David N. Berg, PhD

Interest in healthcare teams has surged in recent years. A majority of the interest has been devoted to teamwork in the interdisciplinary clinical teams that staff operating rooms,1 emergency departments,2 and other inpatient settings.3 Interventions that enhance elements of teamwork like communication, mutual support among team members, and leadership have demonstrated effectiveness.4

Less attention has been paid to improving the success of hospital quality improvement (QI) teams, which gather individuals from different disciplines to improve a defined aspect of care. Studies suggest that QI teams can enable transformational change in healthcare systems,57 and that interdisciplinary representation,8, 9 physician involvement,10, 11 and clear goals12, 13 are associated with successful QI efforts. However, few studies have examined the behaviors of the QI teams that planned and implemented these efforts. Understanding how QI teams work to achieve their goals will allow hospitals to encourage these behaviors, and allow researchers to design interventions to augment these behaviors.

Accordingly, we sought to characterize the behaviors of successful interdisciplinary hospital QI teams. We previously reported on the strategies used by hospitals to reduce door‐to‐balloon times for patients with ST‐elevation myocardial infarction (STEMI)14, 15 to the evidence‐based guideline of 90 minutes.16 Our objective is to examine how QI teams designed and implemented these strategies. We believe that studying high‐performing QI teams is a first step to developing testable hypotheses about the effectiveness of QI team behaviors and mechanisms by which these behaviors might produce positive team outcomes.

METHODS

We designed a qualitative study using in‐depth interviews. We selected a qualitative methodology, since behaviors, social norms, and interpersonal interactions can be most appropriately examined using qualitative methods.17, 18 In addition, we used a positive deviance approach,19 where we focused on hospitals with top performance and the most improvement in door‐to‐balloon times. We sampled from hospitals in the National Registry of Myocardial Infarction (NRMI) who perform percutaneous coronary intervention (PCI, n = 151). We selected hospitals whose median door‐to‐balloon times were 90 minutes (n = 35). Then, we ranked hospitals in descending order according to their improvement during the previous 3 years (19992002). We sampled hospitals in descending order until we reached theoretical saturation where, as recommended for qualitative inquiry,2022 additional site visits did not uncover new concepts or patterns regarding our study questions. All sampled hospitals agreed to participate.

The first contact at each hospital was typically the director of QI. We asked to interview anyone with substantial involvement in the effort to reduce door‐to‐balloon times, and suggested that a wide variety of disciplines and roles be represented. We also used the snowball technique,22 where we asked participants to provide the names of individuals with substantial involvement in the reducing door‐to‐balloon times. Participants had varied levels of participation in QI teams. We purposely asked for minority and dissenting views from all participants.

At least 2 members of the research team conducted in‐depth interviews during hospital site visits. Interviews were conducted individually or in small groups, and lasted 1 to 1.5 hours. All data were audiotaped after verbal consent. Our interviews began with the grand tour question: What, if anything, has this hospital done to reduce its door‐to‐balloon times for patients with STEMI? The research team used standardized probes20, 23 to guide the discussion and achieve a complete understanding of the phenomena under study, including leadership and activities of the QI teams, and recommendations to other hospitals that wished to reduce door‐to‐balloon times. As recommended by experts,23 our interview guide was purposefully open‐ended to capture the range of experiences with QI teams. We did not specifically probe for facilitating or challenging behaviors. Audiotapes were transcribed by an independent, professional transcriptionist.

For this analysis, we defined QI teams as groups of administrators, providers, and staff who designed, implemented, and monitored processes to reduce door‐to‐balloon times. Each analysis team member independently cataloged quotes about team behaviors using a list of concepts (or codes). We then analyzed the quotes to identify recurrent themes relevant to the behaviors of interdisciplinary QI teams. We used the constant comparative method of analysis,20, 24, 25 which stipulates that the initial list of codes is refined as new transcripts are analyzed, and the final list is applied to all the transcripts. The analysis team included experts in QI, medicine, qualitative and health services research, as well as organizational psychology, and one of the interviewers. The presence of diverse perspectives in the analysis team,21 and a detailed audit trail20 to document the emergence of codes and themes, helped enhance researcher neutrality, data accuracy, and validity. We used Atlas.ti version 5.2 (Scientific Software Development GMbH, Berlin, Germany) to assist in the analysis.

