Wednesday, November 26, 2014

Cost-effective preventive care: seeing the forest for the trees

At last month's Family Medicine Education Consortium Northeast Region Meeting, one of my residents presented some research that she had completed under my supervision. Since I left the staff of the U.S. Preventive Services Task Force four years ago, it has been my sense that the Task Force has substantially lowered its evidence "bar" for recommending a preventive service, an impression confirmed in private discussions with colleagues who closely follow the group's activities. In a JAMA editorial published last year, Drs. Steven Woolf and Doug Campos-Outcalt expressed concerns that the Affordable Care Act, by requiring insurers to fully pay for grade "A" and "B" recommended services, would lead to political pressure on the USPSTF to produce more of these favorable recommendations.

My resident and I hypothesized that if this concern turned out to be correct, we would find that a higher proportion of recommendation statements - both new and updated - published in 2011 or later would be grade "A" or "B" rather than "C," "D," or "I." After reviewing the Task Force's portfolio of active recommendations, she concluded that this is absolutely the case. Of course, not being able to attend the meetings or review their minutes (which are unavailable to the public), we could only demonstrate an association, not causation. Another plausible explanation is that research progress over the past several years has produced more evidence and effective interventions to support providing services which weren't recommended before (e.g., lung cancer screening with CT scans, screening for hepatitis C). That's unlikely to be the whole story, though, since the TF would have generated more new "D" (don't do it) recommendations too, which hasn't happened.

Politics aside, the other problem with linking USPSTF decisions to "free" preventive services is that a group that adamantly does not consider cost in assessing the value of a preventive service increases the cost of health care (and health insurance premiums) every time it makes a favorable decision. Dr. Woolf has argued that effective prevention doesn't have to be cost-saving, only cost-effective, and the vast majority of immunizations and recommended screenings meet this criterion. Even CT screening for lung cancer, according to a recent study, would cost $81,000 per quality-adjusted life year (QALY) gained, if appropriately implemented in a high-risk population similar to that in the National Lung Screening Trial.

But cost-effective services can still end up being terrifically expensive. If the estimated 9 million eligible Americans receive "free" annual low-dose CT scans recommended by the USPSTF at $300 per scan, that's $2.7 billion added to the national health care bill each year - and this doesn't count the costs of all of the follow-up CT scans for abnormalities, consultations, biopsies, and treatments that will ensue. If birth cohort screening finds 2 million previously unidentified adults with hepatitis C who subsequently take the new drug sofosbuvir (Sovaldi) at $84,000 per treatment course, it will cost $168 billion to pay for this drug alone, not counting other medications or costs of care.

Even if paying for these screening tests will ultimately provide health benefits to many (though I have qualms about the evidence for both), they will also make health insurance premiums a little less affordable, and it's nobody's job to decide if the benefits are worth the added costs to the population. In an article in Health Affairs, Dr. Mark Pauly and colleagues argue that complete pooling of risk is justifiable only for preventive services that are highly cost-effective. They propose continuing full coverage for the most cost-effective services, increasing patient cost-sharing for less cost-effective services, and discouraging coverage of services that are not cost-effective according to a societally-determined threshold (they suggest $400,000 per QALY).

Plenty of public health and health equity arguments could be made against this proposal, but what I like about it is that it sees the forest for the trees. Not matter how equally we distribute them (and the U.S. does a poorer job at this than most countries), health care resources are limited, and money spent on marginally effective services is money that's not being spent on countless other things that promote health and make life worth living. It's simply not enough to promote evidence-based preventive care by making all of it "free," regardless of the true costs.

Sunday, November 23, 2014

Shared decision-making for lung cancer screening: will it work?

Earlier this month, the Centers for Medicare & Medicaid Services (CMS) officially proposed coverage for annual low-dose computed tomography (LDCT) screening for lung cancer in current or former smokers age 55 to 74 years with at least a 30 pack-year history. In doing so, CMS followed the lead of the U.S. Preventive Services Task Force, which had previously given a "B" grade recommendation for screening in a similar population through age 80 years.

In the November 1st American Family Physician cover article, Dr. Thomas Gates reviewed concepts and controversies in cancer screening. Dr. Gates observed that in the 1960s and 1970s, physicians were misled by lead-time and length-time bias into believing that screening smokers for lung cancer with chest radiography saved lives, when in fact, it did not. He also noted that although LDCT screening has reduced lung cancer and all-cause mortality in a randomized controlled trial, adverse effects include a high false-positive rate, uncertain harms from radiation exposure, and overdiagnosis (leading to potentially unnecessary treatment). For these reasons, the American Academy of Family Physicians decided not to endorse the USPSTF recommendation. In an editorial published earlier this year, AFP Contributing Editor Dean Seehusen, MD, MPH elaborated on arguments against routine LDCT screening.

