Sara J. Singer, MBA, PhD

1) Evaluate structural capabilities associated with the patient-centered medical home (PCMH) model in PCMH pilots in Colorado, Ohio, and Rhode Island; 2) evaluate changes in capabilities over 2 years in the Rhode Island pilot; and 3) evaluate facilitators and barriers to the adoption of capabilities.

We assessed structural capabilities in the 30 pilot practices using a cross-sectional study design and examined changes over 2 years in 5 Rhode Island practices using a pre/post design.

We used National Committee for Quality Assurance’s Physician Practice Connections—Patient-Centered Medical Home (PPC/PCMH) accreditation survey data to measure capabilities. We stratified by high and low performance based on total score and by practice size. We analyzed change from baseline to 24 months for the Rhode Island practices. We analyzed qualitative data from interviews with practice leaders to identify facilitators and barriers to building capabilities.

On average, practices scored 73 points (out of 100 points) for structural capabilities. High and low performers differed most on electronic prescribing, patient self-management, and care-management standards. Rhode Island practices averaged 42 points at baseline, and reached 90 points by the end of year 2. Some of the key facilitators that emerged were payment incentives, “transformation coaches,” learning collaboratives, and data availability supporting performance management and quality improvement. Barriers to improvement included the extent of transformation required, technology shortcomings, slow cultural change, change fatigue, and lack of broader payment reform.

For these early adopters, prevalence of structural capabilities was high, and performance was substantially improved for practices with initially lower capabilities. We conclude that building capabilities requires payment reform, attention to implementation, and cultural change.

	This article examines structural capabilities associated with the patient-centered medical home model in 30 small- and medium-sized primary care practices participating in medical home pilots in Colorado, Ohio, and Rhode Island. The authors found:

		On average, practices earned a total score of 73 points (out of 100) for their structural capabilities at baseline.

		High- and low-performing practices differed most in their achievement on electronic prescribing, patient self-management, and care management standards.

		Rhode Island practices had an average score of 42 points at baseline and 90 points after 24 months.

		Building structural capabilities requires attention to payment reform, implementation and cultural change.

The patient-centered medical home (PCMH) is gaining traction as a promising model for reforming primary care. Some experts believe that PCMHs have the potential to improve the quality, coordination, and efficiency of healthcare and to alleviate the impending manpower crisis in primary care.

 The PCMH has been supported by mounting evidence in 3 areas. First, there is abundant evidence from the United States and other Western countries that links primary care with better health outcomes, lower costs, and greater equity in care.

Second, there is a growing body of evidence around individual components of primary care—particularly those embodied in Wagner’s Chronic Care Model&mdash;linking the use of practice systems with improved quality of care, quality of life, and cost savings for some high-risk populations.

 Third, there is early evidence around the PCMH model itself that suggests it has the potential to enhance patient experience, improve some aspects of clinical quality, and reduce avoidable emergency utilization, although the impact on avoidable hospital utilization is, to date, mixed.

Interest in the PCMH is high, but to achieve maximum potential benefits, primary care practices may be required to invest substantially in structural capabilities associated with this model of care. For example, a population registry can help primary care practices group patients by diagnosis, to facilitate appropriate follow-up; performance management can help providers compare their data with benchmarks and motivate quality improvement; and referral tracking and care management can improve coordination among different professionals and organizations when patients are most vulnerable to medical errors. The Institute of Medicine, the health arm of the National Academy of Sciences, has highlighted the importance of these capabilities and identified their absence as representing a key gap in primary care.

However, studies examining the “readiness” of physician practices to implement structural capabilities show low levels of preparedness. Factors that increase the likelihood of the presence of structural capabilities include: an affiliation with an integrated system, hospital, or other large entity;

the presence of external incentives, such as pay-for-performance;

serving racial/ethnic and economically disadvantaged populations;

and operation in a less-competitive environment.

Practice size seems an especially important factor. 

 Larger practice size is associated with average adoption, with 22% of structural capabilities in place in small- to medium-sized practices (≤9 physicians)

 and 35% in larger practices (>9 physicians).

 found that the capabilities of nurse care managers, disease registries, clinical-information technology, and quality-improvement processes were least developed in smaller practices.

Because 88% of all patient visits occur in small- and medium-sized practices, it is important to understand the experience of these practices with regard to implementing structural capabilities, including which capabilities are easier or more difficult to adopt, whether practice size makes a difference, the time frame for implementation, and the resources and supports they require. This information can help policy makers understand how to best facilitate the transformation of physician practices into PCMHs.

