Impact of a complex chronic care patient case conference on quality and utilization

Impact of a complex chronic care patient case conference on quality and utilization Abstract There is need for effective venues to allow teams to coordinate care for high-risk or high-need patients. In addition, health systems need to assess the impact of such approaches on outcomes related to chronic health conditions and patient utilization. We evaluate the clinical impact of a novel case conference involving colocated trainees and supervisors in an interprofessional academic primary care clinic. The study utilized a prospective cohort with control group. Intervention patients (N = 104) were matched with controls (N = 104) from the same provider’s panel using propensity scores based on age, gender, risk predictors, and prior utilization patterns. Clinical outcomes and subsequent utilization patterns were compared prior to and up to 6 months following the conference. In terms of utilization, intervention patients demonstrated increased visits with primary care team members (p = .0002) compared with controls, without a corresponding increase in the number of primary care providers’ visits. There was a trend towards decreased urgent care and emergency visits (p = .07) and a significant decrease in the rate of hospitalizations (p = .04). Patients with poorly-controlled hypertension saw significant decreases in mean systolic blood pressure from 167 to 146 mm Hg. However, there were no differences between the intervention and control groups. Intervention patients with diabetes demonstrated a nonsignificant trend towards decreased hemoglobin A1c from 9.8 to 9.4, when compared with controls. Interprofessional case conferences have potential to improve care coordination and may be associated with improved disease management, decreased unplanned care, and overall reduced hospitalizations. Implications Practice: An interprofessional case conference for high-risk or high-need patients can change primary care dynamics to increase team-based encounters, with trends toward decreased unplanned care and hospitalizations. Policy: Policymakers should look for opportunities to support the time and effort required to develop recurrent team-based interprofessional case conferences. Research: Future research should be aimed at further evaluating the impact of interprofessional case conferences in controlled trials, as well as at different institutions and systems of care. INTRODUCTION The number of patients with multiple chronic conditions in the USA is increasing (1). This population of potentially high-need, high-cost patients uses more health care resources and requires more coordination to improve appropriate utilization and quality outcomes (2, 3). Primary care providers (PCPs) may be ill-equipped to facilitate coordination of care for such patients (4). At the same time, there is increased emphasis on colocation of different professions as part of Patient-Centered Medical Home and related models, which emphasize practice redesign to improve access, care coordination, and self-management support, particularly of complex patients (5, 6). With these, there are more opportunities for interprofessional collaboration to address patients with complex needs that utilize disproportionate services, and more evidence is needed on effective models of care coordination, and their subsequent impacts on important patient health care outcomes (7, 8). In 2010, the Department of Veterans Affairs (VA) converted to a medical home-like model known as Patient-Aligned Care Teams (PACT) (9). To further opportunities for interprofessional collaboration in this new clinic model, we developed locally the PACT Interprofessional Care Update (PACT ICU) to provide a venue to teach and provide ongoing collaborative team-based care for high-need, high-cost patients. The PACT ICU is an interprofessional care conference with the dual aim of improving care of high-risk patients in primary care, while also offering an opportunity to provide effective team-based care coordination. Educational outcomes and dissemination efforts have been evaluated and are reported separately (10, 11). The aims of this study were to evaluate the impact of an interprofessional case conference on the following: (i) clinical contacts with PCPs and members of the primary care team; (ii) quality of care of chronic diseases; and (iii) utilization of urgent care, emergency department, and hospitalizations. We hypothesized that the quality of care measured by improvements in indices of chronic disease (i.e., hemoglobin A1C and blood pressure) would improve in the intervention group compared with controls, that clinical contacts with the core PACT primary care team would increase in comparison to controls, and that utilization of urgent care, emergency room, and hospitalizations would decrease in the intervention group compared with controls. METHODS Study design We performed a prospective case–control trial with propensity score matching. This study was conducted and written consistent with STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines for observational trials (12). The Institutional Review Boards from the Puget Sound Health Care System and Boise VA Medical Center approved the study protocol. The study was conducted over a 26 month period, with ongoing recruitment for 20 months of patients with a 6 month follow-up period. Setting and participants This study was conducted at an interprofessional Academic PACT primary care clinic based at a VA medical center, with 36 trainees from internal medicine, nurse practitioner (NP), ambulatory pharmacy, and postdoctoral psychology programs. Patients were selected from approximately 2,000 veterans cared for on internal medicine and NP trainee panels. The final study size was a convenience sample, based on the number of intervention patients sequentially presented at the PACT ICU conference, and their matched controls. Patients were included if they were presented at PACT ICU during the study period of January 1, 2013 to October 14, 2015. Patient selection was continual during the study period and facilitated by a clinic-based registered nurse (RN) care manager, who coordinated presenters. The RN coordinator gave presenting PCP trainees, internal medicine and NP residents, and a list of the top 5–10 highest risk patients on their panels, as predicted by the VA-based Care Assessment Need (CAN) registry score. The CAN score effectively estimates the probability of hospitalization or death in the next 90 days in primary care patients within the VA, based on sociodemographic factors, medical conditions, recent clinical data, and prior utilization (13). Each PCP trainee then selected one patient that he or she felt was most appropriate for PACT ICU discussion. Intervention The structure and content of the PACT ICU conference has been previously described (10, 14, 15). In brief, it is an hour-long conference, in which two PCP trainees present selected patients to an interprofessional team of trainees and supervisors from their primary care clinic. The PCPs each select one new patient to present; this is shared with PACT ICU team members via encrypted email for chart review 1 week prior to the conference. Team members include adult-gerontology NPs, ambulatory pharmacists, internal medicine physicians, psychologists, RNs, and social workers who are affiliated with the primary care clinic. As part of a rotating schedule, faculty supervisors from internal medicine, NP, pharmacy, and psychology facilitate the conference using the “EFECT” approach, specifically (a) Eliciting a patient-centered narrative, (b) Facilitating an interprofessional team discussion, (c) Evaluating clinical evidence, (d) Creating a shared care plan, and (e) Tracking outcomes. (16) This is based on an expert consensus clinical approach emphasizing a focus on the patient narrative, evidence-based practice, facilitated interprofessional coordination, and development of a care plan that is followed for progress. A faculty facilitator leads the conference, soliciting input from participating medicine residents and preceptors, NP trainees and supervisors, primary care team nurses, pharmacists, psychologists, and social workers. At the end of the conference, the care plan is documented in the medical record, with specific action items for the team to complete; it is shared electronically with all participating team members. PCP trainees present on average approximately every 6 months, but participate more frequently depending on clinic schedules. Tasks developed in the care plan are shared among team members, who help us to complete the assigned tasks. These may include further chart review and recommendations, patient outreach via phone or monitoring, or a prescheduled appointment. Oftentimes, conferences are used to facilitate “warm hand-offs” which are opportunistic care transfers between colocated members of the team, such as psychology or pharmacy colleagues. Previously presented patients are briefly reviewed to discuss progress with existing care plans. Variables For intervention patients and matched controls, characteristics of gender, age, predicted risk of death or hospitalization, and death within 6 months of presentation, during the intervention period, were recorded. Average of blood pressures obtained in primary care settings for patients with a diagnosis of hypertension were calculated from VA Corporate Data Warehouse (CDW) data for 6 months prior to study enrollment and during the 6 month study follow-up to evaluate the impact on blood pressure control of care plans and team involvement. Due to concern for spurious elevation, blood pressures from ambulatory surgery, episodic care, emergency department, and hospitalization settings were excluded. Diabetes control was evaluated for patients with a diagnosis of diabetes, using the most recent hemoglobin A1c value prior to the PACT conference, and the first value between 90 and 240 days after the conference. All lab work was processed through the same facility laboratory; data were accessed from the electronic health record via the CDW. The VA’s CDW is an often-used source of clinical data collected directly from the VA’s electronic health record and is updated nightly (17). Encounters with primary care providers and associated primary care team members (pharmacy, psychology, and nursing) were tracked, as were urgent care, emergency department, and hospitalizations for 6 months prior to study enrollment and during the 6 month study follow-up of the PACT ICU intervention. These data were also retrieved from the CDW. Power analyses indicate that the sample size attained produced an 80 per cent power to detect small-to-medium effect sizes in the ANOVA (interaction effect size of f = .13 and a main