Chapter 03: Screening Protocols

Evidence-Based Detection and Risk Assessment
弘益人間 · Benefit All Humanity

The Role of Screening in Mental Healthcare

Mental health screening is a critical first step in identifying individuals who may benefit from further assessment and intervention. Early detection through systematic screening can dramatically improve outcomes by enabling timely treatment, preventing crisis situations, and reducing the long-term burden of mental illness. AI-powered screening protocols have the potential to make mental health screening more accessible, consistent, and effective than ever before.

Traditional screening approaches face several limitations: they require trained personnel to administer, they depend on patient self-report which may be influenced by stigma or lack of insight, and they are typically administered only at specific clinical touchpoints rather than continuously. AI systems can address many of these limitations by enabling automated, continuous, and objective screening at scale.

Key Principles of AI-Powered Mental Health Screening:

Validated Screening Instruments

AI screening systems should be based on clinically validated instruments with established psychometric properties. These instruments have been tested across diverse populations and have known sensitivity, specificity, and predictive values. AI systems can administer these instruments, score them automatically, and integrate results with other data sources for enhanced accuracy.

Instrument Condition(s) Screened Items Sensitivity/Specificity AI Implementation
PHQ-9 Major Depression 9 items 88% / 88% Conversational adaptation, passive scoring
GAD-7 Generalized Anxiety 7 items 89% / 82% Chatbot administration, voice-based delivery
PSS-10 Perceived Stress 10 items N/A (continuous) Contextual questioning, temporal tracking
PCL-5 PTSD 20 items 88% / 79% Trauma-sensitive delivery, skip logic
MDQ Bipolar Disorder 13 items + 2 follow-ups 73% / 90% Pattern recognition, mood tracking integration
AUDIT-C Alcohol Use 3 items 95% / 60% Non-judgmental AI delivery, anonymity preservation
Columbia-Suicide Severity Rating Scale Suicidal Ideation/Behavior 6 core items 97% / 95% Crisis detection, immediate escalation protocol

AI-Enhanced Screening Approaches

While traditional instruments remain the foundation of mental health screening, AI enables several enhancements that can improve accuracy, reduce burden, and enable continuous monitoring. These approaches augment rather than replace validated instruments.

1. Conversational Screening

Rather than presenting screening questions in a rigid, survey-like format, AI chatbots can integrate screening into natural conversation. This approach reduces the clinical feel of screening, may increase honest disclosure, and provides contextual information that enhances interpretation of responses.

// Example: Conversational PHQ-9 Implementation
import { ConversationalScreening } from '@wia/mental-002';

class ConversationalPHQ9 {
  constructor() {
    this.screener = new ConversationalScreening({
      instrument: 'PHQ-9',
      conversationStyle: 'empathetic',
      adaptiveBranching: true,
      culturalAdaptation: true
    });
  }

  async conductScreening(userId, conversationContext = {}) {
    const session = await this.screener.initializeSession({
      userId,
      previousConversations: conversationContext.history,
      demographicInfo: conversationContext.demographics
    });

    // Natural conversation flow integrating PHQ-9 items
    const responses = [];

    // Example: Instead of "Little interest or pleasure in doing things?"
    // AI adapts to conversation: "You mentioned you've been staying home more.
    // Have you noticed a change in how much you enjoy your usual activities?"

    for (const item of this.screener.getAdaptedItems()) {
      const response = await this.screener.askItem(item, {
        conversationContext: session.context,
        previousResponses: responses,
        adaptationStrategy: 'contextual'
      });

      responses.push({
        itemId: item.id,
        originalItem: item.standardText,
        adaptedQuestion: response.questionAsAsked,
        userResponse: response.answer,
        clarifications: response.clarifyingExchanges,
        confidence: response.scoringConfidence
      });

      // Update conversation context
      session.context = response.updatedContext;
    }

