Saving AHRQ and the USPSTF

AcademyHealth CEO Aaron Carroll, MD recently submitted testimony to the House Appropriations Subcommittee on Labor, Health and Human Services, Education, and Related Agencies about the dire condition of the Agency for Healthcare Research and Quality (AHRQ), where I spent 4 years as a medical officer early in my career. Dr. Carroll points out the immense return on investment that AHRQ has provided over the years - for example, saving $7.7 billion in U.S. health care costs by reducing hospital-acquired infections from 2014 to 2017 on a budget of around $300 million per year - and its unique, irreplaceable function among federal health agencies:

NIH [National Institutes of Health] studies diseases. AHRQ studies how health care is delivered. These are different missions. NIH can tell us that a treatment works in a clinical trial. AHRQ tells us whether that treatment reaches patients in a rural hospital, whether it is implemented safely, what it costs, and whether a critical access hospital in a rural county can actually use it. No other federal agency performs this function. Eliminating AHRQ does not transfer these capabilities elsewhere. It simply ends them. 

Notably, Congress rejected HHS Secretary Robert F. Kennedy Jr.'s 2025 proposal to eliminate AHRQ. But Dr. Carroll observes that the Trump administration has effectively carried out this plan anyway, by laying off most of the agency's staff and the entire grants management division, crippling its ability to function as a funder of health services research:

AHRQ has not awarded a single new grant since April 2025. An estimated $80 million in FY25 appropriated research funding was allowed to expire unused—a pattern consistent with the Government Accountability Office’s ongoing impoundment investigation. In FY26, the agency has not funded any of the noncompetitive continuing grants it is statutorily obligated to pay. The FY27 congressional justification now explicitly states a policy of “no new grants,” ending AHRQ’s four-decade role as the nation’s primary funder of health services research—a decision Congress never authorized.  

Similarly, former New York City and Philadelphia Health Commissioner Thomas Farley, MD wrote today on his Substack that the U.S. Preventive Services Task Force is being "quietly strangl[ed]" by being deprived of AHRQ support staff, not being convened since March 2025, and not appointing replacements for 5 members whose terms expired on December 31. He cites the recent ACC/AHA dyslipidemia guidelines as an example of what fills the preventive care vacuum when the USPSTF (which wrote its own cholesterol guideline in 2022) is effectively silenced:

Are cholesterol tests for kids and coronary artery scans for adults now scientifically justified? Here’s the problem: I do not know. It takes more expertise and time than I have to sift through all the many complicated studies to figure that out. ... But I do know that (by my count) 12 of the 33 members of the writing committee and 17 of the 29 members of the review committee for the ACC/AHA guidelines have financial ties to biotech companies that are likely to make money from this testing and treatment. (None of the USPSTF members have these conflicts.) And I know this rule: if you’re wondering whether you need a new pair of shoes, don’t ask a shoe salesman.

The muddle about cholesterol testing, statin treatment and coronary artery scans is just one example of what we are losing from the USPSTF’s paralysis. ... Thanks to Kennedy, dozens of other important questions on the USPTF consideration list are also languishing. Each month that the Task Force is in deep freeze our ignorance accumulates. ... Surely we can afford to have a group of experts who are not motivated by profit guiding us on which medical services actually keep us healthy. With the USPSTF dead in the water, the war on science begins to feel like a war on us.

Nearly a year ago, I wrote a Medscape commentary that appealed to readers to "Save the USPSTF." The USPSTF still needs saving. So does AHRQ. So does the entire taxpayer-funded scientific apparatus at HHS devoted to keeping people healthy that RFK Jr. has wrecked.

AI health tools for the general public fall short

A 2025 American Family Physician editorial by Dr. Joel Selanikio discussed how artificial intelligence (AI) tools had accelerated an existing trend of “patients bypassing physicians to diagnose and treat themselves,” which began with over-the-counter drugs and online search engines. This direct-to-consumer health care approach received a boost in January with OpenAI’s launch of ChatGPT Health, which invites users to upload their medical records and health data from apps for personalized recommendations.

AI chatbots can provide helpful responses to health questions in several low-stakes contexts, as outlined in this handout from Dewey Labs: translating medical jargon, brainstorming possible causes of symptoms, summarizing research or test results, and preparing questions for an upcoming doctor’s visit. However, a recent study in Nature Medicine highlighted ChatGPT Health’s significant limitations in triaging patients with acute problems to appropriate levels of care.

