Friday, June 24, 2016

Should divided guideline panels publish minority views?

When I give presentations on the guideline development process of the U.S. Preventive Services Task Force, a question I'm often asked is how many votes of the 16-member panel are needed to approve a recommendation statement. The answer is a two-thirds majority, with a minimum of 10 votes in favor in case of absences or conflict-of-interest recusals. In reality, though, during the four years I attended Task Force meetings, I can't recall a statement passing without overwhelming (15-1 or 14-2) or, more commonly, unanimous support. The feeling among members seemed to be that the lack of a strong consensus on a recommendation suggested that there was something missing about the way they were approaching the evidence.

In contrast, since the vacancy created by the death of associate justice Antonin Scalia, the Supreme Court of the United States has issued several evenly split decisions, with their most recent tie effectively affirming a lower court decision that blocked President Obama's 2014 executive order to shield some undocumented immigrants from deportation. In this case, the Court issued no opinion at all. However, when majorities have decided other cases, the Court provided justices in the minority the option of publishing dissenting opinions, whose legal reasoning sometimes informed future decisions.

Minority opinions rarely accompany medical guidelines. In an unusual case, after the JNC 8 committee published its guideline for management of high blood pressure in adults, five former panel members who disagreed with the guideline's target systolic blood pressure of 150 mm Hg in persons aged 60 years or older formally published their minority view. Even then, this dissenting report appeared some time later, in a different journal than the original guideline.

In a recent article in Mayo Clinic Proceedings, Dr. Daniel Musher, a professor of medicine and infectious diseases at Baylor College of Medicine, made the case for regularly publishing dissenting opinions in medical guidelines. He cited his experience as a member of the Advisory Committee on Immunization Practices (ACIP) working group that recommended the use of 13-valent pneumococcal conjugate vaccine (PCV13) in adults 65 years and older, despite his strong disagreement. As is standard process for the ACIP and most guideline panels, he did not have the opportunity to voice his dissenting opinion and rationale in the text of the guideline. His view did not see publication until more than 18 months later, after the new recommendation had been largely implemented into clinical practice. Dr. Musher wrote:

The perceived problems with publishing dissenting opinions are that this practice would (1) cause confusion within the medical community and (2) diminish the force of the recommendations. Regarding the former, the current situation, in which dissent is not included but in which subsequent articles dispute the formal recommendations or different professional societies publish divergent guidelines, is amply confusing—witness the differing guidelines for screening for breast, lung, or prostate cancer. It defies reason to believe that every member of the American Cancer Society's committee thought that breast cancer screening should begin at age 40 years, whereas all who participated in the US Preventive Services Task Force agreed that screening should wait until age 50 years. Inclusion of dissenting opinions in the final version of published guidelines may well have reduced polarization and confusion by bringing dissent into the recommendation process.


I can't testify to the presence or absence of internal discord on the panels that produced the 2015 ACS or 2016 USPSTF guidelines on breast cancer screening, but my best recollection of the July 2008 meeting where the USPSTF first voted to recommend routine mammography starting at age 50 is that, contrary to Dr. Musher's suggestion, there was no minority view. When there is one, I agree with him and the Slow Medicine bloggers that making a forum available to describe conflicts that occurred within the guideline narrative (including all of the various options that were considered and later discarded) could reduce the intensity of second-guessing and better inform clinicians about the guideline's nuances and potential limitations. Now that I have experience as a voting member of American Academy of Family Physicians guideline panels, though, I wonder if the explication of dissents belongs in the guideline itself, rather than as a separate stand-alone perspective.

Would the airing of minority views within medical guidelines provide useful perspectives for patients, clinicians, or policymakers, or are conflicting guidelines from different organizations already confusing enough as it is?

Tuesday, June 21, 2016

Steroids for severe community-acquired pneumonia: ready for prime time?

A generation ago, one of the major controversies in treatment of infectious diseases was whether or not to prescribe early adjunctive corticosteroids in addition to antibiotics for AIDS patients with presumed pneumocystis pneumonia. Advocates of steroids argued that they would improve outcomes by reducing the body's damaging inflammatory response, but opponents expressed concerns that further suppressing an already impaired immune system could increase the risk for other opportunistic infections. The advocates turned out to be right, as summarized in a 1990 National Institutes of Health consensus statement and this more recent FPIN Clinical Inquiry based on a Cochrane review of six randomized controlled trials that showed decreased mortality in patients receiving steroids.

