Let the Data Serve the Patient: Using Big-Scale Data Collection to Improve Small-Scale Patient Interactions

Abstract

Background and context: Breast cancer is one of the most common cancers in most resource-constrained environments worldwide. Although breast awareness has improved, lack of understanding of the diagnosis and management can cause patient anxiety, noncompliance and ultimately may affect survival through compromised or delayed care. South African women attending government hospitals are diverse, with differing levels of income, education and support available. Often there is a lack of access for them to appropriate information for their cancer care. Aim: A novel bioinformatics data management system was conceived through an innovative close collaboration between Wits Biomedical Informatics and Translational Science (Wits-BITS) and academic breast cancer surgeons. The aim was to develop a platform to allow acquisition of epidemiologic data but synchronously convert this into a personalised cancer plan and “take-home” information sheet for the patient. Strategy/Tactics: The concept of a clinician “customer” was used, in which the “currency” in which they rewarded the database service was accurate data. For this payment they received the “product” of an immediate personalised information sheet for their patient. Program/Policy process: A custom software module was developed to generate individualized patient letters containing a mixture of template text and information from the patient's medical record. The letter is populated with the patient's name and where they were seen, and an personalised explanation of the patient's specific cancer stage according to the TNM system. Outcomes: Through a process of continuous use with patient and clinician feedback, the quality of data in the system was improved. Patients enjoyed the personalised information sheet, allowing patient and family to comprehend and be reassured by the management plan. Clinicians found that the quality of the information sheet was instant feedback as to the comprehensiveness of their data input, and thus assured compliance and quality of data points. What was learned: Using a consumer model, through a process of cross-discipline collaboration, where there is normally poor access to appropriate patient information and poor data entry by overburdened clinicians, a low-cost model of high-quality data collection was achieved, in real-time, by clinicians best qualified to input correct data points. Patients also benefitted from participation in a database immediately, through personalised information sheets improving their understanding of their cancer care.