The trends on obesity are shocking. UK data on the prevalence of obesity in adults and children are provided by the National Obesity Observatory, which also shows the social gradient in obesity.

NICE’s guideline on obesity emphasised prevention, but recommended medical and surgical treatment. Bariatric surgery was an option for those with a body mass index (BMI) >40 or >35 with comorbidies with a cost per quality adjusted life year (£/QALY) of around £10k. Orlistat was also an option, plausibly between £20k and £30k per QALY. Incorporating these treatments into the clinical guideline made them less mandatory. Big geographical variations persist for bariatric surgery. Neither the restrictions on bariatric surgery nor on orlistat were formally estimated. The models used were simple and took minimal account of the impact of surgery on mortality or diabetes.
“Shedding the pounds” extended the NICE cost impact beyond the NHS  to include costs to patients. Based on one small study of 73 patients it showed more of them in work and working longer after surgery compared with before. The report argued that the increased income from employment offset the costs of bariatric surgery—that is, to society if not to the NHS. It acknowledged, however, its method for costing the work aspects were optimistic and assumed people could readily get back to work.

The National Bariatric Surgery Register provides useful data on how bariatric surgery is being done and to whom. It shows 30% being privately funded (mainly by women). This percentage seems to be falling as NHS provides more, but the unfortunately the full report was paywalled. Mortality was 0.1%, complications 2.6%.  28% had Diabetes Mellitus 2, of which 50% were cleared by year 2. The average weight loss was 58% of their body weight. 50% of patients reported a return to no impairment.

I searched in vain for a table showing the relative cost effectiveness of the range of anti obesity interventions. The best I found was the OECD report “Fit not Fat,” which took a microeconomic perspective, arguing that government intervention in lifestyle choices can be justified if the  market “fails.” Causes of market failure include: information gaps, externalities (social and private costs and benefits differ), and, of course, non-rationality (lack of self control, addiction). The OECD work used the WHO chronic disease prevention model to estimate the cost effectiveness of nine interventions covering mass media, food advice, school, work, labelling, physician advice, tax. The impact of the interventions was indirectly via diseases (stroke, heart disease, cancer), rather than directly on obesity over 100 years. The results showed a limited impact on the prevalence of obesity, mortality, or disability adjusted life years (DALYs), especially in first few decades. Assuming (not unreasonably) that DALYs roughly equal QALYs, none of the interventions would meet the NICE threshold of £30k except for tax which was cost saving. The interventions with the largest effects were physician counselling (most costly) and taxation. The importance of this work included the emphasis on tax (not in NICE’s remit) and the key role for present day valuations of future health states (the so called discount rate applied to non monetary benefits).

The health economic implications, based on the £/QALY mainstream approach,  would be to fund those interventions with lowest £/QALY, regardless of disease, emergency, age etc, and to continue until the budget ran out. This would probably imply that the NHS should continue to fund bariatric surgery until the resulting £/QALY fell, and then move on  to fund the next best obesity interventions in terms of £/QALY. Few if any NHS preventive interventions would be funded due to their high £/ QALY (or per DALY).

The high £/QALY for preventive interventions is partly due to the orthodoxy of placing lower values on future health states (discounting). Since the benefits of prevention occur only in the longer run, they are much reduced by discounting. It could be argued that preventive interventions should be discounted at a very low or zero rate. This is what happened with Stern’s economic analysis of climate change. Stern described climate change as the greatest market failure in history. Does obesity rank at number two? And what difference would low discount rates make?

I am driven to conclude that bariatric surgery is a highly cost effective intervention for morbidly obese people who are prepared to have surgery. However, the restrictions imposed by NICE are based on out of date modelling, which took limited account of the impact of surgery on diabetes and mortality. Revised modelling is needed of the cost effectiveness of the entire range of interventions in a disease rather than technology oriented model. This should include effects beyond the NHS, particularly the willingness of many people to pay for bariatric surgery and also the impact that surgery has on privately borne costs, particularly employment.

Methods for assessing the cost effectiveness of obesity interventions need to consider use of low or zero discount rates and consider the parallels between the obesity epidemic and climate change. Taxation (fat taxes) needs to be considered actively despite likely political opposition.

James Raftery is a health economist with several decades experience of the NHS. He is Professor of Health Technology Assessment at Southampton University. A keen ‘NICE-watcher’, he has provided economic input to technical assessment reports for NICE but has never been a member of any of its committees. The opinions expressed here are his personal views.