RESULTS

Our sample (n = 11) included hospitals that varied on several characteristics (eg, geographic location), and median door‐to‐balloon times ranged from 55.5 to 89.5 minutes (Table 1). Hospitals in our sample had higher mean improvements in door‐to‐balloon times compared with non‐sampled NRMI hospitals (n = 140, 24 minutes vs 3 minutes over 3 years). Our interview participants (n = 122) included physicians, nurses, QI personnel, and administrative staff (Table 2). Five behaviors emerged from the data analysis. We found that interdisciplinary QI teams in successful hospitals focused on: (1) motivating involved hospital staff towards a shared goal, (2) creating opportunities for learning and problem‐solving, (3) addressing the impact of changes in care processes on staff, (4) protecting the integrity of the newly developed care processes, and (5) representing each involved clinical discipline effectively. These behaviors were recurrent across our diverse set of hospitals.

Discussion

We identified 5 behaviors of successful interdisciplinary QI teams based on our analysis of hospitals that reduced door‐to‐balloon times for patients with STEMI. These QI teams: (1) motivated involved hospital staff to consider lowering door‐to‐balloon times, a shared goal, (2) created opportunities for learning and problem‐solving, (3) addressed the impact of changes to care processes for patients with STEMI on staff, (4) protected the integrity of new care processes, and (5) represented each clinical discipline effectively by having members with in‐depth knowledge and authority.

Experts suggest that the key elements of effective teamwork in healthcare include prioritizing team over individual goals, mutual understanding, leadership, adaptability, and anticipation of the needs of others.26 These elements are supported by mutual trust and closed‐loop communication. The behaviors of QI teams in our study represent adaptive responses to the unique demands of QI in a complex organization. These teams went beyond an improvement model of identifying and analyzing a problem, and then developing and testing solutions by: (1) motivating and gathering information from each discipline, regardless of interdisciplinary conflicts; (2) responding to the concerns of front‐line staff, while maintaining control over the improvement process; and (3) sharing information across the hospital hierarchy. Table 3 illustrates potential relationships between the team behaviors in our data, the demands on hospital QI teams, and known elements of effective teamwork.

The behaviors in our study suggest effective teamwork strategies for QI. For example, our data suggest that successful interdisciplinary QI teams need effective representation from each involved discipline. This representation is necessary for motivation of front‐line staff, gathering of detailed information about processes, and the effective implementation of changes. Although this level of representation might challenge the cohesiveness of some teams,27 the teams in our sample managed conflict among disciplines without sacrificing the shared goal. By allocating attention and resources to the concerns of each discipline, the teams we studied prioritized team over individual goals and promoted mutual understanding.

Similarly, deciding when to modify the new protocols required leadership, adaptability, and anticipation of the needs of others. Successful QI teams in our sample modified protocols based on data and feedback, and created the mutual trust environment that is known to facilitate learning among disciplines.2830 Their willingness to learn, however, did not deter teams from protecting the integrity of new protocols. Lastly, participants stressed the importance of managing information across hierarchical boundaries. Managing reliable, timely, and accurate information across all levels is crucial to teamwork, and to the power and influence of a team.31

Our conclusions should be interpreted in light of several limitations. First, our study did not include a comparison group of low‐performing hospitals. We followed the recommendations of qualitative research experts23 who recommend sampling those with the most information on, and experience with, the phenomena under study (QI teams in high‐performing hospitals). The hypotheses we present here require further testing in quantitative studies of hospitals with diversity in QI team outcomes. Second, it is possible that sampled participants favored responses that they considered more desirable. To minimize this bias, we interviewed multiple participants per hospital, assured their confidentiality, and asked them to elaborate their responses. We sampled participants with a wide range of clinical and operational roles in each hospital, and also used the snowball sampling method to augment our sample. The range of responses collected, including frank discussions about setbacks, argues against the existence of contrasting behaviors to those captured. Third, although our sample included hospitals of various size and location, our findings might not reflect those of a larger sample of US hospitals. Last, the behaviors of QI teams may differ for other clinical processes.

