Chapter 3: Data Format Specification - Alerts, Medications, and Activities

Building upon the foundational Elder Profile and Vital Signs data structures examined in the previous chapter, this chapter explores three additional critical data formats that complete the comprehensive data model for elder care systems: Alerts, Medications, and Activities. These structures enable the rich, actionable intelligence that transforms raw sensor data and measurements into meaningful care delivery. Alerts notify caregivers and healthcare providers when intervention is needed. Medication records ensure adherence to prescribed treatment regimens. Activity tracking provides holistic views of elder wellbeing beyond physiological measurements alone. Together with profiles and vital signs, these five data structures encompass the full spectrum of information flows in technology-enabled elder care.

The WIA-SENIOR-001 standard's approach to alerts, medications, and activities reflects the same design principles established for profiles and vital signs: comprehensiveness, precision, extensibility, cultural sensitivity, privacy-by-design, and standards alignment. Each structure balances the competing demands of rich functionality and practical implementability, ensuring that standardized formats serve real-world care scenarios without imposing unnecessary complexity. This chapter examines each data structure in detail, exploring the fields, validation rules, and design decisions that enable interoperable elder care systems.

The Alert Data Structure

Alerts represent the critical communication mechanism through which elder care systems notify stakeholders—family caregivers, professional care staff, healthcare providers, emergency services—of conditions requiring attention or intervention. Effective alert systems must balance sensitivity and specificity: generating alerts for all potentially serious conditions (high sensitivity) while minimizing false alarms that create alert fatigue and reduce response reliability (high specificity). The WIA-SENIOR-001 Alert data structure provides the framework for managing this balance through rich alert metadata, severity classification, actionable responses, and lifecycle tracking.

Alert Types and Classifications

The alert type field categorizes alerts into standard classes that enable consistent handling and appropriate routing. VITAL_SIGN_ABNORMAL alerts trigger when vital signs measurements fall outside normal ranges or individual baseline thresholds, for example sustained elevated blood pressure or abnormally low oxygen saturation. FALL_DETECTED alerts indicate that fall detection algorithms have identified a potential fall event, requiring immediate verification of elder safety. MEDICATION_MISSED alerts notify when scheduled medications were not taken within appropriate time windows, enabling intervention to ensure treatment adherence.

EMERGENCY_BUTTON alerts originate from manual activation of emergency buttons or panic alarms, always warranting immediate high-priority response. UNUSUAL_ACTIVITY alerts identify deviations from normal activity patterns that might indicate problems—for example, no movement detected during normally active hours could indicate a fall or medical emergency. DEVICE_OFFLINE alerts notify when critical monitoring devices lose connectivity, creating gaps in care coverage. LOW_BATTERY alerts enable proactive battery replacement before devices fail. GEOFENCE_BREACH alerts trigger when elders with cognitive impairment wander outside designated safe areas. TEMPERATURE_EXTREME alerts identify dangerous environmental temperatures. CUSTOM alerts allow system-specific alert types beyond the standard classifications.

Severity Levels and Prioritization

The severity field classifies alerts into four priority levels that guide response urgency and resource allocation. LOW severity alerts represent informational notifications or minor deviations that require awareness but not immediate action—for example, a single slightly low blood pressure reading in an elder with no history of hypotension. MEDIUM severity alerts indicate conditions warranting timely attention within several hours—for example, a missed medication dose or moderately abnormal vital sign that should be addressed but doesn't constitute an emergency.

HIGH severity alerts demand prompt response within minutes to hours, representing potentially serious conditions that could deteriorate—for example, sustained elevated heart rate, significantly abnormal blood pressure, or detection of unusual inactivity patterns. CRITICAL severity alerts require immediate emergency response, representing life-threatening conditions or situations where minutes matter—for example, fall detection, emergency button activation, or severely abnormal vital signs indicating possible cardiac events or respiratory distress.

Alert severity combines with alert type to enable sophisticated routing and escalation. Critical fall detection alerts might automatically notify emergency services and primary emergency contacts simultaneously. High severity vital sign abnormalities might notify family caregivers immediately while also scheduling nurse follow-up. Medium severity medication missed alerts might send reminders to the elder and notify caregivers if the medication isn't taken within an extended grace period. This multi-tier approach ensures appropriate response without overwhelming stakeholders with unnecessary high-priority notifications.

