Each year, a selection of 100 of the most widely read English-language research journals publish more than 20,000 articles.¹ “Information mastery” is a framework for evaluating the usefulness of medical information.² Based on this framework, the information deemed most useful is highly relevant (leading to improved patient outcomes), highly valid (the study was well-designed and without important biases), and takes little work to access (someone has summarized the results of the study). Studies that meet these criteria and have the potential to change practice are called POEMs, or patient-oriented evidence that matters.² The term “patient-oriented” refers to outcomes that are important to patients, such as decreased morbidity and mortality, and improved quality of life, rather than disease-oriented outcomes, such as changes in physiologic variables like serum glucose levels, glomerular filtration rate, and electrocardiographic findings.

Every month, a team of seven clinicians with expertise in primary care, hospital medicine, women's health, and pharmacology reviews 110 English-language research journals to identify relevant POEMs for primary care physicians. In 2011, approximately 250 studies met the criteria to be designated as a POEM; these studies were summarized in brief, structured critical appraisals. The POEMs are e-mailed to subscribers, and members of the Canadian Medical Association can rate the usefulness and impact of each POEM using a validated tool called the Information Assessment Method.³ From the physicians' ratings, a clinical relevance of information index was developed to identify the relevance of POEMs. The index, which ranges from 0 to 1, is based on the proportion of “totally relevant” and “not relevant” ratings. It also takes into account the total number of responses. The clinical relevance of information index was used to identify an initial list of the top 30 POEMs. POEMs were then excluded if they described a practice guideline rather than an original research study; received fewer than 300 physician ratings; or the editors believed they did not provide clear impetus for practice change.

In the final list of the top 20 POEMs, at least 55 percent of raters identified each POEM as “totally relevant” and less than 21 percent rated each as “not relevant.” This article provides the clinical question and bottom-line answer for each of the POEMs, organized by topic and followed by a brief discussion. Although many journals publish a list of the top studies from the previous year, we believe this approach has greater relevance and validity than a subjective list of studies selected by experts or editors because it is based on the actual ratings of practicing primary care physicians.

Cardiovascular Disease

The new oral anticoagulants, such as dabigatran (Pradaxa), are more convenient, but can cost up to $300 per month. Study 1 (Table 1^4–8 ) found that, for the average patient, dabigatran costs $86,000 for each additional quality-adjusted life-year gained.⁴ The researchers provide a useful diagram, available at http://circ.ahajournals.org/content/123/22/2562/F3.expansion.html, that helps clinicians identify the patients in whom the new drug is most cost-effective, based on the calculated risks of stroke and hemorrhage and time spent in the therapeutic international normalized ratio range.⁴

Studies 2 and 3 address one of the most common tasks performed in primary care: blood pressure measurement.^5,6 Study 2, which evaluated 444 men with hypertension who had a total of more than 100,000 blood pressure measurements over several years, found that each patient's measurements vary significantly. The most reliable way to establish a patient's blood pressure is to take several measurements over several weeks and average them.⁵ Study 3 found that in the office setting, using an automated blood pressure cuff more closely approximates ambulatory blood pressure monitoring than manual measurement (2.3 mm higher versus 6.5 mm higher), especially compared with measurements when the physician is in the room.⁶

Studies 4 and 5 provide some of the best evidence regarding the benefits of statins.^7,8 Study 4 pooled individual patient data from 26 studies and found that in patients with known cardiovascular disease, more intensive statin therapy was slightly more effective than less intensive therapy (number needed to treat [NNT] = 200 per year to prevent one additional cardiovascular event). However, all of the studies used fixed doses of statins, rather than aiming for specific low-density lipoprotein cholesterol targets, as is typically done in clinical practice. In the 14 placebo-controlled trials (including a mix of patients with and without preexisting cardiovascular disease), the NNT to prevent one cardiovascular event was 125 per year.⁷ Study 5 combined data from 76 trials with 170,255 patients and found that all-cause mortality was lower in patients taking a statin compared with patients taking placebo (NNT = 67 to prevent one death over approximately three years).⁸ Neither study found any difference between different types of statins.^7,8

Musculoskeletal Disease

Study 6 (Table 2^9–12 ) summarizes several systematic reviews of the evaluation of back pain. Routine imaging actually increases the risk of surgery, and does not decrease patient anxiety. The indications for imaging are limited to patients at high risk of a serious underlying cause, such as infection, cancer, or cauda equina syndrome, and to those with severe radicular pain.⁹

