In this special edition of the MGMA Executive Session podcast, regular host David N. Gans, MSHA, FACMPE, senior fellow, MGMA, takes a turn being asked the questions by Andrew Hajde, CMPE, senior industry advisor, MGMA, about his recent Data Mine column in the July issue of MGMA Connection magazine.
MGMA members enjoy exclusive access to the print and flipbook versions of the Data Mine column.
Hajde: Before we get started in our main topic for today, do you have any initial insights?
Gans: The most recent Data Mine article, “The secret to staffing success,” appearing in the July print edition of MGMA Connection magazine, looks at how the number of staff relate to the profitability of the practice, and the secret of practices that have better financial success. Is it really having more staff, or is it other things? I think the data are extremely interesting.
Hajde: Dave, as you know, I worked in group practice management for many, many years. When I read your article, I was extremely excited about all the different correlations of data and different things that were going on. When I first read through the article, I wasn't really sure how much to expect to see an exact correlation between that total revenue and total staffing ratios. So, I kind of got some surprises when I looked at some of the scatterplots or other information that were out there. And the other part that kind of surprised me was just the continuing trend of total staffing patterns between physician-owned and hospital-owned practices and how those vary from each other.
Gans: Having worked with the MGMA data for so long, I was able to utilize the raw database of information and create variables that are not normally looked at in practices —looking at, in this case, hundreds of practices that are similar and saying, “Are there patterns of information?”
Hajde: My first question for you, that kind of ties directly with that data — specifically with that gap we were talking about. It was interesting for me to see how large that gap was between the private practice and those integrated health systems, those hospital-owned practices and their overall staffing. It showed that 36% of private practices — a large portion of them — had a total staffing FTE ratio of 5.1 to 6.0 staff members, whereas an integrated system, 35% of those practices had between 2.1 and 3.0, total FTEs. In your opinion, Dave, is that primarily that big variance there primarily because a lot of those integrated practices have a lot of their departments — including credentialing, contracting, billing, or other positions — outside of the practice itself?
Gans: What we're seeing, a statistician would call a normal distribution, or a bell-shaped curve, but we see two different curves. One is that normal distribution for practices that are part of health systems, and these practices have much lower staffing than their private-practice equivalents. We were looking at multispecialty groups with primary and specialty care, which gives some homogeneity to the type of practices we’re observing. The major difference in the data happened to be ownership. And in these practices that were hospital-owned or part of a hospital system, what we saw was more than half of their staffing was less than four employees per doctor, whereas among physician-owned groups, only 4% of the practices staff at this level. In other words, half versus 4% — what a difference! But if you look at the actual graphics displayed in the article, you can say, “Oh, this looks like a pretty normal distribution,” is just that hospital systems, typically staff at lower levels.
Hajde: Absolutely. It's just pretty distinct how much variety there is between those two models.
Gans: Of course, there's some reasons why this occurs. And that's why we oftentimes want to look not only by type of practice, but by ownership as well. For example, in these hospital-system practices, many of these services of medical group have been moved out of the practice and into the health system. A great example would be how imaging and laboratory are seldom found in the hospital medical group, but are found in a centralized imaging center or centralized laboratory. The same thing occurs with certain administrative functions — credentialing, for example, is almost always centralized. Human resources management would be centralized, marketing would be centralized. So consequently, you don't have the need to show the staff in the practice, but you're still getting the services and those staff just appear elsewhere.
Hajde: When you look at that clinical support staff specifically, what kind of differences are you seeing? Because that's usually where the rubber meets the road in that area?
Gans: When I looked specifically at nursing staff, this is the sum of registered nurses (RNs), medical assistants (MAs) and licensed practical nurses (LPNs). What we see is, there's similarity — there's not that much of a difference, but we still see that physician-owned practices typically have more clinical staff supporting their doctors. Now, I think that this is evidence of when you own the practice, and if you want it and you're willing to pay for it, you get it. If you're part of a health system, you have to make those requests through a much more formal bureaucracy. And I think there's some pretty strong evidence that health-system practices starve themselves a little. They oftentimes make do with fewer staff than optimum than if the doctors had those choices themselves.
Hajde: When you see that increased total number of staff, though, what happens in terms of the total revenue? There appears to be a clear correlation between the more total staff you have, the more revenue you have. What are your thoughts are on that?
Gans: In addition to the graphics, the table of information looks at a measure of productivity — total procedures per FTE physician — as well as the financial indicators. And the first place I always want to look at before you start identifying what occurs in finance is what happens in productivity. Because obviously, productivity is highly correlated with total revenue. This was the first thing that I saw that said, we have we have a strong relationship between having staff and having productivity. This is looking across all practices and all types, aggregating both the hospital and physician groups: the lowest level of productivity — 5,430 median total procedures per FTE physician — occurred at the lowest level of staffing, of two to three FTE employees per doctor. The highest level of productivity — 16,000 procedures — was with those that had six or more employees per doctor.
Some of these procedures are the results of, for example, lab procedures or radiology procedures — not necessarily the direct work of hands-on care by a physician. Also, in the context of this analysis, I did not separate out what happens with nonphysician providers, but I found similar ratios of nonphysician providers, nurse practitioners and PAs per doctor. We're basically holding that somewhat constant, but their productivity is included in these total procedures, as, of course, will be the results of their productivity and the revenue they produce, and also their costs. So, it gives us further insights into what's happening. More staffing yields more productivity.
Hajde: One of the things I also noticed in those scatterplots were the outliers. When we look at some of those practices, there's a few of them out there that are generating very high revenues with much lower levels of FTEs. What are those practices doing differently to accomplish that, because I think that's everyone's dream to make that happen?
