Relationship among team dynamics, care coordination and perception of safety culture in primary care

Relationship among team dynamics, care coordination and perception of safety culture in primary care Abstract Background There remains a need to improve patient safety in primary care settings. Studies have demonstrated that creating high-performing teams can improve patient safety and encourage a safety culture within hospital settings, but little is known about this relationship in primary care. Objective To examine how team dynamics relate to perceptions of safety culture in primary care and whether care coordination plays an intermediating role. Research Design This is a cross-sectional survey study with 63% response (n = 1082). Subjects The study participants were attending clinicians, resident physicians and other staff who interacted with patients from 19 primary care practices affiliated with Harvard Medical School. Main Measures Three domains corresponding with our main measures: team dynamics, care coordination and safety culture. All items were measured on a 5-point Likert scale. We used linear regression clustered by practice site to assess the relationship between team dynamics and perceptions of safety culture. We also performed a mediation analysis to determine the extent to which care coordination explains the relationship between perceptions of team dynamics and of safety culture. Results For every 1-point increase in overall team dynamics, there was a 0.76-point increase in perception of safety culture [95% confidence interval (CI) 0.70–0.82, P < 0.001]. Care coordination mediated the relationship between team dynamics and the perception of safety culture. Conclusion Our findings suggest there is a relationship between team dynamics, care coordination and perceptions of patient safety in a primary care setting. To make patients safer, we may need to pay more attention to how primary care providers work together to coordinate care. Academic medicine, edical home, atient safety, primary care, quality of care Introduction Major gaps exist in our knowledge of patient safety in primary care settings (1). Both the volume of care provided in primary care settings and complexity of patients cared for outside the hospital have grown, increasing opportunities for errors. Attention has thus turned to patient safety in primary care settings where there remains a need to create safe, efficient, and sustainable systems (1). Safety culture—‘shared values, attitudes, and patterns of behavior regarding safety’ (2)—is seen as critical for achieving patient safety gains. Tools designed to measure staff members’ perceptions of safety culture in inpatient settings have been adapted for primary care (2–5). A better understanding of the role of safety culture in primary care will allow interested parties to develop and test interventions to improve primary care patient safety (6). Prior studies find that creating high-performing teams is important for encouraging a culture of safety (7–9). Better safety climates, in turn, have been correlated with fewer medication errors, nursing back injuries, readmission rates after acute myocardial infarctions, serious safety events and an array of patient safety indicators (2,8,10). Despite what is known from other industries and from inpatient settings about the importance of teamwork and safety culture, for example decreased error rates (7,9–11), few studies have looked specifically at the relationship between high-functioning teams and perceptions of safety culture in the primary care setting (6). More specifically, few studies consider how team dynamics within primary care practices that have established team-based care (which is a core element of patient-centred medical home initiatives) relate to safety culture. A particular area of concern highlighted by prior primary care patient safety research has been care coordination both within the same team and across different teams, both of which are important for safety culture and patient safety (11–14).Care coordination within a team has the potential to mediate the relationship between team dynamics and perceived safety culture because many errors in the outpatient setting occur as a result of coordination problems: breakdowns in communication, failures to close referral loops and insufficient follow-up of abnormal tests (11–14). In the current study, we are interested in understanding how team dynamics correlate with perceptions of safety culture and whether care coordination plays an intervening role. We conceptualize the relationships between team dynamics, care coordination and perceptions of patient safety as follows (Fig. 1). First, our model suggests that team dynamics—how the team is constituted and how it functions and performs—shape team members’ perceptions of their organization’s culture, specifically its safety culture. Thus, our first hypothesis is that ratings of overall team dynamics will relate positively to perceptions of safety culture (hypothesis 1). For example, we expect that leaders’ support of teamwork and team development and the corresponding actions frontline staff take to enact teamwork (e.g. scheduling regular team meetings, regular huddles prior to patient visits) lead to changes in team member behaviours (e.g. more frequent communication and more thorough test follow-up) and work climate that contribute to a culture that staff report feels more open and safe (2). Figure 1. View largeDownload slide Conceptual model for the relationship between team dynamics, care coordination and perception of safety culture Figure 1. View largeDownload slide Conceptual model for the relationship between team dynamics, care coordination and perception of safety culture Second, given our interest in the role of care coordination in the relationship between team dynamics and safety culture, we hypothesize that care coordination—specifically, the extent to which teams succeed at coordinating, such as by closing communication loops—mediates the positive relationship between team dynamics and perceived safety culture (hypothesis 2). In other words, we expect that care coordination is a mechanism through which teams improve safety. We expect this because working in teams can enable members to better coordinate patient-related activities, and improved care coordination enables teams to accomplish their safety objectives. In this way, experiencing improved coordination in teams would improve team members’ perceptions of safety culture. If so, care coordination would mediate between teamwork and perceived safety culture. Alternatively, it is possible that positive team dynamics merely result in better relationships and more positive affect among team members, without effect on coordination among members. Figure 1 also acknowledges the role of safety culture as a contextual factor shaping team dynamics in the context of teams working together to improve safety. We expect that any perceived improvement in safety culture in turn will also enable teams to function more successfully. In this study, we evaluated the relationship between team dynamics and perceptions of safety culture among patient-facing providers (i.e. any primary care staff member with patient contact) and whether their perceptions of care coordination mediated this relationship. Methods Study design We used a psychometrically validated instrument to conduct a cross-sectional study of 1716 practice personnel (i.e. attending clinicians, resident physicians and other patient-facing staff) who regularly interacted with patients in 2014 (15,16). One-fifth of staff were fully trained attending clinicians (primary care physicians, nurse practitioners and physician assistants) while 53% had other patient-facing staff roles (e.g. nurses, social workers, medical assistants and administrative personnel); 27% were resident physicians. Setting Our survey respondents worked within 19 practices that had established team-based care between 2012 and 2014 and were participating in the Harvard Academic Innovations Collaborative (AIC) initiative to promote Comprehensive, Accessible, Reliable, Exceptional and Safe (CARES) care from 2014 to 2016 (5). The goal of the learning collaborative was to improve patient safety, especially around breast and colorectal cancer screening and care of complex patients. All practice personnel were encouraged to participate in daily team huddles focused on improving information exchange about shared patients, monthly web-based meetings facilitated by practice change experts and quarterly in person meetings that involved curriculum focused on redesigning care processes. Most of the teams consisted of at least one attending clinician or resident physician and at least one of the other type of patient-facing staff such as a medical assistant, nurse and patient coordinator/front desk staff member. Some teams also included a social worker. Survey domains The survey included three key domains that mirrored goals of the AIC intervention: team dynamics, care