PK 1]_rels/PK 1] docProps/PK 1]ppt/PK 1] ppt/_rels/PK 1] ppt/charts/PK 1]ppt/charts/_rels/PK 1]ppt/embeddings/PK 1] ppt/media/PK 1]ppt/slideLayouts/PK 1]ppt/slideLayouts/_rels/PK 1]ppt/slideMasters/PK 1]ppt/slideMasters/_rels/PK 1] ppt/slides/PK 1]ppt/slides/_rels/PK 1] ppt/theme/PK 1]ppt/notesMasters/PK 1]ppt/notesMasters/_rels/PK 1]ppt/notesSlides/PK 1]ppt/notesSlides/_rels/PK 1]vgy,y,[Content_Types].xml PK 1]]] _rels/.rels PK 1]NMwdocProps/app.xml 0 0 Microsoft Office PowerPoint On-screen Show (16:9) 0 22 22 0 0 false Fonts Used 2 Theme 1 Slide Titles 22 Arial Calibri Office Theme Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22 PptxGenJS false false false 16.0000 PK 1]KdocProps/core.xml PptxGenJS Presentation PptxGenJS Presentation PptxGenJS PptxGenJS 1 2026-08-23T06:08:36Z 2026-08-23T06:08:36Z PK 1] 3>ppppt/_rels/presentation.xml.rels PK 1]Oݨ ppt/theme/theme1.xmlPK 1]:[[ppt/presentation.xml PK 1]Xppt/presProps.xml PK 1]ppt/tableStyles.xml PK 1]D >00ppt/viewProps.xml PK 1]H7t!ppt/slideLayouts/slideLayout1.xml PK 1]ђ77,ppt/slideLayouts/_rels/slideLayout1.xml.rels PK 1]>pppt/slides/slide1.xml AI IN DAILY MEDICAL PRACTICEModule 3AI-Supported Clinical Decision MakingSelf-paced + live sessions · ~3 hrs 50 min totalAI in Daily Medical Practice — a curriculum for practicing physiciansPK 1]3 ppt/slides/_rels/slide1.xml.rels PK 1].ppt/notesSlides/notesSlide1.xml 1PK 1]:A*ppt/notesSlides/_rels/notesSlide1.xml.rels PK 1]& GGppt/slides/slide2.xml BEFORE YOU STARTHow this module works1Preview Quiz5 min, ungraded. Orients you before reading.→2Handout / SlidesSection D's case exercise, plus scripts for the roleplay — work them solo first.→3Case Workshop + RoleplayLive session with your cohort — group case work and patient-communication practice.→4Post-Test12 questions, scored. 80% pass, retake anytime.2PK 1]2- ppt/slides/_rels/slide2.xml.rels PK 1]ppt/notesSlides/notesSlide2.xml 2PK 1]xշ*ppt/notesSlides/_rels/notesSlide2.xml.rels PK 1]n)n)ppt/slides/slide3.xml BY THE END OF THIS MODULEYou will be able to1Integrate AI-generated differentials, imaging reads, or risk scores as one input among several in your clinical reasoning — never as a final answer.2Recognize when a tool's validation population differs from your own patient population, and adjust how much you trust its output accordingly.3Communicate AI-assisted findings transparently to patients and colleagues.3PK 1]W/ ppt/slides/_rels/slide3.xml.rels PK 1]K |Őppt/notesSlides/notesSlide3.xml 3PK 1]9 Y*ppt/notesSlides/_rels/notesSlide3.xml.rels PK 1]?I]]ppt/slides/slide4.xml ASECTION AAI as a Second Opinion,Not a VerdictKNOW4PK 1]` ppt/slides/_rels/slide4.xml.rels PK 1]vsppt/notesSlides/notesSlide4.xml 4PK 1]J *ppt/notesSlides/_rels/notesSlide4.xml.rels PK 1] ̀**ppt/slides/slide5.xml SECTION A · THE CORE IDEAOne voice in the room, not the final word5“An AI decision-support tool is one voice in the