Pre-post design was used.

The participant hospitals were selected at convenience because they had previously collaborated with the Training Unit of the ICO in training their health professionals in either electronic learning (e-learning) or in-person courses. The selected hospitals and their characteristics are shown in Table 1.

Each local coordinator recruited hospital workers from all units and departments over 6 months (from September 2015 to March 2016) through informative sessions, leaflets and posters (designed to inform about the training program), and personalized emails. Inclusion criteria were being a hospital worker and having an email account.

The original training program was developed by the online platform e-oncologia (http://www.e-oncologia.org/en/) based on the in-person courses offered during the last 10 years by the Tobacco Control Unit of the ICO. The Catalan and Spanish version of this course has been taken by almost 1000 workers before its adaptation to the Fruitful study. The final curriculum content of the “Brief Intervention for Smoking Cessation Training Program” is composed of four modules, and it is customized with specific data for each country:

• •

  Module 1 describes the tobacco epidemic with specific data for each country.

• •

  Module 2 provides orientation on how to assess tobacco dependence.

• •

  Module 3 explains in detail the 5A's intervention model.

• •

  Module 4 explains the different tobacco cessation treatments available (nicotine replacement, bupropion, varenicline, and other treatments), how to deal with relapses, and so on for each of the countries.

This online training includes: a) slides, b) review exercises, c) practical cases, and d) problem solving exercises. The training provides several materials including slides, online tutorials with an expert tutor, recommended readings, patient cessation brochures, and pharmaceutical pocket guideline, and an organizational recommendation model to facilitate the implementation of tobacco cessation services in the hospital setting. The course has been accredited by the Council of Oncology in Europe (ACOE, www.acoe.be) in support of Continuing Medical Education with 10hours20. The adaptation of the online course was done by local partners and other stakeholders. Extensive information of this process is described elsewhere20, however the main mismatches included: a) language background and literacy level in some of the terms used; b) description of the epidemiology smoking to the reality of Bolivia, Guatemala and Paraguay; c) current tobacco cessation resources in each country; d) simulations and demonstrations were modified to the cultural characteristics of each country; and e) questions and answers of the assessment and evaluation.20

After registering, hospital workers had access to the baseline survey which was compulsory to complete before accessing into the course. All participants were required to provide an e-mail address to receive an invitation to complete the evaluation survey. Only participants, who finished the training successfully, received an invitation to complete the evaluation survey after 6 months. A maximum of five reminders were sent. Besides, local coordinators personally reminded the need to participate in the post-training evaluation.

The pre- and post-evaluation was a 43-item online questionnaire based on an instrument developed by Sheffer et al.21 to measure cognitive, and some behavioral, factors: 1) motivation, 2) knowledge about tobacco cessation, 3) self-efficacy, 4) importance of providing tobacco use interventions, 5) effectiveness of interventions, 6) importance of barriers, and 7) preparedness.

The survey also assessed the self-reported level of performance according to the 5A'S (Ask, Advise, Assess, Assist, and Arrange follow-up) and included questions about responders’ individual characteristics, including sex, professional group (doctors, registered nurses, nurse assistants, others), tobacco use history, previous tobacco cessation education, and characteristics of their organization (public or private, hospital or other). In addition, we included questions identified previously in the literature and suggested by a panel of experts to explore behavioral factors such as: frequency in seeing tobacco-related diseases, using additional resources to intervene, having previous positive experience, and organizational-level characteristics: having records, systematic protocols, access to tobacco cessation pharmacological aids, among others (Cronbach's α = 0.77).22–24 All 43 items were assessed on a discrete scale of 0-10, with 0 being “none or not at all” and 10 being “the most possible”.

All participants were informed about the main objectives of the study and provided informed consent for their voluntary participation. This study protocol was approved by the Ethics Committee of the Hospital Universitari de Bellvitge (PR338/15).

Based on literature regarding gains of training,21,22 we anticipated an increase of 40% in health professionals’ level of knowledge, attitudes, and perception in tobacco cessation from baseline to 6 months. Given this estimate, we would need a minimum of 39 participants per hospital (total 117, α = 0.05, β = 0.2, and 10% dropouts).

Descriptive analyses were conducted on the demographic data. Main outcome variables (performance of 5A's and barriers and opportunities of their performance) did not follow normal distribution in all cases, so we conducted medians and interquartile ranges to describe pre- and post-results. Changes in the frequency of the smoking cessation intervention by each component of the 5A's brief intervention in smoking cessation model (Ask, Advise, Assess, Assist, and Arrange) were analyzed using the non-parametric Wilcoxon rank-sum test for paired samples. As, well as changes in cognitive, behavioral and organizational factors, which support or hamper the smoking cessation intervention, which were also tested using the Wilcoxon rank-sum test. The significant level for all the tests was 0.05.

