Likelihood Ratios and Bayesian Reasoning in Psychiat... | MDster You are offline Back online! [ MDster home ](/ "MDster home") Specialities [ Anesthesiology ](https://mdster.com/speciality/anesthesiology) [ Emergency Medicine ](https://mdster.com/speciality/emergency-medicine) [ Family Medicine ](https://mdster.com/speciality/family-medicine) [ Internal Medicine ](https://mdster.com/speciality/internal-medicine) [ Obstetrics & Gynecology ](https://mdster.com/speciality/obstetrics-gynecology) [ Pediatrics ](https://mdster.com/speciality/pediatrics) [ Psychiatry ](https://mdster.com/speciality/psychiatry) [ Features ](https://mdster.com/features) [ Pricing ](https://mdster.com/pricing) [ Blog ](https://mdster.com/blog) Menu Specialities [ Anesthesiology ](https://mdster.com/speciality/anesthesiology) [ Emergency Medicine ](https://mdster.com/speciality/emergency-medicine) [ Family Medicine ](https://mdster.com/speciality/family-medicine) [ Internal Medicine ](https://mdster.com/speciality/internal-medicine) [ Obstetrics & Gynecology ](https://mdster.com/speciality/obstetrics-gynecology) [ Pediatrics ](https://mdster.com/speciality/pediatrics) [ Psychiatry ](https://mdster.com/speciality/psychiatry) [ Features ](https://mdster.com/features) [ Pricing ](https://mdster.com/pricing) [ Blog ](https://mdster.com/blog) [ Login ](https://mdster.com/auth/login) 1. [ Home ](https://mdster.com) 2. [ Blog ](https://mdster.com/blog) 3. [ Medical Education ](https://mdster.com/blog?category=medical-education) 4. Likelihood Ratios and Bayesian Reasoning in Psychiatry [ Medical Education ](https://mdster.com/blog?category=medical-education) Likelihood Ratios and Bayesian Reasoning in Psychiatry ======================================================== How to turn screening results into post-test probability instead of overcalling diagnoses [ MDster Editorial Team ](https://mdster.com/about) · Apr 15, 2026 · 2 min read · 63 [ Reviewed by Dr. Ali Ragab, MBBCH, MSc, MCAI ](https://mdster.com/medical-reviewers/dr-ali-ragab) [Editorial Policy](https://mdster.com/editorial-policy) | [Corrections Policy](https://mdster.com/corrections) [ Board Review ](https://mdster.com/blog?tag=board-review) [ Medical Education ](https://mdster.com/blog?tag=medical-education) [ Psychiatry ](https://mdster.com/blog?tag=psychiatry) [ Evidence-Based Practice ](https://mdster.com/blog?tag=evidence-based-practice) [ Diagnostic Test Evaluation ](https://mdster.com/blog?tag=diagnostic-test-evaluation) ![Likelihood Ratios and Bayesian Reasoning in Psychiatry](https://mdster.com/storage/blog/images/likelihood-ratios-and-bayesian-reasoning-in-psychiatry.jpg) Share this article Share this post You're on consults. The ED hands you a positive PHQ-9 or PTSD screen and asks, 'So does this patient have the disorder?' If you answer yes, you've already lost the Bayesian plot. In psychiatry, most tools are not diagnostic end points; they are probability shifters. Used well, they sharpen judgment. Used badly, they lead to over-interpretation and wrong diagnosis, exactly the problem classic psychiatry appraisal papers warn about. [\[1\]](#cite-1 "Reference [1]") Start with pretest probability ------------------------------ Bayesian reasoning is simple: begin with how likely the disorder is before the test, then update with the test result. In practice, pretest probability comes from setting, prevalence, symptom quality, time course, collateral, substance use, medical causes, and the MSE—not from vibes. The math is straightforward: `pretest odds = p/(1-p)`, `post-test odds = pretest odds × LR`, and `post-test probability = odds/(1+odds)`. [\[1\]](#cite-1 "Reference [1]") This is why the same questionnaire means