Since 1994, a group of primary care clinicians with expertise in evidence-based practice has performed monthly surveillance of more than 110 English-language research journals.¹ Of approximately 20,000 research studies published during 2015 in these journals, 251 met prespecified criteria for validity, relevance, and practice change. A study was considered valid if it was well designed and avoided important biases such as failure to conceal allocation or failure to mask outcome assessment. Relevance depended on the research question (i.e., is it relevant to a primary care clinician?) and on the outcomes reported; only studies that reported patient-oriented outcomes, such as morbidity, mortality, or quality of life, were considered relevant. Finally, studies with the potential to change practice for a substantial number of physicians were prioritized over those that merely confirmed existing practice (i.e., they matter). Studies that meet all of these criteria are called POEMs, for patient-oriented evidence that matters.² Each POEM is summarized in a structured critical appraisal written by one of the six expert reviewers and peer reviewed by faculty and fellows of the University of Missouri Department of Family Medicine. Writing and disseminating the POEMs is supported by subscriptions, without industry support.

Since 2005, the Canadian Medical Association has sponsored a subscription to POEMs for its members. Each member has the option to receive the daily POEM by e-mail, and the option to rate it using a brief survey through the Information Assessment Method. The questionnaire for the Information Assessment Method is a validated tool that addresses relevance to clinicians, cognitive impact, use for practice, and expected health benefits if the results of the POEM would be applied.³ For this article, we identified the 20 POEMs of 2015 that were rated highest for clinical relevance by Canadian Medical Association members. Each POEM was rated by at least 500 physicians, and at least 52% of respondents rated each of these top 20 POEMs as “totally relevant for at least one of your patients,” whereas less than 18% rated each as “not relevant for at least one of your patients.”

In the fifth installment of this annual series,^4–7 we summarize the clinical question and bottom-line answer for each of the 20 POEMs critically appraising an original research study, organized by topic and followed by a brief discussion. The bottom-line answers have been rewritten slightly from the original to stand alone without the complete synopsis and appraisal. We also briefly discuss POEMs summarizing practice guidelines that were judged to be highly relevant.

Respiratory Tract Infection

Four studies addressed the management of acute respiratory tract infection (Table 1).^8–12 Study 1 was a systematic review of the accuracy of rapid strep tests for streptococcal pharyngitis, and concluded that their accuracy is improving, with a sensitivity of 86%.⁸ Because these tests miss only about one in six cases of streptococcal pharyngitis, the yield of backup cultures is small, particularly in adults, in whom strep is less common. Study 2 is from a British research group that has conducted a series of large trials of antibiotics for acute lower respiratory tract infection when pneumonia is not suspected.¹⁰ The authors randomized 2,061 adults to receive amoxicillin, 1 g three times daily for seven days, or placebo (susceptibility of common respiratory pathogens to amoxicillin remains good in England). The overall findings were a number needed to treat (NNT) of 30 to prevent new or worsening symptoms, and a similar number needed to harm for adverse events caused by the antibiotic. There was a somewhat greater reduction in the duration of symptoms for patients with green sputum, and a greater reduction in symptom severity between days 2 and 4 for those with significant cardiopulmonary morbidities.

Although most systematic reviews examine the benefits of treatment, Study 3 examined the harms of amoxicillin and amoxicillin/clavulanate (Augmentin), which are commonly used to treat respiratory infection.¹¹ The authors identified 25 studies that reported harms in a placebo-controlled trial. They found (not surprisingly) that diarrhea was more common in patients receiving amoxicillin/clavulanate, but not in those receiving amoxicillin alone. On the other hand, one additional case of candidiasis occurred for every 23 patients treated with amoxicillin, with or without clavulanate.

The final study in this group, Study 4, provides insight into the causes of community-acquired pneumonia (CAP).¹² Researchers from the Centers for Disease Control and Prevention identified 2,259 inpatients with CAP at five hospitals. Each had an extensive evaluation for viral and bacterial cases of CAP using culture and polymerase chain reaction testing. At least one viral pathogen was detected in 23% of patients, at least one bacterial pathogen in 11% of patients, one or more of each in 3% of patients, and fungi or mycobacteria in 1% of patients. Despite a careful search, 62% of patients had no pathogen detected.

Back Pain

Back pain was a popular topic, with five studies addressing its diagnosis and treatment (Table 2).^13–17 Study 5 included 5,239 adults 65 years and older presenting to a primary care physician with a new episode of acute low back pain, of whom about one-fourth received imaging within six weeks of presentation (1,174 had plain film radiography and 349 had computed tomography or magnetic resonance imaging).¹³ Propensity score matching, which matches patients who did and did not undergo imaging but were otherwise similar, was used to compare outcomes between patients. Although imaging increased costs, no clinical benefit was observed.

