BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Penn State Health News - ECPv5.2.1//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Penn State Health News
X-ORIGINAL-URL:https://pennstatehealthnews.org
X-WR-CALDESC:Events for Penn State Health News
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20200308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20201101T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20200925T120000
DTEND;TZID=America/New_York:20200925T130000
DTSTAMP:20201026T183347
CREATED:20200914T174152Z
LAST-MODIFIED:20200914T174152Z
UID:45573-1601035200-1601038800@pennstatehealthnews.org
SUMMARY:Webinar: Biostatistics internship presentations
DESCRIPTION:“Assessment of Treatment Effect in a Proof-of-Concept Trial with Longitudinal Measurements” will be presented by Biyi Shen\, a Penn State College of Medicine Biostatistics PhD student who completed a summer internship at Novartis. \nShen’s project tries to extend the generalized MCP-Mod methodology to assess time-response in a Phase 2a trial with longitudinal measurements. Specifically the placebo-adjusted treatment effects and the associated covariance matrix estimated from an MMRM analysis are used to conduct various linear tests to detect: \n\na reduction at the primary time point;\nan average treatment effect across all visits (constant model); and\na time response signal based on a set of candidate models under model uncertainty.\n\nThe project declares a positive proof-of-concept if at least one of the tests is statistically significant after multiplicity adjustment at the MCP stage. Sample size and power calculations as well as simulation results show our approach requires a smaller sample size by taking into consideration of available data from all visits as compared to the conventional method focusing on the primary time point data only. Then in the Mod step\, the predicted time response curve and the corresponding confidence intervals are obtained using the bootstrap model averaging approach. Finally\, the utility of methods that don’t need pre-specification of nonlinear parameter values of the candidate models such as the likelihood-ratio tests in assessing treatment effects over time is explored. \nAlso\, “Comparison of Confidence Intervals for Exposure-adjusted Incidence Rate Differences” will be presented by Lina Yang\, a Biostatistics PhD student who completed a summer internship at the FDA. \nAdverse event rates in clinical trials are often summarized by crude percentages\, which are calculated as the number of subjects experiencing the adverse event of interest divided by the total number of subjects exposed to (or randomized to) the drug. A limitation of crude percentages is that they do not account for treatment exposure or differential follow-up times between subjects and can result in an inaccurate estimate of event rates. The exposure adjusted incidence rate (EAIR) is a measure of risk that attempts to adjust for time of exposure. Yang’s project focused on situations where two groups are compared with respect to event rates when there is either a single event for each subject\, or only the first event is of interest. We compared seven methods for estimation of confidence intervals for differences between two EAIRs. Coverage properties and confidence-interval widths are investigated for these methods under different distributional assumptions for the time to event\, under varying event rates and different fractions of missingness. \nJoin the seminar via Zoom \nParticipants may also join by calling 929-205-6099\, meeting ID 322 043 373. If prompted\, enter passcode 167660.
URL:https://pennstatehealthnews.org/event/webinar-biostatistics-internship-presentations/
LOCATION:Webconference
CATEGORIES:Lectures/Seminars/Academic Presentations
ORGANIZER;CN="Department%20of%20Public%20Health%20Sciences":MAILTO:pennstatepublichealth@phs.psu.edu
END:VEVENT
END:VCALENDAR