FDA-regulated AI Algorithms: Trends, Strengths, and Gaps of Validation Studies

Author(s):  
Shadi Ebrahimian ◽  
Mannudeep K. Kalra ◽  
Sheela Agarwal ◽  
Bernardo C. Bizzo ◽  
Mona Elkholy ◽  
...  
Keyword(s):  
1970 ◽  
Vol 117 (7) ◽  
pp. 1389
Author(s):  
H.M. Ryan ◽  
J.M. Mattingley
Keyword(s):  

1999 ◽  
Vol 38 (04/05) ◽  
pp. 339-344 ◽  
Author(s):  
J. van der Lei ◽  
B. M. Th. Mosseveld ◽  
M. A. M. van Wijk ◽  
P. D. van der Linden ◽  
M. C. J. M. Sturkenboom ◽  
...  

AbstractResearchers claim that data in electronic patient records can be used for a variety of purposes including individual patient care, management, and resource planning for scientific research. Our objective in the project Integrated Primary Care Information (IPCI) was to assess whether the electronic patient records of Dutch general practitioners contain sufficient data to perform studies in the area of postmarketing surveillance studies. We determined the data requirements for postmarketing surveil-lance studies, implemented additional software in the electronic patient records of the general practitioner, developed an organization to monitor the use of data, and performed validation studies to test the quality of the data. Analysis of the data requirements showed that additional software had to be installed to collect data that is not recorded in routine practice. To avoid having to obtain informed consent from each enrolled patient, we developed IPCI as a semianonymous system: both patients and participating general practitioners are anonymous for the researchers. Under specific circumstances, the researcher can contact indirectly (through a trusted third party) the physician that made the data available. Only the treating general practitioner is able to decode the identity of his patients. A Board of Supervisors predominantly consisting of participating general practitioners monitors the use of data. Validation studies show the data can be used for postmarketing surveillance. With additional software to collect data not normally recorded in routine practice, data from electronic patient record of general practitioners can be used for postmarketing surveillance.


2004 ◽  
Vol 4 (2) ◽  
pp. 23-30
Author(s):  
K. Connell ◽  
M. Pope ◽  
K. Miller ◽  
J. Scheller ◽  
J. Pulz

Designing and conducting standardized microbiological method interlaboratory validation studies is challenging because most methods are manual, rather than instrument-based, and results from the methods are typically subjective. Determinations of method recovery, in particular, are problematic, due to difficulties in assessing the true spike amount. The standardization and validation process used for the seven most recent USEPA 1600-series pathogen monitoring methods has begun to address these challenges. A staged development process was used to ensure that methods were adequately tested and standardized before resources were dedicated to interlaboratory validation. The interlaboratory validation studies for USEPA Method 1622, for Cryptosporidium, USEPA Method 1601 for coliphage, and USEPA Method 1605 for Aeromonas assessed method performance using different approaches, due the differences in the nature of the target analytes and the data quality needs of each study. However, the use of enumerated spikes in all of the studies allowed method recovery and precision to be assessed, and also provided the data needed to establish quantitative quality control criteria for the methods.


2016 ◽  
Vol 9 (3) ◽  
pp. 235-243
Author(s):  
Hong Yuan ◽  
Jonathan E. Frank ◽  
Joseph R. Merrill ◽  
Daniel A. Hillesheim ◽  
Mark H. Khachaturian ◽  
...  
Keyword(s):  

Sign in / Sign up

Export Citation Format

Share Document