Subgroup Analysis Using Multiple Linear Regression: Confounding, Interaction, Synergism

When the size of the study permits, important demographic or baseline value-defined subgroups of patients can be studied for unusually large or small efficacy responses; e.g. comparison of effects by age, sex; by severity or prognostic groups. Naturally, such analyses are not intended to “salvage” an otherwise negative study, but may be may be helpful in refining patient or dose selection for subsequent studies.1Most studies have insufficient size to assess efficacy meaningfully in subgroups of patients. Instead a regression model for the primary or secondary efficacy-variables can be used to evaluate whether specific variables are confounders for the treatment effect, and whether the treatment effect interacts with specific covariates. The particular (statistical) regression model chosen, depends on the nature of the efficacy variables, and the covariates to be considered should be meaningful according to the current state of knowledge. In particular, when studying interactions, the results of the regression analysis are more valid when complemented by additional exploratory analyses within relevant subgroups of patients or within strata defined by the covariates.


Regression Weight Linear Regression Line Efficacy Variable Residual Term Residual Standard Deviation 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Department of Health and Human Services, Food and Drug Administration. International Conference on Harmonisation; Guidance on Statistical Principles for Clinical Trials Availability. Federal Register, 63 (179), 1998: 49583–49598.Google Scholar
  2. 2.
    Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: Wiley, 1989.Google Scholar
  3. 3.
    Box Cox, Statistical Software, University Leyden, Netherlands, 1999.Google Scholar
  4. 4.
    Jukema AJ, Zwinderman AH, et al for the REGRESS study group. Effects of lipid lowering by pravastatin on progression and regression of coronary artery disease in symptomatic men with normal to moderately elevated serum cholesterol levels. The Regression Growth Evaluation Statin Study (REGRESS). Circulation. 1995; 91: 2528–40.Google Scholar
  5. 5.
    Rao CR. Linear Statistical Inference and Its Applications. New York: Wiley, 1973.MATHGoogle Scholar

Copyright information

© Springer Science + Business Media B.V. 2009

Personalised recommendations