| |
|
|
 |
When you purchase OCR software/services, you invariably see claims of 99% accuracy. That level is achieved only under ideal conditions where the document was designed to demonstrate high accuracy. Your real-world documents will yield accuracies often below 60%. As important, key search terms are often completely missed. To illustrate, the text below was captured using Adobe® Acrobat®. The inaccurate hidden-text below is what you are actually searching when you settle for off-the-shelf OCR solutions. The bold text beyond this point is actual capture output from Acrobat®: ,. . , .? -' _.. DEPARTMENT OF HEALTH AND HUMAN SERVICES a50 Third Ave PUBLIC HEALTH SERVICE Brooklyn, NY 11232 FOOD AND DRUG ADMINISTRATION 718/340-7000 NPiME OF lNDl"lDUAL TO WHOM REPCJRT lSS"ED PERIOD OF lNSPECTlON c. F. NUMBER '0: Timothy Getek, Ph.D. 10/5 . ..23/9a 24-36921 'ITLE OF lND,"lDUAL TYPE ESTABUSHMMT !NSPECTED &sociate Director Bioanalytical Lab. Bioanalytical Lab 'IRM NAME NAME OF FIRM. BRANCH OR "NIT INSPECTED ?orest/Inwood Laboratories, Inc. Same :TREET ADDRESS STREET ADDRESS OF PREMEES !NSPECTED !20 sea Lane SZilW :,l-Y AND STATE ,zip code, ClTY AND STATE ,zlp code, bxningdala, NY 11735 SallIe l"RlNG AN lNSPECTN5N OF YOUR FIRM ,I, WE, OBSERVED: (Rimantadine), and R/1700/004 (Isosorbide Dinitrate). 1. The SOP #F-O01 entitled "Analytical Run Acceptance Criteria" is not objective in that it permits calibration standards to be deleted so that quality control (QC) value will fall within the acceptable limits. Examples include: b) 2. SOP #F-O01 permits calculation of accuracy of QCs using the validated QC value as the true value, rather than using the theoretical value. Calculation of accuracy using the theoretical value as the true value would result in the majorit analytical runs for 3. The reported assay precision in the analytical runs was not accurate because the precision analysis did not include the standards and QCs which failed to meet the acceptance criteria. For example, inclusion of all standard values in- -increased the coefficient of variation (CV) form(from 2.5 10.2% to 2.515.5%) and for v (from 259.8% to 2512.0%). Similarly, inclusion of all standard ,values i- increased the C\ 'for-(from 3.6 - 9.6% to 3.6 - 18.3%) andB(from 2.2 - 8.0% to 2.2 - 21.9%). SEE REVERSE OF THIS PAGE FORM FDA 48115m51 Lnwrencr F;lrina,Invertigator io/23/9a INSPECTIONAL OBSERVATIONS ,,.,.,... PAOE 1 OF6 PAGES. - End of OCR Output - To view the parent PDF file, click here.
To view the finished output from Document Capture
|
|