As part of a larger study of the impact of the PCMH model on quality, costs, coordination, and consumer and provider satisfaction (using data from 30 primary care practices in 3 pilot initiatives in Rhode Island, Colorado, and Ohio),

 we conducted a quantitative investigation of the practices’ PCMH structural capabilities, using the Physician Practice Connections—Patient-Centered Medical Home Model (PPCPCMH) accreditation survey data from the National Committee for Quality Assurance (NCQA). We used a cross-sectional study design to assess the extent of adoption of structural capabilities in the pilots across the 3 states.

For the Rhode Island pilot, we obtained data on the structural capabilities of participating practices at 2 years post baseline because their baseline data were lower than in the other pilots. This allowed us to explore the change in prevalence of structural capabilities. To understand the facilitators and barriers to the adoption of structural capabilities, we interviewed practice care leaders in the primary care practices involved in the pilot. The study was approved by the Institutional Review Board at the Harvard School of Public Health.

Although there was variation across the pilots in the timing and priority attached to implementing specific components of the medical home, the 3 pilots shared 3 common elements: a self-assessment using the PPC-PCMH survey to obtain PCMH recognition; technical assistance from “transformation coaches” and participation in a learning collaborative; and a per-member, per-month care management fee.

The Rhode Island pilot involved 5 primary care practices with 28 physicians collectively serving 24,000 patients. Historically, there was an underinvestment in primary care, compared to high-performing healthcare systems.

 In 2008, the State Office of the Health Insurance Commissioner Department worked with commercial insurers to increase the portion of spending on primary care and also to support the Rhode Island Chronic Care Sustainability Initiative (with participation from primary care physicians, specialists, professional associations, payers, purchasers, and technical experts) with a graduated, per-member, per-month fee of $3 to $4.50, and funding support for a nurse care manager for each of the 5 pilot practices. The 5 practices were invited to participate in the pilot based on their diverse practice characteristics, history of working with each other, readiness for change, and use of an electronic medical record (EMR) (D. Hurwitz, MBA, letter, August 2013). The expectation was that the practices would strive to achieve certain levels of medical home recognition, participate in shared learning to facilitate practice transformation, and measure their performance.

The Colorado pilot included 14 primary care practices with 50 physicians collectively serving 98,000 patients. This pilot was initiated in May 2009 by HealthTeamWorks and major payers that included Aetna, Anthem-Wellpoint, Cigna, CoverColorado, Colorado Medicaid, Humana, and UnitedHealthcare. The practices were selected from a pool of 23 applications in a competitive process based on their demographics, the culture of the practice (eg, teamwork, collegiality, attitude), quality improvement practices, and electronic capabilities (eg, EMR, disease registry).

 The practices received a monthly fee of $4 to $8 per patient based on their NCQA recognition level, a pay-for-performance payment, assistance from transformation coaches, and participation in a learning collaborative involving all practices in the pilot. Additionally, for all participating practices, HealthTeamWorks brought in a vendor to implement “Reach My Doctor,” an extant Webbased registry featuring a patient portal.

The Ohio pilot had 11 primary care practices with 37 physicians collectively serving 30,000 patients. The pilot was initiated in September 2009 by the Health Improvement Collaborative of Greater Cincinnati and payers that included Anthem, Humana, and United- Healthcare. The practices were selected from a pool of 27 applications in a competitive process based on their diverse practice characteristics, demographics, readiness for change, and EMR capabilities S. Bolton, MPH, letter, August 2013). The practices received a monthly fee of $6 to $7 per patient based on their PCMH recognition level, practice transformation assistance offered mainly by telephone and webinars, and benchmarking data to support improvement.

