Abstract
Introduction:
As care teams adopt team-based models of care, it is important to examine the reach of interdisciplinary services, such as pharmacists providing comprehensive medication management (CMM). This study examined the reach of pharmacist-delivered CMM in the first 10 months of a population health-focused primary care transformation (PCT).
Methods:
Using electronic health record data, descriptive statistics (counts and percentages, as well as means and standard deviations) were quantified to summarize the patients who received CMM in two PCT pilot clinics pre- and post-PCT.
Results:
Patients who had at least one CMM visit increased from 554 during the pre-PCT window to 880 during the post-PCT window. However, when adjusted for the increased pharmacist full-time equivalents (FTE) included as part of the PCT, 462 and 330 patients/FTE were seen in the pre- vs post-PCT periods, respectively. When calculating the percentage of patients who received CMM, this increased from 2.3% of all primary care patients seen in the two pilot clinics before the PCT began to 4.4% after the PCT was implemented. Most patient demographics remained largely the same between the pre- and post-PCT periods. However, CMM patients seen in the post-PCT period had more medication therapy problems across all medication therapy problem categories compared to patients in the pre-PCT period. Additionally, patients receiving CMM had significantly more conditions and medications and higher hospitalizations and emergency department use compared to the general clinic population.
Conclusions:
Reach is an important implementation outcome to determine the representativeness of individuals participating in a given service. This study illustrates that pharmacists providing CMM see complex patients with a high propensity for medication therapy problems. However, opportunities exist to improve the reach of CMM and, in turn, enhance team-based care.
Keywords: primary care, implementation science, medication therapy management
Background
While medications play a pivotal role in improving the health and wellbeing of patients, they are also becoming increasingly complex to manage.1 In primary care, an average office visit lasts 20 minutes, which is often not enough time for a provider to address all of a patient’s medical and medication-related needs.2 Therefore, given their expertise in pharmacotherapy and polypharmacy, pharmacists embedded into primary care teams are well positioned to assist in managing patients’ medications and helping patients achieve their medication-related goals. In fact, when pharmacists deliver comprehensive medication management (CMM) in the primary care setting, it has been shown to reduce health care costs,3 improve clinical outcomes,4-6 and improve patient and provider satisfaction7,8 – all aspects of the Quadruple Aim.9
As health systems undergo primary care transformation (PCT) to become more efficient and effective in care delivery, many PCT efforts are focused on team-based models of care.10-12 Over the years, numerous health systems and clinics around the world have undergone PCT to transition to a team-based Patient-Centered Medical Home (PCMH) model.2-5 Having already achieved medical home certification, a health system in Minnesota, M Health Fairview, is now implementing a population health-focused PCT. However, a question that remains when implementing these new models of care is the reach of interdisciplinary team members, such as pharmacists providing CMM. Reach, which is defined as the proportion and characteristics of individuals who receive an intervention, is a key element of implementation science.13,14 It addresses the representativeness of a study sample by examining the similarity or differences between those who participate in an intervention compared to those who are eligible but do not, and can be evaluated using either qualitative or quantitative methods.14,15 It is important to examine reach because those who receive a particular intervention may not always be those who are most in need.13 Thus, to maximize the benefits of CMM to support care teams and optimize medication use, there is a need to evaluate the reach of these services in new care models. The objective of this study, therefore, was to evaluate the reach of pharmacists providing CMM in a team-based, population health-focused PCT.
Methods
Setting
M Health Fairview is an integrated health system with 8 hospitals and 40 primary care clinics located across Minnesota and Wisconsin. In May 2019, they implemented a population health-focused PCT in two clinics with plans to gradually expand the PCT across all primary care clinics in the coming years. To support a population health model, as part of the PCT, primary care patients are placed into one of five strata, called service bundles, based on their complexity of care. Appointment lengths and services are then tailored to each service bundle. M Health Fairview uses a hybrid approach to assign patients to service bundles. Data from the electronic health record (EHR) are used to assign patients to a service bundle based on pre-determined criteria. However, using their clinical judgement, providers can reassign patients to different service bundles. While the bundle criteria continue to evolve, generally, patients in service bundle 1 are those who decline a relationship with a primary care provider (PCP) and/or patients who have not been seen by primary care for three or more years. Those in service bundle 2 have no active medical issues or they may have chronic conditions that are well controlled. Patients in bundle 3 have multiple chronic conditions that are not at goal. Those in service bundle 4 generally have a mental health diagnosis that is not well controlled and/or have social barriers to health. Finally, service bundle 5 are high-risk patients and high utilizers of the health care system. Given the magnitude of this initiative, there were numerous dissemination, implementation process, integration, capacity building, and scale-up strategies that were employed.16 For example, the system had months of design meetings with various care team members, including a pharmacist providing CMM, to get their input on PCT design and processes. There were also changes made to the EHR to communicate service bundle assignments and appointment scheduling. Daily workstation huddles were also initiated to discuss any implementation or care delivery issues that arise.
