Study design and measures

A descriptive cross-sectional survey was carried out. The widely used and psychometrically sound instrument ‘Nurses Helping Smokers Quit’ was utilised¹⁸. The original questionnaire had a very good level of reliability (α=0.92)¹⁹. It was adapted, tested for validity by experts in nursing in tobacco control and re-tested for reliability (93% of the kappa values were >0.7)^18,20. In this study, all questions/statements were directed to healthcare professionals instead of nurses, while participants were asked to indicate their profession. Otherwise, the tool was left as originally set:

To ensure its validity for distribution amongst healthcare professionals in Malta, the questionnaire was assessed for face validity by a doctor, a nurse, a pharmacist and an allied healthcare professional, and by two public health consultants. No changes were required.

Study participants

Healthcare professionals, who attended one of ten training sessions on brief interventions for smoking cessation, held every month between September 2018 and June 2019 in Malta, were included as potential study participants. Several strategies (mainly mailshots and advertisements) were used for recruiting healthcare professionals, who were interested in tobacco cessation, to these educational sessions. In all, 150 healthcare professionals attended these training sessions and were provided with the questionnaire, which they filled in before the commencement of the training session.

In order to be eligible to this study, they had to be health workers who provided direct patient care to adult patients (aged ≥18 years) during a typical working day. Respondents who were health workers with no direct patient care, or who cared only for paediatric patients, were excluded.

This study was part of a training project on brief smoking cessation interventions for healthcare professionals in the public service, held by the Health Promotion and Disease Prevention Directorate within the Department for Health Regulation/ Superintendence of Public Health, Ministry for Health, Malta. An information letter was attached to each questionnaire, providing details on the aims of the audit and the training session. It also invited participants to write their email address, which was the sole personal identifier, to be used for longterm evaluation of the training project. Filling in the questionnaire was voluntary and not a requirement to attend the educational session. Participants were informed that the Directorate abided by Article 27 (a)(i) of the Public Health Act²¹, the Data Protection Act²², and the General Data Protection Regulation (EU) 2016/679²³ for the processing of personal data. Return of a completed questionnaire implied consent. All responses were kept confidential and later anonymised through the generation of a unique code identifier. No ethical issues were envisaged.

Statistical analysis

All analyses were done using IBM SPSS Statistics Version 26. Descriptive statistics (frequencies, percentages, means with standard deviations) were used to characterise the sample and study variables. As for previous studies^18,20, we defined consistency of performing tobacco cessation interventions by collapsing ‘always/usually’. Similarly, agreement to positively worded opinions and attitudes about counselling patients was denoted as ‘strongly agree/ agree’, while disagreement to negatively worded opinions and attitudes was defined as ‘strongly disagree/disagree’. Univariate logistic regression was carried out to identify any significant associations in these revised, now binary variables by the independent variables, taking note of within group differences with their respective odds ratio (OR), 95% confidence intervals (CI) and p-values. On the other hand, given that some groups of participants were quite small, non-parametric tests were used to identify differences in the other non-binary dependent variables and for associations by years of experience. The MannWhitney U test was used to assess the relationship between years of experience and the binary variables. With regard to the estimated number of patients helped in the past week, Fisher’s exact test was used; whereas for years of experience the Kruskal Wallis test was used. The Kruskal Wallis test (and subsequently Dunn’s pairwise tests with adjustment using the Bonferroni correction) and the Mann-Whitney U test were also utilised to identify associations for the two linear scale questions by the independent variables, except for years of experience, where the Pearson correlation coefficient was computed. Where significant differences were observed in more than one independent variable, a multivariate analysis was conducted to identify and rule out any confounder. A p≤0.05 was considered as the threshold required for all statistically significant associations.

Participants’ characteristics

All (150) participants completed the questionnaire. After excluding those who did not provide direct patient care to adult patients, 133 healthcare professionals were found to be eligible and were included in data analysis. Participants’ characteristics are displayed in Table 1. Most participants were female nurses, who never smoked, and who worked in hospital settings. Mean years of experience were 14.9 ± 12.7. Only 11 participants had received training in smoking cessation within the past 24 months.