Two studies looked at common medical treatments for back pain. Study 6 was a systematic review that identified 13 good-quality randomized controlled trials (RCTs) comparing adequate dosages of acetaminophen with placebo, and found no clinically important short- or long-term benefits for patients with back pain, hip pain, or osteoarthritis.¹⁴ This is disappointing, given the safety of acetaminophen compared with other drugs. Study 7 was an RCT of 323 adults with acute low back pain randomized to naproxen plus either placebo, cyclobenzaprine (Flexeril; 5 mg), or oxycodone/acetaminophen (5 mg/325 mg), all taken as one or two tablets every eight hours.¹⁵ Adding cyclobenzaprine or oxycodone/acetaminophen did not improve any outcomes, but led to more adverse effects (number needed to harm of 8 for the former and 5 for the latter).

What about nondrug therapies? In Study 8, which was an RCT of 220 adults with acute low back pain presenting to their primary care physician, one-half of patients were randomized to receive a referral within 72 hours for physical therapy (PT), whereas the remainder received usual care.¹⁶ The primary outcome was the Oswestry Disability Index, a 100-point measure of disability from back pain. Those receiving PT had a slightly greater improvement in the score (three points), but such a small improvement is unlikely to be noticeable by patients. There were some minor improvements on secondary outcomes, but the overall results do not support routine referral of patients with acute low back pain for PT. Finally, back pain often becomes chronic or recurrent. Study 9 was an Irish study that randomized 246 patients referred for PT to standard PT, a weekly exercise class tailored to patients with back pain, or a program of gradually increased walking.¹⁷ Patients in the walking group were given a pedometer and told to walk at least 10 minutes per day, four days per week, increasing gradually to a target of 30 minutes per day for five days per week. The walkers achieved the best results of the three groups, with the lowest cost and greatest likelihood of sustained adherence to the program. Now, we just need to advocate for our communities to offer safe places to walk.

Screening and Prevention

Several trials evaluated the benefits (and harms) of clinical preventive services (Table 3).^18–20 Study 10 asked how long we should continue nicotine replacement therapy.¹⁸ The researchers randomized 525 smokers to the nicotine patch for eight, 24, or 52 weeks; all received 12 counseling sessions (the first in person and the remainder by telephone). At one year, there was no significant difference between groups in abstinence rates (more than 20% for all three groups).

Falls are an important health problem, especially in older women. Study 11, a Finnish study, included 409 community-dwelling women 70 to 80 years of age who had fallen at least once in the previous year and were not taking vitamin D supplements.¹⁹ Participants were randomized to receive 800 IU of vitamin D per day, twice-weekly exercise, both interventions, or neither. Although the total number of falls did not differ between groups, the number of falls requiring medical attention was halved in those participating in the exercise program.

In Study 12, the 13-year results of the European Randomized Study of Screening for Prostate Cancer were reported.²⁰ The percentage of men dying from prostate cancer was approximately 0.12% lower in the screened group, a number needed to screen of approximately 800. However, there was no effect on all-cause mortality. Despite its large size (182,160 men), the study was still not powered for this outcome.

Diabetes Mellitus

Although intensive glycemic control has long been recommended for patients with type 2 diabetes mellitus, recent studies found benefits and harms (Table 4).^21–23 Study 13, the Veterans Affairs Diabetes Trial, randomized 1,791 veterans with type 2 diabetes to intensive or usual glycemic control, with average A1C values of 6.9% and 8.4%, respectively.²¹ The original study found no difference in cardiovascular events or mortality between groups. The current study reports a longer follow-up of 10 years for cardiovascular events and 12 years for overall mortality. The researchers found an NNT of 116 persons per year to prevent a composite of cardiovascular events (largely because of fewer nonfatal myocardial infarctions), but no difference in the likelihood of cardiovascular death or all-cause mortality. Study 14 provided a longer-term observational follow-up of participants from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial, and found similar small reductions in the likelihood of cardiovascular events.²² However, it is important to remember that the ACCORD trial was stopped early because a data safety and monitoring committee detected a clinically and statistically significant increase in cardiovascular and all-cause mortality in the intensive glycemic control group.

The story is somewhat different for patients with type 1 diabetes. In Study 15, a Swedish study used a registry of patients with type 1 diabetes to demonstrate a strong association between better glycemic control and lower cardiovascular mortality. Even for those with excellent glycemic control, though, cardiovascular mortality was higher than for persons without type 1 diabetes.²³

Cardiovascular Disease

The first of two cardiovascular studies (Table 5^24,25 ) examined the effectiveness of treatment for mild hypertension (140/90 to 159/99 mm Hg). Most systematic reviews combine the aggregate data from studies, but the researchers in Study 16 were able to use individual patient-level data for 15,266 patients randomized to active treatment or placebo.²⁴ Over an average five years of treatment, the NNT to prevent one stroke was 173 (95% confidence interval [CI], 108 to 810), to prevent one cardiovascular death was 95 (95% CI, 55 to 1,188), and to prevent one death overall was 99 (95% CI, 66 to 273). These estimates are clinically and statistically significant, but the CIs are wide.