The debate occurring today is whether steroids benefit patients with severe community-acquired pneumonia (CAP) from other causes. Commenting on a 2015 meta-analysis of 12 trials published in the Annals of Internal Medicine, Dr. Marcos Restrepo and colleagues asserted that it was "time to change clinical practice" and routinely use steroids in patients with severe CAP, with the major research question being how to identify these patients accurately and efficiently. On the other hand, the authors of the Medicine By The Numbers on this topic in the June 1st issue of American Family Physician felt that the supporting evidence was less definitive:

No large, multicenter, methodologically rigorous trials on this topic have been published, making results inconclusive. Small trials like the ones included [in the
Annals review] have significant potential to exaggerate effects, suggesting that large, well-designed trials have the potential to override the findings.

In exchange for 1 in 29 patients developing transient hyperglycemia due to steroids, 1 in 20 avoided mechanical ventilation, 1 in 16 avoided acute respiratory distress syndrome, and there was a nonsignficant trend toward mortality reduction. Drs. Jonathan Fu and Gary Green concluded that "improvements in two patient-oriented outcomes, and no major patient-oriented harms established thus far suggest it may be reasonable to use corticosteroids in patients with CAP while awaiting further data."

**

This post first appeared on the AFP Community Blog.

Monday, June 13, 2016

Tempering unrealistic expectations for prostate cancer treatment

For patients to make informed decisions about treatment for localized prostate cancer, they must receive accurate information about the effects of various therapies and guidance from clinicians who understand their preferences and values. Unfortunately, a study in the May/June issue of Annals of Family Medicine suggested that one or both of these is lacking for many men. Dr. Jinping Xu and colleagues surveyed a population-based sample of 260 men in the metropolitan Detroit area who were newly diagnosed with localized prostate cancer between 2009 and 2010. The survey asked them to estimate how many years they would live without any treatment and with their chosen treatment.

Men greatly underestimated their life expectancy without treatment and overestimated the gain in life expectancy with surgery or radiation. Although 98% of patients with localized prostate cancer who choose active surveillance will not have died from prostate cancer 10 years after diagnosis, only 25% of study participants thought that they would live at least 10 years without treatment. And although the only U.S. study to compare radical prostatectomy to watchful waiting for localized prostate cancer showed no overall or prostate cancer-specific mortality benefit, men in this study who underwent surgery expected to live an average of 12 years longer than if they did not choose surgery. The authors suggested that the typically short time frame for urologists or radiation oncologists to establish patient relationships between diagnosis and treatment discussions may contribute to misconceptions about treatment prognoses. In contrast, they argued that

Primary care physicians, who care for patients over long periods, have the advantage of intimate knowledge of their patients’ approach to clinical decision making and disease management in the course of their prior illnesses. If primary care physicians are included in the decision process following diagnosis, they could begin to focus on helping patients with localized prostate cancer develop realistic expectations and make choices that support their treatment goals.

Whether or not more of these "primary care consultations" have occurred in recent years is unclear, but a New York Times story highlighted the increasing number of men with low-risk prostate cancer who choose active surveillance rather than surgery or radiation, from 10-15 percent five years ago to nearly half of men today. This phenomenon suggests that physicians concerned about overtreatment have been appropriately tempering men's expectations. Delaying or avoiding the adverse effects of traditional therapies should reduce the burden of urinary incontinence and sexual dysfunction in men diagnosed with prostate cancer. American Cancer Society recommendations for the primary care of prostate cancer survivors, including managing common physical and psychosocial issues and encouraging healthy lifestyle choices, were reviewed in a recent article in American Family Physician.

**

This post first appeared on the AFP Community Blog.

Wednesday, June 8, 2016

Why do-it-yourself blood tests are a bad idea

It sounded too good to be true when I first heard about Theranos, a company that promised to revolutionize medical testing by making it possible to perform dozens of tests on a single drop of blood, rather than the several tubes that would typically be required. And that wasn't all. Theranos CEO Elizabeth Holmes, a Stanford dropout and media magnet whose wardrobe seems to consist solely of all-black outfits, promised to empower patients by giving them the ability to order their own tests, rather than needing to ask a doctor to do so. Now that the company is under criminal investigation and under pressure to prove that its technology works, FiveThirtyEight reporter Katherine Hobson pointed out that routine blood testing in healthy people has numerous downsides that Holmes never mentioned, including poor predictive value, false positives, overdiagnosis. Even if the test accurately diagnoses a risk factor such as high blood sugar levels, a United Kingdom study found that persons invited to diabetes screening were no more likely than controls to quit smoking, reduce alcohol consumption, or become more physically active.

