|Thursday, February 21|
|PS1 Poster Session 1 & Opening Mixer||
Thu, Feb 21, 6:30 PM - 8:00 PM
Inverse Prediction: Application to Infectious diseases and Clinical MicrobiologyView Presentation *Jayawant Narayan Mandrekar, Mayo Clinic
Keywords: Inverse Prediction, Regression, CORR, REG, GPLOT
The polymerase chain reaction (PCR) a Nobel-prize winning technology has become an indispensable technique in medical and biological research labs. Real time PCR used in serial monitoring of specimens from solid organ transplant recipients for detection of viruses require estimates and 95% confidence intervals for the viral limits. However, “true” viral loads are unknown and treated as a predictor variable and the results from PCR are treated as the response variable. We use the approach of inverse prediction to estimate or predict X (unknown true concentration from patient sample) from known Y (known concentration from real time PCR) as simply switching the roles of the response and predictor variables to get the desired predictions i.e., regress X on Y, is incorrect. This is because the primary assumption of regression that X is measured without error and Y is a dependent, random and normally distributed variable is violated. A SAS® macro using the PROC CORR, PROC REG, and PROC GPLOT procedures for the calculation of the viral load estimate and its associated 95% confidence limits has been developed. This technology for early detection, monitoring, and medical management of patients with these infections is now used both for clinical practice as well as for laboratory quality monitoring.