markparamlall

21/ What next. • 21 Oct: consultation closes • 16 Feb 2027: final guideline expected • Later batches: no date Preprint (not peer reviewed, yet): osf.io/preprints/psyarxiv/tw… Data and calculations: doi.org/10.17605/OSF.IO/D5EW… Analysis done with @markparamlall over 18 days between 17 Sep 2026 - 05 Oct 2026. Spot an error? Every number has a page reference, so tell us which one. @NICEComms #OCD #BDD #MentalHealth

Who with OCD gets referred? A 10-point deprivation gap in English primary care records osf.io/preprints/psyarxiv/9g… #OCD

9/ Replace, or partially update? • The draft itself says it "will replace" the 2005 guideline (page 101) • NICE's project page says it "will partially update" it • The final scope allowed old recommendations to be kept. The draft keeps none. • Yet the draft still leans on the 2005 guideline, for example in how it identifies OCD. What goes with it: the 2005 advice on poor response and inpatient care. Batch 1 kept 4 of 8 questions. Treatment after first-line care fails was deferred, with no date. doi.org/10.17605/OSF.IO/D5EW… #OCD #BDD #HealthPolicy

8/ The draft relies on terms it never defines. ❌ Matched care matches treatment to "symptom severity". No severity measure is named, and the Y-BOCS appears nowhere in the recommendations. ❌ Who the draft doesn't cover depends on an "adequate trial". Two different 12-week clocks, never reconciled. ❌ No response threshold. The consensus is at least a 35% fall in Y-BOCS. ❌ "Maximum tolerable dose", with no dose targets. NHS England's own national service uses the Y-BOCS (above 30) to decide entry. doi.org/10.17605/OSF.IO/D5EW… #OCD #Psychopharmacology #Psychiatry

7/ For BDD, NICE's own analysis and NICE's own draft disagree. Rank in the analysis, then place in the draft: • Transdiagnostic CBT: 1st, left out (one trial) • Group CBT: 2nd, placed 3rd • MCT: 3rd, left out • Self-help CBT: 4th, placed 2nd • Individual CBT: 5th, placed 1st Group CBT went third because talking about appearance in a group can feel daunting. That's a fair clinical point. But self-help CBT, moved up for being easier to face, had the highest drop-out in NICE's BDD analysis, credibly higher than waitlist. Acceptability moved one option down. The acceptability data argue against the one moved up. doi.org/10.17605/OSF.IO/D5EW… #BDD #BodyDysmorphicDisorder #CBT

6/ The draft says offer options "in the order listed". NICE's own analysis can't separate the top options. • Chance the top option is truly the most cost-effective: 17% • Average rank of 15: individual CBT 2.9, self-help CBT 3.5, CT 3.7, MCT 3.9 • Plausible ranks run from 1st to 8th, or 1st to 10th for CT NICE's own check called the order of options 2 and 3 "sensitive to small and plausible" changes in the evidence. A firm order, from an analysis that can't tell the options apart. doi.org/10.17605/OSF.IO/D5EW… #OCD #CBT #HealthEconomics

5/ NICE's own cost model ranks MCT 3rd of 15. NICE left it out. Net health benefit per 1,000 adults: • Individual CBT: 3,270 • Individual CT: 3,255 • MCT: 3,252 (left out) • Self-help CBT: 3,247 (recommended) • Group CBT: 3,209 (recommended) • EMDR: 3,173 (left out) • SSRIs: 3,028 (recommended) Four recommended options score below a therapy that isn't recommended. doi.org/10.17605/OSF.IO/D5EW… #OCD #HealthEconomics #Psychiatry

4/ Same reasons, different decisions. NICE left out two therapies, citing small samples and wide uncertainty: • MCT: 52 people, 2 trials • EMDR: 59 people, 2 trials Then it recommended one on comparable evidence: • Self-help CBT with therapist support: 61 people, 3 trials • Wider uncertainty than EMDR's (1.39 vs 1.22) The reason given for leaving two out applies just as well to the one put in. doi.org/10.17605/OSF.IO/D5EW… #OCD #CBT #EMDR

3/ Credit where due: NICE pulled together a lot of evidence. Trials in NICE's main analysis: • Adults with OCD: 63 of 82 trials, 6,014 people • Children and young people: 24 of 28 trials, 1,626 people • BDD: 10 of 13 studies, 638 people The weak spots are at the edges. • Adult trials: 72 of 82 were rated at high risk of bias. None was excluded for it. Most adult comparisons are low or very low certainty. • BDD evidence: about a tenth of the adult evidence. Every comparison between two active treatments is indirect, made through waitlist. So those estimates cannot be checked against head-to-head evidence. • BDD costs: NICE's model had no quality-of-life data from people with BDD. It borrowed values from adult OCD and from social anxiety disorder. doi.org/10.17605/OSF.IO/D5EW… #OCD #BDD #EvidenceBasedMedicine

2/ How we did it: Six questions, shaped by the AGREE II domains. Every value in NICE's 4 evidence reviews, 3 network meta-analyses and 3 economic models was transcribed, then checked by script against the text of its source PDF. Every quotation matched verbatim, every DOI checked against Crossref. No new meta-analysis, no re-weighting, no new modelling. As an extraction check we recomputed net health benefit from NICE's published QALYs and costs: all values agreed with NICE's to within 1 QALY per 1,000 people. Comparison guidelines came from a documented PubMed search (30 Sep 2026): 46 records screened, 6 guidelines included, plus NICE's HealthTech guidance and two UK documents. All 46 screening decisions, every extracted value with its source page, and SHA-256 checksums of the source files are on OSF: doi.org/10.17605/OSF.IO/D5EW… #OCD #BDD #OpenScience

It's OCD Awareness Week. Each day this week our group will release one preprint from our OCD work: free to read, data on OSF, open to challenge. Day 1, today: our appraisal of NICE's draft OCD & BDD guideline, read against NICE's own numbers. Analysis done with @markparamlall over 18 days (17/9-5/10) - doi.org/10.17605/OSF.IO/D5EW… Thread below. #OCDweek #OCD

NICE's 2005 OCD/BDD guideline is being updated! The draft is out for consultation until 21 Oct 2026, with publication expected Feb 2027. It replaces stepped care with matched care, a welcome step. But further-line treatment is deferred to a possible future batch, with no date. In the 18 days after #NICE published the draft, @markparamlall and I appraised it against NICE's own published data: a critical appraisal with secondary analysis. Preprint (not peer reviewed) below, highlights in this thread: osf.io/preprints/psyarxiv/tw… #OCD #BDD @Nick_Sireau @LeighWallbank

Wonderful to see a psychiatrist win the Nobel Prize again after 26 years & becoming only the third psychiatrist ever to do so, after Julius Wagner-Jauregg in 1927 (malaria therapy for neurosyphilis) and Eric Kandel in 2000 (signal transduction and memory). Congratulations @KarlDeisseroth on the 2026 Nobel Prize in Physiology or Medicine for #Optogenetics - work that transformed how we study brain circuits, including the circuits underlying #OCD.