Detecting Multi-Item Associations and Temporal Trends Using the WebVDME/MGPS Application DIMACS Tutorial on Statistical and Other Analytic Health Surveillance.

Slides:



Advertisements
Similar presentations
Statistics Review – Part II Topics: – Hypothesis Testing – Paired Tests – Tests of variability 1.
Advertisements

Julianne Gee, MPH Immunization Safety Office
BPS - 5th Ed. Chapter 241 One-Way Analysis of Variance: Comparing Several Means.
Departments of Medicine and Biostatistics
Business Statistics for Managerial Decision
Introduction to Risk Factors & Measures of Effect Meg McCarron, CDC.
1 Chi-Squared Distributions Inference for Categorical Data and Multiple Groups.
Data Mining AERS FDA’s (Spontaneous) Adverse Event Reporting System Division of Drug Risk Evaluation Office of Drug Safety Carolyn McCloskey, M.D., M.P.H.
Adverse Event Reporting at FDA, Data Base Evaluation and Signal Generation Robert T. O’Neill, Ph.D. Director, Office of Biostatistics, CDER, FDA Presented.
Spring INTRODUCTION There exists a lot of methods used for identifying high risk locations or sites that experience more crashes than one would.
Data Mining in VAERS to Enhance Vaccine Safety Monitoring at the FDA
Chi-square Test of Independence
Basic Business Statistics, 10e © 2006 Prentice-Hall, Inc. Chap 9-1 Chapter 9 Fundamentals of Hypothesis Testing: One-Sample Tests Basic Business Statistics.
Using ranking and DCE data to value health states on the QALY scale using conventional and Bayesian methods Theresa Cain.
Copyright © 2010 Pearson Education, Inc. Publishing as Prentice Hall Statistics for Business and Economics 7 th Edition Chapter 9 Hypothesis Testing: Single.
Chapter 8 Introduction to Hypothesis Testing
Business Statistics: A Decision-Making Approach, 6e © 2005 Prentice-Hall, Inc. Chap 8-1 TUTORIAL 6 Chapter 10 Hypothesis Testing.
1 BA 555 Practical Business Analysis Review of Statistics Confidence Interval Estimation Hypothesis Testing Linear Regression Analysis Introduction Case.
Statistics for Managers Using Microsoft® Excel 5th Edition
Postmarketing Risk Assessment of Drug Products Division of Drug Risk Evaluation Office of Drug Safety Center for Drug Evaluation and Research.
Marshall University School of Medicine Department of Biochemistry and Microbiology BMS 617 Lecture 12: Multiple and Logistic Regression Marshall University.
Bayes, Data Mining and Pharmacovigilance Patrick Graham University of Otago, Christchurch.
Confidence Intervals and Hypothesis Testing - II
Business Statistics, A First Course (4e) © 2006 Prentice-Hall, Inc. Chap 9-1 Chapter 9 Fundamentals of Hypothesis Testing: One-Sample Tests Business Statistics,
Fundamentals of Hypothesis Testing: One-Sample Tests
Chapter 12: Inference for Proportions
Simple Linear Regression
Lesson Carrying Out Significance Tests. Vocabulary Hypothesis – a statement or claim regarding a characteristic of one or more populations Hypothesis.
+ Section 10.1 Comparing Two Proportions After this section, you should be able to… DETERMINE whether the conditions for performing inference are met.
8.1 Inference for a Single Proportion
Chapter 9 Hypothesis Testing and Estimation for Two Population Parameters.
User Study Evaluation Human-Computer Interaction.
Anti-Infective Drugs Advisory Committee Meeting December 15, Data Mining Analysis of Multiple Antibiotics in AERS Jonathan G. Levine, PhD Mathematical.
Chapter 10: Comparing Two Populations or Groups
Issues in the Practical Application of Data Mining Techniques to Pharmacovigilance A. Lawrence Gould Merck Research Laboratories May 18, 2005.
Business Statistics for Managerial Decision Farideh Dehkordi-Vakil.
Copyright © 2014, 2011 Pearson Education, Inc. 1 Chapter 18 Inference for Counts.
