92
The Effects of the Implementation of Patient-Centered Medical Home on Staffing and Productivity in the Community Health Centers

Tuesday, June 14, 2016
Lobby (Annenberg Center)

Author(s): Jeongyoung Park; Xiaoli Wu; Bianca K. Frogner; Patricia Pittman

Discussant:

The patient-centered medical home (PCMH) is a model of primary care delivery that utilizes a collaborative, team-based approach to providing patients and their caregivers with coordinated, high quality care.  This practice redesign requires flexible and novel use of health care workers – i.e., traditional staff often take on expanded roles and new staff types may be added.  High-functioning team care leads to potential gains in productivity as team members can facilitate task delegation and shared decision-making.  As of 2014, more than 45% of federally-funded community health centers (CHCs) have achieved PCMH recognition.  Our objective is to understand the workforce transformation occurring in CHCs that have achieved PCMH recognition, and to assess the relationship of those changes to productivity.

The analyses are conducted using the 2007-2013 Uniform Data System that provides detailed information on each CHC, supplemented with four other datasets.  Our measure of staffing is the number of full-time equivalent (FTE) staff in the following types: primary care physicians; advanced practice staff (nurse practitioners, physician assistants); nurses; other medical staff (medical assistants, nurse aides, quality assurance/electronic health record staff); mental health/substance abuse staff; and enabling service staff (case managers, health educators).  Our measure of productivity is the total number of visits.  The key explanatory variable is whether the CHC adopted a PCMH for a given year.  We also create dummy variables to specify the years after PCMH adoption.  To assess whether a PMCH-related staffing change leads to a change in productivity, we generate the interaction terms between an indicator of PCMH adoption and staff FTE in each of 6 categories.  The main analysis method is a Difference-in-Differences approach which exploits variation in the timing of PCMH adoption, employing CHC fixed effects.  Propensity score matching is used to lessen bias by balancing covariates between PCMHs and their comparison groups.  450 PCMHs and 243 propensity score matched non-PCMHs are analyzed.

We find three important trends.  First, contrary to our expectations, we do not find significant changes in nurses, other medical staff, or enabling service staff.  Second, we find a growth in the use of advanced practice staff and a decline in primary care physician use.  For example, there is an average of 0.2 FTE increase in advanced practice staff in CHCs that adopted the PCMH model when compared to those that had not (p<0.1).  These increases are largest in CHCs that were four years into the transformation (1.2 FTE).  This is also reflected in reduced primary care physicians.  PCMHs had significantly reduced primary care physician FTEs over time compared to those that had not (2.4 FTE decrease after 4 years of PCMH adoption, p<0.01).  Third, we find productivity increases associated with this shift and with the increased use of mental health/substance abuse staff. 

This study provides a preliminary set of findings regarding the effects of PCMHs on staffing, and on the relationship of staffing to productivity.  More work is needed to fully understand these complex relationships and to gain better insight on how CHCs can better serve their diverse populations.