According to a recent article in MedCity News, families with children who have a rare disease wait on average for five years before receiving an accurate diagnosis.
It is estimated that there are approximately six thousand rare diseases affecting almost four percent of the world’s population. The doctor may make every effort to obtain an accurate diagnosis, but the discovery process is challenging. Many doctors are not familiar with the majority of these diseases.
The doctor must rely on analyzing findings from a small number of rare disease patients and compare those findings to the rest of the worldwide population.
If the current diagnostic process can be improved, up to four hundred million people in the world would benefit. Half of these people are children.
Artificial Intelligence (AI) Points the Way
Using AI to its maximum advantage would reduce the amount of time needed to accurately diagnose a rare disease.
Currently, AI is being used in the four main states of drug development:
- Identifying targets for medical procedures
- Locating biomarkers to diagnose a disease
- Efficient identification of evaluable candidates for clinical trials
- Developing new drugs
Algorithms and Machine Learning
The design of clinical trials can be improved through machine learning. For example, if a few candidates are chosen to participate in a clinical trial, but are not qualified, replacing them would not only prolong the trial but be very costly in terms of resources and time.
Machine learning can improve the design of trials through the identification of suitable candidates. Algorithms recognize patterns that separate good and bad candidates as well as notify the research team if a trial is not producing positive results.
Machine learning algorithms can actually see patterns just the way physicians see them, but they need many examples. In fact, they need thousands of examples.
Big Data
Recently there has been a wealth of health data generated by hospitals, industry, and academia.
Yet these groups are reluctant to share this information even though it is the patient who owns his or her health data.
The hold-back is usually due to the amount of time and money invested in arranging clinical trials and collecting the data. This explains the determination of academic hospitals to be the first to publish papers on the final data.
As a result, it may be months or sometimes years before other organizations can analyze the data. Perhaps there should be some incentive for the hospitals to provide this data to the industry or academic institutions.
The WHO and Data Sharing
The World Health Organization has committed to exchanging cross-border sharing of health data. Data sharing has been taking place among researchers and experts in infectious diseases in connection with Covid-19. This bodes well for future data sharing in other areas of disease including rare diseases.
Looking Forward
Researchers continue to search for a system that will enhance a patient’s care but still manage compliance with HIPAA regulations.
A health information system that shares data with the patient’s health providers would be helpful to the patient and the patient’s physicians.
But there is a caveat here that explains the risk of either violating the patient’s health information or worse yet, having it disrupted by a hacker. No matter where the data are stored, an experienced hacker may be able to access it. There is still a myriad of questions surrounding where and how and with whom health information should be shared.
