Patient-Centered Medical Home Recognition and Diabetes Control Among Health Centers: Exploring the Role of Enabling Services

Abstract

The patient-centered medical home (PCMH) model has been considered a promising approach to improve chronic care delivery, particularly among patients with diabetes. There is theoretical support to suggest that certain nonmedical services, such as enabling services (eg, case management, social work, transportation), embedded within PCMH could be contributing to successful model implementation. It remains unclear whether PCMH recognition or enabling services are related to diabetes control. Federally Qualified Health Centers (FQHCs) are an important setting in which to study this relationship given the considerable effort required to implement the PCMH model and the ubiquity of enabling services in these safety net settings. This cross-sectional, population-based study used 2012 data from the Health Resources and Services Administration's Uniform Data System and PCMH Recognition Initiative Dataset to determine whether PCMH recognition status was associated with diabetes control rates among FQHCs, while controlling for covariates including enabling services. The study linear regression model estimated that PCMH recognition was associated with a 1.5% increase in the proportion of patients with controlled diabetes (B = 0.015; 95% CI 0.002, 0.027). Clinic region, patient age, and race/ethnicity groups also were related to diabetes control; however, enabling services were not. These findings suggest there is a positive association between PCMH recognition and diabetes control rates among FQHCs. Future research, using data that accurately reflect the provision and utilization of PCMH primary care functions and related enabling services, is needed to fully understand the relationship between the PCMH model and population health measures such as diabetes control.

Introduction

The patient-centered medical home (PCMH) is a health care delivery model intended to achieve the Triple Aim: improved patient experience, improved population health, and reduced cost of care.^1,2 The model involves 5 key primary care functions: comprehensive care, patient-centered care, coordinated care, accessible services, and quality and safety.³ Collectively these functions aim to increase the value of health care by improving the health of the whole patient, family, community, and population.^4,5

The PCMH model has been considered a promising approach to improve chronic care delivery. In fact, the model has been most efficient among populations with comorbid chronic conditions⁶ such as diabetes, which is apt because diabetes management requires many of the key primary care functions (eg, care coordination, comprehensive care, patient-centered) present in the medical home model.^7,8

Some studies have found that PCMH status was associated with lower rates of health care utilization (ie, emergency department [ED] visits, hospitalizations, readmissions, specialty visits, prescription drug use) and improvements in related cost savings.^1,9

–14 Additional PCMH studies reported improvements in preventive care measures such as breast and cervical cancer screening, as well as diabetes care quality (ie, glycated hemoglobin [HbA1c] testing, low-density lipoprotein cholesterol testing, nephropathy monitoring, eye examinations).^13,15 Alternatively, some PCMH studies found the model had little to no effect on diabetes or hypertension control, or preventive measures (ie, immunizations, cervical cancer screening, well-child visits),^16,17 and was associated with more hospital admissions, readmissions, ED visits, and total Medicare payments.¹⁶

Mixed findings could be related, in part, to the limitations and challenges inherent to PCMH certification, measurement, and evaluation. Medical home definitions and recognition criteria vary by agency and accrediting body. As Shi et al noted in their study of PCMH adoption and related clinical performance, PCMH recognition status alone may not indicate that the practice is functioning as a medical home, and could yield distorted findings; there likely are certain elements of the PCMH model that might contribute to diabetes control more than others.¹⁸ PCMH research has yet to definitively determine which elements of the model are necessary and sufficient to achieve desired health outcomes.

There is theoretical support to suggest that certain nonmedical services (eg, enabling services) embedded within PCMH could be contributing to successful model implementation.^19,20 Enabling services are nonmedical services that are delivered by primary care practices to facilitate access to health care.¹⁹ Enabling services are the cornerstone of Federally Qualified Health Centers (FQHCs), which are outpatient clinics that qualify for specific reimbursement systems under Medicare and Medicaid and provide care to underserved, often disadvantaged, areas and populations. FQHC patients are more likely than overall low-income populations to report only fair or poor health and to have higher rates of chronic conditions.²¹ Research on the effectiveness of enabling services among FQHCs is limited, but studies have indicated that nonmedical services provided by FQHCs, such as social work, case management, transportation, outreach, and patient education, can improve health care utilization and health outcomes related to diabetes management.^22

–25 It remains unclear whether enabling services are a necessary element of the PCMH model or if they play a significant role in facilitating diabetes control among FQHCs.

