Self-Managed Chronic Care
Self-Managed Chronic Care
Self-Managed Chronic Care
Hospitals discharge patients with a plan, and assume someone at home will help carry it out. For adults living alone, no one does. I diagnosed where post-discharge care breaks down, then designed a non-clinical support ecosystem that takes on the caregiver's missing role: daily guidance, human reassurance, and practical help, for the first 30 days at home.
Hospitals discharge patients with a plan, and assume someone at home will help carry it out. For adults living alone, no one does. I diagnosed where post-discharge care breaks down, then designed a non-clinical support ecosystem that takes on the caregiver's missing role: daily guidance, human reassurance, and practical help, for the first 30 days at home.
Hospitals discharge patients with a plan, and assume someone at home will help carry it out. For adults living alone, no one does. I diagnosed where post-discharge care breaks down, then designed a non-clinical support ecosystem that takes on the caregiver's missing role: daily guidance, human reassurance, and practical help, for the first 30 days at home.
The gap
$52.4B a year lost to 30-day readmissions in the US
$52.4B a year lost to 30-day readmissions in the US
The service
3 roles that replace the missing caregiver: guidance, reassurance, and logistics
3 roles that replace the missing caregiver: guidance, reassurance, and logistics
Projected impact
38% → 85%+ follow-ups kept ·
<18% readmissions
Pilot targets
38% → 85%+ follow-ups kept ·
<18% readmissions
Pilot targets
Proof
Tested in a 1-week Wizard-of-Oz pilot · Winner, European Product Design Award 2026
Tested in a 1-week Wizard-of-Oz pilot · Winner, European Product Design Award 2026
Role
Service Designer & Researcher
Service Designer & Researcher
Service Designer & Researcher
Advisor
Prof. Silke Bochat, SCAD
Prof. Silke Bochat, SCAD
Prof. Silke Bochat, SCAD
Duration
6 Months
6 Months
6 Months
Focus cohort
Adults 65+, living alone, heart failure and/or type 2 diabetes, first 30 days after discharge
Adults 65+, living alone, heart failure and/or type 2 diabetes, first 30 days after discharge
Adults 65+, living alone, heart failure and/or type 2 diabetes, first 30 days after discharge
My Role
My Role
Systems diagnosis
Systems diagnosis
Mapped patients, clinicians, pharmacies, community, and policy to find where continuity breaks after discharge, and why.
Mapped patients, clinicians, pharmacies, community, and policy to find where continuity breaks after discharge, and why.
Mapped patients, clinicians, pharmacies, community, and policy to find where continuity breaks after discharge, and why.
Experiential research
Experiential research
Ran a 12-hour bodystorming simulation to feel the coordination burden of recovering alone, not just read about it.
Ran a 12-hour bodystorming simulation to feel the coordination burden of recovering alone, not just read about it.
Ran a 12-hour bodystorming simulation to feel the coordination burden of recovering alone, not just read about it.
Service architecture
Service architecture
Designed three roles, each with a clear job, clear limits, and a safe path to escalate.
Designed three roles, each with a clear job, clear limits, and a safe path to escalate.
Designed three roles, each with a clear job, clear limits, and a safe path to escalate.
Service prototyping
Service prototyping
Built and ran a Wizard-of-Oz pilot, playing the human helpline myself, to test the service before building it.
Built and ran a Wizard-of-Oz pilot, playing the human helpline myself, to test the service before building it.
Built and ran a Wizard-of-Oz pilot, playing the human helpline myself, to test the service before building it.
Design for implementation
Design for implementation
Wrote role specifications, service levels, privacy guardrails, a regulatory stance, and a phased pilot roadmap.
Wrote role specifications, service levels, privacy guardrails, a regulatory stance, and a phased pilot roadmap.
Wrote role specifications, service levels, privacy guardrails, a regulatory stance, and a phased pilot roadmap.
