Mobile Intervention for Adherence to Aromatase Inhibitors Misses Mark in Breast Cancer

Julie Gralow, MD

Twice-weekly text message reminders did not improve adherence to aromatase inhibitors (AIs) for postmenopausal patients with early-stage breast cancer; therefore, increasing the odds of long-term adherence for this population will require more personalized approaches, explained Julie Gralow, MD.

Non-adherence to AIs is not uncommon and increases the risk of recurrence in patients with breast cancer. In the randomized SWOG S1105 trial, investigators evaluated the effect that text messages had on treatment compliance, aiming to help doctors find the optimal therapy for each patient. Non-adherence was defined by a negative urine AI metabolite assay at 36 months.

A total 724 patients were assessed for non-adherence to AIs every 3 months. At 36 months, the observed adherence rate was 55.5% for those who received text messages and 55.4% for those who did not. Data also showed that there was no difference in time to adherence failure by arm, as the 3-year adherence failure rate was 81.9% in the text message–arm vs 85.6% in the control arm (HR, 0.85; 95% CI, 0.76-1.05; P = .18).

Moreover, baseline scores from the Brief Pain Inventory, Functional Assessment of Cancer Therapy-Endocrine Subscale (FACT-ES), the Treatment Satisfaction Questionnaire for Medication, and the Brief Medication Questionnaire were strongly associated with non-adherence.

In an interview with OncLive, Gralow, clinical director of Breast Medical Oncology at Seattle Cancer Care Alliance, professor of medical oncology at the University of Washington School of Medicine, and professor of the Clinical Research Division at Fred Hutchinson Cancer Research Center, discussed exciting research in HER2-positive breast cancer. 

OncLive: Could you spotlight the importance of the SWOG S1105 study?

Gralow: We know that a large number of women who are taking adjuvant endocrine therapy, whether it be AIs or tamoxifen, can't complete the recommended 5 years; other patients can receive 10 years of therapy. It’s widely understood that if patients can take the agent for a longer period of time, that there's added benefit and the toxicities build up.

The SWOG S1105 trial was an interesting study that used text messaging to try to improve adherence in women who are taking adjuvant AIs. This trial showed that twice-weekly text messages, for a total of 3 years, that were not personalized did not improve the adherence to AIs. The adherence of patients to AIs in the text-messaging arm versus the non–text-messaging arm was equal and only at about 60% [for adherence]. A lot of women dropped out for a variety of reasons.

We then went back into the study and we looked at the factors that were collected at baseline that could have predicted those who were going to drop out and those who were unable to tolerate the drug. We decided to focus on that group in terms of future studies and future interventions. 

Notably, we collected a ton of patient-reported outcomes (PROs). At baseline, we asked if they were having endocrine symptoms, joint aches, the pains that come along with AIs. There were brief validated surveys that asked about their thoughts on medications in general.

We found that younger patients tended to be less adherent after 3 years. We also found that those who already had any decreased symptoms or pain, especially joint aches, were less likely to be able to stay on over the 3 years of the study. Moreover, the surveys related to thoughts and beliefs about medications predicted for who would fall off. 

We’ve now found and sorted out a group of patients who are more adherent. Looking forward, more studies may help with adherence. We now have a group of higher-risk patients, and we can start to design interventions that will help these patients stay on the trial. Overall, this is an interesting study that will allow us to help higher-risk patients in future studies.

What did these surveys entail and how were they collected and analyzed?

We did a validated brief pain index and a brief pain inventory. We also FACT-ES in addition to the Treatment Satisfaction Questionnaire for Medication. Notably, the more questions that are asked on surveys, the fewer patients will fill them out.

Could you highlight the importance of PROs in clinical trials?

Traditional toxicity measurements are rated either by lab tests or physicians. However, I believe the real impact of any regimen is what the patient is feeling. Often, our traditional ways of grading toxicity will not capture how the patient actually feels. Thus, this is an attempt to let the patients tell us how they feel opposed to lab values or rates provided by physicians or health care providers. PROs are being increasingly incorporated into trials.

In SWOG, we have a working group on PROs and our patient advocates are included in the development, evaluation, and the conduct of all of our trials. If a trial is submitted at triage and we don't have PROs incorporated into it, they will bring it up. 

There are some very encouraging ongoing studies showing patients using the web and entering the information themselves. While we collect this information over time, we can push it to the care team and have some red flags that go off when the patient is reporting something at a certain level. We can get reports in real time, as opposed to waiting until a patient visits the clinic or when the pain is worse. 

Reference

1. Hershman DL, Moseley A, Arnold KB, et al. Predictive model of aromatase inhibitor non-adherence using patient-reported outcomes in women with breast cancer (SWOG S1105). J Clin Oncol. 2020;38(suppl 15):12019. doi: 10.1200/JCO.2020.38.15_suppl.12019