2014 Volume 56 Issue 1 Pages 21-27
Objectives: A better identification of the determinants of smoking relapse among hospital workers would be helpful in development of more effective interventions to decrease the frequency of relapses in this group. The aim of this study was to determine the predisposing enabling, and reinforcing factors associated with smoking relapse among workers at a university hospital. Methods: This was a case-control study based on a self-administered and structured questionnaire. Cases were all those workers who had relapsed after at least 6 months without smoking, and controls were ex-smokers without relapse for more than 6 months. We obtained the following information: sociode- mographic and tobacco consumption characteristics and a list of predisposing, enabling and reinforcing factors.Results: There were 342 respondents: 114 cases and 228 controls. The variables significantly and independently associated with increased risk of relapse were smoking is my vice (OR=4.02), I’ll be able to quit smoking whenever I want (OR=3.43), I have no intention to quit forever (OR=6.02), celebrations (OR=3.93) and weight gain (OR=10.61), while variables associated with lower risk were age (OR=0.88), health-care worker (OR=0.13), years of abstinence (OR=0.91), smoking is a useless habit (OR=0.19) and illness related to tobacco (OR=0.07).Conclusions: Health programs against smoking in the hospital setting should include measures aimed at preventing relapse through behavioral support therapies and dietary control with particular attention to changes in factors related to lifestyle and false beliefs (predisposing factors).
(J Occup Health 2014; 56: 21–27)
The attitudes of health-care workers regarding smoking have an important influence on the general population, as this group performs supporting, educator and model roles in society1–3). Ex-smokers report that one of the most important reasons to quit smoking is medical advice, which directly depends the health-care worker's attitude toward tobacco4, 5). However, the prevalence of smoking in health-care workers remains high4, 6–8), in hospitals higher than primary care centers9). Reduction of the number of smokers in this particular group would be a model of healthy lifestyle for the general population, and would have a positive influence when advising patients to quit smoking, making the advice much more powerful, believable and effective.
Quitting smoking definitively is not easy due to the frequency of relapses. Approximately, 75% of smokers attempting to quit smoking will have a relapse within the first 4 weeks, and most will relapse in the first six months10). It is estimated that 37% will have a relapse during the 10 years after a year of tobacco abstinence11). According to a meta-analysis study12), the annual incidence of relapse is 10% after the first year of abstinence.
Although it seems that pharmacological therapies (varenicline, bupropion, nicotine) could reduce the smoking relapse probability13, 14), there is insufficient evidence at the moment to support the use of any specific behavioral component or intervention for helping smokers who have successfully quit to avoid relapsing and smoking again14, 15). Some authors have suggested that until new positive evidence becomes availabel, it may be more efficient to focus resources on supporting initial cessation attempts rather than on extended relapse prevention interventions16).
A better identification of the determinant factors of smoking relapse among hospital workers would help to develop more effective prevention programs and consequently to decrease the prevalence of smokers in this group. The PRECEDE model17, 18) can provide an adequate framework for this identification. This model proposes that health behaviors are influenced by predisposing factors (characteristics that motivate and lead to a behavior, including knowledge and beliefs), enabling factors (characteristics that facilitate or they are necessary to carry out the behavior, such as personal aptitudes or environmental resources) and reinforcing factors (rewards or punishments).
In the present article, we examine the differences, including predisposing, reinforcing, and enabling factors, between those workers who relapsed in smoking and those who did not relapse at our institution.
We performed a case-control study based on the application of a structured questionnaire in a consecutive sample of workers seen at the Department of Preventive Medicine, University Hospital “Dr. Peset”, in Valencia (Spain) all through 2011, including nurses, physicians, technicians, orderlies and administrative staff. This institution is a 529-bed, tertiary care, public, academic medical center with approximately 2,100 workers. According to our records (data not published), 32% of the hospital workers at this institution were smokers and 29% were ex-smokers. The questionnaire included the following information: age, sex, university education (yes/no), occupation (healthcare worker/not health-care worker), working duration (years), morning shift (yes/no), type of contract (permanent / temporary), current relation with tobacco (smoker/ex-smoker), smoking duration (years), age of smoking onset, number of cigarettes a day, smoking burden (year-packs), abstinence duration (years), Fagerström score before quitting, and the frequency (never or rarely / quite a lot or high frequency) the last time they attempted to quit smoking of each of the 52 predisposing, enabling and reinforcing factors identified in a previous study19).
