Breast Cancer
Conditions
Keywords
Breast Cancer, Vasomotor Symptoms, Machine Learning
Brief summary
Vasomotor symptoms (VMS) are a common consequence of systemic therapies for breast cancer. Breast cancer treatments can cause VMS in approximately 30% of postmenopausal women and 95% of premenopausal women with early stage breast cancer (EBC). There are many non-estrogen-based interventions available to manage VMS, including; lifestyle modifications, complementary and alternative medicine (CAM) therapies. However, a recent systematic review and meta-analysis of pharmacological and CAM interventions conducted by our team, found no single optimal treatment for VMS management in breast cancer patients. Given the complex patient, cancer and treatment variables influencing the experience of VMS, the numerous potentially effective VMS interventions available and the varying expectations for an effective intervention, the investigators believe Machine Learning (ML) is ideally suited to the analysis of this common and bothersome treatment related toxicity. The EPIC electronic medical record, and MyChart application has provided both clinicians and patients with increased tools for the documentation of health related outcomes. The investigators believe that the MyChart platform, and ML techniques can be utilized to collect, and analyze outcome data for breast cancer patients experiencing VMS.
Detailed description
Vasomotor symptoms (VMS) are a common consequence of systemic therapies for breast cancer. Breast cancer treatments can cause VMS in approximately 30% of postmenopausal women and 95% of premenopausal women with early stage breast cancer (EBC). In addition to their negative impact on quality of life, unmanaged VMS are the most common reason for discontinuation of potentially curative treatment in 25-60% of EBC patients. Estrogen replacement is a common treatment for VMS in the general population, however, it is contraindicated in breast cancer patients. There are many non-estrogen-based interventions available to manage VMS, including; lifestyle modifications, complementary and alternative medicine (CAM) therapies. However, a recent systematic review and meta-analysis of pharmacological and CAM interventions conducted by our team, found no single optimal treatment for VMS management in breast cancer patients. The investigators recently conducted a survey in 373 patients with EBC which found that while the majority of patients were interested in receiving an intervention to mitigate their symptoms, only 18% received a treatment for this problem. In addition, more than one third of patients experiencing VMS report that they are not routinely asked about their symptoms in routine follow up. Given the complex patient, cancer and treatment variables influencing the experience of VMS, the numerous potentially effective VMS interventions available and the varying expectations for an effective intervention, the investigators believe Machine Learning (ML) is ideally suited to the analysis of this common and bothersome treatment related toxicity. Prior breast cancer studies have successfully applied to ML models to examine risk of developing breast cancer, as well as breast cancer prognosis. The EPIC electronic medical record, and MyChart application has provided both clinicians and patients with increased tools for the documentation of health related outcomes. The investigators believe that the MyChart platform, and ML techniques can be utilized to collect, and analyze outcome data for breast cancer patients experiencing VMS.
Interventions
Interventions will consist of 4 classes of standard of care treatments, namely, lifestyle modifications, complementary and alternative medicine (CAM) therapies, prescription medications, or adjustment of anti-cancer therapy.
Sponsors
Ottawa Hospital Research Institute
Study design
Allocation
NA
Intervention model
SINGLE_GROUP
Primary purpose
SUPPORTIVE_CARE
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Patients over the age of 18 who have histologically confirmed breast cancer, of any stage * Patients experiencing vasomotor symptoms * While the study is intended to evaluate the feasibility of the MyChart platform, patients without a MyChart account, who are interested in participating in the study, will have access to a paper or electronic email version. As participation in the MyChart program has benefits outside of this intended study, all patients without a MyChart account will be encouraged to sign up for the service
Exclusion criteria
* Those who are unable to complete questionnaires in English
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Patient Engagement (MyChart Accessibility and User Experience) | 3 Months | Patient engagement will be defined by 60% of patients approached agreeing to participate in the study. |
| Physician Engagement (MyChart Accessibility and User Experience) | 3 Months | Physician engagement will be defined by 60% of those completing the study log to approach patients for participation in study. |
| Patient Accrual (MyChart Accessibility and User Experience) | 3 Months | Patient accrual will be defined by accruing 50 participants within 3 months. |
| MyChart Utilization | Baseline and 6 weeks | MyChart utilization will be defined as 85% of participants completing both questionnaires (the Hot Flash Problem Score and the Composite Hot Flash Score) on the MyChart interface, and 50% of enrolled participants completing both questionnaires as per study protocol. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Hot Flash Severity (MyChart Feasibility) | 3 Months | Hot flash severity (MyChart feasibility) will be assessed by the Hot Flash Problem Score, a composite score of the perceived distress, interference, and problematic nature of vasomotor symptoms (VMS) in daily life and by the composite hot flash score (assess hot flashes on a daily basis for 7 days). The researchers will assess the feasibility of using MyChart to complete hot flash severity assessments by determining the percentage of participants who complete the tools as per protocol, including the percentage of patients who complete daily assessments over the 7 day period. |
| Predicting effectiveness of interventions for VMS - machine learning | 3 Months | Utilize machine learning models, including classification and regression trees, with comparison against standard regression models, to assess for improvements in predictive power for hot flash severity. The researchers will use model explainability techniques, such as conditional dependence plots, to study the impact of specific features on the hot flash severity outcomes. |
| MyChart Feasibility in assessing effectiveness of interventions for VMS | 3 months | The investigators will assess the effectiveness of an intervention by assessing change in hot flash severity scores using the Hot Flash Problem Score, and composite hot flash score from baseline to 6 weeks post intervention. |
| Effectiveness of Interventions for VMS - Traditional Statistical Modeling | 3 Months | Analyze MyChart questionnaire response data, using traditional statistical modelling (including linear and logistic regression models) to predict change in hot flash severity outcomes in response to interventions for VMS. The severity outcomes will be based on two validated clinical tools. These tools consist of the Hot Flash Problem Score (a composite score of the perceived distress, interference, and problematic nature of VMS in daily life), and Composite Hot Flash Score (this assess hot flashes on a daily basis for 7 days). |
| Effectiveness of interventions for VMS (MyChart feasibility) | 3 Months | Effectiveness of interventions for VMS (MyChart feasibility) will be assessed by frequency of nocturnal awakenings, and toxicity data. Data will be analyzed using traditional statistics and machine learning techniques to create a preliminary model predicting VMS treatment response in individuals. |
Countries
Canada
Outcome results
None listed