Medical publishing · Clinical calculators · Risk scores · Nomograms · Shared decision making · Case-based CME · Webinars
Doctors are drowning in content, starved of tools.
Orakle is a medical publisher and education provider. We build the calculators, risk scores and case-based learning clinicians reach for in the middle of a consultation.
There is no shortage of courses, modules and mandatory box-ticking. There is a shortage of things that change what happens next.
Validated clinical calculators, risk scores, nomograms and shared decision making tools, serving health professionals worldwide, built for use inside the consultation.
Case-based learning modules, webinars and podcasts. The education clinicians choose to finish, rather than the education they are required to complete.
01 Evidence
02 Calculation
03 Decision
A number becomes a decision about one person
How we work
Reference tools, not black boxes.
Evidence moves slowly into practice. We build and validate the scores and tools that close that gap.
01
Reference, not diagnosis
We reproduce published scores, we do not interpret them for you
02
Nothing stored, nothing sent
No patient data ever reaches a server of ours
03
Logins, never patients
DxTx accounts hold preferences, not people
04
Every number traceable
Each score links back to its source publication
05
Built to be readable
WCAG standards, on the oldest clinic phone
Every output remains yours to interpret.
The evidence test
What the literature actually says.
Sixteen questions from Cochrane reviews, meta-analyses and landmark trials, about what medical education really does to physician behaviour, and to patients. The findings are rarely the ones any of us would predict in advance.
Choose an answer. The evidence and citation appear immediately.
01 / 16
A 2021 Cochrane review directly compared interactive educational meetings with didactic (lecture-based) educational meetings for improving health professionals' compliance with desired practice. What did the review conclude about this comparison?
Correct
Not quite — the answer is A
Based on 7 trials, adjusted risk difference in compliance favoured interactive meetings by 4.18% (95% CI 3.87 to 4.49), but the review rated this as very low-certainty evidence, stating: "we are uncertain of the effects on compliance with desired practice of interactive educational meetings compared with didactic educational meetings." Notably, a separate meta-regression across a different, larger set of studies pointed the opposite way (didactic favoured by 1.21%, 95% CI 0.85 to 1.57) — an inconsistency the authors discuss, attributing it partly to difficulty classifying real-world meetings as purely interactive or didactic.
Forsetlund L, O'Brien MA, Forsén L, Mwai L, Reinar LM, Okwen MP, Horsley T, Rose CJ. Continuing education meetings and workshops: effects on professional practice and healthcare outcomes. Cochrane Database of Systematic Reviews 2021, Issue 9. Art. No.: CD003030. DOI: 10.1002/14651858.CD003030.pub3.
02 / 16
A systematic review pooled 20 comparisons between physicians' self-assessment of their own skill or knowledge and their actual, externally observed competence. What was the most common pattern found?
Correct
Not quite — the answer is B
Of the 20 comparisons reviewed, 13 (65%) showed little, no, or an inverse relationship between self-rated and observed competence; only 7 showed a positive relationship. Accuracy was consistently worst among the physicians who were both least skilled and most confident — a Dunning-Kruger pattern. This directly undercuts a foundational premise of self-directed CME: that clinicians can reliably identify their own learning gaps.
Davis DA, Mazmanian PE, Fordis M, Van Harrison R, Thorpe KE, Perrier L. Accuracy of Physician Self-Assessment Compared With Observed Measures of Competence: A Systematic Review. JAMA. 2006;296(9):1094-1102. doi:10.1001/jama.296.9.1094. PMID: 16954489.
03 / 16
A systematic review examined 62 evaluations of the relationship between physicians' years of clinical experience and measures of care quality or knowledge. What did it find?
Correct
Not quite — the answer is C
52% of the 62 evaluations found declining performance with more years since training across all outcomes studied, and another 21% found decline on at least some outcomes — only 4% found a positive association overall. Every one of the 12 studies specifically measuring physician knowledge found it decreased the longer a physician had been out of training. A related study cited in the review found roughly a 0.5% increase in patient mortality after acute MI for each additional year since the treating physician's medical school graduation.
Choudhry NK, Fletcher RH, Soumerai SB. Systematic Review: The Relationship between Clinical Experience and Quality of Health Care. Ann Intern Med. 2005;142(4):260-273. doi:10.7326/0003-4819-142-4-200502150-00008. PMID: 15710959.
04 / 16
A study of Medicare prescribing data linked to the U.S. Open Payments database looked at physicians who received industry-sponsored meals (average value under $20, usually promoting a single brand-name drug) versus physicians who received none. What was the association?
