Longitudinal Analysis of Electronic Health Record Adoption on Staffing Mix in Community Health Centers ASHEcon 6th Biennial Conference Philadelphia, PA June 15, 2016 Bianca K. Frogner, PhD Associate Professor, Department of Family Medicine Director, Center for Health Workforce Studies School of Medicine

Acknowledgements Co-authors: Xiaoli Wu, MS Research Associate, GW Milken Institute School of Public Health Jeongyoung Park, PhD Assistant Professor, GW School of Nursing Patricia Pittman, PhD Associate Professor, GW Milken Institute School of Public Health Funding: Cooperative Agreement for a Regional Center for Health Workforce Studies (1 U81HP26493-01-00) from the Health Resources and Services Administration awarded to The George Washington University (PI: Patricia Pittman) Additional data provided by The George Washington University Geiger Gibson Program in Community Health Policy

Background Community health centers (CHCs) have long operated with challenges in staffing recruitment and retention Although electronic health records (EHRs) are slow to proliferate across primary care settings, CHCs have been leaders in the adoption of EHRs among primary care providers However, some CHCs are faster adopters than others and many remain without EHRs Limited understanding on whether there is something unique about the staffing in CHCs that lend themselves to adopt EHRs May be useful for other ambulatory care settings to learn as they increasingly adopt EHRs – workforce is often a significant barrier Add my cross sectional study results and Jeongyoung PCMH results

Related Studies Community health centers (CHCs) have adaptable staffing patterns, and manage to maintain similar productivity levels across varying staffing models (Ku et al, 2015, Health Affairs) CEOs of CHCs frequently consider environmental factors (e.g., local supply, scope of practice regulations, relative wages, and reimbursement arrangements) in hiring decisions; also consider personal attributes, clinical experience, quality of professional education, and likelihood of retention in determining distribution and roles (Pittman et al, under review) CHC adoption of the Patient Centered Medical Home Model (PCMH) found to be associated with significant growth in use of advanced practice staff (Park et al, under review) Productivity of physicians in CHCs significantly improved with more years of experience with EHRs, while nurses had a notable but not statistically significant negative impact with EHR adoption (Frogner et al, under review)

Data and Sample Study Question: Data: Sample How does EHR adoption change staffing mix of CHCs over time? Data: 2007 to 2013 Uniform Data System (UDS) Survey responses from “Readiness for Meaningful Use and HIT and Patient Centered Medical Home Recognition Survey” Sample CHCs operating in the fifty states and DC that were operating for the entire study period of 2007 to 2013 CHCs with identifiable year of EHR adoption N=5250 CHC-year observation (750 unique CHCs)

Analytic Approach Fractional Multinomial Logit (fmlogit) to estimate how adoption of EHR impacts the change in proportion of provider types Example: How the proportion of one type of provider category such as physicians, shifts the proportion of another type of provider category, such as nurses, keeping all else constant Controls: CHC size, geographic location, presence of a patient-centered medical home model, and other local market conditions quasi maximum likelihood dependent variables that each must range between 0 and 1 and must always, for each observation, add up to 1

Baseline Community Health Center Profile by EHR Adoption Status, 2007 Never Adopted EHRs (N=314)   EHR Adopter (N=311) Always had EHRs (N=125) Female 57.2% 58.5% 57.8% Age Age 18 and under 31.2% 34.8% 31.3% Age 19-64 61.4% 73.2% 63.0% Age 65 and over 7.7% 8.5% 7.9% Race/Ethnicity White 59.1% 54.2% 51.8% Black 20.5% 22.9% 19.1% Hispanic 24.5% 24.1% Asian/Pacific Islander 2.4% 3.0% 4.3% American Indian/Alaska Native 3.1% 0.8% Other/Unknown 23.% 24.2% 30.1% Insurance Type Uninsured** 41.4% 46.1% 39.6% Medicaid 30.2% 28.7% 31.7% Medicare 9.4% 10.4% 9.5% Other Public* 1.4% 1.9% 2.5% Private 19.8% 17.9% 18.9% Patients with limited English proficiency 19.0% 19.9% 20.0% Patients at 100% or below poverty level 46.9% 50.0% 49.9% Total number of patients* 13,200 14,212 18,002 Rural (Non-Metro) 53.0% 52.0% 59.0% Located in Health Professional Shortage Area (Measured by Whole County) 50.3% 51.4% 55.2% Local Market Conditions Total Active Physicians per 1,000 Population** 2.7 2.8 3.5 Total Nurse Practitioners w/ National Provider Identifier per 1,000 Population* 0.3 0.4

Predicted Shares of Staffing Type by Adoption of EHRs in CHCs

Key Findings Over the seven-year study period, 17% CHCs had EHRs for the entire period, 42% CHCs never adopted EHRs, and 41% adopted at some point over the study period CHCs with EHRS had significantly lower share of physicians compared to CHCs without EHRs throughout the study period CHCs with EHRs had significantly higher share of other medical staff among early adopters and late adopters The trend in share of other medical staff suggests the need for adequate support staff to get an EHR system to successfully “go live” 2011 peak year of adoption

Conclusions and Policy Implications CHCs with and without EHRs experienced similar trends over time in their staffing configuration But CHCs with EHRs allocate their total medical staff differently than CHCs without EHRs. CHCs with EHRs appear to elevate the use of other medical staff over all other staffing Especially in the early years of adoption and again in later years. The finding with regards to other medical staff appears to confirm early studies that suggest that adequate support staff is necessary to get an EHR system to successfully “go live” Although this study does not confirm it, this findings provide support for the hypothesis that EHR adoption in CHCs allows for greater flexibility among staff types

QUESTIONS? CONTACT: Bianca Frogner, PhD bfrogner@uw.edu CHWS Website https://depts.washington.edu/fammed/chws/ Twitter @biancafrogner @uwchws

Nurse Practitioners/Physician Assistants Regression Coefficients on Key Variables Estimating Share of Staffing (N=3498) Coefficient Robust SE P Value Nurse Practitioners/Physician Assistants EHR 0.104 0.220 0.635 PCMH 0.227 0.451 0.615 EHR * Year 0.004 0.022 0.864 PCMH * Year -0.025 0.037 0502 EHR * PCMH -0.085 0.089 0.340 Year 0.023 0.013 0.082 Nurses 0.079 0.207 0.703 -0.269 0.519 0.604 0.020 0.854 0.042 0.601 -0.150 0.087 0.086 -0.011 0.012 0.360 Other Medical Staff 0.239 0.199 0.229 -0.371 0.435 0.394 -0.017 0.379 0.026 0.035 0.464 0.116 0.084 0.168 0.063 0.000 Omitted category: physicians; Control vars: CHC size, urban geography, number of physicians and nurse practitioners per thousand capita in county