Translating these findings into practice will require future studies of the impact of QI team behaviors on sustainability of quality gains. Since QI teams are not typically permanent, additional research is needed to identify behaviors associated with sustainable improvements. In addition, we must test whether the relationship between behaviors and team outcomes depends on whether the QI team strives to reach an evidence‐based goal or to improve a process as much as possible. Our sample demonstrated a combined approach, where the evidence‐based goal was followed by a desire to continue to further reduce door‐to‐balloon times. Similarly, the relationship between behaviors and team outcomes might depend on the catalyst for improvement (eg, regulatory pressure, an adverse event). The confluence of strong evidence and regulatory pressure that fueled these teams might not be true for other measures. Lastly, studies of teamwork in QI teams will require objective measures of team behaviors. A combination of surveys and direct team observation will likely be required to measure these behaviors, especially effective representation.

Our study highlights behaviors common to successful interdisciplinary QI teams in high‐performing hospitals. Previous studies have identified elements of teamwork and the importance of teams to QI, but have not examined team behaviors. In the era of an ever‐growing list of quality measures and of movement toward performance‐based reimbursement models,3234 hospitals have embraced the use of interdisciplinary teams as a key component of QI efforts. Our findings suggest that hospitals could enhance QI team effectiveness by promoting behaviors associated with successful interdisciplinary teams. When applied to QI teams, teamwork training could be supplemented with knowledge, attitudes, and skills regarding information‐gathering, problem‐solving, and communication across disciplines and levels of the hospital hierarchy.

Interest in healthcare teams has surged in recent years. A majority of the interest has been devoted to teamwork in the interdisciplinary clinical teams that staff operating rooms,1 emergency departments,2 and other inpatient settings.3 Interventions that enhance elements of teamwork like communication, mutual support among team members, and leadership have demonstrated effectiveness.4

Less attention has been paid to improving the success of hospital quality improvement (QI) teams, which gather individuals from different disciplines to improve a defined aspect of care. Studies suggest that QI teams can enable transformational change in healthcare systems,57 and that interdisciplinary representation,8, 9 physician involvement,10, 11 and clear goals12, 13 are associated with successful QI efforts. However, few studies have examined the behaviors of the QI teams that planned and implemented these efforts. Understanding how QI teams work to achieve their goals will allow hospitals to encourage these behaviors, and allow researchers to design interventions to augment these behaviors.

Accordingly, we sought to characterize the behaviors of successful interdisciplinary hospital QI teams. We previously reported on the strategies used by hospitals to reduce door‐to‐balloon times for patients with ST‐elevation myocardial infarction (STEMI)14, 15 to the evidence‐based guideline of 90 minutes.16 Our objective is to examine how QI teams designed and implemented these strategies. We believe that studying high‐performing QI teams is a first step to developing testable hypotheses about the effectiveness of QI team behaviors and mechanisms by which these behaviors might produce positive team outcomes.

METHODS

We designed a qualitative study using in‐depth interviews. We selected a qualitative methodology, since behaviors, social norms, and interpersonal interactions can be most appropriately examined using qualitative methods.17, 18 In addition, we used a positive deviance approach,19 where we focused on hospitals with top performance and the most improvement in door‐to‐balloon times. We sampled from hospitals in the National Registry of Myocardial Infarction (NRMI) who perform percutaneous coronary intervention (PCI, n = 151). We selected hospitals whose median door‐to‐balloon times were 90 minutes (n = 35). Then, we ranked hospitals in descending order according to their improvement during the previous 3 years (19992002). We sampled hospitals in descending order until we reached theoretical saturation where, as recommended for qualitative inquiry,2022 additional site visits did not uncover new concepts or patterns regarding our study questions. All sampled hospitals agreed to participate.

The first contact at each hospital was typically the director of QI. We asked to interview anyone with substantial involvement in the effort to reduce door‐to‐balloon times, and suggested that a wide variety of disciplines and roles be represented. We also used the snowball technique,22 where we asked participants to provide the names of individuals with substantial involvement in the reducing door‐to‐balloon times. Participants had varied levels of participation in QI teams. We purposely asked for minority and dissenting views from all participants.