Alert Structure and Metadata

Each alert includes comprehensive metadata that provides context for appropriate response. The alert ID uniquely identifies each alert for tracking and lifecycle management. Elder ID references the affected elder. Timestamp records when the alert was generated. Title provides a human-readable brief description. Message contains detailed information about the alert condition and context. The optional data field can include type-specific structured information—for a vital sign abnormal alert, this might include the specific vital sign values; for a fall detected alert, it might include accelerometer data and confidence scores.

{
  "id": "alert-9876",
  "type": "VITAL_SIGN_ABNORMAL",
  "severity": "HIGH",
  "elderId": "elder-123",
  "title": "Elevated Heart Rate Detected",
  "message": "Heart rate has been elevated above 110 bpm for 15 minutes. Current reading: 118 bpm (irregular rhythm). Elder has no history of tachycardia.",
  "timestamp": "2025-01-15T14:30:00Z",
  "data": {
    "vitalSign": "heartRate",
    "value": 118,
    "threshold": 100,
    "duration": "15 minutes",
    "rhythm": "irregular",
    "previousReadings": [112, 115, 118, 116, 119]
  },
  "resolved": false,
  "resolvedAt": null,
  "resolvedBy": null,
  "actions": [
    {
      "id": "action-1",
      "label": "Contact Elder",
      "actionType": "call",
      "data": {"phoneNumber": "+1-555-0100"}
    },
    {
      "id": "action-2",
      "label": "Notify Physician",
      "actionType": "notify",
      "data": {"contactId": "physician-456"}
    },
    {
      "id": "action-3",
      "label": "Mark as False Alarm",
      "actionType": "resolve",
      "data": {"reason": "false-positive"}
    }
  ]
}

Alert Lifecycle Management

Alerts progress through a defined lifecycle from creation through resolution, with comprehensive tracking ensuring accountability and enabling quality improvement. The resolved boolean indicates whether the alert has been addressed. resolvedAt timestamp records when resolution occurred. resolvedBy identifies the user who resolved the alert, essential for audit trails and accountability. Resolution reasons can be captured to distinguish true positive alerts that required intervention from false positives that should inform alert algorithm refinement.

The actions array provides actionable response options tailored to each alert type and severity. Actions might include directly contacting the elder via phone or video call, notifying specific caregivers or healthcare providers, dispatching emergency services, marking the alert as acknowledged but not requiring intervention, or marking as a false positive. This action-oriented approach transforms alerts from passive notifications into interactive care coordination tools that facilitate appropriate response while capturing data about alert handling that enables continuous improvement.

The Medication Data Structure

Medication management represents one of the most impactful applications of elder care technology, addressing the serious public health problem of medication non-adherence. The WIA-SENIOR-001 Medication data structure captures comprehensive information about prescribed medications, dosing schedules, administration instructions, and adherence tracking. This structure integrates with medication dispensing devices, reminder systems, and electronic health records to create comprehensive medication management ecosystems.

Medication Identification and Dosing

Each medication record begins with unique identification and core medication information. The medication ID provides unique reference for tracking adherence and associating administration records. The name field captures the medication name—preferably using standardized names from RxNorm, the National Library of Medicine's nomenclature for clinical drugs, which enables interoperability with pharmacy systems and electronic health records. Dosage specifies the amount to be taken, including strength and formulation (for example, "10mg tablet" or "5ml liquid").

The frequency field uses standardized categories for common dosing schedules: once-daily, twice-daily, three-times-daily, four-times-daily, as-needed (PRN medications taken only when specific conditions occur), or custom for irregular schedules. The schedule array specifies exact administration times in 24-hour HH:MM format, enabling precise automated reminders. Start date and optional end date define the medication's active period—acute medications like antibiotics have defined end dates, while chronic disease medications may continue indefinitely.