Gans: Looking at the table with information where we have median revenue, total medical revenue, median total operating costs and median total revenue after operating cost, the same pattern that occurred in productivity occurs here —the practices with the most employees have the greatest amount of revenue, they also have, by far, the greatest amount of expense, as you would expect Having more employees means you have to have higher salaries, higher fringe benefits, more space, more equipment and other things. Fortunately, we see these practices have higher net revenue after operating cost.
Now, when we get into the scatterplots, it's really fascinating. We can see some practices’ staffing at very low levels — one or 1.5 or two FTE employees per doctor; others staff at much higher levels: eight, nine or 10 employees per physician. You can see that, for a certain level of staffing, what was that practice’s total revenue, for example? In general, practices that have more staff have more revenue. But more important, there is variance; it's not consistent. Some practices have relatively modest staffing and very high revenue. Others have greater staffing and lower revenues, but there's a pattern.
Hajde: That was kind of kind of what I imagined in terms of some of the variability, but most of them were falling right in the middle where you'd expect. What were some of the other items that you were surprised about when you did all your data analysis?
Gans: It’s very interesting to see how, among revenue, there was far greater clustering as practices have more staff. There's a statistical measure that looks at the amount of variation in the data. Statistically, we refer to this as a coefficient of determination, and it's oftentimes expressed as a term, “R-squared.” This tells the degree of variation that can be explained by the linear regression line. Since the horizontal axis is number staff, we can use this to say, “How well does the number of staff explain the amount of variance in the vertical axis?” — in this case of, revenue expense or revenue after expense.
The strongest overall relationship was in revenue; we had an R-squared of 0.428 —almost 43% of the variation in revenue can be attributed to FTE support staff. I thought this was a this is very telling. At the same time, we started looking at expenses. You see a very strong cluster of data that shows, as more staff increase, we get higher expenses. There are a number of outliers, where some practices have expenses twice as much per doctor than the than the bulk of the others. This causes the statistical representation to get much lower: 0.195, which says only about 20% of the variance in expense can be attributed to staffing. That sort of makes sense, because you have high cost-of-living areas, low cost-of-living areas; you have higher salaries, lower salaries for staff. It may depend what staff you have; it may also depend on what are you spending on other costs in the practice.
Lastly, when we look at the revenue after operating expense, clusters get a little tighter, and we show about an R-squared of 0.318, or about 32% of the variance in profitability can be attributed to staffing. But even so even at this point, we see a substantial amount of variance at the same levels of staffing. My suggestion, and I point out in the article, median staffing levels for multispecialty groups with primary and specialty care is five FTE employees per doctor. And at that one level, we saw a practice that had revenue after operating cost of just over $100,000. Now, obviously, this is a hospital-owned practice that will run an operational loss after you reduce this amount by the compensation of the doctors and other providers. But there are other practices here at virtually the same staffing level — they have more than $700,000 of revenue after operating cost per doctor. This is a sevenfold difference. So again, substantial amount of variance, even though statistically we can explain the opportunity for profit by having more staff.
Hajde: One of my key questions with this whole analysis you did was, what are those key takeaways you’d bring to an executive about if they're considering adding more staff to grow their revenue? What would you recommend?
Gans: First off, study the problem. Research what other organizations are doing. You're not just adding more staff, you want to add the right staff, doing the right things. Why do we get this variance? Well, half of it is having more staff. The other half is how you train the staff., What incentives do they have to work hard? What is the degree of supervision? What are their skill levels? How well are they trained? All these other factors that contribute to the productivity the doctor.
The other factor is, what is the culture of the practice? Some practices have a culture that may be focused more on lifestyle of the providers, the doctors — other practices have a high-productivity focus. A doctor who would thrive in one type of practice would not do well in the other. For an executive wanting to look at how do we optimize our productivity and our revenues, I would look at your staffing levels and look carefully at what these staff are doing.
Now, fortunately, MGMA has several publications — much of the data in this article comes from a book called Staffing the Medical Practice, by Dr. Deborah Walker-Keegan and Elizabeth Woodcock. They went into staffing models and staffing strategies that would give excellent ideas to an executive. I'd strongly urge people to look at that text, and what they what they're suggesting, then put it in the context of your practice. Don't just look at having more staff, but getting the right staff doing the right things.
Hajde: One final question for you. What else would you want the audience to know about using benchmarking and data from companies like MGMA to increase their market share efficiency, or to help give them a leg up on their competition?
Gans: Number one, benchmarking is a tool that allows you to say what are other people doing. In the context of this article and in much of the MGMA DataDive software, there’s opportunity to do comparative benchmarking — benchmarking your performance to that of similar practices and look where the variance occurs. Are you better or worse than other practices? If you're better, pat yourself on the back; more importantly, congratulate your staff. If you're worse in an area, focus your attention — your most valuable resources, your time —where you have problems.
Now, benchmarking is not only comparing yourself to other practices, but also comparing yourself to the processes those practices use. MGMA has expanded our DataDive Practice Operations data to include information on what other organizations are doing, such as hours of staffing, appointment schedules, appointment waiting times. You can model what practices are doing as well. Our benchmarking also lets us look at best practices. I talk about some of this, along with my co-author, Greg Feltenberger, in the book Benchmarking Success. We get into a lot of detail on how to best benchmark and to use the information. One of our mantras has been, “if you can't measure it, you can't manage it.” Well, measurement is the key to good management, and that's what benchmarking is, as well as using the data in this context to look at what your staffing could be.
