OBJECTIVES Despite advances in chronic lymphocytic leukemia (CLL) treatment, roughly 25% of first and 50% of second line patients experience relapse. Relapse, however, is not well coded in claims data and is not well documented in EMR data due to under reporting of patient status, variability in terminology used to report patient status, and change in disease progression over time.  The goal of this analysis was to develop an algorithm to identify relapsed patients when patient status is not clearly documented. METHODS CLL patients in the MarketScan® Oncology EMR Database with recorded patient status were identified.  Relapse was explored using two methods: 1) recorded patient status of relapse; 2) changes in laboratory data.   For the first phase of algorithm development, both indications of relapse were compared to the date of treatment initiation. Laboratory data included lymphocytes, platelets, and hemoglobin. RESULTS Of 18,334 patients with CLL, 7,865 (43%) had any patient status reported. 528 had any mention of either relapse or remission and 73 (1%) patients had a record for relapse on the same date as a CLL diagnosis and no evidence of any other cancer types.  For these 73 patients, the date of new treatment initiation had no relationship with the date of the first recorded relapse.   Among these same patients, declines in hemoglobin and platelets, and increases in lymphocytes preceded treatment initiation by several days.  CONCLUSIONS Patient status does not appear to be updated regularly and documented status may not indicate decision to treat.  This preliminary work suggests that lab data provide a viable source for algorithm development as they are regularly reported in the EMR and for CLL are likely linked to decision to treat.  Next steps include determining the specific rule for identifying the change in lab values that triggers treatment initiation or resumption.