โ† DMDA ํ—ˆ๋ธŒDMDA Hub
๐Ÿ“– DMDA ยท ์‚ฌ์šฉ์ž ๋งค๋‰ด์–ผUser Manual
์‚ฌ์šฉ์ž ๋งค๋‰ด์–ผ ยท v1.0User Manual ยท v1.0

DMDA๋ฅผ 5๋ถ„ ์•ˆ์— ์ดํ•ดํ•˜๊ธฐ Understand DMDA in 5 Minutes

Progressive DSM-5-TR Multi-Source Diagnostic Algorithm Progressive DSM-5-TR Multi-Source Diagnostic Algorithm

1 DMDA๋ž€ ๋ฌด์—‡์ธ๊ฐ€?What is DMDA?

DMDA๋Š” DSM-5-TR Multi-Source Diagnostic Algorithm์˜ ์•ฝ์ž์ž…๋‹ˆ๋‹ค. ํ•œ ๋งˆ๋””๋กœ "์—ฌ๋Ÿฌ ๋ฐ์ดํ„ฐ ์†Œ์Šค๋ฅผ ํ†ตํ•ฉํ•ด์„œ DSM-5-TR ์ง„๋‹จ์„ ์‚ฐ์ถœํ•˜๋Š” ์•Œ๊ณ ๋ฆฌ์ฆ˜"์ด์—์š”. DMDA stands for DSM-5-TR Multi-Source Diagnostic Algorithm. In short: "an algorithm that integrates multiple data sources to produce a DSM-5-TR diagnosis."

๐Ÿ’ก ๊ธฐ์กด ํ‰๊ฐ€์™€์˜ ์ฐจ์ดHow It's Different

๊ธฐ์กด: ์„ค๋ฌธ์ง€ ํ•˜๋‚˜๋กœ๋งŒ ์ง„๋‹จ (์ฃผ๊ด€์  ์ž๊ธฐ๋ณด๊ณ ์— 100% ์˜์กด)
DMDA: ์„ค๋ฌธ + ํƒ€์ดํ•‘ ํŒจํ„ด + ํฐ ์ˆ˜๋™ ๋ฐ์ดํ„ฐ + qEEG + HRV + ์–ผ๊ตด + ์Œ์„ฑ + ERP๋ฅผ ๋™์‹œ์— ํ†ตํ•ฉ. ์ฃผ๊ด€๊ณผ ๊ฐ๊ด€์ด ๋งŒ๋‚˜์š”.
Traditional: Diagnosis from a single questionnaire (100% subjective self-report)
DMDA: Survey + typing + phone passive + qEEG + HRV + face + voice + ERP integrated simultaneously. Subjective meets objective.

ํ•ต์‹ฌ ๊ฐœ๋…: Progressive (์ ์ง„์ ) Key Concept: Progressive

"Progressive"๋ผ๋Š” ์ด๋ฆ„์€ ๋ฐ์ดํ„ฐ๊ฐ€ ์ ์  ๋Š˜์–ด๋‚จ์— ๋”ฐ๋ผ ์ง„๋‹จ ์‹ ๋ขฐ๋„๊ฐ€ ๋‹จ๊ณ„์ ์œผ๋กœ ์ƒ์Šนํ•œ๋‹ค๋Š” ๋œป์ด์—์š”. ์ฒ˜์Œ์—” ์„ค๋ฌธ๋งŒ (Tier 1, 60-75% ์‹ ๋ขฐ๋„) โ†’ qEEG ์ถ”๊ฐ€ (Tier 2, 75-88%) โ†’ ์ƒ๋ฆฌํ•™์  ์ง€ํ‘œ ์ถ”๊ฐ€ (Tier 3, 88-95%) โ†’ ์—ฐ๊ตฌ๊ธ‰ ๋ฐ์ดํ„ฐ (Tier 4, 95-98%). "Progressive" means diagnostic confidence increases stepwise as more data is added. Start with just survey (Tier 1, 60-75% confidence) โ†’ add qEEG (Tier 2, 75-88%) โ†’ add physiological (Tier 3, 88-95%) โ†’ research-grade (Tier 4, 95-98%).

