Abstract 231P
Background
Few studies have explored the patient experience in early breast cancer (eBC). This study sought to understand patient experiences (e.g. treatment experience, emotional impact, etc.) as directly expressed by patients living with eBC in public online conversations.
Methods
A mixed-methods approach was used to examine patient conversations found in public online forums to identify and evaluate themes related to eBC. Among 60,000 posts related to eBC published between September 2014 and September 2019, text from a random subset of 15,000 posts was extracted and grouped into linguistically similar and mutually exclusive clusters using an advanced natural language processing (NLP) algorithm. Each cluster represented a theme and was assigned labels using keywords identified by the NLP algorithm and input from study investigators. Clusters were characterized using four metrics: betweenness centrality (linguistic similarity to other areas of the cluster network), sentiment (general attitude toward a topic), recency (average date of posts), and volume (total number of posts). An aggregate score of the four metrics was used to rank clusters.
Results
A total of 3,906 unique users were represented in the analysis (67% and 33% obtained from cancer-specific and general health/non-health forums, respectively). Twenty-seven clusters were identified. “Emotional support from peers” (12%), “Surgical procedures” (11%), and “Understanding diagnosis & prognosis” (7.4%) were highest in volume. “Discussing recurrence & progression”, “Anxiety before treatments”, and “Skepticism of healthcare system” were most central within the network. The highest scored clusters were “Discussing recurrence & progression” and “Understanding diagnosis & prognosis”.
Conclusions
This study represents a novel application of NLP techniques to capture and process a large amount of patient-reported information. Several major themes in the patient experience related to recurrence risk, diagnosis, monitoring, and treatment options were identified. Findings suggest that patient and clinician partnerships may be strengthened by a greater emphasis on communicating the risk of disease recurrence and shared decision-making.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Pfizer Inc.
Funding
Pfizer Inc.
Disclosure
E.H. Law: Shareholder/Stockholder/Stock options, Full/Part-time employment: Pfizer Inc. P.A. Spears: Advisory/Consultancy: Pfizer Inc. R. Winnette: Shareholder/Stockholder/Stock options, Full/Part-time employment: Pfizer Inc. All other authors have declared no conflicts of interest.