Copyright © 2013, 2010 and 2007 Pearson Education, Inc. Section Inference about Two Means: Independent Samples 11.3.
Multiple Regression and Model Building Chapter 15 Copyright © 2014 by The McGraw-Hill Companies, Inc. All rights reserved.McGraw-Hill/Irwin.
Inference for Regression Simple Linear Regression IPS Chapter 10.1 © 2009 W.H. Freeman and Company.
9-1 Using SafetyAnalyst Module 4 Countermeasure Evaluation.
Statistics for Managers Using Microsoft Excel, 4e © 2004 Prentice-Hall, Inc. Chap 8-1 Chapter 8 Fundamentals of Hypothesis Testing: One-Sample Tests Statistics.
Chap 8-1 A Course In Business Statistics, 4th © 2006 Prentice-Hall, Inc. A Course In Business Statistics 4 th Edition Chapter 8 Introduction to Hypothesis.
BPS - 5th Ed. Chapter 221 Two Categorical Variables: The Chi-Square Test.
Section 3.3: The Story of Statistical Inference Section 4.1: Testing Where a Proportion Is.
Chapter 8 Evaluating Search Engine. Evaluation n Evaluation is key to building effective and efficient search engines  Measurement usually carried out.
Issues concerning the interpretation of statistical significance tests.
FDA Risk Management Workshop – Day #3 April 11, 2003 Robert C. Nelson, Ph.D. RCN Associates, Inc Annapolis, MD, USA.
Chap 8-1 Fundamentals of Hypothesis Testing: One-Sample Tests.
Learning to Detect Events with Markov-Modulated Poisson Processes Ihler, Hutchins and Smyth (2007)
Business Statistics, A First Course (4e) © 2006 Prentice-Hall, Inc. Chap 9-1 Chapter 9 Fundamentals of Hypothesis Testing: One-Sample Tests Business Statistics,
Good Pharmacovigilance Practices
Data Mining Consultant GlaxoSmithKline: US Pharma IT
Point Pattern Analysis
+ The Practice of Statistics, 4 th edition – For AP* STARNES, YATES, MOORE Chapter 10: Comparing Two Populations or Groups Section 10.1 Comparing Two Proportions.
INTRODUCTION TO CLINICAL RESEARCH Introduction to Statistical Inference Karen Bandeen-Roche, Ph.D. July 12, 2010.
Copyright © 2013, 2009, and 2007, Pearson Education, Inc. Chapter 10 Comparing Two Groups Section 10.1 Categorical Response: Comparing Two Proportions.
Statistical inference Statistical inference Its application for health science research Bandit Thinkhamrop, Ph.D.(Statistics) Department of Biostatistics.
Signal identification and development I.Ralph Edwards.
Postmarketing Pharmacovigilance Practice at FDA Lanh Green, Pharm.D., M.P.H. Office of Surveillance and Epidemiology June 21, 2006.
Learning Objectives After this section, you should be able to: The Practice of Statistics, 5 th Edition1 DESCRIBE the shape, center, and spread of the.
The Practice of Statistics, 5th Edition Starnes, Tabor, Yates, Moore Bedford Freeman Worth Publishers CHAPTER 11 Inference for Distributions of Categorical.
Hypothesis Testing. Statistical Inference – dealing with parameter and model uncertainty  Confidence Intervals (credible intervals)  Hypothesis Tests.
Copyright © 2013 Pearson Education, Inc. Publishing as Prentice Hall Statistics for Business and Economics 8 th Edition Chapter 9 Hypothesis Testing: Single.
Building Valid, Credible & Appropriately Detailed Simulation Models
Daniel S. Yates The Practice of Statistics Third Edition Chapter 12: Significance Tests in Practice Copyright © 2008 by W. H. Freeman & Company.
11/12 9. Inference for Two-Way Tables. Cocaine addiction Cocaine produces short-term feelings of physical and mental well being. To maintain the effect,
Measures of disease frequency Simon Thornley. Measures of Effect and Disease Frequency Aims – To define and describe the uses of common epidemiological.
PV-Trend: A JSL Application for Trending Topics for Pharmacovigilance
Data Mining AERS FDA’s (Spontaneous) Adverse Event Reporting System Division of Drug Risk Evaluation Office of Drug Safety Carolyn McCloskey, M.D., M.P.H.
Presentation transcript:

