David Finitsis

PhD, MA, AB

Hartford Healthcare Cancer Institute

Senior Scientist, Quality of Life

IDATA (Learn more about this study)

IDATA-5

Aug 31, 2016

Assessment as intervention: Examining behavioral and anthropometric response to longitudinal dietary and exercise monitoring in older adults.

This proposal requests access to data from the IDATA study in order to test hypotheses relating to the effect of assessment on diet and exercise behavior. In general, individuals will behave differently when they know they’re being observed compared to when they believe they are not being observed. In the field of psychology this is referred to as the Hawthorne effect. The Hawthorne effect can be a uniquely difficult to account for source of bias in observational studies. Viewed orthogonally, the Hawthorne effect can also serve as an unintended intervention to promote behavior change. Health behaviors such as diet and exercise may influence individuals’ cancer risk, response to treatment, and recovery from illness. It is possible that the condition of undergoing nutritional and physical fitness assessments itself promotes improvement in dietary intake and physical activity. Health behavior measurement in both research and clinical settings has historically relied on self-report. Self-report via interview or questionnaires is a low-cost, low-barrier way to collect data irrespective of the healthcare system’s available resources. These advantages are offset however by significant risk of bias. Individuals asked to recall past behavior typically exhibit error due to imperfections of memory (reporting bias). In addition, when the behaviors have an emotional or moral valence, respondents experience additional pressure to publicly misreport, minimizing deleterious behavior and inflating their report of beneficial behavior (social desirability bias). These sources of error are generally recognized as limitations to research; they may also suppress or mask the unintended intervention effect the assessment unwittingly imparts. There is a pressing need to investigate the impact of measuring dietary intake and physical activity in order to better understand what impact if any the Hawthorne effect has on these behaviors.
If approved, the proposed project will conduct secondary analyses of data from the IDATA study. Structural equation modeling and time series analysis will model change in dietary intake and physical activity as assessed by the IDATA protocol. These analyses will test the following hypotheses: (1) variables of dietary quality and physical activity will exhibit a positive slope over time indicating improvement with respect to these variables; (2) these slopes will be increased during those months when participants underwent the wearable accelerometer and biomarker assessments; (3) owing to bias, slopes will be greater in subjective self-report variables than in variables obtained through objective data collection; (4) the magnitude and duration of these positive slope inflections will vary by anthropometric sub-type at baseline. Particular attention will be given to the sub-group of participants reporting anthropometrics carrying a higher cancer risk profile.
Increasing attention has been given to the role of physical activity and dietary intake in contributing to optimal and comprehensive cancer care. Current and future protocols and policies will assess patients’ diet and exercise using objective measures alongside or in place of subjective measurement. The above proposed secondary analyses are relevant to research and practice in this area and may help to better understand the impact of assessment itself on these variables of interest.