At least 2 members of the research team conducted in‐depth interviews during hospital site visits. Interviews were conducted individually or in small groups, and lasted 1 to 1.5 hours. All data were audiotaped after verbal consent. Our interviews began with the grand tour question: What, if anything, has this hospital done to reduce its door‐to‐balloon times for patients with STEMI? The research team used standardized probes20, 23 to guide the discussion and achieve a complete understanding of the phenomena under study, including leadership and activities of the QI teams, and recommendations to other hospitals that wished to reduce door‐to‐balloon times. As recommended by experts,23 our interview guide was purposefully open‐ended to capture the range of experiences with QI teams. We did not specifically probe for facilitating or challenging behaviors. Audiotapes were transcribed by an independent, professional transcriptionist.

For this analysis, we defined QI teams as groups of administrators, providers, and staff who designed, implemented, and monitored processes to reduce door‐to‐balloon times. Each analysis team member independently cataloged quotes about team behaviors using a list of concepts (or codes). We then analyzed the quotes to identify recurrent themes relevant to the behaviors of interdisciplinary QI teams. We used the constant comparative method of analysis,20, 24, 25 which stipulates that the initial list of codes is refined as new transcripts are analyzed, and the final list is applied to all the transcripts. The analysis team included experts in QI, medicine, qualitative and health services research, as well as organizational psychology, and one of the interviewers. The presence of diverse perspectives in the analysis team,21 and a detailed audit trail20 to document the emergence of codes and themes, helped enhance researcher neutrality, data accuracy, and validity. We used Atlas.ti version 5.2 (Scientific Software Development GMbH, Berlin, Germany) to assist in the analysis.

RESULTS

Our sample (n = 11) included hospitals that varied on several characteristics (eg, geographic location), and median door‐to‐balloon times ranged from 55.5 to 89.5 minutes (Table 1). Hospitals in our sample had higher mean improvements in door‐to‐balloon times compared with non‐sampled NRMI hospitals (n = 140, 24 minutes vs 3 minutes over 3 years). Our interview participants (n = 122) included physicians, nurses, QI personnel, and administrative staff (Table 2). Five behaviors emerged from the data analysis. We found that interdisciplinary QI teams in successful hospitals focused on: (1) motivating involved hospital staff towards a shared goal, (2) creating opportunities for learning and problem‐solving, (3) addressing the impact of changes in care processes on staff, (4) protecting the integrity of the newly developed care processes, and (5) representing each involved clinical discipline effectively. These behaviors were recurrent across our diverse set of hospitals.

Discussion

We identified 5 behaviors of successful interdisciplinary QI teams based on our analysis of hospitals that reduced door‐to‐balloon times for patients with STEMI. These QI teams: (1) motivated involved hospital staff to consider lowering door‐to‐balloon times, a shared goal, (2) created opportunities for learning and problem‐solving, (3) addressed the impact of changes to care processes for patients with STEMI on staff, (4) protected the integrity of new care processes, and (5) represented each clinical discipline effectively by having members with in‐depth knowledge and authority.

Experts suggest that the key elements of effective teamwork in healthcare include prioritizing team over individual goals, mutual understanding, leadership, adaptability, and anticipation of the needs of others.26 These elements are supported by mutual trust and closed‐loop communication. The behaviors of QI teams in our study represent adaptive responses to the unique demands of QI in a complex organization. These teams went beyond an improvement model of identifying and analyzing a problem, and then developing and testing solutions by: (1) motivating and gathering information from each discipline, regardless of interdisciplinary conflicts; (2) responding to the concerns of front‐line staff, while maintaining control over the improvement process; and (3) sharing information across the hospital hierarchy. Table 3 illustrates potential relationships between the team behaviors in our data, the demands on hospital QI teams, and known elements of effective teamwork.

The behaviors in our study suggest effective teamwork strategies for QI. For example, our data suggest that successful interdisciplinary QI teams need effective representation from each involved discipline. This representation is necessary for motivation of front‐line staff, gathering of detailed information about processes, and the effective implementation of changes. Although this level of representation might challenge the cohesiveness of some teams,27 the teams in our sample managed conflict among disciplines without sacrificing the shared goal. By allocating attention and resources to the concerns of each discipline, the teams we studied prioritized team over individual goals and promoted mutual understanding.