    // Calculate PHQ-9 score with confidence intervals
    const scoring = await this.screener.calculateScore(responses);

    return {
      instrumentScores: {
        phq9Total: scoring.totalScore,
        confidenceInterval: scoring.confidence,
        itemScores: scoring.itemBreakdown,
        suicidalIdeation: scoring.item9Score // Critical item
      },
      clinicalInterpretation: {
        severity: this.categorizeSeverity(scoring.totalScore),
        recommendations: this.generateRecommendations(scoring),
        requiresImmediateAction: scoring.item9Score >= 2,
        requiresFollowUp: scoring.totalScore >= 10
      },
      conversationalQuality: {
        naturalness: session.naturalness,
        userEngagement: session.engagement,
        completionQuality: scoring.completeness
      }
    };
  }

  categorizeSeverity(score) {
    if (score >= 20) return 'severe';
    if (score >= 15) return 'moderately-severe';
    if (score >= 10) return 'moderate';
    if (score >= 5) return 'mild';
    return 'minimal';
  }

  generateRecommendations(scoring) {
    const recommendations = [];

    if (scoring.item9Score >= 2) {
      recommendations.push({
        priority: 'immediate',
        action: 'suicide-risk-protocol',
        description: 'Activate crisis intervention protocol'
      });
    }

    if (scoring.totalScore >= 15) {
      recommendations.push({
        priority: 'urgent',
        action: 'clinical-referral',
        description: 'Refer for immediate clinical evaluation'
      });
    } else if (scoring.totalScore >= 10) {
      recommendations.push({
        priority: 'high',
        action: 'clinical-assessment',
        description: 'Schedule comprehensive clinical assessment'
      });
    } else if (scoring.totalScore >= 5) {
      recommendations.push({
        priority: 'moderate',
        action: 'monitoring',
        description: 'Continue monitoring, consider self-help resources'
      });
    }

    return recommendations;
  }
}

// Usage
const phq9Screener = new ConversationalPHQ9();
const screeningResult = await phq9Screener.conductScreening('user-789', {
  history: previousConversations,
  demographics: { age: 28, gender: 'female', culture: 'western' }
});

console.log(screeningResult.clinicalInterpretation);
// Output: { severity: 'moderate', recommendations: [...], requiresFollowUp: true }

2. Passive Screening Through Digital Phenotyping

AI systems can conduct passive screening by analyzing digital biomarkers without requiring active user participation. This approach enables continuous monitoring and can detect changes in mental health status before individuals are aware of or willing to report symptoms.

3. Adaptive Screening

Adaptive screening uses AI to dynamically select questions based on previous responses, reducing burden while maintaining accuracy. This approach, based on computerized adaptive testing (CAT) principles, can achieve equivalent accuracy to full instruments with 50% fewer items.

Risk Stratification and Prediction

Beyond identifying current mental health conditions, AI screening systems can stratify individuals by risk level and predict future mental health crises. This enables proactive intervention and efficient allocation of clinical resources to those most in need.

Risk Level Characteristics Monitoring Frequency Intervention Pathway
Critical (Immediate) Active suicidal ideation, severe symptoms, acute crisis Continuous real-time Immediate crisis intervention, emergency services
High Risk Severe symptoms, functional impairment, risk factors present Daily check-ins Urgent clinical referral, intensive monitoring
Moderate Risk Moderate symptoms, some impairment, screening positive Weekly assessments Clinical evaluation, structured treatment
Low Risk Mild symptoms or risk factors, minimal impairment Monthly monitoring Self-help resources, watchful waiting
Prevention No current symptoms, general population screening Quarterly screening Wellness support, early detection focus

Crisis Detection and Response

One of the most critical applications of AI screening is detecting individuals at immediate risk of harm to themselves or others. Crisis detection systems must balance sensitivity (catching all true crises) with specificity (avoiding excessive false alarms) while ensuring rapid response when genuine crises are identified.