CMM at M Health Fairview
M Health Fairview has a robust CMM program of 45 pharmacists providing CMM within 52 primary care and specialty clinics. When pharmacists perform CMM, they meet one-on-one with patients to assess all of their medications to ensure each one is indicated, effective, safe, and that the patient is taking the medication as intended (i.e., adherent).17 If the pharmacist identifies any medication therapy problems (MTPs) during that visit, an individualized care plan is developed in consultation with the patient and the rest of the care team. The pharmacist then follows up and continues to meet with the patient to ensure they are achieving their desired outcomes. Pharmacists follow the taxonomy of MTPs developed by the Pharmacy Quality Alliance (PQA).18 This taxonomy outlines four medication-related needs (indication, effectiveness, safety, and adherence) with related MTPs (e.g., unnecessary medication therapy, dose too low). As part of the PCT, pharmacist full-time equivalents (FTEs) were increased at the two PCT pilot clinics from 0.4 and 0.8 to 1 and 1.8 FTE, respectively. While there are several different methods employed by M Health Fairview to identify CMM patients (e.g., patients can self-refer, patients discharged from the hospital who are high-risk for a readmission and meet specified criteria5 are automatically referred), the majority of patients are identified through provider referral. Provider referral remained the primary mechanism for identifying CMM patients after the initiation of the PCT, with the exception of service bundle 5. The pharmacist was considered integral to the care for those patients and therefore, schedule permitting, the pharmacist was included in the initial visits for all bundle 5 patients.
Study design and data source
This study used a single group, pre-test post-test study design to compare patients receiving CMM services at the two PCT pilot clinics. EHR data was obtained for patients utilizing CMM services at these two clinics. These data included all services provided by the health system, both from the pilot clinics as well as any additional outpatient, inpatient, emergency department, laboratory services, and CMM delivery provided by the health system at other clinics not participating in the PCT. The PCT began May 20, 2019; therefore, the post-PCT period was defined as May 20, 2019 to March 13, 2020. March 13th was selected as the cut-off date because care delivery changed significantly after that date due to changes made within the health system to address the COVID-19 pandemic. Given the potential for seasonal differences in health service utilization, the same time period in the year prior to the PCT (May 20, 2019 to March 13th, 2019) was defined as the pre-period.
Reach was analyzed from three perspectives: (1) those who received CMM in the pre-PCT period versus those who received CMM in the post-PCT period, (2) those who received CMM in the post-PCT period versus those who were scheduled to receive CMM but did not result in a completed visit (i.e., no-shows), and (3) those who received CMM in the post-PCT period versus patients who were seen by primary care in the post-PCT period but did not receive CMM.
Most CMM visits are generated from referrals placed by providers who use their discretion and clinical judgement to determine who would benefit from CMM. Therefore, when defining patients eligible for CMM, all primary care patients could be considered candidates for CMM or eligible patients could be defined as those with a scheduled CMM visit (e.g., a provider or payer deemed them eligible). Thus, these two different approaches were applied when calculating the percentage of patients reached by CMM.
The University of Minnesota Institutional Review Board approved this study.
Patient population
Patients were included in analysis if they had a scheduled CMM visit during the pre- or post-PCT period. To evaluate the patients who were seen by primary care in the post-PCT period but did not receive CMM, patients were included if they had at least one visit with a primary care provider (office visit, phone visit, or virtual visit) during the post-PCT period.
Patients were excluded if they were deceased, missing a service bundle assignment, or if CMM documentation was incomplete. Patients were also excluded if they resided in a long-term care facility because there is a different pharmacist outside of the PCT providing CMM in this setting.
Analysis
Descriptive statistics were calculated including counts and percentages for categorical variables and means and standard deviations for any continuous variables. Bivariate statistics were used to examine differences across cohorts. T-tests were used to compare continuous variables and chi-square tests were used to compare categorical variables. An alpha of 0.05 or lower was deemed significant for all statistical analyses. Descriptive characteristics, such as sex, age, and race were assessed for each cohort. In addition, variables such as health care utilization (e.g., number of emergency room visits) and number and type of MTPs were analyzed as a measure of patient complexity. Using the Elixhauser list of chronic conditions, select conditions were chosen to examine the prevalence of certain chronic conditions in the population.19 Data were analyzed using SAS v 9.4 software.20
Results
Delivery of CMM pre vs post-PCT
The number of patients who received CMM increased from 554 in the pre-PCT period to 880 in the post-PCT period (Table 1). However, when adjusted for the increased pharmacist FTE that occurred because of the PCT, this equated to 462 patients seen per pharmacist FTE in the pre-PCT period compared to 330 patients seen in the post-PCT period. Patient characteristics such as sex, age, and race did not change significantly between the pre-PCT period to the post-PCT period nor did the occurrence of specific conditions, such as diabetes and hypertension. Interestingly, there was a significant increase in the mean number of conditions CMM patients had in the pre- vs post-PCT period (15.3 vs 18.3; p-value<0.001), yet there was no significant difference in the mean number of medications (14 in the pre-PCT period vs 14.2 in the post-PCT period; p-value=0.5279). There was no significant difference in health care utilization among CMM patients in the pre-PCT period vs the post-PCT period in terms of mean count of hospitalizations (0.4 vs 0.4; p-value=0.521) or emergency department (ED) visits (0.4 vs 0.5; p-value=0.399), yet there was an increase in the number of in-person PCP visits. Table 2 illustrates the characteristics of patients seen in the post-PCT period broken down by service bundle.
