This is a proof-of-concept trial that will employ a randomised control trial design to assess the potential efficacy of personal financial incentives on pregnant women smoking cessation rates; combined with a qualitative analysis to investigate the acceptability of the intervention to participants and clinic staff. The study will involve a baseline survey; eight intervention sessions; and a post-intervention telephone interview of participating women and semi-structured interviews with clinic staff. Self-reported smoking status will be validated through saliva cotinine analysis using NicAlert® semi-quantitative dipstick assay. Continuous or long-term smoking cessation will be additionally assessed through hair analysis of patient’s cotinine levels using gas chromatography – mass spectrometry.
The study will take place at the antenatal clinic within a large public teaching hospital in New South Wales, Australia. The antenatal clinic operates a 5 days-a-week service with most women presenting for an average of 6 to 8 clinic visits over the course of their pregnancy. The hospital manages approximately 4,300 deliveries each year with the majority of these women receiving antenatal care predominantly through the hospital antenatal clinic.
A sample of ninety (90) consenting pregnant women will be recruited to participate. Women will be considered eligible if they are: (1) aged at least 16 years; (2) presenting for their first antenatal visit; (3) less than 31 weeks gestation; (4) sufficient in English language to complete the survey; and (5) a current smoker (defined as having smoked in last 7 days and a return a positive saliva sample for cotinine). Pregnant women will be ineligible to participate if they have elected to receive shared antenatal care with their GP and/or are considered by clinic staff to have a severe cognitive or psychiatric disorder; currently being treated for chemical dependency other than alcohol or tobacco; or have quit smoking before their first antenatal appointment.
Consenting women will be randomly allocated using block randomisation into one of three groups: control vs. smaller incentive ($AUS20) vs. larger incentive ($AUS40). The unit of randomisation will be the day and session time (i.e. morning or afternoon) of the patient’s first antenatal visit to minimise potential contamination and deception. Randomisation will be performed offsite, by an independent statistician prior to study initiation. Allocation of days of the week to groups will be done using Proc Plan in SAS.
Two intervention arms will be assessed: (1) a $AUD20 incremental personal financial incentive; and (2) a $AUD40 incremental personal financial incentive. Women from both intervention groups will have an opportunity to receive a PFI at eight study intervention sessions contingent upon smoking abstinence. The amount of the monetary reward will begin at the base amount (either $AUD20 of $AUD40) and will increase by the base amount (either $AUD20 of $AUD40) at each follow-up that participants are found to abstain from smoking. Smoking abstinence will be assessed through self-report and confirmed by saliva cotinine analysis. Participants in the $AUD20 intervention group, who quit smoking and maintain cessation for the entire eight intervention sessions, will be potentially eligible to receive a total of $AUD720. Participants in the $AUD40 group will be potentially eligible to receive a total of $AUD1440 if they maintain smoking abstinence for the 8 intervention sessions. If a participant fails to abstain at one intervention session or does not present for a scheduled visit they will not receive the incentive for that period. However, the reward amount these women will be eligible for during their next intervention session will be the amount they missed during their previous session. Their potential reward amount will remain at this amount until they have successfully stopped smoking.
All women who visit the hospital antennal clinic will undergo their usual antenatal interview by the clinic nurse. During this interview patients are asked about their smoking status and basic demographic characteristics. At this time potentially eligible patients will be identified and informed that the study is being conducted within the clinic. If women are interested they will be directed to a Research Assistant (RA) who will be stationed within the clinic. The RA will provide interested women with information describing the study. If patients would like to take part in the study the RA will obtain their written informed consent. Women will then be asked to provide a saliva sample to confirm their smoking status and complete a 15-minute baseline Touchscreen computer survey, which will also ask for their on-screen consent.
If a woman’s initial antenatal appointment has been randomised to one of the two intervention groups, the RA will verbally explain the details of their incentive scheme, including, the schedule for monitoring, contingencies being offered and the rules of payment. All participants will be provided with a study package to take home with them, which will contain information about the study, referrals to Quit smoking self-help documents and a Quit telephone help number (provided by clinic midwives) and a letter of support to pass onto their partner. Consenting women will be asked to attend 8 intervention sessions, at which participants in the two intervention groups will be eligible to receive a reward. The 8 intervention sessions will occur approximately fortnightly, during participants’ scheduled antenatal appointments.
During each of the 8 intervention sessions, all women will be asked to indicate their current smoking status and whether they have quit smoking in the last 7 days. Women will also be asked to provide a saliva sample which will be analysed to confirm their self-report smoking abstinence. Those women in the intervention groups who self-report non-smoking and return a negative cotinine reading (level 0) will be provided with their monetary incentive. A similar follow-up procedure will occur seven more times (i.e. a total of 8 intervention sessions) during women’s scheduled antenatal appointment.
