A diagnosis of kidney failure, or End Stage Renal Disease (ESRD), means that a patient's kidneys can no longer filter blood — an essential life function. In order to survive, people with ESRD have two choices: to receive ongoing dialysis treatment or get a kidney transplant. The average wait time to get a transplant is 3-5 years, though in some states in can be more than 10 years. Currently, there are about 100,000 people on a kidney transplant center wait list. Most patients live on dialysis while they wait for their new kidney.
Dialysis is life consuming. Most patients with kidney failure spend four hours every other day receiving treatment at a dialysis center. Not only does dialysis impact patients' quality of life, the United States government spends nearly $28 billion dollars each year on dialysis treatment. Medical research consistently shows that patients who receive a kidney transplant have better long-term health outcomes and improved quality of life over those on dialysis. It's also been shown that the government could save billions of dollars if more ESRD patients were to receive a transplant.
So why doesn't everyone get a transplant? While twenty thousand kidneys are donated from deceased donors each year, twenty thousand new patients are diagnosed with ESRD each year. And with a hundred thousand patients already waiting for a kidney, the demand continuously outpaces the supply. Because of this, eleven people die every day waiting for a kidney transplant in the United States. To make matters worse, of all donated kidneys, about 19% are discarded instead of transplanted into patients.
How might we decrease the discard rate to increase access to transplant for thousands of patients each year? Our design team was approached by a national health agency to conduct human-centered design with this goal in mind. Our aim was to uncover the root causes of kidney discard and ideate solutions. After months of interviews and co-creation with kidney transplant stakeholders, we created multiple solutions to the kidney discard problem.
Our investigation into the national kidney discard problem began with immersive design research in the form of interviews, ethnography, data analysis, and scientific literature reviews. We studied emerging evidence from leading researchers, conducted a data analysis of existing reasons for discard and offer refusal, and interviewed dozens of transplant surgeons, nephrologists, transplant coordinators, organ procurement organizations, epidemiologists, kidney transplant thought-leaders, and patients — over 90 interviewees across the States. We sat in the mission control room of an organ procurement organization as organ donation calls came pouring in from donor hospitals. We watched couriers pack "HUMAN ORGAN" coolers onto trucks headed for transplant centers. We observed serology lab technicians testing kidneys for infectious diseases. We walked through an organ procurement organization's hall of remembrance which honored organ donors whose death meant life for someone else. We sat with transplant surgeons and their teams as they received organ offers; one surgeon's phone rang every twenty minutes — she hadn't gotten a full night of sleep in days. Each observation, data point, and quote filled in another piece of the kidney discard puzzle.
Once the research phase of our project was complete, we'd gained a holistic view of the organ transplant ecosystem and were able to start identifying the root causes of discard. We synthesized our research into two major findings:
The system designs of the national allocation and evaluation processes hold multiple opportunities to decrease discard through shared technology, policy, and performance incentives.
These structural opportunities were the largest method to cascade change into the shared options for all users. Currently, providers must navigate multiple systems for donor and recipient information with manual data entry at almost every step of the organ procurement and allocation process. A single technology source for transplant centers, organ procurement organizations, and donor hospitals to share donor data and organ offers with predictive analytics and real-time feedback could increase speed and decision-making, decreasing cold ischemic time of kidneys.
Another structural opportunity lies within defaults of the applications. For example, the current system requests users to give a "provisional yes" to an organ offer, versus a committed yes or no, which causes additional delays. In transplant center evaluation, we also discovered opportunities to create incentives, rather than just disincentives, which could inspire, rather than restrict, the natural curiosity and patient-centered drive of clinicians. Lastly, US policies could benefit from improved the nuance in matching, such as stratifying "old for old" — matching elderly donors with elderly recipients — existing policy in nations like France. These measures could decrease clinician burden and increase the opportunity to procure an expanded criteria of donors.
We also noticed opportunities for the system to influence individual and local behavior. We applied a behavioral science lens to understand why individuals across this ecosystem make the decisions they make every day. This led to our second research insight:
The system can leverage behavioral, or psychological, science to provide guardrails that steer smart, well-meaning clinicians to make difficult proximal decisions for long-term gains.
