ABSTRACT
OBJECTIVE: Improve the safety of methotrexate use in nursing home residents by reducing methotrexate errors. DESIGN: Concurrent cohort analysis. SETTING: Long term care facilities. PARTICIPANTS: Residents who received methotrexate from January 1, 2007, to December 31, 2009. INTERVENTION: A 3-pronged approach involving modification to dispensing systems and practices, mandatory staff training, and measurement was implemented in June 2008 and monitored through December 2009. Software programming to the pharmacy operating systems occurred forcing a mandatory second clinical review of all methotrexate orders during the pharmacist verification process, before dispensing. Pharmacists were required to call and clarify orders that failed to fulfill prespecified safety criteria before approving the prescription for dispensing. All pharmacists were required to complete a brief, concise, focused, mandatory training program that emphasized the proper use, adverse effects, boxed warnings, appropriate dosing schedules, and new dispensing requirements for methotrexate. MEASUREMENTS: On a daily basis, methotrexate orders from the previous day were summarized and forwarded to a Clinical Intervention Center for analysis and measurement. Prescriptions that triggered preestablished safety concerns were triaged back to their respective pharmacies for verification or modification. The results of the Methotrexate Safety Program were measured by tracking the number of prescriptions filled, number of patients treated, number of sentinel events, and number of safety variances identified. RESULTS: All assigned pharmacists (n = 2293) completed the mandatory training between June and December 2008. In 2009, a total of 369 new employees completed the training. The prescriptions per year and patients treated per year remained comparable, whereas the number of sentinel events decreased from 3 in 2007 and 4 in 2008 to 0 following program implementation. The most prevalent variance was daily dosing of methotrexate when weekly was intended. The measurement process detected and averted 497 variances in 2008 and 693 variances in 2009 that could have resulted in sentinel events. CONCLUSION: Implementation of intensification of dispensing practices, mandatory training, and measurement eradicated sentinel events associated with methotrexate in nursing homes.