Monday, August 06, 2007

Hawthorne effect in transfusion research - a thorny issue?

I first came across the Hawthorne effect (HE) when taking a Masters of Education after teaching for 10 years. Teaching and then getting formal qualifications may seem bass-ackwards but that's how many health professionals do it. You get into teaching because you like it, then take another degree once you decide to make it a career.

Recently, I was surprised to see reference to the HE appear in a recent paper in the Supplement to the August 2007 issue of Transfusion. The Transfusion supplement reports the proceedings of this conference:
  • The Role of Blood Centers in Transfusion Recipient Care. Second Joint Conference of America's Blood Centers and the European Blood Alliance
The HE is usually defined as the tendency of research subjects to act atypically as a result of their awareness of being studied, as opposed to any actual treatment that has occurred, and has come to be one of the confounding variables that applies to behavioral research.

The HE got its name from a project (1924 - 1932) by researchers from Harvard Business School on the impact of improved working conditions on productivity done in the Hawthorne Plant of the Western Electric Company (now Lucent) near Chicago.

A major finding - the one that became known as the HE - was that, regardless of the experimental manipulation, worker production seemed to improve. Researchers concluded that the workers were pleased to receive attention from the researchers, who expressed an interest in them, and therefore worked harder.

Since then the HE (which I have come to call the on stage effect) has been broadened to relate to almost every kind of human behavioral research, including educational, clinical, and transfusion medicine research.

Note that the original Hawthorne research was flawed and its findings suspect. For example, the most famous and longest study (1927 - 32), and the one giving rise to the HE, involved only five women, two of whom were replaced mid-study for insubordination and slow work with faster, more compliant workers. See:

Regardless of the validity of the original research, the principle of the HE, that observation may have a significant influence on a study's outcome, seems to have life as a possible confounder in studies involving human behavior.

The HE is mentioned in this paper of the Transfusion supplement:

Tinmouth A. Reducing the amount of blood transfused by changing clinicians' transfusion practices. Transfusion 2007 Aug;47 (s2):132S-136S.

The research involved a systematic review to evaluate published literature to determine the relative effectiveness of interventions to improve transfusion practice. Researchers identified 25 studies from 1983 to 2005 that examined intervention effectiveness and provided data both before and after the introduction of the intervention. The most commonly used interventions were guidelines, education, and audit with feedback. Tinmouth concluded that

  • Published literature suggests that even simple interventions may be effective
  • Data are very limited because of the poor quality of the studies
  • Confounders include lack of controls, publication bias, and the HE

Tinmouth recommends that more randomized controlled studies be done to properly evaluate strategies to improve physician practice.

The conference proceedings includes many papers that will interest laboratory, nursing, and medical transfusion specialists and is highly recommended. Sample breakout sessions include:

  • Right patient, right blood
  • Improving blood utilization
  • Optimizing blood center-hospital relationships
  • Better blood recipient, inventory & supply chain management
As to the HE, I suspect that it illustrates something that I have always believed and regularly practice:

  • Never let the truth get in the way of a good story
Cheers, Pat