This cross-sectional, population-based study sought to determine whether PCMH recognition status is associated with diabetes control rates among FQHCs. This study builds on previous research on medical homes^18,26,27 by using the aforementioned enabling services as a potential covariate to help explain the relationship between medical home recognition and diabetes control. FQHCs are an important setting for this study given the considerable effort required to implement the PCMH model and the ubiquity of enabling services in these safety net settings.

Methods

Data

This study used data reported by all FQHCs (N = 1198; serving nearly 24 million patients) in 2012 as part of the Uniform Data System (UDS), a reporting requirement of the US Health Resources and Services Administration (HRSA). These data included FQHC staffing and service characteristics, patient demographics, and clinical measures including diabetes control.²⁸ HRSA's 2012 PCMH Recognition Initiative Dataset, collected independently from UDS data, provided PCMH recognition status data. This data set included all FQHCs with medical home recognition (of any level) from the Accreditation Association for Ambulatory Health Care, The Joint Commission, or the National Commission for Quality Assurance.

Measures

FQHC characteristics

Health center characteristics described the study population and were potential covariates in the analysis. FQHCs were characterized by patient volume, total full-time equivalents (FTEs), service area (ie, rural, urban), and region (ie, West, Midwest, South, Northeast, other US territories [American Samoa, Federated States of Micronesia, Guam, Marshall Islands, Puerto Rico, Palau, Virgin Islands]). Enabling services included in this analysis were social work, case management, education, outreach, and transportation, and were expressed as a proportion of FTEs (eg, case manager FTEs/total health center FTEs).

Patient characteristics

FHQC patient characteristics included sex, age, race/ethnicity, income (ie, percentage of poverty), and insurance status (ie, Medicaid, Medicare, other public insurance, private insurance, uninsured). Patient characteristics were expressed at the health center level as a proportion (ie, 0–100%) of total patients per health center.

Dependent measure

The primary dependent measure was diabetes control. Diabetes control was expressed as a continuous proportion (ie, 0–100%) of patients per health center, aged 18 to 74 years, with a diagnosis of type I or type II diabetes, whose HbA1c was ≤7%. The denominator was the total number of adult patients with a diagnosis of type I or II diabetes, who had been seen in the health center for medical services at least twice during the reporting year, and who did not meet any of the exclusion criteria (eg, evidence of gestational or steroid-induced diabetes). Diabetes control data reflected the last HbA1c measurement taken in 2012, and was obtained from chart audits (random sampling) and/or the use of electronic health records. The 2012 UDS Manual contains a list of exclusion criteria and a complete description of data collection methods.²⁹

Independent measure

PCMH recognition (ie, PCMH recognized, non-recognized) was the primary independent measure. PCMH recognized health centers included FQHCs with recognition (at any level) from any of the aforementioned PCMH accrediting bodies. Health center organizations were counted as recognized if at least 1 clinic site was recognized as a PCMH.

Covariates

Covariates included various health center and patient characteristics and were selected on the basis of their known and theoretical association with diabetes and diabetes control.^19,30

–33 They included health center region, total FTEs, enabling services, patient sex, age, race/ethnicity, income, and insurance status.

Analytic strategy

Chi-square and t tests were used to make bivariate comparisons between PCMH recognized and non-recognized FQHCs. Two generalized linear models were fit to assess the relationship between PCMH recognition and diabetes control. The first model controlled for health center service region, total FTEs, and patient demographics. In order to determine whether enabling services influence the association between PCMH recognition and diabetes control, the second model controlled for social workers, case managers, education, outreach, and transportation services in addition to health center service region, total FTEs, and patient demographics.

