In recent years, telemedicine has been seen as a promising area of innovation in healthcare delivery. But for reasons that include licensing, reimbursement, and regulatory restrictions, its adoption had remained largely limited to small-scale programmes targeting remote locations, late hours, or specific conditions (Tuckson et al. 2017). For example, in the US, remote visits accounted for less than 1% of primary care visits before 2020 (Dorsey and Topol 2016).
This landscape abruptly changed in 2020. In the wake of the Covid-19 pandemic, facilitated by rapid changes in regulation and increased demand, telemedicine use rate rose sharply. At the height of the pandemic, telemedicine visits accounted for a substantial share of visits, and the share has remained persistently higher than the baseline ever since (Mehrotra et al. 2020, Patel et al. 2021).
This watershed moment will likely change the landscape of primary care in the coming years, with new models of healthcare delivery that mix in-person with remote care. But while remote care is clearly convenient, there remain concerns that it might reduce care quality or increase costs, as it could lead to unnecessary visits, involve decreased diagnostic accuracy, and result in duplicate visits and overuse of downstream services such as testing and specialist care. Existing evidence about quality measures, such as the proper use of antibiotics, remain mixed (e.g. Shi et al. 2018, Ray et al. 2019).
In a new paper, we study the impact of increased access to telemedicine on care cost and outcomes (Zeltzer et al. 2021). We use rich administrative data from Israel, which had widespread telemedicine adoption during the first (March–April 2020) COVID-19 lockdown, followed by a temporary return to near normalcy after the lockdown resulted in a (short-lived) successful mitigation of the virus.
Figure 1 shows both the daily number of COVID-19 cases and the overall share of telemedicine visits via phone or video as a percentage of all primary care visits. Before the first lockdown, telemedicine accounted for about 5% of all primary care visits. It peaked at around 40% during the lockdown and remained high, at around 20%, during the post-lockdown period, which was characterised by a low incidence of COVID-19.
Figure 1 COVID-19 and telemedicine use in Israel
Source: Clalit Health Services.
The impact of telemedicine on care cost and outcomes
The fact that some primary care physicians adopted telemedicine more extensively than others allows us to compare primary care episodes of patients of high adopters to patients of low adopters. To account for potential differences in patient mix and practice patterns between high and low adopters, we adjust for each group’s pre-lockdown baseline. And since patients’ choice of remote versus in-person setting may depend on their medical conditions, we consider visits across all settings and providers, thus measuring the overall impact of increased access to telemedicine on healthcare cost and outcomes.
We find that access to telemedicine results in a slight increase in primary care utilisation (3.5% over the pre-COVID-19 baseline) and no significant increase in overall costs. The evidence is inconsistent with the concerns of increased care intensity, as we find that patients with access to telemedicine appear to receive less intensive treatment; they receive slightly fewer prescriptions and referrals and spend overall 5% less on healthcare services during the 30 days following an initial primary care visit. Although patients with access to telemedicine are 8% more likely to follow up with a physician within a week of the initial visit, the bulk of these follow-ups are with the same physician that provided the initial visit and they tend to be done remotely, so care continuity does not seem to be compromised.
We analyse more granularly the impact of increased access to telemedicine on the diagnosis of specific medical conditions – urinary tract infection, heart attack, and trauma (conditions that are common, unrelated to COVID-19, and may result in serious consequences if misdiagnosed). We find no evidence for decreased accuracy or increased likelihood of adverse events. Taken together, these results suggest that the increased convenience of telemedicine does not compromise care quality or raise costs.
The road ahead
More research is required to understand the many aspects of this unprecedented and universal shift in healthcare delivery. But just like remote work will likely change the labour market landscape even after the pandemic ends (Kakkad et al. 2021), telemedicine will likely be part of the future of healthcare delivery. There remain many open questions about the role of supporting technologies, the design of optimal reimbursement policies, and the optimal ways to combine telemedicine and in-person care for different conditions.
Dorsey, E R, and E J Topol (2016), “State of telehealth”, New England Journal of Medicine 375(2): 154–61.
Kakkad, J, C Palmou, D Britto and J Browne (2021), “Anywhere jobs and the future of work,” VoxEU.org, 10 July.
Mehrotra, A, M Chernew, D Linetsky, D Hatch and D Cutler (2020), “The impact of the COVID-19 pandemic on outpatient visits: A rebound emerges”, To the Point (blog), Commonwealth Fund, updated 19 May.
Patel, S Y, A Mehrotra, H A Huskamp, L Uscher-Pines, I Ganguli and M L Barnett (2021), “Trends in outpatient care delivery and telemedicine during the COVID-19 pandemic in the US”, JAMA Internal Medicine 181(3): 388–91.
Ray, K N, Z Shi, C A Gidengil, S J Poon, L Uscher-Pines and A Mehrotra (2019), “Antibiotic prescribing during pediatric direct-to-consumer telemedicine visits”, Pediatrics 143(5).
Shi, Z, A Mehrotra, C A Gidengil, S J Poon, L Uscher-Pines and K N Ray (2018), “Quality of care for acute respiratory infections during direct-to-consumer telemedicine visits for adults”, Health Affairs 37(12): 2014–23.
Tuckson, R V, M Edmunds and M L Hodgkins (2017), “Telehealth”, New England Journal of Medicine 377(16): 1585–92.
Zeltzer, D, L Einav, J Rashba and R D Balicer (2021), “The Impact of Increased Access to Telemedicine”, NBER Working Paper 28978.