Predictors of retention in a randomised trial of smoking cessation in low-socioeconomic status Australian smokers☆,☆☆
Introduction
Participant attrition (Toerin, Brookes, Metcalfe, et al., 2009) is a potential problem in interpreting the findings of clinical trials especially when participants permanently drop out of a study (Lugtig, 2014). There are two principal types of participant attrition: (i) drop out/withdrawal (i.e. participants that no longer wish to participate in any further data collection/study demands); and (ii) loss to follow-up (i.e. participants who are not retained/or lost without reason) (Goldberg, Francois Chastang, Zins, et al., 2006). It is a common problem in clinical trials. For example, a review of health care intervention randomised control trials (RCTs) in six major journals (Toerin et al., 2009), found that 48% of trials that reported a sample size calculation failed to retain adequate numbers at outcome assessment (Toerin et al., 2009, Severi et al., 2011).
Excessive loss to follow-up can prolong recruitment, reduce statistical power, threaten the internal validity of study findings, compromise the generalisability of study outcomes, and waste money (Leon et al., 2007, Szklo and Nieto, 2012). Study results can be biased when participants retained differ from those who are not (Robinson, Dennison, Wayman, et al., 2007) and bias may be even more pronounced when loss to follow-up differs between intervention and comparison groups (Sprauge, Leece, Bhandri, et al., 2003). Assessment of the characteristics/factors associated with attrition is needed (Goldberg et al., 2006) to assess for selection biases and loss of statistical power (Goldberg et al., 2006, Ellenberg, 1994, Hunt and White, 1998) and these need to be considered in study data analysis and interpretation (Goldberg et al., 2006, Shih, 2002, Twisk and de Vente, 2002). As a rule of thumb, some suggest that loss to follow-up under 5% will result in little bias but over 20% loss can significantly threaten study validity (Severi et al., 2011, Sprauge et al., 2003). Studies indicate that often those participants with incomplete follow-up data, while similar at baseline to those with complete data, may be systematically different at follow-up (Leak et al., 2015, Woolard et al., 2004). Consequently this may limit generalisability of the results and lead to incorrect inferences about treatment effects (Leak et al., 2015). It is imperative that researchers get as close to complete follow-up data as possible (Severi et al., 2011, Sprauge et al., 2003).
Low-socioeconomic status (low-SES) populations have lower participation rates and higher loss to follow-up rates (Bonevski, Randell, Paul, et al., 2014). Low-SES populations also have characteristics that make follow-up more difficult, including substance abuse and mental health disorders, housing instability, intermittent telephone access, incarceration, and less understanding of and exposure to research (Blumental et al., 1995, Cunningham et al., 2008, Ramos-Gomez et al., 2008). Lower education, low health literacy, and financial stress are also associated with incomplete research follow-up (Leak et al., 2015).
Systematic review evidence shows that few behavioural interventions for smoking cessation have been undertaken for low low-SES smokers (Bonevski et al., 2014, Bryant et al., 2011, Courtney et al., 2015). For example, a recent review found only one Australian study that reported abstinence rates for smokers with a psychotic disorder (Bryant et al., 2011); however the factors associated with the high 83% retention obtained at 12-month follow-up was not evaluated (Baker et al., 2006, Baker et al., 2007). The most recent review examining attrition rates in smoking cessation studies found only nine studies (Belita & Sidani, 2015), and none had examined a low-SES or low income population. Consequently, this study is the first to describe and evaluate the factors associated with retention for low-SES smokers enrolled in a pragmatic RCT.
An increasing body of evidence indicates the challenge facing disadvantaged populations is staying quit, rather than forming the goal of quitting and trying (Borland, 2013). Low-SES smokers are understudied (Courtney et al., 2015) and they face some unique challenges that may reduce the likelihood of study retention. For example, low-SES smokers tend to have higher nicotine dependence (Bobak et al., 2000, Hyland et al., 2006, Siahpush et al., 2006), in addition to more smokers in their social networks and stress in their day-to-day lives (Paul, Ross, Bryant, et al., 2010), but these factors association with retention are yet to be tested for low-SES smokers. In the general smoking population, smoking-related, socio-demographic, behavioural, and health-related factors have been linked to retention, but little is known about the role of these factors and recruitment source in retention of low-SES smokers. Length of previous quit attempts (Borrelli et al., 2002, Leeman et al., 2006) and confidence in quitting (Nevid, Javier, & Moulton, 1996) are associated with study retention but evidence is mixed for cigarettes smoked per day (Nevid et al., 1996, Bowen et al., 2000, Curtin et al., 2000). On the whole, the association between study retention and other socio-demographic characteristics (e.g. age, (Leeman et al., 2006, Fortman and Killen, 1994), education level, (Borrelli et al., 2002, Curtin et al., 2000) sex, (Greenberger & Knab, 2000) and number of dependent children) (Leeman et al., 2006), behavioural/psychological factors (e.g. weight concerns (Leeman et al., 2006), feelings of guilt, IQ (Beaver, 2013, Lynham et al., 1993)) and health-related factors (e.g. depression (Curtin et al., 2000), body mass index (BMI) and other health risk behaviours) (Goldberg et al., 2006, de Graaf et al., 2000, Deeg et al., 2002, Morrison et al., 1997, Siddiqui et al., 1996) is conflicting.
