Does the availability of more imaging providers per person result in more patient convenience?

Presenter: Mythreyi Chatfield, American College of Radiology

Abstract

Rationale:

Diagnostic imaging has been growing rapidly in recent years, and the share of non-radiologist treating physicians in high-tech, high-cost imaging has grown particularly rapidly. For a given medical condition, self-referral in imaging has been observed to result in higher rates of imaging utilization. Imaging self-referral is justified (i) as a patient convenience – the patient can get the imaging procedure with an office visit instead of making a separate trip, and (ii) as benefitting health—the patient’s treatment can begin on the same day rather than being delayed until after a visit to an imaging facility and then a return visit to the treating physician. Thus, two non-monetized benefits are propounded.

Objectives:

The goal of this study is to examine whether healthcare markets with more non-radiologists performing imaging have higher rates of imaging on the same day as the office visit.

Methodology:

Using an imaging claim as the unit of analysis, we will measure the effect of per beneficiary numbers of radiologists and non-radiologist imagers in the Hospital Referral Region (HRR, there are 306 HRR’s in the U.S., as defined by the Dartmouth Atlas of Health Care) on the probability of an imaging procedure being accompanied by an evaluation and management (E&M) visit on the same day and with the same physician.

Imaging claims from the 2005 Medicare 5% Research Identifiable Files will be used. For each claim, we will merge the area-level numbers of radiologists per Medicare enrollee and non-radiologist imagers per Medicare enrollee for each modality.

We will use logistic regression. Controls will include imaging modality, physician’s specialty (as a proxy measure for severity of disease), per beneficiary counts of radiologists in the area, patient characteristics (age, gender, comorbidities), and other area level characteristics that might influence characteristics of healthcare service (median income, urban area, etc.)

Results:

Of the self-referred images performed in offices, only 14.4% of MRIs, 25.4% of CTs, and 12.3% of nuclear medicine procedures were accompanied by an office visit on the same day with the same physician as the imaging procedure. For most straightforward x-rays, nearly three-fourths of the in-office self-referred imaging was performed on the same day. For less straightforward x-rays and for ultrasounds, the rates of same-day imagine were between 28.4% and 35.3%. Preliminary regression analysis indicates that while the odds of same-day imaging are lower for high-tech images than for x-rays, an increase in the per-beneficiary number of non-radiologists performing imaging does significantly increase the odds of a patient having an image on the same day as an E&M visit.

Conclusions:

If imaging by treating physicians is found to be associated with higher rates of same-day imaging then policies aimed at restricting financial self-interest in imaging should consider the potential implications for patient convenience and health outcomes. If, however, imaging by treating physicians is not found to increase same-day imaging, policy makers will have more freedom in making policy to restrict physicians’ financial self-interest in imaging.