Weighing the Cost of Success
All people are created equal - quality measures are not.
Some measures are as easy as placing a patient on a scale and documenting the height. Other measures dive deep into the socio-economic fabric of our communities to determine how to improve the birth weight of the children delivered in our practice. It can feel like trying to do better at the gym and switching unexpectedly to heavier and lighter weights.
In the realm of HRSA (and other) Quality Incentive programs, however, this is not taken into consideration. In other words, if you can improve an easy measure by 15% you get the payout. If you improve a hard measure by 15% you get the payout. Therefore, it behooves the clinic staff to think through improving as many easy measures as possible to improve the financial return on investment.
The grand idea behind measuring quality, however, is to provide an objective scientific basis for the best health care we have to offer. Who wouldn’t want that? The problem comes in deciding how to measure that and which measures are actually the most important.
HRSA has intentionally tried to align the UDS Quality Measures with other standardized measures to decrease the burden on health centers through the UDS Modernization Effort. This has resulted in some measures being combined and others being updated or removed.
Disclaimer
Before we go any further, I want to make the following disclaimers:
I assume that all of the Clinical Quality Measures (CQMs) are important for measuring quality care for our patients.
I assume that health centers want to improve in all of these measures and likely want to improve the ones that have the most impact on the lives of their patients.
I assume that health centers have limited staff, resources and time to improve these measures and are often caught between their desire to do better and the limitations to do so.
The following analysis is provided to help health centers focus their efforts on improving individual measures in order to improve quality incentive payments. This does not replace the vital role of the clinical care team who focuses on individual patients and their comprehensive needs.
With that in mind, I present a review of 20 HRSA quality measures that are used to calculate the QIA payments as previously outlined, along with an overview of a resource allocation optimization algorithm that we can use to help determine how the practice should focus on quality measures to improve their chances of receiving a quality incentive.
Measuring the Measures
Let’s say we wanted to maximize our financial payout based on HRSA QIAs. Let’s assume, as I said above, that we want to be the best at all measures right now. But, because we only have so many staff, we have to either pick being the best at some measures or the best at all measures later or being ok at all measures right now. To maximize our QIA payouts, we have to ensure that we receive a quality award, and we should try to receive each award that pays a bonus for each patient served. The National Quality Leaders pay the most, followed by the Health Center Leaders, followed by the Clinical Improver award.
Given this set up, we need a way to estimate how much work or how hard it is to change a particular measures. In decision analysis like the Analytic Hierarchy Process, decision makers are left with a difficult challenge of quantifying something that doesn’t feel quantifiable. We could do a pair-wise comparison to come up with an estimate that may be more scientific. I have chosen to use a rough intuition from being in the health center to provide a relative estimate of time based on four factors:
Is the measure a single item or more than one item?
Does the measure require a records review?
Does the measure require outbound care coordination?
Does the measure require patient behavior change?
I assign a 1 for yes, and 0 for no to each of these questions for each measure, and I came up with the following:
For each measure you can see that I have given a relative score for each of the 4 categories and then summed the total into the Complexity column. For some measures, the particular item required a number higher than 1. For example, because the Immunization measure requires 10 immunizations over 2 years, the composite piece was ranked higher than 1. In addition, controlling hypertension and diabetes requires a larger patient behavior change. Finally, the Low Birth Weight was ranked highest on patient behavior change because of how interconnected the effects of poverty, culture, background, and upbringing play are for the birthweight of a child.
After these were scored, the complexity was adjusted up by 1 to make everything have a minimum value of 1. This was labeled as the Relative Value Unit (RVU). In this first iteration, I assumed that 1 RVU was equivalent to 1 hour of time to change this measure for 1 patient. As I went through the measures, I felt like this was a reasonable starting place for the analysis.
I used these weights in the optimizer described below and discovered something curious. The optimizer rarely wanted to tackle measures that affected a large number of people because of how “costly” the measure was in terms of time. In my experience, however, for something like BMI a one time fix of an EMR rule or standing order can often benefit a large number of people. To account for this, I added a column of fixed cost and separated out the relative cost and adjusted for what helped the optimizer find more realistic solutions. Here are the adjusted results:
The changes for the Weight, Tobacco, Prenatal, Diabetes and Hypertension measures acknowledges that there are base line solutions that can benefit a number of patients that need to be considered. For example, with the BMI measure an EMR reminder or a standing order can help make a difference for 1,000 patients just as easily as it can for 100 patients. The prenatal and chronic conditions acknowledge that these are difficult measures and require a few weeks of someone’s time to draw up a game plan and solve.
Picking a Winning Solution
Now that we have the relative weights for each of the problems, we can now jump into a solution. Just like my kids preparing for camping, we often want to bring everything. Just like I tell my kids for camping that they can “bring what they can carry”, we have to look at quality improvement as being limited to what we can “carry”. This allocation of work time can be modeled and understood as a knapsack problem. In essence, we can calculate how much time it would take for us to become eligible for each of the HRSA QIA incentives. Then, we can estimate how much time we have total. Finally, we can figure out how to (metaphorically) put the most valuable awards in our knapsack at the end of the year.
Let’s look at an example. This particular health center served 35,640 in 2019. We’re going to assume they have about 18 providers, which means they probably have 1 or 2 people who have some responsibilities for quality improvement. We’re going to estimate that it’s about 1.75 FTE or just under 4,000 hours of time (this can be adjusted for real cases, of course). The clinic had moderate performance, with most measures in the 3rd Quartile Rank. Running the optimizer shows how this clinic could earn more than $100,000 just in quality bonuses (not to mention the PCMH or Access awards).
The solution indicates that by taking a first step of investing 637 hours of someone’s time, could result in almost $95,000 in QIA payouts. Instead of focusing on all of the measures, the staff should tackle:
Asthma Medication (though not in 2021 because the measure was discontinued)
Childhood Immunizations
BMI for adults
Use of Aspirin for IVD patients
Use of Statin for CVD patients
Dental Sealants
With this use of time, the Quality Leader for the organization could help review the staff members’ time to allocate it over the course of the year to ensure that these priorities were met. In Steps 2 - 6, there are marginal improvements as additional measures are improved, and finally, there are still 1,000 hours left over, that could be used to focus on access or other payor resources.
This ideal version assumes that focused attention can pay off. I firmly believe that clinics can make this kind of concerted effort if they will allow staff to focus on a small set of items with a clear goal in mind.
If you would like to learn more or have insights into how this might work for you, please let me know.