Correct
Not quite — the answer is D
Among 279,669 physicians, receiving even a single sponsored meal was associated with higher rates of prescribing the promoted drug. The association was dose-dependent: physicians receiving four or more meals prescribed rosuvastatin at 1.8x the rate, nebivolol at 5.4x, olmesartan at 4.5x, and desvenlafaxine at 3.4x the rate of physicians who received no industry meals — driven by a gift often costing less than a takeout lunch.
DeJong C, Aguilar T, Tseng CW, Lin GA, Boscardin WJ, Dudley RA. Pharmaceutical Industry–Sponsored Meals and Physician Prescribing Patterns for Medicare Beneficiaries. JAMA Intern Med. 2016;176(8):1114-1122. doi:10.1001/jamainternmed.2016.2765. PMID: 27322350.
05 / 16
A study of over 3,600 general internists linked their Maintenance of Certification (MOC) exam scores to the quality of care received by more than 220,000 of their Medicare patients. Comparing physicians who scored in the top quartile to those in the bottom quartile, what did the study find?
Correct
Not quite — the answer is A
Top- vs. bottom-quartile MOC scores were weakly associated with better diabetes composite care (OR 1.17, 95% CI 1.07–1.27) and mammography screening (OR 1.14, 95% CI 1.08–1.21) — but showed essentially no association with lipid testing for cardiovascular disease (OR 1.00, 95% CI 0.91–1.10). Even where statistically significant, the odds ratios sit close to 1.0, a thin evidence base for the assumption that passing a recertification exam ensures better patient care.
Holmboe ES, Wang Y, Meehan TP, Tate JP, Ho SY, Starkey KS, Lipner RS. Association Between Maintenance of Certification Examination Scores and Quality of Care for Medicare Beneficiaries. Arch Intern Med. 2008;168(13):1396-1403. PMID: 18625919.
06 / 16
A Cochrane review pooled 140 studies of audit-and-feedback interventions (where clinicians are shown data on their own performance) for changing professional practice. What was the median effect size found?
Correct
Not quite — the answer is B
The weighted median absolute improvement in compliance with desired practice was only 4.3% (interquartile range 0.5% to 16%), and the effect varied widely depending on baseline performance and how the feedback was delivered. Audit and feedback is often treated as a reliable quality-improvement lever, but the median real-world effect is modest and inconsistent — comparable in scale to the didactic-vs-interactive findings in Q1.
Ivers N, Jamtvedt G, Flottorp S, et al. Audit and Feedback: Effects on Professional Practice and Healthcare Outcomes. Cochrane Database of Systematic Reviews. 2012;6:CD000259. doi:10.1002/14651858.CD000259.pub3.
07 / 16
A meta-analysis pooled 31 studies (61 interventions) and calculated separate effect sizes for how CME affects physician knowledge, physician performance, and patient health outcomes. What was the general pattern?
Correct
Not quite — the answer is C
The overall sample-size-weighted effect size across all interventions was r = 0.28. Broken out by domain, the effect on physician knowledge was medium-sized (up to r ≈ 0.33 for interactive methods), while the effect on both physician performance and patient health outcomes was small. In the authors' own words: "the effect size of CME on physician knowledge is a medium one; however, the effect size is small for physician performance and patient outcome." The further an outcome sits from a multiple-choice test, the weaker the evidence gets.
Mansouri M, Lockyer J. A meta-analysis of continuing medical education effectiveness. J Contin Educ Health Prof. 2007;27(1):6-15. doi:10.1002/chp.88. PMID: 17385735.
08 / 16
By 2015, researchers had accumulated 39 systematic reviews of CME effectiveness spanning nearly 40 years (dating back to 1977). Synthesizing all of them, what did the authors conclude about the strength of evidence for CME improving physician performance versus improving patient health outcomes?
Correct
Not quite — the answer is D
After synthesizing 39 systematic reviews, the authors concluded that "CME does improve physician performance and, to a lesser and less consistent extent, patient health outcomes." Even after nearly four decades of research and dozens of reviews, the link between CME and what actually happens to patients remains the weakest, least consistent part of the evidence base — the part CME is ultimately supposed to justify.
Cervero RM, Gaines JK. The Impact of CME on Physician Performance and Patient Health Outcomes: An Updated Synthesis of Systematic Reviews. J Contin Educ Health Prof. 2015;35(2):131-138. doi:10.1002/chp.21290. PMID: 26115113.
09 / 16
A landmark systematic review of 100 studies examined computerized clinical decision support systems (CDSS) — drug-dosing calculators, diagnostic algorithms, reminder systems — assessing both practitioner performance and patient outcomes. What was found?
Correct
Not quite — the answer is A
Practitioner performance was assessed in 97 studies and improved in 62 of them (64%). Patient outcomes were assessed in far fewer trials — just 52 — and improved in only 7 (13%). Reminder systems fared best for changing behaviour (76%), diagnostic systems worst (40%). The gap between "changed what the clinician did" and "changed what happened to the patient" is enormous — and most of these tools are marketed and adopted on the strength of the first number, not the second.
