Re: Financial incentives for smoking cessation in pregnancy: randomised controlled trial
Authors reply: Criticisms can be political rhetoric and poor science
We thank the authors of two recent rapid responses for taking the time to respond to some of the issues raised in our article.
In terms of a response, it is helpful to start by emphasing that our trial (1) was a phase II therapeutic exploratory trial. ‘Phase II is usually considered to start with the initiation of studies in which the primary objective is to explore therapeutic efficacy in patients.’(2) ‘Additional objectives of clinical trials conducted in Phase II may include evaluation of potential study endpoints.’ We think it is important for all to understand that, although it had striking findings, this trial (1) was not designed as a phase III definitive trial and the critique offered by Braillon and Bewley (3) seems to miss this point entirely.
An additional focus of the responses is the validity of the primary outcome. We believe this is robust and provide further detail here.
The primary outcome was self-report of smoking cessation in late pregnancy corroborated by cotinine in saliva or urine (Russell Standard) (4). All participants were contacted in late pregnancy by specialist advisers from the NHS Stop Smoking Helpline call centre. These same advisers had gained telephone consent from participants to take part in the trial and provided concealed random allocation. At that time the study database randomly allocated a date from 34-38 weeks gestation from which advisers would start to try to contact the participant for follow up. This date was not known to the adviser or the participant so, although participants knew they would be followed up within a 4 week window, they did not know exactly when they would be phoned and asked about smoking status for primary outcome assessment. Women who self reported as quit were visited by a research nurse to collect a urine and saliva sample to test for cotinine and validate their reports. Additionally, for the final 200 participants, with a change of consent after a substantial ethics amendment, we performed a second validation of women’s self-reported smoking by testing residual routine blood samples taken, during routine ante natal care over the last 10 weeks of pregnancy. This extra, second validation was reassuring; we found that, in the intervention group, 4 / 18 (20%) women who, by our research processes were considered as quitters, had a blood cotinine level which indicated they were actually smoking but the figure in the control group was very similar, 1 / 5 (20%). This suggests that any bias in primary outcome assessment affected intervention and control group outcomes to a similar degree and will not have impacted on the relative difference in cessation outcomes between trial groups. The primary outcome is therefore quite solid and findings clearly demonstrate that, in this population, incentives were effective in achieving smoking cessation in pregnancy, at least. Relatively few trials of smoking cessation interventions in pregnancy have followed women up for long after pregnancy and we agree that, for incentives, as for other interventions, evidence for longer term efficacy is needed. Participants lost to follow-up were also examined using residual routine blood samples from late pregnancy assayed for cotinine and all tested in both intervention and control groups were smokers.
Cope (5) suggests using a quick test for cotinine rather than carbon monoxide breath testing to decide if self report quitters should receive financial incentives. We are familiar with his work and sympathetic to the need for more routine affordable biochemical validation of smoking status. Indeed, we have tried to apply this in one of our previous studies (6) which was a pragmatic observational study looking at ways of identifying smoking in pregnancy and routine referral to cessation services. However a key challenge is that existing stop smoking services in the UK (and maternity services) are simply currently not set up to use quick cotinine tests to verify abstinence. Instead they use CO testing which has some limitations as we acknowledge, but still provides manageable and very useful biochemical validation of smoking status. It is worth emphasising CO testing was not used to define the primary outcome of this study, which utilised cotinine validation.
With regards to Nicotine Replacement Therapy (NRT) in pregnancy, Braillon and Bewley cite an observational study which has findings which differ from those derived from RCTs; two large randomised placebo controlled trials (7, 8) in pregnancy both showed no significant difference in quit rate and the second titrated NRT to saliva cotinine levels prior to the quit attempt. The second study showed an increase in diastolic blood pressure associated with NRT. There is at present no evidence of the efficacy of NRT during pregnancy, however well it works in the non-pregnant population.
‘Is the improvement of healthcare providers’ motivational interviewing skills really so difficult?’ (3) We (Tappin) have significant firsthand experience of implementing a randomised controlled trial using motivational interviewing (MI) to help pregnant smokers quit during pregnancy in Glasgow (9). It was a challenge to teach a group of research midwives motivational interviewing skills (10) but with the support of a great teacher (Jeff Allison) who spent 8 days full time teaching the midwives and then 1 day per month honing their skills by discussing difficult cases the midwives became good practitioners. To make sure that the motivational interviewing skills were implemented, every intervention interview (625 undertaken in participants’ homes) was digitally recorded. A random 1 in 10 sample (63) were transcribed and sent with the audio recording to a motivational interviewing rating house at the Center on Alcoholism, Substance Abuse and Addictions(CASAA), at the University of New Mexico. The ratings indicated that the midwives could be described as “experts” at providing motivational interviewing in this setting. Outcome assessment for the trial, which included 762 pregnant smokers 351 intervention and 411 controls, was self report of quit in late pregnancy corroborated by cotinine estimation in residual routine late pregnancy blood samples. 4% of pregnant smokers had quit in the control group and 4% had quit in the intervention group. On the basis of our previous well conducted motivational interviewing (9) and NRT (7) trials in pregnancy we therefore take particular exception to the jibe that ‘Financial incentives look more akin to lazy bribery’ (3).
We do however agree that the case for implementing financial incentives across the NHS is far from proved, and that a phase III definitive trial of financial incentives to support smoking cessation during pregnancy should be carried out in the UK. Financial incentives are potentially a highly cost effective intervention to help pregnant smokers stop and to improve the health and wealth of the poorest in our society. It is important that this potentially highly effective health promoting intervention is not lost in a haze of political rhetoric and poor science.
David Tappin, Linda Bauld and Tim Coleman, on behalf of the CPIT study team.
1 Tappin D, Bauld L, Purves D, et al. Financial incentives for smoking cessation in pregnancy: randomised controlled trial. BMJ 2015 27;350:h134.
2 International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. General considerations for clinical trials, 1997.
Guideline E8, section 188.8.131.52. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Eff...
3 Braillon A, Bewley S. Neither carrots nor sticks should replace fairness and autonomy. http://www.bmj.com/content/350/bmj.h134/rr-3
4 West R, Hajek P, Stead L, Stapleton J. Outcome criteria in smoking cessation trials: proposal for a common standard. Addiction 2005;100:299-303.
5 Cope GF. Re: Financial incentives for smoking cessation in pregnancy: randomised controlled trial http://www.bmj.com/content/350/bmj.h134/rr-2
6 Bauld, L. Hackshaw, L, Ferguson, J et al. Implementation of routine biochemical validation and an ‘opt out’ referral pathway for smoking cessation in pregnancy, Addiction, 107, S2: 53-60.
7 Coleman T, Cooper S, Thornton JG, et al. A randomized trial of nicotine-replacement therapy patches in pregnancy. N Engl J Med 2012;366:808–18.
8 Berlin I, Grange G, Jacob N, Tanguy ML. Nicotine patches in pregnant smokers: randomised, placebo controlled, multicentre trial of efficacy. BMJ 2014;348:g1622.
9 Tappin DM, Lumsden MA, Gilmour WH, et al. Randomised controlled trial of home based motivational interviewing by midwives to help pregnant smokers quit or cut down. BMJ 2005;331:373-7.
10 Tappin D. Authors Reply: Teaching midwives motivational interviewing for a randomised controlled trial. http://www.bmj.com.ezproxy.lib.gla.ac.uk/rapid-response/2011/10/31/autho...
Competing interests: No competing interests