A host of behavioral opportunities abound in the current system. Risk aversion, for example, plays a major role in shaping how organ offer decision-makers (typically transplant surgeons) view kidney offers. Surgeons don't want to transplant a kidney they believe could put their patient or transplant program at risk: these risks seem more temporally powerful than the patient's long-term gains. Cognitive burden among surgeons is also evident; while the organ allocation technology provides clinical information on the organ donor, the surgeon must use their best clinical judgment to compare that data to the clinical characteristics of their patient. Frequently, there is no clear answer; without the aid of assisted clinical decision-making, it is a multivariate decision that usually comes down to intuition.
Labeling bias is also present in decision-making. Each kidney that enters the allocation system is assigned a quality score called the Kidney Donor Profile Index (KDPI). Kidneys with scores above 85, likely from an older, less healthy donor, are more likely to be rejected and ultimately discarded if surgeons and patients decline them. However, research has shown that even kidneys with high KDPI are, in most cases, better for a patient's health than staying on dialysis. The predictive value of KDPI is marginally better than a coin toss. Yet the KDPI label can sway surgeons and patients to reject a kidney. Other, smaller details in the kidney offer — like a patient's number on the match list — also inflict labeling bias in clinicians. Information cascade takes effect when it's obvious that many other surgeons have already declined that kidney.
Availability bias also takes effect. Surgeons make their best determination based on the information most readily available to them. One nephrologist told us: "Transplant surgeons' decisions will depend on their last five surgeries." In other words, if their last few transplants, even of marginal kidneys or high-risk patients, were successful, they were more likely to have wider acceptance criteria when evaluating subsequent kidney offers. This is availability bias at work: the tendency to rely on immediately available examples that come to mind when evaluating a decision. We found this created either a virtuous or vicious cycle of patient and kidney selection.
Applying the research
We used our research findings to create a series of concepts and best practices to address these systemic and behavioral opportunities, with the understanding that structural design influences individual behavior. First, we visually mapped the current journey of donated kidneys, from donor death through transplant. This revealed that the most likely point of discard along the donated kidney's path occurs at the point of organ offer between the organ procurement organization and transplant centers. If the offer is declined for the top patient (it is in most cases), the offer pings back and forth from the organ procurement organization and transplant center to transplant center.
Once we communicated our insights to our client, we shifted our focus from describing the problem to solving the problem. We hosted a co-creation workshop with kidney transplant experts and kidney patients from across the country to further immerse ourselves in their experience and facilitate collaborative idea generation. After two days of collaboration, more than two dozen ideas emerged as having the highest potential for reducing discard while balancing the demands of implementation within a complex system. Our team refined these ideas and evolved them into detailed solutions. To drive home the impact of these concepts, we presented our client with a visual artifact displaying the kidney care ecosystem of the future, a world in which these concepts had been fully realized. These concepts included designs for:
- Assisted clinical decision-making tools, built on a robust ecosystem of data science and behavioral science with improved defaults and feedback, so clinicians can make confident organ acceptance decisions that are evidence-based and best for their patients;
- An elimination of KDPI as an indicator of kidney value, replaced with a kidney-patient match score, which communicates donor-patient compatibility as well as a weighted value for the scarcity of kidneys and the opportunity cost of staying on dialysis;
- A shared kidney database, where a clinician can receive real-time feedback and "follow" (as on Twitter) the kidneys they have refused and transplanted; researchers can also study these data on decision-making and acceptance for policy and program design;
- A transplant center evaluation and incentive design, to inspire innovation, volume, and patient / kidney acceptance, prizing the centers who achieve the most public good;
- A method of experience-sampling patients to capture their ongoing preferences and – as one surgeon called it – their "hot miserable score" which would be folded into the matching value, so that clinicians can discern true demand for kidneys;
- A best-practices playbook of mindsets and behaviors among high-performing centers within the existing restraints; and more.
Our work inspired change and communication across the kidney care landscape; our client introduced us to other agencies in the space to share learnings and ownership of next steps. The agency sponsoring our work also requested that our team design and pilot prototypes for the kidney transplant community, including a video to educate patients on high-risk and high-KDPI kidneys (which are frequently their likeliest option to get a transplant sooner). These prototypes are being prepared by the client to be tested in the field measured for impact. Next, our team will be collaborating with other agencies to improve the organ allocation system design, and disseminating best practices for local mindsets at transplant centers and organ procurement organizations across the country.
Our team remains deeply inspired, indebted, and appreciative of the many astounding transplant surgeons, nephrologists, researchers, transplant teams, and organ procurement organizations who shared their insights, ideas, pain points, and criteria for these designs, and who work tirelessly for patients across our county. We hope our work can continue to help shape the strategy and outcomes for these patients and organizations.