Family Practice Advance Access originally published online on June 21, 2007
Family Practice 2007 24(4):395-400; doi:10.1093/fampra/cmm028
Assessing the feasibility of proactive recruitment of smokers to an intervention in general practice for smoking cessation using computer-tailored feedback reports
a Department of Primary Care & Population Sciences, Royal Free and University College Medical School
b Institute of Public Health, University of Cambridge, UK
Correspondence to: Hazel Gilbert, Department of Primary Care & Population Sciences, Royal Free and University College Medical School, London NW3 2PF, UK; Email: hazel.gilbert{at}ucl.ac.uk
Received 25 September 2006; Revised 21 February 2007; Accepted 30 April 2007.
 |
Abstract |
|---|
 Top Abstract BackgroundMethodResultsDiscussionDeclarationReferences |
|---|
Background. Specialist National Health Service clinics for smoking cessation have increased in number, but most smokers prefer less intensive self-help and many smokers have no serious intentions to attempt to quit. Computer-tailored self-help materials can be adapted to provide advice to less motivated smokers, and can also take into account features such as level of education and socio-economic circumstance.
Objective. To assess the feasibility of delivering tailored feedback to a large population by identifying smokers from general practice records, with the aim of informing a large-scale trial of effectiveness.
Method. Questionnaires were sent to a random sample of smokers (n = 876) aged between 18 and 65 years, identified from records in four practices. Smokers returning the questionnaire (n = 78) were randomized to receive standard information, or to receive standard information plus computer-tailored feedback reports. Follow-up questionnaires were sent 3 months after the return of the baseline questionnaire.
Results. The recruitment strategy yielded a response rate of 8.9%, and a 66.7% follow-up rate. There were no significant differences in outcome between the two conditions, and no significant differences in outcome between practices. In the Intervention group significantly more of those who remembered receiving the tailored advice letter had made a quit attempt (6[60%]/3[21.4%], P < 0.05).
Conclusion. This pilot study demonstrated the feasibility of carrying out such a trial to evaluate the effectiveness of delivering an intervention for smoking cessation in primary care, and highlighted issues that should be addressed in considering the design of a large-scale trial.
Keywords. Feasibility, primary care, smoking cessation, tailored self-help.
|
Background |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionDeclarationReferences |
|---|
Specialist clinics in the NHS offering intensive face-to-face treatment for smoking cessation have increased in number over the last few years. While such interventions can produce relatively high abstinence rates, they are limited by low participation1,2 and participants are unrepresentative of smokers in the general population.3 Most smokers prefer less intensive self-help and minimal contact strategies,4–7 and despite well-publicized health risks, as many as 70% of smokers have no serious intentions to attempt to quit in the next year.8,9 There is an urgent need to target smokers less motivated to quit with a variety of minimal and low intensive approaches.
Self-help materials mainly of the generic ‘one-size-fits-all’ variety are widely available. The effectiveness of these interventions for smoking cessation is low partly because they lack the personal element that is central to the success rates of the clinical approach. Tailored self-help materials, intended to meet the needs of one specific person,10 are more effective in supporting behaviour change than standard materials.11 This tailored feedback can be adapted to provide tailored advice to less motivated smokers, to increase the awareness of smoking as a problem and increase the likelihood of taking action. Furthermore, with a strong socio-economic gradient in smoking prevalence,12 exacerbated by smoking cessation literature that is written at a level beyond the literacy skills of many smokers,13 advice can also be adapted to an appropriate style of language and interaction to account for features such as level of education and socio-economic circumstance.
Reactive recruitment, involving promotions and advertised services, attract only smokers who actively seek them out, resulting in participation rates of 1–5% of the target population. Evidence suggests that smokers are receptive to mass-mailed messages regarding cessation,14–16 and a strategy of proactive recruitment, where smokers are contacted directly and offered a service, can result in a more demographically representative sample and higher participation rates of 15–35%.7,17,18 Proactive targeting of the entire population of smokers, offering computer-tailored materials that combine the individualized behavioural intervention principles used in clinics with the participation rates of public health campaigns, has the potential to recruit smokers with low motivation to quit and those who prefer not to participate in traditional clinic-based research.