2 4๊ฐ€์ง€ Tier ์ดํ•ดํ•˜๊ธฐUnderstanding the 4 Tiers

Tier ์ด๋ฆ„Name ๋ฐ์ดํ„ฐ ์†Œ์ŠคData Sources ์‹ ๋ขฐ๋„Confidence ํ™œ์šฉUse Case
Tier 1 ์›๊ฒฉRemote ์„ค๋ฌธ + ํƒ€์ดํ•‘ + ํฐ ์ˆ˜๋™Survey + typing + phone passive 60โ€“75% ์›๊ฒฉ ์Šคํฌ๋ฆฌ๋‹, ์ดˆ๊ธฐ ์„ ๋ณ„Remote screening
Tier 2 ์ž„์ƒClinical Tier 1 + qEEGTier 1 + qEEG 75โ€“88% ์ž„์ƒ ํ™•์ •, ๋ณดํ—˜ ์ฒญ๊ตฌConfirmation, insurance
Tier 3 ์ข…ํ•ฉComprehensive Tier 2 + HRV + ์–ผ๊ตด + ์Œ์„ฑTier 2 + HRV + face + voice 88โ€“95% ์น˜๋ฃŒ ๊ณ„ํš, ๋ฐ˜์‘ ์ถ”์ Treatment planning
Tier 4 ์—ฐ๊ตฌResearch Tier 3 + ERP + ์ข…๋‹จ ์ถ”์ Tier 3 + ERP + longitudinal 95โ€“98% FDA 510(k), ์ถœํŒFDA 510(k), publication
โš ๏ธ Tier ์„ ํƒ ์‹œ ์ฃผ์˜When Choosing a Tier

์ฒ˜์Œ ์˜ค๋Š” ํด๋ผ์ด์–ธํŠธ๋Š” ๋ณดํ†ต Tier 2๋กœ ์‹œ์ž‘ํ•˜๊ณ , ์น˜๋ฃŒ ์ง„ํ–‰ ์ค‘ 3๊ฐœ์›”๋งˆ๋‹ค Tier 3๋กœ ํŒ”๋กœ์—…ํ•˜๋Š” ํŒจํ„ด์ด ํšจ๊ณผ์ ์ž…๋‹ˆ๋‹ค. Tier 1์€ ์›๊ฒฉ ํ‰๊ฐ€๋‚˜ ๋ณดํ—˜์ด ์ปค๋ฒ„๋˜์ง€ ์•Š๋Š” ์ดˆ๊ธฐ ์ƒ๋‹ด์šฉ. For new clients, Tier 2 is typically the starting point, with Tier 3 follow-ups every 3 months during treatment. Tier 1 is best for remote screening or uninsured initial consults.

3 ์‚ฌ์šฉ ์ˆœ์„œ (Workflow)Workflow

ํด๋ผ์ด์–ธํŠธ ํ•œ ๋ช…์˜ DMDA ํ‰๊ฐ€๋Š” ์•„๋ž˜ ์ˆœ์„œ๋กœ ์ง„ํ–‰๋ฉ๋‹ˆ๋‹ค: A DMDA assessment for one client proceeds in this order:

01
IntakeIntake
ํด๋ผ์ด์–ธํŠธ ๋“ฑ๋ก + 90๋ฌธํ•ญ ์„ค๋ฌธ + ํƒ€์ดํ•‘ ๋ฐ”์ด์˜ค๋งˆ์ปค ์ž๋™ ์ˆ˜์ง‘Register client + 90-item survey + auto-captured typing biomarkers
intake.html
02
AssessmentAssessment
Tier 2+: qEEG, HRV, Face, Voice, Phone ๋ฐ์ดํ„ฐ ์ž…๋ ฅTier 2+: enter qEEG/HRV/Face/Voice/Phone data
assessment.html
03
EngineEngine
์—”์ง„ ์‹คํ–‰ โ†’ ์ง„๋‹จ + ์‹ ๋ขฐ๋„ + ๊ทผ๊ฑฐ ์ž๋™ ์‚ฐ์ถœRun engine โ†’ diagnosis + confidence + evidence
engine.html
04
ReportReport
์ž„์ƒ ๋ฆฌํฌํŠธ ๊ฒ€ํ†  ยท ๋ฉ”๋ชจ ยท ์ธ์‡„ / PDFReview report, add notes, print/PDF
report.html
โœ… ํŒ: ์ž๋™ ๋„ค๋น„๊ฒŒ์ด์…˜Tip: Auto Navigation

๊ฐ ๋‹จ๊ณ„ ๋์— ๋‚˜์˜ค๋Š” ๋ฒ„ํŠผ์ด ์ž๋™์œผ๋กœ ๋‹ค์Œ ํŽ˜์ด์ง€(ํ•ด๋‹น ์„ธ์…˜ ID ํฌํ•จ)๋กœ ์ด๋™์‹œ์ผœ์ค๋‹ˆ๋‹ค. URL์„ ๋”ฐ๋กœ ๋ณต์‚ฌํ•  ํ•„์š” ์—†์–ด์š”. Buttons at the end of each step automatically navigate to the next page with the session ID attached. No need to copy URLs.