Table 1:
Characteristics of patients who received comprehensive medication management pre- and post-PCT
| Characteristic | Pre-PCT (n= 554) |
Post-PCT (n= 880) |
p-value |
|---|---|---|---|
| Sex n (%) | 0.9844 | ||
| Female | 300 (54.2%) | 477 (54.2%) | |
| Age n (%) | 0.1184 | ||
| 18-34 | 15 (2.7%) | 43 (4.9%) | |
| 35-50 | 76 (13.7%) | 122 (13.9%) | |
| 51-65 | 217 (39.2%) | 308 (35.0%) | |
| 65+ | 246 (44.4%) | 407 (46.3%) | |
| Race n (%) | 0.2470 | ||
| White | 456 (82.3%) | 704 (80.0%) | |
| Black or African American | 44 (7.9%) | 78 (8.9%) | |
| Asian | 28 (5.1 %) | 54 (6.1%) | |
| Other | 7 (1.3%) | 10 (1.1%) | |
| Unknown | 19 (3.4%) | 34 (3.9%) | |
| Mean number of conditions (SD) | 15.3 (9.55) | 18.3 (11.8) | <0.001 |
| Select Elixhauser Conditions | |||
| Diabetes | 127 (22.9%) | 238 (27.0%) | 0.0811 |
| Hypertension | 134 (24.2%) | 234 (26.6%) | 0.3103 |
| Chronic pulmonary disease | 118 (21.3%) | 196 (22.3%) | 0.6644 |
| Depression | 122 (22.0%) | 214 (24.3%) | 0.3175 |
| Obesity | 202 (36.5%) | 323 (36.7%) | 0.9261 |
| Mean number of medications (SD) | 14 (8.14) | 14.2 (8.15) | 0.5279 |
| Number of in-person PCP visits n (%) | <0.001 | ||
| 0 | 216 (39.0%) | 237 (26.9%) | |
| 1 | 58 (10.5%) | 126 (14.3%) | |
| 2 | 65 (11.7%) | 168 (19.1%) | |
| 3 | 60 (10.8%) | 118 (13.4%) | |
| 4+ | 155 (28.0%) | 231 (26.3%) | |
| Mean count of hospitalizations (SD) | 0.4 (1) | 0.4 (1.01) | 0.5214 |
| Mean count of ED visits (SD) | 0.4 (1.06) | 0.5 (1.24) | 0.3986 |
PCT primary care transformation, PCP primary care provider, ED emergency department
Table 2.
Characteristics of patients who received comprehensive medication management in the post-PCT period across the different service bundles
| Characteristic | SB1 (n=7) |
SB2 (n=49) |
SB3 (n=649) |
SB4 (n=119) |
SB5 (n=56) |
|---|---|---|---|---|---|
| Sex n (%) | |||||
| Female | 3 (42.9%) | 23 (46.9%) | 354 (54.6%) | 61 (51.3%) | 36 (64.3%) |
| Age n (%) | |||||
| 18-34 | 2 (28.6%) | 8 (16.3%) | 22 (3.4%) | 6 (5.0%) | 5 (8.9%) |
| 35-50 | 0 (0%) | 8 (16.3%) | 82 (12.6%) | 29 (24.4%) | 4 (7.1%) |
| 51-65 | 3 (42.9%) | 24 (49.0%) | 215 (33.1%) | 48 (40.3%) | 18 (32.1%) |
| 65+ | 2 (28.6%) | 9 (18.4%) | 330 (50.8%) | 36 (30.3%) | 29 (51.8%) |
| Race n (%) | |||||
| White | 3 (42.9%) | 36 (73.5%) | 518 (79.8%) | 99 (83.2%) | 48 (85.7%) |
| Black or African American | 2 (28.6%) | 5 (10.2%) | 57 (8.8%) | 11 (9.2%) | 3 (5.4%) |
| Asian | 2 (28.6%) | 2 (4.1%) | 41 (6.3%) | 7 (5.9%) | 2 (3.6%) |
| Other | 0 (0%) | 0 (0%) | 6 (0.9%) | 2 (1.7%) | 2 (3.6%) |
| Unknown | 0 (0%) | 6 (12.2%) | 27 (4.2%) | 0 (0%) | 1 (1.8%) |
| Mean number of conditions (SD) | 12.1 (14.8) | 8.4 (5.6) | 18.8 (11.1) | 12.4 (8.7) | 34 (12.1) |
| Specific Elixhauser Conditions | |||||
| Diabetes | 2 (28.6%) | 13 (26.5%) | 172 (26.5%) | 41 (34.5%) | 10 (17.9%) |
