Concern about potential telehealth surge in low-value care unsupported by real-world data

Key takeaways:

• Medium and high telehealth use decreased some low-value services like thyroid testing.
• The results are reassuring due to concerns that telehealth might increase unnecessary visits or tests, a researcher said.

Health care practices that used telehealth services more frequently had a lower propensity to engage in some low-value services, results from a cohort study published in JAMA Network Open showed.

Data from the retrospective analysis revealed no links between the frequency of telehealth services offered by practices and ordering of low-value health care services.

The findings “provide some reassurance that higher levels of telehealth were not associated with uncontrolled use of low-value care on average,” Terrence Liu, MD, MS, a clinical instructor at the University of Michigan Institute for Healthcare Policy and Innovation, told Healio. “We initially hypothesized that more of the low-value laboratory tests would have decreased among practices with higher levels of telehealth, so our results were a little surprising, but not totally unexpected.”

Healio previously reported that low-value services contribute to a significant amount of annual U.S. health care spending and can result in psychological and physical harm to patients.

In their analysis, Liu and colleagues assessed the effect of high, medium and low telehealth usage on the rates of several low-value tests and procedures relevant to primary care providers. These services included:

• colon cancer screening in people older than 85 years;
• cervical cancer screening in women older than 65 years;
• triiodothyronine testing for patients with hypothyroidism;
• vitamin D testing;
• CT scans, head imaging and back imaging for people with uncomplicated sinus infections, headaches or low back pain, respectively; and
• blood-based prostate cancer screening in men older than 75 years.

The analysis used data from 577,928 Michigan residents insured through Medicare (mean age, 76 years; 57% women) who saw providers across 2,552 primary care practices between 2019 and 2022.

The researchers found that high practice-level telehealth use corresponded with reduced rates in average marginal effects (AME) of 2.9 (95% CI, 5.3 to 0.4) per 1,000 beneficiaries for low-value cervical cancer screenings and 40 (95% CI, 70 to 9) per 1,000 beneficiaries for low-value thyroid testing.

The rates of low-value cervical cancer screenings (AME, 2.2; 95% CI, 4.2 to 0.3) and low-value thyroid testing (AME, 57; 95% CI, 88 to 26) also decreased in practices with medium telehealth use.

“It was not surprising that low-value cervical cancer screening decreased among practices with higher levels of telehealth given that this is typically performed as an in-person examination,” Liu said.

The rates of the rest of the services remained unchanged regardless of telehealth usage.

“Since our study examines telehealth at the practice level, it is hard to directly translate our results to the individual primary care provider,” Liu told Healio. “Nonetheless, we hope our findings can highlight areas in which telehealth can influence individual clinician medical decision-making. We hope that our work encourages PCPs and patients to have informed discussions of both the benefits and limitations of telehealth.”

The researchers acknowledged some study limitations. For example, they were not able to examine all low-value care services, such as low-value medication prescriptions. In addition, the study results may not be generally applicable to the broader Medicare beneficiary population because study sample only included Michigan residents.

However, “we hope to repeat our analyses beyond Medicare beneficiaries living in Michigan [to] all traditional Medicare beneficiaries nationwide to see if our results are similar or different when examining a broader population,” Liu said.

References:

• Does more virtual care mean more low-value care? Study suggests no. Available at: https://www.eurekalert.org/news-releases/1063961. Published Nov. 7, 2024. Accessed Nov. 7, 2024.
• Liu T, et al. JAMA Netw Open. 2024;doi:10.1001/jamanetworkopen.2024.45436.