Yet the fascination with do-it-yourself medical testing continues. A New York Times article published earlier this week led with the story of Kristi Wood, a 49 year-old woman who was experiencing fatigue and cognitive problems. Rather than seeing a doctor, she turned to a direct-to-consumer testing service which told her that her vitamin D levels were too high, apparently because she had been overdosing on vitamin D supplements. Once she reduced her supplement dose, "she almost immediately felt better." She credited the testing service for making this (obvious) diagnosis and now has a bunch of blood tests repeated every 4 months.

Although Ms. Wood would do well to read Ms. Hobson's FiveThirty Eight article and the U.S. Preventive Services Task Force statement that found insufficient evidence for adults younger than age 65 to take supplemental vitamin D for any reason, at least the results were clear-cut and actionable. That isn't true for most abnormal results, which require clinical context, careful interpretation, and sometimes additional testing, to distinguish a false from a true positive. A normal laboratory range means that the vast majority - but not all - of healthy people's results will be found between these values. Statistically, 1 in every 20 tests is likely to be abnormal simply by chance. Since blood tests are usually ordered in panels, I estimate that about half of my own patients' result reports have at least one item flagged, with nearly all of them being false positives. In the absence of an informed explanation and reassurance from a health professional who spent 4 years in medical school, 3 years in residency, and 12 years in practice, these results could be alarming and/or trigger unnecessary action like starting a potentially harmful testosterone supplement, as did another do-it-yourself testing patient in the New York Times article.

On a related note, a recent blog post by family physician Jennifer Middleton raised some good questions about do-it-yourself screening for ovarian cancer. For $295, a woman concerned about her ovarian cancer risk can request on a commercial website that an instructional kit be shipped directly to her home. According to the website, the test is intended to be "routine." But there's absolutely nothing routine about it. The ongoing randomized trial evaluating the test's effectiveness hasn't yet determined if it causes more good than harm. For women at low risk, the American College of Obstetricians and Gynecologists and the Society of Gynecologic Oncology recommend against being tested. The moral of this story: if you aren't feeling well or worry about getting sick in the future, don't seek out do-it-yourself testing. Don't be duped by companies such as Life Line Screening. Make an appointment to see a family doctor instead.

Sunday, June 5, 2016

The problem of pain

My patients lie to me every day. Some tell me that they have been taking their medications regularly when they haven't. Some say that they have been eating a healthy diet and exercising for at least 30 minutes every day and don't know where the extra pounds are coming from. Some lie that they are using condoms every time they have sex, that they have quit smoking, and if they drink alcohol at all, it's only a single glass of wine with dinner. They bend the truth for many reasons: because they want to please their doctor, because they don't like to admit lapses of willpower, or because they are embarrassed to tell me that they can't afford to pay for their medications. I forgive them; it's part of my job to understand that patients (and health professionals) are only human. The only lies that I find hard to forgive are the lies about pain.

Like many doctors, I have complicated feelings about prescribing for chronic pain. On one hand, I recognize that relieving headaches, backaches, arthritis and nerve pain has been a core responsibility of the medical profession for ages. On the other hand, deaths and emergency room visits from overdoses of prescription painkillers have skyrocketed over the past 25 years, and I have inherited many patients with narcotic addictions that resulted from a prior physician's well-intentioned generosity with his prescription pad.

Even worse, I've had patients I trusted turn out to be junkies in need of a fix. An earnest, well-dressed young man once came to my office complaining of a common chronic condition that, he said, had not been relieved by high doses of over-the-counter painkillers. He convinced me to to prescribe him narcotic pills, and didn't bat an eye when I asked him to sign a pain contract that required him to return every month for refills and only receive prescriptions from me in person. For the better part of a year, he never missed an appointment, and seemed genuinely receptive to unrelated preventive care that I recommended based on his age and risk factors. His deceit was exposed only after he stumbled, intoxicated, into an acute care facility staffed by a doctor who knew me and requested an early refill of a prescription for a different brand of painkillers prescribed by a third doctor for another imaginary condition. My colleague told him the gig was up, and I've never seen him again.

I believe that drug addiction is a disease. So why do I find this patient's lies (and those from others like him) so hard to forgive? Because they have consequences for people who are truly in pain. For patients' convenience, I transmit virtually all prescriptions electronically to the pharmacy, but I'm not allowed to do this with "controlled substances" such as painkillers. Wary of encouraging drug abuse, some insurers impose arbitrary limits on the number of pills a patient may be prescribed in one month, which I can only override by spending hours on the phone or not at all. One chain pharmacy recently started demanding signed copies of chart notes that included the pain-causing diagnosis before they would dispense painkillers (a practice that I believe to be an illegal invasion of privacy, but they didn't budge an inch when I told them so). And worst of all, doctors like me who have been burned before are that much more likely to view our patients with suspicion.