| Hypertension | 2 (28.6%) | 15 (30.6%) | 171 (26.3%) | 31 (26.05%) | 15 (26.8%) |
| Chronic pulmonary disease | 1 (14.3%) | 10 (20.4%) | 150 (23.1%) | 25 (21.0%) | 10 (17.9%) |
| Depression | 3 (42.9%) | 12 (24.5%) | 161 (24.8%) | 28 (23.5%) | 10 (17.9%) |
| Obesity | 4 (57.1%) | 21 (42.9%) | 232 (35.7%) | 45 (37.8%) | 21 (37.5%) |
| Mean number of medications (SD) | 9.4 (9) | 7.1 (4.37) | 14.6 (7.65) | 11.3 (6.5) | 23.3 (10.18) |
| Number of in-person PCP visits n (%) | |||||
| 0 | 5 (71.4%) | 12 (24.5%) | 182 (28.0%) | 38 (31.9%) | 0 (0%) |
| 1 | 1 (14.3%) | 18 (36.7%) | 85 (13.1%) | 21 (17.6%) | 1 (1.8%) |
| 2 | 0 (0%) | 13 (26.5%) | 127 (19.6%) | 24 (20.2%) | 4 (7.1%) |
| 3 | 0 (0%) | 4 (8.2%) | 87 (13.4%) | 14 (11.8%) | 13 (23.2%) |
| 4+ | 1 (14.3%) | 2 (4.1%) | 168 (25.9%) | 22 (18.5%) | 38 (67.9%) |
| Mean count of hospitalizations (SD) | 0.7 (1.5) | 0.1 (0.32) | 0.4 (1.09) | 0.1 (0.25) | 0.8 (1.1) |
| Mean count of ED visits (SD) | 0.6 (0.79) | 0.1 (0.44) | 0.5 (1.31) | 0.2 (0.62) | 1.1 (1.67) |
| Number of CMM visits n (%) | |||||
| 1 | 2 (28.6%) | 22 (44.9%) | 261 (40.2%) | 44 (37.0%) | 13 (23.1%) |
| 2 | 1 (14.3%) | 19 (38.8%) | 165 (25.4%) | 31 (26.1%) | 8 (14.3%) |
| 3 | 3 (42.9%) | 6 (12.2%) | 90 (13.9%) | 15 (12.6%) | 13 (23.2%) |
| 4+ | 1 (14.3%) | 2 (4.1%) | 133 (20.5%) | 29 (24.4%) | 22 (39.3%) |
| Mean number of MTPs identified (SD) | 2 (1.83) | 1.8 (1.16) | 2.8 (1.74) | 2.4 (1.75) | 4 (1.49) |
| MTP Classification n (%) | |||||
| Indication | |||||
| Unnecessary Drug Therapy | 1 (14.3%) | 6 (12.2%) | 164 (25.3%) | 24 (20.2%) | 21 (37.5%) |
| Needs additional therapy | 3 (42.9%) | 29 (59.2%) | 441 (68.0%) | 80 (67.2%) | 52 (92.9%) |
| Effectiveness | |||||
| Different Drug Needed | 0 (0%) | 7 (14.3%) | 157 (24.2%) | 30 (25.2%) | 21 (37.5%) |
| Dose too low | 3 (42.9%) | 19 (38.8%) | 334 (51.4%) | 57 (47.9%) | 38 (67.9%) |
| Safety | |||||
| Adverse drug reaction | 2 (28.6%) | 11 (22.4%) | 240 (37.0%) | 36 (30.3%) | 31 (55.4%) |
| Dose too high | 1 (14.3%) | 11 (22.4%) | 198 (30.5%) | 25 (21.0%) | 33 (58.9%) |
| Adherence | 2 (28.6%) | 15 (30.6%) | 260 (40.1%) | 34 (28.6%) | 35 (62.5%) |
SB service bundle, PCP primary care provider, ED emergency department, CMM comprehensive medication management, MTP medication therapy problem
When evaluating CMM delivery pre- vs post-PCT (Table 3) there was no significant difference in the mean number of medications taken by CMM patients, although there was a significant difference in the mean number of MTPs identified (2.2 vs 2.7; p-value<0.001). When pharmacists identify an MTP during CMM, there is a taxonomy used to categorize MTPs.18 With the exception of the “different drug needed” MTP category, there was a statistically significant increase in all types of MTPs identified in the post-PCT period.