Detecting Multi-Item Associations and Temporal Trends Using the WebVDME/MGPS Application DIMACS Tutorial on Statistical and Other Analytic Health Surveillance Methods 18 June 2003 Richard Ferris

Pharmaceutical post-marketing surveillance Companies and regulatory agencies collect databases of spontaneous adverse reaction reports Relevant exposure data not readily available (the “denominator problem”) Can drug-event combinations of potential interest be identified from internal evidence alone? Approach: – Use an internally defined “denominator” – Construct set of “expected” counts using a stratified independence model

Computation of Expected Counts The expected count for a given drug-event combination is determined by the overall count for the particular drug (across all events) and the overall count of the particular event (across all drugs) For example, if 2% of all reports have PROZAC as a drug, and 3% of all reports have RASH as an event, then one would expect that 0.06% (0.02*0.03) of the reports will include this combination (PROZAC in combination with RASH) (MGPS carries out this computation separately for each distinct “stratum” and sums the strata-specific expected counts to obtain an overall expected count)

1.Comparing Observed and Expected Counts: Relative Reporting Rate Relative Report Rate (RR): RR ij = N ij / E ij Easy to interpret, easy to compute Statistically unstable if N is small or E is very small The following all have RR = 100: – N = 1000,E = 10 – N = 100,E = 1 – N = 10,E = 0.1 – N = 1,E = 0.01

2.Comparing Observed and Expected Counts: Statistical Significance What is the probability that N ij would be observed by chance (“sampling error”) when expected value is E ij ? (p-value for testing a null hypothesis) Harder to interpret (not expressed in same units as RR) Results in computation of absurdly small probabilities that have no meaning – N=100, E=1 produces ! Small RR can be very significant (small p-value) when sample size is very large: – N = 2000, E = 1000, RR = 2 is more “significant” than – N = 10, E = 0.1, RR =100

3.Comparing Observed and Expected Counts: Empirical Bayes Multi-Item Gamma Poisson Shrinker Try for best of both previous approaches – interpretability of relative rate – adjust properly for sampling variation Focus on the distribution across the set of drug-event combinations of the ratios: – Estimate ij =  ij /E ij, where  ij ~ Poisson(  ij ) Fit a parameterized “prior distribution” function (mixture of two gamma functions) to the empirical distribution of the ’s Find posterior distribution of after observing N = some value n Use this to obtain posterior estimate of expectation value of given observation of  ij This posterior estimate is what we call EBGM (Empirical Bayes Geometric Mean); also get lower and upper 95% confidence bounds (EB05, EB95). EBGM is termed the “shrinkage estimate” for RR

Multi-Item Associations vs. Pairwise Associations Consider the case of an item triplet; e.g. 2 drugs and an event RR ijk = N ijk /E ijk where E ijk is based on independence model EBGM ijk = shrinkage estimate of RR ijk Suppose a particular itemset (drug A, drug B, event C = kidney failure) is unusually frequent (EBGM for the triplet is >> 2) Important to ask: – Is this merely the result of one or more of the pairs (AB, AC, BC) being unusually frequent?OR – Is this a drug-drug interaction Compare Empirical Bayes estimate of the frequency count of the triplet to the prediction from the all-2-factor log-linear model – EXCESS2 = (EBGM * E ) – E All2F – E is the expected count from independence – Computation of E All2F uses shrinkage estimates of pairwise counts – EXCESS2 is an estimate of how many “extra” cases were observed over what was expected using the all-2-factor model Alternate approach: Define E ijk from predictions of all-2-factor model in which case resulting EBGM directly measures divergence of observed count from all- 2-factor prediction

Health Authority Adoption of Signal Detection Technologies FDA – CDER: Experimented in Office of Biostatistics with GPS for several years Validated GPS Moving to production Have published data mining results on internal web for almost all products – CBER: initial GPS implementation (VAERS) – CRADA between Lincoln and FDA to further develop methodology and tools CDC – Collaborative GPS methodology development with FDA – Includes simulation capability WHO Uppsala Monitoring Centre – Production safety signal generation mechanism using BCPNN

FDA/GPS Validation Activities Positive controls – Examine data mining results for drug-event combinations corresponding to known “labeled” adverse reactions Negative controls – Examine data mining results for several drugs (with differing safety profiles) given for the same indication “Roll back” database in time to determine when method would have provided first signal

Databases of Spontaneous AE Reports FDA Spontaneous Report System (SRS) – Post-Marketing Surveillance of all Drugs since 1969 – Dates from mid-60’s thru 1997 – 1.5 Million Reports – Encoded in COSTART FDA Adverse Event Reporting System (AERS) – US cases, serious unlabeled events from all manufacturers. – All products sold in the US ~5000 Rx’s – Replaced SRS in 1997 – Reactions coded as MedDRA PTs – Quarterly Updates, 4-6 month delay – Drugs are Verbatim – Includes initial and some follow-up reports – Includes Demographics, Reactions, Drugs, Outcomes, etc. FDA/CDC Vaccine Adverse Events (VAERS) – Stricter Laws for Vaccine Adverse Event Reporting