different things in different clinics. A positive depression screen in a high-risk CL population carries a different post-test probability than the same score in low-prevalence community screening. LRs are more portable than predictive values because they are less tied to prevalence, but they only become clinically meaningful once you combine them with a patient-specific pretest probability. [\[2\]](#cite-2 "Reference [2]") LR+ and LR-: know what each actually does ----------------------------------------- **LR+** tells you how much to increase the odds when the test is positive: `sensitivity / (1 - specificity)`. **LR-** tells you how much to decrease the odds when the test is negative: `(1 - sensitivity) / specificity`. An LR of 1 changes nothing; the farther LR+ is above 1, or LR- below 1, the more useful the test. [\[2\]](#cite-2 "Reference [2]") Likelihood ratioBoard-style interpretationPractical meaning**LR+ >10** or **LR- <0.1**Large shiftOften strong rule-in or rule-out**LR+ 5-10** or **LR- 0.1-0.2**Moderate shiftHelpful, but not usually definitive alone**LR+ 2-5** or **LR- 0.2-0.5**Small shiftAdds context more than certainty**LR ≈1**No shiftThe test adds little These cutoffs are heuristics, not laws, but they are the ones boards expect and the ones AHRQ and classic Users' Guides teaching treat as clinically informative. [\[3\]](#cite-3 "Reference [3]") The common exam trap is to confuse a highly sensitive test with a strong rule-out in every patient. Wrong. Rule-out power is captured by **LR-**, and even a good LR- may leave substantial residual probability if the pretest probability was high to begin with. Likewise, a respectable LR+ does not rescue a test used in a population where the diagnosis was unlikely from the start. [\[2\]](#cite-2 "Reference [2]") Turn a psychiatry screen into post-test probability --------------------------------------------------- The PHQ-9 is a good teaching example. A meta-analysis found that at the usual cutoff of **≥10**, pooled sensitivity was **0.85** and specificity **0.89** for major depression, which gives an approximate **LR+ of 7.7** and **LR- of 0.17**. Using those pooled estimates, if your pretest probability is 30%, a positive PHQ-9 moves the post-test probability to about 77%, while a negative result drops it to about 7%. If the pretest probability is only 5%, the same positive result reaches only about 29%. Same test, different patient, different meaning. [\[4\]](#cite-4 "Reference [4]") PTSD screening makes the same point. The VA's PC-PTSD-5 is explicitly a **screen**; positive results should be followed by further assessment, preferably a structured interview. In a recent veteran sample, the PC-PTSD-5 showed **LR+ about 6.1** and **LR- about 0.07** overall, reminding you that some instruments are much better at moving probability down after a negative result than they are at confirming diagnosis by themselves. [\[5\]](#cite-5 "Reference [5]") Fagan nomogram: Bayes without the arithmetic -------------------------------------------- The **Fagan nomogram** is simply Bayes' theorem drawn as three vertical scales: pretest probability on the left, likelihood ratio in the middle, post-test probability on the right. Draw a straight line through the first two, and you read the third. That is the whole concept. It matters because it converts abstract statistics into a bedside decision tool when you need to know whether a screen result crosses a management threshold. [\[1\]](#cite-1 "Reference [1]") > **Clinical Pearl:** Never ask whether a questionnaire is good. Ask what it does to probability **in this patient, in this setting, at this cut-point**. If you do not want to juggle odds in your head, use the