Table 3:
Characteristics of CMM delivery of patients who received CMM pre- and post-PCT
| Pre-PCT (N= 554) |
Post-PCT (N= 880) |
p-value | |
|---|---|---|---|
| Number of CMM visits n (%) | 0.1668 | ||
| 1 | 245 (44.2%) | 342 (38.9%) | |
| 2 | 120 (21.7%) | 224 (25.5%) | |
| 3 | 71 (12.8%) | 127 (14.4%) | |
| 4+ | 118 (21.3%) | 187 (21.3%) | |
| Mean number of MTPs identified (SD) | 2.2 (1.73) | 2.7 (1.75) | <0.001 |
| MTP Classification n (%) | |||
| Indication | |||
| Unnecessary Drug Therapy | 76 (13.7%) | 216 (24.5%) | <0.001 |
| Needs additional therapy | 329 (59.4%) | 605 (68.8%) | 0.0003 |
| Effectiveness | |||
| Different Drug Needed | 124 (22.4%) | 215 (24.4%) | 0.3739 |
| Dose too low | 251 (45.3%) | 451 (51.3%) | 0.0284 |
| Safety | |||
| Adverse drug reaction | 161 (29.1%) | 320 (36.4%) | 0.0043 |
| Dose too high | 140 (25.3%) | 268 (30.5%) | 0.0341 |
| Adherence | 167 (30.1%) | 346 (39.3%) | 0.0004 |
CMM comprehensive medication management, PCT primary care transformation, MTP medication therapy problem
Patients without a completed CMM visit
When comparing patients in the post-PCT period who were scheduled for CMM but did not result in a completed visit vs patients who received CMM (Table 4), almost half of CMM patients were older than 65 years, yet among those who did not complete CMM, 24.4% were older than 65 years. Other than obesity (36.6% in CMM patients vs 43.6% in patients without a completed CMM visits; p-value=0.031), there was not a significant difference in specific Elixhauser conditions between CMM patients and those without a completed CMM visit. There also was no significant difference in the mean count of ED visits (0.5 vs 0.4; p-value=0.530) between patients who were seen by CMM vs those who did not have a completed visit. However, there was a statistically significant difference in number of conditions (18.3 vs 15.3; p-value<0.001) and mean number of medications (14.2 vs 10.7; p-value<0.001) between the two groups. There was also a significant difference in mean count of hospitalizations (0.4 vs 0.3; p-value=0.049).
Table 4.
Characteristics of patients in the post-PCT period who received CMM vs those that were scheduled for CMM but did not complete a CMM visit vs all patients ≥ 18 years-old who were seen by primary care during the post-PCT period and did not receive CMM
| Characteristic | Completed CMM (n=880) |
Did not complete CMM (N=319) |
p-value | All Post-PCT (N=20130) |
p-value |
|---|---|---|---|---|---|
| Sex n (%) | 0.9166 | 0.2385 | |||
| Female | 477 (54.2%) | 174 (54.5%) | 11462 (56.9%) | ||
| Age n(%) | <0.0001 | <0.0001 | |||
| 18-34 | 43 (4.9%) | 29 (9.1%) | 5410 (26.9%) | ||
| 35-50 | 122 (13.9%) | 92 (28.8%) | 6162 (30.6%) | ||
| 51-65 | 308 (35.0%) | 120 (37.6%) | 5218 (25.9%) | ||
| 65+ | 407 (46.3%) | 78 (24.4%) | 3340 (16.6%) | ||
| Race n(%) | 0.5569 | <0.0001 | |||
| White | 704 (80.0%) | 256 (80.3%) | 16221 (80.6%) | ||
| Black or African American | 78 (8.9%) | 29 (9.1%) | 1165 (5.8%) | ||
| Asian | 54 (6.1%) | 13 (4.1%) | 1346 (6.7%) | ||
| Other | 10 (1.1%) | 4 (1.3%) | 117 (0.6%) | ||
| Unknown | 34 (3.9%) | 17 (5.3%) | 1281 (6.4%) | ||
| Service bundle assignment | <0.0001 | <0.0001 | |||
| 1 | 7 (0.8%) | 2 (0.6%) | 64 (0.3%) | ||
| 2 | 49 (5.6%) | 43 (13.5%) | 9915 (49.3%) | ||
| 3 | 649 (73.8%) | 224 (70.2%) | 6761 (33.6%) | ||
| 4 | 119 (13.5%) | 47 (14.7%) | 3372 (16.8%) | ||
| 5 | 56 (6.4%) | 3 (0.9%) | 18 (0.1%) | ||
| Mean number of conditions (SD) | 18.3 (11.8) | 15.3 (9.29) | <0.0001 | 8.1 (6.69) | <0.0001 |
| Specific Elixhauser Conditions | |||||
| Diabetes | 238 (27.0%) | 89 (27.90%) | 0.7691 | 2719 (13.51%) | <0.0001 |
| Hypertension | 234 (26.6%) | 82 (25.71%) | 0.7584 | 4624 (22.97%) | 0.0127 |
| Chronic pulmonary disease | 196 (22.3%) | 68 (21.32%) | 0.7240 | 3536 (17.57%) | 0.0003 |
| Depression | 214 (24.3%) | 91 (28.53%) | 0.1392 | 4653 (23.11%) | 0.4075 |
| Obesity | 323 (36.7%) | 139 (43.57%) | 0.0308 | 6289 (31.24%) | 0.0006 |
| Mean number of medications (SD) | 14.2 (8.15) | 10.7 (7.78) | <0.0001 | 4.7 (4.63) | <0.0001 |
