usual care, which did not include e-learning and pharmacological interventions. Methods Analytical approach:A discrete-event simulation evaluated the impact of the intervention and comparator, on 50-year-old patients, who were modelled until all of them had died. Eight types of patient, each with 1,000 simulations, with risk characteristics that could change over time, were modelled. The authors stated that a UK NHS perspective was adopted. Effectiveness data:The relative effectiveness of e-learning was measured by decreases in body mass index (BMI), derived from a published random-effects meta-analysis of e-learning studies, which produced a weighted mean difference. The primary driver of the model was the patient's BMI, which translated into increased risks of cardiovascular disease and type 2 diabetes, via published algorithms. Starting annual bodyweight increases were estimated using a published UK National Institute for Health and Clinical Excellence guideline on obesity. Monetary benefit and utility valuations:The utility values were based on a large UK study that assessed the relationship between BMI, other health issues, age, and EQ-5D utility scores. In the model, the development of two long-standing illnesses was permitted; cardiovascular disease and type 2 diabetes. Measure of benefit:Quality-adjusted life-years were the summary measure of benefit. Future QALYs were discounted at 3.5% annually. Cost data:The cost data were grouped into two categories: specific events, and the initial 12 months of treatment. The costs for cardiovascular disease were from a published drug study that reported one-off costs for fatal and non-fatal events, and an annual cost for survivors. Type 2 diabetes costs were from a published drug study that reported general practice visits, drug treatment for high blood pressure, and statin treatment. Intervention costs were calculated using the meta-analysis of e-learning studies, which included a one-time fixed cost for each patient, for the web service. The costs for health care visits, drugs, and slimming clubs were included. All costs were reported in UK £, and reflated to 2009, using the Personal and Social Services Research Unit (PSSRU) indices. Future costs were discounted at 3.5% annually. Analysis of uncertainty:Uncertainty was analysed in one-way sensitivity analyses, and a probabilistic sensitivity analysis, with 1,000 simulations for each of the eight patient scenarios. The results for the most cost-effective scenario were reported. The probabilistic sensitivity analysis results were presented on a cost-effectiveness acceptability frontier. An analysis was conducted including pharmacological interventions, as well as e-learning and usual care. Results The eight types of patient all resulted in incremental cost-effectiveness ratios over £100,000 per QALY gained. Non-smoking, non-diabetic males, with a BMI of 30 at baseline, had the best ratio at £102,112 per QALY gained. The ratios for other types ranged from £112,628 to £232,911 per QALY gained. Sensitivity analysis showed that the likelihood of being cost-effective was low. The only cost-effective result was created by reducing the fixed cost per patient for e-learning to zero, which made e-learning both more effective and less costly than usual care. The most important variables, in the sensitivity analyses, were the fixed cost of the e-learning devices, the relative treatment effect, and the duration of the treatment effect. An expected value of perfect information analysis, assuming a UK incidence of 308,000 obese persons annually, and time horizons of 10 and two years, with thresholds between £20,000 and £30,000 per QALY gained, resulted in a value between £37 million and £170 million (10 years), or £8 million and £36 million (two years). Authors' conclusions The authors concluded that e-learning devices were unlikely to be cost-effective, without much lower fixed costs or much more effective designs, but the value of further research was high. CRD commentary Interventions:The e-learning intervention was based on a UK study, with a reasonable justification given for the choice of study. The use of usual care, as the control, was appropriate and included variance in practice. Effectiveness/benefits:The effectiveness data were appropriately from a published systematic review of e-learning interventions. The algorithm used to predict the risk of cardiovascular disease, by BMI, was from a published study and was reported. The authors acknowledged that the choice of risk equation for type 2 diabetes was arbitrary, but they used the published risk equation in a sensitivity analysis. The methods used to identify the risk algorithms and equations were not reported, nor were their choices justified. As a result, it is not clear if they were the best available options. Costs:The sources and types of costs were generally well reported. Costs that were not from the price year of the analysis were appropriately reflated to 2009 values, using PSSRU indices. The costs were presented as means for events, with no breakdown of the content of these events. For most costs, such as fatal cardiovascular disease events, this was reasonable, but the annual cost of usual care should have been broken down and justified. Overall, the costing was adequate. Analysis and results:The model was an individual sampling model, which was appropriate for a condition where patients have dynamic risks over time. As acknowledged by the authors, the number of risk factors in the model was small, and adding further risk factors could provide additional information. Obesity is a risk factor for several diseases, so limiting the risks to cardiovascular disease and type 2 diabetes, and not including other diseases might have underestimated the value of the intervention. A thorough sensitivity analysis, with generally good reporting, was presented. Appropriate distributions were used for the costs and the assumptions were transparent. The expected value of perfect information analysis was appropriately conducted and concluded. A thorough evaluation of the strengths and weaknesses of the study was presented, with comparisons with other studies and a thorough explanation for any differences in the results. Concluding remarks:The study was generally well reported and reached appropriate conclusions. Funding Funded by the UK NIHR Health Technology Assessment programme. Bibliographic details Miners A, Harris J, Felix L, Murray E, Michie S, Edwards P. An economic evaluation of adaptive e-learning devices to promote weight loss via dietary change for people with obesity. BMC Health Services Research 2012; 12: 190 Indexing Status Subject indexing assigned by NLM MeSH Body Mass Index; Computer-Assisted Instruction /economics /methods; Female; Health Promotion; Humans; Internet; Male; Obesity /diet therapy; User-Computer Interface; Weight Loss AccessionNumber 22013003052 Date bibliographic record published 01/02/2013 Date abstract record published 13/08/2013 |
|
|
|