effect size of f = .28). These sizes were deemed appropriate to detect clinically relevant measures of change in all measures of utilization (i.e., provider visits, team visits, urgent/episodic care, emergency department visits, and hospitalizations). Data sources/measurement Data were obtained from the CDW and reviewed for accuracy by study authors. The CDW is a relational database of clinical information that interfaces directly with the VA electronic health record and is updated nightly. Procedures for documentation of encounters are closely adhered to by clinicians and regularly reviewed by administrative staff; during the building of the database query for this study, there was continual cross-checking with clinical faculty. Chart review and communication among the clinical staff, database manager, and statistician provided an iterative review process. Data for individual participants were extracted for a baseline period of 6 month prior study enrollment. These variables were collected for the 6 month follow-up period to the PACT ICU, to evaluate the impact of the care conference intervention. Statistical analyses Data were analyzed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). Patient baseline characteristics were compared with t-test and chi-squared analysis. Utilization was broken down into 2 month segments to test for trends. Changes in utilization were evaluated with repeated ANOVA tests using “proc GLMMIXED” function, selecting a Poisson distribution appropriate for count data. Utilization during pre-PACT ICU and post-PACT ICU presentation time periods was compared, as were trends in utilization, using 2 month blocks. To evaluate the utilization data for each patient presented in PACT ICU, one of the four patients not selected for presentation was assigned to be a control using a propensity score based on prior utilization, CAN score, and age (18). This technique of modeling the selection process using propensity scores can mitigate some of the bias introduced by allowing the PCP to select the patients to present. Of note, only patients that were not selected at any time for the PACT ICU intervention and that did not die during study period were eligible for selection as a control. Among eligible controls, the single patient with the highest propensity score was chosen as the propensity score–matched control resulting in a 1:1 match. RESULTS Participants and descriptive data During the observation period, 104 patients were presented in the PACT ICU; 104 controls were selected using the propensity score–matching process described above (Fig. 1). The patients studied tended to be older males, with a baseline 90 day risk of death or hospitalization as predicted by CAN score, of approximately 22 per cent (Table 1). Of note, there were no significant differences in gender, age, CAN score, or overall propensity score. There were no statistically significant differences in baseline utilization, although patients presented in PACT ICU conference displayed a trend of increased utilization prior to the conference (Table 2). Fig 1 View largeDownload slide Selection process for Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) patient participants and controls. 1Trainees include nurse practitioner and internal medicine residents. Over the study time period, panel sizes changed, but there were approximately 2,000 patients at any given time. 2Six PACT ICU intervention patients were dropped because they died within 6 months of the intervention date. 3Only the first presentation to PACT ICU was used (19 dropped). 4Fifty-one possible control patients were dropped because they died within the 6 months of intervention date and 114 were dropped because they were later presented to PACT ICU. 5Of 403 possible controls, 104 were selected based on 1:1 propensity match with PACT ICU intervention patients. Fig 1 View largeDownload slide Selection process for Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) patient participants and controls. 1Trainees include nurse practitioner and internal medicine residents. Over the study time period, panel sizes changed, but there were approximately 2,000 patients at any given time. 2Six PACT ICU intervention patients were dropped because they died within 6 months of the intervention date. 3Only the first presentation to PACT ICU was used (19 dropped). 4Fifty-one possible control patients were dropped because they died within the 6 months of intervention date and 114 were dropped because they were later presented to PACT ICU. 5Of 403 possible controls, 104 were selected based on 1:1 propensity match with PACT ICU intervention patients. Table 1 Demographics of controls and PACT ICU intervention participants at baseline Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 aMeasures included in the calculation of propensity score for matching. View Large Table 1 Demographics of controls and PACT ICU intervention participants at baseline Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 aMeasures included in the calculation of propensity score for matching. View Large Table 2 Number of visits at baseline (prior to PACT ICU conference) for participants and matched controls PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 These utilization measures were included in the calculation of the propensity score for matching. View Large Table 2 Number of visits at baseline (prior to PACT ICU conference) for participants and matched controls PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 These utilization measures were included in the calculation of the propensity score for matching. View Large Utilization Both patients selected for PACT ICU presentation and matched controls from the same PCP tended to have increased PCP visits around the time of the conference (Fig. 2). Following the index date of the conference, both PACT ICU and control patients showed a trend to a decline in the number of visits over the 6 month follow-up period postintervention (p = .08). Overall, there was no significant difference between the number of PCP visits (Fig. 3) for the two groups (p = ns for trend). Conversely, there was an increase in PACT team encounters (e.g., pharmacy, behavioral health, and RN care managers) immediately preceding and following the PACT ICU conference, which was significantly different (p = .0002 for trend). This appeared to persist for at least 6 months, the follow-up period, following the conference intervention, after which it began to decrease. There was a nonsignificant trend towards decreased urgent care and emergency visits for the PACT ICU patients (Fig. 4) when compared with controls (p = .07), and a significantly decreased rate of hospitalizations (Fig. 5) over the 6 month follow-up period in the PACT ICU patients when compared with controls (p = .04). Fig 2 View largeDownload slide Number of Patient–Aligned Care Team (PACT) visits of intervention patients prior to and following presentation at PACT Interprofessional Care Update (ICU) conference, as well as those of matched controls. Post-intervention there were significant intervention effects (p = .0002). Patient encounters by pharmacy, nursing, social work, and psychology providers working with primary care providers (PCPs) in the same PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 2 View largeDownload slide Number of Patient–Aligned Care Team (PACT) visits of intervention patients prior to and following presentation at PACT Interprofessional Care Update (ICU) conference, as well as those of matched controls. Post-intervention there were significant intervention effects (p = .0002). Patient encounters by pharmacy, nursing, social work, and psychology providers working with primary care providers (PCPs) in the same PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 3 View largeDownload slide Number of primary care encounters of intervention patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference, as well as those of matched controls. Postintervention there were no significant intervention effects (p = ns). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 3 View largeDownload slide Number of primary care encounters of intervention patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference, as well as those of matched controls. Postintervention there were no significant intervention effects (p = ns). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 4 View largeDownload slide Number of emergency departments and urgent care visits of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were no significant effects (p = ns), but there was a nonsignificant trend towards decrease (p = .07). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 4 View largeDownload slide Number of emergency departments and urgent care visits of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were no significant effects (p = ns), but there was a nonsignificant trend towards decrease (p = .07). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 5 View largeDownload slide Number of hospitalizations of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were significant intervention effects (p = .04). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 5 View largeDownload slide Number of hospitalizations of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were significant intervention effects (p = .04). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Quality of care In terms of quality of care, there were no significant differences between the PACT ICU patients and matched controls for hypertension or glycemic control. There were improvements in hypertension following the PACT ICU conference date, but these were present in both populations, suggesting regression to the mean. Looking specifically at poorly controlled hypertensive patients (those with a baseline systolic blood pressure of 150 mm Hg or greater, n = 25), the mean systolic blood pressure of patients in both groups was 167 mm Hg (SD 17.5) prior to the index date, decreasing to an average of 146 mm Hg (SD 18.1) in the follow-up period (p < .0001). There were trends towards differences at baseline among PACT ICU patients with diabetes and hemoglobin A1c >7.5 per cent (n = 20); the most recent value prior to the conference averaged 9.8 per cent (SD 2.0) with a postconference hemoglobin A1c of 9.4 per cent (SD 1.6, p = .34); similarly matched controls (n = 30) had overall better baseline hemoglobin A1c (mean 8.7%, SD 2.0) with no change following the index date of the conference (mean 8.7%, SD 2.0. p = .99). A post hoc power analysis indicates that given the observed effect D = .21 in the PACT ICU group, it would require twice the number of participants to have 80 per cent power to detect a significant difference between pre- and post-measures A1c or mm Hg measures of blood pressure. Other analyses In terms of mortality during the study period, 9.1 per cent of all potential controls died during the 6 month follow-up, compared with 3.9 per cent of patients in the intervention group, presented in the PACT ICU conference (p = .05; Table 1). Given concerns regarding selection bias as well as right censored utilization data, these patients were appropriately excluded from analysis if they died during the follow-up period. However, sensitivity analyses of utilization outcomes, including and excluding deceased patient utilization data, did not affect results (data not displayed). DISCUSSION These findings suggest that interprofessional case conferences such as PACT ICU may facilitate increased team involvement in patient care rather than PCP-only visits. The increased involvement by associated team members persisted for 6 month follow-up period, following participation in the PACT ICU conference intervention. The trends towards reduced urgent/emergent care services and an overall reduced rate of hospitalizations suggest that the intervention may have helped us to avert unplanned medical care. Although there was improvement in hypertension following the PACT ICU conference, this was not isolated to the intervention arm and likely represents limitations of sample size, regression to the mean, or effects of independent efforts to address poorly controlled hypertension. Similarly, there were no statistically significant differential effects between the two groups related to glycemic control; this was confounded by poorer baseline control in the intervention arm. Limitations Limitations of the current study include the observational nature of the data and the inherent bias in selecting higher risk patients for inclusion. Although such bias is problematic for statistical analysis and assigning causality to observed effects, we felt that it was essential for proper patient selection and provider buy-in. There is evidence that although risk predictors are increasingly useful, they still do not fully account for complexity, particularly when compared with estimation by the empaneled PCP (19). Other sources of bias were considered: the most prominent being regression to the mean, a common problem in pre- or post-cohort studies in which participants are selected during a period of increased risk or utilization (20). Such bias is common in studies of high-risk patient interventions (21). To address both regression to the mean and selection bias, we employed a propensity-matched control group mentioned above (22, 23). Using a propensity score partially based on prior utilization is likely correlated with subsequent utilization; however, this would bias any subsequent difference in utilization towards the null, making any finding of change more robust. Finally, the convenience sampling approach of patients in this evaluation yielded small numbers of patients with poorly controlled chronic conditions; accordingly, we were underpowered to detect meaningful differences. Unfortunately, other quality or process measures that are likely affected by the PACT ICU, such as opioid treatment, frequency of benzodiazepine use, management of symptoms of depression, and advanced directive completion, were not prespecified or easily measured in our data set. Future studies could take these measures into account. In addition, death was common in both arms due to the population studied and may have created a bias due to censoring. However, sensitivity analysis completed with and without intervention patients who died did not detect this bias. Further limitations include ongoing efforts to improve team-based care via mechanisms and training other than the PACT ICU. It is not uncommon for one intervention to take place as part of a larger cultural change related to team-based care, allowing “cointervention” bias to occur (21). Additional analyses by our study team suggest that there was an increase in within-team consults in the months following the PACT ICU conference for all patients. Although this supports improved collaboration, it may have decreased the relative effect size of the increase in encounters by PACT team members, compared with controls. As is common in VA-based studies, there were small numbers of females, which may allow for introduction of bias in the findings, and decreased generalizability. The trend towards difference in mortality between intervention patients and selected controls was of interest, but we were unable to determine whether this was due to selection bias or direct impact of improved care and coordination. Providers were specifically counselled to select patients that would benefit from the case conference and future care coordination; accordingly, they were encouraged not to select patients already on hospice, or with diseases so progressed they were unlikely to affect change by the primary care team. Consequently, any difference in mortality may be related to a selection bias. There is evidence that the common sense approach of adding clinical triage to automated risk prediction remains necessary to account for more complex and nuanced predictors (24), particularly in identifying high risk primary care patients most appropriate for case conferences (25). Finally, although our study was rooted in a practical application of proactive, prepared team–based care as advocated in the Chronic Care Model and implemented in Patient-Centered Medical Homes and Patient Aligned Care Teams (26), we did not have a guiding theoretical model to refine our specific implementation and evaluation. Future efforts to design such an intervention in different contexts might benefit from a theoretical model such as the Relational Coordination Framework (27), which emphasizes an improved understanding of how collaborative care can be supported to improve quality and efficiency of care, so important in this population. Interpretation Although this conference required significant effort (several trainees and supervisors from different professions meeting on a regular basis), these preliminary results suggest that it may be worth the investment. Investing time into collaboration on high-need, high-cost patients is an appropriate allocation of resources, which can help us to change utilization patterns appropriately for such patients. This has been identified as a particularly important modality for patients with comorbid mental illness commonly found in multimorbidity patients (28). Interpreting these data in the context of other studies, it appears that there are heterogeneous effects of different types of care models. More evidence supports the idea that interprofessional approaches, centered around primary care teams, are more effective at improving quality and utilization than external approaches, those based on care transitions, and those not featuring interprofessional teams (8). To the best of our knowledge, there has not been a comparison between internal (such as PACT ICU) and externally located care coordination interventions; this would be an interesting research question for future studies. Similarly, comparing outcomes of primary care team–based care coordination and specialty care coordination interventions would be of interest. Several traditional and innovative models of specialty care coordination models exist, but these tend to have a different focus than PACT ICU, which emphasizes longitudinal primary care team coordination. More traditional case conferences such as tumor board or interdisciplinary surgery specialties tend to focus on a singular set of decisions regarding a known or suspected disease state, instead of long-term coordination of broader primary care issues. Other specialty-primary care collaborative conferences may focus on educating and consulting with PCPs about a disease state by experts at a distance (29), or other conferences focusing on methods to improve transitions of care for hospitalized patients with specialty input (30). Integration of behavioral health providers into primary care clinics appears to have improved quality of care and decreased costs in a large integrated health care system (31). Case conferences such as PACT ICU provide a venue for behavioral health providers (in our study, psychologists) to collaborate with the rest of the care team, providing opportunities for better integration in the care of these patients in specific, and the team in general. Other authors have reported outcomes related to case conferences, but these are often limited by the lack of an appropriate control arm (21, 32). Given the impact of regression to the mean in such studies, it is important to identify a comparator population, whether through matched controls, or a randomization system which allows for PCP selection of appropriate patients. Effective team-based care requires training and education to allow for the team to develop trust, allow effective communication, and understand different roles and responsibilities. This is particularly important for trainees who have increased exposure to interprofessional education, but may not have explicit exposure to collaborative clinical care. The National Academy of Medicine has called for a dual approach, encouraging workplace-learning opportunities for interprofessional care as well as evaluation beyond educational impacts (33). Our intervention helps us to extend the evaluation of the impact of interprofessional educational interventions beyond learner satisfaction or knowledge, identifying impacts on participant behavior changes and subsequent patient impacts. Overall, the PACT ICU conference provides a practical approach to learning interprofessional care in a collaborative practice model. It allows PCPs to select appropriate patients and empowers the team to develop a shared care plan that can alter patient utilization. Generalizability Although the data from this study are limited to one location, the PACT ICU model has been successfully exported to other academic training clinics within the VA (11, 15). Potentially important elements include a designated care coordinator, interprofessional teams (often featuring behavioral health such as psychology), a facilitator to ensure a coordinated care plan is agreed upon, and means of following up on patient outcomes. However, challenges exist to support this model in other settings, particularly those that require fee-for-service billing to support such efforts (34). There are good examples of community-based programs, which often lack controlled evaluations, yet may be useful examples (35); also, guidance for developing such conferences is also in development. (36) Although support of value-based care or accountable care organization contracts may make this more feasible, new Medicare billing codes that support complex chronic patient care and interprofessional conferences allow this to expand into more traditional billing structures (37), particularly to support behavioral health integration (38). Compliance with Ethical Standards Conflict of Interest: None declared. Primary Data: The authors have full control of the primary data and can submit deidentified data (e.g., minus protected health identifiers, in keeping with VA IRB requirements) to the journal for review upon request. The findings in this report have not been previously published, nor are they being simultaneously submitted elsewhere. Preliminary results related to this project have been presented orally at Northwest Society of General Internal Medicine and National Society of General Internal Medicine; no publications resulted from these presentations. Ethical Approval: No animals were involved in or harmed by this research project. Informed Consent: Waived as part of approval from both Puget Sound Health Care System and Boise VA Medical Center Institutional Review Board Acknowledgments Time for this project was indirectly supported via the VA Office of Academic Affiliations Centers of Excellence in Primary Care Education. References 1. Gerteis J , Izrael D , Deitz D et al. Multiple Chronic Conditions Chartbook . Rockville, MD : Agency for Healthcare Research and Quality ; 2014 . 