nomogram. That is not laziness; clinicians commonly misestimate how much testing should change probability. The nomogram keeps you honest. [\[1\]](#cite-1 "Reference [1]") Clinical Correlations --------------------- Bayesian reasoning matters in psychiatry because our symptoms are promiscuous. Insomnia, concentration problems, irritability, psychomotor change, avoidance, and agitation appear across mood, trauma-related, substance-induced, medical, and sleep disorders. Don't let a positive instrument outrank the history. Use the scale to update the probability you already earned from good interviewing, collateral, and exclusion of mimics. That is how you avoid one of the biggest diagnostic errors in psychiatry: over-calling a questionnaire and under-reading context. [\[1\]](#cite-1 "Reference [1]") Key Takeaways ------------- - Start with a **pretest probability**, not with the score. [\[1\]](#cite-1 "Reference [1]") - Use **LR+** for positive results and **LR-** for negative results. [\[2\]](#cite-2 "Reference [2]") - Think of **LR+ >10** and **LR- <0.1** as strong shifts; values near 1 are basically noise. [\[3\]](#cite-3 "Reference [3]") - A screening tool changes probability; it usually does **not** make the diagnosis alone. [\[5\]](#cite-5 "Reference [5]") - The **Fagan nomogram** is the exam-friendly shortcut from pretest probability and LR to post-test probability. [\[1\]](#cite-1 "Reference [1]") Conclusion ---------- Likelihood ratios rescue you from score worship. In psychiatry, where diagnostic testing is often symptom-based and context-sensitive, Bayesian reasoning is the disciplined move: estimate the prior, apply the LR, and decide whether the new probability is high enough to act, low enough to drop, or still uncertain enough to interview better. [\[1\]](#cite-1 "Reference [1]") Frequently Asked Questions ---------------------------- ### Why are likelihood ratios usually more useful than predictive values on exams? Because LR values are less dependent on prevalence than PPV and NPV, so they travel better across settings; you then combine the LR with the patient's pretest probability to get a post-test probability. [\[2\]](#cite-2 "Reference [2]") ### Can a negative screen rule out depression or PTSD by itself? Only if the **LR-** is low enough and the pretest probability is not already high. A negative result lowers probability; it does not erase strong clinical concern. [\[2\]](#cite-2 "Reference [2]") ### Do I need to calculate odds every time in clinic? No. The Fagan nomogram is designed to bypass mental odds math by linking pretest probability, likelihood ratio, and post-test probability graphically. [\[1\]](#cite-1 "Reference [1]") ### Can I multiply likelihood ratios from two tests? Only if the tests are reasonably independent. If they overlap heavily, multiplying LRs can overstate certainty. [\[6\]](#cite-6 "Reference [6]") References (11) ------------------- 1. 1. [ www.cambridge.org/core/journals/advances-in-psychiatric-treatment/article/clinicians-guide-to-evaluating-diagnostic-and-screening-tests-in-psychiatry/AC98F8BBECBDB2171868C5AFA6B7FBAB ](https://www.cambridge.org/core/journals/advances-in-psychiatric-treatment/article/clinicians-guide-to-evaluating-diagnostic-and-screening-tests-in-psychiatry/AC98F8BBECBDB2171868C5AFA6B7FBAB) [↩](#cite-ref-1-1 "Back to text") 2. 2. [ www.cebm.ox.ac.uk/resources/ebm-tools/likelihood-ratios ](https://www.cebm.ox.ac.uk/resources/ebm-tools/likelihood-ratios) [↩](#cite-ref-2-1 "Back to text") 3. 3. [ effectivehealthcare.ahrq.gov/products/test-performance-metrics/appendixes ](https://effectivehealthcare.ahrq.gov/products/test-performance-metrics/appendixes) [↩](#cite-ref-3-1 "Back to text") 4. 