Telehealth use in the United States expanded rapidly during the COVID-19 pandemic, supported by temporary federal and commercial coverage flexibilities that allowed virtual care to substitute for in-person visits across primary care and many specialties.1 Since then, telehealth use has declined across all specialties,2 with mental health care retaining the highest proportion of visits conducted by telehealth.3 Policy discussions now focus on which flexibilities should be made permanent and how telehealth affects access and equity, particularly across different populations.4 Understanding how telehealth has trended in the post-acute pandemic period and how patterns differ by geography, age, and preferred language can inform operational planning, interpreter services, and policy decisions.
We studied 411 million primary care visits between July 2022 and October 2025. For each encounter, we identified the patient’s rurality based on their most recent ZIP Code, their age at the time of the encounter, and their preferred language. We evaluated each subgroup for the proportion of primary care they had conducted by telehealth.
Across all primary care encounters, telehealth use declined from mid-2022 to mid-2023 and then stabilized at around 6% of visits. Telehealth accounted for just over 8% of encounters in July 2022 and just under 6% by October 2025, representing a roughly 30% reduction by the end of the study period. Since 2023, the overall telehealth share has remained around 6–7%, suggesting a new steady state in the balance between virtual and in-person care.
Telehealth remained more common for patients from metropolitan areas than for those from less urban settings throughout the study period, as seen in Figure 1. While the telehealth proportion fell in all rurality categories, the relative urban–rural gradient persisted: metropolitan patients consistently had about twice the telehealth rate of patients in rural or small-town areas, with micropolitan areas in between.
Age patterns were stable over time, with telehealth concentrated among working-age adults and less frequently used for very young children and the oldest adults. Children aged 0–2 years consistently had telehealth shares below 2%, while adults aged 25–39 years had rates above 10% across the study window.
Telehealth use was consistently higher among patients whose preferred language was not English, and this pattern persisted despite overall declines, as seen in Figure 3. Chinese-, Portuguese-, Russian-, Persian-, and Spanish-speaking patients started from substantially higher baseline telehealth rates, and those rates remained substantially higher than those of English speakers by the end of the study period.