In order to assess whether we could feasibly model the impact of RP on prices alone, we assessed whether or not RP affected NSAID quantities. Based on the data displayed in , the introduction of RP was not associated with any marked drops in the use of NSAIDs. We formally tested this hypothesis by estimating the parameters of (1)
(which are displayed along with other parameter estimates in an online-only appendix) and, using these estimates, testing the hypotheses that (2)(3)
were both equal to 0. We could not reject these hypotheses at the 5 percent level (). Moreover, the estimated values of ((2)
) and ((3)
), also displayed in , were modest: the introductions of Type 1 RP and Type 2 RP were associated with reductions of only 1.3 and 0.04 days of NSAID therapy per senior per month, respectively.
Days of Analgesic Therapy Dispensed per 1,000 Seniors, by Analgesic Type and Month
Effect of RP on Days of Analgesic Therapy Dispensed per 1,000 Seniors over the Post-RP Period, by Analgesic Type
Notes: We fit a moving average trend line through the data points. The vertical lines indicate the introduction of the three major NSAID cost control policies in the 1990s; see text for details of these policies.
While Type 2 RP appeared to have little effect on the total volume of NSAIDs dispensed, it did have a substantial effect on reimbursement prices (see and ). Pharmacare expenditure per day of NSAID therapy dispensed dropped by almost half, from $0.80 to $0.44, after Type 2 RP, and dropped to $0.40 after the “second line restricted” NSAIDs were delisted in November 1996 (). Type 1 RP, by contrast had a much smaller effect on prices. About one-quarter of the savings that Pharmacare realized from the introduction of RP represented additional costs to nonexempt patients who elected to use higher priced NSAIDs. Patients paid nothing pre-RP but paid about $0.10 per day of NSAID therapy after RP (). Most of the savings from the introduction of RP were therefore attributable to substitutions of low cost unrestricted for higher cost restricted NSAIDs. The price differences between these NSAIDs were considerable. Prior to the introduction of RP, the unrestricted NSAIDs cost Pharmacare $0.23 per day whereas the other NSAIDs cost up to six times as much (). As indicates, rates of prescribing unrestricted NSAIDs doubled after the introduction of Type 2 RP. This increase was entirely because of the increased use of naproxen; indeed,
indicates that rates of use of the other two unrestricted NSAIDs, ibuprofen and ASA, actually declined after the introduction of RP. Before the introduction of RP, the restricted NSAID diclofenac was the most widely used NSAID; a year after Type 2 RP was introduced, however, rates of use dropped to about one-quarter of its pre-RP rate of use.
Average Expenditure per Day of NSAID Therapy Dispensed, by Month and Payer (Pharmacare and Patients)
Average Pharmacare Expenditure per Day of Therapy Dispensed, by Analgesic and Time Period
Days of NSAID Therapy Dispensed per 1,000 Seniors, by NSAID Reimbursement Category and Month
Average Monthly Number of Days of Therapy Dispensed per 1,000 Seniors, by Analgesic and Time Period
Multiplying the price reductions attributable to the introduction of RP by the quantity of NSAIDs dispensed post-RP produced (undiscounted) total savings estimates of $7.5 million for Type 1 RP and $22.7 million for Type 2 RP (). The annualized savings are $1 million (95 percent CI: $0.6 to $1.5 million) and $4 million (95 percent CI: $3.6 to $4.4 million), respectively, or about 11 percent and 44 percent, respectively, of the $9.1 million Pharmacare spent on NSAIDs for seniors in the 12 months prior to Type 2 RP. Some of the savings attributed to Type 2 RP are actually because of the delistings of the second line restricted NSAIDs. This latter policy was responsible for about $0.04 or 10 percent of the $0.40 reduction in Pharmacare reimbursement prices that eventually accrued after the introduction of Type 2 RP (). Hence, the delistings policy produced savings of about $400,000 annually (10 percent ×$4,000,000).
Effect of RP on Pharmacare NSAID Expenditures
Pharmacare savings were partially offset by increased total patient spending of $92,000 and $820,000 annually after the introduction of Type 1 and Type 2 RP, respectively (). Patient spending increased rapidly immediately after the introduction of Type 2 RP but dropped off thereafter, stabilizing at about $150 per 1,000 seniors a year after the introduction of the policy (). The rapid increase and then decrease in patient spending likely reflects payments by seniors who elected to pay out of their pocket when filling their first refill prescription for a restricted NSAID after the introduction of the policy, but who subsequently received exemption or switched to a lower cost NSAID. These figures exclude out-of-pocket costs for those seniors who did not receive a special authority exemption for second line restricted drugs used after November 1996. (As Pharmacare did not pay for any portion of these drugs, data on their use and expenditures are not available.) Moreover, our data exclude patient spending on nonreimbursed over-the-counter analgesics, such as low strength ASA.
Effect of RP on Patient NSAID Expenditures
Patient Out-of-Pocket Drug Expenditures per 1,000 Seniors, by Analgesic Category and Month
Pharmacare savings from RP might also be mitigated by increased spending on more costly analgesics, such as opiates, which cost Pharmacare about $1 per day of therapy prior to the introduction of Type 1 RP (). According to the estimates presented in , and as is evident from , Type 1 RP had only negligible effects on the prescribing of other analgesics, but there was an increase in rates of opiate use after the introduction of Type 2 RP in the order of 492 days of therapy per month per 1,000 patients. It is unclear from our aggregate data if opiates were being used as a substitute for NSAIDs, or if they were increasing for other reasons. Whatever the cause, the increased opiate use did not translate into increased Pharmacare expenditures because of offsetting reductions in Pharmacare expenditure per day of opiate therapy over the same time period ().
We next assessed the effects of Type 2 RP on the pricing decisions of the firms whose drugs were used to set the reference price (the unrestricted NSAIDs) or whose drugs were no longer fully reimbursed under the policy (the first and second line restricted NSAIDs). On the basis of the data displayed in , there is no evidence that RP increased retail prices of either the unrestricted or restricted drugs. Indeed, Type 1 RP was associated with a small but sustained drop in the prices of unrestricted NSAIDs, and a larger but transitory drop in prices of second line restricted NSAIDs. Type 2 RP had virtually no effect on drug prices although the delisting of the various NSAIDs in November 1996 did coincide with a marked drop in the second line restricted drug prices. Based on the results of the models of the impact of RP on drug prices, displayed in , the introduction of Type 2 RP was associated with an average price drop of $0.23 per unit per month over the post–Type 2 RP period. As we did not control for the effect of the delistings policy in the regressions, this estimate can be interpreted as the combined influence of Type 2 RP and the delistings policy. Presumably, if this price reduction is a consequence of the delistings policy alone, then the price drop associated with this policy alone would be larger.
Average Retail Unit Prices of NSAIDs, by Their Type 2 RP Reimbursement Status and Month
Effect of RP on the Average Monthly Unit Prices of Restricted and Unrestricted NSAIDs