By Professor Jean Adams

It’s a time of year when it’s nice think about the successes of the last twelve months, but sometimes things aren’t easy or straightforward in science, and it’s important to reflect on those occasions too.

Earlier this month, one of our papers (Pell et al., 2021)¹ on the Soft Drinks Industry Levy was retracted by the BMJ and a correction published in BMJ Open (Rogers et al., 2023)².

Our original BMJ paper reported that at 1 year post implementation, the levy was associated with a reduction in household purchasing of sugar from all soft drinks combined of 30g (or 10%) per household per week, with no change in volume of all soft drinks purchased.

While revisiting the data during subsequent analysis, we found that the analysis in the original paper contained an error. During preparation for the original analysis, a weighting variable was incorrectly calculated. This variable was also redundant, replicating a second weighting variable.

Whilst she didn’t conduct the original analysis, the error was found by Nina Rogers. To her credit, Nina not only worked out that there was an error, but also found what was causing it, and conducted all of the corrected analysis. Without Nina’s careful work and integrity we may never have found or solved this problem. That’s why the corrected paper is Rogers et al, whilst the original paper was Pell et al.

We know the scientific literature is full of errors and false leads: as a body of work it becomes self-correcting as we move falteringly towards discovery. But when we discover an error in our own work we also have to be part of that correction.

So when we found the error in this paper, we alerted The BMJ and submitted a rapid response so that readers would also be aware and that a full correction would follow.

The corrected version published in BMJ Open replicates the original analysis without the redundant and incorrectly calculated weighting variable. Most of the results in the paper are impacted by the correction – some by a negligible amount, others more substantively.

Overall, the estimate of change in purchasing of sugar from all soft drinks combined at 1 year post implementation of the levy reduces from -30g (or -10%) per household per week to -8g (or -3%) per household per week. Alongside, the estimate of change in volume of soft drinks purchases increased from no change to a 189ml (or 3%) increase per household per week.

Although the effect in the correct paper is smaller, we still believe these results have important public health implications. The SDIL may have had a public health benefit (by reducing household purchasing of sugar from soft drinks) whilst potentially benefiting industry (by increasing total volume of soft drinks purchased).

On top of that, our wider evaluation of the levy has been pointing to a range of other promising findings in relation to reformulation,³ obesity⁴ and dental health⁵ in children, public opinion in adults⁶ and adolescents,⁷ and industry turnover⁸ and stock market value.⁹ We’ve also documented the response of industry¹⁰ and parliamentarians¹¹ to the levy. The levy is a complex intervention – there was never going to be a single finding that could make sense of all its effects.

It’s taken a long time from originally finding the error to the corrected version being published. The BMJ commissioned two extra peer review processes. The first to inform their decision about whether to retract and republish a correction; versus issuing a correction on the original. The second a normal peer review of the corrected paper in BMJ Open.

We knew that this correction could have repercussions, so as well as working with the journal, we’ve emailed corresponding authors of all papers that cited the original paper. We briefed colleagues working in civil society, and those in the UK government working on public health and health taxes. Our funders and steering group have also been kept updated throughout. We thank them for their understanding, support and constructive engagement with us. Retraction Watch contacted us and we were happy to speak to them.

This has been a difficult experience for all of the authors. But, as ICMJE say, “honest errors are part of science and publishing”.

The key thing I’ve learnt (remembered) through all of this is to remain humble and that everyone makes mistakes. Making mistakes is to be expected. What matters is that when we find them, we own up to them and fix them. We’re glad to be able to correct the record.

1. Pell D, Mytton O, Penney TL, Briggs A, Cummins S, Penn-Jones C, et al. Changes in soft drinks purchased by British households associated with the UK soft drinks industry levy: controlled interrupted time series analysis. BMJ 2021;372:n254. doi: 10.1136/bmj.n254
2. Rogers N, Pell D, Mytton O, Penney T, Briggs A, Cummins S, et al. Changes in soft drinks purchased by British households associated with the UK soft drinks industry levy: a controlled interrupted time series analysis. BMJ Open 2023;13(12):e077059. doi: 10.1136/bmjopen-2023-077059
3. Scarborough P, Adhikari V, Harrington R, Elhussein A, Briggs A, Rayner M, et al. The impact of the announcement and implementation of the UK Soft Drinks Industry Levy on sugar content of soft drinks in the UK, 2015-18: controlled interrupted time series analysis. PLoS Med 2020;17(2):e1003025.
4. Rogers NT, Cummins S, Forde H, Jones CP, Mytton O, Rutter H, et al. Associations between trajectories of obesity prevalence in English primary school children and the UK soft drink industry levy: an interrupted time series analysis of surveillance data. medRxiv 2022:2022.09.16.22280030. doi: 10.1101/2022.09.16.22280030
5. Rogers N, Conway D, Mytton O, Roberts C, Rutter H, Sherriff A, et al. Trends in childhood hospital admissions for carious tooth extractions in England in relation to the UK soft drink industry levy: an interrupted time series analysis of Hospital Episode Statistics. medRxiv 2023:2023.02.27.23286504. doi: 10.1101/2023.02.27.23286504
6. Adams J, Pell D, Penney TL, Hammond D, Vanderlee L, White M. Public acceptability of the UK Soft Drinks Industry Levy: repeat cross-sectional analysis of the International Food Policy Study (2017–2019). BMJ Open 2021;11(9):e051677. doi: 10.1136/bmjopen-2021-051677
7. Jones CP, Armstrong-Moore R, Penney TL, Cummins S, Armitage S, Adams J, et al. Adolescents’ perspectives on soft drinks after the introduction of the UK Soft Drinks Industry Levy: A focus group study using reflexive thematic analysis. Appetite 2022;179:106305. doi: https://doi.org/10.1016/j.appet.2022.106305
8. Law C, Cornelsen L, Adams J, Pell D, Rutter H, White M, et al. The impact of UK Soft Drinks Industry Levy on manufacturers’ domestic turnover. Econ Hum Biol 2020
9. Law C, Cornelsen L, Adams J, Penney T, Rutter H, White M, et al. An analysis of the stock market reaction to the announcements of the UK Soft Drinks Industry Levy. Econ Hum Biol 2020;38:100834. doi: 10.1016/j.ehb.2019.100834 [published Online First: 2020/02/23]
10. Jones C, Penney T, Forde H, Cummins S, Adams J, Law C, et al. Industry views of the Soft Drinks Industry Levy: a thematic analysis of elite interviews with UK food and drink industry professionals, 2018-20. In preparation for submission to Social Science & Medicine 2022
11. Penn-Jones C, Lawlor E, Forde H, Penney T, Cummins S, White M. OP40 ‘There is no silver bullet’ how parliamentary debate on the UK soft drinks industry levy changed over time (2014–2020): an applied thematic analysis. J Epidemiol Community Health 2021;75(Suppl 1):A19-A19. doi: 10.1136/jech-2021-SSMabstracts.40