Further, there is an absence of data from smoking cessation clinical trials in socially disadvantaged populations (Bonevski et al., 2014). Many studies have failed to analyse the independent contributions of these factors to follow-up (Leak et al., 2015). Little effort has been made to investigate other factors that may be more salient in low-SES population groups, for example mental health disorders and poorer physical health (Leak et al., 2015). If factors associated with drop out in smoking cessation trials in low-SES populations are identifiable at study commencement, measures can be taken to enhance retention (Nowak, Sharif, Eischen, et al., 2014).
Our aims were to: (1) describe the retention rates in the Financial Interventions for Smoking Cessation Among Low-Income Smokers (FISCALS) RCT and (2) identify whether smoking-related, health-related, behavioural, socio-demographic characteristics, or recruitment source were associated with retention at 2- or 8-month follow-ups.
Section snippets
Study design
The FISCALs pragmatic RCT was funded (APP1021862) by the Australian National Health and Medical Research Council (NHMRC) (Courtney, Bradford, Martire, et al., 2014). The trial design was in accordance with the CONSORT statement; the trial is registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12612000725864). In brief, the current trial tested an innovative approach to improving smoking cessation outcomes in low-SES smokers by providing financial education and support
Sample
In total, 2557 people were assessed for eligibility, of which 1047 (41%) participants were randomised. The main exclusion from being randomised was where the applicant was not low-SES (n = 555, 22%). Of those randomised, 946 (90%) completed the 2-month follow-up and 880 (84%) completed the final 8-month follow-up. Details of recruitment and retention are presented in the CONSORT diagram (as shown in Supplementary Fig. 1). Retention rate at 8-month follow-up were similar for the control (86%) and
Discussion
The > 80% retention of low-SES participants herein is the highest recorded in any community-based smoking cessation trial (Belita & Sidani, 2015). The study results suggested that retention was less associated with general and mental health issues than might be expected and appears to be linked more with engagement and commitment to smoking cessation. We found that there were no significant differences in retention for those with co-morbid mental health conditions, smokers receiving additional
Acknowledgments
We would like to thank Emma Black, Veronica Boland, Deborah Bradford, Danya Braunstein, Philip Clare, Jaimi Iredale, Sundresan Naicker, Joel Tibbetts, Lauren Touyz and Emily Upton for their work on the project. We acknowledge the assistance of Quitline services, and Australian Government Department of Human Services Centrelink Customer Service Centres, for assisting with recruitment and staff at the HRF for their diligence with data collection. This research is funded by a grant (APP1021862)
References (55)
- et al.
Characteristics of smokers with a psychotic disorder and implications for smoking interventions
Psychiatry Research
(2007) Intelligence and selective attrition in a nationally representative and longitudinal sample of Americans
Personality and Individual Differences
(2013)- et al.
Attrition in the Longitudinal Aging Study Amsterdam. The effect of differential inclusion in side studies
Journal of Clinical Epidemiology
(2002) - et al.
Higher magnitude cash payments improve research follow-up rates without increasing drug use or perceived coercion
Drug and Alcohol Dependence
(2008) - et al.
Health problems were the strongest predictors of attrition during follow-up of the GAZEL cohort
Journal of Clinical Epidemiology
(2006) - et al.
Tracking and follow-up of 19,915 adolescents: minimizing attrition bias
Controlled Clinical Trials
(1997) - et al.
Recruitment and retention stratagies in longitudinal clinical studies with low-income populations
Contemporary Clinical Trials
(2011) - et al.
Factors affecting attrition in a longitudinal smoking prevention study
Preventive Medicine
(1996) - et al.
Attrition in longitudinal studies. How to deal with missing data
Journal of Clinical Epidemiology
(2002) - et al.
