The Arkansas State Employee Benefits Division (EBD) is a self-insured program comprising public school and other state employees, their spouses, and dependents. Previous research published in JMCP (2006) showed drug cost savings of $2.20 per member per month (PMPM; 37.6%) or annualized savings of $3.4 million associated with a benefit design change and coverage of the proton pump inhibitor (PPI) omeprazole over-the-counter (OTC) beginning in March 2004. On May 1, 2005, brand esomeprazole was excluded from coverage, with current users grandfathered for 4 months until September 2005. Reference pricing for PPIs, including esomeprazole but excluding generic omeprazole, was implemented on September 1, 2005, and the beneficiary cost share for all PPIs except generic omeprazole was determined from comparison of the PPI actual price to the $0.90 omeprazole OTC reference price per unit. To examine PPI utilization and drug costs before and after (a) excluding esomeprazole from coverage (with grandfathering current users) and (b) implementing a therapeutic maximum allowable cost (TMAC), or reference-pricing benefit design, for the PPI class in a large state employee health plan with fairly stable enrollment of approximately 127,500 members in 2005 through 2008 and approximately 128,000 members in 2009 Q1. The pharmacy claims database for the EBD was used to examine utilization and cost data for PPIs in a longitudinal analysis for the 61-month period from March 1, 2004, through March 31, 2009. Pharmacy claims data were compared for the period 14 months prior to esomeprazole exclusion (preperiod), 4 months during the esomeprazole exclusion (postperiod 1), and the ensuing 43 months of PPI reference pricing (postperiod 2). PPI cost and utilization data for the intervention group of approximately 127,500 beneficiaries were compared with a group of 122 self-insured employers with a total of nearly 1 million beneficiaries whose pharmacy benefits did not include reference pricing for PPIs. Despite 79% of existing esomeprazole users being grandfathered during the 4-month esomeprazole-exclusion period (postperiod 1), the share of omeprazole OTC claims increased from 35.2% to 42.5% (+ 7.3 percentage points) of all PPI claims, and esomeprazole claims decreased from 16.7% to 12.0% (-4.7 percentage points), with little change in the use of other PPIs. The average allowed charge (price) per day of PPI drug therapy decreased in postperiod 1 by 8.9% from $2.81 to $2.56, while utilization increased by 2.2% from 1.83 days PMPM to 1.87 days PMPM; the net plan cost PMPM decreased by $0.40 PMPM from $3.78 to $3.38 (-10.6%), representing a reduction in spending of $35,664 per month while the average member copayment per claim was essentially unchanged. In the 43 months of reference pricing in postperiod 2, PPI utilization was essentially unchanged at 1.82 days PMPM compared with the preperiod (1.83 days PMPM) and 2.7% lower than the esomeprazole-exclusion period (1.87 days PMPM); however, price (charge per day) decreased by 38.4% during refer- RESEARCH ence pricing to $1.73 from $2.81 in the preperiod and by 32.4% compared with $2.56 in the esomeprazole-exclusion period, despite an increase in the average pharmacy dispensing fee to $5.21 per PPI claim. Net plan cost decreased by $1.87 PMPM (49.5%) to $1.91 PMPM during reference pricing compared with the preperiod ($3.78 PMPM) and by $1.47 PMPM (43.5%) compared with the esomeprazole-exclusion period 1 ($3.38 PMPM). Beneficiary costs (copayment per claim) for PPIs decreased to $1.24 PMPM ($23.27 per claim) compared with the preperiod ($1.37 PMPM, $24.95 per claim) and compared with the esomeprazole-exclusion period ($1.40 PMPM, $25.06 per claim). The reductions in net plan costs represented lower plan spending for the 43 months of reference pricing (postperiod 2) of approximately $9.4 million or an average of approximately $219,500 per month compared with the preperiod or $7.9 million (approximately $183,900 per month) compared with the esomeprazole-exclusion period. Compared with a group of self-insured health plans without pharmacy benefit reference pricing of PPIs, the cost savings over the 43-month period from September 1, 2005, through March 31, 2009, were approximately $7.2 million or $1.31 PMPM. For this state employee health plan, the policy change that excluded esomeprazole from coverage but grandfathered current users was associated with a relatively small reduction in PMPM spending on PPIs compared with the subsequent policy change that applied reference pricing to the PPI class based on the price (drug cost plus dispensing fee) for omeprazole OTC. Over 43 months of reference pricing, net plan costs fell dramatically by 49.5% PMPM compared with the preperiod or decreased by 43.5% compared with the esomeprazole-exclusion period. While utilization was essentially unchanged compared with the 18 months before reference pricing, the average pharmacy dispensing fee per PPI claim increased, and beneficiary costs PMPM decreased.