coordination and perceptions of safety culture. Survey items in the domains of interest were all measured on a 5-point Likert scale (1 = strongly disagree, 3 = neither agree nor disagree, 5 = strongly agree). The survey was administered to staff via their work email addresses a total of three times between August and December 2014. The first part of the survey was a 29-item validated Primary Care Team Dynamics instrument, designed to assess team dynamics in ambulatory settings (15,16). Originally, this instrument included seven factors of primary care team dynamics: conditions for team effectiveness, shared understanding, processes for accountability, processes for communication and information exchange, processes for conflict resolution, acting and feeling like a team and perceived team effectiveness. The psychometric reliability and validity of this factor structure were demonstrated in the AIC population in previous administrations of the survey (15). We modified the previously validated team dynamics survey (15) by adding a new factor on learning processes so that the survey contained eight total domains related to team dynamics (Supplementary Material Survey Items). We also verified that these survey modifications did not adversely affect internal consistency (i.e. Cronbach’s alphas remained in the appropriate range 0.71–0.91) or discriminant validity (Pearson correlations between factors were lower than their internal consistency). The second part of the survey included a maximum of five items designed to assess care coordination (e.g. ease of knowing if critical tests were completed, there was follow-up on positive test results, whether a referral had been placed, what consulting physicians recommend for patients, and communication with specialists had occurred). All five care coordination questions were applicable to attending clinicians and residents; four (all but the last item) were applicable to other patient-facing staff. The Cronbach’s alpha for each set of items was 0.79 for attending clinicians and residents, and 0.89 for the other patient-facing staff. The third part of the survey assessed perceptions of safety culture using four previously validated questions from an inpatient survey (e.g. whether respondents agreed that practice managers have a clear picture of safety risks, that they have enough resources to provide safe patient care, they feel safe speaking up about a problem, and there is response when someone speaks up about patient safety problems). Cronbach’s alpha coefficient was 0.84 for these items and the same for attending and resident clinicians and other patient-facing staff. Survey item non-response For team dynamics and perceptions of safety culture items, non-response was extremely low at <1.5%. Non-response for care coordination items was at 6.3% overall, driven by other patient-facing staff who were less likely than attending clinicians and residents to respond to these questions. We used descriptive statistics to evaluate demographic characteristics of the study sample, including age, gender, race/ethnicity, job title and practice site. Analysis Descriptive statistics provided the frequency and average of response options stratified by professional role: attending clinicians, residents and other patient-facing staff. We used t-tests and ANOVA to compare differences between role groups. For each respondent, we calculated an overall team dynamics score by taking the mean response across all 33 items and for each of the eight team dynamic domains, removing not-applicable and missing responses where appropriate. We followed the same process for our dependent variable domain—perceptions of safety culture, and for our hypothesized mediator, care coordination. We used linear regression to examine the relationships between the independent variable, overall team dynamics, and the dependent variable, perceptions of safety culture. We controlled for all available demographic characteristics of respondents including age, gender and race/ethnicity. We also adjusted for whether the primary care practice site was a hospital-based or community-based practice. To determine whether care coordination was a potential mediator between team dynamics and perceptions of safety culture, we performed a mediation analysis using the process described by Baron and Kenny (17). The goal of the mediation test was to assess the total effect of overall team dynamics, our independent variable, on perceptions of safety culture through both direct and indirect effects. To do this, we ran two additional linear regressions, one looking at the relationship of the independent variable (overall team dynamics) with the mediator (care coordination) and the other assessing the relationship of the mediator (care coordination) with the dependent variable (perceptions of safety culture) while controlling for the independent variable (overall team dynamics). We also tested for the significance of the indirect effect or the mediating effect of care coordination on the relationship between team dynamics and perceptions of safety culture and calculated the per cent of total effect mediated, using the Sobel test of mediation (18). The Sobel test is used to test the significance of the mediating effect by determining whether the beta for the independent variable (overall team dynamics) is significantly different when the mediator is in the model. For all of the analyses mentioned earlier, we conducted unadjusted analyses on the group as a whole and for each of the different professional roles: attending clinicians, resident physicians and other patient-facing staff. When available, we adjusted for age, sex, race/ethnicity and practice site (hospital-based practice versus community-based practice). For all analyses, we used heteroskedasticity robust standard errors clustered by primary care practice. We performed all analyses using SAS statistical software version 9.4. Subgroup analysis Because resident physicians often spend less time at their primary care practice site and have fewer opportunities to be connected with team dynamics and to be aware of patient safety culture, we also repeated the analyses excluding resident physicians. Results Response rates Of the 1716 staff who received the survey, 1082 responded for a response rate of 63.1%. We received 1082 responses from 256 attending clinicians, 253 resident physicians and 573 other patient-facing staff. Study population The majority of the respondents were females (74.8%); women comprised 64.8% of attending clinicians, 56.5% of resident physicians and 87.3% of other patient-facing staff. The mean (SD) age for attending clinicians was 48.4 (11.4), and the mean age for other patient-facing staff was 41.4 (12.8). The majority of attending clinicians were non-Hispanic White (73.8%), and about half (49.2%) of other patient-facing staff were non-Hispanic White (Table 1). Table 1. Demographic characteristics of respondents to a survey assessing primary care team dynamics, safety culture and care coordination in primary care, all respondents and by role on the primary care team Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) aAttending clinicians refer to all Doctors of Medicine (MDs), Physician Assistants (PAs) and Nurse Practitioners (NPs). PAs and NPs account for 12% of the attending clinicians. bInformation on age and race/ethnicity was not collected in the resident physician survey. View Large Table 1. Demographic characteristics of respondents to a survey assessing primary care team dynamics, safety culture and care coordination in primary care, all respondents and by role on the primary care team Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) aAttending clinicians refer to all Doctors of Medicine (MDs), Physician Assistants (PAs) and Nurse Practitioners (NPs). PAs and NPs account for 12% of the attending clinicians. bInformation on age and race/ethnicity was not collected in the resident physician survey. View Large Level of team dynamics, safety culture and care coordination Ratings of team dynamics, perceptions of safety culture and care coordination were all on the positive side of neutral—ratings of perceptions of safety culture were the most positive while ratings of care coordination were most neutral. For all respondents (n = 1082), the mean (SD) of overall team dynamics was neutral to positive at 3.79 (0.60) on a 5-point scale. Mean levels of reported team dynamics were significantly different between the three professional role groups (F = 7.83, P < 0.001). Attending clinicians reported the lowest average level of team dynamics at 3.68 (0.60), compared with residents at 3.76 (0.58) (t = 1.61, P = 0.11), and other patient-facing staff reported the highest levels at 3.84 (0.60) (t = 3.67, P = 0.001) (Table 2). Table 2. Average levels of team dynamics, care coordination and perceptions of safety culture among survey respondents, all respondents and by role on the primary care team Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) All items used a 5-point Likert