Notably, CMS has proposed to pay for not only the LDCT itself, but also for a "counseling and shared decision making visit" with a physician to review benefits and harms of lung cancer screening and emphasize smoking cessation (in current smokers) and continued smoking abstinence (in ex-smokers). This element is critical, as Dr. Gates observed in his article:

Perhaps the most important issue with low-dose CT screening is that it is a costly, high-tech response to what is essentially a behavioral and lifestyle problem. Smoking is responsible for 85% of lung cancers; convincing persons to quit smoking (or to not start) is far more effective in preventing lung cancer deaths than low-dose CT screening.

Shared decision-making is increasingly recommended by screening guidelines, but I worry that these difficult discussions may not actually take place, even if family physicians are paid to initiate them with patients eligible for LDCT screening. Will clinicians merely go through the motions and just order the test, as happened with prostate-specific antigen testing for prostate cancer and screening mammography for women in their 40s? What do you think?


This post first appeared on the AFP Community Blog.

Monday, November 17, 2014

Public Speaking Update

Since I began blogging at Common Sense Family Doctor in July 2009, its posts have been featured in widely read blogs such as, The Doctor Weighs InThe Health Care Blog, and Gary Schwitzer's HealthNewsReview, and the websites of major national health and news outlets such as Proto Magazine, the Costco Connection, the New York Times, the Wall Street Journal, USA Today, and the Boston Globe. I also wrote the consumer health blog Healthcare Headaches for U.S. News and World Report from August 2010 through September 2011.

Like the vast majority of physicians who blog, I write in my spare time. I have never accepted advertising or paid web links on Common Sense Family Doctor, and the choices of topics for posts are my own and not influenced by financial or other conflicts of interest. In order to support the time I devote to blogging, and to encourage high-quality medical writing and clinical practice, I give lectures and workshops to medical and non-medical audiences on a variety of topics. These include the uses of social media tools in medicine and education, developing and implementing medical guidelines, and the evidence supporting specific prevention recommendations. If you or your organization would like to invite me to speak, please e-mail me at or

Upcoming events:

Choosing Wisely: Pearls for Primary Care Physicians
- District of Columbia Academy of Family Physicians
- January 21, 2015

Medical Apps: Topic TBD
- International Consumer Electronics Show, Las Vegas, NV
- January 6, 2015

Cancer Screening: An Updated Primer for Journalists
- National Press Foundation's Cancer Issues 2014, Washington, DC
- December 8, 2014

Past events:


Lung and Bronchial Cancer
- American Academy of Family Physicians Assembly, Washington, DC

Policy and Funding for Preventive Care Programs
- Georgetown University Health Systems, Policy, and Public Health Elective


CT Screening for Lung Cancer: Evaluating the Evidence
- National Capital Consortium Family Medicine Residency, Fort Belvoir, VA

Thinking Like An Editor
- Society of Teachers of Family Medicine Annual Spring Conference, Baltimore, MD

Politics of HIV Testing
- Georgetown University School of Medicine

Burnout Prevention for Healthcare Professionals
- Teaching Prevention 2013, Washington, DC

Evidence-Based Literature Searching: A Primer
- National Capital Consortium Family Medicine Residency, Fort Belvoir, VA

Less is More: New Approaches to Cancer Screening in Primary Care
- Primary Care Coalition of Montgomery County, Maryland


Science and Public Policy in Conflict: PSA Screening
- Johns Hopkins University Bloomberg School of Public Health Fall Policy Seminar

Screening Mammography for Women in their 40s: Exploring the Controversy
- National Capital Area Regional Breast Healthcare Improvement Initiative

Why You Should Stop Screening Patients for Prostate Cancer
- Ephrata Community Hospital (PA)

Identifying and Using Good Practice Guidelines
- Temple University School of Medicine 2012 Family Practice Review Course


Cancer Screening: A Primer for Journalists
- National Press Foundation's Cancer Issues 2011

What to Do When Screening Guidelines Conflict: HIV and Mammography
- Grand Rounds, Georgetown University Department of Family Medicine

