In research, stumbling upon answers or insights you weren’t necessarily looking for is always a pleasant surprise. Incidental discoveries can have big impacts – this scientific serendipity is what many find invigorating about the profession. In the case of Dr. Dawn Hershman, MD, MS, an oncologist in the Departments of Medicine and Epidemiology at Columbia University Medical Center, a clinical trial meant to assess whether texting could improve adherence to a drug regimen instead generated valuable insight as to why patients stop taking their medications. Dr. Hershman teamed up with Dr. Joseph Unger, an Associate Professor in the Cancer Prevention Program of the Public Health Sciences Division at Fred Hutchinson Cancer Center, to analyze and interpret this unexpected treasure trove of data. “If we could identify up front those patients at higher risk of non-adherence, we may be able to target certain patient groups with interventions to help them better adhere to their treatment,” said Dr. Unger. Their results were published in the Journal of the National Cancer Institute.
Dr. Hershman and Dr. Unger have a history of working together to combat disparities in cancer treatment. The two doctors are part of a study team in SWOG (Southwest Oncology Group), a member group of the National Cancer Institute’s National Clinical Trials Network, and have previously published studies showing that enrollment in clinical trials can reduce the gap in survival between the poorest and most affluent patients, but that patient outcomes are still positively correlated with socioeconomic status (SES). For this study, they also teamed up with Dr. Julie Gralow, formerly of the SCCA but now Chief Medical Officer and Executive Vice President of the American Society of Clinical Oncology, and Dr. Scott Ramsey, the Director of the Hutchinson Institute for Cancer Outcomes Research (HICOR) at Fred Hutch.
The original goal of clinical trial S1105 was to assess whether texting interventions could improve adherence to aromatase inhibitor (AI) therapies among breast cancer patients. AIs are commonly prescribed following initial treatment of hormone-sensitive breast cancer and have been shown to drastically improve patient outcomes. However, only about half of patients who are prescribed AIs complete the 5-year regimen, and “non-adherence to aromatase inhibitor therapy among breast cancer patients can be detrimental to patients’ health over the long term,” explained Dr. Unger. Ultimately, S1105 found no effect of texting reminders on patient rates of adherence to AI therapy – the authors suggest this is because a one-size-fits-all approach to interventions likely won’t be enough. However, the study did a very detailed job of collecting patient-reported outcomes (PRO) data, as well as collecting a urinary biomarker to measure adherence (more quantitative and reliable than the self-reporting often used in studies). Patient-reported outcomes (PROs) are standardized measurements that doctors can use to understand a patient’s perspective and include things like asking patients for their functional and emotional well-being, or how they would rate their side effects. By combining the biomarker data and PROs, the research team was unexpectedly able to generate a predictive model of adherence based on PROs that they think could identify patients with high-risk of non-adherence in the future and, critically, flag them as needing additional support throughout their AI regimen.
The study analyzed data from 702 patients and found fourteen baseline PRO scores that were statistically significantly associated with nonadherence. These PROs included scores of worst, least, and average pain, pain “right now” versus pain interference, functional, emotional, physical, and social/family well-being, and patient satisfaction with the treatment overall, as well as specifically in terms of side effects and effectiveness. In general, these PROs represent a highly informative quantification of the qualitative, real-life consequences of AI therapy. The authors categorized patients according to how many baseline risk factors they had to create a composite risk model, organized into quartiles. Each quartile had a 46.5% increase in the odds of nonadherence over the previous, meaning that the most high-risk patients were more than 3 times as likely to be nonadherent as low-risk patients by the end of 3 years of data collection. “This study demonstrates that patients at high risk of not adhering to their aromatase inhibitor therapy can be readily identified using patient-reported quality of life measures,” said Dr. Unger. Identifying higher-risk patients earlier on in treatment according to a risk model like this one empowers care teams to identify who might benefit from more patient-provider communication and attention to symptom management, hopefully leading to better patient outcomes overall.
This work was funded by the National Institutes of Health National Cancer Institute, the Conquer Cancer Foundation, and the Breast Cancer Research Foundation.
Cancer Consortium members Dr. Scott Ramsey and Dr. Joseph Unger contributed to this work.
DL Hershman, AI Neugut, A Moseley, KA Arnold, JR Gralow, NL Henry, GC Hillyer, SD Ramsey, and JM Unger. 2021. Patient-reported outcome and long-term nonadherence to aromatase inhibitors. Journal of the National Cancer Institute. 113(8): 989-996.