Risk framing alters perception of benefit for patients, doctors
How risk information is posed influences perceptions of treatment benefit, with absolute survival rates creating the perception of weakest benefit and relative mortality reduction creating the perception of greatest benefit, researchers found.
To determine which risk framing format best conveys information, and to compare framing bias in doctors and in patients, researchers mailed randomized surveys to 2,746 doctors in Geneva, Switzerland, and to 2,204 recently hospitalized patients. Completed surveys were returned by 1,431 doctors (56% response rate) and by 1,121 patients (65% response rate).
Respondents were asked to interpret the results of a hypothetical clinical trial comparing an old drug to a new one. They were randomly assigned to framing formats of absolute survival (96% for the new drug vs. 94% for the old drug), absolute mortality (4% vs. 6%), relative mortality reduction (reduction by a third) or all three (fully informed condition). Results were published in the December 2011 Journal of General Internal Medicine.
The risk presentation format influenced whether doctors rated the new drug as more effective (absolute survival, 51.8%; absolute mortality, 68.3%; relative mortality reduction, 93.8%; and fully informed condition, 69.8%; P<0.001). These proportions were similar in patients (absolute survival, 51.7%; absolute mortality, 66.8%; relative mortality reduction, 89.3%; and fully informed condition, 71.2%; P<0.001). None of the differences between doctors and patients were significant (all P>0.1).
The fully informed condition was similar to the absolute risk format for both doctors (P=0.72) and patients (P=0.23), but differed significantly from the other conditions (all P<0.01). Compared to the fully informed condition, the odds ratio of greater perceived effectiveness was 0.45 for absolute survival (P<0.001), 0.89 for absolute mortality (P=0.29), and 4.40 for relative mortality reduction (P<0.001). Absolute risks constitute the least biased risk format, the authors concluded. In contrast, relative risk reductions create an optimistic bias of a more than fourfold increase in the odds of a positive assessment of the new treatment. Absolute survival proportions caused a pessimistic bias, with a more than twofold decrease in the odds of a favorable assessment.
That doctors and patients had similar vulnerabilities to framing bias was unexpected, wrote the authors, who expected doctors would be more sophisticated in their analysis.
Doctors' understanding of various terms used in medical literature does not translate to forming an objective, criterion-based assessment, they wrote. Most doctors misunderstand numerical data about test accuracy and fail to use relevant numerical information, such as disease prevalence, when they interpret diagnostic test results. The solution is to present risk and benefit information in absolute and relative scales and to report absolute risks in medical research reports and other original sources of medical information used by doctors, they wrote.
Score predicts VTE risk for acutely ill patients at admission
Four clinical factors available at admission can be used to predict risk of venous thromboembolism (VTE) in acutely ill medical inpatients, a study indicates, though external validation is needed.
Researchers collected data from 15,156 patients at 52 hospitals in 12 countries between July 2002 and September 2006. Eligible patients had to be admitted for an acute medical illness, hospitalized for at least three days, and at least 18 years old. They were excluded if they had taken anticoagulant or thrombolytic drugs at or within 48 hours of admission, if they had had major surgery or trauma within three months of admission, if they had been admitted for deep vein thrombosis (DVT) or pulmonary embolism (PE) (or diagnosed within 24 hours of admission), or if follow-up was determined to be impossible. Patients were followed for signs of VTE for three months after admission, and researchers performed multiple regression analysis to identify factors associated with VTE risk. In total, 13,172 patients had three months of follow up data. Results were published in the September 2011 Chest.
One hundred eighty-four patients, or 1.2%, developed VTE by three months after admission; 41 were excluded from the main analyses for upper-limb DVT or VTE at an unknown site. Of the remainder, 76 had PE and 67 had lower-extremity DVT. Cumulative incidence of VTE was 1.0%; 45% of events occurred after discharge. Factors independently associated with VTE and available at admission—along with their point values in the risk score—were previous VTE (3 points), known thrombophilia (3 points), cancer (1 point) and being older than 60 years (1 point). Additional risk factors, apparent during hospitalization, were lower-limb paralysis, immobilization for at least seven days and admission to an ICU or coronary care unit. Patients with a score of 2 to 3 had a three-month VTE symptomatic event rate of 1.5% by the Kaplan-Meier method, while the rate was 5.7% for those with a score of 4 or greater. At admission, 67% of patients had a score of 1 or greater. During hospitalization, 31% had an associative score of 2 or greater. For a score of 2 or 3, observed VTE risk was 1.5%, compared to 5.7% for a score of 4 or greater. During hospitalization, a score of 2 or greater was associated with higher overall and VTE-related mortality. The risk calculator is available online.
The VTE risk factors identified in the study are consistent with those described by the most recent guidelines of the American College of Chest Physicians (ACCP), the authors noted. Once validated with external studies, the risk model should be a useful addition to ACCP guidelines for assessing VTE risk in acutely ill medical patients upon hospital admission, they said. Specifically, it will help assess the risks and benefits of drug prophylaxis (in conjunction with a bleeding risk score), including for high-risk medical patients with risk factors that don't fit into existing group-specific VTE risk categories, they wrote.