Of the total 1198 FQHCs, 4 were excluded because of substantial missing data, which left 1194 FQHCs in the final sample. The remaining missing data were negligible (ie, less than 5% of the sample and likely missing at random); therefore, single mean imputation was used to address item-level nonresponse. This study was approved as exempt research by the University of Kentucky Institutional Review Board.

Results

FQHC characteristics

FQHCs were grouped by PCMH recognition status: health centers that were recognized as PCMHs (n = 562), and health centers that had not achieved PCMH recognition (ie, non-recognized FQHCs) (n = 632); which served an average of 22,800 and 13,115 patients, respectively. Most health centers served urban areas and were concentrated in the Western and Southern regions of the United States. Additional FQHC characteristics are presented in Table 1.

Table 1.

Health Center Characteristics

	PCMH recognition status
FQHC characteristics	Recognized	Non-recognized	P
Total, n (%)	562 (52.9)	632 (47.1)	—
Total FTE, mean (SD)	97.3 (117.1)	51.4 (63.9)	<0.01
Service area, n (%)
Urban	354 (63.1)	390 (61.8)	0.67
Rural	208 (36.9)	241 (38.2)
Region, n (%)
West	150 (26.7)	166 (26.3)	0.01
Midwest	111 (19.8)	118 (18.7)
South	172 (30.6)	240 (38.0)
Northeast	117 (20.8)	92 (14.5)
Other US territories	12 (2.1)	16 (2.5)
Enabling services (% total FTEs), mean (SD)
Social worker	2.4 (2.9)	2.3 (3.3)	0.38
Case manager	5.9 (6.0)	6.4 (5.7)	0.31
Transportation	0.9 (1.8)	1.2 (2.6)	0.01
Outreach	2.5 (2.8)	3.4 (5.5)	<0.01
Patient education	2.6 (3.0)	2.5 (3.3)	0.52

Enabling services are expressed as an average proportion (%) of total FTEs.

FQHC, Federally Qualified Health Center; FTE, full-time equivalent; PCMH, patient-centered medical home; SD, standard deviation.

PCMH recognized and non-recognized FQHCs differed significantly by region (P = 0.01). When compared to non-recognized FQHCs, PCMH recognized FQHCs had more total FTEs (45.8, P < 0.01), but fewer FTEs dedicated to transportation (−.3, P = 0.01) and outreach (−.9, P < 0.01) (Table 1). PCMH recognized FQHCs served a higher volume of patients (9685 vs. 13,115, P < 0.05), had a lower prevalence of diabetes (−.4%, P = 0.03), and greater rates of diabetes control (2.3%, P < 0.01) than non-recognized FQHCs. PCMH recognized FQHCs served a greater percentage of patients between the ages of 0 and 19 years (2.4%, P < 0.01), a smaller percentage of black/African American patients (−3.5%, P = 0.01), a greater percentage of patients insured by Medicaid (3.9%, P < 0.01), and a smaller percentage of uninsured patients (−5.2%, P < 0.01), when compared to FQHCs that were not recognized PCMHs (Table 2).

Table 2.

