1. Norwich Medical School Faculty of Medicine and Health Science Smoking relapse and health-related quality of life: Secondary analysis of data from a trial on smoking relapse prevention Song F 1 Bachmann MO 1 Aveyard P 2 Barton GR 1 Brown TJ 1 Brandon TH 4 Maskrey V 1 Blyth A 1 Notley C 1 Holland R 1 Sutton S 3 1.Norwich Medical School, University of East Anglia, Norwich, Norfolk, UK 2.Nuffield Department of Primary Care Health Sciences, University of Oxford, Oxford, UK 3.Institute of Public Health, University of Cambridge, Cambridge, UK 4.Department of Health Outcomes and Behavior, Moffitt Cancer Center, Tampa FL, USA

2. Background Smoking and smoking cessation are associated with health- related quality of life (HRQoL). Uncertain whether trying and failing to quit increases the level of psychological distress and depression Uncertain about the causal direction between HRQoL and smoking relapse, and the effects may work in both directions

3. Study objective To compare changes in HRQoL in people who maintained abstinence with people who had relapsed to smoking. A secondary analysis using data from a relapse prevention trial

4. Methods: Data source SHARPISH: RCT of self-help booklets for smoking relapse prevention (ISRCTN36980856) Funded by NIHR/HTA programme (09/91/36), to investigate the effect of self-help materials for the prevention of smoking relapse in 4-wk quitters who have used NHS Stop Smoking Services. Experimental: 8 Forever Free booklets Control: Leaflet -learning to Stay Stopped Main findings: No difference between the intervention groups in smoking outcomes by 12 months: continuous abstinence 31% vs. 34% ( P=0.231 )

5. EQ-5D instrument - HRQoL European Quality of Life -5 Dimensions- 3 Level Participants completed the EQ-5D questions at baseline (1 mon after the quit date), and at 3 and 12 mon after the quit date Five dimensions: (1) mobility (2) self-care (3) usual activity (4) pain/discomfort (5) anxiety/depression For each dimension, there are three possible responses: o “no problems”, o “some or moderate problems”, and o “severe or extreme problems”.

6. EQ-5D: data analysis methods Two methods; (1) utility scores, (2) dichotomisation EQ-5D utility score using York A1 tariff (Dolan & Roberts 2002): - Death=0, Full health=1; - Not normally distributed: nonparametric Wilcoxon rank-sum test - Ordinary least squares (OLS) regression (Vogl et al 2012): to explore the association between changes in utility score (dependent variable) and smoking outcomes, after adjusting for multiple baseline variables Dichotomisation of responses to each of the dimensions: - 0 = “no problem” vs. 1 = “some or severe problems” - Chi-square test and logistic regression analysis

7. Results: EQ-5D score and participant characteristics at baseline The number of people who completed the EQ-5D-3L questionnaire was 1348 (95.8%) at baseline, 1290 (91.7%) at the 3 month follow-up and 1171 (83.2%) at the 12 month follow-up. The mean EQ-5D tariff score was 0.8252 at baseline. After adjusting for other baseline characteristics, results of multivariable regression analysis found that EQ-5D scores at baseline were negatively associated with: age (P<0.001) being female (P=0.002) unemployment (P<0.001) first cigarette within 5 minutes of waking (P<0.001),

8. Utility scores by smoking outcome at 12 mon AbstinentRelapsed P value nMeann Baseline4420.82899030.82340.412 At 3 months4510.82478360.82610.873 At 12 months4510.83077200.79240.040

9. Results of multiple variable regression analysis VariableCoefficient (95% CI) P value Continuous abstinence0.0288 (0.0006, 0.0571) 0.045 Age0.0002 (-0.0009, 0.0014) 0.674 Sex (female 1 vs male 0)-0.0039 (-0.0318, 0.0239) 0.782 Married or living with a partner-0.0137 (-0.0425, 0.0150) 0.349 Unemployed0.0124 (-0.0375, 0.0624) 0.626 Low education0.0007 (-0.0289, 0.0303) 0.964 Free prescription-0.0054 (-0.0369, 0.0261) 0.737 First cigarette in 5 min after waking-0.0084 (-0.0365, 0.0197) 0.558 Note: Dependent variable was utility change, which was the difference between utility value at 12 months and utility value at baseline. Utility change >0 indicates the improved utility at 12 months.