Critical Safety Requirements for Crisis Detection:

// Example: Crisis Detection System with Safety Protocols
import { CrisisDetection } from '@wia/mental-002';

class CrisisDetectionSystem {
  constructor() {
    this.detector = new CrisisDetection({
      sensitivityLevel: 'high', // Err on side of caution
      multiFactorVerification: true,
      humanOversight: 'required',
      escalationProtocol: 'immediate'
    });

    this.crisisHotlines = {
      'US': '988',
      'UK': '116123',
      'global': '988' // Default to US
    };
  }

  async assessCrisisRisk(input) {
    // Multi-factor crisis assessment
    const assessments = await Promise.all([
      this.detector.analyzeSuicidalLanguage(input.text),
      this.detector.assessSeverityIndicators(input.screeningScores),
      this.detector.evaluateBehavioralWarnings(input.digitalPhenotype),
      this.detector.checkHistoricalPatterns(input.userId)
    ]);

    // Integrate multiple risk signals
    const riskScore = this.calculateIntegratedRisk(assessments);

    // Immediate action threshold
    if (riskScore.immediate >= 0.7) {
      return await this.activateCrisisProtocol({
        userId: input.userId,
        riskAssessment: riskScore,
        evidenceSummary: this.summarizeEvidence(assessments),
        timestamp: new Date()
      });
    }

    // Elevated risk monitoring
    if (riskScore.elevated >= 0.5) {
      return await this.escalateToHighRiskMonitoring({
        userId: input.userId,
        riskLevel: riskScore,
        monitoringPlan: this.createMonitoringPlan(riskScore)
      });
    }

    // Standard monitoring
    return {
      riskLevel: 'standard',
      score: riskScore,
      action: 'continue-monitoring',
      nextAssessment: this.calculateNextAssessment(riskScore)
    };
  }

  async activateCrisisProtocol(crisisData) {
    // Step 1: Immediate notification to crisis team
    await this.notifyCrisisTeam({
      priority: 'critical',
      userId: crisisData.userId,
      riskScore: crisisData.riskAssessment,
      evidence: crisisData.evidenceSummary
    });

    // Step 2: Connect user to immediate support
    const supportConnection = await this.connectToCrisisSupport(
      crisisData.userId
    );

    // Step 3: Activate safety planning if available
    const safetyPlan = await this.retrieveSafetyPlan(crisisData.userId);

    // Step 4: Notify emergency contacts if authorized
    if (await this.isEmergencyContactAuthorized(crisisData.userId)) {
      await this.notifyEmergencyContacts(crisisData);
    }

    // Step 5: Log all actions for clinical review
    await this.logCrisisEvent({
      ...crisisData,
      actionsT: [
        'crisis-team-notified',
        'support-connected',
        'safety-plan-activated',
        'emergency-contacts-notified'
      ],
      humanReviewRequired: true,
      humanReviewBy: new Date(Date.now() + 5 * 60000) // 5 minutes
    });

    return {
      status: 'crisis-protocol-activated',
      supportConnection,
      safetyPlan,
      nextSteps: [
        'Crisis counselor will contact within 5 minutes',
        'Emergency services on standby',
        'Continuous monitoring activated'
      ]
    };
  }

  calculateIntegratedRisk(assessments) {
    // Weight different risk factors
    const weights = {
      suicidalLanguage: 0.40,
      severityIndicators: 0.25,
      behavioralWarnings: 0.20,
      historicalPatterns: 0.15
    };

    let immediateRisk = 0;
    let elevatedRisk = 0;

    for (const [key, weight] of Object.entries(weights)) {
      const assessment = assessments.find(a => a.type === key);
      immediateRisk += assessment.immediateScore * weight;
      elevatedRisk += assessment.elevatedScore * weight;
    }

    return {
      immediate: immediateRisk,
      elevated: elevatedRisk,
      confidence: this.calculateConfidence(assessments),
      factorsPresent: assessments.filter(a => a.present).map(a => a.type)
    };
  }
}

Population-Level Screening

AI enables mental health screening at population scale, allowing healthcare systems, universities, employers, and other organizations to identify at-risk individuals and understand mental health trends across their populations. Population screening must balance the benefits of early detection with ethical considerations around privacy, consent, and potential stigmatization.

