Telehealth as a Blunt Instrument: Rethinking Virtual Care and Diagnostic Value
A commentary and companion investigation published in JAMA Network Open reveal that virtual annual visits are associated with meaningful reductions in both high- and low-value test ordering and completion—raising urgent questions for primary care physicians and health system leaders navigating the future of telemedicine.
Background and Policy Context
Telemedicine transformed from a fringe modality to a cornerstone of primary care delivery during the COVID-19 pandemic. In the years since, virtual visits have remained widely used across health systems in the United States. Yet the policy foundation supporting their continued reimbursement remains precarious. As of this writing, Congress faces a deadline: without legislative action, Medicare coverage for non-behavioral health telemedicine services was scheduled to lapse on January 31, 2026. Against this volatile backdrop, a rigorous new study published in JAMA Network Open provides timely and clinically consequential evidence about how virtual care affects one of primary care's most fundamental functions—medical testing.
In an invited commentary accompanying the original investigation, Dr. Richard L. Kravitz of the University of California Davis framed the stakes plainly:
"In this policy milieu, additional information on the benefits, risks, and costs of telehealth services is sorely needed."
The Study: Design and Population
The original investigation by Ganguli, Daley, Mehrotra, Rosenthal, and Cutler—affiliated with Harvard University and Brigham and Women's Hospital—drew upon electronic health record (EHR) data from the Mass General Brigham health system, one of the largest integrated delivery networks in New England. The cohort study examined annual visits conducted between January 1, 2022, and October 25, 2023, across 234 primary care practices, ultimately analyzing 22,547 propensity score–matched visits from 20,948 patients (mean age, 51.0 years; 69.2% women) across 87 clinics.
The investigators focused on annual wellness visits deliberately, explaining that these encounters are uniquely suited to studying testing behavior because
"there are a range of point-of-care and scheduled tests that can be, and often are, provided in these visits that are well-defined in prior literature as high value or low value."
Fifteen tests were classified as either high-value or low-value based on established clinical guidelines, including those from the ABIM Foundation's Choosing Wisely campaign and the U.S. Preventive Services Task Force. High-value tests included HbA1c, lipid panels, cervical cancer screening (for eligible women aged 21 to 65), colorectal cancer screening (for patients aged 45 to 75), and mammography (for women aged 40 to 74). Low-value tests included screening electrocardiograms, routine metabolic panels, thyroid-stimulating hormone levels, and cervical and colorectal cancer screening in populations where guidelines do not recommend these interventions.
Key Findings: A Dual Effect on Test Ordering and Completion
The results were striking in both their direction and magnitude. Across all 15 high- and low-value tests, virtual visits were consistently associated with lower rates of test ordering and lower rates of test completion—even among tests that were ultimately ordered. The differential was somewhat more pronounced for low-value tests than for high-value tests, and markedly more pronounced for point-of-care laboratory tests than for scheduled tests such as colonoscopy and mammography.
Specifically:
- High-value tests were 14.3% less likely to be ordered at virtual visits (relative to an in-person ordering rate of 54.8%) and 13.1% less likely to be completed (relative to an in-person completion rate of 87.7%).
- Low-value tests were 19.3% less likely to be ordered (relative to an in-person rate of 27.8%) and 17.3% less likely to be completed (relative to an in-person rate of 91.1%).
- Point-of-care laboratory tests showed the largest declines, with ordering falling 18.5% and completion falling 16.3% at virtual visits relative to in-person encounters.
- Scheduled tests declined less steeply—ordering fell 11.6% and completion fell 6.2% at virtual visits—though the pattern was not uniform across test types.
These effects were not limited to tests that require physical presence in the examination room, such as cervical cancer screening and ECGs—which were nearly never ordered during virtual encounters (−99.6% and −85.4%, respectively). Even laboratory tests that can be completed at an external site after either a virtual or in-person encounter showed clinically meaningful declines in both ordering and completion.