Primary care is a useful setting for a public health-type intervention, where identification of smokers from clinical records and delivery of a version of tailored feedback to a large population is possible. The aim of this pilot study was to assess the feasibility of this approach, with the aim of informing a large-scale trial of effectiveness. The objectives were to assess the willingness of practices to participate in the research, to assess the logistics of the procedure of identifying smokers, to define the components of standard care applied in general practice and to assess the response rates to an assessment using a proactive recruitment strategy.
|
Method |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionDeclarationReferences |
|---|
Tailored feedback
A computer-based system for generating individually tailored feedback reports, developed as a supplement to enhance the effectiveness of telephone counselling,19 was adapted for use in general practice. The content of the tailored feedback uses specific concepts from different theoretical models that have been shown to be relevant to behaviour change, including social cognitive theory20 and the perspectives on change model.21 It aims to change the cognitive determinants of smoking and smoking cessation, by addressing beliefs and expectations, to reduce attitudinal ambivalence towards quitting and to enhance perceived self-efficacy. In addition to this strong theoretical base, it offers information to encourage individuals to bring about the desired cognitive states or behaviour for successful quitting. The tailored feedback reports were also developed in consultation with smoking cessation counsellors and include conventional wisdom, e.g. the importance of setting a quit date. Furthermore, by modifying the questionnaire and tailoring the style, layout and content to suit different educational levels, and by tailoring to include information more appropriate for smokers who are less motivated to quit, we were able to target smokers of all reading levels and of all states of readiness.
Selection of practices
We aimed to recruit four practices in Camden and Islington to represent both large (>7000 list size) and small (<4000 list size), and located in both high and low socio-economic areas based on the Townsend index score. Initially, 12 practices were selected to meet these criteria and approached by letter to ask if they were willing to participate in the trial. Meetings were held with practice staff of those practices agreeing to participate to discuss and determine the logistics of the procedure for identifying smokers and distributing the initial questionnaires.
Selection of patients
Smokers between 18 and 65 were identified by practice staff from records, and a random sample of approximately 200 from each practice selected using Egton Medical Information Systems (EMIS) computer systems. GPs checked the list for suitability for inclusion, excluding people severely or terminally ill or with severe mental impairment (total excluded = 12), and an assessment [the Smoking Behaviour Questionnaire (SBQ)] sent, together with a covering letter from the GP and a Freepost envelope.
Smokers willing to be contacted for future assessment returned the completed questionnaire and signed consent form to the research team. Participants were randomized to receive a standard information booklet sent by post and the standard care offered by their general practice, or to receive a computer-tailored feedback report based on the information obtained at baseline, also sent by post, in addition to the standard information booklet and standard care offered by their practice. The general practices represent a stratifying variable, and so, to reduce any confounding by practice, we used block randomization within each practice to allocate patients to each condition, using a list of random numbers generated by SPSS, to closely balance the numbers in each group. Data from completed assessments were entered and stored in a Microsoft Excel file. A computer programme temporarily combined participants' personal data with the data file and the correct messages from the message file written in Microsoft Word to produce a three-page feedback report for those participants allocated to the feedback condition. Follow-up questionnaires were sent by post to all participants 3 months after the return of the baseline questionnaire, with one reminder to non-responders.
Measures
Variables upon which the feedback letter was tailored were assessed in the SBQ at baseline and included gender, age, current smoking status, intention to quit, dependence in terms of cigarette consumption and time from waking to first cigarette, previous quit attempts, motivation and determination to quit, self-image, reasons for quitting, advantages and disadvantages of quitting, self-efficacy and social environment. Literacy level was assessed by the amount and nature of daily reading, and highest qualification. Demographic information was also gathered so that the sample could be fully described.
The primary outcome measure was self-reported, prolonged abstinence for at least 1 month. Secondary outcome measures were 7-day and 24-hour point prevalence abstinence, quit attempts and cessation methods used [i.e. use of nicotine replacement therapy (NRT), individual counselling or attendance at smoking cessation groups or clinics offered by the practice or elsewhere]. We also included process measures: recall of materials received and perception of the feedback reports. Interviews were carried out with practice nurses, after the initial response and randomization, to define the components of standard care applied in general practice (i.e. amount and availability of care offered by the practice in the form of individual advice or clinics).
Data analysis
All outcome analyses were conducted on an intention-to-treat basis (assuming those lost to follow-up were smoking). Chi-square tests were used for comparison of binary outcomes between the two groups, and independent t-tests for outcome variables that could be treated as continuous. Univariate analyses were carried out to explore possible predictors of abstinence (e.g. demographic variables including practice, dependence, deprivation, literacy level and motivation).
|
Results |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionDeclarationReferences |
|---|
Response
Five practices agreed to take part, although one of these later dropped out due to problems of communication. Five declined to take part for reasons of time and capacity, but indicated that they would be glad to consider such research in the future. One practice objected to the topic of the research and one to the method of recruitment.