Health Center Patient Characteristics

	PCMH recognition status
Patient characteristics	Recognized	Non-recognized	P
Total patients, mean (SE)	22,800 (985.6)	13,115 (592.9)	<0.05
Diabetes, % (SE)
Prevalence	6.7 (0.1)	7.1 (0.1)	0.03
Control	40.5 (0.3)	38.2 (0.5)	<0.01
Sex, % (SE)
Male	42.2 (0.3)	42.8 (0.3)	0.14
Female	57.8 (0.3)	57.2 (0.3)
Age (years), % (SE)
0–19	31.5 (5.1)	29.1 (5.7)	<0.01
20–64	60.3 (0.5)	63.1 (0.6)	<0.01
65+	8.2 (0.2)	7.9 (0.2)	0.47
Race/ethnicity, % (SE)
White	44.9 (1.3)	42.2 (1.2)	0.13
Black/African American	17.9 (1.0)	21.4 (1.0)	<0.05
Hispanic/Latino	17.2 (0.9)	16.0 (0.9)	0.34
Asian/Pacific Islander	3.5 (0.5)	3.8 (0.5)	0.61
American Indian	1.7 (0.3)	2.7 (0.5)	0.09
Other	16.2 (0.8)	14.6 (0.7)	0.15
Income (% poverty), % (SE)
100% and below	52.3 (0.9)	50.3 (1.0)	0.15
101%–150%	11.3 (0.3)	10.7 (0.3)	0.18
151%–200%	5.5 (0.2)	5.0 (0.2)	0.12
>200%	6.3 (0.3)	5.4 (0.3)	0.06
Insurance status, % (SE)
Medicare	9.3 (0.3)	8.7 (0.3)	0.16
Medicaid	36.4 (0.7)	32.5 (0.7)	<0.01
Private	16.6 (0.5)	15.7 (0.5)	0.22
Uninsured	36.0 (0.8)	41.2 (0.9)	<0.01

Patient characteristics, with the exception of Total Patients, are expressed at the health center level as average proportion (%) of patients.

Percentages may not sum to 100 because of missing data or rounding.

PCMH, patient-centered medical home; SE, standard error.

PCMH recognition and diabetes control

Model 1

After adjusting for health center service region, total FTEs, and patient demographics, the first linear regression model estimated that PCMH recognition was associated with a 1.5% increase in the proportion of patients with controlled diabetes (Table 3). Health center region, age, and race/ethnicity groups also were also related to diabetes control. When compared to health centers in the Western region of the US, health centers in the Midwest and Northeast were associated with a 3.5% and 2.4% increase in the proportion of patients with controlled diabetes, respectively, while health centers in other US territories were associated with a decrease in diabetes control rates. Age groups 20–64 years and 65 years and older were associated with a 6.5% and 44.7% increase in the proportion of patients with controlled diabetes, respectively. When compared to white race/ethnicity groups, black/African American race/ethnicity groups and American Indian groups were associated with a 4.3% and 1.6% decrease in the proportion of patients with controlled diabetes, respectively (Table 3).

Table 3.

Models 1 and 2—Health Center and Patient Characteristics Related to Diabetes Control Rates

	Model 1		Model 2
Characteristics	B	95% CI	B	95% CI
PCMH recognition	0.015^a	0.003 to 0.028	0.015^a	0.002 to 0.027
Total FTE	1.1E-5	−5.9E-5 to 8.3E-5	1.2E-5	−5.9E-5 to 8.4E-5
Region
West (Ref)	—	—	—	—
Midwest	0.035^a	0.015 to 0.055	0.034^a	0.014 to 0.055
South	0.012	−0.006 to 0.031	0.012	−0.005 to 0.031
Northeast	0.024^a	0.003 to 0.045	0.023^a	0.002 to 0.044–0.095
Other US territories	−0.095^a	−0.141 to −0.049	−0.095^a	−0.141 to −0.050
Enabling services
Social worker	—	—	0.094	−0.107 to 0.296
Case manager	—	—	−0.050	−0.140 to 0.039
Transportation	—	—	−0.098	−0.379 to 0.181
Outreach	—	—	−0.008	−0.152 to 0.135
Patient education	—	—	0.001	−0.203 to 0.206
Sex
Male (Ref)	—	—	—	—
Female	−4.5E-4	−0.092 to 0.091	−0.127	−0.108 to 0.083
Age (years)
0–19 (Ref)	—	—	—	—
20–64	0.065^a	0.002 to 0.127	0.069^a	0.006 to 0.133
65+	0.447^a	0.267 to 0.627	0.447^a	0.266 to 0.627
Race/ethnicity
White (Ref)	—	—	—	—
Black/African American	−0.043^a	−0.074 to −0.012	−0.041^a	−0.073 to −0.009
Hispanic/Latino	−0.020	−0.057 to 0.015	−0.019	−0.056 to 0.018
Asian/Pacific Islander	−0.026	−0.082 to 0.029	−0.024	−0.081 to 0.031
American Indian	−0.016^a	−0.081 to −0.049	−0.016	−0.082 to 0.050
Other	−0.032	−0.072 to 0.006	−0.030	−0.070 to 0.009
Income (% poverty)
100% and below (Ref)	—	—	—	—
100%–150%	0.069	−0.030 to 0.169	0.066	−0.034 to 0.167
151%–200%	0.036	−0.98 to 0.151	0.043	−0.091 to 0.179
>200%	0.042	−0.044 to 0.129	0.040	−0.047 to 0.127
Insurance status
Medicare (Ref)	—	—	—	—
Medicaid	0.070	−0.0460 0.200	0.072	−0.058 to 0.203
Private	0.019	−0.115 to 0.153	0.017	−0.117 to 0.152
Uninsured	−0.038	−0.156 to 0.079	−0.035	−0.153 to 0.082