Similarly, deciding when to modify the new protocols required leadership, adaptability, and anticipation of the needs of others. Successful QI teams in our sample modified protocols based on data and feedback, and created the mutual trust environment that is known to facilitate learning among disciplines.2830 Their willingness to learn, however, did not deter teams from protecting the integrity of new protocols. Lastly, participants stressed the importance of managing information across hierarchical boundaries. Managing reliable, timely, and accurate information across all levels is crucial to teamwork, and to the power and influence of a team.31

Our conclusions should be interpreted in light of several limitations. First, our study did not include a comparison group of low‐performing hospitals. We followed the recommendations of qualitative research experts23 who recommend sampling those with the most information on, and experience with, the phenomena under study (QI teams in high‐performing hospitals). The hypotheses we present here require further testing in quantitative studies of hospitals with diversity in QI team outcomes. Second, it is possible that sampled participants favored responses that they considered more desirable. To minimize this bias, we interviewed multiple participants per hospital, assured their confidentiality, and asked them to elaborate their responses. We sampled participants with a wide range of clinical and operational roles in each hospital, and also used the snowball sampling method to augment our sample. The range of responses collected, including frank discussions about setbacks, argues against the existence of contrasting behaviors to those captured. Third, although our sample included hospitals of various size and location, our findings might not reflect those of a larger sample of US hospitals. Last, the behaviors of QI teams may differ for other clinical processes.

Translating these findings into practice will require future studies of the impact of QI team behaviors on sustainability of quality gains. Since QI teams are not typically permanent, additional research is needed to identify behaviors associated with sustainable improvements. In addition, we must test whether the relationship between behaviors and team outcomes depends on whether the QI team strives to reach an evidence‐based goal or to improve a process as much as possible. Our sample demonstrated a combined approach, where the evidence‐based goal was followed by a desire to continue to further reduce door‐to‐balloon times. Similarly, the relationship between behaviors and team outcomes might depend on the catalyst for improvement (eg, regulatory pressure, an adverse event). The confluence of strong evidence and regulatory pressure that fueled these teams might not be true for other measures. Lastly, studies of teamwork in QI teams will require objective measures of team behaviors. A combination of surveys and direct team observation will likely be required to measure these behaviors, especially effective representation.

Our study highlights behaviors common to successful interdisciplinary QI teams in high‐performing hospitals. Previous studies have identified elements of teamwork and the importance of teams to QI, but have not examined team behaviors. In the era of an ever‐growing list of quality measures and of movement toward performance‐based reimbursement models,3234 hospitals have embraced the use of interdisciplinary teams as a key component of QI efforts. Our findings suggest that hospitals could enhance QI team effectiveness by promoting behaviors associated with successful interdisciplinary teams. When applied to QI teams, teamwork training could be supplemented with knowledge, attitudes, and skills regarding information‐gathering, problem‐solving, and communication across disciplines and levels of the hospital hierarchy.

Interest in healthcare teams has surged in recent years. A majority of the interest has been devoted to teamwork in the interdisciplinary clinical teams that staff operating rooms,1 emergency departments,2 and other inpatient settings.3 Interventions that enhance elements of teamwork like communication, mutual support among team members, and leadership have demonstrated effectiveness.4

Less attention has been paid to improving the success of hospital quality improvement (QI) teams, which gather individuals from different disciplines to improve a defined aspect of care. Studies suggest that QI teams can enable transformational change in healthcare systems,57 and that interdisciplinary representation,8, 9 physician involvement,10, 11 and clear goals12, 13 are associated with successful QI efforts. However, few studies have examined the behaviors of the QI teams that planned and implemented these efforts. Understanding how QI teams work to achieve their goals will allow hospitals to encourage these behaviors, and allow researchers to design interventions to augment these behaviors.

Accordingly, we sought to characterize the behaviors of successful interdisciplinary hospital QI teams. We previously reported on the strategies used by hospitals to reduce door‐to‐balloon times for patients with ST‐elevation myocardial infarction (STEMI)14, 15 to the evidence‐based guideline of 90 minutes.16 Our objective is to examine how QI teams designed and implemented these strategies. We believe that studying high‐performing QI teams is a first step to developing testable hypotheses about the effectiveness of QI team behaviors and mechanisms by which these behaviors might produce positive team outcomes.

