Methods: The Primary Care patients were attributed using the following methodology: 1. Two or more outpatient visits with a PCP within the same practice in the prior three years; (2) Member of the hospital medical group managed care plan; (3) One or more outpatient visit with a provider where a CPT code (V70.0) for preventive care services was generated as a professional bill.
Results: A total of 3146 patients were attributed for statin initiation. Each Primary Care Physician (PCP) was notified of his or her patients and asked to review the list for accuracy and appropriateness for statin initiation. The PCPs that had large fallouts (>20 patients) were asked to review the first 10 patients and then to notify the quality team of any issues he or she encountered. A total of 629 patients (20%) were reviewed by the PCPs. The remaining cases were reviewed by a support team, including a team of Pharmacists (46%), a Hospitalist (28%) and an RN (6%). A statin protocol was created with definitions to standardize and demonstrate an efficient chart-review workflow for the data validation. Patients were placed in 16 subcategories based on the indication for statins. From the 16 subcategories, 4 major themes emerged: 1. Therapy recommendation (41.99%), 2. Engage the patient at a later time (5.50%), 3. Documentation issues (15.67%) and 4. Reporting query errors (36.84%). The number of hours spent by each provider to validate the data totaled 209 hours, with the breakdown of 42 hours by PCP, 97 hours by the pharmacy team, 58 hours by the Hospitalist and 12 hours by the RN. The total cost for the data validation was $10,000, including $3000 for the Hospitalist and $7000 for the Pharmacy team.
Conclusion: The Statin Improvement Project demonstrated the need for data validation to accurately classify patients for appropriately statin utilization based on the ACC/AHA Guidelines. The number and types of data integrity issues were classified under reporting and query error (36.84%) with 1159 cases having either inaccurate PCP or diagnosis or already been on a statin or unclear indication. The computer algorithm was modified based on this feedback. Overall, in 1321 patients, statin therapy was recommended and should be initiated. The study findings may serve as a model for existing and future performance improvement programs.