The Problem
The Problem
The riskiest part of chronic care isn't in the hospital. It's the first 30 days after leaving it.
The cost
$52.4B a year, the cost of 30-day readmissions in the US
FY 2025 Cost to Hospitals
78.6% of hospitals still faced readmission penalties in FY2025, totaling $320M
FY 2025 Cost to Hospitals
25% Isolated patients return to the ER within 30 days compared to 15% of those in supported living
For hospitals, every avoidable readmission costs three ways:
Loss of Life
Loss of Life
Leads to 7 - 23% mortality rate depending upon the length of follow-up.
Leads to 7 - 23% mortality rate depending upon the length of follow-up.
Loss of Reputation
Loss of Reputation
Hospitals lose patient's trust and public ratings.
Hospitals lose patient's trust and public ratings.
Financial Penalties
Financial Penalties
~1% Deduction as penalty, reducing Medicare fee-for-service payments
~1% Deduction as penalty, reducing Medicare fee-for-service payments
While 1% Doesn't sound much, as of September 2025 this is a
While 1% Doesn't sound much, as of September 2025 this is a
While 1% Doesn't sound much, as of September 2025 this is a
For patients, the cause is simpler
A discharge plan assumes a caregiver at home to manage medications, spot warning signs, book follow-ups, and answer "is this normal?" For adults living alone, that caregiver doesn't exist, so the patient absorbs all of that coordination work, usually while recovering, often under cognitive load.
Scoping the Project
Scoping the Project
Started broad: People living chronic illness + living alone
Started broad: People living chronic illness + living alone
Found during research: Post-discharge is the highest-stakes window
Found during research: Post-discharge is the highest-stakes window
Narrowed to: Adults 65+ living alone, first 30 days post-discharge (patient harm + hospital penalty)
Narrowed to: Adults 65+ living alone, first 30 days post-discharge (patient harm + hospital penalty)
Problem Statement
Problem Statement
In the U.S., adults over the age of 65 with a Heart condition and/or type 2 Diabetes who lack a regular in-home caregiver face a fragile first 30 days after hospital discharge:
they must secure meds, set up devices, spot warning signs, and make early follow-ups with fragmented support, which leads to confusion, missed care, and preventable ED returns.
The Research
The Research
Every research method was chosen to answer a specific question. Here's what we needed to know, how we found out, and what it changed.
Every research method was chosen to answer a specific question. Here's what we needed to know, how we found out, and what it changed.
01
01
What we needed to know
What we needed to know
Where does continuity break?
Where does continuity break?
Method
Method
Stakeholder interviews with patients, caregivers, clinicians, a pharmacist, and discharge staff (n=12) , stakeholder mapping, and journey maps of recovery with and without a caregiver.
I mapped everyone involved in post-discharge care, then removed one role: the at-home caregiver.
I mapped everyone involved in post-discharge care, then removed one role: the at-home caregiver.
What we learned
What we learned
The caregiver is the highest-impact connector in the whole system. Take them away and links break or weaken immediately, because their coordination work silently shifts to the patient. Recovery quality depends on whether someone is there.
The caregiver is the highest-impact connector in the whole system. Take them away and links break or weaken immediately, because their coordination work silently shifts to the patient. Recovery quality depends on whether someone is there.
02
02
What we needed to know
What we needed to know
What does it feel like to recover alone?
What does it feel like to recover alone?
Method
Method
A 12-hour bodystorming simulation (weighted constraints for reduced strength, simulated age-related vision loss, timed real-world errands), plus online ethnography in patient and caregiver communities.
A 12-hour bodystorming simulation (weighted constraints for reduced strength, simulated age-related vision loss, timed real-world errands), plus online ethnography in patient and caregiver communities.
A non-clinical, self-run simulation, designed to pressure-test coordination burden, not to replace lived experience.
What we learned
What we learned
Routine tasks become high-friction without support. Small barriers, unreadable labels, limited mobility, logistics, compound into cognitive load, making it harder to stay confident about what to do next.
