Context and Problem
Chronic care management depends on patient adherence to care plans (medication schedules, lifestyle changes, monitoring routines). When patients fall off track, outcomes worsen and costs rise. But care teams often lack visibility into patient progress between appointments, which means interventions happen too late or not at all.
[Company] needed a system where patients could set goals, track progress, and surface status updates to their care team. The core problems:
- Patients set goals but did not report progress consistently
- Care coordinators had no real-time view of who was struggling vs succeeding
- Intervention triggers were manual (weekly reviews, chart hunting)
- Dashboards showed history but did not highlight actionable next steps
Users and Constraints
Primary users
- Patients (setting goals, logging progress, viewing trends)
- Care coordinators (monitoring panels, prioritizing outreach, documenting interventions)
- Clinicians (reviewing status, adjusting care plans)
Constraints
- Wide range of digital literacy
- Care coordinators managing 50 to 100+ patients each
- Needed to integrate with existing patient records and care workflows
- HIPAA-compliant data handling required
What I Shipped
Patient Goal Setting
- Simple goal setup flow (goal type, target, check-in cadence)
- Optional milestone triggers (example: alert care team after missed check-ins)
Progress Tracking and Status Reporting
- Patient check-ins via mobile and SMS
- Progress visuals designed for clarity, not complexity
- Status summaries surfaced to care team for fast review
Intervention Trigger System
- Rules-based alerts for drift (example: missed consecutive check-ins, declining trends)
- Dashboard surfaced alerts with recommended next actions
- Escalation logic when outreach did not get a response within [X days]
Care Coordinator Dashboard
- Prioritized list of patients needing attention
- One-click actions: message, log outreach, snooze with reason
- Drilldown to history when needed
Clinician Summary View
- Fast visibility into goal status during patient touchpoints
- Exportable summaries for clinical documentation
How AI Showed Up
AI was used as an internal accelerator for analysis and workflow design, with strict constraints around compliance and human oversight. Where we prototyped AI-assisted features, outputs remained draft-only and required human confirmation.
- Intervention recommendations based on historical patterns (care coordinators could accept, modify, or ignore)
- Progress summarization using NLP (plain-language summaries for review)
Guardrails: no automated outreach, auditable recommendations, opt-out supported.
Outcomes
- [X]% lift in patient adherence (measured by check-in completion rate)
- [Y] days earlier intervention timing (at-risk patients contacted sooner on average)
- [Z]% reduction in manual triage time (alert system replaced weekly chart reviews)
- [N] patients enrolled across [M] practices
- [X]% of clinicians reported improved visibility into patient progress (post-launch survey)
My Role
What I owned
- Product strategy, scoped around adoption and measurable outcomes
- Requirements and workflows (journeys, triggers, escalation rules)
- SQL analysis to identify drop-off patterns and tune thresholds
- User research with care coordinators and patients
- Dashboard design direction focused on prioritization and action
What I collaborated on
- Engineering on trigger architecture and data pipelines
- Clinical advisors to validate intervention logic
- Client success on onboarding and training
What I Learned and What I Would Do Next
- Simplicity drives adoption: 1 to 2 clear goals outperform complex setups.
- Prioritization beats completeness: teams need the top patients to contact today.
- Timing matters: earlier outreach often beats changing outreach method.
Next: personalized reminder timing, goal templates, and proactive risk scoring to flag disengagement earlier.