The case and control definitions were adapted from the World Health Organization classification20) All workers that returned to smoking after a six-month period of abstinence were considered cases. Controls were all ex-smoking workers with at least 6 months without relapsing, which means they were regular smokers and had gone more than 6 months without smoking at the time of the survey.
Since there was no main predictor variable, the following were assumed: an expected odds ratio of 2, 50% exposed cases and 2 controls per case. Therefore, at least 111 cases and 222 controls were required in order to get a statistical power of 80% with a significance level of 5%. All workers who went to the Preventive Medicine Service for any reason (such as for vaccinations, checkups, and biological exposures) and who met the case or control criteria were consecutively recruited to achieve the required sample size. Respondents participated freely and without any particular reason in the study, filling out the questionnaires themselves anonymously (without identifying data of the worker). This work received prior approval from the Research Committee of the institution where the study was carried out.
Regarding statistical analysis, qualitative variables were described using absolute and relative frequencies (percentages), and quantitative variables were described with means and standard deviations. The statistical significance of comparisons was assessed through the Pearson x2 test for proportions and Student's t-test for means. Odds ratios (OR) with 95% confidence intervals (CI) were also calculated. Significance was set at p<0.05.
We performed a multivariate logistic regression analysis using the forward stepwise method to select variables on the basis of the likelihood ratio statistic for assessment of the association of the independent variables with the probability of relapse. Variables that were statistically significant in bivariate comparisons were considered for inclusion into the model. The adjusted OR and 95% CI were estimated.
The final respondents were 342 workers: 114 cases and 228 controls. Table 1 shows the sociodemo- graphic and tobacco consumption characteristics of cases compared with controls. The proportion of men among cases was significantly higher than in controls. The mean age, working duration, age of smoking onset and years of abstinence were significantly lower in cases, who, on the other hand, had higher Fagerström test scores than controls. Likewise, controls had a higher frequency of morning shift, university education and health-care worker job. However, there were no significant differences regarding type of contract, smoking duration, number of cigarettes and tobacco burden.
The numerous predisposing factors that were significantly associated with smoking relapse according to the questionnaire responses are shown in Table 2, which highlights that some beliefs, such as I have no intention to quit forever; it's just one, so nothing bad is going to happen; and if I smoke one, I can control it, had high odds ratios. On the other hand, smoking damages your health and I have smoked for a long time were not associated with relapse.
Regarding the enabling factors associated with relapse (Table 3), the following circumstances were remarkable: living with smokers, being offered tobacco, celebrations and social gatherings. Nevertheless, smoking ban, family pressure, poor social perception of smoking and playing sports did not influence relapse.
