Crowd-sourcing medical diagnoses: A possible future of healthcare?

20170320110000, Brenda Lau
Moderators, all of whom are licensed physicians, monitor conversation to ensure conclusions are made based on peer-reviewed science.
14 years ago, Carly Heyman fell ill with a rare combination of symptoms - depression, weight gain and long hours of sleep interrupted by nightmares or suicidal thoughts. She sought help from many doctors, many of whom treated her symptomatically as they could not identify the cause of her illness.

After several years, a doctor finally diagnosed her with Fragile X syndrome and all that was needed was a hormone patch.

Inspired by Carly, her brother Jared Heyman founded CrowdMed, a crowd-sourcing platform, reportedly made up of 20,000 doctors, nurses, medical students and others - with no formal healthcare credentials - for individuals with unexplainable medical conditions to pinpoint a likely diagnosis.

The San Francisco-based company was launched in 2013 and claims to have helped more than 2,000 patients to get closer to finding the right diagnosis. It also has pilot programs with four major insurance firms to determine if crowd-sourced medicine can be economically effective for patients and insurers.

If a condition is unusual, Heyman says that patients "often end up in this cycle where you just get referred from specialist to specialist searching for that needle-in-a-haystack person who might have familiarity with what you have."

Shortening the average time for a diagnosis

A 2013 research showed approximately 8% of Americans are affected by rare diseases, and it takes an average of 7½ years to get a diagnosis. In Britain, it takes an average of 5.5 years - and doctors often get it wrong. A separate survey in Europe looked at eight rare diseases and uncovered that 40% of patients initially received an incorrect diagnosis.

CrowdMed aims to reduce the timeline of a diagnosis. The idea works by a patient paying USD50 to create an anonymous patient account, then answering a series of questions about their symptoms and medical history - with an option to upload their anonymous medical records.

The information is then uploaded to the site for "medical detectives" to collaborate with the patient and each other to offer suggestions about possible diagnoses.

Most "medical detectives" are doctors, but many are also individuals with no medical background. These individuals are ranked based on the number of diagnoses they get "right" - based on what the patient deems correct after consultation with his or her physician - and rankings from their peers on the site.

There are also moderators, all of whom are licensed physicians, who monitor conversation to ensure conclusions are made based on peer-reviewed science. The CrowdMed algorithm can also determine those with the highest probability of being right.

Patients then receive a list of potential diagnoses, ranked from the most to least probable diagnoses based on crowd opinion and the algorithmic assessment. Those from the crowd who get the diagnosis "right" share a cash reward that the site has reserved from the patient's fees.

CrowdMed proves to have economical impact

Studies have reportedly shown that the impact of CrowdMed has been substantial as after cases were resolved, patient healthcare visits dropped by 50% and average medical costs dropped by USD200 per month.

The controversial approach proved to address the need of individuals suffering from a number of rare diseases that are difficult for general practitioners or family physicians to recognise - the National Institutes of Health, America's medical agency, recognises 7,000 different rare diseases.

Thus, getting a correct diagnosis, even if no medical treatment is available for the condition, is often a triumph for patients who need validation that something is truly wrong, says Ronald DeBellis, chief scientific officer for the National Organisation for Rare Disorders, a non-profit organisation in Connecticut.

Empowered patients forces physicians to change ways

But this service does not come cheap. Patients pay fees of up to USD749 per month - higher fees buy higher-ranked detectives. The diagnosis is also not foolproof. The site and its detectives avoid liability issues by ensuring that suggestions are just offered, rather than medical advice.

There are also concerns that the CrowdMed detectives do not have tools to recognise if a symptom is from a physical condition or a manifestation of a patient's psychology - somatoform disorders.

This presents "a limitation of not being able to physically examine a patient," says Kyle Walker, a CrowdMed moderator and general practice physician in Ohio.

While the idea presents a game changer for insurance companies and a possible peace of mind to patients, it also forces physicians to acknowledge that patients are becoming more empowered and that the way online information is received and perceived, should evolve accordingly - and CrowdMed is just the beginning. MIMS

Read more:
Telehealth services: Are they truly cost effective?
Algorithmic disease surveillance: Staying ahead of an epidemic
Orphan Diseases and Orphan Drugs: What a pharmacist should know

Sources:
http://www.nbcnews.com/mach/science/how-crowdsourcing-transforming-diagnosis-rare-disorders-n728306
http://www.economist.com/news/science-and-technology/21651765-many-hands-make-diagnosis-work-crowd-will-see-you-now
http://www.popsci.com/crowdsourcing-diagnoses-medical-mysteries