Patients with delayed diagnoses want digital tools that carry their story across visits

Three patient advocates described why digital tools should carry their accounts across visits and give patients a way to correct what is missing.

Author
HDM Staff
Save
Share

    LAS VEGAS — Patara Harper said it took about ten years to get a Crohn’s disease diagnosis. She described recurring pain, repeated visits and clinicians who attributed her symptoms to stress or depression. By the time one emergency physician admitted her and began to help, Harper said, her children were with her and she was in a wheelchair.

    Harper’s account pointed to a design problem for health systems and technology companies. A patient may have a long record of tests, visits and treatments, yet the person seeing her next may have only a short account of the immediate complaint. If the patient has to reconstruct the history at each visit, the clinician may miss how an initially plausible explanation has held up over time.

    Jen Horonjeff, co-founder and CEO of Real Patients, moderated Believe Her Sooner on the Garage Stage at HLTH USA. Harper, a community advocate and breast cancer survivor, joined patient advocates Heather “Joy” Ali and Erica Olenski. They described what it takes to be heard, how patients and caregivers use technology between visits and what a tool cannot do on its own.

    Repeated encounters can hide a single unresolved problem

    Harper said the process took “about ten years.” She described going to several clinicians and eventually returning to emergency departments because of the pain. From her perspective, each new visit came with the burden of explaining what she had already tried and why she was back.

    Horonjeff asked what was different about the clinician who ultimately listened to her. Harper credited the doctor with seeing her as a whole person. She also asked whether she would have been treated differently if she appeared healthier on arrival. Her own prescription for clinicians was direct: “Treat me fairly.”

    Horonjeff gave an example from her own experience. When she went to the emergency department with severe headaches that she said were later linked to a brain tumor, she recalled her partner worrying that clinicians would take her less seriously because she could still chat with them. It was another instance, distinct from Harper’s, in which the patient’s presentation could obscure the severity of what she was reporting.

    A digital intake form often has a narrow task, such as collecting a reason for today’s appointment. The panel’s examples show what can be lost when the form is also expected to represent a longer illness. A clinician may need to see which symptoms persisted after treatment, whether a patient returned repeatedly and why a previous explanation is being questioned. Those details can get separated when records organize encounters one visit at a time.

    What patients are already doing with AI

    Olenski described caring for a medically complex son and using AI to organize a conversation involving several specialists. She used the tool to “draft an agenda” for the conversation. The tool helped her assemble her questions and organize what she wanted to cover. She described doing that work while already carrying the responsibilities of caregiving.

    Olenski also discussed finding people online who described symptoms like hers during her search for an explanation for severe skin problems. She said those conversations helped her feel less isolated and gave her leads to discuss. Her experience illustrated the value of finding a vocabulary for what a patient is experiencing. A discussion among patients, or a result returned by an AI tool, does not by itself confirm what caused the symptoms.

    She pressed designers to preserve a patient’s “ability to make decisions” as they bring AI into healthcare. Her examples suggested a distinction between help that makes people better able to prepare for a clinical conversation and a system that assumes it can speak for them. In the former case, the patient can still review the questions, correct what is missing and decide what to raise with a specialist.

    For developers, that distinction has practical consequences. A summary intended to save time could show which details came from the patient, leave room for corrections and make it possible to retrieve the underlying account. A clinician could then assess a shorter summary without assuming that it contains everything the patient believed was important.

    Bring patients into the design decisions

    Ali said she was 22 when she received a lupus diagnosis. She described how chronic illness had changed her relationship with her body and had made caregiving a part of her marriage. She also discussed the experience of receiving information after a mastectomy and the difficulty of being prepared for what a diagnosis might mean. Those were accounts of what she needed at the time, rather than claims about how another patient would experience the same care.

    Ali urged the audience, “don’t design health care for patients without patients.” Asked what that would mean in practice, she described “lived experience” as a form of intelligence that can inform a design decision. The people using a service can explain where its workflow fits their lives and where they are asked to do work that an organization may never see.

    That input can matter early, when a team decides which problems it wants an intake tool to solve. A patient who has already told a complicated story several times may use a form differently from someone making a first visit. A caregiver may also need to prepare for a conversation in which several specialists know different parts of the file. Each scenario tests whether the tool can carry information forward in a way that is useful to the next person.

    For someone designing a patient portal, there is a difference between giving a patient access to a record and giving her a practical way to describe what she thinks the record misses. A team could ask patients to review its proposed intake process and then compare what the form retains with what they would tell a clinician in their own words. It could also test whether the next clinician can see the correction in time to use it.

    The panel presented personal experiences rather than a comparison of tools or an estimate of diagnostic delays. A health system could use those experiences to ask whether people can add context to a structured record, see what has been passed along and correct summaries before clinicians rely on them. The speakers’ test was whether a new tool leaves patients better able to be heard when they meet the next clinician, with less of the burden of keeping the whole story together on their own.

    Photo credit for both images: HLTH USA 2026.