Building a National CML Registry

Background:Tyrosine kinase inhibitor (TKI) therapy has led to phenomenal improvement in overall survival for patients with CML. Since TKI therapies were introduced and their results observed, very little innovative research has been conducted in CML. However, several important research directions remain, including, evaluating attempts to truly cure CML, i.e., patients off therapy with no detectable disease, evaluating long-term side effects and improving health related quality of life for patients who are unable to discontinue TKIs, and treatment of patients who are resistant to TKIs. Given that CML is a rare disease, this can only be achieved with considerable collaboration amongst CML experts. Such a collaboration is ongoing in the US NIH-funded Life After Stopping TKIs (LAST) study, which directly led to establishing the H. Jean Khoury Cure CML Consortium (HJKC3, named after a founding member, the late H. Jean Khoury, MD). One of the goals of the HJKC3 is to establish a national CML registry. A registry would collect longitudinal patient information to meet the critical need for both early and late outcomes research. A national CML registry is likely the only viable mechanism for identifying this population, enumerating baseline information, and addressing critical questions regarding late outcomes that may otherwise never be answered. This registry would also serve several other goals. The first aim of this registry is to prospectively collect and evaluate patient and physician compliance with recommended monitoring guidelines. The second goal is to collect data on survivorship. Some CML medications are associated with increased cardiovascular risk, nephrotoxicity, and pulmonary complications. Ensuring adequate medical monitoring of these long-term side effects and appropriately intervening is crucial to improve survival in patients with CML. The third goal is collecting patient reported outcome data directly from patients and linking that to compliance, complications, and clinical outcomes data. Lastly, this registry would inform potential clinical trials.Methods:After obtaining IRB approval, a pilot CML registry was developed at our institution with the goal of establishing the architecture for the infrastructure for a national CML registry. During this pilot we had two primary aims: first, to establish the most practical and efficient method for clinical data collection and second, to build and pilot test patient-reported outcome data collection. Retrospective clinical data was collected using REDCap with two primary mechanisms to populate the data into REDCap: 1) direct extraction and transfer of the data from the EHR and 2) manual chart review and data abstraction for complex fields or fields that require clinical interpretation. Broad categories of data that can be extracted from EMR include demographics, prior medical diagnosis, medications, some lab data, and prescription fill history. Examples of those requiring manual chart review include toxicity data, response, and disease progression data. For prospective chart review and patient-reported outcome data collection, informed consent was obtained electronically. Patient-reported outcomes assessments were administered online by emailing the study subjects a link to access a secure REDCap platform hosted by the MCW (CTSI) Clinical and Translational Science Institute. Measures include PROMIS computerized adaptive tests and select additional measures assessing a broad range of physical, emotional, cognitive, and social health domains.Progress:We identified 164 patients with CML to evaluate for inclusion. Of these, 72 (44%) were non-evaluable due to death, not an established patients, or other reasons and were included retrospectively only, while 92 (56%) were identified as potentially eligible for retrospective and prospective data collection. An invitation to join that included consent was sent to the 72 patients who had agreed to be contacted by email. Of those, 50 consented and 45 completed the baseline PROs, representing a 63% response rate. At 6 months, 39 completed the PRO assessment, representing an 87% retention rate. Data regarding human resource utilization are being collected to inform expected overall costs for a national registry. The next step is to expand to other sites in the HJKC3.

Duke Scholars

Author Kathryn Eve Flynn Population Health Sciences