{
  "id": "med-5432",
  "name": "Metformin",
  "rxNormCode": "860975",
  "dosage": "500mg tablet",
  "frequency": "twice-daily",
  "schedule": ["08:00", "20:00"],
  "startDate": "2025-01-01",
  "endDate": null,
  "instructions": "Take with meals to reduce stomach upset. Monitor blood glucose levels daily.",
  "prescribedBy": "Dr. Sarah Mitchell",
  "prescribedDate": "2025-01-01",
  "purpose": "Type 2 Diabetes Management",
  "sideEffects": ["Nausea", "Diarrhea", "Stomach upset"],
  "interactions": ["Alcohol - may increase risk of lactic acidosis"],
  "refillsRemaining": 3,
  "pharmacyInfo": {
    "name": "Springfield Pharmacy",
    "phone": "+1-555-0150",
    "rxNumber": "RX-789456"
  }
}

Administration Instructions and Clinical Information

The instructions field captures important administration guidance that affects medication effectiveness and safety. This might include timing relative to meals ("take with food" or "take on empty stomach"), special preparation instructions ("shake well before use"), administration route details ("apply to affected area" for topical medications), or monitoring requirements ("check blood pressure before taking"). Clear instructions improve adherence and reduce administration errors, particularly important for elders managing complex multi-drug regimens.

Prescribed-by information documents the prescribing clinician, essential for medication reconciliation and questions about therapy. Purpose explains why the medication was prescribed, helping elders understand the importance of adherence. Side effects lists known adverse effects to monitor for, enabling early detection of medication problems. Interactions documents known drug-drug, drug-food, or drug-condition interactions that require caution. This clinical context transforms medication records from simple administration schedules into comprehensive medication knowledge bases that support safe therapy management.

Medication Adherence Tracking

Medication records link to administration events that track actual medication taking behavior. Each time a medication is taken (or missed), a corresponding event is recorded with timestamp, medication ID, scheduled time, actual time taken, and administration status (taken, missed, taken-late). This granular tracking enables calculation of adherence metrics including overall adherence rate (percentage of prescribed doses actually taken), timing accuracy (percentage taken within appropriate time windows), patterns of non-adherence (which medications are missed most frequently, whether misses occur at specific times of day or days of week), and early identification of declining adherence that might indicate problems like side effects, confusion about regimen, or depression affecting self-care motivation.

Integration with automated medication dispensers enables passive adherence monitoring—devices detect when compartments are opened and doses dispensed, automatically recording administration events without requiring manual logging. Smart pill bottles with opening sensors provide similar passive monitoring for medications not suitable for automated dispensers. For medications without automated monitoring, mobile applications or voice assistants can prompt manual confirmation of administration. This multi-modal approach balances comprehensive tracking with practical implementation across diverse medication types and elder capabilities.

The Activity Data Structure

While vital signs provide objective physiological measurements, activity tracking captures behavioral and functional information that reveals crucial dimensions of elder wellbeing not visible in medical measurements alone. Activities of daily living (ADLs) including eating, bathing, dressing, toileting, and mobility represent core functional capabilities that determine independence. Social activities combat isolation and loneliness. Exercise and physical activity affect both physical and mental health. Sleep quality influences numerous health outcomes. The WIA-SENIOR-001 Activity data structure provides standardized formats for recording and analyzing these diverse activity types.

Activity Types and Classification

The activity type field categorizes activities using standard classifications that enable meaningful analysis and comparison. SLEEP activities track rest periods, with start and end times enabling calculation of sleep duration, timing, and patterns. MEAL activities record eating events, important for monitoring nutrition, detecting appetite changes, and managing diabetic meal timing. MEDICATION activities link to medication records, documenting administration events. EXERCISE activities capture physical activity from formal exercise sessions to casual walking, with duration and optional intensity measures.

SOCIAL activities document social engagement including in-person visits, phone calls, video chats, and participation in group activities. BATHROOM activities track toileting frequency, which can reveal urinary tract infections, incontinence issues, or medication side effects. SHOWER and other hygiene activities indicate self-care capability and detect declines in personal care that might signal depression or cognitive impairment. ENTERTAINMENT activities including reading, watching television, hobbies, and games provide baseline activity patterns against which unusual inactivity can be detected. MEDICAL_APPOINTMENT and THERAPY activities document healthcare engagement. OTHER captures activities not fitting standard categories.