Garg AX, Adhikari NK, McDonald H, et al. Effects of Computerized Clinical Decision Support Systems on Practitioner Performance and Patient Outcomes: A Systematic Review. JAMA. 2005;293(10):1223-1238. PMID: 15755945.
10 / 16
Researchers analyzed which specific design features predicted whether a clinical decision support system (e.g. a risk calculator or alert built into an EHR) would actually change clinical practice, across 70 trials and ~130,000 patients. Which single feature had the strongest statistical association with success?
Correct
Not quite — the answer is B
Automatic delivery within existing workflow had an adjusted odds ratio of 112.1 (95% CI 12.9 to infinity, P<0.00001) for predicting success — a larger effect size than almost any drug in medicine. Three other features also predicted success: delivering the recommendation at the time and place of decision-making (OR 15.4), providing an actual recommendation rather than just an assessment (OR 7.1), and being computer-based rather than paper-based (OR 6.3). When all four features were present, 94% of systems (30 of 32) improved practice. In other words, the tool's clinical content mattered far less than whether it was simply shoved in front of the clinician at the right moment.
Kawamoto K, Houlihan CA, Balas EA, Lobach DF. Improving clinical practice using clinical decision support systems: a systematic review of trials to identify features critical to success. BMJ. 2005;330(7494):765. PMID: 15767266.
11 / 16
A large multicentre trial introduced the Ottawa Ankle Rules — a simple bedside decision rule — across emergency departments and measured its effect on X-ray use for acute ankle injuries, compared with usual practice, in nearly 13,000 patients. What happened?
Correct
Not quite — the answer is C
Across 12,777 patients, radiography use fell from 82.8% in usual-care hospitals to 60.9% where the rule was introduced (P<.001), with virtually no difference in fractures missed at ED discharge (0.5% vs. 0.4%). An earlier single-centre study found the same pattern in more detail: 28% relative reduction in ankle X-rays, 100% sensitivity for the fractures that mattered, and shorter ED visits (80 vs. 116 minutes) at lower cost — proof that a well-designed, well-implemented decision rule can cut unnecessary testing without missing significant injuries.
Stiell IG, McKnight RD, Greenberg GH, et al. Implementation of the Ottawa Ankle Rules. JAMA. 1994;271(11):827-832. PMID: 8114236. Stiell IG, Wells GA, Laupacis A, et al. Multicentre trial to introduce the Ottawa ankle rules for use of radiography in acute ankle injuries. BMJ. 1995;311(7005):594-597. PMID: 7663253.
12 / 16
A randomized controlled trial across 19 Canadian hospitals tested whether using a validated pneumonia severity risk score to guide admission decisions could safely shift more low-risk patients to outpatient treatment, compared with usual clinical judgment. What happened?
Correct
Not quite — the answer is D
In this trial of 1,743 patients, outpatient treatment of patients identified as low-risk by the score rose from 31% to 49% (P=.01), average bed-days per patient fell from 6.1 to 4.4, and IV therapy duration fell similarly — while quality-of-life scores and clinical outcomes were statistically equivalent between groups. Used well, a risk score let clinicians safely send roughly one in five more pneumonia patients home who would otherwise have been admitted.
Marrie TJ, Lau CY, Wheeler SL, et al. A Controlled Trial of a Critical Pathway for Treatment of Community-Acquired Pneumonia. JAMA. 2000;283(6):749-755. PMID: 10683053.
13 / 16
A prospective study of over 42,000 children with head trauma tested a clinical prediction rule designed to identify kids at such low risk of significant brain injury that a CT scan could safely be skipped. How well did the rule perform at correctly identifying children who truly had no clinically important brain injury?
Correct
Not quite — the answer is A
In children under 2, the rule's negative predictive value for clinically important brain injury was 100% (95% CI 99.7–100%); in children 2 and older, 99.95% (95% CI 99.81–99.99%). Applying the rule would have safely avoided CT imaging in roughly 20–24% of children who were actually scanned, without missing injuries that mattered. Out of 42,412 children, only 376 (0.9%) had a clinically important brain injury — and the rule caught essentially all of them while sparing thousands of children unnecessary radiation.
Kuppermann N, Holmes JF, Dayan PS, et al. Identification of children at very low risk of clinically-important brain injuries after head trauma: a prospective cohort study. Lancet. 2009;374(9696):1160-1170. PMID: 19759692.
14 / 16
The "Christopher Study" tested an algorithm combining a clinical probability score (Wells score), D-dimer blood testing, and CT scanning to diagnose or exclude pulmonary embolism (PE) in over 3,300 patients. For patients classified as low probability with a normal D-dimer — who were left untreated and had no further imaging — what happened over the next three months?