Overdiagnosed: Making People Sick in the Pursuit of Health
- William J. Bicknell Lecture (panelist)
- Boston University School of Public Health

For Geeks and Geezers: With Social Media Skills You Can Change the World
- Family Medicine Education Consortium Northeast Region Meeting

Screening for Diabetes: What Does the Evidence Say?
- Spanish Catholic Center of Catholic Charities of Washington, DC

Don't Do It! Preventive Health Services That Harm More Than They Help
- District of Columbia Academy of Family Physicians

Using the Medical Literature to Make Decisions About Preventive Health Services
- Medical Library Association Annual Meeting

2008 - 2010

Medical Blogging and Other Professional Uses of Social Media
- Grand Rounds, Virginia Commonwealth University Internal Medicine

Spilling Ink: An Expert's Guide to Getting Your Work Published
- Society of Teachers of Family Medicine Annual Meeting

COPD Update: A Prevention Perspective
- Maryland Academy of Family Physicians

Sunday, November 9, 2014

For homeless patients, housing is preventive health care

Every year, a medical school course that I teach invites two speakers to tell students their compelling stories about how being homeless negatively affected their health. Conversely, I care for patients whose declining health led to homelessness because they were unable to work and fell too far behind on mortgage or rent payments. The American Academy of Family Physicians and other professional societies, such as the American College of Obstetricians and Gynecologists, encourage their members to provide compassionate and unbiased care to homeless persons, and a recent article in American Family Physician reviewed strategies for managing clinical conditions that commonly occur in this population.

The standard approach to chronically homeless persons with mental illness and/or substance dependence has been to improve control of these underlying medical problems before placing them in permanent housing. The trouble is that not knowing where one will eat or sleep from day to day is about the worst possible environment to improve mental health or recover from addiction. Dr. Kelly Doran and colleagues reported in the New England Journal of Medicine on a pilot program that used New York State Medicaid funds to house high-risk homeless patients:

Placing people who are homeless in supportive housing — affordable housing paired with supportive services such as on-site case management and referrals to community-based services — can lead to improved health, reduced hospital use, and decreased health care costs, especially when frequent users of health services are targeted.

New York health officials hope that much of its investment will pay for itself by reducing acute and emergency care visits, but so far has been unable to convince the Centers for Medicare and Medicaid Services (which only pays for nursing homes through Medicaid) to make a similar investment. Despite a lack of federal support, this "Housing First" approach has been successful in other states too, notably Utah, as James Surowiecki recently described in The New Yorker. Like it because it's the decent thing to do, because it saves money, or both, Housing First has garnered support across the political spectrum.

Some may view advocating for Housing First policies to improve the health of homeless persons to be outside of the scope of family medicine, but I don't. I have come to realize that some of my patients will not be able to fully address their chronic health issues until they have roofs over their heads and the stability and security that comes with having a place to call home. As Surowiecki observed, this approach can be viewed as a cost-effective form of preventive health care:

Our system has a fundamental bias toward dealing with problems only after they happen, rather than spending up front to prevent their happening in the first place. We spend much more on disaster relief than on disaster preparedness. And we spend enormous sums on treating and curing disease and chronic illness, while underinvesting in primary care and prevention. This is obviously costly in human terms. But it’s expensive in dollar terms, too. The success of Housing First points to a new way of thinking about social programs: what looks like a giveaway may actually be a really wise investment.


This post first appeared on the AFP Community Blog.

Thursday, November 6, 2014

Birth control pills over-the-counter: debate evidence, not politics

I've waited to address this sensitive topic until after the midterm elections, when political slogans such as the phony "war on women" and trumped-up threats to religious liberty were discarded like so many campaign posters. It was curious to see the American College of Obstetricians & Gynecologists (ACOG) and Planned Parenthood attacking Republican Senate candidates for supporting over-the-counter birth control pills without a prescription - a position that, if the pills were free or the candidates were Democrats, they would probably have cheered. (When the American Academy of Family Physicians quietly supported oral contraceptives over the counter earlier this year, it was careful to specify that such purchases be covered by health insurance.)

As outlined in a 2012 ACOG opinion paper, the rationale for granting over-the-counter status to birth control pills goes something like this: unintended pregnancies are common; visiting a doctor for a prescription is inconvenient and unnecessary; oral contraceptives are safer than many medications already available without a prescription; women can screen themselves for contraindications; and women wouldn't stop seeing doctors for other preventive services. There are, however, very few studies that actually support these arguments; much of the literature simply surveys what other countries do regarding contraceptive access and assumes that outcomes are better (or at least not worse). And surprisingly, there's no research whatsoever that shows making oral contraceptives over-the-counter reduces unintended pregnancies.

