Oral cancer screening is ineffective; will the USPSTF weigh in?

The U.S. Preventive Services Task Force (USPSTF) normally meets in person three times per year: in March, July, and November. This year, they did meet in March, but their meetings have since been suspended as they await the federal government's reopening and Health and Human Services (HHS) Secretary RFK Jr.'s postponed decision on whether to fire them all. Even if they could meet and vote on new or updated recommendations, their support staff at the Agency for Healthcare Research and Quality (AHRQ) has been gutted, and it isn't clear if the Evidence-Based Practice Centers that AHRQ contracts with to produce systematic evidence reviews are still being funded. Is the USPSTF better off remaining intact but non-functional as opposed to the reconstituted Advisory Committee on Immunization Practices, which is technically operating but has been producing nonsense? I don't know. But while the Task Force's activities remain at a standstill, prevention science is moving on.

One of the last topics I worked on as an AHRQ medical officer was screening for oral cancer. The USPSTF's last recommendation statement, concluding that the evidence was insufficient to determine the effectiveness of an oral screening exam in primary care, is nearly 12 years old. At the time, the only direct evidence was a cluster randomized trial performed in an Indian state with an extremely high incidence of oral cancer due in part to widespread betel quid chewing, and the Task Force reasonably concluded that it wasn't possible to extrapolate the mortality benefit seen in that trial to an American primary care population. In 2023, the USPSTF did a quick search for new evidence and determined that there wasn't enough to warrant updating the previous recommendation. The National Cancer Institute's latest summary of the evidence on oral and nasopharyngeal cancer screening, updated in April 2025, reaches a similar conclusion.

Then, the results of a new study were published. It turns out that Taiwan, where betel quid chewing is also common, launched the world's first national oral cancer screening program in 1999. Screening exams were conducted every two years in all male persons aged 30 and older who chew betel quid and/or smoke, as well as in indigenous males starting at age 18. An observational study of the outcomes associated with this screening program found that despite a 2.4 fold increase in early-stage oral cancer diagnoses, there was little change in late-stage incidence and no reduction in oral cancer mortality from 2004 to 2022. The researchers concluded that oral cancer screening is ineffective.

Taiwan isn't the United States, and if members of the USPSTF have reviewed this study, they may have been tempted to conclude that its results aren't any more applicable to our population than the trial from India. My view is different. If oral cancer screening doesn't work in Taiwan, it isn't likely to work anywhere else, especially in a dysfunctional U.S. health system reliant on opportunistic primary care screening in the absence of a robust public health infrastructure. As cigarette smoking and betel nut use have plummeted in the past two decades, this is likely the best evidence that we will ever have on this topic.

This is a small sample of the studies that the USPSTF and its support team should be reviewing but are likely unable due to irresponsible and damaging decisions being made (or intentionally not made) by HHS department leadership. Here's hoping that the Task Force's limbo ends sometime soon and that they emerge with renewed support and funding - but I'm not holding my breath.

Blood pressure drug combinations, comparisons, and therapeutic intensity

Although blood pressure management is the bread and butter of family medicine, selecting antihypertensive medications is not always straightforward. A 2023 American Family Physician article on initial management of hypertension in adults recommended, “Initial antihypertensive treatment should include an angiotensin-converting enzyme (ACE) inhibitor, an angiotensin receptor blocker (ARB), a long-acting dihydropyridine calcium channel blocker, or a thiazide diuretic.” But how should one choose among drug classes and single vs combination therapy?

Compared with monotherapy, combination antihypertensive drugs have the advantage of reaching blood pressure goals quicker, with similar tolerability. Expert consensus suggests that initial combination therapy is preferred “in patients with systolic blood pressure higher than 160 mm Hg or greater than 20 mm Hg above goal, or with diastolic blood pressure higher than 100 mm Hg or greater than 10 mm Hg above goal.” A 2024 study of US adults taking two classes of antihypertensives found that patients on fixed-dose combinations were 1.78 times more likely to have controlled blood pressure than patients on two separate pills. In June 2025, the US Food and Drug Administration first approved a triple antihypertensive drug, a combination of telmisartan, amlodipine, and indapamide.

Specific antihypertensive classes are indicated for special populations (eg, patients with heart failure, chronic kidney disease, diabetes). A randomized trial of more than 11,000 patients with hypertension at high risk of cardiovascular events found that despite similar blood pressure control, benazepril plus amlodipine was superior to benazepril plus hydrochlorothiazide (number needed to treat [NNT] = 45 to prevent a composite cardiovascular end point over 36 months). In the general hypertensive population, a Cochrane review found that over 5 years, thiazide diuretics have small advantages over calcium channel blockers (NNT = 100 to prevent a cardiovascular event; NNT = 84 to prevent heart failure) and ACE inhibitors (NNT = 167 to prevent one stroke).

Although cholesterol-lowering drugs and doses are classified by intensity, until recently no similar schema was available for antihypertensive drugs. A systematic review and meta-analysis of 484 placebo-controlled trials classified the average systolic blood pressure-lowering effects of 57 monotherapies and 189 combinations as low (< 10 mm Hg), moderate (10-19.9 mm Hg), and high (> 20 mm Hg) intensity. Unsurprisingly, most monotherapies had low efficacy, whereas dual or triple therapies generally produced moderate or high effects. Clinicians can use an online calculator derived from the review to estimate the efficacy of any antihypertensive drug and dose combination depending on the patient’s baseline blood pressure.

**

This post first appeared on the AFP Community Blog.

Recent and upcoming presentations and meetings

In September, I traveled to Oxford University for the international Preventing Overdiagnosis conference, where I presented on behalf of a research team that spent several years working to estimate the annual harms of screening colonoscopy overuse in the United States. We concluded that 2.1 million to 3.2 million low-value (unnecessary) screening colonoscopies are performed every year, leading to 9 to 12 thousand preventable cases of severe bleeding or bowel perforation.

Examination Schools, Oxford University

On October 20, I will deliver the cancer screening update lecture at the Temple University Fall Family Medicine Review course. For many years, this live course was delivered in person at a resort conference center in Lancaster, but the virtual format instituted during the pandemic proved so popular that it has continued so that family physicians and other primary care clinicians can join from anywhere in the world. Later that same day, I will lead a study session that allows American Board of Family Medicine Diplomates to complete the Health Counseling and Preventive Care Knowledge Self-Assessment. For readers who are interested, there is still time to register at https://medicine.temple.edu/cme

I've taught in this course since 2012.

Looking ahead, the week before Thanksgiving I will travel to Atlanta for the North American Primary Care Research Group (NAPCRG)'s annual meeting. Although I have never consciously defined myself as a researcher, many of my publications (including the colonoscopy overuse harms study mentioned above) can be classified as such. My relatively new role as the faculty member in charge of resident scholarly activity at the LGH Family Medicine Residency Program has me eager to learn from fellow educators about how to motivate learners to move beyond point-of-care references and artificial intelligence tools for answering clinical questions to designing an original research study or performance improvement project.