Who is your doctor?
That should be an easy question to answer. It’s not.
Most of the time, it is a fairly straightforward question. A person’s primary care provider (PCP) is responsible for quarterbacking the individual’s care pathways and is supposed to be looking at the big picture as well as the day to day flows of a patient. For most people who see their doctor and thus have a claim, the person that the claims analysis system says is their PCP and who they say is their PCP is the same individual. But not always.
There are a couple of cases where there is significant areas of conflict. The first is when a person indicates that Provider A is their PCP but they have been seeing Provider B, who fits normal PCP criteria, numerous times in the recent pass. This could be driven because their is a slight misclassification where Provider B is acting as a specialist. That is not too uncommon for Ob-Gyn and some infectious disease specialties where a provider will be dual classified as both a specialist and a PCP. Geriatricians are slowly becoming more likely to be dual credentialed as a PCP and a common specialty like cardiology.
Another case of confusion is when a person does not routinely see a provider for primary care. My dad did not see a PCP for most if not all of my childhood. My wife has not had a PCP appointment in years as she gets her only regular interaction with the medical system through her OB/Gyn or at the urgent care for an annual flu shot. These are not profoundly unusual situations. Most people in most years barely touch the medical system. Finally, there are cases where people are in flux for their care patterns where there is a major discontinuity in their utilization. This can be done through either an insurance switch or a disease burden transition.
A new study** highlights the problems of trying to guess who a person’s doctor really is.
The study sample included 15,624 patient-years from 3658 older adults with diabetes. Agreement was higher for algorithms based on primary care visits (range, 78.0% for majority match without a tie breaker to 85.9% for majority match with the longest time from first to last visit) than for claims to all visits (range, 25.4% for majority match without a tie breaker to 63.3% for majority match with the amount billed tie breaker). Percent agreement was lower for nonwhite individuals, those enrolled in Medicaid, individuals experiencing a PCP change, and those with >10 physician visits.
85.9% is pretty good but the last 14% is a major policy problem.
Why does this matter?
As we move towards alternative payment methodologies, accountable care organizations, global budgets, capitation systems and other provider side risk carrying payment methods, the question becomes important. Who is the doctor financially responsible for and who is at risk of a major claims blow-up? There are three major error types from a provider group perspective
The first is a patient is assigned to a group incorrectly. The group is getting paid for the patient’s care but they are not in a position to influence the individual’s care as they don’t see the patient on a regular basis. They will not complain too loudly if this is a patient who retrospectively and prospectively looks to be a highly profitable patient who uses very few services. If the patient has a chronic condition which requires expensive and chargeable to the global budget treatments, the practice group will scream.
The second error is when they routinely see a patient and get paid for that patient on a fee for service basis but the patient is not “their” patient. The provider is performing the role of PCP in most cases but is offloading performance risk to another practice. Here they will fight to get the patient re-assigned to their risk bearing entity if the patient is a probable profitable patient. If the patient is likely to be a time, cost and risk sink, the treating physician will not seek patient responsibility to be transferred to them if they can get away with it legally and ethically. This is effectively the mirror of the first scenario.
Finally, there are the patients who no one knows anything about. They are new to the data systems because they changed providers or they changed insurers or they just don’t go to the hospital. Providers will push back on assignment if they are being held to quality metrics where these people make up part of their denominator and there is absolutely no data on them. They will also push back if they are panel slot constrained and these ghosts occupy panel slots that the provider thinks they can fill with more claims paying individuals.
85% match rates are pretty awesome, but the study is using a single large academic medical center’s data set and consists of traditional Medicare patients. Academic medical centers tend to have fairly deep data sets as they have multiple specialties and usually multiple PCP clinics under the same roof. Smaller medical practices will have fuzzier and shallower data. Medicare patients by default are high touch patients. Being old means being in touch with the medical system. Younger, healthier and more transient populations will have fewer and less frequent touches in any given look-back period.
So who is your doctor?
** Dugoff, E. H., Walden, E., Ronk, K., Palta, M., & Smith, M. (2017). Can Claims Data Algorithms Identify the Physician of Record? Medical Care, 1. doi:10.1097/mlr.0000000000000709