Participating practices ranged in list size from 6400 to 15 979, and identified between 8% and 23% of their patients aged between 18 and 65 as smokers. An overall response rate of 8.9% was heavily influenced by the low response of 4.8% in one of the practices. Townsend index scores, a measure of deprivation, were matched to the postcodes of all those sent an assessment questionnaire. The mean was calculated for each practice, and an ANOVA revealed a significant difference between the practices. A high deprivation score within a practice tended to be associated with a high proportion of smokers (Table 1). There was no significant difference in mean Townsend index score between the responders and non-responders (4.04/4.21).
|
Participant characteristics and attrition
Over half of the participants were highly qualified with a degree or postgraduate qualification, and 71.8% were in non-manual occupations. Most were daily smokers, the mean dependency score of participants was 3.92 (scale of 0–7), and a large proportion of the sample were either planning to quit in the next 6 months, but not in the next 30 days (42.3%), or not planning to quit at all (32.8%). Groups were equivalent on all baseline variables (Table 2).
|
Fifty-two (66.7%) participants returned the follow-up questionnaire, 36 (46.1%) on receipt of the questionnaire and a further 16 (20.5%) after one reminder. There was no difference in attrition between the conditions. Females, and those who were not planning to quit because they wanted to smoke, were significantly more likely to return the follow-up questionnaire (
2 = 3.714, P < 0.05; 2 = 6.406, P < 0.05, respectively).
Outcome
There were no significant differences between the Intervention and Control groups (Table 3). The sample was too small to draw conclusions, but the figures seem to indicate that abstinence and quit attempts are not differentiated by intention to quit at baseline (Table 4).
|
|
Predictors
There were no significant differences in outcome between practices, by demographic variables, deprivation or dependency or by literacy level. Point prevalent 24-hour abstinence was predicted by motivation (4.67/3.49 score on 1–5 scale, t = 5.307, P < 0.0001) and determination to quit (4.38/3.34, t = 3.337, P < 0.005).
Use of resources
In total, 13 participants reported seeking advice from their GP, practice nurse, other health professional or service, and nine reported using some form of pharmacotherapy. Six participants did both.
Process measures
Almost all participants remembered receiving the booklet. Of those who returned the follow-up, significantly more in the Intervention group remembered receiving the tailored advice letter [12 (46.2%)/5 (20%)/P < 0.05]. Of six participants in the Intervention group who commented on the tailored letter in open questions, three gave very positive comments, reinforcing the effect of addressing personal issues, and three were more reserved, suggesting that the method can be attractive to smokers but the content could be more compelling.
Standard care and practice policy
All of the practices reported being active in smoking cessation, and have a protocol for identifying, recording and referring smokers, recommending that they make an appointment to see an advisor in the practice, usually the practice nurse, but making the appointment is left to the patient. Patients are first seen for 20–30 minutes, with 10–15 minute follow-up appointments, for a minimum of 3 months, although appointments become less frequent. The majority of patients use NRT, and use of Zyban varies between the practices. Other resources such as information leaflets are used in all practices, but idiosyncrasies occur according to the preferences of the health professional. For example, some recommend popular reading, and one practice nurse offers complementary therapy in the form of acupuncture, taken up by 40% of smokers seeking advice.
|
Discussion |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionDeclarationReferences |
|---|
This pilot work enabled us to explore the feasibility of engaging the public in a mass mailing of computer-tailored self-help for smoking cessation, and to consider the implications of the findings for the design and procedure of a large-scale trial. We assessed the willingness of practices to participate in a proactive method of recruitment, the logistics of the procedure of identifying smokers and distributing initial questionnaires and the response rates to this method of recruitment, as well as the perception by participants of the questionnaire and feedback.
Overall, the response from practices was encouraging. The practices approached were ready to consider the project, indicating that smoking cessation activity was a high priority. Those that refused indicated their willingness to consider future involvement with such research. Unfortunately, smaller practices were more likely to refuse for reasons of capacity and time; therefore, the objective of including smaller practices was not wholly met.
The identification and selection of smokers using EMIS computer systems was quick and straightforward. However, identification of comparable groups of smokers was difficult as the use of codes within EMIS varies enormously and practices also regularly change and update the codes used, to include, e.g. patients who had received a referral or a prompt to make an appointment and those who had made an appointment and received advice. Exclusion of these different groups was difficult, and resulted in the inclusion in the samples of varying proportions of smokers who had already received multiple prompts and referrals, which may have influenced the response rate in some practices. More standardization of coding would be necessary to implement this as a standard procedure in practices.