Correct
Not quite — the answer is B
32% of the full cohort (1,057 of 3,306 patients) had a "PE unlikely" Wells score plus a normal D-dimer; of the 1,028 who went untreated on this basis, only 5 (0.5%, 95% CI 0.2–1.1%) had a confirmed venous thromboembolism in the following 3 months — a failure rate low enough to be considered clinically safe. The algorithm as a whole completed management using only clinical scoring, D-dimer, and CT (no further testing) in 97.9% of the entire cohort.
van Belle A, Büller HR, Huisman MV, et al. (Christopher Study Investigators). Effectiveness of Managing Suspected Pulmonary Embolism Using an Algorithm Combining Clinical Probability, D-Dimer Testing, and Computed Tomography. JAMA. 2006;295(2):172-179. PMID: 16403929.
15 / 16
The Glasgow-Blatchford Score (GBS) is used to identify patients with upper GI bleeding who are at low enough risk to be managed as outpatients rather than admitted. When implemented prospectively, what happened to hospital admission rates for patients scoring zero (lowest risk) on the scale?
Correct
Not quite — the answer is C
In the prospective implementation cohort, 22% of patients were classified as low-risk (score of 0); of these, 68% were managed entirely as outpatients with no adverse events — no deaths, no rebleeding. Hospital admission for this population fell from 96% to 71% (P<0.00001) once the score was put into routine use. A four-question bedside score safely kept roughly a quarter of a high-stakes patient population out of the hospital altogether.
Stanley AJ, Ashley D, Dalton HR, et al. Outpatient management of patients with low-risk upper-gastrointestinal haemorrhage: multicentre validation and prospective evaluation. Lancet. 2009;373(9657):42-47. PMID: 19091393.
16 / 16
A meta-analysis of 7 randomized controlled trials (604 participants) compared case-based learning combined with problem-based learning against traditional lecture-based teaching in clinical medical education, measuring theoretical knowledge, practical skills, and clinical thinking. How large were the effect sizes found?
Correct
Not quite — the answer is D
The pooled effect sizes were striking: theoretical knowledge SMD 2.16 (95% CI 1.22 to 3.11), practical skills SMD 1.59 (95% CI 1.04 to 2.15), and clinical thinking SMD 3.66 (95% CI 1.75 to 5.57) — all P<.0001 — plus a strong effect on learning satisfaction (0.86, 95% CI 0.81 to 0.91). These are unusually large effect sizes for an education intervention, several times bigger than most CME effects seen elsewhere on this list, though it should be read as case-based-plus-problem-based learning combined against lecture, not case-based learning in isolation, and is based on a relatively small pooled sample (604 participants across 7 trials).
Lu B-R, Shi X-Y, An L, He K, Guo M, Sun Z-G. Effectiveness of case-based learning combined with problem-based learning versus lecture-based learning in clinical medical education: a systematic review and meta-analysis. Postgraduate Medical Journal. 2026;102(1209):615-625.
01 / 16
Out of 16. Never stored, never sent, forgotten the moment you leave this page.
Every figure here is drawn from a peer-reviewed source, cited in full. Nothing is paraphrased from memory, and nothing is ours.
Selected recent activity
The work, and who with.
Calculators, decision aids and accredited case simulations — built with medical societies and patient associations, live in eight countries. Nothing ships untested. We build the platform; the society keeps its own name, its data, its final clinical word.
The data was never the problem — the system was. Guidelines and risk scores get proven, then sit untested at the bedside, doing nothing for the ill they were built for. We close that gap: your research, live in a clinician's hands — accredited, branded, running on almost anything.
Small team, long memories. Between us we have shipped medical education, priced risk, and sat on both sides of a pharmaceutical budget.
Seun Moses
Founder
A veteran of medical publishing. Former account manager at Informa and Touch Briefings, with a master's in health economics and an earlier career trading derivatives.
Joshua Ojegbile
Digital lead
Fifteen years in JavaScript, TypeScript, C++ and Java. Builds everything client-side, so the arithmetic runs on the clinician's own device and the patient's data has nowhere at all it could go.
August Felix
Partnerships
Leads partnerships with medical societies and patient associations, from the first conversation through to signed governance, and answers personally for every claim a finished Orakle tool makes.
Privacy and accessibility
What we collect: nothing.
Privacy
This site sets no cookies, runs no analytics and makes no third-party requests. Fonts are served from our own domain. Our calculators run entirely in your browser; no patient information is ever transmitted to us, and none of it is stored by us at all.
Accessibility
We build to WCAG 2.2 AA and EN 301 549. Keyboard navigable, screen reader tested, motion respects your system settings. Found a barrier? Tell us and we will fix it.