Access as Quality

Earlier this year I couldn't get a medically necessary physical therapy appointment with the clinic that was covered by my insurance. They didn't have any available appointments at all, not even months out. I ended up paying out of pocket to see someone good in my neighborhood, and I was lucky that I could.

Most people on Medicare can't do that. The rules that govern Medicare billing make private-pay arrangements for covered services legally unavailable unless a physician has formally opted out of Medicare entirely. That's a significant commitment most practices cannot afford to make, and one that physical therapists aren't permitted to make at all. When the system doesn't offer what you need, the exit I had (not an equitable or ideal one, either) is sometimes structurally closed off entirely for the patients this newsletter is mostly about, the same patients we all presumably hope to become.

Why is it hard just to get an appointment with someone who has the expertise and training you need? I think the reasons have much more to do with the structure of how health care systems and clinics get paid than with simple supply and demand.

You've probably heard of "fee-for-service" as a payment model. It's intuitive: insurers pay for each service as it's delivered. That's how Medicare has paid since 1966, and it creates predictable problems. More visits, more tests, more prescriptions means more revenue. The value of not doing something, or of taking time to think carefully before doing anything, never shows up anywhere that gets paid.

Value-based care was supposed to fix that, by tying payment to outcomes instead of volume. It's a genuine improvement in theory, and I don't think the people who built it were wrong to try. But there are real issues. The outcomes and processes it tracks, the diabetic eye exam, the blood pressure in range, the cancer screening, whether a patient landed back in the hospital within thirty days, are the same measures applied to thousands of patients at once. They are certainly useful and in some ways an improvement over just quantity, but they can't account for heterogeneity of goals. That is to say, for what an individual person is actually trying to get out of their care.

What current quality metrics often don't track is whether you or your family member could actually get an appointment with someone who knows them in the next month. They don't measure whether that clinician had enough time in the visit, or after it, to think carefully about what you told them. Some emerging quality and aging-health certification programs, despite their good intentions, even add documentation burdens that take away from that time to think and listen. We don't know if there was capacity for a follow-up call the next week, or whether the next available appointment is four months out. These things are not tracked for billing or as a core part of any model; they are left to mid-level management.

I am not the first to think of this. Health services researchers have sorted quality into three categories for many decades now, since a physician named Avedis Donabedian first laid it out: the structure of care (the staffing, the time, the capacity a system has to offer), the process of care (what actually happens between a patient and a clinician), and the outcome (what happens to the patient afterward). Almost everything value-based care measures is process and outcome. Structure, the conditions that make good process and good outcomes possible in the first place, is the piece almost nobody bothers to pay for. Access, continuity, and appointment time are structural measures. They're not new. They're just neglected.

Let me give an example from my own field.

One of the most important things a clinician can do for an older adult on many medications is to cautiously and intentionally take some away (“deprescribing”). Medications accumulate over years and across providers. What made sense at one point may be causing real harm now, quietly, in ways that look to the untrained eye like "aging" rather than like a side effect. Getting someone off a medication that is driving their blood pressure dangerously low, or making them unsteady on their feet, or causing confusion, can change their life. I have seen it, and it is an important and meaningful part of what I can do for patients. But it is not easy, simple, or quick.

Safely removing a medication requires time to think as well as close follow-up. You make a careful change, then you check in the following weeks, and the weeks after, to see what happened. A clinic booked four months out cannot do this safely, especially if you are only seeing a patient one time. And so the thing that would actually help often doesn't happen, not because anyone decided against it, but because the access required to do it safely isn't there. Continuing a medication someone has been on for years is fast, defensible, and financially neutral. Stopping one is time-intensive, cognitively demanding, and poorly reimbursed.

In fee-for-service, clinics get paid less for having longer appointments and seeing fewer patients per day. In value-based care models, they get paid less for taking on higher-risk patients who are more likely to have complications or bad outcomes or for taking on fewer patients so they have the time to think through a long med list. There is no clear financial incentive driving the kind of care we know works better for older patients.

So much of the time, it just doesn't happen. Which medications are right for which patients in which circumstances is too heterogeneous to capture in any existing outcomes measure that I know about. The structural conditions that allow good decisions to happen, though, are not nearly as hard to measure. A patient has to be able to see the same provider multiple times over years, and that provider has to have the time to think carefully. Or, if time is limited, they need the support of a strong interdisciplinary team that also knows the patient over time. This is structural, and it seems a much easier and more reliable target than finding the right outcome measures in a heterogeneous population with diverse and variable needs. Pay for the structure that works.

The research on what actually makes primary care work is not ambiguous. Seeing the same clinician over time is associated with fewer hospitalizations, fewer emergency department visits, lower costs, and lower mortality. This has been studied extensively. The American Board of Family Medicine developed a validated, nationally endorsed quality measure for continuity of care. Value-based payment models presumably incentivize continuity through the outcomes they measure, but they still don't measure or reward building continuity into the structure of health care directly.

What's actually rewarded in the current payment system is seeing a high volume of patients. A primary care physician managing a panel of two thousand people or more, which is common, cannot offer the kind of continuity the research says matters. The schedule built to serve that panel cannot absorb the follow-up visit needed to safely adjust a medication, or the longer conversation needed to understand what someone is actually trying to be able to do.

And here is the part most people don't see. The same conditions that make it hard to provide good care are driving clinicians out of primary care entirely. Nearly half of primary care physicians in the United States report experiencing burnout, and more than a third of those plan to stop seeing patients within the next few years. When that happens, patients lose their doctor. The continuity that the evidence says matters gets severed, the shortage that results puts even more pressure on the clinicians who remain, and the cycle continues.

The conditions that would make primary care work better are also what would make clinicians want to stay in it. Smaller panels. Adequate time per visit. The ability to actually follow up. These are not luxuries. They are the conditions that produce the outcomes everyone says they want.

What would it look like if these conditions of good care (access and continuity) were the quality measures themselves? Not whether a checkbox got completed, but whether a patient could get a timely appointment with someone who knows them, and whether there was capacity to follow through on what was decided together.

I don't know exactly how you would build that into a payment system, and I would genuinely like to know if anyone already is doing this. I think naming that this element of structure matters, and paying for it, would show us us a great deal about where the real gaps are for patients and for the doctors trying to care for them (and move the needle on closing those gaps, all in one.)

I'm not a health policy expert, and this corner of the system is dynamic and genuinely confusing. What I know is what I see from inside the clinic, and what I've read trying to understand why. If you know this landscape better than I do, or know someone who does, or if there's already policy work moving in this direction, I want to hear about it.

Reply and tell me what you think.

-Rachel