AI framework analyzes how online interactions impact treatment decisions for cancer patients

Australian researchers are encouraging physicians to consider how online support groups may impact a cancer patient’s decision-making following a study that used a machine-learning framework to analyze patient interactions within online groups.

“These ecosystem-like interactions indicate the self-sufficient nature of OSGs (online support groups) where patient voices are prominently and equally represented,” study author Daswin De Silva, with La Trobe University, Victoria, Australia, et al. wrote. “Thereby, it is timely and relevant for primary care providers to accept OSGs as an adjunct to cancer care and consider participating in OSGs through artificial intelligence enabled [optimized] moderation and streamlined intervention.”

Researchers with La Trobe University and Austin Hospital in Victoria, Australia, developed the Patient Reported Information Multidimensional Exploration (PRIME) framework, designed to investigate the impact of online social influences on a patient’s decision-making regarding treatments, recovery, side effects and emotions.

Researchers designed it to use machine learning and natural language processing techniques to analyze and collect patient information and specific interactions in the online groups.

Data was collected from 10 heavily active online support groups, with researchers focusing on prostate cancer discussions for the study. For the dataset, nearly 607,000 conversations, from more than 22,200 patients, were collected. Researchers used patients who self-disclosed their treatment and openly discussed their decision-making process. Information from a total of 6,457 patients were selected and used for the study.

The results revealed that most patients fell into three major groups:

  • autonomous (3,883 patients)––included patients who were solely driven by personal preference
  • shared (2,154 patients)––included patients whose decisions were based both on clinician recommendations and personal preferences
  • paternalistic (420 patients)––included patients who strictly adhered to clinician recommendations

While all patient types contributed to the online support groups during the decision-making period, the shared group had the most active, consistent and prolonged interactions, according to the study. They also shared both positive and negative emotions, experience and advice for others.

The paternalistic group was more expressive, especially with negative emotions, but contributed to groups with advice for months after treatment, while the autonomous group only sought advice and contributed minimally to conversations, according to the study.

“Paternalistic and autonomous groups reduce activity soon afterward, but the shared group consistently participate in OSG discussions throughout the 12 months,” the study said.

Based on the results, researchers said online groups play an essential role in the decision-making process for cancer patients, and doctors shouldn’t ignore the impact it has when it comes to care and treatment decisions.  

“Besides fulfilling the essential role of decision support for patient-centered care, OSGs make a further paramount contribution as a medium for post-decision conversations on information exchange and emotional support,” De Silva et al. concluded.

“Although patients with similar experiences provide each other support, OSGs are peer-to-peer and unregulated which can be challenging for optimal healthcare. Therefore, healthcare providers must identify specific patient needs communicated on OSGs, in order to optimize delivery of care and ensure that patients don’t extensively depend on their peers for healthcare advice.”