Using quantitative bias analysis with epidemiologic data: Short course at SACEMA, 21-23 May 2014. Register now!

Posted on Wed, Dec 18 2013 12:32:00

by Matthew Fox, DSc, MPH 1,2,3

1 Center for Global Health and Development, Boston University
2 Department of Epidemiology, Boston University School of Public Health
3 Health Economics and Epidemiology Research Office, University of WitwatersrandIntroduction

US_AID_logo     SACEMA_logo     

Introduction

Dr. Matthew Fox of the Department of Epidemiology and the Center for Global Health and Development at Boston University will be presenting an intensive three-day course on using quantitative bias methods with epidemiological data, at Stellenbosch University under the auspices of the South African DST/NRF Centre for Epidemiological Modeling and Analysis (SACEMA). The course will take place from 9 am to 4 pm daily, 21-23 May 2014, in the STIAS Library, adjacent to SACEMA.

The course fee includes a copy of the recent book on this subject by Lash, Fox and Fink. The fee, including refreshments, lunches and social events, is R3500 for early bird registration by 28 February 2014, and R4400 for later registration. Of this, R500 is a non-returnable registration fee. For international participants, the course fee is 400 euros for early bird registration, and 450 euros for late registration. Of this, 50 euros is a non-returnable registration fee. Accommodation, breakfast and dinner is not included in the course fee, but information about accommodation packages may be obtained from SACEMA.

The Registration Form is available here. Enquiries may be directed to the SACEMA Research Manager, Ms Lynnemore Scheepers (and copy to Matthew Fox) or phone: +27(0)21 808 2589/2893.

Course overview

Students of epidemiology are well versed in ways to reduce systematic error (bias) in the design of their studies and to describe random error in the analysis of their studies through confidence intervals and p values. However students are rarely taught methodologies for quantifying systematic error in their studies. Quantitative bias analysis (QBA) provides a methodology for assessing the impact of bias on study results by making assumptions about the bias parameters. QBA allows for assessment of both the direction and magnitude of systematic error and gives an estimate of effect (or a series of estimates of effect) that would have occurred had the bias been absent, assuming the bias parameters are correct. Such analyses allow investigators to go beyond speculation about the bias in discussion section of manuscripts and can be a powerful tool for quantifying the impact of such biases.

Based on the book co-authored by Dr. Fox*, this 3 day workshop will cover simple and multidimensional bias analysis methods that can be used to gain a better understanding of the impact of unmeasured confounding, selection bias and misclassification (measurement error) on study results. These methods can be applied to nearly any dataset, even summary data presented in the literature. Such approaches lay the foundation for more complicated methods, but by themselves, they act as if the bias parameters are known with certainty. We will then continue with probabilistic bias analysis which requires specification of probability distributions about the bias parameters and then uses Monte Carlo simulations methods to create intervals accounting for the uncertainty in the systematic error. Finally we will finish with methods for combining the systematic error to create simulation intervals that account for the total error (systematic and random) in the study results.

* Lash TL, Fox MP, Fink A. Applying Quantitative Bias Analysis to Epidemiologic Data Springer 2009.

Course details

The emphasis of this course will be on the basic methodologies and on applying the methods to data. We will not emphasize statistical theory, but rather the simple mathematics required to correct for study errors. All aspects of quantitative bias analysis, including probabilistic bias analysis, will be demonstrated using freely available software in Microsoft Excel. A few examples in later sessions will be demonstrated in SAS only to allow participants to understand ways to continue with their training in QBA. However we do not expect students to be proficient in SAS. Examples can be translated to STATA by participants after the course. All participants are expected to bring a laptop with Microsoft Excel. It is preferable if students bring a dataset with them that has sources of random error.

Draft timetable:

(subject to change)

Session

Morning (9:00-12:00)

Afternoon (13:30-15:30)

Wednesday

Introduction to QBA and simple bias analysis methods (selection bias)

Simple bias analysis II (uncontrolled confounding)

Thursday

Simple and Multidimensional Bias Analysis (misclassification/information bias)

Probabilistic bias analysis (probability distributions, individual vs. summary level datasets, creating summary intervals)

Friday

Multiple Bias Analysis and combing systematic and random error

Best practices for quantitative bias analysis/conclusions and similarities to Bayesian analysis

 

The Registration Form is available here.