| Controls | Cases | Statistic* | p value | |
|---|---|---|---|---|
| Male sex,n (%) | 42 (18.3) | 32 (28.1) | 4.25 | 0.039 |
| Age,mean (SD) | 45.33 (9.9) | 38.39 (8.9) | 6.54 | <0.001 |
| Working duration, mean (SD) years | 13.06 (11.7) | 7.69 (9.1) | 4.65 | <0.001 |
| Smoking duration, mean (SD) years | 19.31 (9.9) | 19.69 (9.8) | -0.33 | 0.741 |
| Starting age,mean (SD) | 16.62 (3.9) | 15.76 (3.1) | 2.18 | 0.029 |
| Abstinence duration, mean (SD) years | 9.39 (8.0) | 2.94 (4.4) | 9.62 | <0.001 |
| Burden (year-packs),mean (SD) | 18.13 (15.2) | 19.76 (16.1) | -0.91 | 0.361 |
| Number of cigarettes, mean (SD) | 17.91 (10.5) | 18.63 (10.2) | -0.60 | 0.548 |
| Fagerstrom test score,mean (SD) | 3.39 (2.6) | 4.72 (3.0) | -3.95 | <0.001 |
| Permanent contract, n (%) | 94 (41.0) | 36 (31.6) | 2.90 | 0.089 |
| Morning shift,n (%) | 120 (52.4) | 41 (35.9) | 8.25 | 0.004 |
| University education, n (%) | 144 (63.1) | 40 (35.1) | 24.09 | <0.001 |
| Health-care worker,n (%) | 169 (74.1) | 53 (48.6) | 21.33 | <0.001 |
| Controls n (%) | Cases n (%) | OR | 95%CI | |
|---|---|---|---|---|
| I have no intention to quit forever. | 30 (13.1) | 61 (53.5) | 7.63 | 4.34,13.48 |
| It's just one,so nothing bad is going to happen. | 51 (22.3) | 66 (57.9) | 4.80 | 2.87,8.03 |
| The psychological dependence is so strong. | 124 (54.4) | 96 (84.2) | 4.47 | 2.46,8.22 |
| If I smoke one,I can control it. | 101 (45.5) | 89 (78.1) | 4.26 | 2.47,7.40 |
| Smoking is my vice. | 76 (33.6) | 75 (65.8) | 3.80 | 2.30,6.28 |
| I need it. | 45 (20.1) | 53 (46.5) | 3.46 | 2.05,5.83 |
| I can quit it again whenever I want. | 53 (23.9) | 59 (51.7) | 3.40 | 2.05,5.66 |
| I miss it. | 46 (20.1) | 49 (45.4) | 3.30 | 1.95,5.61 |
| I am going to relapse because I did it before. | 42 (18.5) | 42 (36.8) | 2.57 | 1.50,4.40 |
| Being a social smoker causes no damage. | 82 (35.9) | 67 (58.8) | 2.54 | 1.56,4.13 |
| Greedy personality with relapse possibility | 32 (14.0) | 36 (31.6) | 2.83 | 1.59,5.05 |
| If I bother others,I don’t like to smoke. | 138 (61.9) | 86 (76.1) | 1.96 | 1.14,3.38 |
| A smoker can never change. | 44 (19.2) | 36 (31.6) | 1.94 | 1.12,3.35 |
| I have smoked for a long time. | 141 (62.1) | 80 (70.2) | 1.44 | 0.86,2.39 |
| Smoking damages your health. | 145 (63.9) | 69 (60.5) | 0.87 | 0.53,1.41 |
| I do not have a need to smoke. | 135 (59.5) | 53 (46.5) | 0.59 | 0.37,0.96 |
| I don’t want to be dependent. | 138 (60.8) | 51 (44.7) | 0.52 | 0.32,0.84 |
| Smoking is a disgusting habit. | 145 (63.6) | 52 (45.6) | 0.48 | 0.30,0.78 |
| Smoking is a useless habit. | 184 (81.0) | 62 (54.4) | 0.28 | 0.16,0.47 |
OR, odds ratio; CI, confidence interval.
Despondency sensation, weight gain and liking the taste and smell of tobacco were the reinforcing factors that were positively associated with a probability of relapse with higher odds ratios, while suffering from an illness related to tobacco was negatively associated with relapse in a statistically significant way (Table 4). The feeling of escape produced by tobacco, the reinforcement of self-esteem from not smoking, and considering tobacco to represent a waste were not associated with relapse.