During their eighth or final intervention session (i.e. the last follow-up where women are eligible for a reward) women will be asked to provide a hair sample to allow for cotinine analysis using gas chromatography – mass spectrometry. The hair analysis will allow for long-term assessment of women’s smoking status. Women will also be asked to consent to a post-intervention study approximately six weeks after delivery. Women who consent will be contacted via telephone. A short semi-structured telephone interview will be conducted assessing women’s acceptability of the intervention and their current smoking status. A similar interview will be conducted with clinic staff to assess their acceptability of, and feedback on, the program. Both participants and staff will be asked to provide feedback on whether they believe the intervention was useful in increasing smoking abstinence in pregnant women, whether the monetary reward was sufficient and/or delivered in an appropriate manner, whether they believe this intervention could be easily introduced into standard care and whether they experienced any problems with the intervention.
Primary outcome measures
The RA will maintain detailed records documenting: (1) the number of eligible women asked to take part in the research study and how many of these women consent to take part; (2) the number of participants who withdraw from the study before study completion; and (3) the number of women who complete or refuse to provide a saliva and/or hair sample for cotinine analysis. This data will be used to measure the following three primary outcomes: (i) consent rates; (ii) loss to follow-up rates of study participants and (iii) participant compliance with saliva and hair cotinine analyses for biochemical validation of smoking status
Semi-structured interviews will be conducted with both participants and clinic staff. Women will be asked a serious of previously published and study specific questions about their experiences and perceptions of the intervention trial. Specifically, six questions previously published by Bryant et al.
] will be included in the interviews with study participants, to assess their acceptability of using a touchscreen computer for self-reported smoking status and demographics. Five (5) items previously adapted by the researchers to assess pregnant women’s acceptability of PFI as a strategy to decrease smoking in pregnant women
], will be included in both participant and staff interviews. These questions will be used to investigate the acceptability of incentives of varying magnitude to staff and study participants. Open-ended questions will also be included in both participant and staff interviews to assess their thoughts on the acceptability of using saliva and hair cotinine analysis for biochemical validation of smoking status, as well as any perceived barriers and facilitators to using financial invectives.
Secondary outcome measures
To measure participants’ self-reported smoking status women will complete one baseline and eight follow-up self-report touchscreen computer surveys. The computer survey was programed using Digivey survey software
]. Use of computer surveys to assess smoking status has been shown to be a reliable and accurate method of data collection
], as well as being reported as enjoyable and easy to complete by most participants
].Questions assessing participants’ demographic characteristics and factors previously found to be associated with smoking status will also be included. The following questions will be included in the self-report, touchscreen computer surveys:
• Smoking status and history
will be assessed through standardised questions or questions adapted from previous research
] with many items based on the previous the work of Bryant et al.
] Items will include: current smoking status
], number of cigarettes currently smoked per day
], smoking history, previous quit attempts, reductions in smoking
], strategies used to assist in previous quit attempts
], readiness/stages of change
], financial stress
], whether they smoked during any previous pregnancies
]and whether they attempted to quit or reduced smoking during any previous pregnancies
]. Questions assessing women’s exposure to environmental smoke will also be included
• Demographic characteristics: women will be asked to answer questions relating to their: age, date of birth, highest level of education, Aboriginal or Torres Strait Islander status, current household income, marital status, number of previous pregnancies, gestation period, postcode, number of adults and children (<18 years) residing in their household.
• The Patient Health Questionnaire2
]: Will be included as a short, simple measure of participants’ levels of depression. The PHQ2 consists of two questions, which assess respondent’s level of depressed mood in the previous 2 weeks
]. Respondents rate the frequency of depressed mood and anhedonia on a four point likert scale, ranging from 0 (“not at all”) to 3 (“nearly every day”)
]. Total score on the PHQ2 range from 0 to 6. A score of 4 has recommended as an appropriate cut-point for screening for possible depressed mood in pregnant women
]. The PHQ-2 has evidence of criterion validity; as well as acceptable specificity (92%) and sensitivity (83%)
]. Previous research has found a relationship between depression and smoking status
• Economic questions: Women will be asked to a number of questions relating to their personal out-of-pocket costs, time costs and health care utilisation incurred as a result of the intervention.
• Social characteristics: women will be asked whether their partner or others within their household smoke.
Follow-up surveys: During each of the follow-up visits, women will be asked to complete a brief touchscreen computer survey using a number of items from the baseline survey. Questions will assess their current smoking status, number of cigarettes smoked per day (if they have not abstained), changes in smoking behaviour, partners’ current smoking status, exposure to environmental smoke, the number of quit attempts made, gestation age; the PHQ2 and efforts to engage with quit smoking support services.
The following two biochemical analyses will be used to validate women’s self-reported smoking status:
1. Saliva cotinine analysis
: will be used to biochemically validate self-reported smoking status for all participants at baseline and each of the eight intervention sessions using the NicAlert® semiquantitative dipstick assay. A previous study analysing 167 family practice patients using NicAlert® illustrated acceptable levels of sensitivity (99%) and specificity (96%)
].Cotinine analysis with the NicAlert® dipstick takes approximately 20 minutes and can be performed by untrained persons, highlighting the potential utility of such a test in clinical practice .
