Date of Award


Document Type

Open Access Thesis


Epidemiology and Biostatistics


The Norman J. Arnold School of Public Health

First Advisor

Michael Wirth


Background. There are approximately 15.5 million individuals alive today in the U.S. with a personal history of cancer. For this large and growing population, ill effects associated with cancer diagnosis and treatment include both physical and psychosocial symptoms adversely affecting quality of life. One low-risk alternative to conventional pharmaceutical use in treating these symptoms is mindfulness practice. Research on self-reported measures provides strong evidence that this type of intervention improves quality of life for cancer survivors, but evidence of impact on objective measures is limited. Cortisol, the body’s primary stress hormone, is one relatively easy to measure indicator that has been increasingly favored.

Methods. In this randomized controlled trial, 38 cancer survivors from the Greenville Health System in South Carolina were assigned to either a 4-week mindfulness intervention program or a control group. Six salivary cortisol samples were collected daily at baseline, 6 weeks, and 12 weeks, along with questionnaire data assessing a variety of symptoms and lifestyle behaviors.

Both rhythm, as the change in cortisol over the day, and absolute levels of cortisol were modelled using multivariable linear regression. Specifically, for absolute levels, a treatment effect by categorization at baseline of those with high, normal, and low cortisol levels was modelled using an interaction between treatment and baseline levels of cortisol.

Results. None of the explored self-report or demographic variables assessed were significantly associated with diurnal cortisol decline at baseline in either the crude or adjusted models. Although non-significant, the intervention was associated with a greater cortisol decline over the day (-1.79 {p=0.52} and -2.71 {p=0.23} respectively for 6 and 12-week follow-up) which has been previously associated with better survival outcomes for cancer patients. No significant or consistent directional results were observed when examining effect modification of baseline cortisol categorizations on change in cortisol levels.

Conclusions. Although the effect size was relatively large, our analysis did not show a significant impact of the mindfulness intervention program on change in cortisol decline or in changes in absolute levels of cortisol as a function of baseline categorization. Our results add to the accumulating literature surrounding the use of alternative medicine practices for symptom treatment in cancer survivors and provide recommendations for future work using cortisol to assess impact of mindfulness interventions.


© 2017, Susannah Small

Included in

Epidemiology Commons