Study 7 randomized 401 patients with chronic back pain to massage or usual care and found that weekly massage provides persistent improvement in pain and function.¹⁰ Study 8 found a small increase in the risk of femoral neck or intertrochanteric fracture in women who took bisphosphonates for longer than five years. Because other studies have not shown a clear benefit to treatment for longer than five years, the study authors suggest discontinuing the drugs at that point for most women.¹¹

Diabetes Mellitus

Several large, well-designed randomized trials found no benefit to intensive glucose control (e.g., A1C target less than 6.5 percent) in middle-aged patients with type 2 diabetes.^13–15 Studies 9 and 10 (Table 3^16–18 ) confirm these findings.^16,17 Study 9 was a five-year follow-up of the ACCORD (Action to Control Cardiovascular Risk in Diabetes) study that found persistent elevation in all-cause mortality in persons randomized to intensive glucose control.¹⁶ Study 10 combined the results of 13 trials with more than 34,500 patients and found no improvement in all-cause mortality or cardiovascular death with intensive glucose control; the only benefit was a small reduction in the likelihood of nonfatal myocardial infarction.¹⁷

Preventing diabetes is obviously the best strategy. Study 11 randomized patients to dietary counseling about a Mediterranean-type diet, supported with free mixed nuts and olive oil, or to counseling about a low-fat diet. After four years, patients on a Mediterranean-type diet were less likely to develop diabetes (10 versus 18 percent; NNT = 12).¹⁸

Gastrointestinal Disease

Functional abdominal problems can be frustrating to manage, but several new POEMs provide some useful insights. Study 12 (Table 4^19–21 ) identified patients with normal upper endoscopic findings and no evidence of Helicobacter pylori infection. Those randomized to receive a proton pump inhibitor instead of placebo were more likely to have a significant improvement in symptoms (36 versus 6 percent), but primarily if they had a body mass index greater than 25 kg per m² (53 percent response rate).¹⁹

Study 13 randomized 132 adults with mild to moderate irritable bowel syndrome to receive a probiotic or placebo, and found that those taking the probiotic were more likely to have a clinically significant response (57 versus 21 percent; NNT = 3).²⁰ Study 14 found that although consuming whole prunes leads to a significant improvement in constipation, a person would have to eat about 12 (260 calories worth) per day to benefit.²¹

Screening and Prevention

Study 15 (Table 5^22–24 ) reassures physicians that asking about suicidal ideation does not “plant the thought” in patients with depression. The researchers randomized 443 patients with depression to either be asked six questions about suicidal thoughts and ideas, or six questions about health and lifestyle in general. At a follow-up visit, there was no difference in the proportion who had experienced suicidal ideation.²²

Study 16, a meta-analysis of 14 randomized controlled trials, found that the risk of adverse cardiovascular events was significantly increased in persons taking varenicline (Chantix); however, most studies excluded persons with known heart disease. Of course, if smokers successfully quit smoking, their risk of cardiovascular events decreases, but it does suggest that alternative approaches be considered for those with known cardiovascular disease.²³

In study 17, older adults who were obese and frail were randomized to weight loss, exercise (aerobic, resistance, and flexibility), both, or usual care. Patients in all of the intervention groups had improvements on a physical performance test at the end of one year, and both weight loss groups lost about 20 lb (9 kg), with the greatest improvement occurring in the group that was randomized to both weight loss and exercise.²⁴

Miscellaneous

Almost every treatment has harms as well as benefits. A physician's task is to make sure the likelihood of benefit outweighs the likelihood of harm for most patients most of the time. Study 18 (Table 6^25–27 ) is a systematic review of corticosteroid inhalers in patients with chronic obstructive pulmonary disease. Although there is a known benefit of one fewer exacerbation per year for every six persons who use an inhaler, there is also one fracture every three years for every 83 patients.²⁵

Many physicians prescribe only three packs of oral contraceptives at a time, and some state laws even prevent prescribing more than that at a time. Study 19 linked data from the California Medicaid program to a birth registry and found that women given 12 or 13 packs of oral contraceptives were less likely to become pregnant (1.2 versus 3.3 percent) and less likely to have an abortion (0.18 versus 0.63 percent) than women receiving three packs at a time. After adjusting for potential confounders, there was still a 30 percent reduction in pregnancy and a 46 percent reduction in the likelihood of induced abortion.²⁶

Study 20 randomized 240 adolescents and adults with a plantar wart to up to four sessions of cryotherapy (mean of 1.6 sessions) or eight weeks of daily pumice stone use plus application of salicylic acid 50%. After six months, results were identical (and unimpressive) for both groups, although patients tended to prefer the cryotherapy.²⁷