# Forwards Strategy

*Lead Author(s):* Jeff Martin, MD

## Forwards Strategy for Potential Confounders

In contrast to the

backwards strategy for potential condunders this is what is known as a forward strategy.

## Procedure for Forwards Strategy for Potential Confounders

Forward Strategy start with the variable that has the biggest

**change-in-estimate** impact when evaluated individually.

This is where you start by assessing the effects of the potential confounders one variable at a time.

- To begin, you identify the variable that has the biggest impact, the biggest difference between crude and adjusted when looked in isolation of the other potential confounders.

*THEN:*
- Add the variable with the second biggest impact
- Keep this variable if its presence meaningfully changes the adjusted estimate
- Procedure continues until no other added variable has an important impact

After you find the variable that has the biggest change in the estimate compared to the crude estimate,

- you then jointly adjust for another potential confounder,
- the one with the second biggest change in estimate when evaluated individually.
- If the addition of the this variable results in a meaningfully different change (e.g., 10%) compared to when just the first variable is adjusted for, you would keep this variable.
- If not, drop it and consider the remaining variables.
- This continues until no other added variable has an important impact.

## Advantage: Forwards Strategy

**Advantage:** Avoids the initial sparse cell problem of

backwards approach

This procedure has some advantages like that it gets away from the problem of having many sparse cells when looking at joint confounding of many variables.

## Problem: Forwards Strategy

**Problem:** Does not evaluate joint confounding effects of many variables

It has a **big disadvantage** in that it really does not fully look at the effects of joint confounding.

## Preference

If possible, we always prefer **backward procedures** because they formally assess joint confounding.

Topic revision: r2 - 11 Jun 2009 - 16:55:51 -

MaryB?