Finding HCPs who are managing yet-to-be-diagnosed patients with rare diseases is often said to be analogous to finding a needle in a haystack. A better analogy would be mining for gold; knowing where to dig is only half the problem. The strategy for bringing this value to the surface is another matter.
Identifying the HCPs who have undiagnosed rare disease patients tells us where to dig, however we need to "bring this value to the surface". Simply identifying HCP's having rare disease patients will not alter the rate at which these patients are diagnosed. The delay period between symptom onset and correct diagnosis is not generally correlated with AI-model scores. Undiagnosed rare disease patients may look quite different due to the stage of their progression and/or the diversity of the clinical journeys associated with the disease.
We need to formulate a strategy for accelerating diagnosis through patient activation. As my colleague, Susan Abedi says, "Ensure you have an integrated strategy for engaging HCPs and supporting diagnosis, not just finding patients" (Susan Abedi, "Two Truths and A Lie"). A strategy needs to be put in place to profile patients and then form profile-specific strategies to accelerate diagnosis through HCP outreach, education, etc. Properly done, AI models give us important information we can use in these profiles in order to shorten undiagnosed patient diagnostic delay.
AI-assisted profiling and patient activation strategies addresses the diagnostic delay problem
At 81qd, we recognize that we have various tools in our toolbox to tackle the problem of activation. We find patients in context by combining the outputs of our patient-finding solution with our proprietary clinical network mapping solution. Once patients are identified and attributed to HCPs, we build clinical networks based on our proprietary network mapping solution. Our clinical influence scores combined with our patient-finding precision scores offer state of the art microtargeting.
Comprehensive diagnosed and undiagnosed patient profiling is also necessary. We not only find undiagnosed patients, but also provide the insight needed to build the required diagnosed and undiagnosed patient profiles to assist in the formulation of patient activation strategies. Thus clients can see what is happening right now with known rare disease patients, and also see how undiagnosed and/or diagnosed and untreated rare disease patients look different from those who are diagnosed or diagnosed and treated.
About the author
Tim is a statistician and computational biologist with 16 years of experience in the biochemistry of Drug Discovery, 15 years of experience in Drug Discovery Analytics, and 5 years of experience translating this into AI and Statistical solutions in Healthcare Care Analytics. Tim has an extensive publication record, has received industry awards and been an invited speaker in health care informatics conferences.