 Machine or artificial intelligence depends on complex architectures of neural networks that need to be properly built for every particular task. AI needs large amounts of training data to work - as machines are not yet able to contextually adapt, that is, build reliable models from sparse and noisy data, like humans do. But it is not just artificial neural networks that need to keep improving. Individual intelligence depends on the complexity of neural networks in the brain. These networks consist of almost 100 billion neurons of different types and trillions of flexible connections between neurons engaged in similar tasks. With productive learning, the connections keep evolving, and patterns of electrical activity continue to tone and refine. |   |
Every intelligent entity - whether human or machine - depends not only on the configurations of its neurons, but also connections between itself and others entities, optimized for efficient exchange of information. Hence, better human networks providing training and feedback from others will lead to both smarter humans and better AI.
Just one example of how this could be leveraged in Healthcare.
Just one example of how this could be leveraged in Healthcare.
The biggest obstacle for applying AI in Healthcare is the lack of good data available for computation.
AI can predict heart attacks and strokes more accurately than a doctor - if there are good quality medical records. AI is better in analyzing visual information - if there are tenths of thousands of good quality images annotated for thousands of patients. But in most cases data we need for predicting outcomes is either too expensive to prepare or impossible to get - as it remains in the brains of individuals. And cages accurately monitoring food intake, activities and symptoms so far work only for mice.
AI can predict heart attacks and strokes more accurately than a doctor - if there are good quality medical records. AI is better in analyzing visual information - if there are tenths of thousands of good quality images annotated for thousands of patients. But in most cases data we need for predicting outcomes is either too expensive to prepare or impossible to get - as it remains in the brains of individuals. And cages accurately monitoring food intake, activities and symptoms so far work only for mice.
In machine learning, efficiency can be improved if we pre-train algorithms on cheap and large datasets. This helps to pre-optimize the parameters for working with more expensive data.
In the world of humans, efficiency of medical research could be improved if cheap large datasets focusing on particular medical questions could be easily collected. If clinical trials were more engaging and convenient, generating outcomes clinically meaningful to participants; if the participants could properly design studies by themselves, guided by Software as a Medical Device (SaMD) platform and their own devices (BYOD model), we could collect enough data to get to the next level.
In the world of humans, efficiency of medical research could be improved if cheap large datasets focusing on particular medical questions could be easily collected. If clinical trials were more engaging and convenient, generating outcomes clinically meaningful to participants; if the participants could properly design studies by themselves, guided by Software as a Medical Device (SaMD) platform and their own devices (BYOD model), we could collect enough data to get to the next level.
The success of an N-of-1 trial methodology, hindered by the operational complexity, largely depends on the collaboration of patients and knowledgeable parties. Large scale crowdsourced studies would need innovative software solutions inter-connecting participants and their treatment sequences. And the impact of this platform would be no less important than the race to build an artificial intelligence for everything.
References
Geirhos R, Janssen DH, Schütt HH, Rauber J, Bethge M, Wichmann FA. Comparing deep neural networks against humans: object recognition when the signal gets weaker. arXiv preprint arXiv:1706.06969. 2017 Jun 21.
Scuffham PA, Nikles J, Mitchell GK, Yelland MJ, Vine N, Poulos CJ, Pillans PI, Bashford G, Del Mar C, Schluter PJ, Glasziou P. Using N-of-1 trials to improve patient management and save costs. Journal of general internal medicine. 2010 Sep 1;25(9):906-13.
Lillie EO, Patay B, Diamant J, Issell B, Topol EJ, Schork NJ. The n-of-1 clinical trial: the ultimate strategy for individualizing medicine?. Personalized medicine. 2011 Mar;8(2):161-73.
Sackett DL. Clinician-trialist rounds: 4. Why not do an N-of-1 RCT?.
Li J, Tian J, Ma B, Yang K. N-of-1 trials in China. Complementary therapies in medicine. 2013 Jun 30;21(3):190-4.
Nyman SR, Goodwin K, Kwasnicka D, Callaway A. Increasing walking among older people: A test of behaviour change techniques using factorial randomised N-of-1 trials. Psychology & health. 2016 Mar 3;31(3):313-30.
Federman DG, Shelling ML, Kirsner RS. N-of-1 trials: not just for academics. Journal of general internal medicine. 2011 Feb 1;26(2):115-.
Duan N, Kravitz RL, Schmid CH. Single-patient (n-of-1) trials: a pragmatic clinical decision methodology for patient-centered comparative effectiveness research. Journal of clinical epidemiology. 2013 Aug 31;66(8):S21-8.
Joy TR, Monjed A, Zou GY, Hegele RA, McDonald CG, Mahon JL. N-of-1 (single-patient) trials for statin-related myalgia. Annals of internal medicine. 2014 Mar 4;160(5):301-10.
Shaffer JA, Falzon L, Cheung K, Davidson KW. N-of-1 randomized trials for psychological and health behavior outcomes: a systematic review protocol. Systematic reviews. 2015 Jun 17;4(1):87.
Scuffham PA, Nikles J, Mitchell GK, Yelland MJ, Vine N, Poulos CJ, Pillans PI, Bashford G, Del Mar C, Schluter PJ, Glasziou P. Using N-of-1 trials to improve patient management and save costs. Journal of general internal medicine. 2010 Sep 1;25(9):906-13.
Lillie EO, Patay B, Diamant J, Issell B, Topol EJ, Schork NJ. The n-of-1 clinical trial: the ultimate strategy for individualizing medicine?. Personalized medicine. 2011 Mar;8(2):161-73.
Sackett DL. Clinician-trialist rounds: 4. Why not do an N-of-1 RCT?.
Li J, Tian J, Ma B, Yang K. N-of-1 trials in China. Complementary therapies in medicine. 2013 Jun 30;21(3):190-4.
Nyman SR, Goodwin K, Kwasnicka D, Callaway A. Increasing walking among older people: A test of behaviour change techniques using factorial randomised N-of-1 trials. Psychology & health. 2016 Mar 3;31(3):313-30.
Federman DG, Shelling ML, Kirsner RS. N-of-1 trials: not just for academics. Journal of general internal medicine. 2011 Feb 1;26(2):115-.
Duan N, Kravitz RL, Schmid CH. Single-patient (n-of-1) trials: a pragmatic clinical decision methodology for patient-centered comparative effectiveness research. Journal of clinical epidemiology. 2013 Aug 31;66(8):S21-8.
Joy TR, Monjed A, Zou GY, Hegele RA, McDonald CG, Mahon JL. N-of-1 (single-patient) trials for statin-related myalgia. Annals of internal medicine. 2014 Mar 4;160(5):301-10.
Shaffer JA, Falzon L, Cheung K, Davidson KW. N-of-1 randomized trials for psychological and health behavior outcomes: a systematic review protocol. Systematic reviews. 2015 Jun 17;4(1):87.
RSS Feed