4 ํŽ˜์ด์ง€๋ณ„ ์•ˆ๋‚ดPage-by-Page Guide

๐Ÿ“ intake.html โ€” ํด๋ผ์ด์–ธํŠธ ์ ‘์ˆ˜Client Intake

  1. ํด๋ผ์ด์–ธํŠธ ์ •๋ณด: ์ด๋ฆ„๊ณผ ์ด๋ฉ”์ผ ํ•„์ˆ˜. ์ด๋ฏธ ๋“ฑ๋ก๋œ ์ด๋ฉ”์ผ์ด๋ฉด ์ž๋™ ๋งค์นญ๋จ. Client info: Name + email required. Existing emails auto-matched.
  2. Tier ์„ ํƒ: ๋„ค ์นด๋“œ ์ค‘ ํ•˜๋‚˜ ํด๋ฆญ. Tier selection: Click one of the four cards.
  3. ํƒ€์ดํ•‘ ์บ˜๋ฆฌ๋ธŒ๋ ˆ์ด์…˜: 20์ดˆ ์ด์ƒ ์ž์œ  ํƒ€์ดํ•‘. ํ‚ค์ŠคํŠธ๋กœํฌ ๋ฆฌ๋“ฌยท๋ฐฑ์ŠคํŽ˜์ด์Šค ํŒจํ„ด์ด ๊ฐ๊ด€ ๋ฐ”์ด์˜ค๋งˆ์ปค๊ฐ€ ๋ฉ๋‹ˆ๋‹ค. Typing calibration: Type freely for 20+ seconds. Keystroke rhythm and backspace patterns become objective biomarkers.
  4. 90๋ฌธํ•ญ ์„ค๋ฌธ: ๊ฐ ๋ฌธํ•ญ์— 0โ€“4 Likert. ์‘๋‹ต ์‹œ๊ฐ„๊ณผ ์ˆ˜์ • ํšŸ์ˆ˜๊ฐ€ ์ž๋™ ๊ธฐ๋ก๋จ (์šฐ์ธก ํ•˜๋‹จ ์‹ค์‹œ๊ฐ„ HUD ์ฐธ๊ณ ). 90-item survey: 0โ€“4 Likert each. Response time and edits auto-captured (see bottom-right live HUD).
  5. ์ œ์ถœ: 70% ์ด์ƒ ์‘๋‹ต ์‹œ ์ œ์ถœ ๊ฐ€๋Šฅ. Tier 1์ด๋ฉด ๋ฐ”๋กœ ์—”์ง„์œผ๋กœ, Tier 2+์ด๋ฉด Assessment๋กœ ์ง„ํ–‰. Submit: Enabled at 70%+ answered. Tier 1 โ†’ engine directly; Tier 2+ โ†’ Assessment next.

โšก assessment.html โ€” ๊ฐ๊ด€ ๋ฐ์ดํ„ฐ ์ž…๋ ฅObjective Data Entry

5๊ฐœ ํƒญ (qEEG / HRV / Face / Voice / Phone Passive). ๊ฐ ํƒญ์—์„œ ๋ฐ์ดํ„ฐ ์ž…๋ ฅ ํ›„ "์ €์žฅ" ํด๋ฆญ. ํƒญ์— ์ดˆ๋ก ์ (โ—)์ด ํ‘œ์‹œ๋˜๋ฉด ์™„๋ฃŒ. 5 tabs (qEEG / HRV / Face / Voice / Phone Passive). Enter data in each tab, click Save. Green dot (โ—) on a tab = completed.

๐Ÿ”‘ qEEG ํ•ต์‹ฌ 4๊ฐ€์ง€ ๋งˆ์ปคKey qEEG Markers
  • Frontal Alpha Asymmetry: < -0.3 = ์šฐ์šธ ์‹ ํ˜ธdepression signal
  • Theta/Beta Ratio: > 2.8 = ADHD ์‹ ํ˜ธADHD signal
  • Peak Alpha Freq: < 9 Hz = ์ธ์ง€ ๋‘”ํ™”cognitive slowing
  • High-beta Z: > 1.5 = ๊ณผ๊ฐ์„ฑ (๋ถˆ์•ˆ/PTSD)hyperarousal (anxiety/PTSD)

๐Ÿ”ฌ engine.html โ€” ์—”์ง„ ์‹คํ–‰Engine

ํŽ˜์ด์ง€๊ฐ€ ์ž๋™์œผ๋กœ ์ˆ˜์ง‘๋œ ๋ฐ์ดํ„ฐ๋ฅผ ๋ณด์—ฌ์ค๋‹ˆ๋‹ค. "DMDA ๋ถ„์„ ์‹คํ–‰" ๋ฒ„ํŠผ์„ ๋ˆ„๋ฅด๋ฉด: The page auto-shows collected data. Click "Run Progressive DMDA Analysis" to:

  1. ์—”์ง„์ด ๊ฐ€์šฉํ•œ ๋ฐ์ดํ„ฐ๋งŒ ์‚ฌ์šฉํ•ด Smart Weighting ์ ์šฉ Engine applies Smart Weighting using only available data
  2. Bayesian ์กฐํ•ฉ์œผ๋กœ ๊ฐ ICD-10 ์ฝ”๋“œ์˜ ์ ์ˆ˜ ๊ณ„์‚ฐ Bayesian combination computes scores for each ICD-10 code
  3. ์ตœ๊ณ  ์ ์ˆ˜๊ฐ€ 1์ฐจ ์ง„๋‹จ, ์ƒ์œ„ ํ›„๋ณด๊ฐ€ ๊ฐ๋ณ„ ์ง„๋‹จ Top score = primary; top candidates = differential
  4. ๊ฒฐ๊ณผ๊ฐ€ ํ™”๋ฉด์— ํ‘œ์‹œ๋˜๊ณ  DB์— ์ €์žฅ๋จ (์„ธ์…˜ ์ƒํƒœ = completed) Results displayed; session marked completed in DB

๐Ÿ“„ report.html โ€” ์ž„์ƒ ๋ฆฌํฌํŠธClinical Report

์ธ์‡„ ๊ฐ€๋Šฅํ•œ ์ž„์ƒ ๋ฆฌํฌํŠธ. ํฌํ•จ ๋‚ด์šฉ: ํด๋ผ์ด์–ธํŠธ ์ •๋ณด ยท 1์ฐจ ์ง„๋‹จ + ์‹ ๋ขฐ๋„ ยท ๊ฐ๋ณ„ ์ง„๋‹จ ยท ๋ฐ์ดํ„ฐ ๊ฐ€์ค‘์น˜ ยท ๋ชจ๋“  ๋ฐ”์ด์˜ค๋งˆ์ปค ยท ๊ทผ๊ฑฐ ์ถ”์  ยท Tier ์—…๊ทธ๋ ˆ์ด๋“œ ๊ถŒ์žฅ ยท ์ž„์ƒ๊ฐ€ ๋ฉ”๋ชจ ยท ์„œ๋ช…๋ž€. Printable clinical report. Contains: client info ยท primary dx + confidence ยท differential ยท data weights ยท all biomarkers ยท evidence trail ยท tier upgrade recommendation ยท clinician notes ยท signature block.

์ธ์‡„: ์ƒ๋‹จ "๐Ÿ–จ๏ธ ์ธ์‡„ / PDF ์ €์žฅ" ๋ฒ„ํŠผ. ์–‘๊ตญ์–ด ์ธ์‡„: "๐ŸŒ ์–‘๊ตญ์–ด ํ‘œ์‹œ" ๋ฒ„ํŠผ ๋จผ์ € ํด๋ฆญ. Print: Click "๐Ÿ–จ๏ธ Print / Save PDF" at top. Bilingual print: Click "๐ŸŒ Show both languages" first.

๐Ÿ“‹ admin.html โ€” ๊ด€๋ฆฌ ๋Œ€์‹œ๋ณด๋“œAdmin Dashboard

5 ์ „์ฒด 16๊ฐ€์ง€ ์‹œ๋‚˜๋ฆฌ์˜ค ๊ฒฐ๊ณผAll 16 Scenario Outcomes

4๋ช…์˜ ๊ฐ€์ƒ ํด๋ผ์ด์–ธํŠธ ร— 4๊ฐœ Tier = 16๊ฐ€์ง€ ์กฐํ•ฉ. ์‹œ๋‚˜๋ฆฌ์˜ค ์‹œ๋ฎฌ๋ ˆ์ดํ„ฐ์—์„œ ์ง์ ‘ ์ฒดํ—˜ํ•ด๋ณด์„ธ์š”. 4 fictional clients ร— 4 tiers = 16 combinations. Try them interactively at the Scenario Simulator.