2. Hair cotinine analysis
: will be undertaken during the eighth or final intervention session. Participants will be asked to provide a small sample of hair taken from the nape of the neck. Samples will be transported to an off-site laboratory for analysis. Hair cotinine analysis can reliably quantify long-term exposure to tobacco, with each centimetre of scalp hair representing approximately 1 month of past exposure
]. Gas chromatography – mass spectrometry will be used to analyse hair cotinine levels, as it is considered the standard reference in analysis of cotinine
]. A cut-point of 0.2 ng/mg has been found to distinguish between pregnant active smokers and passive or unexposed with high levels of sensitivity (91%), specificity (94%) and test accuracy (91.7%)
]. We will also validate hair cotinine analysis as a measure of smoking status by comparing its accuracy against women’s self-report smoking status and saliva cotinine throughout the study.
Economic outcomes measures
The feasibility of undertaking a cost-effectiveness analysis will also be undertaken. This will include the collection of data on the costs of the interventions that would include the amount of money received by each participant, and the costs of administering the payments in terms of staff time and other resources used. The intervention may also encourage patients to attend their antenatal visits, and potentially other health care visits associated with their pregnancy, more often than the control group, and so data will also be collected on patients’ utilisation of health services throughout the study. The feasibility of obtaining these data from Medicare, PBS, and hospital separations data, including obtaining patients’ consent, will be investigated and compared to gathering this data directly from patients using a standard questionnaire administered at each visit. The costs incurred by patients are also likely to be affected by the intervention. This will include savings in purchasing cigarettes, and differences in out of pocket, travel and time costs of attending if the intervention influences health care utilisation. These data will be collected from a short patient survey administered at each visit. Out of pocket costs for GP visits and prescribed medication can also be collected from MBS and PBS records. The effectiveness data on quitting will be collected as mentioned above, and combined with the data on costs. Incremental cost-effectiveness ratios (ICERs) will be calculated
A convenience sample size totalling 90 women will be recruited, with 30 women randomised to each of the three study groups. As this is a proof-of-concept trial, the sample size will not be adequately powered to detected small differences however a sample size of 30 women in each group will allow for detection of a difference in smoking rates between groups of 24%, with 5% significance level and 68% power. Based on previous studies, this will be sufficient to provide an estimate of the difference in proportions between groups for the following primary outcomes: (i) consent rates; (ii) lost to follow-up rates; (iii) participant compliance to saliva and hair cotinine analysis; (iv) acceptability of incentives and validation measures, and the secondary outcome variable, smoking cessation. To allow for a 20% drop-out rate, a consent rate of 60% and 10% of women being ineligible, 208 pregnant women who smoke will be approached. Based on a previous study conducted by the investigators, approximately 17% of women seen at this clinic are self-reported smokers
Baseline demographic and social characteristics of patients in all the three treatment groups will be summarised separately. Continuous data will be expressed as mean values with standard deviations (SD), and categorical data will be presented as counts with percentages.
Proportions and frequencies will be calculated to assess the following outcomes: (i) participant consent rates; (ii) loss to follow-up rates; and (iii) participant compliance with saliva and hair cotinine analysis. Proportions and frequencies will also be calculated for all closed-ended questions included in the study participant and clinic staff interviews to allow investigation of: (iv) participant acceptability of touchscreen computers for self-reported smoking rates and demographics; and (v) acceptability of incentives to staff and study participants. Differences between experimental groups in the proportion of each of these five primary outcomes and the secondary outcome (proportion quit smoking) will be assessed through Fisher’s exact test. Differences between participant subgroups and these outcomes will also be assessed using Fisher’s exact test. Although p-values from the above will be reported, the focus will be on providing 95% confidence intervals around point estimates for comparisons of all three treatment arms.
Qualitative analysis will be performed on open-ended questions included in participant and staff interviews investigating the following primary outcomes: (vi) acceptability of saliva and hair cotinine analyses for biochemical validation of smoking status; and (vii) perceived barriers and facilitators to using financial incentives. All interviews will be transcribed verbatim and a thematic analysis will be performed independently by two of the research team members.
Although not adequately powered for detecting small differences in the secondary outcomes, Fisher’s exact test will be used to detect potential differences in smoking cessation rates between groups and to demonstrate potential effectiveness of the intervention. Differences between experimental groups in the secondary outcome - PHQ2 will be assessed using the Kruskal-Wallis test. Fisher’s exact tests will be used to compare conditions on post-partum smoking rates at 6 weeks after delivery.
Validation analysis of saliva and hair cotinine compared to self-report smoking rates
Agreement between self-reported smoking status and the cotinine saliva test strip will be assessed using the Kappa statistic. Agreement between self reported smoking status and hair cotinine results will also be assessed using the Kappa statistic.
Quantities of resources used (e.g. number of visits) will be combined with their unit costs and calculated from existing sources using standardised methods. Differences in the average cost per patient in each arm of the trial will be compared with differences in quitting rates. These are used to calculate incremental cost-effectiveness ratios (ICERs) for each intervention arm compared to the control arm.
Ethics and safety approval
The study will be conducted in accordance with the Helsinki Declaration and has received ethical approval from both the Hunter New England Human Research Ethics Committee (Ref No: 11/09/21/4.05) and the University of Newcastle Human Research Ethics Committee (Ref No: H-2012-0047). Safety approval was also obtained from the University of Newcastle Health and Safety Review Committee (Ref: 26/2012).