room, not the person making the decision.”Form your own impression firstOr in parallel — never only after seeing the AI's answer.Disagreement is normal, not a malfunctionA second opinion you're allowed to override is doing its job even when you override it.The chart reflects your judgmentNoted as one input you weighed — not “the AI said X, so we did X.”PK 1]5 ppt/slides/_rels/slide5.xml.rels PK 1]W8ppt/notesSlides/notesSlide5.xml 5PK 1]Qe*ppt/notesSlides/_rels/notesSlide5.xml.rels PK 1]cxxppt/slides/slide6.xml BSECTION BDifferential-Diagnosis Support Tools —Capabilities and LimitsKNOW6PK 1]ج+ ppt/slides/_rels/slide6.xml.rels PK 1]zppt/notesSlides/notesSlide6.xml 6PK 1]=|*ppt/notesSlides/_rels/notesSlide6.xml.rels PK 1]XS**ppt/slides/slide7.xml SECTION B · WHAT THESE TOOLS DOUseful for widening a differential — not for replacing one7GENUINELY USEFUL FOR• Widening a differential you've already started• A fast second pass on an unfamiliar or atypical presentation• Structuring your own reasoning into an explicit, ranked listNOT BUILT TO DO• Replacing history-taking and physical examination• Accounting for findings never entered into it• Knowing what's common in your setting, unless trained on itPK 1]F ppt/slides/_rels/slide7.xml.rels PK 1])lppt/notesSlides/notesSlide7.xml 7PK 1]|g*ppt/notesSlides/_rels/notesSlide7.xml.rels PK 1]_<$<$ppt/slides/slide8.xml SECTION B · TWO LIMITS THAT FAIL SILENTLYNothing signals that these happened8Garbage in,garbage out,quietlyA vague or incomplete symptom description still produces a confidently ranked differential — the tool doesn't flag that the input was insufficient.Ranking reflectstraining prevalence,not yoursA condition common in the training data ranks higher than an equally plausible, locally-common condition it saw less of.PK 1]6 ppt/slides/_rels/slide8.xml.rels PK 1]iސppt/notesSlides/notesSlide8.xml 8PK 1]pO*ppt/notesSlides/_rels/notesSlide8.xml.rels PK 1]K>ttppt/slides/slide9.xml CSECTION CMultimodal Interpretation:Imaging, Labs, ECG-Assisted ReadsKNOW9PK 1]>$ ppt/slides/_rels/slide9.xml.rels PK 1]qppt/notesSlides/notesSlide9.xml 9PK 1]1*ppt/notesSlides/_rels/notesSlide9.xml.rels PK 1]isf,,ppt/slides/slide10.xml SECTION C · THREE FORMS THIS TAKESWorking across more than one kind of input10Imaging triage /computer-aideddetectionFlags regions on a chest X-ray for closer review — WHO-endorsed for TB screening, directly relevant here. Flags and prioritizes; doesn't issue a final report.ECG-assistedinterpretationSuggests a reading or flags an abnormality — most mature for arrhythmia detection. A second read, not a replacement for correlating with the patient.Lab patternrecognitionFlags a suggestive combination across a panel, not just one out-of-range value — still built on whatever combinations its training data emphasized.PK 1]Ѳ!ppt/slides/_rels/slide10.xml.rels PK 1]O ppt/notesSlides/notesSlide10.xml 10PK 