์ผ€์ด์ŠคCase Tier 1 Tier 2 Tier 3 Tier 4
๐Ÿ˜” ์‚ฌ๋ผ (์šฐ์šธ)Sarah (MDD) F32.2 ยท 69% F32.2 ยท 80% F32.2 ยท 89% F32.2 ยท 96%
๐ŸŒ€ ๋งˆ์ปค์Šค (ADHD)Marcus (ADHD) F90.2 ยท 66% F90.2 ยท 79% F90.2 ยท 89% F90.2 ยท 96%
๐Ÿ˜ฐ ํ”„๋ฆฌ์•ผ (GAD)Priya (GAD) F41.0 ยท 65% F41.1 ยท 77% F41.1 ยท 89% F41.1 ยท 95%
โšก ๋ฐ์ด๋น„๋“œ (PTSD)David (PTSD) F43.0 ยท 67% F43.10 ยท 79% F43.10 ยท 89% F43.10 ยท 96%
๐Ÿ” ์ด ํ‘œ์—์„œ ์ฃผ๋ชฉํ•  ์ Notable Observations
  • Progressive ์ˆ˜๋ ด: ํ”„๋ฆฌ์•ผ๋Š” Tier 1์—์„œ ๊ณตํ™ฉ์žฅ์• (F41.0)๋กœ ์‚ด์ง ์˜ค๋ถ„๋ฅ˜๋˜์ง€๋งŒ, Tier 2์—์„œ qEEG๊ฐ€ ์ถ”๊ฐ€๋˜์ž ์ •ํ™•ํžˆ GAD(F41.1)๋กœ ์ˆ˜๋ ดํ•จ. Progressive convergence: Priya is slightly misclassified as Panic (F41.0) at Tier 1, but converges correctly to GAD (F41.1) once qEEG is added at Tier 2.
  • ๊ธ‰์„ฑ vs ๋งŒ์„ฑ: ๋ฐ์ด๋น„๋“œ๋Š” Tier 1์—์„œ ๊ธ‰์„ฑ์ŠคํŠธ๋ ˆ์Šค(F43.0)๋กœ ๋ณด์ด์ง€๋งŒ Tier 2+์—์„œ ๋งŒ์„ฑ PTSD(F43.10)๋กœ ์ •ํ™•ํžˆ ๋ถ„๋ฅ˜๋จ. Acute vs chronic: David appears as Acute Stress (F43.0) at Tier 1 but correctly classified as chronic PTSD (F43.10) at Tier 2+.
  • ์‹ ๋ขฐ๋„ ์ง„ํ–‰: ๋ชจ๋“  ์ผ€์ด์Šค์—์„œ Tier๊ฐ€ ์˜ฌ๋ผ๊ฐˆ์ˆ˜๋ก ์‹ ๋ขฐ๋„๊ฐ€ ์•ฝ 10%p์”ฉ ์ƒ์Šน. Confidence progression: All cases show ~10pp confidence increase per tier.

6 ๋””์ง€ํ„ธ ๋ฐ”์ด์˜ค๋งˆ์ปค ์ดํ•ดํ•˜๊ธฐUnderstanding Digital Biomarkers

๋””์ง€ํ„ธ ๋ฐ”์ด์˜ค๋งˆ์ปค๋ž€ "์ˆซ์ž๋กœ ์ธก์ • ๊ฐ€๋Šฅํ•œ ๊ฐ๊ด€์  ์‹ ํ˜ธ"์ž…๋‹ˆ๋‹ค. DMDA๋Š” ๋‹ค์Œ์„ ์ž๋™ ์ˆ˜์ง‘ยท๋ถ„์„ํ•ฉ๋‹ˆ๋‹ค: Digital biomarkers are objective signals that can be measured as numbers. DMDA auto-captures and analyzes:

โฑ๏ธ ์‘๋‹ต ๋ฉ”ํƒ€ (Response Meta)Response Meta
์„ค๋ฌธ ๊ฐ ๋ฌธํ•ญ์— ๋‹ตํ•˜๋Š”๋ฐ ๊ฑธ๋ฆฐ ์‹œ๊ฐ„, ๋‹ต ์ˆ˜์ • ํšŸ์ˆ˜, ๊ฑด๋„ˆ๋›ฐ๊ธฐ ํŒจํ„ด.Time to answer each item, number of changes, skip patterns.
โŒจ๏ธ ํƒ€์ดํ•‘ ๋™์—ญํ•™Typing Dynamics
ํ‚ค์ŠคํŠธ๋กœํฌ ๊ฐ„๊ฒฉ, WPM, ๋ฐฑ์ŠคํŽ˜์ด์Šค ๋น„์œจ, ๋‹จ์–ด ๋‚ด ๋ฉˆ์ถค.Inter-keystroke intervals, WPM, backspace rate, mid-word pauses.
๐Ÿ“ฑ ํฐ ์ˆ˜๋™ ๋ฐ์ดํ„ฐPhone Passive
์ˆ˜๋ฉด ์‹œ๊ฐ„, ๊ฑธ์Œ์ˆ˜, ์‚ฌํšŒ์  ํ†ต์‹  ๋นˆ๋„, ์•ฑ ์ „ํ™˜ ๋“ฑ (HealthKit/Google Fit ์—ฐ๋™ ๊ฐ€๋Šฅ).Sleep duration, steps, social comms, app switches (HealthKit/Google Fit integration possible).
๐Ÿง  qEEG
19์ฑ„๋„ EEG ์ •๋Ÿ‰ ๋ถ„์„ (Mitsar + Neuroguide).19-channel EEG quantitative analysis (Mitsar + Neuroguide).
โค๏ธ HRV
์‹ฌ๋ฐ•๋ณ€์ด๋„ (RMSSD, SDNN, LF/HF). ์ž์œจ์‹ ๊ฒฝ ์ง€ํ‘œ.Heart rate variability (RMSSD, SDNN, LF/HF). Autonomic indicator.
๐Ÿ˜ ์–ผ๊ตด ๋ถ„์„Face Analysis
๋“€์…ด ๋ฏธ์†Œ, ๋ฏธ๊ฐ„ ๊ธด์žฅ, ๋†€๋žŒ ๋ฐ˜์‘, ์‹œ์„  ๊ณ ์ •.Duchenne smile, brow tension, startle, fixation.
๐ŸŽ™๏ธ ์Œ์„ฑ ๋ถ„์„Voice Analysis
ํ”ผ์น˜, ์ง€ํ„ฐ, ๋ง์†๋„, ๋ง์†๋„ ๋ณ€๋™.Pitch, jitter, speech rate, rate variance.
โšก ERP
์‚ฌ๊ฑด๊ด€๋ จ์ „์œ„ (P300 ๋“ฑ). HBImed/Kropotov ๊ธฐ๋ฐ˜.Event-related potentials (P300 etc). HBImed/Kropotov based.