| Number of in-person PCP visits n(%) | <0.0001 | <0.0001 | |||
| 0 | 237 (26.9%) | 116 (36.4%) | 500 (2.5%) | ||
| 1 | 126 (14.3%) | 62 (19.4%) | 12467 (61.9%) | ||
| 2 | 168 (19.1%) | 66 (20.7%) | 4272 (21.2%) | ||
| 3 | 118 (13.4%) | 35 (11.0%) | 1726 (8.6%) | ||
| 4+ | 231 (26.3%) | 40 (12.5%) | 1165 (5.8%) | ||
| Mean count of hospitalization (SD) | 0.4 (1.01) | 0.3 (0.74) | 0.0491 | 0.1 (0.29) | <0.0001 |
| Mean count of ED visits (SD) | 0.5 (1.24) | 0.4 (1.1) | 0.5303 | 0.2 (0.61) | <0.0001 |
CMM comprehensive medication management, PCT primary care transformation, PCP primary care provider, ED emergency department
Characteristics of all clinic patients not receiving CMM
Finally, when examining patients who received CMM versus all other clinic patients who were seen by primary care but did not receive CMM in the post-PCT period (Table 4 and Appendix 1), CMM patients were older than the general clinic population (46.3% vs 16.6% > 65 years-old). In addition, a higher proportion of Black or African American patients were seen by CMM compared to the general clinic population (8.9% vs 5.8%). The majority of CMM patients were in service bundle 3 (73.8%) while the majority of clinic patients who did not receive CMM were in service bundle 2 (49.3%). CMM patients had significantly more conditions compared to the general clinic population (18.3 vs 8.1; p-value<0.001), as well as medications (14.2 vs 4.7; p-value<0.001). There was also a statistically significant difference in health care utilization between CMM patients and the general clinic population, including number of in-person PCP visits, mean count of hospitalizations (0.4 vs 0.1; p-value<0.001) and mean count of ED visits (0.5 vs 0.2; p-value <0.001).
Overall reach of CMM
When examining the reach of CMM pre- vs post- PCT, there was a slight increase in the percent of patients who received CMM, both among patients scheduled for CMM and among all primary care patients. Patients seen for CMM among all patients scheduled for a CMM visit increased from 68.6% to 73.4% in the pre- vs post-PCT period, while CMM patients seen among all primary care patients also increased from 2.3% to 4.4% in the pre- vs post-PCT period.
Discussion
This study demonstrated the reach of CMM delivered in the first ten months of a team-based PCT focused on population health. While the reach of CMM nearly doubled among all primary care patients after the PCT began, this still equated to less than 5% of all primary care patients receiving CMM. Given that roughly half of all Americans take one or more medications21 and about 40% of older adults take five or more,22 it is likely that more patients could benefit from CMM. This work suggests that although a team-based, population health model did improve CMM reach, other strategies may be necessary to increase CMM delivery and subsequent benefits to both patients and care teams. These strategies may include different methods for CMM patient identification, enhanced education around care team roles, and increasing the number of pharmacists providing CMM.
Comparing patient characteristics of those who received CMM before the PCT started versus after illustrates any potential changes in reach that occurred as a result of the PCT. While characteristics of patients seen by CMM were quite similar before and after the PCT began, there was a significant increase in the number of MTPs pharmacists identified during a CMM visit after the PCT was initiated. There are a number of factors that may have contributed to this. First, with an emphasis on team-based care and an increase in pharmacist FTEs as part of the PCT, providers may have been more acutely aware of CMM and therefore referred patients with a higher propensity for MTPs. Second, as part of the PCT, CMM appointment lengths were tailored to each service bundle so that patients in higher service bundles had longer CMM appointment lengths. Therefore, the increased appointment lengths that now occur may allow the pharmacists more time to assess patients’ medications leading to the identification of more MTPs.
