Big Data,Artificial Intelligence and machine learning algorithms have revolutionized many fields including medicine. I was wondering how can we use Big data and Artificial Intelligence to improve diagnostic accuracy, Prognosis and Physical Therapy management ?
Excellent question! There's been a recent surge in AI-mediated auditing and AI-supported diagnosis of many conditions. Several studies demonstrated that these methods of imaging analysis and pattern recognition can reliably improve the diagnostic accuracy when compared to a human-based assessment (e.g. in classification of fractures Article Automated detection and classification of the proximal humer...
dermatological assessmentArticle Deep neural networks show an equivalent and often superior p...
or MRI interpretationArticle Accuracy of dementia diagnosis - A direct comparison between...
Article Computer-assisted diagnostic system for neurodegenerative de...
Article Predicting behavioral variant frontotemporal dementia with p...
.Article End-to-end lung cancer screening with three-dimensional deep...
The research on guiding diagnosis, improving accuracy, and clinical outcomes has been somewhat less popular, perhaps because of ethical limitations, poor availability of datasets, and confusing standardisation of symptomatology.
To my knowledge, the first AI use in a large scale auditing of diagnostic accuracy has been done in Frontotemporal dementia (by us): Article Diagnostic accuracy of frontotemporal dementia. An artificia...
This field has, in my opinion, an enormous potential of automation and dramatic improvement in quality of care, for the fraction of resources needed to conduct a full-scale audit.
Naturally, human issues still persist: the implementation of AI requires good note-keeping and universality of databases (which is very difficult to achieve!) and any potential benefits must be implemented by responsible teams and clinicians (same problem as in conventional audits).
Nevertheless, this area of research offers an exciting new route of getting the clinical answers right where we need them.
Since Big data tries to find hidden information from raw data, several new techniques will be considered
In my opinion there will be a lot of progress in the systemic analysis of knowledge that we already have and I am referring in particular to patient files and scientific literature (i. e. PubMed). If I'm not completely misled, IBM Watson Health, for example, works according to this principle.
- Badges
- Science topic