1]T+ppt/notesSlides/_rels/notesSlide10.xml.rels PK 1]J)J)ppt/slides/slide11.xml SECTION C · TWO FAILURE MODES TO NAMESensitivity has a cost11Silent falsenegativeNothing flagged gets read as “nothing's there” — it may just mean the finding fell outside what the tool was trained to detect.Confident falsepositiveA specific-looking confidence score gets mistaken for diagnostic certainty rather than a probability worth checking.PRACTICAL HABITRead a flag the way you'd read a colleague's phone consult: useful, worth taking seriously, still something you check yourself.PK 1];!ppt/slides/_rels/slide11.xml.rels PK 1]s6ӑ ppt/notesSlides/notesSlide11.xml 11PK 1]O+ppt/notesSlides/_rels/notesSlide11.xml.rels PK 1]Mhllppt/slides/slide12.xml DSECTION DCase Workshop: Working a CaseWith and Without AITHINK12PK 1]c!ppt/slides/_rels/slide12.xml.rels PK 1]*)@ ppt/notesSlides/notesSlide12.xml 12PK 1]Fb+ppt/notesSlides/_rels/notesSlide12.xml.rels PK 1] 7Ėppt/slides/slide13.xml SECTION D · STEP 1 — READ THIS FIRSTForm your own differential before reading on13CASEA 34-year-old man presents to your clinic in Dili in early rainy season with four days of fever, frontal headache, diffuse myalgia, and mild retro-orbital pain. He reports wading through floodwater near his home three days before symptoms began. No rash yet. No cough, no dysuria, no diarrhea.Exam: temperature 38.7°C, mildly injected conjunctivae, no rash, no hepatosplenomegaly, no calf tenderness. A rapid malaria test is negative. A full blood count and platelet count are pending.Before turning the page — what's your own differential, and what would you order or watch for next?PK 1]x!ppt/slides/_rels/slide13.xml.rels PK 1]Ї ppt/notesSlides/notesSlide13.xml 13PK 1]yv+ppt/notesSlides/_rels/notesSlide13.xml.rels PK 1]))ppt/slides/slide14.xml SECTION D · STEP 2 — THE AI'S DIFFERENTIALNow compare it to your own14GENERAL-PURPOSE AI TOOL RETURNS:1. Viral syndrome / influenza-like illness — most likely, self-limiting2. Infectious mononucleosis3. Early COVID-194. Dengue fever — possible, recommend monitoring5. Bacterial pharyngitisLeptospirosis does not appear on the list.Now compare that ranking to your own. What's missing, and why might it be missing?PK 1]O!ppt/slides/_rels/slide14.xml.rels PK 1]0 ppt/notesSlides/notesSlide14.xml 14PK 1]?ݤ+ppt/notesSlides/_rels/notesSlide14.xml.rels PK 1]ʾf22ppt/slides/slide15.xml SECTION D · DISCUSSIONWhat this case is built to show151The AI's top-ranked answer reflects general global prevalence, not Dili in rainy season.2Dengue is under-weighted and leptospirosis is missing — despite floodwater exposure being a textbook clue.3A negative malaria test narrowed one possibility; it proved nothing about the others.4Keep dengue and leptospirosis near the top of your own differential regardless of the AI's ranking, based on exposure history and local epidemiology the tool never saw.PK 1]*R!ppt/slides/_rels/slide15.xml.rels PK 1]Qz ppt/notesSlides/notesSlide15.xml 15PK 1]$Q+ppt/notesSlides/_rels/notesSlide15.xml.rels PK 1]*?