assessment of a skin lesion to determine the potential need for biopsy) than others (e.g., prenatal care for a high-risk pregnancy, diagnosing ear pain). The most-effective policies will differentiate between different types of telehealth, different use cases, and different patient populations because impact on quality, costs, and access will vary. Although it would be ideal to pursue all three goals with a particular telehealth service, economists typically argue that it is possible to achieve one, or possibly two, but almost never all three (Carroll, 2012). In the following sections, we consider the evidence for how well telehealth might achieve each of the three goals. Abbreviations ARI acute respiratory infection CARES Act Coronavirus Aid, Relief, and Economic Security Act CBO Congressional Budget Office CMS Centers for Medicare & Medicaid Services COVID-19 coronavirus disease 2019 DEA Drug Enforcement Administration DTC direct-to-consumer FQHC Federally Qualified Health Center HHS U.S. Department of Health and Human Services OUD opioid use disorder SUD substance use disorder 4 Improve Care Quality If one goal of telehealth is to improve care quality, policymakers must take care not to transfer the current flaws of the health care system to a virtual environment. Telehealth should at least be equivalent to in-person care if not superior to it. No matter the modality of care delivery—inperson or telehealth—major struggles in maintaining high-quality health care include (1) ensuring that providers routinely follow agreed-on clinical guidelines and (2) reducing variation in quality across clinicians. One way to compare relative quality between modalities is to examine providers’ test ordering or prescribing patterns using the diagnosis code in a medical claim submitted to a payer to assess whether they reasonably follow guidelines. Various studies have looked specifically at rates of antibiotic prescribing for acute respiratory infections (ARIs), such as bronchitis or sinusitis. These rates are of high interest because inappropriate prescribing of antibiotics can lead to antibiotic resistance. Most of the studies thus far have compared in-person encounters with FIGURE 1 Varieties of Telehealth NOTE: Some people also consider e-consults or such programs as Project ECHO (Extension for Community Healthcare Outcomes) to be forms of telehealth, but these are typically provider-to-provider interactions that do not necessarily involve direct patient care. Varieties of Telehealth Telehealth (sometimes also called telemedicine) can refer to several different types of remote patient-to-provider interactions Remote patient monitoring • Technology-enabled monitoring of patients outside health care settings (e.g., in the home) Asynchronous “store-and-forward” consultation • Collection of patient data, which are stored and analyzed by a provider in a different location at a different time Synchronous interaction • Real-time visit between a patient and a provider • Can be with the patient’s usual provider or a new provider through direct-to-consumer telehealth • Can be done from a clinic (e.g., to provide a specialist consult from a primary care physician’s office) or from home 5 direct-to-consumer (DTC) telehealth services, during which care is provided by clinicians with whom patients do not have an established relationship. These studies have found somewhat mixed results: In some studies, telehealth-based rates of prescribing were found to be generally equivalent to in-person visits (Eze, Mateus, and Hashiguchi, 2020; Halpren-Ruder et al., 2019; Shigekawa et al., 2018). Other studies have found correlation between overprescribing and telehealth in general (Hoffman, 2020), or higher rates of inappropriate prescribing—particularly in children (Ray et al., 2019; Uscher-Pines, Mulcahy, et al., 2016). Still others have found that telehealth providers who treat ARIs deliver less appropriate testing and require more-frequent followup (Shi et al., 2018). It is unclear whether these findings represent potential challenges for telehealth in general or the challenges are more tightly linked to the DTC model for ARIs. At this time, we do not know whether the aforementioned findings would differ substantially for telehealth visits provided by a patient’s usual care team, a model that has become common since the beginning of the COVID-19 pandemic. Clinicians have voiced a variety of concerns about the effects of substituting in-person visits for telehealth visits. Some aspects of high-quality care depend on strong patient-provider rapport, the ability to do a physical exam and quickly order tests, and clinician familiarity with a patient’s history. Furthermore, personal topics—such as asking about drug use, delivering devastating news, or counseling patients about issues not directly related to the reason for their visit (e.g., need for a flu shot)—can be harder to address during a shorter and more-transactional telehealth visit. Patients might not feel as secure in their privacy over a video or telephone call, particularly if the provider is not the patient’s usual clinician. Despite these concerns, telehealth visits can support the delivery of high-quality, guideline-concordant care. Mental health services, for instance, have been shown to be of equivalent quality when delivered remotely (telemental health). An Agency for Healthcare Research and Quality–funded systematic review