4. [ pmc.ncbi.nlm.nih.gov/articles/PMC3281183 ](https://pmc.ncbi.nlm.nih.gov/articles/PMC3281183/) [↩](#cite-ref-4-1 "Back to text") 5. 5. [ www.ptsd.va.gov/professional/assessment/screens/pc-ptsd.asp ](https://www.ptsd.va.gov/professional/assessment/screens/pc-ptsd.asp) [↩](#cite-ref-5-1 "Back to text") 6. 6. [ pmc.ncbi.nlm.nih.gov/articles/PMC6074451 ](https://pmc.ncbi.nlm.nih.gov/articles/PMC6074451/) [↩](#cite-ref-6-1 "Back to text") 7. 7. Fagan TJ. Letter: Nomogram for Bayes's theorem. N Engl J Med. 1975;293(5):257. 8. 8. Jaeschke R, Guyatt G, Sackett DL, et al. Users' Guides to the Medical Literature: How to Use an Article About a Diagnostic Test. JAMA. 1994. 9. 9. Warner J. Clinicians' guide to evaluating diagnostic and screening tests in psychiatry. Advances in Psychiatric Treatment. 2004;10(6):446-454. 10. 10. Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CMAJ. 2011. 11. 11. Bovin MJ, Kimerling R, Weathers FW, et al. Diagnostic accuracy and acceptability of the Primary Care PTSD Screen for DSM-5 among US Veterans. JAMA Netw Open. 2021;4(2):e2036733. Next Build momentum in Psychiatry with a focused study pathway ----------------------------------------------------------- - Exam‑style questions for Psychiatry - Smart review to target weak topics - Progress tracking that keeps you accountable [ Start your free trial ](https://mdster.com/user/dashboard) [ Psychiatry ](https://mdster.com/speciality/psychiatry) [ View pricing ](https://mdster.com/pricing) [ Explore features ](https://mdster.com/features) No credit card required. Full access to all features. No commitment. Cancel anytime. Explore topics: [ # Board Review ](https://mdster.com/blog?tag=board-review) [ # Medical Education ](https://mdster.com/blog?tag=medical-education) [ # Psychiatry ](https://mdster.com/blog?tag=psychiatry) [ # Evidence-Based Practice ](https://mdster.com/blog?tag=evidence-based-practice) [ # Diagnostic Test Evaluation ](https://mdster.com/blog?tag=diagnostic-test-evaluation) [ Back to all posts ](https://mdster.com/blog) Discussion () ----------------- Join the discussion [ Log in ](https://mdster.com/auth/login) or [ Sign up ](https://mdster.com/auth/register) No comments yet Be the first to share your thoughts! ![]() More in Medical Education ------------------------- [ See all ](https://mdster.com/blog?category=medical-education) [### Cervical Ripening Mechanisms: Inflammation, Collagen, Induction 6 min read Apr 14, 2026 ](https://mdster.com/blog/cervical-ripening-mechanisms-inflammation-collagen-induction) [### Depression Treatment Planning in Family Medicine: Aim for Remission 7 min read Apr 13, 2026 ](https://mdster.com/blog/depression-treatment-planning-in-family-medicine-aim-for-remission) [### Malignant Pleural Effusion: Diagnosis, Prognosis, and Management 7 min read Apr 12, 2026 ](https://mdster.com/blog/malignant-pleural-effusion-diagnosis-prognosis-and-management) Related Posts ------------- [ ![Severe Hyponatremia Case Discussion: SIADH, Seizures, and SCLC](https://mdster.com/storage/blog/images/severe-hyponatremia-case-discussion-siadh-seizures-and-sclc.jpg) Case Discussion ### Severe Hyponatremia Case Discussion: SIADH, Seizures, and SCLC A 58-year-old smoker presents with seizure, sodium 108 mmol/L, and a hilar mass. This case reviews SIADH diagnosis, emergency hypertonic saline, ODS prevention, and next steps. 6 min read 0 comments ](https://mdster.com/blog/severe-hyponatremia-case-discussion-siadh-seizures-and-sclc) [ ![Cervical Ripening Mechanisms: Inflammation, Collagen, Induction](https://mdster.com/storage/blog/images/cervical-ripening-mechanisms-inflammation-collagen-induction.jpg) Medical Education ### Cervical Ripening Mechanisms: Inflammation, Collagen, Induction High-yield OBGYN review of cervical ripening mechanisms: how collagen remodeling, sterile inflammation, and prostaglandins prepare the cervix and shape induction choices. 