Model 1, R ² = 0.147, F = 10.15; Model 2, R ² = 0.149, F = 8.21.

P < 0.05.

CI, confidence interval; FTE, full-time equivalent; PCMH, patient-centered medical home; Ref, reference group.

Model 2

The second linear regression model, which added enabling services (ie, health center social worker, case manager, education, outreach, transportation) as covariates, found the same degree of association between PCMH recognition and increase in the proportion of patients with controlled diabetes as Model 1. Health center region, and age and race/ethnicity groups also were related to diabetes control. Model 2 did not indicate that any of the enabling services were significantly associated with diabetes control (Table 3).

Discussion

This paper explores the association between PCMH recognition and diabetes control, while accounting for the influence of enabling services and other potential covariates. This study found that PCMH recognition among FQHCs was associated with greater rates of diabetes control. Health center region, age, and race also were related to diabetes control; however, enabling services were not.

When the study team evaluated multivariable comparisons between PCMH recognized and non-recognized FQHCs and controlled for health center service region, total FTEs, and patient demographics, the team found that health centers recognized as medical homes had greater rates of diabetes control when compared to health centers without medical home recognition – a finding that aligns with existing diabetes and medical home research.^7,8,22 The PCMH model is intended to promote improved chronic disease management. Patient education, promotion of self-management, and the use of case managers are examples of strategies embedded within the PCMH model that have demonstrated improved diabetes control.^22,34

In an effort to better understand what specific PCMH-like practices and services were associated with diabetes control, enabling services were added to the model. It is known that enabling services have the potential to improve a variety of health measures^{19,20,22

–25}; however, the effectiveness of enabling services has been difficult to evaluate among community health centers because the practice and utilization of such services are challenging to measure.

When enabling services were introduced into the adjusted model, it would not have been surprising to observe the relationship between PCMH recognition and diabetes control rates weaken, and a significant association between diabetes control and 1 or more of the enabling services emerge, but this was not the case. Instead, this study found that PCMH recognition, health center region, and patient age and race/ethnicity remained significant; however, enabling services staff (ie, social workers, case managers, transportation, outreach, patient education) were not significantly related to diabetes control rates.

Both models estimated an association between health center region and diabetes control, which aligns with national statistics.³² Minority race/ethnicity groups were negatively associated with diabetes control – a relationship that also is well documented in diabetes literature.^30,35

The models also estimated health centers serving older patient groups were associated with greater rates of diabetes control when compared to health centers serving younger patient groups. This was an unexpected finding as diabetes control can become more challenging with age because of increased insulin resistance, decreased physical activity, and the presence of comorbid conditions.³¹ It also is surprising because there is growing concern that tight control for elderly people may be more harmful than helpful.³⁶ The study findings may be explained in part by patterns in health care seeking behaviors. Data from the 2013 National Health Interview Survey indicate that use of diabetes medication, visits to eye and foot care specialists, and blood pressure and cholesterol checks all increase with age.³⁷ The relationship this study found between health center age groups and diabetes control rates may reflect variation in diabetes prevalence and type by age and health care seeking behaviors; however, such individual characteristics were not measured by the population level data.