03
03
What we needed to know
What we needed to know
Where is the opportunity?
Where is the opportunity?
Method
Method
Mapped the patient's emotional journey across 30 days against the involvement of hospital care teams, pharmacies, and community support.
Mapped the patient's emotional journey across 30 days against the involvement of hospital care teams, pharmacies, and community support.
What we learned
What we learned
Care is front-loaded, then drops off. Pharmacies are recurring but can't answer "is this normal?" Community help is real but unreliable. Support arrives in touchpoints, but recovery is lived in between, and that's where uncertainty turns into avoidable escalation. The opportunity wasn't another clinical tool. It was a low-acuity support layer for the space between touchpoints.
From finding to feature
From finding to feature
Research told us what was broken. This is what we built because of it.
Research told us what was broken. This is what we built because of it.
DISCHARGE PLANS DON'T BECOME DAILY ACTION
Kiko turns the plan into Today, This Week, and If-Then steps
MEDICAL TOOLS FEEL INTIMIDATING
A voice-first, pet-like companion, not a medical device
NO ONE TO ASK "IS THIS NORMAL?"
Medi-Mate, a non-clinical human helpline
UNCERTAINTY TURNS INTO AVOIDABLE ER VISITS
A 3-level escalation ladder
RIDES, REFILLS, AND GROCERIES FALL APART
A vetted volunteer network for small, non-clinical tasks
LIVING ALONE MEANS NO ONE CHECKS IN
A daily "I'm OK" from Kiko, plus volunteer check-ins
The Solution
The Solution
A three-part ecosystem that replaces the missing caregiver's key functions, for 30 days, without replacing medical care.
A three-part ecosystem that replaces the missing caregiver's key functions, for 30 days, without replacing medical care.
Kiko
Kiko
Guidance
A companion that turns the discharge plan into a daily routine: medications, follow-ups, diet, and when to call for help. Kiko stays happy when you take care of yourself, and looks worried when something's missed: motivation without shaming. Sensitive health data stays on the device.
A companion that turns the discharge plan into a daily routine: medications, follow-ups, diet, and when to call for help. Kiko stays happy when you take care of yourself, and looks worried when something's missed: motivation without shaming. Sensitive health data stays on the device.
Medi-Mate
Medi-Mate
Reassurance
A non-clinical human navigator and the patient's single "safe number" for the first 30 days. Answers "is this normal?", translates the plan into doable steps, and routes anything clinical to the right team.
A non-clinical human navigator and the patient's single "safe number" for the first 30 days. Answers "is this normal?", translates the plan into doable steps, and routes anything clinical to the right team.
Volunteer Network
Volunteer Network
Logistics
Vetted, trained community volunteers for rides, first-fill pickups, groceries, and check-ins. Volunteers never see health information and never do clinical work.
Vetted, trained community volunteers for rides, first-fill pickups, groceries, and check-ins. Volunteers never see health information and never do clinical work.
How they connect
Take care of yourself and Kiko stays happy, miss your medications or mess up your routine...he's going to be sad
Take care of yourself and Kiko stays happy, miss your medications or mess up your routine...he's going to be sad
Take care of yourself and Kiko stays happy, miss your medications or mess up your routine...he's going to be sad
Take care of yourself and Kiko stays happy, miss your medications or mess up your routine...he's going to be sad
Need a helping hand? The Volunteer Network is at your Service!
Need a helping hand? The Volunteer Network is at your Service!
Need a helping hand? The Volunteer Network is at your Service!
Need a helping hand? The Volunteer Network is at your Service!