6 min read 0 comments ](https://mdster.com/blog/cervical-ripening-mechanisms-inflammation-collagen-induction) [ ![Tourniquet Pain Under Spinal Anesthesia: A TKA Case Discussion](https://mdster.com/storage/blog/images/tourniquet-pain-under-spinal-anesthesia-a-tka-case-discussion.jpg) Case Discussion ### Tourniquet Pain Under Spinal Anesthesia: A TKA Case Discussion A board-focused anesthesiology case discussion on tourniquet pain during spinal anesthesia for total knee arthroplasty, with differential diagnosis, rescue strategies, and conversion planning. 6 min read 0 comments ](https://mdster.com/blog/tourniquet-pain-under-spinal-anesthesia-a-tka-case-discussion) [ ![Depression Treatment Planning in Family Medicine: Aim for Remission](https://mdster.com/storage/blog/images/depression-treatment-planning-in-family-medicine-aim-for-remission.jpg) Medical Education ### Depression Treatment Planning in Family Medicine: Aim for Remission Board-focused guide to depression treatment planning in Family Medicine, covering PHQ-9 targets, psychotherapy referral, SSRI/SNRI basics, follow-up timing, and relapse prevention in perinatal and late-life care. 7 min read 0 comments ](https://mdster.com/blog/depression-treatment-planning-in-family-medicine-aim-for-remission) [ ![Malignant Pleural Effusion: Diagnosis, Prognosis, and Management](https://mdster.com/storage/blog/images/malignant-pleural-effusion-diagnosis-prognosis-and-management.jpg) Medical Education ### Malignant Pleural Effusion: Diagnosis, Prognosis, and Management Malignant pleural effusion is more than recurrent dyspnea. Learn how to interpret cytology yield, choose IPC versus pleurodesis, and use prognosis to guide bedside care. 7 min read 0 comments ](https://mdster.com/blog/malignant-pleural-effusion-diagnosis-prognosis-and-management) [ ![Autism Spectrum Disorder vs Social Anxiety in a 9-Year-Old](https://mdster.com/storage/blog/images/autism-spectrum-disorder-vs-social-anxiety-in-a-9-year-old.jpg) Case Discussion ### Autism Spectrum Disorder vs Social Anxiety in a 9-Year-Old A board-style child psychiatry case on distinguishing autism spectrum disorder from social anxiety, interpreting meltdowns, and planning neurodiversity-affirming care. 6 min read 0 comments ](https://mdster.com/blog/autism-spectrum-disorder-vs-social-anxiety-in-a-9-year-old) [ MDster home ](/ "MDster home") Master your medical exams with evidence-based learning. [ GET IT ON Google Play ](https://play.google.com/store/apps/details?id=com.mdster.app) Platform - [Home](https://mdster.com) - [Features](https://mdster.com/features) - [Pricing](https://mdster.com/pricing) - [About](https://mdster.com/about) Resources - [Blog](https://mdster.com/blog) - [Dashboard](https://mdster.com/user/dashboard) Support - [Contact](https://mdster.com/contact) - [Legal & Policies](https://mdster.com/legal) - [Medical Reviewers](https://mdster.com/medical-reviewers) © 2026 MDster [ ](https://play.google.com/store/apps/details?id=com.mdster.app) [Terms](https://mdster.com/terms) [Privacy](https://mdster.com/privacy) [Editorial](https://mdster.com/editorial-policy) reCAPTCHA Protected by reCAPTCHA. Google [Privacy Policy](https://policies.google.com/privacy) and [Terms of Service](https://policies.google.com/terms) apply. Cookie Consent -------------- We use cookies to enhance your experience. By continuing to visit this site you agree to our use of cookies. [ Terms of Use ](https://mdster.com/terms) & [ Privacy Policy ](https://mdster.com/privacy) Accept