Personalized Risk Prediction Tool for Deceased Donor Kidney Offers: Stakeholder Perspectives from a Qualitative Study

This qualitative study involving transplant patients, coordinators, and providers at the University of New Mexico reveals that a personalized kidney risk prediction tool must prioritize plain language explanations, contextual education, and workflow-aligned design to effectively support decision-making and potentially reduce organ discard rates.

The Gist

Imagine you are standing at a crossroads. On one path lies a long, exhausting journey on a dialysis machine. On the other, a new life with a kidney transplant. But here's the catch: the signpost pointing to the new life is covered in fog. You can see a kidney is available, but you don't know if it's a "golden ticket" or a "risky gamble."

This is the daily reality for kidney transplant patients and their doctors. Every day, thousands of kidneys are offered, and thousands are thrown away (discarded) because no one is sure if they are safe to use.

This paper is about a team of researchers who tried to build a GPS for this foggy journey. They created a prototype app called the "Kidney Risk Calculator" and asked the people who actually walk this path—patients, nurses, and doctors—what they thought about it.

Here is the story of their findings, explained simply.

The Problem: The "One-Size-Fits-All" Map Doesn't Work

Right now, doctors use standard scores (like KDPI) to judge kidneys. Think of these scores like a weather forecast for a whole country. They might tell you, "It's generally rainy in the Midwest," but they can't tell you specifically if it's going to rain on your head at your house.

These old scores look only at the donor (the person giving the kidney). They don't account for the receiver (the patient). A kidney that might be "risky" for a young, healthy person could be a "lifesaver" for someone who has been on dialysis for ten years and is running out of time.

The Solution: A Personalized Navigation App

The researchers built a digital tool that acts like a personalized navigation system. Instead of just saying "This kidney is 70% risky," it asks: "Who is getting this kidney?" and "How long have they been waiting?" Then, it gives a specific prediction: "For you, this kidney has an 85% chance of working well for 5 years."

What the Users Said: The "Driver's Manual"

The researchers didn't just build the app and walk away. They sat down with 13 people (patients, nurses, and doctors) to see how the app felt to use. Here is what they learned, using some simple metaphors:

1. The Content: "Stop Speaking 'Robot'"

• The Issue: The app was full of medical jargon and scary acronyms. It was like reading a car manual written in a language you don't speak.
• The Fix: Users wanted plain English.
  • The "Hep C" Confusion: Many patients were terrified of kidneys from donors with Hepatitis C, thinking they would get sick. The app needed to explain that modern medicine can cure this easily, like explaining that a "rusty car" can be fixed with a new engine.
  • The "Time" Factor: Patients felt the app didn't capture their urgency. One patient said, "I'm running out of time." The app needs to show the trade-off: "If you wait for a perfect kidney, you might lose your chance entirely. If you take this 'imperfect' one now, you might live longer."

2. The Format: "Make it a Story, Not a Spreadsheet"

• The Issue: The app showed numbers like "0.85" or "85.4%." To a stressed patient, this is confusing. It's like being told your car has a "0.85 probability of starting" instead of "Your car will likely start."
• The Fix:
  • Percentages, Not Decimals: People understand "25% chance" better than "0.25 probability."
  • Step-by-Step: Patients wanted the app to feel like filling out a simple tax form (like TurboTax), where you answer one question at a time, rather than staring at a giant wall of data.
  • Visuals: They wanted simple pictures, not complex charts.

3. The Functionality: "The Rush Hour Dilemma"

• The Issue: Kidney offers happen fast. A doctor gets a call, has 10 minutes to decide, and calls the patient. The patient might be 5 hours away. They can't sit down and study an app for an hour.
• The Fix: The app needs to be a quick reference guide, not a textbook.
  • Confidentiality: Doctors can't tell patients everything about the donor (like their exact age or medical history) due to privacy laws. The app needs to give the summary of the risk without revealing the donor's secrets.
  • Decision Aid, Not Decision Maker: Everyone agreed the app shouldn't make the choice for them. It's like a co-pilot. It points out the potholes and the shortcuts, but the driver (the patient and doctor) still steers the car.

The Big Takeaway

The study concluded that this tool has huge potential to stop kidneys from being wasted. If patients and doctors can see clearly that a "risky" kidney is actually a great match for a specific person, they will be more willing to say "Yes."

In a nutshell:
The researchers built a tool to turn a foggy, scary decision into a clear, personalized map. The users loved the idea but said, "Make it simpler, make it faster, and speak our language." If they fix these things, this app could help save thousands of kidneys and give thousands of people a second chance at life.

Technical Summary

1. Problem Statement

Despite kidney transplantation being the superior treatment for End-Stage Kidney Disease (ESKD), the supply of donor kidneys remains insufficient, leading to rising discard rates (nearly 25% in 2022). A primary driver of this is the uncertainty providers and patients face when evaluating "higher-risk" or "marginal" donor kidneys.

