Early findings from our primary care/medical home benchmarking initiative provide a quantitative snapshot of medical homes investing in more clinical support staff per physician and taking on new site-level functions.
Four key take-aways
Last week we shared early findings from the staffing section of the primary care/medical home benchmarking initiative. Four key take-aways:
- Compared to non-PCMH primary care, PCMHs are altering staff model
- Compared to non-PCMH primary care, PCMHs are altering site functions
- Among PCMHs, greater IT access, larger practice size support a leaner-than-average staffing profile
- Among PCMHs, our profiled sites tend to be more robustly staffed than average
Note that none of these changes we documented are statistically significant. However, the findings agree with what our research team is seeing and hearing anecdotally in conversations with physician practices, hospitals, and health systems nationwide. See further information on sample size at the end of this post.
Compared to non-PCMH primary care, PCMHs are altering staff model
The average PCMH in our survey is six months into its transformation. About half (52%) of these PCMHs added new clinical support staff FTEs specifically to support the transition to medical homes. The mean number of FTEs added was 1.5.
Compared to the non-medical home primary care sites in the sample, the PCMHs had as many or more clinical support staff across all clinical credential categories and overall. Counting NPs, PAs, RNs, LPNs/LVNs, and MAs, non-PCMH primary care had an average total of 2.2 clinical support staff FTEs per physician--versus the PCMHs, which had 2.8.
PCMHs are altering site functions
Beyond the number of clinical support staff, we also see the staff model changes playing out in who does what task -- for example, the non-PCMH sites more frequently reported physicians as the primary owner of tasks such as pre-visit planning, patient self-management support, and population management data entry and data analysis than the PCMH sites did.
The run of data on "who mostly does [medical home] tasks such as..." yielded an interesting side-analysis on whether those tasks are done at all. By looking at which sites said "this service not offered at our site," we found suggestions that the PCMHs in the sample are operating differently from the non-PCMH primary care practices.
In particular, PCMHs more frequently report that someone on the team is doing these key medical home tasks:
- Pre-visit planning: 84% of PCMHs versus 67% of non-PCMH primary care
- Patient self-management support: 95% of PCMHs versus 68% of non-PCMH primary care
- Population management data entry: 98% of PCMHs versus 64% of non-PCMH primary care
- Population management data analysis: 98% of PCMHs versus 68% of non-PCMH primary care
- Among PCMHs, greater IT, larger scale support a leaner staffing profile
We have not yet arrived at the sample size that will support extensive analysis by medical home sub-groupings, but we did take a first look at whether we could see any preliminary patterns by "types" of PCMH. For example, did the data suggest any links between greater or lesser staffing and factors such as being a more mature site, having greater access to different resources, etc? With all caveats in force, and supported by gut-checking these numbers against conversations with real-world PCMHs, my own two most certain conclusions from this run of data relate to access to IT and practice scale.
Access to IT: More IT, Fewer Support Staff
PCMHs with more access to data, quantified as having access to more sources of data (such as EMR, claims data, and external quality reports) had fewer clinical support staff FTEs (2.2 clinical FTEs per physician) compared to lower-IT-access PCMH sites (3.4 clinical FTEs per physician). This makes sense when you look at the work these FTEs are doing in low-IT-access PCMH environments, where--for example--disease registries have to be built, populated, and maintained by hand.
Practice Scale: Bigger Practices, Fewer Support Staff (Per Physician)
A look at trends by practice size, measured by number of FTE physicians per site, supported another common-sense read on the PCMH terrain, which is that economies of scale apply to PCMH staff models.
The 25th percentile-sized PCMH sites ( 0-2.9 physicians) had an average of 3.4 clinical FTEs per physician, versus 2 clinical FTEs per physician at the 75th-percentile-sized sites (5.7 or more physicians per site).
Among PCMHs, our profiled sites tend to be more robustly staffed than average
As backstops to the data, I pulled out a few sites that we "know personally," meaning that we're familiar with the organizations and we know for a fact that these sites are functioning well as PCMHs today. Interestingly, these sites tended to have more robust staffing than the other PCMHs, reporting clinical support staff FTEs well above the average of PCMHs in the sample.
This is good reminder about keeping perspective on benchmarks generally. While it is useful to have in hand a statistic about the average clinical support staff complement for a PCMH, at the end of the day it's just that--an average. For methodological reasons which I won't go into now (but maybe in a future blog post), we didn't put a "quality" filter onto these PCMHs. It is possible that high-performing PCMHs might have greater or lesser support staffing than the average. All organizations will have to consider their own markets, operating contexts, goals, etc. when formulating their PCMH staffing models, using this data only as a point of departure.
On a purely anecdotal level, at this moment in time, I do think our case study institutions are more robustly staffed than the average PCMH and I don't think that is a coincidence. These are institutions that have taken the leap and invested extensive resources up front, in the belief that the model will yield ROI both now and in the future, as these new systems are built and the PCMH begins to make progress against its long-term strategic and operational goals.
However, it is equally possible that in the future, these same case study institutions will have lower staffing per physician FTE than the average PCMH of that time. Why? Because as systems are built and refined (e.g., disease registries are built and workflow re-organized around them), more advanced sites might not need as many clinical support staff FTEs as they have today.
Note on the Sample Size
Depending on the question, this data is based on a sample of as many as 101 sites, with about 68 sites having submitted data on all questions. I would call this respectable for supporting preliminary repiorting, but still relatively small -- especially in certain sub-groups. Many of our most important data points and analyses require multiple questions to construct, so we will need more sites that have answered all the questions in order to boost confidence in conclusions.
More reporting to come.
Benchmarking Initiative Still Open - Submit Data Now
The benchmarking initiative is now open to all participants in the Medical Home Project at no incremental cost -- Medical Home Project participation is part of an organization's membership in the Health Care Advisory Board.
At some point, we will be closing the initiative so as to finalize the data and move on to new analyses. If you would like to participate, please have your primary care and/or PCMH sites submit data now.
All participants will receive site-level benchmarks on staffing models, payer mix, patient access, IT implementation, and other key aspects of primary care transformation, including customized reports comparing their results to others in the project.
Sites do not have to be accredited as medical homes, in fact, they do not need to be medical homes at all. We are actively looking for more non-PCMH medical home sites to complete the survey. This will help us support more detailed analyses between PCMHs and non-PCMHs; the benefit of completing the survey for non-PCMH sites is that they can expect to receive a gap analysis illustrating where they are today versus where the average PCMH is now, in terms of staffing, IT access, etc.
As always, let me know if there's anything we at the Medical Home Project can be doing for you.