Our findings suggest that positive organizational climate is associated with enhanced performance of SBI for alcohol and substance abuse problems. Across 30 primary health care clinics in Brazil, multiple domains of OC were directly associated with self-reported use of SBI. Of seven OC factors, six showed significant positive correlations with SBI. In addition, several OC factors were also directly related to confidence and self-efficacy to perform SBI.
Before discussing the results in detail, it is important to consider a design limitation of the study and the degree to which it affects interpretation of the findings. Although it appears that OC influenced adoption of SBI, we measured OC at the end, rather than the beginning, of the study. It may be that our SBI training also affected OC within those teams, in which case we may be reporting an association that is not a causal relationship. Because we did not measure OC prior to measuring SBI performance, and we did not have a control group, our design does not rule out this possibility. However, we believe this is unlikely. Our classroom training was very focused on the concepts of SBI and how to deliver it in clinics; hence, we do not believe that aspect of the intervention could have strongly affected OC. It is possible our weekly performance feedback sessions could have affected OC somewhat, as these meetings involved interactive discussions of how to incorporate SBI into clinic practice, and at times may have touched on leadership, teamwork, and/or relationships with community members. However, even after the SBI training and feedback, clinics differed widely in their perceived OC, suggesting that the intervention effects on OC, if they occurred, were weak. Hence, we believe the OC values we observed at post training are likely to be similar to those we would have observed at baseline, as only three months elapsed between baseline and follow-up assessment.
Even after training and feedback in all participating clinics, there was some variation in adoption of SBI, and this variation may be due to underlying clinic-specific OC factors. Hence, OC may moderate the impact of training on the adoption of new clinical practices such as SBI. Some authors have found that in environments with a favorable OC, activities become more easy to perform, produce greater job satisfaction, and help workers reach their full potential [17, 28]. Clinicians who work in environments with less rigidity and more functional climates report higher levels of organizational commitment , which is important for incorporating new activities, such as SBI, into work routines.
To better understand and test this hypothesis, it would be useful to include other measures of OC in future trials to ensure all domains are included; i.e., assess OC at the beginning of a controlled trial of SBI training in clinics; include clinics with a wide range of OC in both study arms; and perhaps even design the study to explore the impact of directly addressing OC during SBI training. The analytic plan of such a trial could explicitly plan for analyses that examine the moderating effects of baseline OC on SBI performance. To better understand how specific domains of OC might moderate the effects of SBI training, we further explore the findings of the present study below.
Professional development (an OC factor) was most closely linked to SBI performance. The 11 items comprising this factor assess learning via trainings offered in the unit, improvement in productivity, investment in the personal and technical development of staff, application of newly acquired skills, and other features of environment supportive of development. Our findings accord with numerous studies that have demonstrated that training increases knowledge, builds confidence in addressing clinical topics [27, 29], and helps prevent burnout .
The second factor most frequently associated with SBI was relationship with the community. This factor assesses the strength of relationships within the health teams, between the teams and their communities, and between the teams and the municipal health department. A strong relationship with community residents is vital to the mission of Brazilian primary care because of the daily presence of PHC teams in the community and their mission to make preventive care accessible to all. Teams with strong community relationships may be more likely to implement SBI because they have caring and trusting relationships with families, which enables them to address sensitive issues such as drug problems. Studies show that community health workers (who made up the majority of our study sample) vary in terms of their implementation of drug prevention and treatment activities . Such workers, however, are key in bringing together the community and health services. They have strong potential for developing SBI interventions and for involving families, schools, and philanthropy organizations in early intervention activities .
Leadership was also associated with several measures of SBI implementation. The 17 items comprising this factor include supervision as well as motivation—two important aspects of leadership (e.g., “my manager is able to motivate my team, my manager corrects errors as they occur, my manager rewards good work, communication with my manager is clear and open, my manager keeps informed regarding the goals of the team”). Our findings support the suggestion by Patterson et al.  that clinical supervision is important to producing better services and patient outcomes. They argue it is important because supervision improves OC (and hence performance), supports implementation and quality of evidence-based practice, increases the visibility of professional counseling, and improves patient outcomes.
