Compare Incidence from Time-Varying Exposure

Advantage of Using Incident Rates

There are a number of advantages to using incident rates. One of them is to look at incidence with time-varying exposures in persons while exposed and not exposed..

Need to Determine Time-Varying Exposure

Research question: In a Medicaid database is there an association between use of non-aspirin non-steroidal anti-inflammatory drugs (NSAID) and coronary artery disease (CAD)?

How would you study the relationship between NSAID use and CAD?

This is a common type of problem in cohort studies that examine exposures which can change over time, such as smoking, changes in diet, exercise, medications, etc. How would you analyze such data using the cumulative incidenceapproach?

Calculating Stratified Person-Time Incidence Rates

For persons followed in a cohort some potential risk factors may be fixed but some may be variable Adding up person-time in an exposure category to get a denominator of time at risk is one way to deal with risk factors that change over time.

Numerator and Denominator

For exposures that are time-varying, adding up the total person-time of exposure and non-exposure to form two denominators is a very useful way of dealing with the problem of changing exposure groups.

The numerators for each denominator then become the number of events that occurred among persons at times when they were exposed and similarly for the events among the unexposed.


This approach assumes that the exposure’s effect on the event is predominately during the time the exposure is occurring. It therefore makes sense for exposures whose association with the outcome quickly disappears when the exposure is removed.

This assumption would not work very well for exposures which have lasting effects long after the actual exposure is removed. For an exposure like smoking with known long-term health risks, linking events to time periods when an individual was and was not smoking would probably assume too close a relationship between the time of exposure and the outcome.

Example: NSAIDs and CHD

Ray (2002) analyzed Medicaid recipients in Tennessee over a ten-year period to determine if there is a relationship between use of steroidal anti-inflammatory drugs (NSAIDs) and coronary heart disease (CHD).

They calculated incidence rates during times of use and non-use that account for time of exposure. (The assumption is that NSAID use will have an immediate effect during usage in lowering the incidence of cardiovascular events.)

Rate for NSAID use = 12.02 per 1000 pers-yrs

Rate for non use = 11.86 per 1000 pers-yrs

Rate ratio = 1.01

Concluded no evidence that NSAIDS reduced risk of CHD events.


Ideally, this is a question that should be resolved by a clinical trial, but a clinical trial of this question may never be done. In the absence of a randomized trial, an observational cohort study is the second best choice.

Again, a prospective cohort with better measurements of all the potential confounders, in particular aspirin use, would be preferable, but to get the numbers required would mean a very large cohort followed for a number of years. Possible but very expensive.

Analyzing existing data is less desirable, but it does provide an opportunity to assemble a cohort analysis on a large number over many years at minimal expense.

The question that remains is whether they were able to get adequate control of confounders.

Jeff Martin, MD

-- MaryB - 05 Mar 2009