COVID-19 transformed health care delivery almost overnight. In the United States, stay-at-home orders combined with the removal of certain legal, reimbursement, and regulatory barriers spurred telehealth to become a norm. The unprecedented shift away from in-office medical visits also opened a window into what telemedicine could be in a world evolving from the pandemic. And now, a new study by Stanford economist Liran Einav and his co-authors indicates that related concerns over costs and health outcomes may well be overblown. Their findings, released recently as a Stanford Institute for Economic Policy Research (SIEPR) working paper, rely on rich and granular data from Israel to provide clear empirical evidence of telemedicine’s effects on the quality and cost of care. Looking at primary care visits, the researchers show that the average cost per medical visit fell 5 percent even as telemedicine contributed to a slight (3.5 percent) increase in the combined total of in-person and remote visits. Just as revealing: Einav and his co-authors do not find any evidence of harm — such as misdiagnoses or improper treatments — to patients who received their care remotely. “Policymakers or doctors or patients should find more comfort that scaling up telemedicine is not harmful,” says Einav, a professor of economics at the School of Humanities and Sciences and the Tad and Dianne Taube Healthcare Fellow at SIEPR. “We don’t yet know where to draw the line in terms of the types of care that make sense for telemedicine, but we’re not seeing in our data any huge red flags in how it’s being used today.”
Close-to-ideal study conditions
Einav, who has done extensive research into health care markets, had long wanted to investigate the major concerns about telemedicine adoption, but sufficient data before the pandemic were lacking. When the outbreak hit, the speed with which Israel closed and reopened its economy in those early months provided a unique research opportunity to shed light on two perceived drawbacks of telemedicine: that patients would be more likely to seek care given its convenience and doctors would misdiagnose or overprescribe treatments. Either outcome would drive up costs. Like much of the world, Israel went into lockdown in March 2020 as coronavirus cases surged. On the eve of the pandemic, only about 5 percent of primary care visits took place over phone or video. After stay-at-home orders were issued, that fraction shot to 40 percent. Then when the country moved swiftly to loosen restrictions on economic activity in early May, telemedicine visits, which were mostly by phone, remained fairly high, accounting for about 20 percent of all primary care interactions. Given the ideal setting for analyzing telemedicine rates pre- and post-lockdown, the researchers gained access to data covering 12 million primary care visits from January 2019 to June 2020 at Israel’s largest health care provider, Clalit Health Services. The administrative records cover all of Clalit’s 4.5 million members, who comprise more than half of Israel’s population. The researchers analyzed 10.4 million primary care evaluations that started in the 14-month period before the lockdown against the 560,000 primary care visits — both in-person and remote — that started after the lockdown, between May and June 2020. Focusing on initial primary care visits allowed the researchers to test for the accuracy of diagnoses as telemedicine use went up. They also analyzed costs by tracking patient care for 30 days following the initial visit, which allowed them to capture lab tests, prescriptions, and referrals to specialists, among other treatments.
Accurate diagnoses, less expensive care
Einav and his collaborators find that, while total primary care visits went up 3.5 percent as telemedicine access rose, costs measured over the following 30 days fell by about 5 percent per visit. This decline in costs does not account for any potential cost differences between in-person and remote visits, but is driven by changes in the level of care that telemedicine patients received. For instance, patients of doctors who were more likely to use telehealth received 5 percent fewer prescriptions, 4.6 percent fewer referrals to other physicians, and 9.5 percent fewer referrals to radiology. The study found no differences in referrals for lab tests. Reassuringly, the researchers do not find any evidence that treatment decisions from telemedicine visits compromised patient outcomes. They look at three fairly common conditions — urinary tract infections (UTIs), heart attacks, and bone fractures — with symptoms that are unrelated to COVID-19 and also require additional care if misdiagnosed. For UTIs, for example, they track referrals for urine tests and UTI-specific antibiotic prescriptions. Across the board, the researchers find that physicians are just as likely to diagnose and correctly treat UTIs, heart attacks and broken bones whether the medical visits are in-person or remote. “I don’t think there’s much of an argument here: Medical care that makes sense to shift to telemedicine leads to a better, less expensive experience for everyone,” Einav says. “It’s a win-win as long as quality of care doesn’t suffer.”
Ripe for future research
Einav’s team uncovers other nuggets. They find, for example, that telemedicine patients are more likely to be female, wealthy, and sicker than in-person patients. Doctors who use telemedicine at higher rates tend also to be female, somewhat younger, and family medicine specialists instead of pediatricians. Einav says the reasons for these patterns are worth a closer look. So, too, are other lingering questions around telemedicine, including the role of supporting technologies like home tests and remote sensors, and the design of optimal reimbursement policies. Einav also notes that the study does not examine other potential benefits of telemedicine, including convenience and improved access for wheelchair-bound patients and others for whom in-person visits are challenging. Ultimately, he says, there is a need to figure out the right mix of modes. “COVID has been a watershed moment for telemedicine — much like it has been for remote work or business conferences over Zoom,” Einav says. “In all these instances, we’re looking at a hybrid approach going forward. And for the future of telemedicine specifically, more research and more statistical power is needed to know how we should regulate it and where we need to draw the line on the type of remote care that should be allowed versus not allowed.”
Source: siepr.stanford.edu