7 ์ฃผ์š” ์šฉ์–ด์ง‘Glossary

Smart Weighting
๊ฐ€์šฉํ•œ ๋ฐ์ดํ„ฐ ์†Œ์Šค์— ๋”ฐ๋ผ ๊ฐ ์†Œ์Šค์˜ ๊ฐ€์ค‘์น˜๋ฅผ ์ž๋™ ์žฌ์ •๊ทœํ™” (ํ•ฉ=100%).Auto-renormalizes source weights based on what's available (sum=100%).
Confidence Band
๊ฐ Tier์˜ ์‹ ๋ขฐ๋„ ๋ชฉํ‘œ ๋ฒ”์œ„ (์˜ˆ: Tier 2๋Š” 75โ€“88%).Target confidence range per tier (e.g., Tier 2 = 75โ€“88%).
Evidence Trail
๊ฐ ๋ฐ์ดํ„ฐ ์†Œ์Šค๊ฐ€ ์ง„๋‹จ์— ๊ธฐ์—ฌํ•œ ์ •๋„์™€ ์ด์œ ๋ฅผ ์ถ”์ ํ•œ ๊ธฐ๋ก.Record of how much and why each data source contributed to the diagnosis.
Differential Diagnosis
1์ฐจ ์ง„๋‹จ ์™ธ์— ๊ณ ๋ คํ•ด์•ผ ํ•  ๋Œ€์•ˆ ์ง„๋‹จ ํ›„๋ณด๋“ค (softmax ํ™•๋ฅ ).Alternative dx candidates besides the primary (softmax probabilities).
Tier Upgrade
๋” ๋งŽ์€ ๋ฐ์ดํ„ฐ๋ฅผ ์ถ”๊ฐ€ํ•ด ๋‹ค์Œ Tier๋กœ ์ƒ์Šนํ•˜๊ธฐ. ์‹ ๋ขฐ๋„ ์ƒ์Šน + ๋น„์šฉ ๋ฐœ์ƒ.Moving to next tier by adding more data. Higher confidence + added cost.
Session
DMDA ํ‰๊ฐ€ 1ํšŒ ๋‹จ์œ„. ํ•œ ํด๋ผ์ด์–ธํŠธ๋Š” ์—ฌ๋Ÿฌ ์„ธ์…˜์„ ๊ฐ€์งˆ ์ˆ˜ ์žˆ์Œ.One DMDA assessment unit. One client can have multiple sessions.
BPS-90
Biopsychosocial 90-item survey. 12๊ฐœ ์นดํ…Œ๊ณ ๋ฆฌ(์šฐ์šธ/๋ถˆ์•ˆ/ADHD/์™ธ์ƒ/์ˆ˜๋ฉด ๋“ฑ)๋ฅผ ๋ง๋ผ.90-item biopsychosocial survey spanning 12 categories (depression, anxiety, ADHD, trauma, sleep, etc.).
Client vs Patient
UI์—์„œ๋Š” "Client" ์‚ฌ์šฉ (BCN+PhD non-physician scope์— ๋งž์ถค). DB ํ…Œ์ด๋ธ”๋ช…์€ `patients` legacy ์œ ์ง€.UI uses "Client" (matches BCN+PhD non-physician scope). DB table name `patients` kept as legacy.