Past PCMH research has produced mixed results and focused primarily on cost and quality metrics among pilot programs.^1,38 Few studies have considered the role of nonmedical enabling services in the evaluation of the medical home. Those that have used past data from a time when few FQHCs were recognized medical homes¹⁹ and did not examine specific enabling services. The present study is among the first to use a complete population-level data set to explore the relationship between PCMH recognition and specific enabling services and a critical population health measure – diabetes control.

This study is subject to a number of potential limitations. The PCMH model is comprised of 5 key primary care functions and is measured through formal recognition or accreditation, yet it remains unclear whether the current recognition tools accurately measure the key primary care functions.²⁰ Although the PCMH recognition process may be considered the most viable method to measure the medical home, when interpreting the results, one must consider the distinction between PCMH recognition and implementation, and understand that PCMH recognition does not measure implementation or practice of specific PCMH functions or related services. Despite the limitations inherent to PCMH measurement and recognition, an association was found between medical homes and diabetes control; thus, these findings may be an underestimation of the true association. Continued research is needed to identify and understand the relationship between PCMH recognition, specific PCMH practices such as enabling services, and diabetes management.

Health centers choose whether or not to pursue recognition as a PCMH and the centers that made that decision may be systematically different from the ones that did not. Many of these differences were controlled for in the analyses, but not all differences could be accounted for. The American Academy of Family Physicians' National Demonstration Project, which tested the PCMH model in a diverse sample of family practices, discovered that practices most successful in adopting the model possessed advanced adaptive reserve. Adaptive reserve refers to an internal capability for organizational learning and development and involves strong leadership, communication, trust, and a commitment to adaptation and growth.³⁹ Such nuanced differences between health centers, beyond PCMH recognition status, may explain some of the differences in clinic management, care strategies, and ultimately diabetes control rates.³⁹

The present study's nonsignificant findings may be attributable to the method with which enabling services data were measured and collected. HRSA requires that FQHCs provide all appropriate enabling services in order to receive federal funding⁴⁰; reporting of FTEs attributed to enabling services in the UDS likely reflects that requirement. Measuring enabling services by FTE calculations (eg, FTEs dedicated to social work) may not be the most reliable approach as it does not guarantee the service was provided or utilized. Data that accurately measure enabling service provision and utilization are required to detect any effect enabling services may have on diabetes management and related diabetes control rates. Another possibility is that this study's findings reflect the true lack of association between enabling services and diabetes control (HbA1c <7%). Future related research could explore other health conditions as there may be different health measures and outcomes that are more amenable to enabling services.

Conclusion

This study's findings suggest that there is a positive association between PCMH recognition and diabetes control rates among FQHCs. Future research, using data that accurately reflect the provision and utilization of PCMH primary care functions and related enabling services, is needed to fully understand the relationship between the PCMH model and population health measures such as diabetes control.

Footnotes

Acknowledgments

We thank the Health Research and Services Administration (HRSA) for the data used in this study. This study reflects the work and views of the authors. HRSA, or any other federal agency, did not have a role in the study design, analysis, interpretation or decision to publish this study.

Author Disclosure Statement

Drs. Dobbins, Peiper, Jones, Clayton, Peterson, and Phillips declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: In addition to their stated affiliation, Dr. Dobbins is currently employed by Humana Inc., Dr. Peiper is employed by RTI International, and Dr. Jones is employed by the US Department of Health and Human Services. The authors received no financial support for this article.

References

Nielsen

, Buelt

, Patel

, Nichols

. The patient-centered medical home's impact on cost and quality: annual review of evidence, 2014–2015. February 2016. https://www.pcpcc.org/resource/patient-centered-medical-homes-impact-cost-and-quality-2014-2015 Accessed September 4, 2016 .