This method of recruitment is intended to reach all smokers regardless of their interest in and readiness to quit. Volunteers and enrollers to clinics traditionally tend to be from advantaged educational and social backgrounds,7 and recruitment can be related to readiness to quit.14,22 Nevertheless, deliberate targeting of more deprived and low socio-economic areas allowed recruitment of smokers from these groups, approximately one-quarter of respondents were of low literacy level. In addition, three-quarters of respondents were not planning to quit in the near future; thus, the method was successful in terms of engaging these smokers in an activity requiring them to become more aware of their smoking and their reasons for smoking. Furthermore, these apparently unmotivated quitters were also more likely to return the follow-up questionnaire. It is also important that light smokers and occasional smokers be encouraged to quit. As clinic enrollers and help-seekers are more dependent,23 this recruitment method was able to recruit an important subgroup, not normally targeted. The response rate of approximately 10% was lower than expected, but could be increased by the use of other strategies, e.g. a reminder and more attractive printed questionnaires and information leaflets7,24 not used in this pilot study for reasons of economy.
In view of the pattern of response, one would expect the overall level of commitment to be less than that of volunteer clinic samples and therefore would expect the quit rates to be lower.4 However, the focus with such groups is on motivating smokers to engage in quitting activity, to increase the perception and awareness of smoking as a problem and to increase the likelihood of taking action. This study suggested that smokers with low motivation to quit can be persuaded to make a quit attempt, as attempts were made by those who were planning to quit in the next 6 months as well as those who said they were planning to quit in the next 2 weeks or 30 days. While there is an apparent negative outcome for short-term 24-hour point prevalent abstinence, the small sample size limits the validity of the outcome analysis, and the wide confidence intervals do not rule out a possible positive intervention effect. However, the number of 24-hour point prevalent quitters also suggests that the receipt of the questionnaire alone can prompt a quit attempt regardless of the intervention. With only four practices, it was not possible to estimate the intra-practice correlation for these binary outcomes. While we cannot rule out a modest amount of intra-practice correlation, we think it is unlikely to be detectable even in a study involving large numbers of practices, since randomization is at an individual rather than at a practice level.
One might expect that the letter would yield more enrollments in clinical therapy, but few respondents used these resources, the majority using no aid at all, which indicates the need to emphasize and encourage the use of support services for behavioural therapy and the use of pharmacotherapy in addition to the behavioural strategies suggested.
Implications for the design and procedure of a large-scale randomized controlled trial
This pilot study has demonstrated the feasibility and acceptability of delivering computer-tailored feedback in primary care. The study was designed to assess the feasibility and was not powered to compare cessation between the two conditions. Thus, a large-scale trial is needed to evaluate the effectiveness and cost-effectiveness of this intervention. In considering the design of such a trial, certain issues should be addressed. While willingness to take part was demonstrated, the selection of practices needs to ensure that sufficient low socio-economic areas are targeted to increase the number of participants from these areas. Performance measures (Quality Outcomes Framework) now ensure that most practices have a high rate of recording of patient's smoking status, and standardized coding of patient's engagement in the quitting process would enable the identification procedure to be more efficient, targeting smokers who are most likely to benefit from a brief intervention. Simplifying and standardizing the procedure would also enable more small practices to participate, thus ensuring a more representative sample of practices, and justifying randomization at individual rather than practice level. The ease of completion, and acceptability of the questionnaire has been demonstrated,19 but more emphasis should be placed on methods to encourage greater response by, e.g. optimizing the balance between the attractiveness of the materials and the use of strategies to give a professional appearance.
In conclusion, computer-tailored feedback, adapted to reading levels and readiness to quit, is a simple intervention that can be widely implemented and delivered to a large proportion of the smoking population. The approach is designed to make minimal demands on the time of practice staff, in contrast to other approaches that involve practice staff in the delivery of the intervention. A modest success rate could have a large effect on public health given its recruitment potential, making a valuable contribution to lowering smoking prevalence, while offering an efficient method of integrating smoking cessation counselling into a busy primary care practice.
|
Declaration |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionDeclarationReferences |
|---|
Funding: The North Central London Research Consortium (NoCLoR).
Conflicts of interest: None.
|
Acknowledgments |
|---|
This research was reviewed by the Camden & Islington Community Health Services Local Research Ethics Committee.
|
Notes |
|---|
Gilbert H, Nazareth I and Sutton S. Assessing the feasibility of proactive recruitment of smokers to an intervention in general practice for smoking cessation using computer-tailored feedback reports. Family Practice 2007; 24: 395–400.
|
References |
|---|
|
TopAbstractBackgroundMethodResultsDiscussionDeclarationReferences |
|---|
1 Bauld L, Chesterman J, Judge K, Pound E, Coleman T. On behalf of the English Evaluation of Smoking Cessation Services. Impact of UK National Health Service smoking services: variations in outcomes in England. Tob Control (2003) 12:296–301.