This hypothesis would be relatively straightfoward to test in a randomized clinical trial. Enroll, say, five hundred non-pregnant, sexually active, pre-menopausal women without contraindications to oral contraceptives who don't want to become pregnant in the next 12 months. Randomly assign half of them to receive birth control pills without a prescription at a convenient pharmacy, and assign the other half to obtain contraceptives the usual way, by requesting a prescription from their family doctor or gynecologist. After a year, compare the numbers of unintended pregnancies and adverse events (deep venous thromboses, strokes, sexually transmitted infections) in each group. Other outcomes could include contraceptive adherence, appropriate use, and use of recommended preventive health care such as immunizations and screenings.

Why hasn't this study been performed already? Some physicians have told me that this question doesn't need to be studied because it's obvious that over-the-counter access to contraceptives would lead to fewer pregnancies. Others have insinuated that even asking the question is "anti-woman" and insensitive to the long history of gender bias in health and men using fertility to control and oppress women.

I say bull. This isn't only a political question, it's also a scientific one. Otherwise, why stop at putting oral contraceptives over the counter? Why not, for example, make it easier for millions of women and men with poorly controlled ("unintended") high blood pressure to treat themselves by making anti-hypertensive drugs over-the-counter? In fact, self-monitoring and self-titration of blood pressure medications is a strategy that is being seriously considered in high-risk populations. A recent randomized trial published in JAMA compared this strategy to usual care in five hundred primary care patients with hypertension and a history of stroke, coronary heart disease, diabetes, or chronic kidney disease. After 12 months, the mean blood systolic blood pressure of the intervention group was 9 points lower than that of the control group, with no difference in adverse events.

The outcome of the hypertension study wasn't obvious. It might easily have gone the other way. And for that reason, it was a question that deserved to be rigorously studied. Similarly, over-the-counter birth control need not be an evidence-free debate. Regardless of where you stand on this issue personally or politically, it's time to stop with the slogans and inform the discussion with science.

Sunday, November 2, 2014

The natural history of symptoms in primary care

Not long ago, I was sitting in my office catching up on some electronic charting when I began to feel chilly, achy, and weak. I went home, skipped dinner, and went straight to bed. Although I felt mostly better the next morning, my appetite didn't fully return until later in the day. My self-diagnosis was a probable viral infection. But the truth was that I had no idea if my symptoms were related to any kind of disease.

Medical education trains physicians to approach patients' symptoms foremost as manifestations of an underlying cause. Only "treating the symptoms," in contrast, can often feel like a sort of failure. But as Dr. Kurt Kroenke reported in a narrative review published in the Annals of Internal Medicine, at least one-third of common physical symptoms evaluated in primary care (including pain, fatigue, dizziness, sleep disturbances, and gastrointestinal symptoms) are "medically unexplained," meaning that they are never connected to a disease-based diagnosis after an appropriate history, physical examination, and testing.

Dr. Kroenke further asserted that viewing symptoms as purely disease-oriented influences the language physicians use to describe them:

The lack of a definitive explanation for many symptoms is further underscored by the use of adjectival modifiers indicating what a symptom is not ("noncardiac" chest pain or "nonulcer" dyspepsia) or implying causal explanations that are weakly defensible ("tension" headache, "mechanical" low back pain, or "irritable" bowel syndrome).

Not only do some symptoms have no obvious causes, but others have multiple possible causes which may be unproductive to approach separately. For example, why does a patient with congestive heart failure, anemia and depression feel tired all the time? Also, symptoms usually occur in a group, rather than in isolation; for example, a classic symptom cluster in cancer patients is SPADE (sleep  / pain / anxiety / depression / energy).

Studies show that about a quarter of symptoms that present to primary care eventually become chronic. Fortunately, very few of these patients harbor a serious missed diagnosis such as an occult infection or cancer. As family physicians know, even if we are uncertain about if or when a particular symptom might improve, communication still has great therapeutic value. "Is this normal, doctor?" is the question I hear most frequently from my patients who have persistent symptoms without diagnoses. I usually respond that there is a wide range of "normal," and what's more important to me is working with him or her to make these particular symptoms more manageable.


This post first appeared on the AFP Community Blog.