response scale (1 = strongly disagree to 5 = strongly agree). aOverall team dynamics is an average of the questions that make up the eight factors assessing primary care team dynamics. View Large Table 2. Average levels of team dynamics, care coordination and perceptions of safety culture among survey respondents, all respondents and by role on the primary care team Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) All items used a 5-point Likert response scale (1 = strongly disagree to 5 = strongly agree). aOverall team dynamics is an average of the questions that make up the eight factors assessing primary care team dynamics. View Large Overall, perceptions of safety culture were also on the positive side of neutral at 3.85 (0.72) on a 5-point scale and differences between professional groups only bordered on statistical significance (F = 2.78, P = 0.06). Across the study population, care coordination was rated 3.55 (0.90), indicating slightly less agreement with statements. In contrast to the other survey domains, other patient-facing staff reported the lowest average level of care coordination at 3.40 (1.09), residents remained in the moderate position at 3.66 (0.64), and attending clinicians at 3.76 (0.77) had the highest levels. The difference between all groups (F = 17.5, P < 0.001) as well as other patient-facing staff and both attending and resident clinicians were statistically significant (t = 6.32 and t = 4.19 P < 0.001, respectively). Team dynamics and perceptions of safety culture Overall, for every 1-point increase in overall team dynamics, perception of safety culture increased 0.76 point on a 5-point Likert scale [95% confidence interval (CI) 0.70–0.82, P < 0.001]. This relationship was significant for all three professional groups (all P < 0.001) in both unadjusted and adjusted models (Table 3). Table 3. Results from unadjusted and adjusted analyses of the relationship between overall team dynamics and perceived safety culture in primary care, all respondents and by role on the primary care team All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 CI, confidence interval. Dependent variable = perception of safety culture. Standard errors were clustered by primary care practice. aAll respondent adjusted model and resident physician models were only adjusted for sex, provider type and practice site (community versus hospital based) given that these were the only covariates available for all respondents and resident physicians. The attending clinicians and other patient-facing staff models were adjusted for age, sex, race/ethnicity and practice site. *P < 0.001. View Large Table 3. Results from unadjusted and adjusted analyses of the relationship between overall team dynamics and perceived safety culture in primary care, all respondents and by role on the primary care team All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 CI, confidence interval. Dependent variable = perception of safety culture. Standard errors were clustered by primary care practice. aAll respondent adjusted model and resident physician models were only adjusted for sex, provider type and practice site (community versus hospital based) given that these were the only covariates available for all respondents and resident physicians. The attending clinicians and other patient-facing staff models were adjusted for age, sex, race/ethnicity and practice site. *P < 0.001. View Large Patient care coordination as a mediator of team dynamics and perception of safety culture For all respondents, both the relationship of overall team dynamics and care coordination (95% CI 0.33–0.47, P < 0.001) and the relationship of care coordination and perceptions of safety culture were significant (95% CI 0.058–0.128, P < 0.001). Results from the Sobel test indicated that the association between overall team dynamics and the perception of safety culture is significantly mediated by care coordination (t = 4.51, P < 0.001) for the sample overall (Table 4, Supplementary Figure S1). However, care coordination only partially mediated this relationship. That is, the direct effect of overall team dynamics on perceptions of safety culture when controlling for care coordination remained significant for the full sample (Supplementary Figure S1). Table 4. Results of a mediation analysis investigating whether care coordination mediates the relationship between team dynamics and safety culture in primary care, all respondents and by role on the primary care team Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) All results were adjusted for gender, practice site (hospital based versus community based). For the analyses for the attending clinicians and other patient-fronting providers, results were also adjusted for age and race/ethnicity. Standard errors were also heteroskedasticity robust and clustered by primary care practice. CI, confidence interval. aTotal effect of team dynamics on perception of safety culture. bTotal effect of team dynamics on care coordination. cEffect of care coordination on perception of safety culture controlling for team dynamics. dDirect effect of team dynamics on perception of safety culture controlling for care coordination. *P < 0.05, **P < 0.01. View Large Table 4. Results of a mediation analysis investigating whether care coordination mediates the relationship between team dynamics and safety culture in primary care, all respondents and by role on the primary care team Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) All results were adjusted for gender, practice site (hospital based versus community based). For the analyses for the attending clinicians and other patient-fronting providers, results were also adjusted for age and race/ethnicity. Standard errors were also heteroskedasticity robust and clustered by primary care practice. CI, confidence interval. aTotal effect of team dynamics on perception of safety culture. bTotal effect of team dynamics on care coordination. cEffect of care coordination on perception of safety culture controlling for team dynamics. dDirect effect of team dynamics on perception of safety culture controlling for care coordination. *P < 0.05, **P < 0.01. View Large In our analysis stratified by professional role, we found that care coordination mediated the relationship between team dynamics and perceptions of safety culture for each role group. Additionally, while care coordination was found to be a partial mediator, it only explained a small amount of the total effect on the relationship between team dynamics and perceptions of safety culture across all role groups (4.67% for the whole cohort, 6.95% for attending clinicians, 8.15% for resident physicians and 5.60% for other patient-facing staff) (Table 4). Subgroup analysis When we excluded residents from the cohort, we found no differences in observed relationships. That is, we again found a strong positive relationship between team dynamics and perceptions of safety culture and that care coordination remained a significant partial mediator. Discussion Our study finds a significant and positive relationship between primary care practice personnel’s perceptions of overall team dynamics and their perceptions of safety culture, and this relationship is partially mediated by ratings of care coordination. To our knowledge, this is the first examination of the relationship between team dynamics, perceptions of safety culture and care coordination within the primary care setting (7,9,19,20). It is innovative because it examines the role that care coordination—a major focus in health care today—plays in mediating the relationship between team dynamics and perceived safety culture. Our findings are consistent with prior work on team dynamics and safety culture from other industries and other areas of medicine (7,19,21) and extend our understanding of the degree to which care coordination—perhaps by improving processes of care delivery and patterns of behaviour such as through closing communication loops, confirming that critical tests are completed or following up abnormal test results—makes providers feel that their practice is prioritizing safety and delivering safer and more reliable care. Our results, however, suggest that care coordination accounted for a small portion of the relationship between team dynamics and perceptions of safety culture. One potential explanation for why care coordination only partially mediated this relationship is that team dynamics may improve safety culture through mechanisms other than care coordination. For example, high-functioning teams could improve safety culture by increasing feelings of mutual respect and trust that team members can and will keep patients safe. High-functioning teams may also decrease hierarchy and increase psychological safety in ways that promote willingness of team members to speak up about safety concerns. These mechanisms may account for the relationship between team dynamics and perceptions of safety culture beyond the benefits of teams to care coordination. This study has