Unpacking the Mechanisms
Why would a virtual visit suppress test completion even for scheduled tests that require a separate laboratory or imaging appointment regardless of visit modality? The authors offer several mechanistic hypotheses. For point-of-care laboratory tests, the friction of not being physically present in the office—where a clinician might hand an order directly to a patient and direct them to an adjacent phlebotomy suite—may explain much of the gap. For scheduled tests such as mammography, the loss of in-person clinical workflow dynamics, including front-desk staff who routinely facilitate scheduling at check-out, may play an important role.
Dr. Kravitz's commentary introduced an additional layer of interpretation grounded in social psychology. Research on communication modality and compliance suggests that video-based interactions may make it easier for both parties to defer or negotiate around requests.
"In keeping with prior social psychological research, telehealth may blunt such testing by making it easier for physicians to defer, deny, or otherwise negotiate around patients' requests."
This observation has practical clinical relevance. As social media health influencers and direct-to-consumer laboratory services continue to drive demand for marginally indicated or frankly low-value tests, the virtual encounter may—perhaps inadvertently—serve as a structural buffer against unnecessary testing. However, that same buffer operates with equal indifference on tests that patients genuinely need.
The High-Value Care Gap: The Most Urgent Clinical Concern
While reduced low-value testing carries some theoretical benefit—particularly in containing care cascades that amplify spending and patient anxiety—the parallel decline in high-value testing is far more concerning. Suboptimal rates of guideline-recommended screening remain a persistent and well-documented problem in primary care settings across the United States. The finding that virtual visits are associated with significantly fewer HbA1c tests, lipid panels, mammograms, and colorectal cancer referrals suggests that a modality increasingly relied upon by patients and health systems alike may be quietly contributing to preventive care gaps.
Dr. Kravitz acknowledged the tension inherent in this dual finding:
"Telehealth has many virtues, but with respect to test ordering, it is not a precision instrument. While continuing to monitor the implications of virtual care for health care costs and outcomes is important, future research should focus on finding ways to deliver care when, where, and how it is needed, regardless of modality."
Implications for Practice: Targeted Interventions and Emerging Solutions
The authors of the primary study are clear that their findings should not be read as a wholesale indictment of telehealth. Rather, they identify a structural opportunity. Physicians and health systems can take deliberate steps to counteract the inadvertent suppression of high-value testing during virtual encounters. These include formalized front-desk follow-up protocols after virtual visits to facilitate test scheduling, the expanded use of EHR-embedded care gap alerts during virtual encounter workflows, and patient-portal–based reminders triggered by visit type or screening eligibility status.
Looking further ahead, Dr. Kravitz envisions a role for artificial intelligence agents to address the post-visit follow-through gap:
"It is also possible to imagine postvisit chats orchestrated by artificial intelligence (AI) agents in which patient understanding of key action items is assessed and supported."
The analogy he offers is instructive: if AI-powered travel platforms can coordinate complex international itineraries, deploying similar technology to ensure a patient schedules a colonoscopy or HbA1c draw after a virtual annual visit is entirely within reach.
Limitations and Generalizability
The study carries important limitations that clinicians should weigh. The Mass General Brigham health system is large and academically affiliated, serving a predominantly White, English-speaking population insured primarily through commercial plans. Generalizability to safety-net settings, rural practices, or more racially and socioeconomically diverse populations remains to be established. There is also the inherent possibility of residual confounding: patients who choose virtual annual visits may differ in their health engagement and follow-through behaviors in ways that available EHR variables cannot fully capture. Finally, the study did not assess downstream clinical outcomes associated with changes in test ordering and completion—a critical gap for future research.
Conclusion
This investigation offers the most granular EHR-based evidence to date regarding how telehealth modality influences both clinician test-ordering behavior and patient follow-through. Its central message for practicing physicians is sobering and actionable in equal measure: telehealth reduces both low- and high-value testing, and the system is not well-designed to distinguish between them. As health systems expand virtual care infrastructure and policymakers debate continued Medicare coverage, practice leaders must invest in the complementary tools—technological, administrative, and clinical—that preserve the value of high-value care regardless of how the visit is delivered.
.png)
.png)
.png)
.png)
.png)
.png)