| Controls n (%) | Cases n (%) | OR | 95%CI | |
|---|---|---|---|---|
| Smoking family members at home | ||||
| 1 member | 52 (22.7) | 34 (29.8) | 2.73 | 1.49,5.01 |
| 2 or more members | 31 (13.5) | 45 (39.5) | 6.06 | 3.23,11.4 |
| In social reunions, I really want to smoke. | 35 (15.4) | 53 (46.5) | 4.77 | 2.76,8.25 |
| Somebody offers me tobacco. | 28 (12.2) | 44 (38.9) | 4.58 | 2.56,8.21 |
| Celebrations | 42 (18.6) | 57 (50.0) | 4.38 | 2.59,7.43 |
| I am exposed to smokers. | 72 (31.4) | 75 (65.8) | 4.19 | 2.53,6.96 |
| Stress makes me smoke. | 41 (17.9) | 49 (42.9) | 3.44 | 2.02,5.86 |
| During exams,I want to smoke. | 30 (13.6) | 36 (33.0) | 3.12 | 1.73,5.65 |
| Some friends are smokers | 177 (77.3) | 101 (88.6) | 2.28 | 1.14,4.64 |
| Smokers at work | 164 (71.6) | 98 (85.9) | 2.43 | 1.28,4.64 |
| Drinking coffee makes me want to smoke. | 47 (20.6) | 49 (42.9) | 2.90 | 1.73,4.89 |
| I would like to smoke at work. | 35 (16.1) | 39 (34.8) | 2.78 | 1.58,4.89 |
| My nonsmoking partner helps me not to smoke. | 118 (58.7) | 78 (72.9) | 1.89 | 1.10,3.26 |
| When I drink,I would like to smoke. | 47 (20.6) | 34 (32.1) | 1.82 | 1.05,3.16 |
| Smoking ban at work | 121 (55.7) | 58 (50.9) | 0.82 | 0.51,1.33 |
| No smoking laws | 114 (53.0) | 53 (46.9) | 0.78 | 0.48,1.27 |
| My family pressures me to quit smoking. | 114 (51.1) | 49 (42.9) | 0.72 | 0.45,1.16 |
| There is a bad perception of smokers. | 133 (58.3) | 57 (50.0) | 0.71 | 0.44,1.15 |
| Practicing a sport helps me to quit smoking. | 128 (59.5) | 55 (48.2) | 0.68 | 0.42,1.11 |
| Other people from my environment are quitting. | 141 (62.9) | 57 (50.0) | 0.59 | 0.36,0.95 |
OR, odds ratio; CI, confidence interval.
| Controls n (%) | Cases n (%) | OR | 95%CI | |
|---|---|---|---|---|
| Despondency sensation without tobacco | 16 (6.9) | 47 (41.2) | 9.34 | 4.78,18.46 |
| I like the taste of tobacco. | 15 (6.8) | 39 (34.2) | 7.07 | 3.53,14.32 |
| Weight gain | 28 (12.9) | 58 (50.8) | 6.99 | 3.94,12.47 |
| I like the smell of tobacco. | 17 (7.7) | 31 (27.2) | 4.46 | 2.24,8.94 |
| Smoking is my reward. | 15 (6.7) | 20 (17.5) | 2.95 | 1.37,6.38 |
| Smoking relaxes me. | 58 (25.3) | 50 (43.9) | 2.30 | 1.39,3.81 |
| I like smoking. | 84 (37.0) | 60 (52.6) | 1.89 | 1.17,3.06 |
| Smoking helps me to escape. | 37 (16.2) | 27 (23.7) | 1.60 | 0.88,2.90 |
| Smoking helps me to concentrate. | 48 (20.9) | 27 (23.7) | 1.17 | 0.66,2.07 |
| Not smoking reinforces my self-esteem. | 120 (53.1) | 52 (45.6) | 0.74 | 0.46,1.19 |
| Tobacco represents a waste. | 176 (80.0) | 82 (71.9) | 0.64 | 0.37,1.12 |
| Relatives died because of tobacco. | 70 (30.6) | 25 (21.9) | 0.64 | 0.36,1.11 |
| I think about the example that I am giving to my kids. | 142 (62.0) | 51 (44.7) | 0.38 | 0.22,0.66 |
| Illness related to tobacco | 52 (22.7) | 5 (4.4) | 0.16 | 0.05, 0.41 |
OR, odds ratio; CI, confidence interval.