Institute for Healthcare Improvement. IHI triple aim initiative. February 2016. http://www.ihi.org/engage/initiatives/tripleaim/pages/default.aspx Accessed September 4, 2016 .

Agency for Healthcare Research and Quality; Patient-Centered Medical Home Resource Center.. Defining the PCMH. https://pcmh.ahrq.gov/page/defining-pcmh Accessed September 4, 2016 .

Moreno

, Peikes

, Krilla

Necessary But Not Sufficient: The HITECH Act and Health Information Technology's Potential to Build Medical Homes. AHRQ Publication no. 10-0080-EF. Rockville, MD: Agency for Healthcare Research and Quality, June 2010.

Arend

, Tsang-Quinn

, Levine

, Thomas

. The patient-centered medical home: history, components, and review of the evidence. Mt Sinai J Med, 2012; 79:433–450.

Flottemesch

, Anderson

, Solberg

, Fontaine

, Asche

. Patient-centered medical home cost reductions limited to complex patients. Am J Manag Care, 2012; 18:677–686.

Bojadzievski

, Gabbay

. Patient-centered medical home and diabetes. Diabetes Care, 2011; 34:1047–1053.

Jortberg

, Miller

, Gabbay

, Sparling

, Dickinson

. Patient-centered medical home: how it affects psychosocial outcomes for diabetes. Curr Diab Rep, 2012; 12:721–728.

Pines

, Keyes

, Van Hasselt

, McCall

. Emergency department and inpatient hospital use by Medicare beneficiaries in patient-centered medical homes. Ann Emerg Med, 2015; 65:652–660.

10.

Raphael

, Cooley

, Vega

, et al. Outcomes for children with chronic conditions associated with parent-and provider-reported measures of the medical home. J Health Care Poor Underserved, 2015; 26:358–376.

11.

Rosenthal

, Sinaiko

, Eastman

, Chapman

, Partridge

. Impact of the Rochester medical home initiative on primary care practices, quality, utilization, and costs. Med Care, 2015; 53:967–973.

12.

Maeng

, Khan

, Tomcavage

, Graf

, Davis

, Steele

. Reduced acute inpatient care was largest savings component of Geisinger health system's patient-centered medical home. Health Aff (Millwood), 2015; 34:636–644.

13.

Friedberg

, Rosenthal

, Werner

, Volpp

, Schneider

. Effects of a medical home and shared savings intervention on quality and utilization of care. JAMA Intern Med, 2014; 175:1362–1368.

14.

Farrell

, Tomoaia-Cotisel

, Scammon

, et al. Impact of an integrated transition management program in primary care on hospital readmissions. J Healthc Qual, 2015; 37:81–92.

15.

Jones

, Finison

, McGraves-Lloyd

, et al. Vermont's community-oriented all-payer medical home model reduces expenditures and utilization while delivering high-quality care. Popul Health Manag, 2016; 19:196–205.

16.

Friedberg

, Schneider

, Rosenthal

, Volpp

, Werner

. Association between participation in a multipayer medical home intervention and changes in quality, utilization, and costs of care. JAMA, 2014; 311:815–825.

17.

Christensen

, Zickafoose

, Natzke

, McMorrow

, Ireys

. Associations between practice reported medical homeness and health care utilization among publicly insured children. Acad Pediatr, 2015; 15:267–274.

18.

Shi

, Lock

, Lee

, et al. Patient-centered medical home capability and clinical performance in HRSA-supported health centers. Med Care, 2015; 53:389–395.

19.

Clarke

, Tseng

, Brook

, Brown

. Tool used to assess how well community health centers function as medical homes may be flawed. Health Aff (Millwood), 2012; 31:627–635.

20.

Escaron

, Weir

, Stanton

, Clarke

. Defining and rating the effectiveness of enabling services using a multi-steakholder expert panel approach. J Health Care Poor Underserved, 2015; 26:554–576.

21.

