Changes in Drug Pricing After Drug Shortages in the United States

Abstract

Background: Prescription drug shortages may result in substitution of less effective drugs, delays in necessary treatments, and omission of or reductions in doses (1). These shortages cause an estimated $230 million in additional costs each year (1) because of the rising prices of drugs under shortage and the higher costs of substitute drugs (1, 2). Objective: To assess the association between price changes and drug shortages across the marketplace and to evaluate how price changes differ according to the number of manufacturers that supply each drug. Methods and Findings: Using the U.S. Food and Drug Administration drug shortages database (3), we identified an active shortage of 917 drugs between December 2015 and December 2016 and extracted their generic names, National Drug Code (NDC) numbers, and shortage start dates. We used pricing data from AnalySource (First Databank) between 2005 and 2016 to extract monthly wholesale acquisition costs (WACs) for each NDC number in the month the shortage began and 12 months before and after. When we excluded NDC numbers with shortage start dates that were not listed or WACs that were unavailable the month the shortage began (n= 300 for 73 drug products), our sample included 617 NDC numbers for 90 drug products. For each NDC number, we calculated and plotted the ratio between the monthly WAC and the WAC 12 months before the shortage began. We constructed a linear mixed model that regressed this ratio against fixed effects for continuous time, the period after the shortage began, whether the drug was supplied by more than 3 manufacturers, and the route of administration; the second-order interactions between continuous time and each of these variables, respectively; and the third-order interactions among continuous time, the period after the shortage began, whether the drug was supplied by more than 3 manufacturers, and the route of administration, respectively. Random effects included a random intercept and a random slope for drug. We specified an autoregressive covariance structure to model the correlation between repeated measures for each NDC number. We selected the cutoff of more than 3 manufacturers because the U.S. Food and Drug Administration expedites review of generic drug applications in markets served by 3 or fewer manufacturers. Analyses were conducted with SAS, version 9.4 (SAS Institute). All price increases reported are averages that were calculated on the basis of average WACs. Prices of all drugs increased by 7.3% and 16.0% in the 11 months before and after the shortage began, respectively (we report increases 11 months before and after shortage initiation because the shortage may have already affected WACs in the month it began [Figure 1]). Prices of drugs supplied by 3 or fewer manufacturers increased by 12.1% and 27.4% in the 11 months before and after the shortage began, respectively. For drugs supplied by more than 3 manufacturers, prices increased by 2.5% and 4.8%. Mixed models showed that the shortage had a greater effect on prices of drugs supplied by 3 or fewer manufacturers (P= 0.002). Figure 1. Observed increases in average WACs in the 12 mo before and after the shortage began. Error bars indicate 95% CIs. The WAC of tiopronin increased by 1900% 11 mo after the shortage began, which created an outlier. When we excluded this drug from the sample, the WACs of all drugs increased by 28% (95% CI, 23%34%) and 29% (CI, 23%34%) of the original WACs 11 and 12 mo after the shortage began, respectively; WACs of drugs supplied by 3 manufacturers increased by 46% (CI, 36%56%) and 46% (CI, 36%56%), respectively. WAC = wholesale acquisition cost. In the 11 months after the shortage began, the expected price increase for all drugs was 20% compared with 9% in the absence of a shortage (Figure 2). Expected price increases under shortages were also 13 percentage points higher than those in the absence of a shortage for drugs supplied by 3 or fewer manufacturers (an increase from 17% to 30%) and 8 percentage points higher for drugs supplied by more than 3 manufacturers (an increase from 1% to 9%). Figure 2. Expected increases in average WACs with and without drug shortages. Error bars indicate 95% CIs. WAC = wholesale acquisition cost. Discussion: Prices for drugs under shortage between 2015 and 2016 increased more than twice as quickly as they were expected to in the absence of a shortage. We could not assess reasons for these increases; however, they may reflect manufacturers opportunistic behavior during shortages, when the imbalance between supply and demand increases willingness to pay. Because we used WACs, whether rebates increased in concert with price changes is not known; nevertheless, it is unlikely that rebates increased exactly when shortages began. Our results are consistent with other research showing increases in patient and insurer pharmaceutical spending related to shortages (4). Prescription drug shortages are a public health crisis that requires attention from policymakers, who should be particularly concerned about the prices charged by remaining suppliers. If manufacturers are observed using shortages to increase prices, public payers could set payment caps for drugs under shortage and limit price increases to those predicted in the absence of a shortage (2).