Create a patient-reported outcomes program: instruments, cadence, and actioning.
- User
Create a patient-reported outcomes program: instruments, cadence, and actioning.
Assistant
* User's Request: Create a patient-reported outcomes (PRO) program. The request asks for three specific components: 1. Instruments: Which questionnaires/surveys to use? 2. Cadence: How often should patients fill them out? 3. Actioning: What do we do with the data (alerts, workflows, clinical decision support)?
* **Context:** The user hasn't specified a specific disease area (e.g., oncology, cardiology, orthopedics) or setting (primary care, specialty clinic, research). Therefore, I need to provide a *framework* that is adaptable but specific enough to be useful. I should probably choose a hypothetical scenario or offer a modular approach to make it concrete. * **Target Audience:** Likely a healthcare administrator, clinical operations manager, or a clinician looking to implement a system. * **Structure of the Response:** 1. **Executive Summary/Philosophy:** Briefly explain *why* we are doing this (value-based care, symptom management). 2. **Phase 1: Selection of Instruments (The "What"):** Categorize them (Generic vs. Disease-Specific). Give examples. 3. **Phase 2: Cadence & Workflow (The "When"):** Define the timing based on care pathways. 4. **Phase 3: Actioning & Triage (The "How"):** This is the most critical part for clinical utility. Alerting, clinical review, population health. 5. **Technology/Implementation Considerations:** EHR integration, patient engagement. 6. **Example Use Case:** To tie it all together, I'll create a hypothetical example (e.g., Chronic Pain or Depression) or a general structure. Let's go with a Modular Framework that applies to a general population with chronic conditions. * **Drafting Content - Section by Section:** * **Instruments:** * *Generic:* Need a baseline health metric. PROMIS-10 or SF-36 are standards. PROMIS is better for modern EHR/integration. Let's suggest PROMIS Global Health. * *Disease-Specific:* I can't list every disease. I should mention the "Gold Standards" (e.g., PHQ-9 for depression, KOOS for knees, FACT-G for cancer). * *Experience:* Don't forget patient experience (eNPS). * **Cadence:** * Needs to be balanced. Too much = survey fatigue. Too little = missed deteriorations. * *Touchpoints:* New patient intake, pre-visit, post-intervention, routine monitoring (every 3-6 months). * **Actioning (The "So What?"):** * *Tier 1 (Red):* Immediate danger (Suicidality, severe pain). Action: Immediate alert to RN/MD. * *Tier 2 (Yellow):* Moderate issue. Action: Flag for discussion during visit. * *Tier 3 (Green):* Stable. Action: Continue monitoring, offer self-management resources. * *Feedback loops:* Show the patient their trends. That creates engagement. * **Refining the Output:** * Make it actionable. Use bullet points and bold text. * Add a section on "Governance" because a PRO program needs a team to run it. * Add a note on "Burden vs. Value." * **Self-Correction during drafting:** * *Initial thought:* List 50 different surveys. *Correction:* Too overwhelming. Categorize them into "Universal" vs. "Specific." * *Initial thought:* Just say "do it every month." *Correction:* That depends on stability. Suggest a dynamic cadence (stable patients do it less often, active patients do it more). * **Final Polish:** Ensure the tone is professional, clinical, and operational. Structure it as a "Program Design Document." * *(Self-Correction on specific instruments)*: Ensure I mention PROMIS (Patient-Reported Outcomes Measurement Information System) as it's the modern standard for interoperability. * *Actioning details:* Be specific about *who* receives the alerts. It shouldn't always be the doctor; often a nurse navigator or medical assistant is more appropriate for the initial triage. * Let's construct the response. **Structure:** 1. **Program Strategy** (The "Why") 2. **Component 1: Instruments** (The Toolkit) 3. **Component 2: Cadence** (The Schedule) 4. **Component 3: Actioning** (The Response Algorithms) 5. **Implementation Blueprint** (Tech & Governance) 6. **Case Study Example** (e.g., Musculoskeletal/Ortho) to make it concrete. * *Check:* Did I cover all three user requests? Yes: Instruments, Cadence, Actioning.</think>Creating a Patient-Reported Outcomes (PRO) program requires a balance between scientific rigor (valid data) and operational feasibility (not burdening the patient or staff).
Below is a comprehensive blueprint for designing a PRO program, adaptable for a specialty clinic, value-based care model, or clinical research.
1. Program Strategy & Scope
Before selecting surveys, define the goal.
- Clinical Goal: Symptom monitoring (e.g., chemo side effects) or Functional improvement (e.g., post-physical therapy).
- Population: Chronic care (long-term relationship) vs. Acute care (episodic).