Delivery of tobacco cessation interventions to patients

As displayed in Table 2, whereas most participants consistently (always/usually) asked (76.3%; n=100), advised (83.5%; n=111), and assessed (70.5%; n=93) patients for tobacco cessation, fewer participants assisted (40.9%; n=54) and arranged follow-up (24.2%; n=31). Moreover, few participants routinely carried out the other listed tobacco cessation interventions, except for ‘recommend to patients and family members the importance of creating a smoke-free home environment’ (65.7%; n=86). Most participants (46.6%; n=62) had counselled 1–2 patients over the previous week.

Opinions about counselling patients to quit smoking and attitudes about healthcare professionals and smoking, and their role in tobacco control

Table 3 shows the responses to the 13 opinions about counselling patients to quit. Most respondents agreed (strongly agree/agree) to the first eight positive opinions (ranked in a descending order of agreement), except for ‘asking patients about smoking increases the likelihood that they will quit’, where only 48 participants agreed (37%). Most disagreed (strongly disagree/disagree) to the other negatively worded statements (ranked in a descending order of disagreement) with the exception of ‘it is difficult for me to get people to quit smoking’, and ‘I have insufficient time to counsel patients about quitting smoking’ (18.2%, n=24; and 39.3%, n=51; respectively). Almost all participants (93.1%; n=121) agreed that they need more training to help patients to quit smoking.

As evidenced in Table 4, the majority of the respondents agreed with the attitudes presented or rated them as important. Again, the need for additional training and skills in tobacco control was highlighted by almost all participants (94.8%; n=126).

Differences by gender, smoking status, and professional characteristics

Following the multivariate analyses, where confounders were identified, the variables that remained statistically significant by personal and professional characteristics (Supplementary file, Tables S1–S3), were further analysed for associations within the groups.

Reflection on main findings and highlights from the literature

The inconsistent delivery of all of the 5As was acknowledged in international literature^24-28. A similar trend in the delivery of the 5As was reported in a study amongst hospital health workers in Catalonia, Spain²⁶. Where the acceptable performance of the 5As was defined as a mean value ≥5, professionals reportedly asked (6.4 ± 3.1), advised (7.1 ± 2.7), and assessed (6.3 ± 2.8) patients for cessation, but few provided assistance (4.4 ± 2.9), and arranged follow-up (3.2 ± 3.3)²⁶. Most probably, healthcare professionals tend to ‘assist’ and ‘arrange follow-up’ only to those who are interested in quitting. Nevertheless, guidelines recommend the delivery of all the 5As, irrespective of whether smokers are motivated to quit or not⁶.

Unlike what was reported in some studies^24,26-28, doctors in our study did not outperform the other healthcare professionals in all the 5As. Nonetheless, doctors were more likely to have counselled an increased number of patients, than other professionals during the previous week. Doctors, being very much involved in the assessment and diagnosis of health interactions, may be more likely to tie in smoking cessation advice in their management.

Although females were more likely to report having sufficient time for counselling, they were less active in advice-giving. Similarly, in the Girvalaki et al.¹⁶ study, male general practitioners were significantly more likely to ‘advise’ patients (adjusted odds ratio, AOR=2.88; 95% CI: 1.06– 7.86; p<0.05)¹⁶. In contrast, this difference was not observed in the Martinez et al.²⁶ study. Also, no significant differences in the delivery of brief interventions were noted by years of experience²⁶. However, one must note that unlike our study, almost all participants (80.8%) had previous tobacco cessation training²⁶. Training, which is known to have a positive impact on delivery of the 5As^29-31, could have impacted on the above-mentioned results, eliminating differences by gender and years of experience. Although in our study professionals with more years of experience were more likely to see smoking cessation counselling as an inefficient use of their time, it could be that they had other nonclinical roles such as in management.

