Actual versus projected cost avoidance for clinical pharmacy specialist-initiated medication conversions in a primary care setting in an integrated health system

BACKGROUND: Primary care clinical pharmacy specialists (PCCPSs) are positioned to promote effective, safe, and affordable medication use. Documentation of performed interventions is difficult because the diversity of performed interventions in a variety of disease states in some practice settings. Validation of cost-avoidance projections is also difficult because traditional projection methods have several limitations. OBJECTIVE: To (1) compare projected medication cost avoidance (MCA) to actual MCA for medication conversions related to hyperlipidemia, hypertension, depression, and chronic pain initiated by PCCPS, and (2) estimate medication discontinuation that might be attributable to serious adverse drug events (ADEs) possibly associated with medication conversions. METHODS: This was a retrospective, longitudinal study conducted in a not-for-profit, integrated health system comprising approximately 470,000 members. Using a portable documentation tool, PCCPSs recorded projected annual MCA for medication conversions in 4 disease conditions (i.e., hypertension, dyslipidemia, depression, and chronic pain) in the 6-month period from December 1, 2003, through May 31, 2004. Actual annual MCA for these interventions for a 1-year follow-up period was calculated using integrated, electronic data from an administrative pharmacy database. Comparisons were made between projected MCA and actual MCA. Cost was defined as actual drug acquisition cost. In addition, an assessment of serious ADEs potentially related to the conversions was undertaken by reviewing electronic medical records of converted, nonpersistent patients. RESULTS: There were 704 medication conversions for 656 patients, of which 47 (6.7%) were for members who disenrolled in the health plan during the 12 months following the medication conversion date. The total projected MCA was $327,337 in 2004 dollars, or an average of $465 per conversion. For the 657 medication conversions in 609 patients that were evaluable (i.e., the member remained enrolled through 12 months follow-up), 466 (70.9%) persisted at 12 months, 138 (21.0%) discontinued the medication or converted to an alternative therapy, and 53 (8.1%) reverted to the original medication. Drug cost information was not available for some members, leaving approximately half (n = 331, 50.4%) of the 657 evaluable medication conversions with complete cost information available. For these 331 conversions, the overall projected MCA overestimated the actual MCA by 14.1% ($24,888 in aggregate or an average of $75 per conversion, P < 0.001). For persistent medication conversions with complete cost information (n = 278), the projected MCA ($160,225) was not significantly different compared with the actual MCA ($166,546, P = 0.477). For medication conversions that reverted to previous therapy (n = 53), the projected MCA ($41,644) overestimated by 4-fold the actual MCA ($10,435, P < 0.001). There were no emergency department visits or hospital admissions related to nonpersistent medication conversions. Compared with patients who were either nonpersistent or disenrolled at the 12-month follow-up, persistent patients did not significantly differ in chronic disease score but were slightly older (mean = 62.6 years, standard deviation = 13.1 for persistent patients vs. 59.2 [SD = 15.5] for nonpersistent or disenrolled patients). CONCLUSIONS: Projected medication cost avoidance for pharmacistinitiated medication conversions is valid for the 66% of medication conversions that persist but not for nonpersistent conversions or for patients who leave the health care system. The projected medication cost avoidance overestimated the actual cost avoidance by approximately 14%, suggesting that there is opportunity for improvement in the tool used to document medication conversions to more accurately measure cost outcomes from clinical pharmacy interventions.