A randomized controlled trial of a smoking cessation intervention among people with psychotic disorder
The American Journal of Psychiatry
(2006)
Attrition in smoking cessation studies
The Canadian Journal of Nursing Research
Recruitment and retention of subjects for a longitudinal cancer prevention study in an inner-city black community
Health Services Research
Smoke intake among smokers is higher in lower socioeconomic groups
Tobacco Control
Comparing socially disadvantaged smokers who agree and decline to participate in a randomised smoking cessation trial
BMJ Open
Reaching the hard-to-reach: a systematic review of strategies for improving health and medical research with socially disadvantaged groups
BMC Medical Research Methodology
Understanding hard to maintain behaviour change: A dual process approach
Predictors of quitting and dropout among women in a clinic-based smoking cessation program
Psychology of Addictive Behaviors
Recruiting women into a smoking cessation program: Who might quit?
Women & Health
A systematic review and meta-analysis of the effectiveness of behavioural smoking cessation interventions in selected disadvantaged groups
Addiction
Proactive recruitment predicts participant retention to end of treatment in a secondhand smoke reduction trial with low-income maternal smokers
Translational Behavioral Medicine
A randomized clinical trial of a financial education intervention with nicotine replacement therapy (NRT) for low socio-economic status Australian smokers: a study protocol
Addiction
Smoking cessation among low-socioeconomic status and disadvantaged population groups: A systematic review of research output
International Journal of Environmental Research and Public Health
Tracking inner city subtance users from the emergency department: how many contacts does it take?
Academic Emergency Medicine
Determinants of attrition from cessation treatment in smokers with a history of major depressive disorder
Psychology of Addictive Behaviors
Psychiatric and sociodemographic predictors of attrition in a longitudinal study: The Netherlands Mental Health Survey and Incidence Study (NEMESIS)
American Journal of Epidemiology
Recruitment and retention of low-income minority women in a behavioural intervention to reduce smoking, depression, and intimate partner violence during pregnancy
BMC Public Health
Selection bias in observational and experimental studies
Statistics in Medicine
Cited by (21)
Causal relationships between pain, medical treatments, and knee osteoarthritis: A graphical causal model to guide analyses
2024, Osteoarthritis and CartilagePatterns and predictors of nicotine replacement therapy use among alcohol and other drug clients enrolled in a smoking cessation randomised controlled trial
2021, Addictive BehaviorsCitation Excerpt :Given education was not a consistent predictor of NRT use across timepoints or types of use there may be a couple of explanations for this result. There is some evidence to indicate that higher levels of education are associated with retention in treatment among smoking cessation trials (Courtney et al., 2017; Patterson et al., 2003; Urbanoski, Kenaszchuk, Inglis, Rotondi, & Rush, 2018; Veldhuizen, Zawertailo, Noormohamed, Hussain, & Selby, 2021). However, as the NRT use question in the 6.5-month follow-up survey assessed any use over the past 6-months rather than use at that specific follow-up timepoint it may also be plausible that those with a higher level of education better understood what the survey item was asking and/or they were more accurately able to recall the information.
Predictors of attrition in a smoking cessation trial conducted in the lung cancer screening setting
2021, Contemporary Clinical TrialsCitation Excerpt :An understanding of the barriers to tobacco treatment is a necessary first step to increasing the quit rates among those undergoing LCS, which can lead to an increase in the public health impact of screening [6]. Study results support prior findings that lower education [15,30] and heavier smoking [12,13,31] are associated with greater trial attrition. Importantly, novel findings indicated that trial attrition was associated with factors specific to the LCS setting.
Single-arm trial of the second version of an acceptance & commitment therapy smartphone application for smoking cessation
2017, Drug and Alcohol DependenceCitation Excerpt :Response choices ranged from “Not at all” (1) to “Very much” (5). To detect a potential signal in this short-term trial, the two-month post-enrollment follow-up survey assessed 7-day and 30-day point prevalence abstinence (Brunette et al., 2015; Courtney et al., 2016; Salepci et al., 2016). Due to cost and low demand characteristics for false reporting, the SRNT Subcommittee on Biochemical Verification recommends biochemical confirmation is unnecessary in population-based studies with limited face-to-face contact and studies where the optimal data collection methods are through the mail or telephone (Benowitz et al., 2002).
Predictors of client retention in a state-based tobacco quitline
2020, Journal of Smoking Cessation
- ☆
Clinical trial registration: trial registered with the Australian and New Zealand Clinical Trials Registry (ACTRN) 12612000725864.
- ☆☆
Declaration of interests: Professor Billie Bonevski has received research funding from a Pfizer Investigator Initiated Grant. Professor Robert West undertakes research and consultancy for companies that develop and manufacture smoking cessation medications. He is unpaid co-director of the National Centre for Smoking Cessation and Training, a not-for-profit organisation involved in training and assessing stop smoking practitioners.