Review Question - QID 108449

This improvement in patients who received no intervention is most likely explained by the Hawthorne effect, which states that participants will change their behavior when they know that they are watched.

Errors in studies due to bias can be divided broadly into three categories: patient selection, study performance, and result interpretation. In this case the question stem provides that the patients are selected randomly and are then randomized into groups with no significant differences between groups. There is also no reason to suspect therefore that any issues arose during the selection process. The question stem does not address the intervention group nor does it propose any correlation or conclusions. Therefore, it is unlikely that the effect is due to bias in result interpretation. Bias sustained during study performance then consists of recall bias, measurement bias, procedure bias, and observer-expectancy bias. Of these, only measurement bias could have occurred because the others are not relevant to this study design (further explanation is provided in the incorrect answers section). One type of measurement bias is the Hawthorne effect, in which subjects change their behavior when they know they are observed. In this case, the subjects likely adhered more strictly to their dietary regimen because they knew they had to report it.

Incorrect Answers:
Answer 1: The confounding effect is an error in result interpretation in which a third unmeasured factor accounts for the observed effects. In this case, no conclusions were drawn between groups and no significant differences were noted between groups on multivariate analysis.

Answer 3: Procedure bias is when two groups of subjects in a study are treated differently in ways that affect the outcome of the study. Here the question examines only one group and makes no claim about the relative effect between the two groups. Procedure bias is therefore not applicable.

Answer 4: The Pygmalion effect is where the researcher's beliefs about the efficacy of an intervention leads to behavioral changes that produce a self-fulfilling prophecy. In this case the effect was found in group with no interventions, which would be expected to be similar to the general population.

Answer 5: Recall bias is when subjects with certain outcomes are more likely to remember preceding factors. In this case, the food journal was confirmed using objective markers.

Bullet Summary:
Bias can be divided into three categories and pertain only when those appear to be compromised in a question. The Hawthorne effect requires that subjects know they are being studied and changes their behavior due to that knowledge.