VF in the News

Pathway to Diagnosis (P2D) Plans an Innovative Study with the Goal of Reducing Delays in Diagnosis

P2D Objective

Using big data to accurately predict a diagnosis using predictive analytics and machine learning methods.

The Reality: Delays in diagnosis for patients with vasculitis

Despite remarkable medical advances, diagnostic delays for patients with vasculitis remain extremely common and are a major cause of increased morbidity and mortality.

This project directly responds to, and aligns with, the top research priorities of patients with vasculitis; specifically, to conduct research that will reduce the delay in diagnosis for new patients and improve clinical outcomes.

Why are there still delays in diagnosis?

The delays in diagnosis are frequently attributed to:

› The rarity of the disease

› Providers’ lack of awareness of the disease › Inadequate medical knowledge

› Practice referral patterns

With recent increases in clinical and administrative demands on clinicians, there is often insufficient time to understand, gather or even piece together patients’ complex and often fragmented medical histories.

Each patient then becomes a data challenge with a vast amount of complicated information that is not always easily accessible for care providers.

Potential Solution: Harnessing the power of “big data”

What can be done when patients become a data challenge with immense amounts of complicated information that isn’t easily accessible for the care providers trying to piece together the bigger picture and find a correct diagnosis?

In today’s technology driven world, large consumer companies like Amazon can suggest to you merchandise you are likely to buy based on an algorithm.

This algorithm is designed to collect large amounts of data about you (e.g. what you like to buy, where you live, how old you are, etc.) and accurately predicts what items you are likely to buy.

In the healthcare industry, large amounts of data from lab results to medications are collected about each person, similar to how data is collected about one’s buying habits. The same basic principle of using an algorithm, or model, to predict what one might purchase can also be used to sift through, and make sense of an individual’s health data.

Testing a machine learning model to achieve a diagnosis Machine learning capabilities and predictive analytics offer a potential solution for assisting busy providers in sorting through an individual patient’s “big data” to establish accurate diagnoses and reduce diagnostic delays.