Shin

, Alvarez

, Sharac

, et al. A profile of community health center patients: implications for policy. December. 2013. https://kaiserfamilyfoundation.files.wordpress.com/2013/12/8536-profile-of-chc-patients.pdf Accessed November 27, 2016 .

22.

Akroyd

, Wexler

. Effectiveness of diabetes interventions in the patient-centered medical home. Curr Diab Rep, 2014; 14:471–487.

23.

DeCoster

VA.

Challenges of type 2 diabetes and role of health care social work: a neglected area of practice. Health Soc Work, 2001; 26:26–37.

24.

Syed

, Gerber

, Sharp

. Traveling toward disease: transportation barriers to health care access. J Community Health, 2013; 35:976–993.

25.

Tang

, Funnell

, Sinco

, et al. Comparative effectiveness of peer leaders and community health workers in diabetes self-management support: results of a randomized controlled trial. Diabetes Care, 2014; 37:1525–1534.

26.

Gao

, Nocon

, Gunter

, et al. Characteristics associated with patient-centered medical home capability in health centers: a cross-sectional analysis. J Gen Intern Med, 2016; 31:1041–1051.

27.

Jones

, Zur

, Dobbins

. Laying the foundation: factors associated with medical home recognition among health centers. Am J Prev Med, 2016; 51:e21–e26.

28.

Health Resources and Services Administration. Uniform Data System. Rockville, MD: US Department of Health and Human Services, 2012.

29.

Bureau of Primary Health Care.. UDS Reporting Instructions for Health Centers. December 1,. 2012. http://bphc.hrsa.gov/datareporting/reporting/2012udsmanual.pdf Accessed August 25, 2016 .

30.

Spinakis

, Golden

. Race/ethnic differences in diabetes and diabetic complications. Curr Diab Rep, 2013; 13:814–823.

31.

Muller

, Elahi

, Tobin

, Andres

. The effect of age on insulin resistance and secretion: a review. Semin Nephrol, 1996; 16:289–298.

32.

Beckles

, Chou

. Diabetes–United States, 2006–2010. Morb Mortal Wkly Rep, 2013; 62:99S–104S.

33.

Centers for Disease Control and Prevention; United States Diabetes Surveillance System.. Diagnosed diabetes. Updated December 16,. 2016. https://gis.cdc.gov/grasp/diabetes/DiabetesAtlas.html Accessed December 17, 2016 .

34.

Tricco

, Ivers

, Grimshaw

, et al. Effectiveness of quality improvement strategies on the management of diabetes: a systematic review and meta-analysis. Lancet, 2012; 379:2252–2261.

35.

Golden

, Brown

, Cauley

, et al. Health disparities in endocrine disorders: biological, clinical, and nonclinical factors—an Endocrine Society scientific statement. J Clin Endocrinol Metab, 2012; 97:E1579–E1639.

36.

American Diabetes Association.. Tight diabetes control. May 13,. 2015. http://www.diabetes.org/living-with-diabetes/treatment-and-care/blood-glucose-control/tight-diabetes-control.html Accessed February 9, 2017 .

37.

Villarroel

, Vahratian

, Ward

. Health care utilization among US adults with diagnosed diabetes,. 2013. NCHS Data Brief No. 183. February 2015. http://www.cdc.gov/nchs/data/databriefs/db183.htm Accessed September 4, 2016 .

38.

Nielsen

, Gibson

, Buelt

, Grundy

, Grumbach

. The patient-centered medical home's impact on cost and quality: annual review of evidence, 2013–2014. January. 2015. https://www.pcpcc.org/resource/patient-centered-medical-homes-impact-cost-and-quality-annual-review-evidence-2013-2014 Accessed September 4, 2016 .

39.

Nutting

, Crabtree

, Miller

, Stange

, Stewart

, Jaen

. Transforming physician practices to patient-centered medical homes: lessons from the national demonstration project. Health Aff (Millwood), 2011; 30:439–445.

40.

Health Resources and Services Administration. Summary of program requirements. http://bphc.hrsa.gov/programrequirements/summary.html Accessed September 4, 2016 .