2. Hong CS , Siegel AL , Ferris TG . Caring for high-need, high-cost patients: What makes for a successful care management program ? Issue Brief (Commonw Fund) . 2014 ; 19 : 1 – 19 . Google Scholar PubMed 3. Blumenthal D , Chernof B , Fulmer T , Lumpkin J , Selberg J . Caring for high-need, high-cost patients – An urgent priority . n Engl j Med . 2016 ; 375 ( 10 ): 909 – 911 . Google Scholar CrossRef Search ADS PubMed 4. Osborn R , Moulds D , Schneider EC , Doty MM , Squires D , Sarnak DO . Primary care physicians in ten countries report challenges caring for patients with complex health needs . Health Aff (Millwood) . 2015 ; 34 ( 12 ): 2104 – 2112 . Google Scholar CrossRef Search ADS PubMed 5. American College of Physicians, American Association of Pediatrics, American Association of Famiy Physicians, American Osteopathic Association . Joint Principles of the Patient Centered Medical Home . Available at https://www.acponline.org/acp_policy/policies/joint_principles_pcmh_2007.pdf. Accessibility verified January 21, 2018. 6. Ladebue AC , Helfrich CD , Gerdes ZT , Fihn SD , Nelson KM , Sayre GG . The experience of Patient Aligned Care Team (PACT) members . Health Care Manage Rev . 2016 ; 41 ( 1 ): 2 – 10 . Google Scholar CrossRef Search ADS PubMed 7. Zulman DM , Grant RW . Transforming care for complex patients: Addressing interconnected medical, social, and behavioral challenges . j Gen Intern Med . 2016 ; 31 ( 3 ): 263 – 264 . Google Scholar CrossRef Search ADS PubMed 8. McCarthy KJ , Ryan J , Klein S. Models of Care for High-Need, High-Cost Patients: An Evidence Synthesis . In: Lorber D , ed. New York, NY: Commonwealth Fund ; 2015 . 9. Piette JD , Holtz B , Beard AJ et al. ; Ann Arbor PACT Steering Committee . Improving chronic illness care for veterans within the framework of the Patient-Centered Medical Home: experiences from the Ann Arbor Patient-Aligned Care Team Laboratory . Transl Behav Med . 2011 ; 1 ( 4 ): 615 – 623 . Google Scholar CrossRef Search ADS PubMed 10. Weppner WG , Davis K , Sordahl J et al. Interprofessional care conferences for high-risk primary care patients . Acad Med . 2016 ; 91(6) : 798 – 802 . Google Scholar CrossRef Search ADS 11. King I , Strewler A , Wipf JE et al. Translating innovation: Exploring dissemination of a unique case conference . J Interprof Educ Pract . 2017 ; 6 : 55 – 60 . Google Scholar CrossRef Search ADS 12. von Elm E , Altman DG , Egger M , Pocock SJ , Gøtzsche PC , Vandenbroucke JP ; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies . Ann Intern Med . 2007 ; 147 ( 8 ): 573 – 577 . Google Scholar CrossRef Search ADS PubMed 13. Wang L , Porter B , Maynard C et al. Predicting risk of hospitalization or death among patients receiving primary care in the Veterans Health Administration . Med Care . 2013 ; 51 ( 4 ): 368 – 373 . Google Scholar CrossRef Search ADS PubMed 14. Buu J , Fisher A , Weppner W , Mason B . Impact of patient aligned care team interprofessional care updates on metabolic parameters . Fed Pract . 2016 ; 33 ( 2 ): 44 – 48 . 15. Pearson M , Patel S , Kim C et al. Interprofessional Case Conferences: Rewards and Challenges . Paper presented at: Society of General Internal Medicine Annual Meeting . San Diego, California ; 2014 . 16. Bitton A , Pereira AG , Smith CS , Babbott SF , Bowen JL ; Society of General Internal Medicine Patient-Centered Medical Home Education Summit . The EFECT framework for interprofessional education in the patient centered medical home . Healthc (Amst) . 2013 ; 1 ( 3-4 ): 63 – 68 . Google Scholar CrossRef Search ADS PubMed 17. Corporate Data Warehouse (CDW); VA Informatics and Computing Infrastructure. Health Services Research and Development; U.S. Department of Veterans Affairs . Available at https://www.hsrd.research.va.gov/for_researchers/vinci/cdw.cfm. Accessibility verified December 15, 2017 . 18. Pirracchio R , Resche-Rigon M , Chevret S . Evaluation of the propensity score methods for estimating marginal odds ratios in case of small sample size . bmc Med Res Methodol . 2012 ; 12 : 70 . Google Scholar CrossRef Search ADS PubMed 19. Grant RW , Ashburner JM , Hong CS et al. Defining patient complexity from the primary care physician’s perspective: A cohort study . Ann Intern Med . 2011 ; 155 ( 12 ): 797 – 804 . Google Scholar CrossRef Search ADS PubMed 20. Morton V , Torgerson DJ . Regression to the mean: Treatment effect without the intervention . j Eval Clin Pract . 2005 ; 11 ( 1 ): 59 – 65 . Google Scholar CrossRef Search ADS PubMed 21. Lynch CS , Wajnberg A , Jervis R et al. Implementation Science Workshop: A novel multidisciplinary primary care program to improve care and outcomes for super-utilizers . j Gen Intern Med . 2016 ; 31 ( 7 ): 797 – 802 . Google Scholar CrossRef Search ADS PubMed 22. Cottone F , Efficace F , Apolone G , Collins GS . The added value of propensity score matching when using health-related quality of life reference data . Stat Med . 2013 ; 32 ( 29 ): 5119 – 5132 . Google Scholar CrossRef Search ADS PubMed 23. Rosenbaum PR , Rubin DB . The central role of the propensity score in observational studies for causal effects . Biometrika . 1983 ; 70(1) : 41 – 55 . Google Scholar CrossRef Search ADS 24. Hwang AS , Atlas SJ , Hong J et al. Defining team effort involved in patient care from the primary care physician’s perspective . j Gen Intern Med . 2017 ; 32 ( 3 ): 269 – 276 . Google Scholar CrossRef Search ADS PubMed 25. Watts B , Lawrence R , Carter C et al. How Risky is Watching and Waiting? A Pilot Study Combining Clinical Triage and Automated Risk Modeling to Enhance Allocation of Ambulatory Intensive Case Management Resources for High-Risk Patients . Poster presented at: Society of General Internal Medicine . Hollywood, Florida ; 2016 . 26. Wagner EH , Coleman K , Reid RJ et al. Guiding Transformation: How Medical Practices Can Become Patient-Centered Medical Homes . New York, NY: The Commonwealth Fund , February 2012 . 27. McDonald KM , Sundaram V , Bravata DM et al. Care coordination . Vol 7. In: Shojania KG , McDonald KM , Wachter RM , Owens DK , eds. Closing the Quality Gap: A Critical Analysis of Quality Improvement Strategies. Technical Review 9 (Prepared by the Stanford University-UCSF Evidence-based Practice Center under contract 290-02-0017) . AHRQ Publication No. 04(07)-0051-7. Rockville, MD : Agency for Healthcare Research and Quality ; June 2007 . 28. Ward MC , Miller BF , Marconi VC , Kaslow NJ , Farber EW . The role of behavioral health in optimizing care for complex patients in the primary care setting . j Gen Intern Med . 2016 ; 31 ( 3 ): 265 – 267 . Google Scholar CrossRef Search ADS PubMed 29. Department of Veterans Affairs . Fact Sheet: Specialty Care Access Network-Extension for Community (SCAN-ECHO) . Available at https://www.va.gov/HEALTH/docs/Specialty-Care-Access-Network.pdf, Accessibility verified November 14, 2017 . 30. ClinicalTrials.gov . Integrating Care After Exacerbations of COPD (InCasE) , PI David Au; Identifier NCT02021955. Available at https://clinicaltrials.gov/ct2/show/NCT02021955. Accessibility verified November 14, 2017 . 31. Reiss-Brennan B , Brunisholz KD , Dredge C et al. Association of integrated team-based care with health care quality, utilization, and cost . Jama . 2016 ; 316 ( 8 ): 826 – 834 . Google Scholar CrossRef Search ADS PubMed 32. Ritchie C , Andersen R , Eng J et al. Implementation of an interdisciplinary, team-based complex care support health care model at an Academic Medical Center: Impact on health care utilization and quality of life . Plos One . 2016 ; 11 ( 2 ): e0148096 . Google Scholar CrossRef Search ADS PubMed 33. Institute of Medicine: Measuring the Impact of Interprofessional Education on Collaborative Practice and Patient Outcomes . Institute of Medicine of the National Academies: Washington, DC ; 2015 . https://www.ncbi.nlm.nih.gov/books/NBK338360/ 34. Hong CS , Abrams MK , Ferris TG . Toward increased adoption of complex care management . n Engl j Med . 2014 ; 371 ( 6 ): 491 – 493 . Google Scholar CrossRef Search ADS PubMed 35. Warning W II , Wood W , Letcher A et al. Working With the Super-Utilizer Population: The Experience and Recommendations of Five Pennsylvania Programs ; 2014 . Available at http://www.aligning4healthpa.org/pdf/High_Utilizer_report.pdf. Accessibility verified January 21, 2018. 36. O’Brien BC , Patel SR , Pearson M et al. 12 Tips for Delivering Successful Interprofessional Case Conferences . Med Teach . 2017 ; 39 ( 12 ): 1214 – 1220 . Google Scholar CrossRef Search ADS PubMed 37. Center for Medicare and Medicaid Services . Chronic Care Management Services . Department of Health and Human Services ; 2016 . Available at https://www.cms.gov/About-CMS/Agency-Information/OMH/equity-initiatives/chronic-care-management.html. Accessibility verified January 21, 2018. 38. Press MJ , Howe R , Schoenbaum M et al. Medicare payment for behavioral health integration . n Engl j Med . 2017 ; 376 ( 5 ): 405 – 407 . Google Scholar CrossRef Search ADS PubMed Published by Oxford University Press on behalf of the Society of Behavioral Medicine 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Translational Behavioral Medicine Oxford University Press

Impact of a complex chronic care patient case conference on quality and utilization

Loading next page...
 
/lp/ou_press/impact-of-a-complex-chronic-care-patient-case-conference-on-quality-MZkuZR8b5L
Copyright
Published by Oxford University Press on behalf of the Society of Behavioral Medicine 2018.
ISSN
1869-6716
eISSN
1613-9860
D.O.I.