2 Cokkinides VE, Ward E, Jemal A, Thun M. Under-use of smoking-cessation treatments: results form the National Health Interview Survey, 2000. Am J Prev Med (2005) 28:119–122.[CrossRef][ISI][Medline]
3 Brown S, Owen N. Population versus clinical perspectives on smoking behaviour. Behav Change (1990) 7:120–125.
4 Lando HA, Pirie PL, McGovern PG, Pechacek TF, Swim J. A comparison of self-help approaches to smoking cessation. Addict Behav (1991) 16:183–193.[CrossRef][ISI][Medline]
5 Fiore MC, Novotny TE, Pierce JP, Giovino GA, Hatziandreu EJ. Methods used to quit smoking in the United States. Do cessation programs help? JAMA (1990) 263:2760–2765.[Abstract]
6 Glynn TJ, Boyd GM, Gruman JC. Essential elements of self-help/minimal intervention strategies for smoking cessation. Health Educ Q (1990) 17:329–345.[ISI][Medline]
7 Schmid TL, Jeffery RW, Hellerstedt WL. Direct mail recruitment to home-based smoking and weight control programs: a comparison of strategies. Prev Med (1989) 18:503–517.[CrossRef][ISI][Medline]
8 Etter J-F, Perneger TV, Ronchi A. Distribution of smokers by stage: international comparison and association with smoking prevalence. Prev Med (1997) 26:580–585.[CrossRef][ISI][Medline]
9 Wewers ME, Stillman FA, Hartman AM, Shopland DR. Distribution of daily smokers by stage of change: current population survey results. Prev Med (2003) 26:710–720.
10 Kreuter M, Farrel D, Olevitch L, Brennan L. Tailoring Health Messages: Customizing Communication with Computer Technology (2000) Mahwah, NJ: Lawrence Earlbaum.
11 Lancaster T, Stead LF. Self-help interventions for smoking cessation. Cochrane Database of Syst Rev (2005) doi: 10.1002/14651858.CD001118.pub2.
12 Jarvis MJ. , , eds. Patterns and predictors of smoking cessation in the general population. In Bolliger CT, Fagerstrom KO (eds). The Tobacco Epidemic Progr Respir Res, Basel: Karger, 1997; 28: 151–164.
13 Meade CD, Byrd JC. Patient literacy and the readability of smoking education literature. Am J Public Health (1989) 79:204–206.
14 Cunningham JA, Ferrence R, Cohen J, Adlaf EM. Interest in self-help materials among a general population sample of smokers. Addict Behav (2003) 28:811–816.[CrossRef][ISI][Medline]
15 Fiore MC, McCarthy DE, Jackson TC, et al. Integrating smoking cessation treatment into primary care: an effectiveness study. Prev Med (2004) 38:412–420.[CrossRef][ISI][Medline]
16 Paul CL, Wiggers J, Daly JB, Green S, Walsh RA. Direct telemarketing of smoking cessation interventions: will smokers take the call? Addiction (2004) 99:907–913.[CrossRef][ISI][Medline]
17 Lennox AS, Osman LM, Reiter E, et al. Cost effectiveness of computer tailored and non-tailored smoking cessation letters in general practice: randomised controlled trial. BMJ (2001) 322:1396.
18 Etter J-F, Perneger TV. Effectiveness of a computer-tailored smoking cessation program: a randomized trial. Arch Intern Med (2001) 161:2596–2601.
19 Sutton S, Gilbert H. Effectiveness of individually-tailored smoking cessation advice letters as an adjunct to telephone counselling and generic self-help materials: a randomised controlled trial. Addiction (2007) 107:994–1000.
20 Bandura A. Social Foundations of Thought and Action: A Social Cognitive Theory (1986) New York: Prentice-Hall.
21 Borland R, Balmford J, Hunt D. The effectiveness of personally tailored computer-generated advice letters for smoking cessation. Addiction (2004) 99:369–377.[CrossRef][ISI][Medline]
22 Lichtenstein E, Hollis J. Patient referral to a smoking cessation program: who follows through? J Fam Pract (1992) 34:739–744.[ISI][Medline]
23 Gilbert H, Sutton S, Sutherland G. Who calls QUIT? The characteristics of smokers seeking advice via a telephone helpline compared with those attending a clinic and with smokers in the general population. Public Health (2005) 119:933–939.[CrossRef][ISI][Medline]
24 Edwards P, Robert I, Clarke M, et al. Increasing response rates to postal questionnaires: systematic review. BMJ (2002) 324:1183–1211.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||