Unlike in some international studies^24-26, delivery of the 5As was not found to be significant by smoking status (Supplementary file, Table S1), probably due to the small number of participating smokers. Nonetheless, as was found in the Yan et al.²⁵ study (where mean attitudes were: never smokers 18.2 ± 1.8 versus ever smokers 16.1 ± 2.7; p<0.05), never smokers were more likely to agree to positive statements towards counselling.

When compared to previous studies, where it was also agreed that smoking cessation is an important professional responsibility^24,26, and that healthcare professionals should be role models²⁵, and be involved in tobacco control activities²⁵, our study provided a wider understanding of the participants’ opinions and attitudes towards tobacco cessation. Few healthcare professionals, nurses, and allied healthcare professionals in particular, agreed that ‘asking’ patients increases the likelihood that they will quit. This highlights the participants’ lack of knowledge in this evidence-based practice¹³. Health providers working in residential care settings reported less frequent tobacco cessation interventions and fewer positive opinions and attitudes of such role. Probably they perceived little or no benefits for the elderly clients.

Although most participants tended to agree that they had insufficient time to counsel patients, allied healthcare professionals were less likely to. While this is encouraging, as it suggests that allied healthcare professionals might be in a better position to deliver smoking cessation interventions; lack of sufficient time deserves its due attention. Given the benefits of opportunistic brief smoking cessation advice, it is recommended that healthcare management addresses this perceived organisational barrier.

As was found in international literature^26,32, almost all participants expressed the need for training in smoking cessation. Training healthcare professionals in smoking cessation is more likely to result in positive changes in the delivery of the 5As, such as: asking patients to set a quit date (p<0.0001); providing follow-up appointments (p<0.00001); counselling smokers (p<0.00001); providing selfhelp material (p<0.0001); and prescribing a quit date (p<0.00001)³¹. All health workers, including allied healthcare professionals, should be provided with training opportunities on tobacco cessation interventions. Given the limited time in carrying out cessation practices, training sessions can help equip professionals with the required skills to deliver interventions, which are brief yet effective. Such initiatives can support professionals who feel that it is difficult for them to help smokers quit and provide skills for addressing barriers. Training should also provide an overview of local community services, especially for those professionals who may be less aware of available services. Moreover, in view of the highlighted benefits of opportunistic smoking cessation amongst older adults³³, training initiatives should be also provided to professionals working in residential care settings. In addition, given that former smokers were more confident in helping patients quit, engaging their experiences within any training activities would be an asset.

National clinical guidelines or protocols, which have been found to improve the provision of brief cessation interventions³⁴, can also guide professionals to fully engage in the 5As evidencebased framework. Given that smoking was found to be a risk factor for disease progression of coronavirus disease, COVID-19^35,36, 29.8% of smokers compared with 17.5% of non-smokers experienced disease progression (OR=1.91; 95% CI: 1.42–2.59; p=0.001)³⁶, the rapid increase of which has made the WHO European Region the centre of this pandemic, promoting use of brief tobacco cessation interventions through training efforts and issuing of protocols/guidelines would be of further significance for the current times.

Strengths and limitations

This is the first time that the extensively used tool by Sarna and Bialous et al.¹⁸ has been used amongst a wide array of healthcare professionals, not only providing a detailed description of practices, opinions and attitudes, but also identifying differences by professional characteristics. There are some limitations. Our survey was completed by a small number of healthcare professionals, working in Malta, who attended a training session. These participants may have been more likely to engage in tobacco cessation interventions than those who did not participate, limiting the generalizability of the findings to the total healthcare population. A largerscale study in another European country is thus recommended. Notwithstanding the small sample size, the training sessions were carried out in four different settings, using different recruitment strategies to encourage participation, and including a wide array of participating healthcare professionals (except for pharmacists) to enable comparison between groups of professionals. No oral healthcare professionals participated in this study, however various other studies have included these professionals^24,27,28. As the data on the delivery of tobacco cessation interventions were self-reported, the results may not be an accurate assessment of the actual interventions. There might have been other factors, other than those studied that could have influenced the results presented. Nonetheless, multivariate analyses were conducted, ruling out identified confounders.