We used the PPC-PCMH accreditation survey data to measure structural capabilities in the pilot practices. The PPC-PCMH survey asks 170 questions (elements) regarding the capabilities of physician practices on 9 standards: 1) access and communication, 2) patient tracking and registry, 3) care management, 4) patient self-management support, 5) electronic prescribing, 6) test tracking, 7) referral tracking, 8) performance reporting and improvement, and 9) interactive website. The questions identify the presence or absence of the elements, which are coded as present, absent, or not applicable. The score for each standard represents the proportion of weighted points, up to a possible cumulative score of 100 points (see 

To gain recognition by the NCQA, a practice selfassessed its capabilities using the PPC-PCMH web-based survey and submitted the required data to the NCQA. The NCQA then evaluated and scored the application based on the information received, and it conducted an audit of a sample of the applications as well. Pilot practices received 1 of 3 levels of recognition based on their total score and achievement on 10 “must pass” areas considered foundational for undergoing medical home transformation: 20 Level 1 (a total score of 25 to 49 points and at least 50% achievement on 5 out of the 10 “must pass” elements); Level 2: (a total score of 50 to 74 points and at least 50% achievement on the 10 “must pass” elements); or Level 3 (a total score of 75 to 100 points and at least 50% achievement on the 10 “must pass” elements).

We obtained the PPC-PCMH baseline data from the NCQA for all 30 primary care practices in the Rhode Island, Colorado, and Ohio pilots. Additionally, since the 5 practices in the Rhode Island pilot began the study at below Level 3 status, they had the opportunity to improve their recognition level. Accordingly, we also obtained their PPC-PCMH data at 24 months post baseline. We used practice addresses to identify the state involved and calculated each practice’s size based on its number of physicians.

To understand the implementation experiences as they sought to transform into medical homes, we conducted interviews at baseline, 18 months, and 30 months with a practice leader in each medical home who had responsibility for facilitating the transformation from primary care physician practices to patient-centered medical homes. Most often, they were physicians with a formal leadership role within the practice (eg, medical director), although sometimes they were informal champions of medical home transformation. The interviews lasted approximately 1 hour and were conducted by 2 or more authors. The interview protocol included open-ended questions with guided probes that explored facilitators and barriers to implementing the PCMH model (please see 

To understand the prevalence of structural capabilities, we calculated the mean and range of the total score, as well as the average percentage score by standard for the 30 pilot practices. To understand which capabilities are easiest to develop and which ones are more difficult, we stratified practices by their total score and compared the capabilities of the 10 lowest- and highest-performing practices by standard. To understand how the prevalence of capabilities differs by practice size, we also analyzed performance by 3 practice size categories: solo, 2 to 3 physicians, and 4 or more physicians. Finally, to understand how structural capabilities change over time, we analyzed the mean score and the average percentage by standard at baseline and 24 months for the 5 Rhode Island practices. Due to the small sample size, we did not examine statistical differences among groups.

We used thematic analysis to analyze the qualitative data.

 First, we analyzed responses to each interview question. The investigators discussed and developed a preliminary list of codes to represent salient and recurring themes describing facilitators and barriers to the adoption of structural capabilities. The primary author then applied the codes deduced through the initial analysis to data from the 60 interviews using qualitative software (ATLAS.ti, version 6.2). We established coding reliability by having the project’s research assistant code approximately 20% of the quotes extracted from the transcripts (kappa = 0.95, percent agreement = 0.96). We selected quotes to illustrate the key themes emerging from the interviews. All authors confirmed agreement with codes once applied.

The characteristics of the practices are summarized in 

. The majority of practices were in Colorado (47%), followed by Ohio (37%), and Rhode Island (17%). All were small- to medium-sized practices, 16 with the typical practice being a private family medicine, single-specialty practice, with 3 or fewer physicians.

The summary of structural capabilities at baseline appears in 

. The mean practice score across the 30 pilot practices was 72.7 points (range from 30 to 97 points) (eAppendix B). The majority of the practices had achieved Level 3 recognition (22 practices), while a small percentage had achieved Level 1 (4 practices) and Level 2 recognition (4 practices). The practices with lower PPC-PCMH status were predominantly those in Rhode Island.

The difference in structural capabilities between the 10 lowest- and highest-performing practices is shown in 

. There was least difference between the low- and high-performing practices in their achievement on the access and communication, patient tracking and registry, and referral tracking standards. The greatest difference in their achievement was on the electronic prescribing, patient self-management support, and care-management standards.

Solo practices averaged the highest PPC-PCMH score (87.9 points) compared with practices in other categories (average score for 2 to 3 physicians: 72 points; for 4 or more physicians: 69.5 points). Solo practices also scored the highest on each of the 9 standards (see eAppendix).

Articles

Thirty medical home pilot primary care practices had high structural capabilities at baseline and performance improved substantially after 24 months in practices starting with lower capabilities.

Sara J. Singer, MBA, PhD

Articles

Structural Capabilities in Small and Medium-Sized Patient-Centered Medical Homes