limitations. It is a cross-sectional study based on perceptions among primary care personnel in the USA. Perceptions are, however, important, especially where team dynamics are concerned because dynamics are based on individuals’ experiences, as indicated by the differences experienced between the physician and non-physician professional groups. Future work would benefit from efforts to link perceptions to alternate measures of safety (e.g. event counts) and coordination (e.g. patients’ perspective). Also, our questions for care coordination were general, but can provide a foundation for more aspects of coordination (e.g. the quality of the systems supports for appropriate handoffs, or who specifically does what when coordinating across different teams, professional groups and health systems) as opposed to the interpersonal features of team relationships. Additionally, some of the care coordination questions may be physician centred and therefore may not have been applicable to the daily tasks of the other patient-facing providers. Further research to clarify the role of care coordination in safety culture could support new item development. Lastly, although this study was conducted in the USA, we expect the general relationship between team dynamics, care coordination and safety culture to be applicable and important to non-US settings. This is especially true as health care systems move towards team-based care as a means of better managing chronic disease in ambulatory settings. Future work should test this relationship both nationally and internationally. Our findings support a positive relationship between overall perceptions of team dynamics and perceptions of safety culture. Further research is needed to clarify what other factors mediate this relationship. Regardless of the directionality of the relationship between team dynamics and perceptions of safety culture, the relationship is positive. Ultimately, investment by primary care to improve care coordination remains critical for its potential to improve both safety culture and quality and delivery of primary care. Supplementary Material Supplementary data are available at Family Practice online. Declaration Funding: This research was supported by a National Research Service Award grant T32 HP10251 and the authors acknowledge funding for the 2-year AIC CARES Initiative from CRICO (Harvard Risk Management Foundation). Ethical approval: All research locations had institutional review board (IRB) via the IRB at the T.H. Chan Harvard School of Public Health. Individual data collection sites were deemed not to need IRB involvement because they did not have any identifiable data. Conflict of interest: The authors declare that they do not have any conflicts of interest related to this research. References 1. Wynia MK , Classen DC . Improving ambulatory patient safety: learning from the last decade, moving ahead in the next . JAMA 2011 ; 306 : 2504 – 5 . Google Scholar CrossRef Search ADS PubMed 2. Singer SJ , Vogus TJ . Reducing hospital errors: interventions that build safety culture . Annu Rev Public Health 2013 ; 34 : 373 – 96 . Google Scholar CrossRef Search ADS PubMed 3. Singer S , Meterko M , Baker L et al. Workforce perceptions of hospital safety culture: development and validation of the patient safety climate in healthcare organizations survey . Health Serv Res 2007 ; 42 : 1999 – 2021 . Google Scholar CrossRef Search ADS PubMed 4. Modak I , Sexton JB , Lux TR , Helmreich RL , Thomas EJ . Measuring safety culture in the ambulatory setting: the safety attitudes questionnaire—ambulatory version . J Gen Intern Med 2007 ; 22 : 1 – 5 . Google Scholar CrossRef Search ADS PubMed 5. Bitton A , Ellner A , Pabo E et al. The Harvard Medical School Academic Innovations Collaborative: transforming primary care practice and education . Acad Med 2014 ; 89 : 1239 – 44 . Google Scholar CrossRef Search ADS PubMed 6. Verbakel NJ , Langelaan M , Verheij TJM , Wagner C , Zwart DLM . Effects of patient safety culture interventions on incident reporting in general practice: a cluster randomised trial . Br J Gen Pract 2015 ; 65 : e319 – 29 . Google Scholar CrossRef Search ADS PubMed 7. Firth-Cozens J . Cultures for improving patient safety through learning: the role of teamwork . Qual Health Care 2001 ; 10 ( suppl 2 ): ii26 – 31 . Google Scholar PubMed 8. Berry JC , Davis JT , Bartman T et al. Improved safety culture and teamwork climate are associated with decreases in patient harm and hospital mortality across a hospital system . J Patient Saf 2016;1–7 . doi: 10.1097/PTS.0000000000000251 . 9. Baker DP , Salas E , Barach P . Medical Teamwork and Patient Safety: The Evidence-Based Relation. 2013. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.136.9246&rep=rep1&type=pdf (accessed 28 January 2016). 10. Hofmann DA , Mark B . An investigation of the relationship between safety climate and medication errors as well as other nurse and patient outcomes . Pers Psychol 2006 ; 59 : 847 – 69 . Google Scholar CrossRef Search ADS 11. Gandhi TK , Kachalia A , Thomas EJ et al. Missed and delayed diagnoses in the ambulatory setting: a study of closed malpractice claims . Ann Intern Med 2006 ; 145 : 488 – 96 . Google Scholar CrossRef Search ADS PubMed 12. Gandhi TK , Sittig DF , Franklin M et al. Communication breakdown in the outpatient referral process . J Gen Intern Med 2000 ; 15 : 626 – 31 . Google Scholar CrossRef Search ADS PubMed 13. Gandhi TK , Graydon-Baker E , Huber CN , Whittemore AD , Gustafson M . Closing the loop: follow-up and feedback in a patient safety program . Jt Comm J Qual Patient Saf 2005 ; 31 : 614 – 21 . Google Scholar CrossRef Search ADS PubMed 14. Singh H , Giardina TD , Meyer AND et al. Types and origins of diagnostic errors in primary care settings . JAMA Intern Med 2013 ; 173 : 418 – 25 . Google Scholar CrossRef Search ADS PubMed 15. Song H , Ryan M , Tendulkar S et al. Team dynamics, clinical work satisfaction, and patient care coordination between primary care providers: a mixed methods study . Health Care Manage Rev 2015; 42: 28–41 . doi: 10.1097/HMR.0000000000000091 . 16. Song H , Chien AT , Fisher J et al. Development and validation of the primary care team dynamics survey . Health Serv Res 2015 ; 50 : 897 – 921 . Google Scholar CrossRef Search ADS PubMed 17. Baron RM , Kenny DA . The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations . J Pers Soc Psychol 1986 ; 51 : 1173 – 82 . Google Scholar CrossRef Search ADS PubMed 18. Mackinnon DP , Dwyer JH . Estimating mediated effects in prevention studies . Eval Rev 1993 ; 17 : 144 – 58 . Google Scholar CrossRef Search ADS 19. Singer SJ , Falwell A , Gaba DM et al. Identifying organizational cultures that promote patient safety . Health Care Manage Rev 2009 ; 34 : 300 – 11 . Google Scholar CrossRef Search ADS PubMed 20. Singer S , Lin S , Falwell A , Gaba D , Baker L . Relationship of safety climate and safety performance in hospitals . Health Serv Res 2009 ; 44 ( 2 pt 1 ): 399 – 421 . Google Scholar CrossRef Search ADS PubMed 21. Mitchell P , Wynia M , Golden R et al. Core Principles & Values of Effective Team-Based Health Care . Washington, DC: Institute of Medicine ; 2012 . © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Family Practice Oxford University Press

Relationship among team dynamics, care coordination and perception of safety culture in primary care

Loading next page...
 
/lp/ou_press/relationship-among-team-dynamics-care-coordination-and-perception-of-F0HHSd0xaQ
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0263-2136
eISSN
1460-2229
D.O.I.
10.1093/fampra/cmy029
Publisher site
See Article on Publisher Site

Abstract

Abstract Background There remains a need to improve patient safety in primary care settings. Studies have demonstrated that creating high-performing teams can improve patient safety and encourage a safety culture within hospital settings, but little is known about this relationship in primary care. Objective To examine how team dynamics relate to perceptions of safety culture in primary care and whether care coordination plays an intermediating role. Research Design This is a cross-sectional survey study with 63% response (n = 1082). Subjects The study participants were attending clinicians, resident physicians and other staff who interacted with patients from 19 primary care practices affiliated with Harvard Medical School. Main Measures Three domains corresponding with our main measures: team dynamics, care coordination and safety culture. All items were measured on a 5-point Likert scale. We used linear regression clustered by practice site to assess the relationship between team dynamics and perceptions of safety culture. We also performed a mediation analysis to determine the extent to which care coordination explains the relationship between perceptions of team dynamics and of safety culture. Results For every 1-point increase in overall team dynamics, there was a 0.76-point increase in perception of safety culture [95% confidence interval (CI) 0.70–0.82, P < 