Dr. Ashwin Ramaswamy and colleagues compared the chatbot’s responses to “60 clinician-authored vignettes across 21 clinical domains under 16 factorial conditions (960 total responses)” to triage levels assigned independently by three physicians: non-urgent, semi-urgent, urgent, and emergency. ChatGPT Health performed well in triaging semi-urgent and urgent clinical situations, but it over-triaged 65% of non-urgent situations and under-triaged 52% of true emergencies. For example, it recommended evaluation in 24 to 48 hours for patients with diabetic ketoacidosis and impending respiratory failure rather than sending them directly to the emergency department. Just as concerning, patients with suicidal ideation were less likely to receive crisis interventions when they had identified a method of self-harm than when they had no identified method:

The crisis guardrail finding may be the most consequential failure mode exhibited in the entire study. … A guardrail that fires for ‘haven’t thought through how I would do it’ but not for ‘thought about taking a lot of pills’ is not calibrated to clinical risk and users have no basis to anticipate when it will or will not fire. The capability to recognize mental health crises and connect users with crisis resources is a basic prerequisite for any consumer health platform. Our data show this prerequisite has not been reliably met.

In another study, three AI chatbots were provided with 10 detailed medical scenarios and tested on their ability to diagnose the condition and recommend appropriate management. In the United Kingdom, 1,298 adults were provided the scenarios and randomized to use one of the chatbots or a usual source of their choice (typically an online search engine). When researchers input the full scenarios, the chatbots diagnosed 95% of the conditions and correctly managed them 56% of the time. However, when intervention participants shared elements of the scenarios in live conversations, the chatbots performed much worse, correctly diagnosing 34% of the time and recommending appropriate management in 44%; this result is no better than control participants using a search engine. Researchers observed that participants often failed to provide enough information to make the diagnosis, and slight changes in symptom emphasis or wording of questions frequently led to dramatic differences in advice.

Bottom line: For patient-facing chatbots such as ChatGPT Health to diagnose and triage problems appropriately and safely, it isn’t enough to passively process the incomplete clinical data they are provided. They will need to get much better at asking the right questions to elicit information that patients may not be aware is relevant.

**

This post first appeared on the AFP Community Blog.

Universal resource menu more effective than social needs screening

In January, I wrote a Medscape commentary about the mixed evidence that routinely screening patients for social needs improves health outcomes. Although many patients could benefit from assistance with necessities like food, housing, or transportation, screening for social needs only works if patients are prepared to accept help and realistic options exist to provide it:

There are also downsides to screening, which takes time and can distract from the purpose of the visit. ... Many of my patients decline social work referrals, leading me to wonder why we screened them in the first place if they knew that they would not want assistance. For others, referrals may be wasting their time and giving them false hope. Social workers can’t magically produce affordable housing in the midst of a national housing shortage or provide regular access to healthy meals — especially after Supplemental Nutrition Assistance Program benefits were cut by hundreds of millions of dollars.

In his 2020 Presidential campaign, entrepreneur Andrew Yang famously endorsed a different strategy. Rather than spending millions on inefficient bureaucracies whose sole purpose is to ensure that only the neediest qualify for government assistance, Yang advocated providing universal basic income, a "Freedom Dividend," to every citizen to spend or save as they saw fit. This strategy would eliminate the stigma of being identified as having social needs.

Could an analogous approach to social risk work in health care - offering every patient a menu of resources to select from rather than screening first and only assessing the needs of persons who screen positive? Dr. Danielle Cullen and colleagues at the Children's Hospital of Philadelphia studied this question in the Socially Equitable Care by Understanding Resource Engagement (SECURE) randomized trial. In this study, 3949 caregivers of children and young adults in the emergency department were randomized to one of 3 groups: 1) completing a social risk screener; 2) receiving a resource menu to indicate desired assistance; and 3) no social assessment. All caregivers were provided with an electronic "resource map" after the initial intervention. The primary outcome was reported desire for resources in any of 5 domains (housing, transportation, childcare, food security, household heat and electricity). Caregivers assigned to the resource menu group were significantly more likely to meet the primary outcome than the screening group (38.4% vs 29%).

Although this study's findings need to be replicated in primary care and patients of all ages, it contributes an important piece to the puzzle of how health care institutions can best elicit and assist with health-related social needs.