||ppt/slides/slide16.xml ESECTION EValidation Gaps: When the ToolWasn't Built for Your PatientsFIND16PK 1] !ppt/slides/_rels/slide16.xml.rels PK 1]  ppt/notesSlides/notesSlide16.xml 16PK 1] y+ppt/notesSlides/_rels/notesSlide16.xml.rels PK 1] @.@.ppt/slides/slide17.xml SECTION E · WHY THIS HAPPENSTrust has to be calibrated to fit17Disease prevalence mismatchConditions common where the training data came from outrank conditions common where you practice.Demographic and phenotypic mismatchImaging models can underperform across skin tones, body habitus, or nutrition patterns underrepresented in training data.Baseline and comorbidity mismatchRisk scores calibrated on a different baseline population don't automatically transfer their accuracy here.Infrastructure assumptionsA tool can assume a prior scan or lab reference range that simply doesn't exist for a given patient here.PK 1]檋!ppt/slides/_rels/slide17.xml.rels PK 1]r. ppt/notesSlides/notesSlide17.xml 17PK 1]BI+ppt/notesSlides/_rels/notesSlide17.xml.rels PK 1]ciippt/slides/slide18.xml FSECTION FTalking to Patients AboutAI-Assisted FindingsTHINK18PK 1]Yh!ppt/slides/_rels/slide18.xml.rels PK 1] ppt/notesSlides/notesSlide18.xml 18PK 1]N+ppt/notesSlides/_rels/notesSlide18.xml.rels PK 1]d"++ppt/slides/slide19.xml SECTION F · DISCLOSURE AND DISAGREEMENTFraming matters as much as accuracy19A screening tool flagged something on imagingSay: “A screening tool flagged something on your X-ray I want to look at more closely with you. It's a first pass, not a diagnosis.”Avoid: “The computer says you have TB.”When you disagree with the AI's outputSay: “A tool suggested [X], but based on your history and exam, I don't think that fits — here's what I think is going on.”Avoid: Quietly overriding the tool without mentioning it existed.PK 1]~!ppt/slides/_rels/slide19.xml.rels PK 1]0[ ppt/notesSlides/notesSlide19.xml 19PK 1]i+ppt/notesSlides/_rels/notesSlide19.xml.rels PK 1]I5ppt/slides/slide20.xml SECTION F · A PATIENT WORRIED ABOUT “A MACHINE”You're still the one deciding20Reassuring without dismissing the concernSay: “I'm the one making the decision about your care. This tool is one more piece of information — like a colleague's opinion or a lab result. It doesn't replace my judgment or my exam of you.”ROLEPLAY REMINDERIn the live roleplay session, you'll practice your own version of one of these scripts, in the language(s) you actually use with patients, with a colleague playing the patient.PK 1]g_!ppt/slides/_rels/slide20.xml.rels PK 1]7 ppt/notesSlides/notesSlide20.xml 20PK 1]SB+ppt/notesSlides/_rels/notesSlide20.xml.rels PK 1]iwk*k*ppt/slides/slide21.xml BEFORE YOU MOVE ONModule 3 summaryKNOWWhy AI decision support is a second opinion, not a verdict; what differential-support and multimodal