• Limitations of Current Tools: Standard indices like the Kidney Donor Risk Index (KDRI) and Kidney Donor Profile Index (KDPI) provide population-level risk assessments but fail to incorporate recipient-specific factors or generate individualized survival estimates.
• Decision-Making Gap: There is a lack of decision-support tools that integrate donor and recipient data to provide cognitively accessible, real-time risk projections during the high-pressure, time-sensitive window of a kidney offer.
• Need: A tool is required to bridge the gap between complex statistical models and the practical, shared decision-making needs of patients, coordinators, and providers, particularly regarding complex offers (e.g., Hepatitis C positive donors).

2. Methodology

The study employed a qualitative, user-centered design approach to evaluate a prototype web-based "Kidney Risk Calculator" app.

• Study Design: Formative qualitative study using focus groups and semi-structured individual interviews.
• Setting: University of New Mexico Hospital (UNMH) Kidney Transplant Center.
• Participants (N=13):
  • Patients (n=5): 4 transplant candidates and 1 patient advocate (stratified by sex, race/ethnicity, and dialysis status).
  • Coordinators (n=3): Transplant coordinators involved in offer evaluation.
  • Providers (n=5): 3 attending physicians and 2 advanced practice practitioners.
• Data Collection: Sessions (45–60 mins) were conducted via Zoom between Dec 2022 and May 2023. Participants viewed a live demonstration of the prototype app.
• Analysis: Researchers used inductive reflexive thematic analysis. Two analysts independently reviewed transcripts, developed a codebook, and coded data in NVivo. Inter-rater reliability was not calculated, consistent with reflexive thematic analysis principles. Matrix queries were used to visualize code distribution across stakeholder groups.

3. Key Contributions

The study contributes to the field of transplant medicine and health informatics by:

1. Defining User Requirements: Identifying specific content, format, and functionality needs for a kidney offer decision-support tool from the perspectives of all three key stakeholder groups.
2. Highlighting Workflow Constraints: Emphasizing that decision support tools must function within the unique constraints of kidney offers (phone-based, time-pressured, strict donor confidentiality).
3. Reframing Risk Communication: Proposing a shift from population-level indices to personalized, plain-language narratives that address specific misconceptions (e.g., Hepatitis C risks in the Direct-Acting Antiviral era).
4. Equity-Oriented Design: Providing a framework for designing tools that accommodate varying levels of health literacy, numeracy, and digital access (e.g., rural populations).

4. Key Results

Stakeholders affirmed the value of personalized projections but identified critical areas for refinement across three domains:

A. Content (What to show)

• Education: High demand for plain-language explanations of Hepatitis C (HCV) risks/benefits (specifically that HCV+ kidneys have excellent outcomes with DAAs) and Public Health Service (PHS) risk criteria to reduce refusal based on fear of infection.
• Contextual Variables: Participants requested inclusion of variables often missing in standard indices, such as time on dialysis, dialysis compliance, smoking history, and specific donor metrics (e.g., last creatinine vs. average).
• Trade-offs: The tool must explicitly frame the trade-off between "time on dialysis" and "risk of graft failure/death."
• HCV Nuance: Some providers suggested HCV status might not need to be a negative weighting factor given modern treatment outcomes.

B. Format & Features (How to show it)

• Plain Language: Users strongly preferred narrative summaries (e.g., "25% survival rate") over raw decimals or dense charts.
• Visuals: Simple visuals, percentages (not decimals), and minimized acronyms (e.g., spelling out CPRA) were requested to reduce cognitive load.
• Input Flow: Patients preferred a stepwise, "TurboTax-like" input flow rather than a single complex form.
• Units: Support for U.S. customary units (inches/pounds) was critical for patient usability.

C. Functionality (How it fits the workflow)

• Decision Aid, Not Decision Maker: The tool must augment clinical judgment, not replace it.
• Confidentiality & Access: A major barrier is that patients often cannot access the app during the offer call because they live far away, lack internet, or cannot be given specific donor details (age, sex, history) due to confidentiality laws. The tool must function as an interpretive layer for providers to explain risks without revealing protected donor data.
• Mobile/Offline: Needs for mobile app availability and offline/low-bandwidth functionality were highlighted.

D. Barriers and Facilitators

• Barriers: Digital divide (rural access), health literacy/numeracy challenges, unpredictable clinical events (e.g., primary graft nonfunction), and strict confidentiality limits.
• Facilitators: The tool's potential to increase provider confidence in accepting higher-risk kidneys, reassure patients, and serve as an educational scaffold for post-visit review.

5. Significance and Conclusion

This study demonstrates that the success of a kidney offer decision-support tool depends less on predictive accuracy alone and more on clarity, context, and workflow alignment.

• Clinical Impact: By integrating individualized predictions with plain-language education, such tools can help reduce the "bypassing" of medically complex kidneys, potentially lowering discard rates and increasing transplant access for vulnerable populations.
• Implementation Strategy: Future iterations must prioritize:
  • Updated modeling reflecting current HCV outcomes.
  • Accessible outputs (percentages, narratives, printable views).
  • Deployment strategies that account for digital equity and the specific constraints of phone-based offer calls.
• Future Work: The authors call for prospective evaluation to assess the tool's impact on patient comprehension, communication quality, and actual organ acceptance rates across diverse settings.

The paper concludes that a well-designed, stakeholder-informed tool can transform the kidney offer process from a source of uncertainty into a structured, shared decision-making opportunity.