There’s a big debate in the world of obesity treatment. Does it help to inform someone they’re obese, or does it cause unnecessary distress?
Those who believe the former will be buoyed by a new study, published in the Lancet medical journal on Tuesday, which attracted excited news coverage. With usual restraint, the Sun shouted, ‘HONESTY IS BEST POLICY: GPs must TELL patients they’re fat’, while the Guardian announced, ‘GPs should not worry about offending obese patients’. Both took their lead from the Lancet’s own press release, which stated ‘the findings should provide reassurance to doctors who rarely talk to patients about their weight for fear of causing offense’.
But does the study support this conclusion? And can, as the BBC declared, a 30-second chat reallytrigger weight loss? The answer as ever requires delving beyond the headlines.
The study, led by a team from the University of Oxford, was a randomised controlled trial of GP-led referral to commercial weight loss programmes. Patients with an obese body mass index were recruited while waiting to see their GP, and given an envelope to take into their appointment. The GP opened the envelope, and depending on whether the patient had been randomly assigned to the intervention or control group, reacted accordingly. In both situations, the GP spent around 30-seconds recommending that the patient lose weight. Those allocated for intervention were referred to a commercial weight loss programme and helped to book their first appointment, while controls were left to their own devices.
That difference – whether the patient was actively helped to book onto a commercial weight loss programme or was left to their own devices – is what the study primarily tested. Not the effectiveness of the 30-second chat, and certainly not whether it’s helpful telling patients they’re obese.
The trial’s reported success is because those who received the intervention lost an extra 1.4kg (3.1lb) after 12-months compared to those who received the control. But since all the participants received a 30-second chat, it’s impossible to say whether this was itself useful, as there was nothing to compare it to. Although the control group were 1kg (2.2lb) lighter at the end of the study than they were at the beginning, this might not be that surprising, because of a statistical quirk called regression-to-the-mean, first described by statistician Francis Galton, who found that tall parents had children who were, on average, shorter than their parents. This law of measurement occurs because any extreme value (one far away from average) is more likely to be a high water point than a true extreme. Think how often you’ve discovered an amazing new dish at a restaurant, only to find it disappointing the second time around. Along with the placebo-effect, this is one of the key reasons why a control group is so important.
So, to the question of whether a 30-second chat can trigger weight loss, the answer is maybe, maybe not. We know onlythat a 30-second chat followed by immediate referral to a commercial weight loss programme can lead to modest weight loss over a period of 12-months. It’s a substantial leap from there to concluding that GPs can talk freely to their patients about their weight without fear of causing offense.
The study did ask participants whether they found the chat ‘appropriate’ and ‘helpful’, and almost all participants agreed. But this does not prove that all such conversations are appropriate and helpful. First because the results can’t be interpreted beyond the brief and highly scripted discussions in the study. Second because questionnaires are highly susceptible to bias, depending on factors such as wording and how, where, and by whom questions are asked. To avoid this, the gold-standard approach to testing the pros and cons of a new intervention are in-depth (qualitative) interviews. Research of this type was planned in the study protocol, but was not included in the final publication. Qualitative research is considerably more difficult to publish in leading medical journals. The Lancet will hopefully welcome these results when they are available, and afford them with same publicity as they have the numerical findings.
Another difference between the protocol and the publication includes what’s been explicitly highlighted as a potential ‘competing interest’. In the protocol the authors state under ‘competing interests’ that, ‘Slimming World and Rosemary Conley… agreed to fund free treatment courses for NHS patients participating in the trial’. In the article, they state under ‘acknowledgments’ that, ‘the weight loss programmes… provided by Slimming World and Rosemary Conley Health…were donated to the NHS’. Semantic differences aside, the study has provided excellent marketing in the media and in the wording of the ‘typical’ GP conversation, which asked, “Did you know the best way to lose weight is to go to [Slimming World or Rosemary Conley]…?”. Rival weight-loss companies may dispute this, especially those that advocate more drastic methods such as meal-replacements. But the bigger issue, and perhaps the key limitation with spending just 30-seconds on something as complex as obesity, is what was not said.
As most who’ve tried to change their weight will affirm, the biggest challenge isn’t so much the initial weight-loss, but keeping it off. In fact the single best predictor of future weight gain is to currently be on a diet. How then should a GP respond when the same patient returns after a couple of years having regained their weight, or gained even more? And what impact might this have on the patient’s self-esteem, and their trust in their doctor? Until such questions are answered, I think GPs are right to pause and think carefully about how and when to raise something as sensitive as obesity.
Dr Peter Tennant is a University Academic Fellow in Healthcare Research and Statistical Epidemiologist at the University of Leeds. His research primarily studies the effects of obesity and diabetes in pregnancy. You can follow him on Twitter @pwgtennant.
Can a 30-second chat with a GP really trigger weight loss?