When comparing patients who received CMM to patients who were scheduled for CMM, but did not complete a CMM visit (i.e., no-shows), there was a difference in age between the two groups. Many patients who completed a CMM visit were over the age of 65 years, while over 75% of patients who did not complete a scheduled CMM visit were 65 years or younger. A potential reason for this could be that patients younger than 65 years are more likely to still be employed; therefore, taking time off work could be a barrier to completing CMM visits. In addition, patients who did not complete a CMM visit were less likely to have had an in-person primary care physician visit, suggesting that they could be less engaged in primary care in general. Lastly, those that did not complete a CMM visit had, on average, fewer medical conditions and medications, as well as fewer hospitalizations and ED visits than CMM patients, indicating that there may be less complex patients. From the perspective of reach, this is a positive finding because it illustrates that the most medically complex patients are receiving CMM. However, CMM is also a preventative service that can help decrease hospitalizations and improve patient outcomes, so further research is needed to determine if there is a difference in long-term health outcomes among patients who do not complete CMM and similar patients who do.
While pharmacists saw more patients in the post-PCT period compared to the pre-PCT period, when this was adjusted for the increase in pharmacist FTEs that occurred as part of the PCT, the number of patients seen per pharmacist FTE decreased. There are a number of reasons that may have contributed to this. For example, pharmacists are considered an integral member of the bundle 5 care team and, given these patients’ complexity, they require more time to assess and manage their medications. Therefore, seeing more complex patients, like bundle 5, may limit the pharmacists’ ability to see more patients. In addition, two new pharmacists were hired at the beginning of the PCT to support CMM in the pilot clinics. As these pharmacists were working to build their practices and grow their patient panels, this may have also limited reach in the early stages of the PCT. Additionally, enhanced team-based care requires a shift in thinking surrounding professional roles, relationships, and responsibilities.11 Therefore, more education to care team members about when to engage CMM may be necessary to maximize the reach of CMM.
Finally, examining reach in the early stages of implementation is important to inform the integration of interdisciplinary services, such as CMM, as PCT models are spread across health systems. For example, the results of this work illustrated that as service bundle assignment increased, the number of MTPs identified by pharmacists also increased. This suggests that as patient complexity increases, CMM may become a more necessary service and therefore care teams should consider the pharmacist as an essential member of the care team for these patients. Furthermore, pharmacists providing CMM saw less than 10% of all bundle 3 patients. As the PCT was designed, bundle 3 patients are patients with multiple chronic conditions who are not meeting their desired goals. Health systems should pay particular emphasis to this patient group and consider providing higher levels of CMM to these patients to ensure that their health does not advance to a higher service bundle, such as bundle 5. Therefore, there is potentially greater opportunity to expand the reach of CMM among these patients, which should be examined in this PCT and within other health systems.
Limitations
As with any observational study, the potential for unmeasured confounding to influence the relationships tested (e.g., severity of medical conditions and MTPs on number of CMM and PCP visits) cannot be ruled out. Additionally, a limitation of this study is that the data reflect patients who were seen at two clinics located in suburban Minneapolis. If clinics in other geographic regions had been included, patient characteristics and the reach of CMM may have been different. In addition, there are certain limitations to using EHR data that should be considered. For example, patient medication lists may contain medications the patient is no longer taking. To account for this, we quantified the number of medications ordered within primary care during the respective time period (pre-PCT or post-PCT period), plus an additional two months preceding the time period to allow for a 12-month prescribing window. However, this method does not account for any over-the-counter products a patient may be taking and may also contain acute medications (e.g., antibiotics). Finally, the post-PCT period was 10 months to avoid the confounding nature of COVID-19 and the disruptions to care delivery and patient health that ensued. However, if reach of CMM had been examined for a longer period of time following PCT implementation, there may have been differences in CMM patient characteristics and uptake of CMM.
Conclusion
Pharmacists providing CMM play a key role in optimizing medications and medication-related outcomes within health care teams. The results of this work illustrated that when a team-based, population health-focused PCT was implemented, there was a slight increase in the percent of patients reached by CMM. Although reach increased, there are likely more patients who would benefit from CMM services. Therefore, to maximize the benefits pharmacists can provide, further opportunities should be explored to enhance the reach of CMM to support health care teams in caring for complex patients and improving patient care.
Supplementary Material
Table 5.
Percent of patients reached by CMM pre- and post-PCT
| Pre-PCT | Post-PCT | |
|---|---|---|
| CMM patients seen among patients with a scheduled CMM visit | 68.6% | 73.4% |
| CMM patients seen among all primary care patients | 2.3% | 4.4% |
Funding
This research was supported by the Agency for Healthcare Research and Quality (AHRQ) and Patient-Centered Outcomes Research Institute (PCORI), grant K12HS026379 and the National Institutes of Health’s National Center for Advancing Translational Sciences, grant KL2TR002492. Additional support for MN-LHS scholars is offered by the University of Minnesota Office of Academic Clinical Affairs and the Division of Health Policy and Management, University of Minnesota School of Public Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of AHRQ, PCORI, or Minnesota Learning Health System Mentored Career Development Program (MN-LHS).
