askill
care-pathway-summarization

care-pathway-summarizationSafety 95Repository

Summarize patient care journeys across encounters into concise clinical pathway narratives with key milestones, transitions, and outcome tracking. Use when reviewing longitudinal patient records, preparing care coordination summaries, generating transition-of-care documents, or analyzing patient journeys across multiple providers and settings.

0 stars
1.2k downloads
Updated 2/7/2026

Package Files

Loading files...
SKILL.md

Care Pathway Summarization

Overview

Synthesize longitudinal patient records spanning multiple encounters, providers, and care settings into concise, clinically actionable pathway summaries. This skill traces the patient journey from initial presentation through diagnosis, treatment, and outcomes — highlighting key decision points, care transitions, complications, and adherence to evidence-based protocols.

When to Use

  • Preparing transition-of-care or handoff summaries
  • Reviewing patient journeys for care coordination meetings
  • Generating case summaries for utilization review or peer review
  • Analyzing treatment trajectories for quality improvement
  • Building patient timeline visualizations for clinical dashboards
  • Supporting continuity of care across provider changes

Required Inputs

InputDescriptionFormat
Patient encounter historyChronological list of encounters with notesArray of encounter objects
Problem listActive and resolved conditionsICD-10 coded list
Medication historyCurrent and historical medicationsRxNorm-coded list with dates
Care setting contextPrimary care, specialty, hospital, post-acuteEnum string
Summary purposeHandoff, UR, quality review, patient-facingEnum string

Methodology

Step 1: Temporal Encounter Mapping

  1. Order all encounters chronologically
  2. Classify each encounter by type: ambulatory, inpatient, ED, observation, post-acute, telehealth
  3. Identify care episodes by grouping related encounters (e.g., surgery then post-op visits then rehab)
  4. Map care transitions: setting changes, provider handoffs, level-of-care changes

Step 2: Clinical Thread Extraction

For each active problem, trace its thread through the encounter history:

  • Onset/Presentation: When and how the condition first appeared
  • Diagnostic workup: Tests ordered, results, differential narrowing
  • Treatment initiation: First-line therapy, rationale, response
  • Treatment modifications: Escalations, switches, adverse reactions
  • Current status: Stable, improving, worsening, resolved

Step 3: Key Milestone Identification

Extract significant clinical events:

Milestone Categories:

  • Diagnosis confirmed (with supporting evidence)
  • Major procedure or intervention
  • Hospitalization (admission and discharge)
  • Medication change (new, discontinued, dose adjustment)
  • Adverse event or complication
  • Care setting transition
  • Specialist referral and outcome
  • Goals-of-care discussion
  • Disease progression or remission marker

Step 4: Pathway Quality Assessment

Evaluate the care pathway against applicable standards:

  • Guideline adherence: Were evidence-based protocols followed?
  • Timeliness: Were interventions delivered within recommended timeframes?
  • Care gaps: Were recommended screenings, tests, or follow-ups missed?
  • Coordination quality: Were transitions smooth with appropriate information transfer?

Step 5: Summary Generation

Produce a structured summary tailored to the stated purpose:

  • Clinical handoff: Focus on active problems, pending items, anticipated needs
  • Utilization review: Emphasize medical necessity, level-of-care appropriateness
  • Quality review: Highlight guideline adherence, outcomes, variance analysis
  • Patient-facing: Plain language, key dates, action items, medication list

Output Specification

The structured output includes:

patient_identifier: MRN or de-identified ID

summary_period: start date and end date

encounter_timeline: total_encounters, by_type (inpatient, outpatient, ed, telehealth), care_episodes (episode_id, primary_condition with description and icd10, encounters with date/type/provider/setting/key_actions, status as active/resolved/ongoing)

clinical_threads: condition (description, icd10), onset_date, thread_summary narrative, milestones (date, event, significance, details), current_status, current_treatment

key_milestones: date, event description, category, clinical_significance (high/medium/low)

active_problems: condition (description, icd10), since date, current_management, pending_actions

medications: current (name, dose, indication, start_date), recently_changed (name, change, date, reason)

care_quality: guideline_adherence (guideline, status, details), care_gaps (gap, recommendation, priority), transition_quality (score 0-100, issues)

summary_narrative: concise prose summary

pending_items: item, owner, due_date, priority

Analysis Framework

Care Episode Classification

Episode TypeDefinitionTypical Duration
AcuteSingle illness or injury eventDays to weeks
Chronic managementOngoing condition monitoringMonths to years
SurgicalPre-op through post-op recoveryWeeks to months
PreventiveScreening and wellnessAnnual or per-guideline
Palliative/HospiceComfort-focused careVariable

Transition Quality Scoring

Evaluate each care transition on:

  1. Information transfer (0-25): Was a complete summary transmitted?
  2. Medication reconciliation (0-25): Were medications reconciled at transition?
  3. Follow-up plan (0-25): Were follow-up appointments scheduled?
  4. Patient understanding (0-25): Was teach-back or patient education documented?

Examples

Input: 12 encounters over 6 months for a patient with new-onset heart failure.

Summary Output (abbreviated):

  • Episode: Heart failure diagnosis and management (Oct 2025 to Mar 2026)
  • Milestones: ED visit with dyspnea (Oct 5) > Echo showing EF 30% (Oct 6) > Cardiology consult (Oct 8) > Started on sacubitril/valsartan + carvedilol (Oct 10) > 30-day follow-up showing improvement (Nov 8) > Repeat echo EF 38% (Jan 2026) > ICD evaluation (Feb 2026)
  • Guideline adherence: ACC/AHA HFrEF pathway — GDMT initiated appropriately, ICD evaluation per guidelines
  • Gap: Cardiac rehab referral not documented

Guidelines

  1. Maintain chronological integrity — present events in temporal order within each thread
  2. Distinguish documented facts from inferences — clearly label any inferred connections
  3. Prioritize clinically significant events — omit routine stable follow-ups unless they mark a change
  4. Tailor language to audience — clinical terminology for clinicians, plain language for patients
  5. Highlight pending and unresolved items prominently — these drive the next action

Validation Checklist

  • All encounters in the input appear in the timeline
  • Clinical threads trace each active problem from onset to current status
  • Milestones are ordered chronologically with accurate dates
  • Medication list reflects current state with recent changes highlighted
  • No encounter or clinical event is attributed to the wrong date or provider
  • Summary purpose drives the appropriate level of detail and language
  • Pending items are actionable with clear ownership

HIPAA Compliance Notes

  • Patient identifiers must be handled per minimum necessary standard
  • When generating patient-facing summaries, exclude sensitive diagnoses per 42 CFR Part 2 (substance use) and state-specific regulations
  • De-identify summaries used for quality review or analytics unless a valid authorization or permitted use applies
  • Maintain audit trails for all summary generation activities involving PHI
  • Apply role-based access controls — not all summary types should be accessible to all roles

Install

Download ZIP
Requires askill CLI v1.0+

AI Quality Score

95/100Analyzed 2/10/2026

A comprehensive and highly structured skill for clinical data synthesis. It provides clear inputs, a multi-step methodology, detailed output schemas, and critical safety/compliance considerations.

95
98
90
98
95

Metadata

Licenseunknown
Version-
Updated2/7/2026
Publisherwassemgtk

Tags

observability