10. Proportion of participants with “some/severe problems” in different EQ-5D dimensions by continuous abstinence In each EQ-5D dimension the “some problems” and “severe problems” were combined as “some/severe problems”. Differences between the relapse and abstinent was statistically significant only for anxiety/depression dimension at 12 months (P=0.001).

11. “Some/severe problems” in anxiety/depression and smoking status at 3 and 12 months Point smoking was defined as any smoking during the 7 days before the follow-up. The first and second digits in the bracket refer to smoking at 3 and 12 months respectively: “0” indicates no smoking and “1” smoking. For example: smoking (0, 0) indicates no smoking at 3 and 12 mon; and smoking (0, 1) indicates no smoking at 3 mon and smoking at 12 mon.

12. Discussion In the current study, the mean EQ-5D score was 0.8252 (SD 0.2594) at baseline and 0.8072 (SD 0.2456) at 12 months, which is similar to the score reported for a representative sample of English former smokers 0.8225 (SD 0.2512) At follow-up, those who had relapsed had a lower quality of life, while those maintaining abstinence had a small increase in HRQoL. The difference in HRQoL between the relapsers and continuous quitters was due to changes in anxiety/depression, with no evidence of changes in other EQ-5D dimensions.

13. Causal direction of smoking relapse and anxiety/depression Evidence from the current study: o Return to smoking by 3 mon was associated with somewhat greater anxiety/depression at baseline (1 month after the quit date) o Anxiety/depression at 1 or 3 mon were not associated with smoking relapse at 12 mon o Relapse to smoking by 3 mon was associated with an increase in anxiety/depression at that time, while returning to abstinence by 12 mon was associated with reduced anxiety/depression at 12 mon The relation was possibly bidirectional (Cohen & Lichtenstein 1990) Effects of anxiety/depression on smoking relapse occur with little delay Relapse to smoking itself is likely a stressful incident Beneficial effects of smoking cessation on anxiety/depression may not be perceived immediately, and often complicated by withdrawal symptoms

14. Limitations Results may not be generalizable to ethic minority groups, pregnant women, and young people. As HRQoL data were collected infrequently, this study alone cannot clarify the causal pathway. The current study was observational in nature, and we could not rule out the existence of other factors or events that were associated with both relapse to smoking and changes in EQ-5D responses.

15. Conclusions Relapse to smoking is associated with reduced quality of life measured by EQ-5D, which seems mainly attributable to problems of anxiety/depression. The relationship between relapse to smoking and anxiety/depression is likely to be bidirectional.

16. Main references Blyth A, Maskrey V, Notley C, et al. Self-help educational materials for the prevention of smoking relapse: randomised controlled trial (SHARPISH). Health Technology Assessment 2015. (Forthcoming). Vogl M, Wenig CM, Leidl R, Pokhrel S: Smoking and health-related quality of life in English general population: implications for economic evaluations. BMC public health 2012, 12:203. Shahab L, Andrew S, West R: Changes in prevalence of depression and anxiety following smoking cessation: results from an international cohort study (ATTEMPT). Psychological medicine 2014, 44(1):127-141. Dolan P, Roberts J: Modelling valuations for EQ-5D health states: an alternative model using differences in valuations. Med Care 2002, 40(5):442-446. Taylor G, McNeill A, Girling A, Farley A, Lindson-Hawley N, Aveyard P: Change in mental health after smoking cessation: systematic review and meta-analysis. BMJ 2014, 348:g1151. Cohen S, Lichtenstein E: Perceived stress, quitting smoking, and smoking relapse. Health psychology : official journal of the Division of Health Psychology, American Psychological Association 1990, 9(4):466-478. Taylor G, McNeill A, Aveyard P: Does deterioration in mental health after smoking cessation predict relapse to smoking? BMC public health 2015, 15:1150.

17. Acknowledgements This project was funded by the NIHR HTA programme (Project 09/91/36). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Department of Health. We thank stop smoking advisors from NHS stop smoking services in Norfolk, Lincolnshire, Suffolk, Hertfordshire, Great Yarmouth and Waveney for recruiting quitters to the study. The trial was conducted in collaboration with the Norwich Clinical Trials Unit whose staff provided input into the design, conduct and analysis (Tony Dyer – randomisation and data management). We thank Laura Vincent for providing administrative support, data entering and data checking. Declaration of interest: PA has done ad hoc consultancy and research for the pharmaceutical industry on smoking cessation. THB has received research funding and study medication from Pfizer, Inc. No other competing interest declared; no other relationships or activities that could appear to have influenced the submitted work.