- Platform: EHR-integrated (Epic, Cerner) vs. Standalone Digital Platform (REDCap, Qualtrics).
2. The Instruments (The "What")
Avoid "survey fatigue" by using the fewest questions necessary to capture the health domain. Use a Core + Modular approach.
A. The "Core" Instrument (Universal)
Used for every patient to allow for population health comparisons.
- Recommendation: PROMIS Global-10 (or PROMIS-29).
- Why: It creates a standardized "Health Score" (Physical and Mental) that can be compared across different disease states.
B. Disease-Specific Instruments (Targeted)
Select based on the clinical pathway.
Domain Gold Standard Instrument Best For Mental Health PHQ-9 (Depression) / GAD-7 (Anxiety) Primary Care, Integrated Behavioral Health Musculoskeletal KOOS (Knee) / HOOS (Hip) / ODI (Spine) Orthopedics, Sports Medicine Oncology PRO-CTCAE (Toxicity) / EORTC QLQ-C30 Cancer centers, Infusion clinics Chronic Disease CCQ (COPD) / MLHFQ (Heart Failure) Pulmonology, Cardiology General Satisfaction NPS (Net Promoter Score) Practice Management C. The "Functional" Layer
- Activity Tracking: Integrate passive data (steps per day) via Apple Health/Google Fit to corroborate survey data.
3. Cadence (The "When")
The frequency of surveys should match the volatility of the patient's condition.
Strategy: Trigger-Based Cadence
- Event-Based (The "Visit"): Sent 48 hours prior to every appointment.
- Logic: Ensures the provider has the data in the chart before the patient walks in.
- Interval-Based (The "Watch"): Sent during long gaps in care (e.g., every 4 weeks).
- Logic: Catches deterioration in chronic patients who aren't scheduling visits.
Specific Cadence Examples
- Pre-Procedure (e.g., Surgery):
- Baseline (T-0): Consultation.
- Pre-Op: 1 week before surgery.
- Post-Op: 2 weeks, 6 weeks, 3 months, 6 months, 1 year.
- Chronic Management (e.g., Depression/Anxiety):
- Every 2 weeks initially.
- Monthly once stable.
- Adaptive: If PHQ-9 score spikes, system auto-prompts for a weekly check-in.
4. Actioning (The "How")
Data without workflow is waste. You must build Alert Algorithms and Closed-Loop Workflows.
Tier 1: The "Red" Alert (Immediate Safety)
- Trigger: Specific questionnaire responses (e.g., "I have thoughts of killing myself" on PHQ-9, or Pain > 9/10 with specific keywords).
- Action:
- Automated: SMS/Email sent to patient: "A member of your care team will contact you shortly."
- Human: Alert routed immediately to a Triage Nurse or On-Call Provider via EHR in-basket or Pager.
- SLA (Service Level Agreement): Contact within 2 hours.
Tier 2: The "Yellow" Alert (Clinical Review)
- Trigger: Significant change in score (e.g., >5 point drop in PROMIS Physical Function or moderate increase in symptoms).
- Action:
- Automated: Flag placed on patient chart for next visit OR routed to a Nurse Navigator/Medical Assistant (MA).
- Human: Navigator calls patient to assess if a sooner appointment is needed or if medication adjustment is required.
Tier 3: The "Green" Zone (Maintenance)
- Trigger: Scores stable or improving.
- Action:
- Automated: "High-Five" message sent to patient ("Great progress!").
- Educational: Push relevant educational content (articles, videos) to the patient portal.
5. Implementation & Governance Plan
The "Feedback Loop" (Engagement)
To keep patients filling out forms, they must see value.
- The Visualization: After submission, the patient should see a graph of their progress over time on the portal.
- The Conversation: Providers must reference the data.
- Script: "I see your pain score dropped from an 8 to a 4 since we started this medication. Is that accurate?"
Measurement of Success (Program KPIs)
Do not just measure patient health; measure the program efficiency.
- Completion Rate: Target >65% of sent surveys completed.
- Alert Response Time: Average time for staff to respond to "Red" alerts.
- Clinical Correlation: Does the PRO score correlate with clinical outcomes (e.g., fewer ER visits)?
Summary Checklist for Launch
- Select Instruments: 1 Global + 1 Disease Specific.
- Build EHR Flowsheets: Ensure data flows into the patient chart, not just a separate database.
- Define Routing: Who gets the email when a bad score comes in? (MA, RN, or MD?)
- Patient opt-in: Explain why you are asking ("To help us treat you better between visits").
- Pilot: Test on 1 provider for 2 weeks before rolling out clinic-wide.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.