8FAQ

DMDA๋Š” ์‹ค์ œ ์ง„๋‹จ ๋„๊ตฌ์ธ๊ฐ€์š”?Is DMDA a real diagnostic tool?
DMDA๋Š” ์ž„์ƒ ์˜์‚ฌ๊ฒฐ์ • ์ง€์› ๋„๊ตฌ (Clinical Decision Support)์ž…๋‹ˆ๋‹ค. ์ž๊ฒฉ ์žˆ๋Š” ์ž„์ƒ๊ฐ€๊ฐ€ ๋ฐ˜๋“œ์‹œ ๊ฒ€ํ† ยท์„œ๋ช…ํ•ด์•ผ ํ•˜๋ฉฐ, ๋‹จ๋…์œผ๋กœ ํ™•์ • ์ง„๋‹จ์œผ๋กœ ์‚ฌ์šฉํ•˜์ง€ ์•Š์Šต๋‹ˆ๋‹ค.DMDA is a Clinical Decision Support tool. It must be reviewed and signed by a qualified clinician and is not used as a standalone diagnosis.
Tier 1๋งŒ์œผ๋กœ ์ถฉ๋ถ„ํ•œ๊ฐ€์š”?Is Tier 1 enough?
Tier 1์€ 60โ€“75% ์‹ ๋ขฐ๋„๋กœ ์Šคํฌ๋ฆฌ๋‹์— ์ ํ•ฉํ•˜์ง€๋งŒ, ์น˜๋ฃŒ ๊ณ„ํš์—๋Š” Tier 2+ (qEEG ํฌํ•จ)๋ฅผ ๊ถŒ์žฅํ•ฉ๋‹ˆ๋‹ค.Tier 1 is good for screening at 60โ€“75% confidence, but Tier 2+ (with qEEG) is recommended for treatment planning.
์ผ๋ถ€ ๋ฐ์ดํ„ฐ๊ฐ€ ์—†์œผ๋ฉด ์–ด๋–ป๊ฒŒ ๋˜๋‚˜์š”?What if some data is missing?
Smart Weighting์ด ๊ฐ€์šฉํ•œ ๋ฐ์ดํ„ฐ๋งŒ์œผ๋กœ ๊ฐ€์ค‘์น˜๋ฅผ ์ž๋™ ์žฌ์ •๊ทœํ™”ํ•ฉ๋‹ˆ๋‹ค. ์ง„๋‹จ์€ ์—ฌ์ „ํžˆ ๊ฐ€๋Šฅํ•˜์ง€๋งŒ ์‹ ๋ขฐ๋„๊ฐ€ ํ•ด๋‹น Tier ๋ฐด๋“œ ๋‚ด์—์„œ ํ•˜ํ–ฅ ์กฐ์ •๋ฉ๋‹ˆ๋‹ค.Smart Weighting auto-renormalizes using only available data. Diagnosis still possible, but confidence adjusts lower within the tier band.
๊ฐ๋ณ„ ์ง„๋‹จ์ด ์—ฌ๋Ÿฌ ๊ฐœ ๋‚˜์˜ค๋ฉด?What if there are multiple differentials?
์ •์ƒ์ž…๋‹ˆ๋‹ค. ๊ฐ๋ณ„ ์ง„๋‹จ์€ ๊ณต์กด์งˆํ™˜ ๊ฐ€๋Šฅ์„ฑ์ด๋‚˜ ๋Œ€์•ˆ ์ง„๋‹จ ํ›„๋ณด๋ฅผ ๋ณด์—ฌ์ฃผ๋ฉฐ, ์ž„์ƒ๊ฐ€์˜ ์ตœ์ข… ํŒ๋‹จ์ด ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. ์ฐจ์ด๊ฐ€ ์ž‘์œผ๋ฉด Tier ์—…๊ทธ๋ ˆ์ด๋“œ๊ฐ€ ๋„์›€๋ฉ๋‹ˆ๋‹ค.Normal. Differentials show comorbidity possibilities or alternative candidates; clinician's final judgment needed. Small gaps โ†’ tier upgrade helps.
๋ฆฌํฌํŠธ์— ์ž„์ƒ๊ฐ€ ๋ฉ”๋ชจ๋ฅผ ์ถ”๊ฐ€ํ•˜๋ ค๋ฉด?How do I add clinician notes to the report?
Admin ํŒจ๋„ โ†’ ์„ธ์…˜ ํ–‰์˜ โœ๏ธ ๋ฒ„ํŠผ ํด๋ฆญ โ†’ ๋ฉ”๋ชจ ํŽธ์ง‘ โ†’ ์ €์žฅ. ๋‹ค์Œ ๋ฆฌํฌํŠธ ์—ด๊ธฐ๋ถ€ํ„ฐ ๋ฐ˜์˜๋ฉ๋‹ˆ๋‹ค.Admin panel โ†’ โœ๏ธ button in session row โ†’ edit โ†’ save. Reflected next time the report is viewed.
ํด๋ผ์ด์–ธํŠธ๊ฐ€ ํ•œ ๋ฒˆ ๋” ํ‰๊ฐ€๋ฐ›์œผ๋ ค๋ฉด?How does a client get re-assessed?
intake.html์—์„œ ๊ฐ™์€ ์ด๋ฉ”์ผ๋กœ ๋‹ค์‹œ ์ ‘์ˆ˜ํ•˜์„ธ์š”. ๊ธฐ์กด ํด๋ผ์ด์–ธํŠธ๋กœ ์ธ์‹๋˜์–ด ์ƒˆ ์„ธ์…˜์ด ์ƒ์„ฑ๋ฉ๋‹ˆ๋‹ค (๊ธฐ์กด ์„ธ์…˜ ๋ณด์กด).Register again via intake.html with the same email. Recognized as existing client, new session created (old sessions preserved).
์‹œ์Šคํ…œ ์–ธ์–ด๋ฅผ ์˜์–ด๋กœ ์™„์ „ํžˆ ๋ฐ”๊พธ๋ ค๋ฉด?How to switch entire system to English?
๊ฐ ํŽ˜์ด์ง€ ์ƒ๋‹จ ์šฐ์ธก์˜ "English" ๋ฒ„ํŠผ. ํŽ˜์ด์ง€๋ณ„๋กœ ๊ฐœ๋ณ„ ์„ค์ •๋ฉ๋‹ˆ๋‹ค. URL์— ?lang=en์„ ์ถ”๊ฐ€ํ•˜๋ฉด ์‹œ์ž‘๋ถ€ํ„ฐ ์˜์–ด๋กœ ์—ด๋ฆฝ๋‹ˆ๋‹ค."English" button at top-right of each page. Set per-page. Add ?lang=en to URL to open in English from start.