tools do and don't do; the failure modes specific to this category, including silent false negatives.THINKA trained habit of forming your own impression first, practiced on a locally-grounded case, plus disclosure language for patients and colleagues.FINDHow to calibrate trust in a tool by checking its validation population, and which locally-important diagnoses to keep in mind regardless of its ranking.21PK 1]z!ppt/slides/_rels/slide21.xml.rels PK 1]89| ppt/notesSlides/notesSlide21.xml 21PK 1]HΩ+ppt/notesSlides/_rels/notesSlide21.xml.rels PK 1]%ppt/slides/slide22.xml NEXT STEPSTake the Post-Test12 scored questions, 80% pass mark. After the case workshop and roleplay session, and once Module 3 is evaluated, Module 4 — Critical Appraisal, Bias & Clinical Autonomy — opens next.Module 3 of 6 — AI-Supported Clinical Decision Making22PK 1]dV!ppt/slides/_rels/slide22.xml.rels PK 1]cQc ppt/notesSlides/notesSlide22.xml 22PK 1]aeZ+ppt/notesSlides/_rels/notesSlide22.xml.rels PK 1]K !ppt/slideMasters/slideMaster1.xml PK 1]N),ppt/slideMasters/_rels/slideMaster1.xml.rels PK 1]6TT!ppt/notesMasters/notesMaster1.xml 7/23/19Click to edit Master text stylesSecond levelThird levelFourth levelFifth level‹#›PK 1]s **,ppt/notesMasters/_rels/notesMaster1.xml.rels PK 1]_rels/PK 1] $docProps/PK 1]Kppt/PK 1] mppt/_rels/PK 1] ppt/charts/PK 1]ppt/charts/_rels/PK 1]ppt/embeddings/PK 1] ppt/media/PK 1]Bppt/slideLayouts/PK 1]qppt/slideLayouts/_rels/PK 1]ppt/slideMasters/PK 1]ppt/slideMasters/_rels/PK 1]  ppt/slides/PK 1]3ppt/slides/_rels/PK 1] bppt/theme/PK 1]ppt/notesMasters/PK 1]ppt/notesMasters/_rels/PK 1]ppt/notesSlides/PK 1]ppt/notesSlides/_rels/PK 1]vgy,y,P[Content_Types].xmlPK 1]]] /_rels/.relsPK 1]NMw2docProps/app.xmlPK 1]K\:docProps/core.xmlPK 1] 3>pp{=ppt/_rels/presentation.xml.relsPK 1]Oݨ (Mppt/theme/theme1.xmlPK 1]:[['nppt/presentation.xmlPK 1]X}ppt/presProps.xmlPK 1]ppt/tableStyles.xmlPK 1]D >00ppt/viewProps.xmlPK 1]H7t!Hppt/slideLayouts/slideLayout1.xmlPK 1]ђ77,6ppt/slideLayouts/_rels/slideLayout1.xml.relsPK 1]>pppt/slides/slide1.xmlPK 1]3 şppt/slides/_rels/slide1.xml.relsPK 1].ѡppt/notesSlides/notesSlide1.xmlPK 1]:A*ppt/notesSlides/_rels/notesSlide1.xml.relsPK 1]& GGppt/slides/slide2.xmlPK 1]2- ppt/slides/_rels/slide2.xml.relsPK 1]ppt/notesSlides/notesSlide2.xmlPK 1]xշ*ppt/notesSlides/_rels/notesSlide2.xml.relsPK 1]n)n)ppt/slides/slide3.xmlPK 1]W/ s&ppt/slides/_rels/slide3.xml.relsPK 1]K |Ő(ppt/notesSlides/notesSlide3.xmlPK 1]9 Y*L/ppt/notesSlides/_rels/notesSlide3.xml.relsPK 1]?I]]_1ppt/slides/slide4.xmlPK 1]` Hppt/slides/_rels/slide4.xml.relsPK 1]vsJppt/notesSlides/notesSlide4.xmlPK 1]J *Qppt/notesSlides/_rels/notesSlide4.xml.relsPK 1] ̀**Sppt/slides/slide5.xmlPK 1]5 ~ppt/slides/_rels/slide5.xml.relsPK 1]W8ppt/notesSlides/notesSlide5.xmlPK 1]Qe*gppt/notesSlides/_rels/notesSlide5.xml.relsPK 1]cxxzppt/slides/slide6.xmlPK 1]ج+ %ppt/slides/_rels/slide6.xml.relsPK 