0.001]. Care coordination mediated the relationship between team dynamics and the perception of safety culture. Conclusion Our findings suggest there is a relationship between team dynamics, care coordination and perceptions of patient safety in a primary care setting. To make patients safer, we may need to pay more attention to how primary care providers work together to coordinate care. Academic medicine, edical home, atient safety, primary care, quality of care Introduction Major gaps exist in our knowledge of patient safety in primary care settings (1). Both the volume of care provided in primary care settings and complexity of patients cared for outside the hospital have grown, increasing opportunities for errors. Attention has thus turned to patient safety in primary care settings where there remains a need to create safe, efficient, and sustainable systems (1). Safety culture—‘shared values, attitudes, and patterns of behavior regarding safety’ (2)—is seen as critical for achieving patient safety gains. Tools designed to measure staff members’ perceptions of safety culture in inpatient settings have been adapted for primary care (2–5). A better understanding of the role of safety culture in primary care will allow interested parties to develop and test interventions to improve primary care patient safety (6). Prior studies find that creating high-performing teams is important for encouraging a culture of safety (7–9). Better safety climates, in turn, have been correlated with fewer medication errors, nursing back injuries, readmission rates after acute myocardial infarctions, serious safety events and an array of patient safety indicators (2,8,10). Despite what is known from other industries and from inpatient settings about the importance of teamwork and safety culture, for example decreased error rates (7,9–11), few studies have looked specifically at the relationship between high-functioning teams and perceptions of safety culture in the primary care setting (6). More specifically, few studies consider how team dynamics within primary care practices that have established team-based care (which is a core element of patient-centred medical home initiatives) relate to safety culture. A particular area of concern highlighted by prior primary care patient safety research has been care coordination both within the same team and across different teams, both of which are important for safety culture and patient safety (11–14).Care coordination within a team has the potential to mediate the relationship between team dynamics and perceived safety culture because many errors in the outpatient setting occur as a result of coordination problems: breakdowns in communication, failures to close referral loops and insufficient follow-up of abnormal tests (11–14). In the current study, we are interested in understanding how team dynamics correlate with perceptions of safety culture and whether care coordination plays an intervening role. We conceptualize the relationships between team dynamics, care coordination and perceptions of patient safety as follows (Fig. 1). First, our model suggests that team dynamics—how the team is constituted and how it functions and performs—shape team members’ perceptions of their organization’s culture, specifically its safety culture. Thus, our first hypothesis is that ratings of overall team dynamics will relate positively to perceptions of safety culture (hypothesis 1). For example, we expect that leaders’ support of teamwork and team development and the corresponding actions frontline staff take to enact teamwork (e.g. scheduling regular team meetings, regular huddles prior to patient visits) lead to changes in team member behaviours (e.g. more frequent communication and more thorough test follow-up) and work climate that contribute to a culture that staff report feels more open and safe (2). Figure 1. View largeDownload slide Conceptual model for the relationship between team dynamics, care coordination and perception of safety culture Figure 1. View largeDownload slide Conceptual model for the relationship between team dynamics, care coordination and perception of safety culture Second, given our interest in the role of care coordination in the relationship between team dynamics and safety culture, we hypothesize that care coordination—specifically, the extent to which teams succeed at coordinating, such as by closing communication loops—mediates the positive relationship between team dynamics and perceived safety culture (hypothesis 2). In other words, we expect that care coordination is a mechanism through which teams improve safety. We expect this because working in teams can enable members to better coordinate patient-related activities, and improved care coordination enables teams to accomplish their safety objectives. In this way, experiencing improved coordination in teams would improve team members’ perceptions of safety culture. If so, care coordination would mediate between teamwork and perceived safety culture. Alternatively, it is possible that positive team dynamics merely result in better relationships and more positive affect among team members, without effect on coordination among members. Figure 1 also acknowledges the role of safety culture as a contextual factor shaping team dynamics in the context of teams working together to improve safety. We expect that any perceived improvement in safety culture in turn will also enable teams to function more successfully. In this study, we evaluated the relationship between team dynamics and perceptions of safety culture among patient-facing providers (i.e. any primary care staff member with patient contact) and whether their perceptions of care coordination mediated this relationship. Methods Study design We used a psychometrically validated instrument to conduct a cross-sectional study of 1716 practice personnel (i.e. attending clinicians, resident physicians and other patient-facing staff) who regularly interacted with patients in 2014 (15,16). One-fifth of staff were fully trained attending clinicians (primary care physicians, nurse practitioners and physician assistants) while 53% had other patient-facing staff roles (e.g. nurses, social workers, medical assistants and administrative personnel); 27% were resident physicians. Setting Our survey respondents worked within 19 practices that had established team-based care between 2012 and 2014 and were participating in the Harvard Academic Innovations Collaborative (AIC) initiative to promote Comprehensive, Accessible, Reliable, Exceptional and Safe (CARES) care from 2014 to 2016 (5). The goal of the learning collaborative was to improve patient safety, especially around breast and colorectal cancer screening and care of complex patients. All practice personnel were encouraged to participate in daily team huddles focused on improving information exchange about shared patients, monthly web-based meetings facilitated by practice change experts and quarterly in person meetings that involved curriculum focused on redesigning care processes. Most of the teams consisted of at least one attending clinician or resident physician and at least one of the other type of patient-facing staff such as a medical assistant, nurse and patient coordinator/front desk staff member. Some teams also included a social worker. Survey domains The survey included three key domains that mirrored goals of the AIC intervention: team dynamics, care coordination and perceptions of safety culture. Survey items in the domains of interest were all measured on a 5-point Likert scale (1 = strongly disagree, 3 = neither agree nor disagree, 5 = strongly agree). The survey was administered to staff via their work email addresses a total of three times between August and December 2014. The first part of the survey was a 29-item validated Primary Care Team Dynamics instrument, designed to assess team dynamics in ambulatory settings (15,16). Originally, this instrument included seven factors of primary care team dynamics: conditions for team effectiveness, shared understanding, processes for accountability, processes for communication and information exchange, processes for conflict resolution, acting and feeling like a team and perceived team effectiveness. The psychometric reliability and validity of this factor structure were demonstrated in the AIC population in previous administrations of the survey (15). We modified the previously validated team dynamics survey (15) by adding a new factor on learning processes so that the survey contained eight total domains related to team dynamics (Supplementary Material Survey Items). We also verified that these survey modifications did not adversely affect internal consistency (i.e. Cronbach’s alphas remained in the appropriate range 0.71–0.91) or discriminant validity (Pearson correlations between factors were lower than their internal consistency). The second part of the survey included a maximum of five items designed to assess care coordination (e.g. ease of knowing if critical tests