A 2011 meta-analysis of nine studies with more than 100,000 participants found fewer major cardiovascular events (NNT = 253 over seven years) but more major bleeding complications (number needed to harm = 261 over seven years) for patients without known cardiovascular disease who were randomized to aspirin as primary prevention.²⁶ Study 17 was a large Japanese trial that randomized adults 60 years and older with cardiovascular risk factors but no known cardiovascular disease to aspirin or control and then followed them for five years.²⁵ There were fewer nonfatal myocardial infarctions and transient ischemic attacks in the treatment group, but there were also more bleeding complications and no difference in all-cause mortality. Overall, the authors concluded that there was no net benefit of aspirin.

Miscellaneous

Several top POEMs did not fit into one of the previous categories, but they provide important guidance for family physicians (Table 6).^27–30 Study 18 is a systematic review of 143 studies comparing 14 different regimens with standard triple therapy (a proton pump inhibitor plus clarithromycin [Biaxin] plus either metronidazole [Flagyl] or amoxicillin) for Helicobacter pylori eradication.²⁷ Three regimens, described in Table 6,^27–30 emerged as more effective than the others. Study 19 included 493 patients presenting to a primary care practice in Australia who were randomized to minor skin surgery with sterile gloves vs. boxed, nonsterilized gloves.²⁸ There was no difference in the likelihood of infection (9.3% in those treated with sterile gloves vs. 8.7% in those treated with nonsterile gloves). Finally, Study 20, a population-based prospective cohort study, followed 1,449 women 42 to 52 years of age to determine the duration of vasomotor symptoms during menopause.³⁰ The duration was longer for women whose symptoms began before menopause compared with those whose symptoms began after menopause (11.8 vs. 3.4 years). This information allows family physicians to better advise their patients.

Guidelines

These top 20 POEMs are based on an original research study or systematic review. Several POEMs that summarized practice guidelines were also highly rated. Guidelines from the U.S. Preventive Services Task Force included recommendations to not screen for thyroid disease,³¹ to confirm the diagnosis of mild to moderate hypertension with ambulatory or home blood pressure monitoring,³² and to screen patients who are obese for diabetes.³³ The American College of Physicians released guidelines that recommended against screening adults not at increased risk of cardiovascular disease,³⁴ and largely affirmed the recommendations of other organizations regarding cervical cancer screening.³⁵

The Department of Veterans Affairs and the Department of Defense have created guidelines for the use of lipid-lowering therapies³⁶ that are somewhat less aggressive than those of the American College of Cardiology and the American Heart Association (ACC/AHA),³⁷ recommending therapy for those with a 12% or higher 10-year risk of cardiovascular events, shared decision making for those between 6% and 12%, and no therapy for those with a 10-year risk less than 6%. They also recommend using a moderate, fixed-dose statin for most patients and not checking or following lipid levels.

Finally, the ACC/AHA guidelines on atrial fibrillation recommend that physicians use the CHA₂DS₂-VASc score (congestive heart failure; hypertension; age 75 years or older; diabetes; prior stroke, transient ischemic attack, or thromboembolism; vascular disease; age 65 to 74 years; sex category) to assess risk, prescribe warfarin (Coumadin) or novel oral anticoagulants, and consider rate control as an option.³⁸ The novel oral anticoagulants provide greater convenience with greater cost.

editor's note: This article was cowritten by Dr. Mark Ebell, who was a member of the U.S. Preventive Services Task Force (USPSTF) from 2012 to 2015 and currently serves as a consultant to the USPSTF. This article does not necessarily represent the views and policies of the USPSTF. Dr. Ebell is deputy editor for American Family Physician (AFP) and cofounder and editor-in-chief of Essential Evidence Plus, published by Wiley-Blackwell, Inc. The POEMs described in this article stem from work that Dr. Ebell and his colleagues have been doing for the past two decades. Medical journals occasionally publish an article summarizing the best studies in a certain field from the previous year; however, those articles are limited by being one person's idiosyncratic collection of a handful of studies. In contrast, this article by Drs. Ebell and Roland Grad is validated in two ways: (1) the source material (POEMs) was derived from a systematic review of thousands of articles using a rigorous criterion-based process, and (2) these “best of the best” summaries were rated by thousands of Canadian primary care physicians for relevance and benefits to practice.

Because of Dr. Ebell's dual roles and ties to Essential Evidence Plus, the concept for this article was independently reviewed and approved by a group of AFP's medical editors. In addition, the article underwent peer review and editing by four of AFP's medical editors. Dr. Ebell was not involved in the editorial decision-making process.—Jay Siwek, MD, Editor, American Family Physician

The authors thank Wiley-Blackwell, Inc., for giving permission to excerpt the POEMs; Drs. Allen Shaughnessy, Henry Barry, David Slawson, Nita Kulkarni, and Linda Speer for their work in selecting and writing the original POEMs; the academic family medicine fellows and faculty of the University of Missouri–Columbia, for their work as peer reviewers; Pierre Pluye, PhD, for his work in codeveloping the Information Assessment Method; and Maria Vlasak for her assistance with copyediting the POEMs for the past 22 years.