Footnotes
Conflict of interest
The authors have no conflicts of interest to disclose.
References
- 1.Cipolle RJ, Strand L, Morley P. Pharmaceutical Care Practice: The Patient-Centered Approach to Medication Management Services. 3rd ed. New York: McGraw-Hill Medical; 2012. [Google Scholar]
- 2.Linzer M, Bitton A, Tu S-P, et al. The End of the 15-20 Minute Primary Care Visit. J Gen Intern Med. 2015;30(11):1584–1586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Isetts BJ, Schondelmeyer SW, Artz MB, et al. Clinical and economic outcomes of medication therapy management services: the Minnesota experience. J Am Pharm Assoc JAPhA. 2008;48(2):203–211. [DOI] [PubMed] [Google Scholar]
- 4.Brummel AR, Soliman AM, Carlson AM, de Oliveira DR. Optimal diabetes care outcomes following face-to-face medication therapy management services. Popul Health Manag. 2013;16(1):28–34. [DOI] [PubMed] [Google Scholar]
- 5.Budlong H, Brummel A, Rhodes A, Nici H. Impact of Comprehensive Medication Management on Hospital Readmission Rates. Popul Health Manag. 2018;21(5):395–400. [DOI] [PubMed] [Google Scholar]
- 6.Isetts BJ, Brummel AR, de Oliveira DR, Moen DW. Managing drug-related morbidity and mortality in the patient-centered medical home. Med Care. 2012;50(11):997–1001. [DOI] [PubMed] [Google Scholar]
- 7.Moczygemba LR, Barner JC, Brown CM, et al. Patient satisfaction with a pharmacist-provided telephone medication therapy management program. Res Social Adm Pharm. 2010;6(2):143–154. [DOI] [PubMed] [Google Scholar]
- 8.Funk KA, Pestka DL, Roth McClurg MT, Carroll JK, Sorensen TD. Primary Care Providers Believe That Comprehensive Medication Management Improves Their Work-Life. J Am Board Fam Med. 2019;32(4):462–473. [DOI] [PubMed] [Google Scholar]
- 9.Bodenheimer T, Sinsky C. From triple to quadruple aim: care of the patient requires care of the provider. Ann Fam Med. 2014;12(6):573–576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gill JM, Bagley B. Practice Transformation? Opportunities and Costs for Primary Care Practices. Ann Fam Med. 2013;11(3):202 LP–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Miller R, Weir C, Gulati S. Transforming primary care: scoping review of research and practice. J Integr care. 2018;26(3):176–188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Miller R, Khullar D, Challa K, Jenkins C, Chokshi DA. Primary Care Transformation Across The Largest Safety-Net Health System In The United States . https://www.healthaffairs.org/do/10.1377/hblog20180226.969530/full/. Accessed December 7, 2020. [Google Scholar]
- 13.Glasgow RE, Vogt TM, Boles SM. Evaluating the public health impact of health promotion interventions: the RE-AIM framework. Am J Public Health. 1999;89(9):1322–1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Glasgow RE, Harden SM, Gaglio B, et al. RE-AIM Planning and Evaluation Framework: Adapting to New Science and Practice With a 20-Year Review. Front Public Heal. 2019;7:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Reach – RE-AIM. https://www.re-aim.org/about/what-is-re-aim/reach/. Accessed October 25, 2020.
- 16.Leeman J, Birken SA, Powell BJ, Rohweder C, Shea CM. Beyond “implementation strategies”: classifying the full range of strategies used in implementation science and practice. Implement Sci. 2017;12(1):125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Patient-Centered Primary Care Collaborative. The Patient-Centered Medical Home: Integrating Comprehensive Medication Management to Optimize Patient Outcomes. Washington D.C.; 2012. https://www.pcpcc.org/sites/default/files/media/medmanagement.pdf. [Google Scholar]
- 18.Pharmacy Quality Alliance. PQA Medication Therapy Problem Categories Framework. https://www.pqaalliance.org/assets/Measures/PQA MTP Categories Framework.pdf. Published August 2017. Accessed October 28, 2020.
- 19.Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. [DOI] [PubMed] [Google Scholar]
- 20.SAS [computer program]. 2013. Cary, NC. [Google Scholar]
- 21.FastStats - Therapeutic Drug Use. https://www.cdc.gov/nchs/fastats/drug-use-therapeutic.htm. Accessed July 2, 2021.
- 22.Charlesworth CJ, Smit E, Lee DSH, Alramadhan F, Odden MC. Polypharmacy Among Adults Aged 65 Years and Older in the United States: 1988-2010. J Gerontol A Biol Sci Med Sci. 2015;70(8):989–995. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.