were completed, there was follow-up on positive test results, whether a referral had been placed, what consulting physicians recommend for patients, and communication with specialists had occurred). All five care coordination questions were applicable to attending clinicians and residents; four (all but the last item) were applicable to other patient-facing staff. The Cronbach’s alpha for each set of items was 0.79 for attending clinicians and residents, and 0.89 for the other patient-facing staff. The third part of the survey assessed perceptions of safety culture using four previously validated questions from an inpatient survey (e.g. whether respondents agreed that practice managers have a clear picture of safety risks, that they have enough resources to provide safe patient care, they feel safe speaking up about a problem, and there is response when someone speaks up about patient safety problems). Cronbach’s alpha coefficient was 0.84 for these items and the same for attending and resident clinicians and other patient-facing staff. Survey item non-response For team dynamics and perceptions of safety culture items, non-response was extremely low at <1.5%. Non-response for care coordination items was at 6.3% overall, driven by other patient-facing staff who were less likely than attending clinicians and residents to respond to these questions. We used descriptive statistics to evaluate demographic characteristics of the study sample, including age, gender, race/ethnicity, job title and practice site. Analysis Descriptive statistics provided the frequency and average of response options stratified by professional role: attending clinicians, residents and other patient-facing staff. We used t-tests and ANOVA to compare differences between role groups. For each respondent, we calculated an overall team dynamics score by taking the mean response across all 33 items and for each of the eight team dynamic domains, removing not-applicable and missing responses where appropriate. We followed the same process for our dependent variable domain—perceptions of safety culture, and for our hypothesized mediator, care coordination. We used linear regression to examine the relationships between the independent variable, overall team dynamics, and the dependent variable, perceptions of safety culture. We controlled for all available demographic characteristics of respondents including age, gender and race/ethnicity. We also adjusted for whether the primary care practice site was a hospital-based or community-based practice. To determine whether care coordination was a potential mediator between team dynamics and perceptions of safety culture, we performed a mediation analysis using the process described by Baron and Kenny (17). The goal of the mediation test was to assess the total effect of overall team dynamics, our independent variable, on perceptions of safety culture through both direct and indirect effects. To do this, we ran two additional linear regressions, one looking at the relationship of the independent variable (overall team dynamics) with the mediator (care coordination) and the other assessing the relationship of the mediator (care coordination) with the dependent variable (perceptions of safety culture) while controlling for the independent variable (overall team dynamics). We also tested for the significance of the indirect effect or the mediating effect of care coordination on the relationship between team dynamics and perceptions of safety culture and calculated the per cent of total effect mediated, using the Sobel test of mediation (18). The Sobel test is used to test the significance of the mediating effect by determining whether the beta for the independent variable (overall team dynamics) is significantly different when the mediator is in the model. For all of the analyses mentioned earlier, we conducted unadjusted analyses on the group as a whole and for each of the different professional roles: attending clinicians, resident physicians and other patient-facing staff. When available, we adjusted for age, sex, race/ethnicity and practice site (hospital-based practice versus community-based practice). For all analyses, we used heteroskedasticity robust standard errors clustered by primary care practice. We performed all analyses using SAS statistical software version 9.4. Subgroup analysis Because resident physicians often spend less time at their primary care practice site and have fewer opportunities to be connected with team dynamics and to be aware of patient safety culture, we also repeated the analyses excluding resident physicians. Results Response rates Of the 1716 staff who received the survey, 1082 responded for a response rate of 63.1%. We received 1082 responses from 256 attending clinicians, 253 resident physicians and 573 other patient-facing staff. Study population The majority of the respondents were females (74.8%); women comprised 64.8% of attending clinicians, 56.5% of resident physicians and 87.3% of other patient-facing staff. The mean (SD) age for attending clinicians was 48.4 (11.4), and the mean age for other patient-facing staff was 41.4 (12.8). The majority of attending clinicians were non-Hispanic White (73.8%), and about half (49.2%) of other patient-facing staff were non-Hispanic White (Table 1). Table 1. Demographic characteristics of respondents to a survey assessing primary care team dynamics, safety culture and care coordination in primary care, all respondents and by role on the primary care team Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) aAttending clinicians refer to all Doctors of Medicine (MDs), Physician Assistants (PAs) and Nurse Practitioners (NPs). PAs and NPs account for 12% of the attending clinicians. bInformation on age and race/ethnicity was not collected in the resident physician survey. View Large Table 1. Demographic characteristics of respondents to a survey assessing primary care team dynamics, safety culture and care coordination in primary care, all respondents and by role on the primary care team Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) Characteristics All respondents (n = 1082) Attending clinicians (n = 256)a Resident physicians (n = 253)b Other patient-facing staff (n = 573) Gender, n (%) Male 255 (23.6) 88 (34.4) 110 (43.5) 57 (9.9) Female 809 (74.8) 166 (64.8) 143 (56.5) 500 (87.3) Declined to answer 18 (1.7) 2 (0.8) 0 16 (2.8) Age <30, n (%) 107 (9.9) 2 (0.8) – 105 (18.3) 30–35 123 (11.4) 35 (13.7) – 88 (15.4) 36–50 251 (23.2) 102 (39.8) – 149 (26.0) >50 233 (21.5) 96 (37.5) – 137 (23.9) Declined to answer 368 (34.0) 21 (8.2) – 94 (16.4) Race or ethnicity, n (%) White 471 (43.5) 189 (73.8) – 282 (49.2) Hispanic 151 (14.0) 14 (5.5) – 137 (23.9) African American 59 (5.5) 9 (3.5) – 50 (8.7) Asian/Pacific Islander 45 (4.2) 26 (10.2) – 19 (3.3) American Indian/Alaska Native 6 (0.6) 1 (0.4) – 5 (0.9) Others 42 (3.9) 8 (3.1) – 34 (5.9) Declined to answer 308 (28.5) 9 (3.5) – 46 (8.0) Practice site, n (%) Hospital-based practice 599 (55.4) 150 (58.6) 199 (78.7) 250 (43.6) Community-based practice 483 (44.6) 106 (41.4) 54 (21.3) 323 (56.4) aAttending clinicians refer to all Doctors of Medicine (MDs), Physician Assistants (PAs) and Nurse Practitioners (NPs). PAs and NPs account for 12% of the attending clinicians. bInformation on age and race/ethnicity was not collected in the resident physician survey. View Large Level of team dynamics, safety culture and care coordination Ratings of team dynamics, perceptions of safety culture and care coordination were all on the positive side of neutral—ratings of perceptions of safety culture were the most positive while ratings of care coordination were most neutral. For all respondents (n = 1082), the mean (SD) of overall team dynamics was neutral to positive at 3.79 (0.60) on a 5-point scale. Mean levels of reported team dynamics were significantly different between the three professional role groups (F = 7.83, P < 0.001). Attending clinicians reported the lowest average level of team dynamics at 3.68 (0.60), compared with residents at 3.76 (0.58) (t = 1.61, P = 0.11), and other patient-facing staff reported the highest levels at 3.84 (0.60) (t = 3.67, P = 0.001) (Table 2). Table 2. Average levels of team dynamics, care coordination and perceptions of safety culture among survey respondents, all respondents and by role on the primary care team Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) All items used a 5-point Likert response scale (1 = strongly disagree to 5 = strongly agree). aOverall team dynamics is an average of the questions that make up the eight factors assessing primary care team dynamics. View Large Table 2. Average levels of team dynamics, care coordination and perceptions of safety culture among survey respondents, all respondents and by role on the primary care team Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) Variables All respondents (n = 1082) Attending clinicians (n = 256) Resident physicians (n = 253) Other patient-facing staff (n = 573) Mean (SD) Mean (SD) Mean (SD) Mean (SD) Primary care team dynamics factors  Conditions for team effectiveness 3.66 (0.70) 3.53 (0.74) 3.62 (0.66) 3.74(0.69)  Shared understanding 3.89 (0.62) 3.81 (0.61) 3.86 (0.62) 3.94 (0.61)  Processes for communication and  information exchange 3.72 (0.73) 3.48 (0.75) 3.70 (0.71) 3.84 (0.71)  Processes for accountability 3.53 (0.98) 3.17 (1.05) 3.54 (0.94) 3.70 (0.93)  Processes for conflict resolution 3.46 (0.87) 3.25 (0.85) 3.58 (0.73) 3.52 (0.91)  Acting and feeling like a team 4.07 (0.62) 4.13 (0.56) 4.10 (0.62) 4.02 (0.64)  Perceived team effectiveness 3.79 (0.75) 3.70 (0.77) 3.73 (0.75) 3.83 (0.74)  Learning activities 3.62 (0.75) 3.49 (0.75) 3.51 (0.76) 3.73 (0.73)  Overall team dynamicsa 3.79 (0.60) 3.68 (0.60) 3.76 (0.58) 3.84 (0.60) Care coordination 3.55 (0.90) 3.77 (0.55) 3.66 (0.64) 3.40 (1.09) Perception of safety culture 3.85 (0.72) 3.76 (0.77) 3.90 (0.59) 3.87 (0.74) All items used a 5-point Likert response scale (1 = strongly disagree to 5 = strongly agree). aOverall team dynamics is an average of the questions that make up the eight factors assessing primary care team dynamics. View Large Overall, perceptions of safety culture were also on the positive side of neutral at 3.85 (0.72) on a 5-point scale and differences between professional groups only bordered on statistical significance (F = 2.78, P = 0.06). Across the study population, care coordination was rated 3.55 (0.90), indicating slightly less agreement with statements. In contrast to the other survey domains, other patient-facing staff reported the lowest average level of care coordination at 3.40 (1.09), residents remained in the moderate position at 3.66 (0.64), and attending clinicians at 3.76 (0.77) had the highest levels. The difference between all groups (F = 17.5, P < 0.001) as well as other patient-facing staff and both attending and resident clinicians were statistically significant (t = 6.32 and t = 4.19 P < 0.001, respectively). Team dynamics and perceptions of safety culture Overall, for every 1-point increase in overall team dynamics, perception of safety culture increased 0.76 point on a 5-point Likert scale [95% confidence interval (CI) 0.70–0.82, P < 0.001]. This relationship was significant for all three professional groups (all P < 0.001) in both unadjusted and adjusted models (Table 3). Table 3. Results from unadjusted and adjusted analyses of the relationship between overall team dynamics and perceived safety culture in primary care, all respondents and by role on the primary care team All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 CI, confidence interval. Dependent variable = perception of safety culture. Standard errors were clustered by primary care practice. aAll respondent adjusted model and resident physician models were only adjusted for sex, provider type and practice site (community versus hospital based) given that these were the only covariates available for all respondents and resident physicians. The attending clinicians and other patient-facing staff models were adjusted for age, sex, race/ethnicity and practice site. *P < 0.001. View Large Table 3. Results from unadjusted and adjusted analyses of the relationship between overall team dynamics and perceived safety culture in primary care, all respondents and by role on the primary care team All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 All respondents Attending clinicians Resident physicians Other patient-facing staff Overall team dynamics Unadjusted (95% CI) 0.75 (0.69–0.81)* 0.77 (0.67–0.87)* 0.65 (0.53–0.77)* 0.78 (0.68–0.88)* n = 1074 n = 256 n = 250 n = 568 Adjusted (95% CI) 0.76 (0.70–0.82)*a 0.75 (0.63–0.87)* 0.66 (0.54–0.78)* 0.80 (0.70–0.90)* n = 991 n = 246 n = 250 n = 517 CI, confidence interval. Dependent variable = perception of safety culture. Standard errors were clustered by primary care practice. aAll respondent adjusted model and resident physician models were only adjusted for sex, provider type and practice site (community versus hospital based) given that these were the only covariates available for all respondents and resident physicians. The attending clinicians and other patient-facing staff models were adjusted for age, sex, race/ethnicity and practice site. *P < 0.001. View Large Patient care coordination as a mediator of team dynamics and perception of safety culture For all respondents, both the relationship of overall team dynamics and care coordination (95% CI 0.33–0.47, P < 0.001) and the relationship of care coordination and perceptions of safety culture were significant (95% CI 0.058–0.128, P < 0.001). Results from the Sobel test indicated that the association between overall team dynamics and the perception of safety culture is significantly mediated by care coordination (t = 4.51, P < 0.001) for the sample overall (Table 4, Supplementary Figure S1). However, care coordination only partially mediated this relationship. That is, the direct effect of overall team dynamics on perceptions of safety culture when controlling for care coordination remained significant for the full sample (Supplementary Figure S1). Table 4. Results of a mediation analysis investigating whether care coordination mediates the relationship between team dynamics and safety culture in primary care, all respondents and by role on the primary care team Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) All results were adjusted for gender, practice site (hospital based versus community based). For the analyses for the attending clinicians and other patient-fronting providers, results were also adjusted for age and race/ethnicity. Standard errors were also heteroskedasticity robust and clustered by primary care practice. CI, confidence interval. aTotal effect of team dynamics on perception of safety culture. bTotal effect of team dynamics on care coordination. cEffect of care coordination on perception of safety culture controlling for team dynamics. dDirect effect of team dynamics on perception of safety culture controlling for care coordination. *P < 0.05, **P < 0.01. View Large Table 4. Results of a mediation analysis investigating whether care coordination mediates the relationship between team dynamics and safety culture in primary care, all respondents and by role on the primary care team Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) Team dynamics on perception of safety culturea Team dynamics on care coordinationb Care coordination on perception of safety culture given team dynamicsc Team dynamics on perception of safety culture given care coordinationd Per cent of total effect mediated P-Value Sobel test (t value) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) Estimate (95% CI) % – All respondents 0.761 (0.702–0.820)** 0.399 (0.326–0.472)** 0.093 (0.058–0.128)** 0.733 (0.684–0782)** 4.67 <0.001 (4.50) Attending Clinicians 0.752 (0.634–0.870)** 0.337 (0.200–0.474)** 0.166 (0.025–0.307)* 0.703 (0.580–0.826)** 6.95 0.038 (2.08) Residents 0.659 (0.539–0.781)** 0.450 (0.297–0.603)** 0.130 (0.120–0.140)* 0.608 (0.506–0.710)** 8.15 0.019 (2.34) Other patient-facing staff 0.802 (0.698–0.906)** 0.427 (0.315–0.539)** 0.093 (0.046–0.140)** 0.781 (0.656–0.906)** 4.73 0.001 (3.48) All results were adjusted for gender, practice site (hospital based versus community based). For the analyses for the attending clinicians and other patient-fronting providers, results were also adjusted for age and race/ethnicity. Standard errors were also heteroskedasticity robust and clustered by primary care practice. CI, confidence interval. aTotal effect of team dynamics on perception of safety culture. bTotal effect of team dynamics on care coordination. cEffect of care coordination on perception of safety culture controlling for team dynamics. dDirect effect of team dynamics on perception of safety culture controlling for care coordination. *P < 0.05, **P < 0.01. View Large In our analysis stratified by professional role, we found that care coordination mediated the relationship between team dynamics and perceptions of safety culture for each role group. Additionally, while care coordination was found to be a partial mediator, it only explained a small amount of the total effect on the relationship between team dynamics and perceptions of safety culture across all role groups (4.67% for the whole cohort, 6.95% for attending clinicians, 8.15% for resident physicians and 5.60% for other patient-facing staff) (Table 4). Subgroup analysis When we excluded residents from the cohort, we found no differences in observed relationships. That is, we again found a strong positive relationship between team dynamics and perceptions of safety culture and that care coordination remained a significant partial mediator. Discussion Our study finds a significant and positive relationship between primary care practice personnel’s perceptions of overall team dynamics and their perceptions of safety culture, and this relationship is partially mediated by ratings of care coordination. To our knowledge, this is the first examination of the