D-dimer plus decision rule or gestalt usually safe to exclude PE
Combining D-dimer testing with either a decision rule or physician's unstructured estimate (gestalt) is usually safe when evaluating adults for suspected pulmonary embolism (PE).
Researchers sought to compare the test characteristics of gestalt and clinical decision rules for evaluating adults with suspected PE, and assess the failure rate of gestalt and rules when combined with D-dimer testing. They searched articles in MEDLINE and EMBASE between 1966 and June 2011, and found 52 prospective studies of 55,268 inpatients, emergency department patients and referred patients who were thought to have PE. The studies estimated the probability of PE using gestalt or a decision rule, and verified the diagnosis using a reference standard. They identified five sets of clinical decision rules in the studies: the Wells rules (with cutoff values of <2 or ≤4) and simplified Wells rules, the Geneva rules (original, revised and simplified revised), the Pisa rules (original and revised), the Charlotte rule and the PERC rule.
Meta-analysis was performed on those studies that used gestalt (15 studies; sensitivity, 0.85; specificity, 0.51); the Wells rule with a cutoff value <2 (19 studies; sensitivity, 0.84; specificity, 0.58) or ≤4 (11 studies; sensitivity, 0.60; specificity, 0.80); the Geneva rule (5 studies; sensitivity, 0.84; specificity, 0.50); and the revised Geneva rule (4 studies; sensitivity, 0.91; specificity, 0.37). The other decision rules weren't included in the meta-analysis because they had been used in fewer than four studies. The analysis found that increased prevalence of PE was associated with higher sensitivity and lower specificity. Neither the clinical decision rules nor gestalt were sensitive enough to safely exclude PE on their own, but all could do so when combined with sensitive D-dimer testing. The meta-analysis was published in the Oct. 4, 2011 Annals of Internal Medicine.
The authors recommended using a decision rule instead of gestalt, since “physicians who use gestalt tend to assign a higher probability to PE to avoid missing it, thus causing more false-positive results and exposing more patients to unnecessary pulmonary imaging.” It's important to remember that the sensitivity of a decision rule increases, and specificity decreases, as PE prevalence increases, they noted. Physicians should use the strategy that fits their situation best, they concluded: in a high-prevalence setting, as with referred patients, a rule with higher specificity is desirable; in a lower-prevalence setting, rules with higher sensitivity are more desirable.
Residents' burnout, exhaustion and debt affect their medical learning
Burnout and educational debt are associated with lower scores among residents taking the Internal Medicine In-Training Examination (IM-ITE), according to a Sept. 7, 2011 Journal of the American Medical Association study.
To measure how well-being correlates with medical knowledge during residency, researchers used 2008 and 2009 IM-ITE scores and the 2008 IM-ITE survey. There were 16,394 residents surveyed, representing 74.1% of all eligible internal medicine residents, including U.S. and international medical graduates.
Overall quality of life was measured by self-assessment on a scale of 1 (“As bad as it can be”) to 5 (“As good as it can be”). Low quality of life was defined as the lowest two categories. Burnout was assessed using two measures adapted from a longer survey. Emotional exhaustion was assessed by the question, “How often do you feel burned out from your work?” and depersonalization by the question, “How often do you feel you've become more callous toward people since you started your residency?” Each question was answered on a 7-point Likert scale with response options ranging from “never” to “daily.”
About 15% of residents classified their quality of life as “as bad as it can be” or “somewhat bad”. At least weekly symptoms of emotional exhaustion were reported by 7,394 residents, or 45.8%. Weekly symptoms of depersonalization were reported by 4,541, or 28.9%. Overall, at least one symptom of burnout was present in 8,343 of 16,192 residents (51.5%).
Greater educational debt was associated with the presence of at least one symptom of burnout (61.5% of those with debt more than $200,000 vs. 43.7% of those with no debt; odds ratio, 1.72; P<0.001).
Decreased quality of life and increased frequency of burnout symptoms were associated with lower IM-ITE scores, most notably in the 242 residents (1.5%) reporting quality of life that was “as bad as it can be” (mean score, 57.6 vs. 60.3 for residents reporting quality of life “as good as it can be”; difference, 2.7 points; P<0.001). IM-ITE scores also were lower for the 600 residents (3.7%) with daily feelings of emotional exhaustion (mean score, 57.8 vs. 62.0 for residents never feeling burned out from work; difference, 4.2 points; P<0.001).
Residents reporting debt greater than $200,000 had mean IM-ITE scores 5.0 points (99% CI, 4.4 to 5.6; P<0.001) lower than those with no debt. These differences were as large as the increases normally seen as residents progress from their first to second postgraduate year (4.1 points) and their second to third postgraduate year (2.6 points).
Burnout was less common among international medical graduates than among U.S. medical graduates (45.1% vs. 58.7%; odds ratio, 0.70; P<0.001). This effect persisted after adjusting for debt. Researchers suggested that international medical graduates in the U.S. may be more resilient because they've already successfully completed a complex and highly competitive selection process for U.S. residency positions.
The study authors also noted that all members of the study cohort began training after duty hour limits went into effect in 2003, so burnout remains an issue despite these regulations.