| ORa | 95% CI | p-value | |
|---|---|---|---|
| Weight gain | 10.61 | 3.27,34.42 | <0.001 |
| I have no intention to quit forever. | 6.02 | 1.82,19.91 | 0.003 |
| Smoking is my vice. | 4.02 | 1.46,11.03 | 0.006 |
| Celebrations | 3.93 | 1.46,10.53 | 0.006 |
| I’ll be able to quit whenever I want. | 3.43 | 1.36,8.63 | 0.008 |
| Years of abstinence | 0.91 | 0.84,0.99 | 0.034 |
| Age | 0.88 | 0.83,0.93 | <0.001 |
| Smoking is a useless habit. | 0.19 | 0.07,0.51 | 0.001 |
| Health-care worker | 0.13 | 0.05,0.35 | <0.001 |
| Illness related to tobacco | 0.08 | 0.01,0.37 | 0.001 |
ORa, adjusted odds ratio; CI, confidence interval.
The results of multivariate logistic regression analysis (Table 5) showed that the factors positively associated with relapse independently were smoking is my vice, I’ll be able to quit whenever I want, I have no intention to quit forever, celebrations and weight gain. In contrast, the factors negatively associated were age, health-care worker, years of abstinence, the belief that smoking is useless, and illness related to tobacco.
Due to the convenience of working with groups of high influence, such as health-care professionals, it is important to plan strategies or interventions to reduce the relapse rate for tobacco consumption among health-care workers. The first step in designing these strategies must be based on a behavioral diagnosis as established by the PRECEDE model, in order to develop interventions adapted to the particular needs of this group that are more effective, and to provide specific advice according to individual characteristics.
The current study deals with phase 4 of the PRECEDE model (educational and organizational diagnosis) and it is the first study carried out using this model to evaluate smoking relapse reasons, related to predisposing, enabling and reinforcing factors in hospital workers. Incorporating the PRECEDE model probably has allowed us to consider all the important reasons for smoking relapse. A comprehensive analysis of the determinants of relapse was performed, by comparing these factors between a group of hospital workers who had relapsed in smoking and others who had not relapsed. The study data were self-reported, so the possibility of recall bias and misclassification cannot be ruled out. Nonetheless, error in reporting was likely minimal, because the subjects were not asked to provide complex details but rather were asked to answer “yes” or “no” to simple questions from a structured questionnaire.
According to the results obtained in our study, a low score for the Fagerström test and late age at onset of smoking are factors that prevent relapse, in keeping with the results of another study21) conducted in the general population. Furthermore, our findings about years of abstinence, age and education level being associated with a lower risk of relapse are consistent with those of another paper11).
We identified false beliefs, smoking environment, stressful events, celebrations and social gatherings as main predisposing and enabling factors associated with smoking relapse, in accordance with results from other studies22–24). This would correspond to the social learning model25), which suggests a strong association between behavior and environment. Additionally the relapses related to stressful situations may be due to a lack of skills26, 27). There are high-risk situations where relapse is most likely to occur and where the way in which the person faces such a situation plays an essential role.
Various studies28–31) have identified some predisposing and precipitating factors focused on the relapse process in the general population. They described predisposing factors such as age, occupation, socioeconomic status, coping skills, moods, attitudes, health, smoking duration, previous quit attempts and withdrawal symptoms, while the precipitating factors were environmental factors and high-risk or stressful situations. The predisposing and precipitating factors of these studies are similar to those predisposing and enabling factors of the PRECEDE model obtained in our work. Moreover, it is important to highlight weight gain as the main reinforcing factor on which we could act and thereby reduce its influence.
However, there are still some major questions about the factors that have been analyzed in our study; these include their roles in this phase of the PRECEDE model and the interrelation among them and this phase with other phases of the model, which should be addressed in further research in the future. It would also be necessary to perform an intervention study in order to test the effectiveness of a relapse prevention program that has influence over all the modifiable important factors identified in this work, especially through behavioral support therapies and control of diet and lifestyle.