METHODS

We designed a qualitative study using in‐depth interviews. We selected a qualitative methodology, since behaviors, social norms, and interpersonal interactions can be most appropriately examined using qualitative methods.17, 18 In addition, we used a positive deviance approach,19 where we focused on hospitals with top performance and the most improvement in door‐to‐balloon times. We sampled from hospitals in the National Registry of Myocardial Infarction (NRMI) who perform percutaneous coronary intervention (PCI, n = 151). We selected hospitals whose median door‐to‐balloon times were 90 minutes (n = 35). Then, we ranked hospitals in descending order according to their improvement during the previous 3 years (19992002). We sampled hospitals in descending order until we reached theoretical saturation where, as recommended for qualitative inquiry,2022 additional site visits did not uncover new concepts or patterns regarding our study questions. All sampled hospitals agreed to participate.

The first contact at each hospital was typically the director of QI. We asked to interview anyone with substantial involvement in the effort to reduce door‐to‐balloon times, and suggested that a wide variety of disciplines and roles be represented. We also used the snowball technique,22 where we asked participants to provide the names of individuals with substantial involvement in the reducing door‐to‐balloon times. Participants had varied levels of participation in QI teams. We purposely asked for minority and dissenting views from all participants.

At least 2 members of the research team conducted in‐depth interviews during hospital site visits. Interviews were conducted individually or in small groups, and lasted 1 to 1.5 hours. All data were audiotaped after verbal consent. Our interviews began with the grand tour question: What, if anything, has this hospital done to reduce its door‐to‐balloon times for patients with STEMI? The research team used standardized probes20, 23 to guide the discussion and achieve a complete understanding of the phenomena under study, including leadership and activities of the QI teams, and recommendations to other hospitals that wished to reduce door‐to‐balloon times. As recommended by experts,23 our interview guide was purposefully open‐ended to capture the range of experiences with QI teams. We did not specifically probe for facilitating or challenging behaviors. Audiotapes were transcribed by an independent, professional transcriptionist.

For this analysis, we defined QI teams as groups of administrators, providers, and staff who designed, implemented, and monitored processes to reduce door‐to‐balloon times. Each analysis team member independently cataloged quotes about team behaviors using a list of concepts (or codes). We then analyzed the quotes to identify recurrent themes relevant to the behaviors of interdisciplinary QI teams. We used the constant comparative method of analysis,20, 24, 25 which stipulates that the initial list of codes is refined as new transcripts are analyzed, and the final list is applied to all the transcripts. The analysis team included experts in QI, medicine, qualitative and health services research, as well as organizational psychology, and one of the interviewers. The presence of diverse perspectives in the analysis team,21 and a detailed audit trail20 to document the emergence of codes and themes, helped enhance researcher neutrality, data accuracy, and validity. We used Atlas.ti version 5.2 (Scientific Software Development GMbH, Berlin, Germany) to assist in the analysis.

RESULTS

Our sample (n = 11) included hospitals that varied on several characteristics (eg, geographic location), and median door‐to‐balloon times ranged from 55.5 to 89.5 minutes (Table 1). Hospitals in our sample had higher mean improvements in door‐to‐balloon times compared with non‐sampled NRMI hospitals (n = 140, 24 minutes vs 3 minutes over 3 years). Our interview participants (n = 122) included physicians, nurses, QI personnel, and administrative staff (Table 2). Five behaviors emerged from the data analysis. We found that interdisciplinary QI teams in successful hospitals focused on: (1) motivating involved hospital staff towards a shared goal, (2) creating opportunities for learning and problem‐solving, (3) addressing the impact of changes in care processes on staff, (4) protecting the integrity of the newly developed care processes, and (5) representing each involved clinical discipline effectively. These behaviors were recurrent across our diverse set of hospitals.

Discussion

We identified 5 behaviors of successful interdisciplinary QI teams based on our analysis of hospitals that reduced door‐to‐balloon times for patients with STEMI. These QI teams: (1) motivated involved hospital staff to consider lowering door‐to‐balloon times, a shared goal, (2) created opportunities for learning and problem‐solving, (3) addressed the impact of changes to care processes for patients with STEMI on staff, (4) protected the integrity of new care processes, and (5) represented each clinical discipline effectively by having members with in‐depth knowledge and authority.