Implementation Considerations

Screening in Special Populations

Mental health screening protocols may need adaptation for special populations including children and adolescents, older adults, pregnant and postpartum women, individuals with cognitive impairment, and culturally diverse populations. AI systems should incorporate population-specific instruments, norms, and interpretation guidelines.

Pediatric Screening Considerations:

Quality Assurance and Continuous Improvement

AI screening systems require ongoing monitoring and improvement to maintain accuracy and clinical utility. Quality assurance should include tracking of key performance metrics, regular validation against clinical outcomes, and incorporation of feedback from users and clinicians.

Key Takeaways

Review Questions

  1. What are the advantages and limitations of AI-powered screening compared to traditional clinician-administered screening?
  2. Explain how conversational screening differs from standard questionnaire administration. What are the potential benefits and challenges?
  3. Describe the risk stratification framework presented in this chapter. How should monitoring frequency and intervention intensity differ across risk levels?
  4. What safety protocols are essential for AI-based crisis detection systems? Why is human oversight particularly critical in this domain?
  5. Discuss the ethical considerations involved in population-level mental health screening. How can organizations balance early detection with privacy and autonomy?
  6. How might screening protocols need to be adapted for pediatric populations? What additional data sources should be incorporated?
  7. What key performance metrics should be tracked to ensure ongoing quality of AI screening systems?
  8. How can AI screening systems be designed to minimize false positives while maintaining high sensitivity for detecting true cases?

The principle of 弘益人間 calls us to make mental health screening universally accessible, removing barriers of cost, geography, stigma, and availability. Yet we must implement screening with wisdom, recognizing that identification of need is meaningful only when coupled with pathways to effective care. Our screening systems must be designed not merely to detect illness, but to connect people with the support they need to flourish. We commit to screening approaches that respect human dignity, protect privacy, and ultimately serve the wellbeing of all.

Korea Standardization Infrastructure Mapping

Korea operates a comprehensive standards governance system through inter-ministerial cooperation. National Standards Council (under Prime Minister's Office, per Framework Act on National Standards Article 5) coordinates KATS (Korean Agency for Technology and Standards), MFDS (Ministry of Food and Drug Safety), MOTIE (Ministry of Trade, Industry and Energy), MSIT (Ministry of Science and ICT), MOIS (Ministry of the Interior and Safety), MOE (Ministry of Environment), MOHW (Ministry of Health and Welfare), MND (Ministry of National Defense), MCST (Ministry of Culture, Sports and Tourism), MOFA (Ministry of Foreign Affairs), MOJ (Ministry of Justice), and FSC (Financial Services Commission). Accreditation and Testing: KOLAS (Korea Laboratory Accreditation Scheme) accredits 800+ testing laboratories. KAS (Korea Accreditation System) accredits 50+ certification bodies. KTC (Korea Testing Certification), KTR (Korea Testing & Research Institute), KTL (Korea Testing Laboratory), and KCL (Korea Conformity Laboratories) provide conformance testing. Telecom and Cyber: KCC (Korea Communications Commission), KCA (Korea Communications Agency), TTA (Telecommunications Technology Association), IITP (Institute for Information & Communications Technology Planning & Evaluation), NIPA (National IT Industry Promotion Agency), KISA (Korea Internet & Security Agency), KCMVP (Korea Cryptographic Module Validation Program), NIS (National Intelligence Service), NSR (National Security Research Institute), and NCSC (National Cyber Security Center). National R&D Centers: KIST, ETRI, KAIST, Seoul National University, Yonsei University, Korea University, POSTECH, UNIST, GIST, DGIST, KISTI, KIER, KIMM, KRICT, KFRI, KRIBB. International Standards Cooperation: ISO TC/SC Korean secretariats, IEC TC/SC Korean secretariats, ITU-T Study Group Korean chairs, 3GPP RAN/SA Korean chairs, IEEE 802 Korean chairs, W3C Korea office, OASIS Korea office, IETF Korea cooperation, OECD CSTP, UN ESCAP, APEC SCSC Korean cooperation. Korean Industrial Standards (KS) Catalog: KS X (Information) 25,000+, KS A (Basic) 15,000+, KS B (Machinery) 25,000+, KS C (Electrical) 18,000+, KS D (Metallurgy) 12,000+, KS E (Mining) 5,000+, KS F (Construction) 18,000+, KS H (Food) 8,000+, KS I (Environment) 5,000+, KS J (Biology) 3,000+, KS K (Textile) 15,000+, KS L (Ceramics) 7,000+, KS M (Chemistry) 12,000+, KS P (Medical) 5,000+, KS Q (Quality Mgmt) 4,000+, KS R (Transport) 12,000+, KS S (Service) 3,000+, KS T (Packaging) 4,000+, KS V (Shipbuilding) 5,000+, KS W (Aerospace) 3,000+ — totaling 220,000+ Korean Industrial Standards. Key Acts: Personal Information Protection Act (Act 19234, effective Sept 15, 2024), Electronic Government Act, Electronic Signature Act, Act on Promotion of Information and Communications Network Utilization and Information Protection, Information and Communications Infrastructure Protection Act, Data Industry Act, Public Data Act, AI Framework Act (Act 20212, effective July 2026), Industrial Technology Innovation Promotion Act, Framework Act on Science and Technology — 70+ Korean standardization-related laws.