1]z1ppt/notesSlides/notesSlide6.xmlPK 1]=|*ppt/notesSlides/_rels/notesSlide6.xml.relsPK 1]XS**ppt/slides/slide7.xmlPK 1]F (ppt/slides/_rels/slide7.xml.relsPK 1])l4ppt/notesSlides/notesSlide7.xmlPK 1]|g*ppt/notesSlides/_rels/notesSlide7.xml.relsPK 1]_<$<$ppt/slides/slide8.xmlPK 1]6 ppt/slides/_rels/slide8.xml.relsPK 1]iސppt/notesSlides/notesSlide8.xmlPK 1]pO*\ppt/notesSlides/_rels/notesSlide8.xml.relsPK 1]K>ttoppt/slides/slide9.xmlPK 1]>$ )ppt/slides/_rels/slide9.xml.relsPK 1]q"+ppt/notesSlides/notesSlide9.xmlPK 1]1*1ppt/notesSlides/_rels/notesSlide9.xml.relsPK 1]isf,,4ppt/slides/slide10.xmlPK 1]Ѳ!appt/slides/_rels/slide10.xml.relsPK 1]O -cppt/notesSlides/notesSlide10.xmlPK 1]T+ippt/notesSlides/_rels/notesSlide10.xml.relsPK 1]J)J)lppt/slides/slide11.xmlPK 1];!ppt/slides/_rels/slide11.xml.relsPK 1]s6ӑ ppt/notesSlides/notesSlide11.xmlPK 1]O+lppt/notesSlides/_rels/notesSlide11.xml.relsPK 1]Mhllppt/slides/slide12.xmlPK 1]c!!ppt/slides/_rels/slide12.xml.relsPK 1]*)@ /ppt/notesSlides/notesSlide12.xmlPK 1]Fb+ppt/notesSlides/_rels/notesSlide12.xml.relsPK 1] 7Ėppt/slides/slide13.xmlPK 1]x!ppt/slides/_rels/slide13.xml.relsPK 1]Ї ppt/notesSlides/notesSlide13.xmlPK 1]yv+ppt/notesSlides/_rels/notesSlide13.xml.relsPK 1]))ppt/slides/slide14.xmlPK 1]O!ppt/slides/_rels/slide14.xml.relsPK 1]0 ppt/notesSlides/notesSlide14.xmlPK 1]?ݤ+ppt/notesSlides/_rels/notesSlide14.xml.relsPK 1]ʾf22ppt/slides/slide15.xmlPK 1]*R!Rppt/slides/_rels/slide15.xml.relsPK 1]Qz Tppt/notesSlides/notesSlide15.xmlPK 1]$Q+[ppt/notesSlides/_rels/notesSlide15.xml.relsPK 1]*?||]ppt/slides/slide16.xmlPK 1] !uppt/slides/_rels/slide16.xml.relsPK 1]  wppt/notesSlides/notesSlide16.xmlPK 1] y+p~ppt/notesSlides/_rels/notesSlide16.xml.relsPK 1] @.@.ppt/slides/slide17.xmlPK 1]檋!ppt/slides/_rels/slide17.xml.relsPK 1]r. ppt/notesSlides/notesSlide17.xmlPK 1]BI+ַppt/notesSlides/_rels/notesSlide17.xml.relsPK 1]ciippt/slides/slide18.xmlPK 1]Yh!ppt/slides/_rels/slide18.xml.relsPK 1] ppt/notesSlides/notesSlide18.xmlPK 1]N+eppt/notesSlides/_rels/notesSlide18.xml.relsPK 1]d"++zppt/slides/slide19.xmlPK 1]~!Jppt/slides/_rels/slide19.xml.relsPK 1]0[ X ppt/notesSlides/notesSlide19.xmlPK 1]i+'ppt/notesSlides/_rels/notesSlide19.xml.relsPK 1]I5<ppt/slides/slide20.xmlPK 1]g_!l3ppt/slides/_rels/slide20.xml.relsPK 1]7 z5ppt/notesSlides/notesSlide20.xmlPK 1]SB+I<ppt/notesSlides/_rels/notesSlide20.xml.relsPK 1]iwk*k*^>ppt/slides/slide21.xmlPK 1]z!hppt/slides/_rels/slide21.xml.relsPK 1]89| kppt/notesSlides/notesSlide21.xmlPK 1]HΩ+qppt/notesSlides/_rels/notesSlide21.xml.relsPK 1]%sppt/slides/slide22.xmlPK 1]dV! ppt/slides/_rels/slide22.xml.relsPK 1]cQc ppt/notesSlides/notesSlide22.xmlPK 1]aeZ+ppt/notesSlides/_rels/notesSlide22.xml.relsPK 1]K !ppt/slideMasters/slideMaster1.xmlPK 1]N),Ѵppt/slideMasters/_rels/slideMaster1.xml.relsPK 1]6TT!ٶppt/notesMasters/notesMaster1.xmlPK 1]s **,lppt/notesMasters/_rels/notesMaster1.xml.relsPK{{#