10.1093/tbm/ibx082
Publisher site
See Article on Publisher Site

Abstract

Abstract There is need for effective venues to allow teams to coordinate care for high-risk or high-need patients. In addition, health systems need to assess the impact of such approaches on outcomes related to chronic health conditions and patient utilization. We evaluate the clinical impact of a novel case conference involving colocated trainees and supervisors in an interprofessional academic primary care clinic. The study utilized a prospective cohort with control group. Intervention patients (N = 104) were matched with controls (N = 104) from the same provider’s panel using propensity scores based on age, gender, risk predictors, and prior utilization patterns. Clinical outcomes and subsequent utilization patterns were compared prior to and up to 6 months following the conference. In terms of utilization, intervention patients demonstrated increased visits with primary care team members (p = .0002) compared with controls, without a corresponding increase in the number of primary care providers’ visits. There was a trend towards decreased urgent care and emergency visits (p = .07) and a significant decrease in the rate of hospitalizations (p = .04). Patients with poorly-controlled hypertension saw significant decreases in mean systolic blood pressure from 167 to 146 mm Hg. However, there were no differences between the intervention and control groups. Intervention patients with diabetes demonstrated a nonsignificant trend towards decreased hemoglobin A1c from 9.8 to 9.4, when compared with controls. Interprofessional case conferences have potential to improve care coordination and may be associated with improved disease management, decreased unplanned care, and overall reduced hospitalizations. Implications Practice: An interprofessional case conference for high-risk or high-need patients can change primary care dynamics to increase team-based encounters, with trends toward decreased unplanned care and hospitalizations. Policy: Policymakers should look for opportunities to support the time and effort required to develop recurrent team-based interprofessional case conferences. Research: Future research should be aimed at further evaluating the impact of interprofessional case conferences in controlled trials, as well as at different institutions and systems of care. INTRODUCTION The number of patients with multiple chronic conditions in the USA is increasing (1). This population of potentially high-need, high-cost patients uses more health care resources and requires more coordination to improve appropriate utilization and quality outcomes (2, 3). Primary care providers (PCPs) may be ill-equipped to facilitate coordination of care for such patients (4). At the same time, there is increased emphasis on colocation of different professions as part of Patient-Centered Medical Home and related models, which emphasize practice redesign to improve access, care coordination, and self-management support, particularly of complex patients (5, 6). With these, there are more opportunities for interprofessional collaboration to address patients with complex needs that utilize disproportionate services, and more evidence is needed on effective models of care coordination, and their subsequent impacts on important patient health care outcomes (7, 8). In 2010, the Department of Veterans Affairs (VA) converted to a medical home-like model known as Patient-Aligned Care Teams (PACT) (9). To further opportunities for interprofessional collaboration in this new clinic model, we developed locally the PACT Interprofessional Care Update (PACT ICU) to provide a venue to teach and provide ongoing collaborative team-based care for high-need, high-cost patients. The PACT ICU is an interprofessional care conference with the dual aim of improving care of high-risk patients in primary care, while also offering an opportunity to provide effective team-based care coordination. Educational outcomes and dissemination efforts have been evaluated and are reported separately (10, 11). The aims of this study were to evaluate the impact of an interprofessional case conference on the following: (i) clinical contacts with PCPs and members of the primary care team; (ii) quality of care of chronic diseases; and (iii) utilization of urgent care, emergency department, and hospitalizations. We hypothesized that the quality of care measured by improvements in indices of chronic disease (i.e., hemoglobin A1C and blood pressure) would improve in the intervention group compared with controls, that clinical contacts with the core PACT primary care team would increase in comparison to controls, and that utilization of urgent care, emergency room, and hospitalizations would decrease in the intervention group compared with controls. METHODS Study design We performed a prospective case–control trial with propensity score matching. This study was conducted and written consistent with STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines for observational trials (12). The Institutional Review Boards from the Puget Sound Health Care System and Boise VA Medical Center approved the study protocol. The study was conducted over a 26 month period, with ongoing recruitment for 20 months of patients with a 6 month follow-up period. Setting and participants This study was conducted at an interprofessional Academic PACT primary care clinic based at a VA medical center, with 36 trainees from internal medicine, nurse practitioner (NP), ambulatory pharmacy, and postdoctoral psychology programs. Patients were selected from approximately 2,000 veterans cared for on internal medicine and NP trainee panels. The final study size was a convenience sample, based on the number of intervention patients sequentially presented at the PACT ICU conference, and their matched controls. Patients were included if they were presented at PACT ICU during the study period of January 1, 2013 to October 14, 2015. Patient selection was continual during the study period and facilitated by a clinic-based registered nurse (RN) care manager, who coordinated presenters. The RN coordinator gave presenting PCP trainees, internal medicine and NP residents, and a list of the top 5–10 highest risk patients on their panels, as predicted by the VA-based Care Assessment Need (CAN) registry score. The CAN score effectively estimates the probability of hospitalization or death in the next 90 days in primary care patients within the VA, based on sociodemographic factors, medical conditions, recent clinical data, and prior utilization (13). Each PCP trainee then selected one patient that he or she felt was most appropriate for PACT ICU discussion. Intervention The structure and content of the PACT ICU conference has been previously described (10, 14, 15). In brief, it is an hour-long conference, in which two PCP trainees present selected patients to an interprofessional team of trainees and supervisors from their primary care clinic. The PCPs each select one new patient to present; this is shared with PACT ICU team members via encrypted email for chart review 1 week prior to the conference. Team members include adult-gerontology NPs, ambulatory pharmacists, internal medicine physicians, psychologists, RNs, and social workers who are affiliated with the primary care clinic. As part of a rotating schedule, faculty supervisors from internal medicine, NP, pharmacy, and psychology facilitate the conference using the “EFECT” approach, specifically (a) Eliciting a patient-centered narrative, (b) Facilitating an interprofessional team discussion, (c) Evaluating clinical evidence, (d) Creating a shared care plan, and (e) Tracking outcomes. (16) This is based on an expert consensus clinical approach emphasizing a focus on the patient narrative, evidence-based practice, facilitated interprofessional coordination, and development of a care plan that is followed for progress. A faculty facilitator leads the conference, soliciting input from participating medicine residents and preceptors, NP trainees and supervisors, primary care team nurses, pharmacists, psychologists, and social workers. At the end of the conference, the care plan is documented in the medical record, with specific action items for the team to complete; it is shared electronically with all participating team members. PCP trainees present on average approximately every 6 months, but participate more frequently depending on clinic schedules. Tasks developed in the care plan are shared among team members, who help us to complete the assigned tasks. These may include further chart review and recommendations, patient outreach via phone or monitoring, or a prescheduled appointment. Oftentimes, conferences are used to facilitate “warm hand-offs” which are opportunistic care transfers between colocated members of the team, such as psychology or pharmacy colleagues. Previously presented patients are briefly reviewed to discuss progress with existing care plans. Variables For intervention patients and matched controls, characteristics of gender, age, predicted risk of death or hospitalization, and death within 6 months of presentation, during the intervention period, were recorded. Average of blood pressures obtained in primary care settings for patients with a diagnosis of hypertension were calculated from VA Corporate Data Warehouse (CDW) data for 6 months prior to study enrollment and during the 6 month study follow-up to evaluate the impact on blood pressure control of care plans and team involvement. Due to concern for spurious elevation, blood pressures from ambulatory surgery, episodic care, emergency department, and hospitalization settings were excluded. Diabetes control was evaluated for patients with a diagnosis of diabetes, using the most recent hemoglobin A1c value prior to the PACT conference, and the first value between 90 and 240 days after the conference. All lab work was processed through the same facility laboratory; data were accessed from the electronic health record via the CDW. The VA’s CDW is an often-used source of clinical data collected directly from the VA’s electronic health record and is updated nightly (17). Encounters with primary care providers and associated primary care team members (pharmacy, psychology, and nursing) were tracked, as were urgent care, emergency department, and hospitalizations for 6 months prior to study enrollment and during the 6 month study follow-up of the PACT ICU intervention. These data were also retrieved from the CDW. Power analyses indicate that the sample size attained produced an 80 per cent power to detect small-to-medium effect sizes in the ANOVA (interaction effect size of f = .13 and a main effect size of f = .28). These sizes were deemed appropriate to detect clinically relevant measures of change in all measures of utilization (i.e., provider visits, team visits, urgent/episodic care, emergency department visits, and