In a 2014 article in Health Affairs, Janice Schuster described a health odyssey that began with a seemingly minor surgical procedure and ended with her becoming "one of the estimated 100 million American adults who live with chronic pain" - in this case neuropathic pain, or pain from nerve damage that in my experience can be the most difficult type to treat. She wrote about how health system restrictions designed to discourage abuse created obstacles to her obtaining adequate pain relief, and about a lack of compassion from her primary care physician (who "dismissed my symptoms") and her surgeon (who "said again and again that he had not heard of a patient experiencing such pain"). As the author of a popular self-help book for persons facing serious illness, Schuster understood better than most the public health crisis posed by prescription painkillers, but that understanding offered little consolation as she navigated "the maze of pain management" that has evolved to deal with it:

Pain patients like me often feel trapped between the clinical need to treat and manage pain and the social imperative to restrict access to such drugs and promote public safety. ... When I am not overwhelmed by pain, or depressed by it, I am furious at the attitudes I encounter, especially among physicians and pharmacists. It has been stigmatizing and humiliating. ... Surely, we can find better ways to ease the suffering and devise treatments and strategies that do more good than harm and that do not shame and stigmatize those who suffer.


A few of my colleagues have become so disillusioned with the dilemmas of pain management that they have sworn off prescribing narcotic painkillers entirely. As often as I've been tempted to take that path, I won't abandon patients in pain, for whom the services of caring and competent family physicians are needed now more than ever.

**

This post first appeared on Common Sense Family Doctor on August 13, 2014.

Wednesday, June 1, 2016

From precision medicine to community vital signs

Last fall, I expressed concerns that the federal government's push to accelerate developments in "precision medicine" will draw attention and funding away from population health initiatives that could make a positive difference here and now. I still feel that way, particularly in light of preliminary CDC findings suggesting that the age-adjusted U.S. mortality rate rose last year for the first time since 2005. The bad news affects different groups of Americans in different ways. A previous analysis led to news stories asking, what's going on with middle-aged white Americans in rural areas? The next question is, what can health care or social services do to help them as well as traditionally disadvantaged populations?

Family doctors like me are generally comfortable with screening patients for physical conditions such as high blood pressure and cholesterol, infectious diseases, and cancer. Unless patients volunteer this information, I am far less adept at asking about "social" conditions such as alcohol misuse and intimate partner violence, even though the U.S. Preventive Services Task Force has concluded that both of these screenings can improve health. The prospect of asking my patients if they experience other negative social determinants, such as financial insecurity, housing insecurity, joblessness, and childcare challenges is even more intimidating.

But that's exactly what three family medicine clinics in Albuquerque, New Mexico did with an 11-question instrument called WellRx. As researchers reported in the May/June issue of the Journal of the American Board of Family Medicine, these clinics administered WellRx to more than three thousand patients over a 90-day pilot period and found that 46% screened positive for at least one area of social need, with most having multiple needs. Critically, every patient with a positive WellRx screen was referred to a community health worker (CHW) trained to address identified needs:

Services offered by CHWs included helping patients to access resources like food banks and to fill out job applications, accompanying patients to apply for food stamps, conducting home visits, or arranging family meetings with the health care team. A patient's level of need dictated the intensity and duration of the CHW's interaction with that patient—usually 1 contact a month and usually for less than 3 months.


Although the researchers reported that their university hospital was convinced enough by this pilot study to implement screening for social needs in all 7 of its primary care clinics, it remains to be seen if these screenings and CHW referrals will lead to health benefits. Even though 11 questions doesn't seem like a lot, given how many (less useful) administrative tasks are already required of primary care, many practices may be unable to routinely collect this information. One possible alternative is leveraging big data, argued Dr. Lauren Hughes and colleagues in the same issue of JABFM. Rather than collecting information about social determinants from patients one by one, they advocated "linking aggregated population health data" with patients' home addresses in electronic health records to provide physicians with an easily accessible set of community vital signs.

Either approach to data collection about social needs faces challenges and skepticism from physicians who aren't yet convinced of its utility. In response to an AAFP News article about community vital signs, one physician commented, "Fantastic! This is just what underworked and overpaid family physicians need- a massive data dump into their EMRs of public health info that they are expected, coerced, and eventually forced to analyze and act upon." Another wrote, "I can imagine it might be interesting to have some of this data occasionally but practically it would have essentially zero impact on a physician interaction with a patient." Zero impact? I have to disagree. Although the jury is out on the potential benefits of clinically-enabled community health interventions, my feeling is that they will be a lot more impressive than those of precision medicine.