Primary care supply and access challenges around the world

Although as an American family physician, I tend to focus on the primary care crisis in the United States, a recent New York Times story highlighted the familiar challenges of recruiting family doctors to rural communities in Canada. In a "Hunger Games"-style competition, cash-strapped communities with doctor shortages outbid each other with signing bonuses of up to $51,000 ($70,000 Canadian) each to foreign-trained primary care physicians who agree to move to town and start new practices. Stettler, an Alberta town that couldn't keep its emergency room open 24/7 due to a dearth of doctors, shelled out $300,000 to lure seven young physicians from Nigeria and Ghana. A Milbank Quarterly perspective by my family physician colleague Bob Phillips (the namesake of Georgetown's Robert L. Phillips, Jr. Health Policy Fellowship) and primary care leaders in the United Kingdom, Australia, Canada, the Netherlands, and New Zealand outlined how policies that deprive primary care of funding and other necessary resources have compromised care and access around the world, even in otherwise high-performing health systems.

Producing enough primary care clinicians to meet population demand is a problem worldwide, but North Americans are at a disadvantage compared to peer countries, according to the results of a Commonwealth Fund survey in this month's issue of Health Affairs. In 2022, 44 and 47 percent of U.S. and Canadian primary care physicians reported burnout, respectively, trailing only New Zealand. U.S. physicians also reported high rates of dissatisfaction and stress from 2012-2022, with a higher burden in females than males. These factors cause physicians to cut back on patient care hours or retire early, further depleting the primary care pool.

How many more primary care clinicians do we need? Traditional calculations which divide the eligible patient population by a standard panel size of 2,000 for a full-time physician fail to account for the aging population with multiple chronic conditions and the increasing complexity of medicine in general. Two recent studies used novel methods to approach this question at a national and state level. In JAMA Internal Medicine, Dr. Katherine Morgan and colleagues examined trends in the numbers of primary care clinicians accepting new patients from 2013 to 2021, with availability defined as billing >11 new visit evaluation & management codes for fee-for-service Medicare beneficiaries annually. They found a steady decline in primary care physician availability that was incompletely offset by a rise in advanced practice clinician availability. It's not only harder to find a new doctor than it was a decade ago, it's harder to find a new primary care clinician of any kind.

Dr. Jacqueline Britz and colleagues measured primary care workforce capacity in Virginia from 2016 to 2021, and their data told a similar story. Of the 4,508 primary care physicians (PCPs) practicing in the state, more than one-third are older than age 60, while only 24 percent are younger than age 50. The median number of unique patients each PCP saw in 2021 was 1,290, with each patient being seen twice per year. In order for every Virginia resident to have a PCP, researchers calculated that 1,305 additional PCPs would be needed, a 29 percent shortfall compared to an 18 percent shortfall in 2016.

One policy solution for Virginia and other states facing widening gaps in primary care supply is to open more training programs. After all, the majority of family physicians still choose to practice in the state where they were trained. (I rejoined that group when I moved back to Pennsylvania 4 years ago.)  But Dr. Morgan Stickler and colleagues pointed out in a commentary that the family medicine residency Match isn't exactly a "Field of Dreams": if you build it, they don't come. Since 2012, the number of family medicine positions offered in the Match has almost doubled to 5,357, but the number of unfilled slots has quintupled from 142 to 805, involving 288 unfilled programs. The absolute number of students attempting to match in Family Medicine decreased by 16.6 percent from 2019 to 2024.

As Dr. Arch Mainous reminded us in a 2025 Family Medicine editorial, the increasing production of nonphysician clinicians and capabilities of artificial intelligence means that the family physician-dominated model of primary care could rapidly become as archaic as the "Great Horse Manure Crisis of 1894" (which ended with the onset of cars and other motorized vehicles). That doesn't mean that doctors will all be replaced by chatbots or that we don't need to find ways to remove ever-multiplying tasks from overwhelmed physicians' plates (the "everything bagel" problem). It does mean that we will need to question old assumptions about primary care scarcity and come up with creative solutions to meet the primary care needs of the future, without resorting to zero-sum competitions.

Smartwatch screening for hypertension

An analysis of 2021-2023 National Health and Nutrition Examination Survey (NHANES) data found that among nonpregnant adults with hypertension (defined as systolic blood pressure 130 mm Hg or higher, diastolic blood pressure 80 mm Hg or higher, or use of antihypertensive medication), nearly 80% had a mean of three blood pressure (BP) measurements higher than 130/80 mm Hg. Of this group with uncontrolled BP, based on some practice guidelines, fewer than 40% were taking antihypertensive medications. Although undertreatment of high BP may result from therapeutic inertia, many adults do not have a regular source of care and simply do not know that they have hypertension.

