Interval likelihood ratios improved correct management decision rates by 11.2 to 18.3 percentage pointsTrial shows interval likelihood ratios improve decisions for heart conditions

This multicenter randomized trial evaluated the impact of different diagnostic information presentation formats on the management decisions of 611 healthcare providers, including internal medicine residents, attending physicians, and advanced practice providers. The study specifically focused on two clinical scenarios: pulmonary embolism and heart failure. The primary objective was to determine if presenting test results as interval likelihood ratios (ILRs) would improve decision accuracy compared to traditional single-threshold or binary test characteristics.

The intervention group received diagnostic information presented as ILRs, while the comparator groups received information in either a single-threshold format or a binary format. These formats were used to present data regarding pulmonary embolism and heart failure cases to the participants to assess their management choices.

Regarding the primary outcome of correct management decision rates, the study found that the ILR group achieved a 93.4% correct decision rate compared to 82.2% in the single-threshold group. This represents an improvement of 11.2 percentage points (95% CI, 6.9-15.6; P<0.001). When comparing ILR to binary test characteristics, the ILR group achieved a 93.4% correct decision rate compared to 75.1% in the binary group, an improvement of 18.3 percentage points (95% CI, 13.5-23.1; P<0.001).

Secondary outcomes included specific analysis of case types and subgroup comparisons. For pulmonary embolism cases, the correct response rates were 91.9% for the ILR group, 70.9% for the threshold group, and 64.7% for the binary group. For heart failure cases, the correct response rates were 94.9% for the ILR group, 93.4% for the threshold group, and 85.6% for the binary group. These results indicate that ILRs consistently outperformed other formats across both clinical scenarios.

Safety and tolerability data were not reported in this study as it involved a decision-making assessment rather than an intervention involving medication or physical procedures. No adverse events or discontinuations were recorded.

These results suggest that the method of presenting diagnostic information significantly influences clinician decision-making. While previous literature has explored various management strategies for heart failure and pulmonary embolism, this study specifically addresses the educational and communication aspect of clinical data presentation. The randomized trial design suggests a causal link between the presentation format (ILR) and the resulting accuracy in clinical decisions.

A primary limitation of this study is that it utilized fictional vignettes rather than real-time clinical scenarios to assess decision-making. This may impact the generalizability of the findings to high-pressure, real-world clinical environments where patients are present. Furthermore, while the trial showed significant improvements for ILRs, the specific factors that make ILRs more intuitive for clinicians were not explored in depth.

For clinical practice, these results suggest that incorporating interval likelihood ratios into diagnostic reports may improve decision accuracy among residents and advanced practitioners. However, because the study used fictional vignettes, the transition to real-world application should be monitored. Questions remain regarding whether this improvement persists across a wider range of medical specialties or in more complex, multi-morbidity cases beyond those presented in the vignette format.