gms | German Medical Science

50. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds)
12. Jahrestagung der Deutschen Arbeitsgemeinschaft für Epidemiologie (dae)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie
Deutsche Arbeitsgemeinschaft für Epidemiologie

12. bis 15.09.2005, Freiburg im Breisgau

Demonstrating intervention effects in clinical databases using change point methods

Meeting Abstract

Search Medline for

  • Tim Friede - Novartis Pharma AG, Basel, Schweiz
  • Robin Henderson - University of Newcastle upon Tyne, Newcastle, UK
  • Chung Feng Kao - Lancaster University, Lancaster, UK

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. Deutsche Arbeitsgemeinschaft für Epidemiologie. 50. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds), 12. Jahrestagung der Deutschen Arbeitsgemeinschaft für Epidemiologie. Freiburg im Breisgau, 12.-15.09.2005. Düsseldorf, Köln: German Medical Science; 2005. Doc05gmds095

The electronic version of this article is the complete one and can be found online at: http://www.egms.de/en/meetings/gmds2005/05gmds262.shtml

Published: September 8, 2005

© 2005 Friede et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by-nc-nd/3.0/deed.en). You are free: to Share – to copy, distribute and transmit the work, provided the original author and source are credited.


Outline

Text

In some medical conditions it is difficult or even impossible to study intervention effects in randomized controlled trials for ethical or practical reasons. For instance, a clinically meaningful outcome in hip replacement surgery is time to revision of the prostheses. However, a follow-up time of at least 10 years is required in order to demonstrate any treatment effects. Therefore, a randomized comparison using time to revision as an endpoint is not feasible though clinically relevant. Instead, surrogates are used in short-term trials. Alternatively, treatments are evaluated using clinical databases. The advantage of the latter is that long-term effects can be studied under real-life conditions. Problems arise through the observational nature of the data. Intervention effects are possibly confounded and their estimates therefore possibly biased. Especially in situations where a new treatment was introduced at a certain time point and all patients treated after this time point got this new treatment and all patients treated before this time point got the old treatment, naïve comparison with historic controls can be subject to considerable bias due to calendar time effects other than the intervention itself. In this situation Heuer and Abel [1] recommended to search for intervention effects over the whole study period and only conclude that the intervention had an effect if an effect could be demonstrated close in time to the introduction of the new treatment. Technically this can be done using change point methods. Recently, we proposed a method for such analyses with time-to-event data and illustrated the application using data from a clinical database of hip replacement surgeries [2]. In this presentation we will also propose a method for binary data and will apply it to a clinical database of pin site infections during use of an external fixator after orthopedic surgery.


References

1.
Heuer C, Abel U. The analysis of intervention effects using observational data bases. In: Abel U, Koch A. Nonrandomized comparative clinical studies. Duesseldorf : Symposium Publishing; 1998: 101-108
2.
Friede T, Henderson R. Intervention e_ects in observational survival studies withan application in total hip replacements. Statistics in Medicine 2003; 22:3725-3737