Remuneration was also related to SBI. This supports numerous studies that have linked remuneration, and perceived remuneration, with work performance and burnout [31, 32]. A study conducted among mental health professionals showed that the perceived balance between the effort to perform tasks and the reward was directly related to performance of evidence-based practice . Another study conducted among mental health workers found the perception of high reward was a protective factor for risk of burnout .
The factors least associated with SBI practices were strategy and team spirit. This result was different from the findings of several studies of OC that support the importance of teamwork forperformance and quality of services [33, 34]. It is important to highlight that remuneration and strategy were associated with some attitudes for substance use prevention but showed no relationship with SBI measures. Many professionals develop approaches to reducing the consumption of alcohol or other drugs, without necessarily using structured and standardized instruments such as the ASSIST. Similarly, some PHC workers could provide advice for behavior change even when they do not follow all the recommended steps in evidence-based BI.
In general, although self-reported SBI practices related to a number of OC factors, in terms of objectively measured behaviors, only professional development and relationship with community were significantly related to performance of SBI. The other OC dimensions of leadership, relationship, team spirit, worksite safety, team strategy, and remuneration did not correlate with SBI performance. Perhaps only teams that valued and participated in professional development were motivated to adopt new interventions such as SBI. Also, teams with strong links to the community might be more aware of the impact of drug use on their communities, less willing to judge and/or condemn patients with drug problems, and more willing to adopt new interventions that could benefit their communities, such as SBI.
Interestingly, professional development did not relate to confidence or self-efficacy to perform SBI, and relationship with community only related to confidence to perform SBI. Professional development activities for our participants probably had not ever addressed the treatment of drug abuse, because it has never been a mandate to address it in primary health care in Brazil. Perhaps having a strong relationship with the community would enable professionals to overcome reservations about addressing a sensitive and often illegal behavior, for the good of their patients.
This study had several limitations. First, this paper relates perceived OC with performance of BI. However, many more administrators and higher level clinic staff participated in the intervention than were included in the final OC survey. Our data on OC is based on 149 participants for whom we had complete survey data. These were mainly community health workers. However, a total of 230 health care providers completed the classroom part of the training, and most health care providers and administrators in participating clinics attended the post-training weekly clinic meetings in which we assessed ASSIST and BI performance and provided feedback. Hence, a much greater proportion of providers and administrators were probably exposed to the training and site intervention compared with the number who completed the OC assessment. The differences between the intervention population and the survey population may have reduced how representative the survey results were, may have introduced bias into the findings, and limited the number and type of analyses we were able to perform.
Second, given the nature of the data, it would have been better to use an analysis framework, such as mixed effects regression, that directly addressed the clustering and hierarchy. The sample in this study included 149 professionals distributed in 30 teams, including teams with less than five individuals observed in four municipalities. Although the method of aggregation we used was simpler and involved fewer assumptions, some statistical power may have been lost compared to a true multilevel analysis. Future studies should include greater representation within teams and examine whether clustering and hierarchy have effects on relationships between OC and SBI performance.
Third, our prospective measures of SBI performance relied on health-care provider administration of the ASSIST. It is possible that providers misreported screening rates to comply with the research protocol. We believe this is unlikely, as it would be difficult and time-consuming to fill out multiple ASSIST forms for nonexistent patients given that staff are extremely busy, and because we saw a great deal of variance in SBI performance across and within teams.
Fourth, factors other than OC may affect performance of SBI. Future studies could relate OC with infrastructure or other pre- or co-occurring conditions that could affect SBI performance, such as private versus public health care contexts, resources available to clinics, or specialist versus primary health care providers.
Last, although all of our participants attended the same SBI training, we do not know what other professional development activities they have attended or the extent to which their clinics supported ongoing professional development. Futures studies should assess the extent to which ongoing professional development opportunities contribute to receptiveness and ability to innovate among providers.