Signal Detection Demonstration Using VAERS Data

“Significant” EBGM and even extremely conservative EB05 with small N

Simple Rankings by Signal Strength

Evolution of Signals Over Time

Multi-Symptom Syndromes (Higher Order Associations)

The “Serotonin Syndrome” Could MGPS be used to identify unknown syndromes? Try mining the AERS data for “significant” event triples using a known syndrome. "The symptoms of the serotonin syndrome are: euphoria, drowsiness, sustained rapid eye movement, overreaction of the reflexes, rapid muscle contraction and relaxation in the ankle causing abnormal movements of the foot, clumsiness, restlessness, feeling drunk and dizzy, muscle contraction and relaxation in the jaw, sweating, intoxication, muscle twitching, rigidity, high body temperature, mental status changes were frequent (including confusion and hypomania - a "happy drunk" state), shivering, diarrhea, loss of consciousness and death. (The Serotonin Syndrome, AM J PSYCHIATRY, June 1991)

Using Simulation to Test the Signal Detection Process

Interpreting Simulation Parameters 1. As R  P and (Q-R)  (1-P) => “No Signal” 2. As R  P and (Q-R) “Strong Signal” 3. When R “No Signal” 4. When R “Rare event” Q1-Q 1-P-Q+R Q-R P-RR Outcome Exposure Yes No Yes No P 1-P 1

Using Simulation to Create a Receiver Operating Characteristic (ROC) Curve for EBGM An ROC curve displays the true-positive rate (sensitivity) versus the false-positive rate (1 – specificity) for a statistic Ran a 20 iteration simulation using P = Q = and R = (RR = 10) to check the true-positive rate Ran a 20 iteration simulation using P = 0.003, Q = and R = (RR = 1) to check the false-positive rate

ROC Curve Based on Simulated Injection of Signals

Simulating a Rare Event Sample 100,000 records from VAERS data Set P = 0.003, Q = 0.001, R = Iterate 20 Monte Carlo simulations Expect (on average): – x 100,000 = 300 “Rare Exposures” – x 100,000 = 100 “Rare Outcomes” – x 100,000 = 3 “Rare Exposure + Rare Outcome” combinations – E = (300 x 100) / 100,000 = 0.3 – RR = 3/ 0.3 = 10

Base Simulation on VAERS Data

Sample Cases From VAERS

Sample 100,000 Cases

P = Q = R =

20 Monte Carlo Iterations

RareExposure Expected N = 300

RareOutcome Expected N = 100

RareExposure + RareOutcome Expected N = 3 Expected RR = 10

Technical Details William DuMouchel. Bayesian Data Mining in Large Frequency Tables (with Discussion). The American Statistician (1999) pp William Dumouchel and Daryl Pregibon. Empirical Bayes Screening for Multi-Item Associations. Proceedings of KDD 2001.

Methodology History and Key Contributors Stephan Evans – MCA, UK – Proportional reporting ratio (PRR) with Chi 2 analyses – Simple, highly intuitive, can be calculated by hand Bate, Lindquist, Edwards et. al. – WHO Uppsala Monitoring Centre – Bayesian neural network method for adverse drug reaction signal generation Ana Szarfman, FDA (CDER) and Bill DuMouchel (ATT) – Empiric Bayes, more robust than PRR for small n MGPS method: statistical parameter is EGBM William DuMouchel. Bayesian Data Mining in Large Frequency Tables (with Discussion). The American Statistician (1999) pp William Dumouchel and Daryl Pregibon. Empirical Bayes Screening for Multi-Item Associations. Proceedings of KDD – Multidimensional analyses possible Interactions, gender and other demographic associates, syndrome identification – Can directly compare EBGM values of different drugs, as well as for a specific drug

Key Contributors (continued) WHO Collaborating Center for Internat’l Drug Monitoring: M Lindquist, M Stahl, A. Bate, R. Edwards, RH Meyboom. – Bayesian confidence propagation neural network (BCPNN). Information Component (IC) statistic is the measure of the strength of D:E relationship – Iterative approach L. Gould. Comparison and refinement of Bayesian approaches for evaluating spontaneous reports of ADRs. DIA Annual meeting, July 2001, (Denver) – EB vs BCPNN = similar results Thakrar, BT, Blesch, KS, Sacks, ST, Wilcock, K (2001) – (ISPE, Pharmacoepid. & Drug Safety 10), – PRR vs. EB= similar sensitivity, EB better at ranking events based on small N.