hospitalizations). Data sources/measurement Data were obtained from the CDW and reviewed for accuracy by study authors. The CDW is a relational database of clinical information that interfaces directly with the VA electronic health record and is updated nightly. Procedures for documentation of encounters are closely adhered to by clinicians and regularly reviewed by administrative staff; during the building of the database query for this study, there was continual cross-checking with clinical faculty. Chart review and communication among the clinical staff, database manager, and statistician provided an iterative review process. Data for individual participants were extracted for a baseline period of 6 month prior study enrollment. These variables were collected for the 6 month follow-up period to the PACT ICU, to evaluate the impact of the care conference intervention. Statistical analyses Data were analyzed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). Patient baseline characteristics were compared with t-test and chi-squared analysis. Utilization was broken down into 2 month segments to test for trends. Changes in utilization were evaluated with repeated ANOVA tests using “proc GLMMIXED” function, selecting a Poisson distribution appropriate for count data. Utilization during pre-PACT ICU and post-PACT ICU presentation time periods was compared, as were trends in utilization, using 2 month blocks. To evaluate the utilization data for each patient presented in PACT ICU, one of the four patients not selected for presentation was assigned to be a control using a propensity score based on prior utilization, CAN score, and age (18). This technique of modeling the selection process using propensity scores can mitigate some of the bias introduced by allowing the PCP to select the patients to present. Of note, only patients that were not selected at any time for the PACT ICU intervention and that did not die during study period were eligible for selection as a control. Among eligible controls, the single patient with the highest propensity score was chosen as the propensity score–matched control resulting in a 1:1 match. RESULTS Participants and descriptive data During the observation period, 104 patients were presented in the PACT ICU; 104 controls were selected using the propensity score–matching process described above (Fig. 1). The patients studied tended to be older males, with a baseline 90 day risk of death or hospitalization as predicted by CAN score, of approximately 22 per cent (Table 1). Of note, there were no significant differences in gender, age, CAN score, or overall propensity score. There were no statistically significant differences in baseline utilization, although patients presented in PACT ICU conference displayed a trend of increased utilization prior to the conference (Table 2). Fig 1 View largeDownload slide Selection process for Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) patient participants and controls. 1Trainees include nurse practitioner and internal medicine residents. Over the study time period, panel sizes changed, but there were approximately 2,000 patients at any given time. 2Six PACT ICU intervention patients were dropped because they died within 6 months of the intervention date. 3Only the first presentation to PACT ICU was used (19 dropped). 4Fifty-one possible control patients were dropped because they died within the 6 months of intervention date and 114 were dropped because they were later presented to PACT ICU. 5Of 403 possible controls, 104 were selected based on 1:1 propensity match with PACT ICU intervention patients. Fig 1 View largeDownload slide Selection process for Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) patient participants and controls. 1Trainees include nurse practitioner and internal medicine residents. Over the study time period, panel sizes changed, but there were approximately 2,000 patients at any given time. 2Six PACT ICU intervention patients were dropped because they died within 6 months of the intervention date. 3Only the first presentation to PACT ICU was used (19 dropped). 4Fifty-one possible control patients were dropped because they died within the 6 months of intervention date and 114 were dropped because they were later presented to PACT ICU. 5Of 403 possible controls, 104 were selected based on 1:1 propensity match with PACT ICU intervention patients. Table 1 Demographics of controls and PACT ICU intervention participants at baseline Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 aMeasures included in the calculation of propensity score for matching. View Large Table 1 Demographics of controls and PACT ICU intervention participants at baseline Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 Controls PACT ICU intervention p -Value Gendera 90% male 88% male .55 Agea 69.2 (SD 12.6) 67.0 (SD 12.3) .49 CAN score (risk of death or hospitalization in the next 90 days)a 21% (SD 14%) 22% (SD 14%) .85 aMeasures included in the calculation of propensity score for matching. View Large Table 2 Number of visits at baseline (prior to PACT ICU conference) for participants and matched controls PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 These utilization measures were included in the calculation of the propensity score for matching. View Large Table 2 Number of visits at baseline (prior to PACT ICU conference) for participants and matched controls PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 PACT ICU patients Baseline (0–2 months prior) (n = 104) Matched controls Baseline (0–2 months prior) (n = 104) p-Value Mean (SD) Number of visits Mean (SD) Number of visits PCP visit 1.1 (1.0) 113 1.0 (1.1) 106 .35 PACT team visit 2.6 (3.2) 266 1.8 (2.3) 187 .09 Urgent/episodic care/ER 0.6 (1.1) 62 0.4 (0.7) 42 .44 Hospitalizations 0.2 (0.5) 22 0.2 (0.6) 23 .64 These utilization measures were included in the calculation of the propensity score for matching. View Large Utilization Both patients selected for PACT ICU presentation and matched controls from the same PCP tended to have increased PCP visits around the time of the conference (Fig. 2). Following the index date of the conference, both PACT ICU and control patients showed a trend to a decline in the number of visits over the 6 month follow-up period postintervention (p = .08). Overall, there was no significant difference between the number of PCP visits (Fig. 3) for the two groups (p = ns for trend). Conversely, there was an increase in PACT team encounters (e.g., pharmacy, behavioral health, and RN care managers) immediately preceding and following the PACT ICU conference, which was significantly different (p = .0002 for trend). This appeared to persist for at least 6 months, the follow-up period, following the conference intervention, after which it began to decrease. There was a nonsignificant trend towards decreased urgent care and emergency visits for the PACT ICU patients (Fig. 4) when compared with controls (p = .07), and a significantly decreased rate of hospitalizations (Fig. 5) over the 6 month follow-up period in the PACT ICU patients when compared with controls (p = .04). Fig 2 View largeDownload slide Number of Patient–Aligned Care Team (PACT) visits of intervention patients prior to and following presentation at PACT Interprofessional Care Update (ICU) conference, as well as those of matched controls. Post-intervention there were significant intervention effects (p = .0002). Patient encounters by pharmacy, nursing, social work, and psychology providers working with primary care providers (PCPs) in the same PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 2 View largeDownload slide Number of Patient–Aligned Care Team (PACT) visits of intervention patients prior to and following presentation at PACT Interprofessional Care Update (ICU) conference, as well as those of matched controls. Post-intervention there were significant intervention effects (p = .0002). Patient encounters by pharmacy, nursing, social work, and psychology providers working with primary care providers (PCPs) in the same PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 3 View largeDownload slide Number of primary care encounters of intervention patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference, as well as those of matched controls. Postintervention there were no significant intervention effects (p = ns). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 3 View largeDownload slide Number of primary care encounters of intervention patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference, as well as those of matched controls. Postintervention there were no significant intervention effects (p = ns). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 4 View largeDownload slide Number of emergency departments and urgent care visits of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were no significant effects (p = ns), but there was a nonsignificant trend towards decrease (p = .07). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 4 View largeDownload slide Number of emergency departments and urgent care visits of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were no significant effects (p = ns), but there was a nonsignificant trend towards decrease (p = .07). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 5 View largeDownload slide Number of hospitalizations of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were significant intervention effects (p = .04). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Fig 5 View largeDownload slide Number of hospitalizations of intervention and matched control patients prior to and following presentation at Patient–Aligned Care Team Interprofessional Care Update (PACT ICU) conference. Postintervention there were significant intervention effects (p = .04). Patient encounters by primary care providers (PCPs) in a PACT primary care clinic. PACT ICU intervention patients were presented in conference; controls were matched by propensity score from the same PCP and same date of conference. Quality of care In terms of quality of care, there were no significant differences between the PACT ICU patients and matched controls for hypertension or glycemic control. There were improvements in hypertension following the PACT ICU conference date, but these were present in both populations, suggesting regression to the mean. Looking specifically at poorly controlled hypertensive patients (those with a baseline systolic blood pressure of 150 mm Hg or greater, n = 25), the mean systolic blood pressure of patients in both groups was 167 mm Hg (SD 17.5) prior to the index date, decreasing to an average of 146 mm Hg (SD 18.1) in the follow-up period (p < .0001). There were trends towards differences at baseline among PACT ICU patients with diabetes and hemoglobin A1c >7.5 per cent (n = 20); the most recent value prior to the conference averaged 9.8 per cent (SD 2.0) with a postconference hemoglobin A1c of 9.4 per cent (SD 1.6, p = .34); similarly matched controls (n = 30) had overall better baseline hemoglobin A1c (mean 8.7%, SD 2.0) with no change following the index date of the conference (mean 8.7%, SD 2.0. p = .99). A post hoc power analysis indicates that given the observed effect D = .21 in the PACT ICU group, it would require twice the number of participants to have 80 per cent power to detect a significant difference between pre- and post-measures A1c or mm Hg measures of blood pressure. Other analyses In terms of mortality during the study period, 9.1 per cent of all potential controls died during the 6 month follow-up, compared with 3.9 per cent of patients in the intervention group, presented in the PACT ICU conference (p = .05; Table 1). Given concerns regarding selection bias as well as right censored utilization data, these patients were appropriately excluded from analysis if they died during the follow-up period. However, sensitivity analyses of utilization outcomes, including and excluding deceased patient utilization data, did not affect results (data not displayed). DISCUSSION These findings suggest that interprofessional case conferences such as PACT ICU may facilitate increased team involvement in patient care rather than PCP-only visits. The increased involvement by associated team members persisted for 6 month follow-up period, following participation in the PACT ICU conference intervention. The trends towards reduced urgent/emergent care services and an overall reduced rate of hospitalizations suggest that the intervention may have helped us to avert unplanned medical care. Although there was improvement in hypertension following the PACT ICU conference, this was not isolated to the intervention arm and likely represents limitations of sample size, regression to the mean, or effects of independent efforts to address poorly controlled hypertension. Similarly, there were no statistically significant differential effects between the two groups related to glycemic control; this was confounded by poorer baseline control in the intervention arm. Limitations Limitations of the current study include the observational nature of the data and the inherent bias in selecting higher risk patients for inclusion. Although such bias is problematic for statistical analysis and assigning causality to observed effects, we felt that it was essential for proper patient selection and provider buy-in. There is evidence that although risk predictors are increasingly useful, they still do not fully account for complexity, particularly when compared with estimation by the empaneled PCP (19). Other sources of bias were considered: the most prominent being regression to the mean, a common problem in pre- or post-cohort studies in which participants are selected during a period of increased risk or utilization (20). Such bias is common in studies of high-risk patient interventions (21). To address both regression to the mean and selection bias, we employed a propensity-matched control group mentioned above (22, 23). Using a propensity score partially based on prior utilization is likely correlated with subsequent utilization; however, this would bias any subsequent difference in utilization towards the null, making any finding of change more robust. Finally, the convenience sampling approach of patients in this evaluation yielded small numbers of patients with poorly controlled chronic conditions; accordingly, we were underpowered to detect meaningful differences. Unfortunately, other quality or process measures that are likely affected by the PACT ICU, such as opioid treatment, frequency of benzodiazepine use, management of symptoms of depression, and advanced directive completion, were not prespecified or easily measured in our data set. Future studies could take these measures into account. In addition, death was common in both arms due to the population studied and may have created a bias due to censoring. However, sensitivity analysis completed with and without intervention patients who died did not detect this bias. Further limitations include ongoing efforts to improve team-based care via mechanisms and training other than the PACT ICU. It is not uncommon for one intervention to take place as part of a larger cultural change related to team-based care, allowing “cointervention” bias to occur (21). Additional analyses by our study team suggest that there was an increase in within-team consults in the months following the PACT ICU conference for all patients. Although this supports improved collaboration, it may have decreased the relative effect size of the increase in encounters by PACT team members, compared with controls. As is common in VA-based studies, there were small numbers of females, which may allow for introduction of bias in the findings, and decreased generalizability. The trend towards difference in mortality between intervention patients and selected controls was of interest, but we were unable to determine whether this was due to selection bias or direct impact of improved care and coordination. Providers were specifically counselled to select patients that would benefit from the case conference and future care coordination; accordingly, they were encouraged not to select patients already on hospice, or with diseases so progressed they were unlikely to affect change by the primary care team. Consequently, any difference in mortality may be related to a selection bias. There is evidence that the common sense approach of adding clinical triage to automated risk prediction remains necessary to account for more complex and nuanced predictors (24), particularly in identifying high risk primary care patients most appropriate for case conferences (25). Finally, although our study was rooted in a practical application of proactive, prepared team–based care as advocated in the Chronic Care Model and implemented in Patient-Centered Medical Homes and Patient Aligned Care Teams (26), we did not have a guiding theoretical model to refine our specific implementation and evaluation. Future efforts to design such an intervention in different contexts might benefit from a theoretical model such as the Relational Coordination Framework (27), which emphasizes an improved understanding of how collaborative care can be supported to improve quality and efficiency of care, so important in this population. Interpretation Although this conference required significant effort (several trainees and supervisors from different professions meeting on a regular basis), these preliminary results suggest that it may be worth the investment. Investing time into collaboration on high-need, high-cost patients is an appropriate allocation of resources, which can help us to change utilization patterns appropriately for such patients. This has been identified as a particularly important modality for patients with comorbid mental illness commonly found in multimorbidity patients (28). Interpreting these data in the context of other studies, it appears that there are heterogeneous effects of different types of care models. More evidence supports the idea that interprofessional approaches, centered around primary care teams, are more effective at improving quality and utilization than external approaches, those based on care transitions, and those not featuring interprofessional teams (8). To the best of our knowledge, there has not been a comparison between internal (such as PACT ICU) and externally located care coordination interventions; this would be an interesting research question for future studies. Similarly, comparing outcomes of primary care team–based care coordination and specialty care coordination interventions would be of interest. Several traditional and innovative models of specialty care coordination models exist, but these tend to have a different focus than PACT ICU, which emphasizes longitudinal primary care team coordination. More traditional case conferences such as tumor board or interdisciplinary surgery specialties tend to focus on a singular set of decisions regarding a known or suspected disease state, instead of long-term coordination of broader primary care issues. Other specialty-primary care collaborative conferences may focus on educating and consulting with PCPs about a disease state by experts at a distance (29), or other conferences focusing on methods to improve transitions of care for hospitalized patients with specialty input (30). Integration of behavioral health providers into primary care clinics appears to have improved quality of care and decreased costs in a large integrated health care system (31). Case conferences such as PACT ICU provide a venue for behavioral health providers (in our study, psychologists) to collaborate with the rest of the care team, providing opportunities for better integration in the care of these patients in specific, and the team in general. Other authors have reported outcomes related to case conferences, but these are often limited by the lack of an appropriate control arm (21, 32). Given the impact of regression to the mean in such studies, it is important to identify a comparator population, whether through matched controls, or a randomization system which allows for PCP selection of appropriate patients. Effective team-based care requires training and education to allow for the team to develop trust, allow effective communication, and understand different roles and responsibilities. This is particularly important for trainees who have increased exposure to interprofessional education, but may not have explicit exposure to collaborative clinical care. The National Academy of Medicine has called for a dual approach, encouraging workplace-learning opportunities for interprofessional care as well as evaluation beyond educational impacts (33). Our intervention helps us to extend the evaluation of the impact of interprofessional educational interventions beyond learner satisfaction or knowledge, identifying impacts on participant behavior changes and subsequent patient impacts. Overall, the PACT ICU conference provides a practical approach to learning interprofessional care in a collaborative practice model. It allows PCPs to select appropriate patients and empowers the team to develop a shared care plan that can alter patient utilization. Generalizability Although the data from this study are limited to one location, the PACT ICU model has been successfully exported to other academic training clinics within the VA (11, 15). Potentially important elements include a designated care coordinator, interprofessional teams (often featuring behavioral health such as psychology), a facilitator to ensure a coordinated care plan is agreed upon, and means of following up on patient outcomes. However, challenges exist to support this model in other settings, particularly those that require fee-for-service billing to support such efforts (34). There