Korea Digital Transformation Detailed Mapping

Korea operates digital transformation through a comprehensive governance system. Digital Government: Digital Platform Government Committee (established September 2022, under the President)·Ministry of the Interior and Safety Digital Government Bureau·e-Government Support Center·Gov.kr·National Citizen Service·KDIS (Korea Digital Information Society)·NIA (National Information Society Agency)·MOIS (Ministry of the Interior and Safety). K-DNS Infrastructure: Korea Internet & Security Agency (KISA) Korea Internet Center·KISA DNS Root Server·KRNIC (Korea Network Information Center)·BGP Korea·National Cyber Security Center (NCSC)·KCC (Korea Communications Commission)·MSIT (Ministry of Science and ICT)·NIA·NIPA. Korean Cloud Infrastructure: KT Cloud·NAVER Cloud (NCloud)·Samsung SDS Cloud·LG U+ Cloud·NHN Cloud·Kakao Enterprise Cloud·SK Telecom Cloud·KISA Cloud Security Assurance Program (CSAP)·KCMVP-validated cloud·ISMS-P (Information Security & Personal Information Management System). Korean Security Certifications: KISA ISMS-P certification·KCMVP (Korean Cryptographic Module Validation Program)·NIS (National Intelligence Service) "National Cryptographic Technology Operation Standards"·NCSC "National Cyber Security Strategy 2024-2028"·CC (Common Criteria) Korean evaluation bodies·EAL4·EAL5·KS X ISO/IEC 15408·19790·24759 Korean Profile. Korean Data Standards: NIA AI Hub·National Data Standardization Committee·Statistics Korea (KOSTAT)·MyData 4 Designated Combination Specialists (Samsung SDS, KICI, KOSTAT, KFTC)·National Institute of Korean Language·National Law Information Center·National Spatial Information Platform·National Spatial Data Center·Korean Spatial Information Standards. Finance and Fintech Standards: FSC (Financial Services Commission)·FSS (Financial Supervisory Service)·FIU (Financial Intelligence Unit)·BOK (Bank of Korea)·FSEC (Financial Security Institute)·KFTC (Korea Financial Telecommunications)·KSD (Korea Securities Depository)·KRX (Korea Exchange) 8-agency cooperation. 5G/6G Communications Infrastructure: 5G subscribers 35 million (2024)·5G base stations 350,000·6G commercialization target 2028·5G dedicated networks 16 operators·6G Acceleration Council (MSIT, 2024). K-Content: KOCCA (Korea Creative Content Agency)·MCST (Ministry of Culture, Sports and Tourism)·KCA (Korea Communications Agency)·Korea Culture Information Service Agency·Korean Film Archive·Korea Publishing Industry Promotion Agency. Data 3 Acts (Personal Information Protection Act·Credit Information Act·Telecommunications Network Act, 2020 enforcement)·Data Industry Act (2021)·Public Data Act (2013)·AI Framework Act (2026)·Digital Platform Government Framework Act (2024 proposed) — Korea digital transformation core legislation.