Not sure if your symptoms need medical attention? No worries! The Medi-mate has it covered
Not sure if your symptoms need medical attention? No worries! The Medi-mate has it covered
Not sure if your symptoms need medical attention? No worries! The Medi-mate has it covered
Not sure if your symptoms need medical attention? No worries! The Medi-mate has it covered
A system to support, care and nourish
A system to support, care and nourish
A system to support, care and nourish
Safe by design
Every concern is routed by urgency
Every concern is routed by urgency
Question
Answer and log
Symptoms/medication confusion
Pharmacist or nurse, same-day callback.
Red flags or falls
Emergency services, care team notified
This is a non-clinical support layer. It supports clarity, coordination, and safe escalation, and never replaces medical care.
This is a non-clinical support layer. It supports clarity, coordination, and safe escalation, and never replaces medical care.
Validation
Validation
I built & tested the support ecosystem through scenario-based prototyping and early feasibility checks.
I built & tested the support ecosystem through scenario-based prototyping and early feasibility checks.
Kiko
Medi-Mate
Volunteers
Hypothesis:
Voice guidance boosts clarity + independence when reading is hard.
The Test:
Today / Week / If–Then plan + meds + follow-ups.
Method:
Voice GPT prototype tested (N=9) with post-discharge scenarios.
Outcome:
Higher confidence + clearer next steps; voice removed “small text” friction.
Projected Impact
Projected Impact
Where Patients are today and where they aim to be in 30 days
Where Patients are today and where they aim to be in 30 days
10–15% fewer
10–15% fewer
Unplanned ER visits
Today: ~18%
+1 point on a Likert scale.
+1 point Target, on a Likert scale.
Confidence to self-manage
Today: moderate (PAM-10, ~50–60 typical)
Value for each stakeholder
Patients get a safer 30 days with clear next steps. Hospitals get fewer readmissions and better patient experience scores. Payers reduce avoidable spend. Volunteers earn service hours and credits. Social work students get real practicum experience.
Deep Dive
Read the process book here
Read the process book here
What's Next
What's Next
Phase 1 : Mobile-first pilot
50 cardiology patients. Kiko as a mobile app, Medi-Mate as a human team. Goal: validate the service and prove the KPIs.
50 cardiology patients. Kiko as a mobile app, Medi-Mate as a human team. Goal: validate the service and prove the KPIs.
Phase 2 : Hardware rollout
Partner with a companion-robot maker. Kiko robot, Medi-Mate team, volunteer app. Goal: validate the technology at scale.
Partner with a companion-robot maker. Kiko robot, Medi-Mate team, volunteer app. Goal: validate the technology at scale.
Phase 3 : Full integratioN
Deep EHR integration. Goal: a new standard of post-discharge care, and system-wide readmission reduction.
Deep EHR integration. Goal: a new standard of post-discharge care, and system-wide readmission reduction.
What it takes to deploy
Start as a Class I wellness device, with FDA 510(k) clearance only if clinical decision support is added. HIPAA-aware design, on-device processing for health data, business associate agreements with partners, and volunteers kept fully outside health information.
Start as a Class I wellness device, with FDA 510(k) clearance only if clinical decision support is added. HIPAA-aware design, on-device processing for health data, business associate agreements with partners, and volunteers kept fully outside health information.
Reflection
Reflection
Recovery breaks in the mundane
The biggest risks weren't clinical. They were unreadable labels, a missed ride, a question with no one to ask.
The biggest risks weren't clinical. They were unreadable labels, a missed ride, a question with no one to ask.
Human reassurance is a feature, not a fallback
Automation works best for clarity and navigation. When someone is scared, they want a person.
Automation works best for clarity and navigation. When someone is scared, they want a person.
Community must augment the safety net, not become it.
Volunteers are powerful, but only if the service never depends on them for anything safety-critical.
Volunteers are powerful, but only if the service never depends on them for anything safety-critical.
Next time
I'd run the pilot with recently discharged patients from day one, and track time-to-answer and ER visits from the start, so the next phase can measure impact, not project it.
I'd run the pilot with recently discharged patients from day one, and track time-to-answer and ER visits from the start, so the next phase can measure impact, not project it.