Identifying undiagnosed hypertension is the goal of the Apple Watch’s hypertension notification feature, which was cleared in September 2025 by the US Food and Drug Administration for use in nonpregnant people 22 years or older. The watch is not a substitute for a home BP monitor, and it does not provide individual BP measurements. Instead, it uses an optimal sensor that collects photoplethysmography (PPG) data during waking hours over 30-day intervals, analyzes that data with a machine-learning algorithm, and alerts the wearer if hypertension is suspected. The alert encourages patients to monitor their BP at home for 7 days and consult their doctor about next steps.

A clinical validation study involving more than 2,000 participants found that the Apple Watch correctly identified 214 of 585 people with BP of 130/80 mm Hg or higher (sensitivity = 41.2%) with a specificity of 92.3%. The alert’s positive and negative likelihood ratios were 5.35 and 0.64, respectively. In a research letter in JAMA, Dr. Jordana Cohen and colleagues applied the Apple Watch’s screening performance characteristics to the nationally representative NHANES population. The researchers calculated the posttest probabilities of hypertension after a positive or negative alert based on the prevalence of hypertension in different subgroups:

Among eligible individuals aged younger than 30 years, an alert increases the probability of hypertension from 0.14 (95% CI 0.11-0.17) to 0.47 (95% CI 0.39-0.56), wherein the absence of an alert lowers it to 0.10 (95% CI 0.07-0.12). Among individuals aged 60 years or older, an alert increases the probability of hypertension from 0.45 (95% CI 0.39-0.50) to 0.81 (95% CI 0.76-0.86) and the absence of an alert lowers it to 0.34 (95% CI 0.29-0.39).

Similar to the heart-rhythm monitoring and obstructive sleep apnea detection features of various wearable consumer devices, the Apple Watch’s hypertension screening feature may nudge some patients to seek appropriate care for an undiagnosed chronic condition. For others, it may cause unnecessary anxiety or provide false reassurance. Given the popularity of these devices, though, smartwatch screening for hypertension will become increasingly common.

**

This post first appeared on the AFP Community Blog.

Continuity of care: health benefits and compensation

Continuity of care—a trusting relationship with a personal physician based on recurring interactions over time—is an essential ingredient of the secret sauce of family medicine. Numerous studies have demonstrated that higher care continuity is associated with more appropriate care and lower costs. Data also support positive effects on health outcomes. A retrospective analysis from Alberta, Canada, reported that physician and clinic continuity (seeing a clinical partner when one’s primary physician was unavailable) correlated with fewer emergency department visits across all levels of patient complexity and fewer hospitalizations for highly complex patients. A similar US cohort study of Medicare fee-for-service claims from a nationally representative sample of 4,940 medical practices found that patients with higher primary care physician continuity were 5.5% to 6.8% less likely to be hospitalized and 4.9% to 6.3% less likely to visit the emergency department.

In 2023, the Accreditation Council for Graduate Medical Education changed the emphasis of family medicine resident training requirements from volume of encounters to continuity of care. In a special article published in Family Medicine, Dr. Gregory Garrison and colleagues described five domains of care continuity. In addition to traditional clinician-patient or “relational” continuity, they discussed benefits of informational (medical record), managerial (interdisciplinary care), family (caring for multiple family members), and geographical (care in multiple locations or environments, such as home, office, and hospital) continuity.

Unfortunately, the percentage of the US population who experience continuity of care has been decreasing since the turn of the century. A 2015 Graham Center Policy One-Pager found that fewer adults reported a personal physician (as opposed to a practice or facility) as their usual source of health care. From 2000 to 2019, the percentage of US patients who reported having a usual source of care declined from 84% to 74%.

To support continuity and compensate primary care clinicians for the additional complexity and cognitive load involved in these longitudinal relationships, the Centers for Medicare and Medicaid Services (CMS) introduced the add-on G2211 code in January 2024 and refined its use in 2025. Some private health insurers also recognize and pay for this code. A 2024 Graham Center analysis projected that applying the G2211 code to all eligible Medicare evaluation and management (E/M) visits would increase annual reimbursement by an average of $2,667 per physician. Attaching the G2211 code to E/M visits with 25 modifiers (approved by CMS in 2025) resulted in another $990 per physician. This year, G2211 codes also can be applied to home-based primary care visits. Relative to the established benefits of care continuity, the added compensation seems modest at best. In addition, a research letter published this week in JAMA found that in 2024, the largest share of G2211 codes (43%) was billed by specialist physicians, not primary care physicians (40%), many for diagnoses "that may not qualify as serious or complex" as CMS originally intended.

**

This post first appeared on the AFP Community Blog.