9 ๋ฌธ์ œ ํ•ด๊ฒฐTroubleshooting

โŒ "Supabase key not set" ์—๋Ÿฌerror

์›์ธ: Supabase anon key๊ฐ€ ํŽ˜์ด์ง€์— ์ฃผ์ž…๋˜์ง€ ์•Š์Œ.
ํ•ด๊ฒฐ: ๊ฐ ํŽ˜์ด์ง€์— <script>window.DMDA_SUPABASE_KEY = '...'</script> ์ถ”๊ฐ€. ๋ฐฐํฌ ๊ฐ€์ด๋“œ(README.md) Step 3 ์ฐธ์กฐ.
Cause: Supabase anon key not injected.
Fix: Add <script>window.DMDA_SUPABASE_KEY = '...'</script> to each page. See README.md Step 3.

โŒ ์„ธ์…˜ ID๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค ์—๋Ÿฌ"Session ID required" error

assessment/engine/report ํŽ˜์ด์ง€๋Š” URL์— ?session=...๊ฐ€ ํ•„์š”ํ•ฉ๋‹ˆ๋‹ค. intake.html์—์„œ ์ •์ƒ์ ์œผ๋กœ ์ง„ํ–‰ํ•˜๋ฉด ์ž๋™์œผ๋กœ ๋ถ™์Šต๋‹ˆ๋‹ค. assessment/engine/report pages need ?session=... in the URL. Automatically attached when you progress from intake.html properly.

โš ๏ธ ์ง„๋‹จ ์‹ ๋ขฐ๋„๊ฐ€ Tier ๋ฐด๋“œ๋ณด๋‹ค ๋‚ฎ์ŒConfidence below tier band

์ˆ˜์ง‘๋œ ๋ฐ์ดํ„ฐ๊ฐ€ ๋ถ€์กฑํ•˜๊ฑฐ๋‚˜ ์ฆ์ƒ ํŒจํ„ด์ด ์• ๋งคํ•ฉ๋‹ˆ๋‹ค. Tier๋ฅผ ์—…๊ทธ๋ ˆ์ด๋“œํ•˜๊ฑฐ๋‚˜, ์„ค๋ฌธ ์‘๋‹ต๋ฅ ์„ ๋†’์ด์„ธ์š” (์ตœ์†Œ 85% ์ด์ƒ). Insufficient data collected or ambiguous symptom pattern. Upgrade tier or increase survey completion (aim for โ‰ฅ85%).

๐Ÿ’ก ์—”์ง„์ด ์˜ˆ์ƒ๊ณผ ๋‹ค๋ฅธ ์ง„๋‹จ์„ ๋ƒ„Engine gave unexpected diagnosis

1) Evidence Trail ์„น์…˜์—์„œ ์–ด๋–ค ๋ฐ์ดํ„ฐ๊ฐ€ ๊ฐ€์ค‘์น˜๊ฐ€ ์ปธ๋Š”์ง€ ํ™•์ธ. 2) ๊ฐ๋ณ„ ์ง„๋‹จ์˜ 2-3์œ„๋„ ๊ฒ€ํ†  (์ ์ˆ˜ ์ฐจ์ด๊ฐ€ ์ž‘์œผ๋ฉด ๊ณต์กด์งˆํ™˜ ๊ฐ€๋Šฅ). 3) ์˜์‹ฌ๋˜๋ฉด ๋” ๋†’์€ Tier๋กœ ์žฌํ‰๊ฐ€. 1) Check Evidence Trail to see which data weighted most. 2) Review 2nd-3rd differentials (small gaps = possible comorbidity). 3) If uncertain, re-assess at higher tier.