are good examples of community-based programs, which often lack controlled evaluations, yet may be useful examples (35); also, guidance for developing such conferences is also in development. (36) Although support of value-based care or accountable care organization contracts may make this more feasible, new Medicare billing codes that support complex chronic patient care and interprofessional conferences allow this to expand into more traditional billing structures (37), particularly to support behavioral health integration (38). Compliance with Ethical Standards Conflict of Interest: None declared. Primary Data: The authors have full control of the primary data and can submit deidentified data (e.g., minus protected health identifiers, in keeping with VA IRB requirements) to the journal for review upon request. The findings in this report have not been previously published, nor are they being simultaneously submitted elsewhere. Preliminary results related to this project have been presented orally at Northwest Society of General Internal Medicine and National Society of General Internal Medicine; no publications resulted from these presentations. Ethical Approval: No animals were involved in or harmed by this research project. Informed Consent: Waived as part of approval from both Puget Sound Health Care System and Boise VA Medical Center Institutional Review Board Acknowledgments Time for this project was indirectly supported via the VA Office of Academic Affiliations Centers of Excellence in Primary Care Education. References 1. Gerteis J , Izrael D , Deitz D et al. Multiple Chronic Conditions Chartbook . Rockville, MD : Agency for Healthcare Research and Quality ; 2014 . 2. Hong CS , Siegel AL , Ferris TG . Caring for high-need, high-cost patients: What makes for a successful care management program ? Issue Brief (Commonw Fund) . 2014 ; 19 : 1 – 19 . Google Scholar PubMed 3. Blumenthal D , Chernof B , Fulmer T , Lumpkin J , Selberg J . Caring for high-need, high-cost patients – An urgent priority . n Engl j Med . 2016 ; 375 ( 10 ): 909 – 911 . Google Scholar CrossRef Search ADS PubMed 4. Osborn R , Moulds D , Schneider EC , Doty MM , Squires D , Sarnak DO . Primary care physicians in ten countries report challenges caring for patients with complex health needs . Health Aff (Millwood) . 2015 ; 34 ( 12 ): 2104 – 2112 . Google Scholar CrossRef Search ADS PubMed 5. American College of Physicians, American Association of Pediatrics, American Association of Famiy Physicians, American Osteopathic Association . Joint Principles of the Patient Centered Medical Home . Available at https://www.acponline.org/acp_policy/policies/joint_principles_pcmh_2007.pdf. Accessibility verified January 21, 2018. 6. Ladebue AC , Helfrich CD , Gerdes ZT , Fihn SD , Nelson KM , Sayre GG . The experience of Patient Aligned Care Team (PACT) members . Health Care Manage Rev . 2016 ; 41 ( 1 ): 2 – 10 . Google Scholar CrossRef Search ADS PubMed 7. Zulman DM , Grant RW . Transforming care for complex patients: Addressing interconnected medical, social, and behavioral challenges . j Gen Intern Med . 2016 ; 31 ( 3 ): 263 – 264 . Google Scholar CrossRef Search ADS PubMed 8. McCarthy KJ , Ryan J , Klein S. Models of Care for High-Need, High-Cost Patients: An Evidence Synthesis . In: Lorber D , ed. New York, NY: Commonwealth Fund ; 2015 . 9. Piette JD , Holtz B , Beard AJ et al. ; Ann Arbor PACT Steering Committee . Improving chronic illness care for veterans within the framework of the Patient-Centered Medical Home: experiences from the Ann Arbor Patient-Aligned Care Team Laboratory . Transl Behav Med . 2011 ; 1 ( 4 ): 615 – 623 . Google Scholar CrossRef Search ADS PubMed 10. Weppner WG , Davis K , Sordahl J et al. Interprofessional care conferences for high-risk primary care patients . Acad Med . 2016 ; 91(6) : 798 – 802 . Google Scholar CrossRef Search ADS 11. King I , Strewler A , Wipf JE et al. Translating innovation: Exploring dissemination of a unique case conference . J Interprof Educ Pract . 2017 ; 6 : 55 – 60 . Google Scholar CrossRef Search ADS 12. von Elm E , Altman DG , Egger M , Pocock SJ , Gøtzsche PC , Vandenbroucke JP ; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies . Ann Intern Med . 2007 ; 147 ( 8 ): 573 – 577 . Google Scholar CrossRef Search ADS PubMed 13. Wang L , Porter B , Maynard C et al. Predicting risk of hospitalization or death among patients receiving primary care in the Veterans Health Administration . Med Care . 2013 ; 51 ( 4 ): 368 – 373 . Google Scholar CrossRef Search ADS PubMed 14. Buu J , Fisher A , Weppner W , Mason B . Impact of patient aligned care team interprofessional care updates on metabolic parameters . Fed Pract . 2016 ; 33 ( 2 ): 44 – 48 . 15. Pearson M , Patel S , Kim C et al. Interprofessional Case Conferences: Rewards and Challenges . Paper presented at: Society of General Internal Medicine Annual Meeting . San Diego, California ; 2014 . 16. Bitton A , Pereira AG , Smith CS , Babbott SF , Bowen JL ; Society of General Internal Medicine Patient-Centered Medical Home Education Summit . The EFECT framework for interprofessional education in the patient centered medical home . Healthc (Amst) . 2013 ; 1 ( 3-4 ): 63 – 68 . Google Scholar CrossRef Search ADS PubMed 17. Corporate Data Warehouse (CDW); VA Informatics and Computing Infrastructure. Health Services Research and Development; U.S. Department of Veterans Affairs . Available at https://www.hsrd.research.va.gov/for_researchers/vinci/cdw.cfm. Accessibility verified December 15, 2017 . 18. Pirracchio R , Resche-Rigon M , Chevret S . Evaluation of the propensity score methods for estimating marginal odds ratios in case of small sample size . bmc Med Res Methodol . 2012 ; 12 : 70 . Google Scholar CrossRef Search ADS PubMed 19. Grant RW , Ashburner JM , Hong CS et al. Defining patient complexity from the primary care physician’s perspective: A cohort study . Ann Intern Med . 2011 ; 155 ( 12 ): 797 – 804 . Google Scholar CrossRef Search ADS PubMed 20. Morton V , Torgerson DJ . Regression to the mean: Treatment effect without the intervention . j Eval Clin Pract . 2005 ; 11 ( 1 ): 59 – 65 . Google Scholar CrossRef Search ADS PubMed 21. Lynch CS , Wajnberg A , Jervis R et al. Implementation Science Workshop: A novel multidisciplinary primary care program to improve care and outcomes for super-utilizers . j Gen Intern Med . 2016 ; 31 ( 7 ): 797 – 802 . Google Scholar CrossRef Search ADS PubMed 22. Cottone F , Efficace F , Apolone G , Collins GS . The added value of propensity score matching when using health-related quality of life reference data . Stat Med . 2013 ; 32 ( 29 ): 5119 – 5132 . Google Scholar CrossRef Search ADS PubMed 23. Rosenbaum PR , Rubin DB . The central role of the propensity score in observational studies for causal effects . Biometrika . 1983 ; 70(1) : 41 – 55 . Google Scholar CrossRef Search ADS 24. Hwang AS , Atlas SJ , Hong J et al. Defining team effort involved in patient care from the primary care physician’s perspective . j Gen Intern Med . 2017 ; 32 ( 3 ): 269 – 276 . Google Scholar CrossRef Search ADS PubMed 25. Watts B , Lawrence R , Carter C et al. How Risky is Watching and Waiting? A Pilot Study Combining Clinical Triage and Automated Risk Modeling to Enhance Allocation of Ambulatory Intensive Case Management Resources for High-Risk Patients . Poster presented at: Society of General Internal Medicine . Hollywood, Florida ; 2016 . 26. Wagner EH , Coleman K , Reid RJ et al. Guiding Transformation: How Medical Practices Can Become Patient-Centered Medical Homes . New York, NY: The Commonwealth Fund , February 2012 . 27. McDonald KM , Sundaram V , Bravata DM et al. Care coordination . Vol 7. In: Shojania KG , McDonald KM , Wachter RM , Owens DK , eds. Closing the Quality Gap: A Critical Analysis of Quality Improvement Strategies. Technical Review 9 (Prepared by the Stanford University-UCSF Evidence-based Practice Center under contract 290-02-0017) . AHRQ Publication No. 04(07)-0051-7. Rockville, MD : Agency for Healthcare Research and Quality ; June 2007 . 28. Ward MC , Miller BF , Marconi VC , Kaslow NJ , Farber EW . The role of behavioral health in optimizing care for complex patients in the primary care setting . j Gen Intern Med . 2016 ; 31 ( 3 ): 265 – 267 . Google Scholar CrossRef Search ADS PubMed 29. Department of Veterans Affairs . Fact Sheet: Specialty Care Access Network-Extension for Community (SCAN-ECHO) . Available at https://www.va.gov/HEALTH/docs/Specialty-Care-Access-Network.pdf, Accessibility verified November 14, 2017 . 30. ClinicalTrials.gov . Integrating Care After Exacerbations of COPD (InCasE) , PI David Au; Identifier NCT02021955. Available at https://clinicaltrials.gov/ct2/show/NCT02021955. Accessibility verified November 14, 2017 . 31. Reiss-Brennan B , Brunisholz KD , Dredge C et al. Association of integrated team-based care with health care quality, utilization, and cost . Jama . 2016 ; 316 ( 8 ): 826 – 834 . Google Scholar CrossRef Search ADS PubMed 32. Ritchie C , Andersen R , Eng J et al. Implementation of an interdisciplinary, team-based complex care support health care model at an Academic Medical Center: Impact on health care utilization and quality of life . Plos One . 2016 ; 11 ( 2 ): e0148096 . Google Scholar CrossRef Search ADS PubMed 33. Institute of Medicine: Measuring the Impact of Interprofessional Education on Collaborative Practice and Patient Outcomes . Institute of Medicine of the National Academies: Washington, DC ; 2015 . https://www.ncbi.nlm.nih.gov/books/NBK338360/ 34. Hong CS , Abrams MK , Ferris TG . Toward increased adoption of complex care management . n Engl j Med . 2014 ; 371 ( 6 ): 491 – 493 . Google Scholar CrossRef Search ADS PubMed 35. Warning W II , Wood W , Letcher A et al. Working With the Super-Utilizer Population: The Experience and Recommendations of Five Pennsylvania Programs ; 2014 . Available at http://www.aligning4healthpa.org/pdf/High_Utilizer_report.pdf. Accessibility verified January 21, 2018. 36. O’Brien BC , Patel SR , Pearson M et al. 12 Tips for Delivering Successful Interprofessional Case Conferences . Med Teach . 2017 ; 39 ( 12 ): 1214 – 1220 . Google Scholar CrossRef Search ADS PubMed 37. Center for Medicare and Medicaid Services . Chronic Care Management Services . Department of Health and Human Services ; 2016 . Available at https://www.cms.gov/About-CMS/Agency-Information/OMH/equity-initiatives/chronic-care-management.html. Accessibility verified January 21, 2018. 38. Press MJ , Howe R , Schoenbaum M et al. Medicare payment for behavioral health integration . n Engl j Med . 2017 ; 376 ( 5 ): 405 – 407 . Google Scholar CrossRef Search ADS PubMed Published by Oxford University Press on behalf of the Society of Behavioral Medicine 2018. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Journal

Translational Behavioral MedicineOxford University Press

Published: May 23, 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