relationship between team dynamics, perceptions of safety culture and care coordination within the primary care setting (7,9,19,20). It is innovative because it examines the role that care coordination—a major focus in health care today—plays in mediating the relationship between team dynamics and perceived safety culture. Our findings are consistent with prior work on team dynamics and safety culture from other industries and other areas of medicine (7,19,21) and extend our understanding of the degree to which care coordination—perhaps by improving processes of care delivery and patterns of behaviour such as through closing communication loops, confirming that critical tests are completed or following up abnormal test results—makes providers feel that their practice is prioritizing safety and delivering safer and more reliable care. Our results, however, suggest that care coordination accounted for a small portion of the relationship between team dynamics and perceptions of safety culture. One potential explanation for why care coordination only partially mediated this relationship is that team dynamics may improve safety culture through mechanisms other than care coordination. For example, high-functioning teams could improve safety culture by increasing feelings of mutual respect and trust that team members can and will keep patients safe. High-functioning teams may also decrease hierarchy and increase psychological safety in ways that promote willingness of team members to speak up about safety concerns. These mechanisms may account for the relationship between team dynamics and perceptions of safety culture beyond the benefits of teams to care coordination. This study has limitations. It is a cross-sectional study based on perceptions among primary care personnel in the USA. Perceptions are, however, important, especially where team dynamics are concerned because dynamics are based on individuals’ experiences, as indicated by the differences experienced between the physician and non-physician professional groups. Future work would benefit from efforts to link perceptions to alternate measures of safety (e.g. event counts) and coordination (e.g. patients’ perspective). Also, our questions for care coordination were general, but can provide a foundation for more aspects of coordination (e.g. the quality of the systems supports for appropriate handoffs, or who specifically does what when coordinating across different teams, professional groups and health systems) as opposed to the interpersonal features of team relationships. Additionally, some of the care coordination questions may be physician centred and therefore may not have been applicable to the daily tasks of the other patient-facing providers. Further research to clarify the role of care coordination in safety culture could support new item development. Lastly, although this study was conducted in the USA, we expect the general relationship between team dynamics, care coordination and safety culture to be applicable and important to non-US settings. This is especially true as health care systems move towards team-based care as a means of better managing chronic disease in ambulatory settings. Future work should test this relationship both nationally and internationally. Our findings support a positive relationship between overall perceptions of team dynamics and perceptions of safety culture. Further research is needed to clarify what other factors mediate this relationship. Regardless of the directionality of the relationship between team dynamics and perceptions of safety culture, the relationship is positive. Ultimately, investment by primary care to improve care coordination remains critical for its potential to improve both safety culture and quality and delivery of primary care. Supplementary Material Supplementary data are available at Family Practice online. Declaration Funding: This research was supported by a National Research Service Award grant T32 HP10251 and the authors acknowledge funding for the 2-year AIC CARES Initiative from CRICO (Harvard Risk Management Foundation). Ethical approval: All research locations had institutional review board (IRB) via the IRB at the T.H. Chan Harvard School of Public Health. Individual data collection sites were deemed not to need IRB involvement because they did not have any identifiable data. Conflict of interest: The authors declare that they do not have any conflicts of interest related to this research. References 1. Wynia MK , Classen DC . Improving ambulatory patient safety: learning from the last decade, moving ahead in the next . JAMA 2011 ; 306 : 2504 – 5 . Google Scholar CrossRef Search ADS PubMed 2. Singer SJ , Vogus TJ . Reducing hospital errors: interventions that build safety culture . Annu Rev Public Health 2013 ; 34 : 373 – 96 . Google Scholar CrossRef Search ADS PubMed 3. Singer S , Meterko M , Baker L et al. Workforce perceptions of hospital safety culture: development and validation of the patient safety climate in healthcare organizations survey . Health Serv Res 2007 ; 42 : 1999 – 2021 . Google Scholar CrossRef Search ADS PubMed 4. Modak I , Sexton JB , Lux TR , Helmreich RL , Thomas EJ . Measuring safety culture in the ambulatory setting: the safety attitudes questionnaire—ambulatory version . J Gen Intern Med 2007 ; 22 : 1 – 5 . Google Scholar CrossRef Search ADS PubMed 5. Bitton A , Ellner A , Pabo E et al. The Harvard Medical School Academic Innovations Collaborative: transforming primary care practice and education . Acad Med 2014 ; 89 : 1239 – 44 . Google Scholar CrossRef Search ADS PubMed 6. Verbakel NJ , Langelaan M , Verheij TJM , Wagner C , Zwart DLM . Effects of patient safety culture interventions on incident reporting in general practice: a cluster randomised trial . Br J Gen Pract 2015 ; 65 : e319 – 29 . Google Scholar CrossRef Search ADS PubMed 7. Firth-Cozens J . Cultures for improving patient safety through learning: the role of teamwork . Qual Health Care 2001 ; 10 ( suppl 2 ): ii26 – 31 . Google Scholar PubMed 8. Berry JC , Davis JT , Bartman T et al. Improved safety culture and teamwork climate are associated with decreases in patient harm and hospital mortality across a hospital system . J Patient Saf 2016;1–7 . doi: 10.1097/PTS.0000000000000251 . 9. Baker DP , Salas E , Barach P . Medical Teamwork and Patient Safety: The Evidence-Based Relation. 2013. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.136.9246&rep=rep1&type=pdf (accessed 28 January 2016). 10. Hofmann DA , Mark B . An investigation of the relationship between safety climate and medication errors as well as other nurse and patient outcomes . Pers Psychol 2006 ; 59 : 847 – 69 . Google Scholar CrossRef Search ADS 11. Gandhi TK , Kachalia A , Thomas EJ et al. Missed and delayed diagnoses in the ambulatory setting: a study of closed malpractice claims . Ann Intern Med 2006 ; 145 : 488 – 96 . Google Scholar CrossRef Search ADS PubMed 12. Gandhi TK , Sittig DF , Franklin M et al. Communication breakdown in the outpatient referral process . J Gen Intern Med 2000 ; 15 : 626 – 31 . Google Scholar CrossRef Search ADS PubMed 13. Gandhi TK , Graydon-Baker E , Huber CN , Whittemore AD , Gustafson M . Closing the loop: follow-up and feedback in a patient safety program . Jt Comm J Qual Patient Saf 2005 ; 31 : 614 – 21 . Google Scholar CrossRef Search ADS PubMed 14. Singh H , Giardina TD , Meyer AND et al. Types and origins of diagnostic errors in primary care settings . JAMA Intern Med 2013 ; 173 : 418 – 25 . Google Scholar CrossRef Search ADS PubMed 15. Song H , Ryan M , Tendulkar S et al. Team dynamics, clinical work satisfaction, and patient care coordination between primary care providers: a mixed methods study . Health Care Manage Rev 2015; 42: 28–41 . doi: 10.1097/HMR.0000000000000091 . 16. Song H , Chien AT , Fisher J et al. Development and validation of the primary care team dynamics survey . Health Serv Res 2015 ; 50 : 897 – 921 . Google Scholar CrossRef Search ADS PubMed 17. Baron RM , Kenny DA . The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations . J Pers Soc Psychol 1986 ; 51 : 1173 – 82 . Google Scholar CrossRef Search ADS PubMed 18. Mackinnon DP , Dwyer JH . Estimating mediated effects in prevention studies . Eval Rev 1993 ; 17 : 144 – 58 . Google Scholar CrossRef Search ADS 19. Singer SJ , Falwell A , Gaba DM et al. Identifying organizational cultures that promote patient safety . Health Care Manage Rev 2009 ; 34 : 300 – 11 . Google Scholar CrossRef Search ADS PubMed 20. Singer S , Lin S , Falwell A , Gaba D , Baker L . Relationship of safety climate and safety performance in hospitals . Health Serv Res 2009 ; 44 ( 2 pt 1 ): 399 – 421 . Google Scholar CrossRef Search ADS PubMed 21. Mitchell P , Wynia M , Golden R et al. Core Principles & Values of Effective Team-Based Health Care . Washington, DC: Institute of Medicine ; 2012 . © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Family PracticeOxford University Press

Published: May 18, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off