Overall 202 health workers who finished the course, half of them were doctors (54.0%), and the rest were other health professionals (including nurses, psychologists, and technicians) (Table 2). The majority of them were women (72.3%), equally distributed by age groups (≤35 years old and or >35 years/old). Most reported they had never smoked (73.8%) and 11.9% were current smokers. Overall, 9.4% had previous training in smoking cessation. There were 76 participants in Bolivia, 79 in Guatemala and 47 in Paraguay who finished the training. We observed differences between the profiles of the participants by country: in Guatemala and Paraguay participants were mainly doctors; while in Bolivia other professionals were more prevalent (p <0.001) (Table 2).

From all the participants, we selected those who completed both pre- and post-evaluation (n = 99). Those that completed the two evaluations were similar to those who did not complete them except for country, thus hospital workers from Bolivia participated less in the pre-post evaluation than those from Guatemala and Paraguay (p = 0.018). In addition, we observed that those who completed the pre- and post- evaluations were more likely to Ask than those who did not completed both evaluations, but performed similarly in the rest of the 5A's model (Table 3).

Finally, taking into account the participation level (54.0% were doctors and 46% other health professionals) we decided to perform bivariate analysis between these two groups.

Table 4 displays the self-reported frequency of providing the 5A's smoking cessation intervention to patients before and after the training, overall and by independent variables.

Prior to the training the most frequent intervention performed by overall participants was Ask (median score of 7.5 out of 10) and the less frequent Arrange (median score of 0.5 out of 10). Generally, doctors reported higher frequency of performance of the three first components of the 5A's intervention model (Ask, Advise and Assess) than other health professional group (Table 4).

Six months after the training, participants reported higher and statistically significant performance scores in each of the 5A's components of the brief smoking cessation intervention model (p <0.001). The scores of doctors were on all cases higher before and after the training (Table 4). Former smokers obtained higher scores after the training in each of the 5A's components compared to smokers and non-smokers. However, never-smokers were the ones who gained more level of performance after the training compared to baseline (all 5A's components; p <0.001). In addition, those with previous smoking cessation training scored higher after the training but these changes were not statistically significant. Participants without previous training increased the most in the performance of each of the 5A's after the training (p <0.001).

By country, participants from Paraguay scored the highest in each of the 5A's both at the baseline and after the training compared to those from Bolivia and Guatemala. However, the difference in the performance of the 5A's was notorious between Ask, Advise and Assess compared to the scores obtained in Assist and Arrange (Table 4). In spite of this, there was a rise in the performance of all components by country (except for participants from Paraguay who already had high scores in Ask at baseline and the increase was not statistically significant after the training).

Table 5 summarizes some of the cognitive, behavioral and organizational factors that literature report to influence in the performance of the 5A's before and after the training.

At baseline, several factors scored ≤5, being identified as barriers. These were self-reported preparedness, preparedness in using drugs, competency in assisting patients, lack of knowledge and skills, use of resources, familiarity with resources, previous positive experiences and the three variables explored for assessing organizational and support factors (Table 5). On the other hand, there were a number of other factors scored over ≥7, identified as opportunities, such as having motivation to help patients to quit, reporting that smoking cessation is important in their job, desiring more training, believing that smoking cessation is part of their job, and seeing patients with tobacco-related diseases.

After the training, five of the identified barriers improved and showed statistically significant changes, all related to cognitive and behavioral factors (p <0.001) (Table 5). Thus, participants reported having gained preparedness, drug preparedness, competency in assisting smokers to quit, using additional resources and having positive experience (p <0.05). However, it was not observed a change in any of the organizational and support factors explored (Table 5).

After the training some factors identified as opportunities (scored ≥7) improved. Hence, participants reported being motivated to quit help patients to quit, thinking that smoking cessation is important for their job and seeing tobacco-related diseases patients more frequently (p <0.001).

By groups, we observed differences among doctors and other professionals (Table 5). The scores were higher among doctors in all the factors explored except for lack of knowledge and organizational and support factors. In addition, only other health professional group expressed a gain in their perception of having more recognition in their task after the training (from 0 to 3; p = 0.001).

Several limitations should be noted for this study. First, this study relies on self-reported responses, and we did not explored patient outcomes as it is suggested by experts in how integrate learning into practice33, that should explored in future research Anyhow in this phase of the FRUIFUL study, we were interested in testing the training course as a tool to gain performance of the 5A's among health care professionals in three Latin American countries. Second, our participants may not be representative of the general profile of the hospital workers in these three countries. Third, only half of our participants completed the pre-post evaluations even though they successfully completed the online course. Despite of the low level of compliance with the two questionnaires, the level of participation is similar to that in previous pre-post training evaluations that have used the Internet (around 50%).27,31 In addition, participants in our course might have a different level of computer skills and Internet usage than non-participants; given that certain computer skills were necessary. Online training programs also require devices for its use −for instance computers, tablets or smart-phones, etc.− and high-speed Internet connection. Internet service and technologic devices are rapidly growing in low-income countries but are not of high quality yet.34 However, this study explores, for the first time in three Latin American Countries, the impact of online smoking cessation training on changing the level of implementation of the 5A's model. Finally, our sample size was small (99 subjects) but enough for pre-post paired comparisons.