Korea Industrial, Research, Education Infrastructure Mapping

Korea operates its industrial ecosystem and standardization system through the following core infrastructure. Korea Top 5 Groups: Samsung, Hyundai Motor, LG, SK, Lotte. Each group operates standardization committees and ISO/IEC TC Korean secretariats. Samsung Electronics (semiconductors, displays, home appliances, telecom)·Hyundai Motor (automobiles, mobility)·LG Electronics (home appliances, displays, OLED)·SK hynix (memory)·LG Energy Solution·Samsung SDI (batteries)·POSCO Future M (materials)·Hyundai Mobis (parts). Korean IT Big Tech: NAVER (search, cloud, AI HyperCLOVA)·Kakao (messenger, payment, mobility, banking)·Coupang (e-commerce, logistics)·Karrot Market·Toss·Woowa Brothers. Korea Telcos: SK Telecom·KT·LG U+. 5G·5G dedicated networks·B2B cloud·AI businesses operating. Korea Top 7 Research Universities: Seoul National University·KAIST·POSTECH·Yonsei University·Korea University·UNIST·DGIST·GIST. All serve as standardization R&D bases and ISO/IEC/IEEE Korean chairs. Korea Government-affiliated National Research Institutes (26): KIST, KAERI, KIMM, KIER, KFRI, KRICT, KRIBB, KARI, KASI, KIGAM, KICT, KISTI, KETI, ETRI, NIMS, KIMS, KISDI, KOTRA, STEPI, KOEN, KICCE, KIET, KIPF, KIHASA, KICJ, KLRI. Korea Industrial Complexes / Tech Valleys: Pangyo Techno Valley·Dongtan·Gwanggyo·Songdo IBD·Yeouido·Gangnam·Sihwa·Banwol·Gumi·Ulsan·Changwon·Geoje·Yeosu·Onsan·Cheongju·Iksan·Gwangyang·POSCO Gwangyang Steel Mill·Asan Bay·Seosan·Songdo·Incheon Airport·Sejong·Cheongna·Geomdan. Korea Trade and Finance Infrastructure: Korea International Trade Association (KITA)·Korea Trade-Investment Promotion Agency (KOTRA)·Export-Import Bank of Korea (KEXIM)·Bank of Korea·Kookmin Bank·Shinhan·Hana·Woori·NH Nonghyup·IBK Industrial Bank·SC First Bank·Citi Bank Korea·HSBC Korea·DBS Korea — 14 Korean major banks and foreign banks. Korea K-POP / K-Content: HYBE·SM·YG·JYP 4 major entertainment companies·CJ ENM·tvN·MBC·KBS·SBS·EBS·YTN·Yonhap News TV·JTBC Korean broadcasting·NETFLIX Korea·Disney Plus·TVING·Wavve·Watcha·Coupang Play. Korea Gaming Industry: Nexon·NCsoft·Krafton·Netmarble·Kakao Games·Pearl Abyss·Com2uS·Gamevil·NHN·Smilegate·Webzen. Korea Automotive / Battery: Hyundai Motor·Kia·Genesis·LG Energy Solution·Samsung SDI·SK On·POSCO Future M·EcoPro·L&F battery cathode material suppliers. Korea Semiconductor: Samsung Electronics (HBM3E·HBM4)·SK hynix (HBM3E 12-Hi)·DB HiTek·SK siltron·SK Enpulse·Dongjin Semichem·Seoul Semiconductor·Simmtech·Samsung Display·LG Display.