The physical exam as a diagnostic tool, not an empty formality.
The traditional system-based exam teaches you to go through the motions. BedsideDx helps you actually use the physical exam to make a diagnosis. Every test comes with clear technique guidance, the evidence behind it, and a calculator so you can know exactly what it means for your diagnostic probability.
Ben Wilson, MD
Ben Wilson, MD, founder of BedsideDx, is a board-certified, actively-practicing family physician who has been teaching medicine for more than 10 years. He is passionate about improving the lives of patients through excellent bedside diagnosis.
- Comprehensive organ-system exams, performed systematically
- Every maneuver in the sequence, every visit
- Designed for thoroughness in a controlled teaching environment
- Findings documented — but rarely weighted by diagnostic value
- The gold standard, built for completeness
- Every maneuver is a diagnostic test with a likelihood ratio
- Start with a clinical question, choose the exam that answers it
- Know whether a positive or negative result actually shifts the probability
- Technique guidance for real clinical settings, not just ideal ones
- Evidence at the point of care, when you need it
From the ideal to the practical.
Medical school teaches the comprehensive physical exam for good reason — it builds the foundation of clinical pattern recognition and trains systematic thinking. That approach is the right way to learn medicine, and the thoroughness it instills is genuinely valuable.
In practice, though, most clinicians work in conditions that don't always allow for the full textbook exam: time constraints, a focused clinical question, a patient who came in for something specific. The result is often an exam that follows the motions of what was taught rather than being driven by the diagnostic question at hand. Or sometimes the physical exam gets dropped altogether.
BedsideDx is designed to bridge that gap. The technique guidance for each test reflects how to get an accurate, clinically valid result in realistic conditions — so the exam you actually perform is purposeful, not just perfunctory.
The evidence behind every finding.
Each entry pairs the clinical maneuver with the published diagnostic accuracy data: likelihood ratios, confidence intervals, and a plain-language read on whether a positive or negative result is actually worth acting on.
The evidence behind the labels
For each finding, we search the published diagnostic-accuracy literature and report the likelihood ratios from the best available source. Every number on the site links back to a citation you can verify, and every finding is tagged with one of four evidence categories so you know how much to trust the number at a glance:
- Strong — from a rigorous existing synthesis (Cochrane DTA, JAMA Rational Clinical Examination, or another meta-analysis with formal risk-of-bias appraisal).
- Moderate — anchored on a substantial single-cohort primary study, used when the available systematic reviews exist but do not separately pool that specific finding-diagnosis pair. Uncommon on the site.
- Limited — a single primary diagnostic-accuracy study, or a defensible extraction from within a higher-tier review.
- Insufficient Evidence — the finding is included because clinicians traditionally look for it, but the literature search returned no source that met our inclusion criteria. We don’t invent a number.
The helpfulness label on each finding (Very helpful, Helpful, Somewhat helpful, Minimally helpful, Not helpful) translates the LR into a plain-language read of whether the result is actually worth acting on. We follow the standard Sackett interpretation thresholds and flag findings whose confidence interval crosses 1.0 as Not helpful, regardless of how the point estimate looks — a result that’s statistically indistinguishable from a useless test shouldn’t shift your decisions.