Experts suggest that the key elements of effective teamwork in healthcare include prioritizing team over individual goals, mutual understanding, leadership, adaptability, and anticipation of the needs of others.26 These elements are supported by mutual trust and closed‐loop communication. The behaviors of QI teams in our study represent adaptive responses to the unique demands of QI in a complex organization. These teams went beyond an improvement model of identifying and analyzing a problem, and then developing and testing solutions by: (1) motivating and gathering information from each discipline, regardless of interdisciplinary conflicts; (2) responding to the concerns of front‐line staff, while maintaining control over the improvement process; and (3) sharing information across the hospital hierarchy. Table 3 illustrates potential relationships between the team behaviors in our data, the demands on hospital QI teams, and known elements of effective teamwork.

The behaviors in our study suggest effective teamwork strategies for QI. For example, our data suggest that successful interdisciplinary QI teams need effective representation from each involved discipline. This representation is necessary for motivation of front‐line staff, gathering of detailed information about processes, and the effective implementation of changes. Although this level of representation might challenge the cohesiveness of some teams,27 the teams in our sample managed conflict among disciplines without sacrificing the shared goal. By allocating attention and resources to the concerns of each discipline, the teams we studied prioritized team over individual goals and promoted mutual understanding.

Similarly, deciding when to modify the new protocols required leadership, adaptability, and anticipation of the needs of others. Successful QI teams in our sample modified protocols based on data and feedback, and created the mutual trust environment that is known to facilitate learning among disciplines.2830 Their willingness to learn, however, did not deter teams from protecting the integrity of new protocols. Lastly, participants stressed the importance of managing information across hierarchical boundaries. Managing reliable, timely, and accurate information across all levels is crucial to teamwork, and to the power and influence of a team.31

Our conclusions should be interpreted in light of several limitations. First, our study did not include a comparison group of low‐performing hospitals. We followed the recommendations of qualitative research experts23 who recommend sampling those with the most information on, and experience with, the phenomena under study (QI teams in high‐performing hospitals). The hypotheses we present here require further testing in quantitative studies of hospitals with diversity in QI team outcomes. Second, it is possible that sampled participants favored responses that they considered more desirable. To minimize this bias, we interviewed multiple participants per hospital, assured their confidentiality, and asked them to elaborate their responses. We sampled participants with a wide range of clinical and operational roles in each hospital, and also used the snowball sampling method to augment our sample. The range of responses collected, including frank discussions about setbacks, argues against the existence of contrasting behaviors to those captured. Third, although our sample included hospitals of various size and location, our findings might not reflect those of a larger sample of US hospitals. Last, the behaviors of QI teams may differ for other clinical processes.

Translating these findings into practice will require future studies of the impact of QI team behaviors on sustainability of quality gains. Since QI teams are not typically permanent, additional research is needed to identify behaviors associated with sustainable improvements. In addition, we must test whether the relationship between behaviors and team outcomes depends on whether the QI team strives to reach an evidence‐based goal or to improve a process as much as possible. Our sample demonstrated a combined approach, where the evidence‐based goal was followed by a desire to continue to further reduce door‐to‐balloon times. Similarly, the relationship between behaviors and team outcomes might depend on the catalyst for improvement (eg, regulatory pressure, an adverse event). The confluence of strong evidence and regulatory pressure that fueled these teams might not be true for other measures. Lastly, studies of teamwork in QI teams will require objective measures of team behaviors. A combination of surveys and direct team observation will likely be required to measure these behaviors, especially effective representation.

Our study highlights behaviors common to successful interdisciplinary QI teams in high‐performing hospitals. Previous studies have identified elements of teamwork and the importance of teams to QI, but have not examined team behaviors. In the era of an ever‐growing list of quality measures and of movement toward performance‐based reimbursement models,3234 hospitals have embraced the use of interdisciplinary teams as a key component of QI efforts. Our findings suggest that hospitals could enhance QI team effectiveness by promoting behaviors associated with successful interdisciplinary teams. When applied to QI teams, teamwork training could be supplemented with knowledge, attitudes, and skills regarding information‐gathering, problem‐solving, and communication across disciplines and levels of the hospital hierarchy.