{
  "id": "activity-8765",
  "elderId": "elder-123",
  "type": "EXERCISE",
  "subtype": "Walking",
  "startTime": "2025-01-15T09:00:00Z",
  "endTime": "2025-01-15T09:30:00Z",
  "duration": 1800,
  "location": "Neighborhood park",
  "intensity": "moderate",
  "metrics": {
    "distance": "1.2 miles",
    "steps": 2400,
    "averageHeartRate": 95,
    "calories": 120
  },
  "notes": "Walked with neighbor Jane. Weather was pleasant. No discomfort or fatigue.",
  "completedBy": "self",
  "assistanceRequired": false,
  "mood": "positive",
  "deviceSources": ["fitness-tracker-001", "smartphone-gps"]
}

Temporal and Contextual Information

Activity records capture rich temporal and contextual metadata that enables sophisticated analysis. Start time and optional end time define the activity period, with duration calculated automatically or recorded directly for activities where precise timing isn't captured. Location specifies where the activity occurred, valuable for understanding elder mobility patterns and detecting whether elders are venturing outside the home. Notes provide free-text space for recording qualitative observations, details, or context not captured in structured fields.

Completed-by indicates who participated in or performed the activity—"self" for activities the elder completed independently, specific caregiver identifiers for assisted activities, or "group" for social group activities. Assistance-required boolean flags activities where help was needed, enabling tracking of functional independence over time. Mood captures affective state during or immediately after the activity, providing mental health insights. Device sources documents which sensors, devices, or applications generated the activity record, important for data quality assessment and sensor fusion when multiple devices observe the same activity.

Activity Pattern Analysis

Individual activity records become most valuable when aggregated and analyzed for patterns that reveal health status and trends. Normal activity pattern baselines established through machine learning on historical data enable detection of deviations that might indicate problems. Significantly reduced activity levels might signal depression, illness, pain, or medication side effects. Increased nighttime bathroom activity could indicate urinary tract infections, prostate problems, or poorly controlled diabetes. Declining social activity suggests increasing isolation requiring intervention.

Changes in sleep patterns—difficulty falling asleep, frequent nighttime waking, or excessive daytime sleep—can indicate pain, anxiety, sleep disorders, or medication effects. Reduced meal activity or changes in eating patterns might reveal appetite loss from illness or depression. Decreased shower and hygiene activities often represent early indicators of functional decline or cognitive impairment. Activity-based algorithms can generate alerts when patterns deviate significantly from baselines, enabling proactive intervention before problems become severe.

Data Integration and Cross-Structure Analytics

While each data structure serves specific purposes, their true power emerges through integration and combined analysis. Vital signs, alerts, medications, and activities together create comprehensive multi-dimensional pictures of elder health and wellbeing that enable sophisticated care delivery impossible with any single data stream alone.

Correlating Vital Signs with Activities

Combining vital signs and activity data reveals important relationships and contexts. Elevated heart rate during exercise activities is expected and normal, while the same elevation during sleep might indicate cardiac problems. Blood pressure measured during stressful activities requires different interpretation than resting measurements. Blood glucose patterns analyzed relative to meal timing enables fine-tuning of diabetes management. Sleep activity data combined with overnight vital signs can detect sleep apnea through patterns of oxygen desaturation coinciding with sleep periods.

Medication Effects on Vital Signs and Activities

Correlating medication records with vital signs and activities enables monitoring for therapeutic effects and adverse reactions. Blood pressure vital signs analyzed relative to antihypertensive medication timing reveals medication effectiveness and appropriate dosing. Heart rate changes after starting new cardiac medications should be monitored. Activity level changes after initiating antidepressant therapy can indicate treatment response. Unusual vital sign patterns or activities coinciding with medication changes suggest possible adverse effects requiring clinical review.

Alert Generation from Combined Data Streams

The most sophisticated alerts combine information across multiple data structures. An alert algorithm might generate a fall detection alert only when accelerometer data indicates sudden impact AND subsequent activity tracking shows prolonged immobility AND the elder doesn't dismiss an automatic "Are you okay?" prompt within a timeout period. Medication adherence alerts can be suppressed if activity tracking shows the elder is away from home during scheduled medication times, automatically rescheduling reminders appropriately. Vital sign abnormal alerts can be contextualized with recent activities—elevated heart rate immediately after exercise is less concerning than the same reading during sedentary periods.