Payer restriction policies for pregabalin: Prior authorization
A few studies have evaluated cost impacts of PA restriction policies for pregabalin. Margolis et al. compiled data from US Medicaid records (2005–2006) of patients with pDPN or PHN and compared states with (2 states,
n = 424) versus without (4 states;
n = 5153) PA policies [
27]. States with PA policies had a smaller increase in pregabalin use compared with unrestricted states (+9.2% vs +13.6%, respectively, for a difference in difference of −4.4%,
P = 0.01), with an estimated marginal effect of a 4.0% decrease in the probability of any pregabalin use (
P = 0.02). However, the likelihood of opioid use in the restricted states (vs. unrestricted states) was significantly higher by 6.5% (
P < 0.01). In addition, other non-opioid analgesics, “other antidepressants” (i.e. bupropion, citalopram, duloxetine, paroxetine, trazodone, venlafaxine), and anxiolytics had significantly higher probability of being used in PA restricted versus unrestricted states (all
P < 0.05). In restricted versus unrestricted states, the overall annual per patient pDPN and PHN healthcare costs were $270 higher and after controlling for baseline characteristics, the relative cost increase was estimated to be $418 per patient (both
P < 0.01). Thus, PA restricted states had relative increases (rather than cost savings) in total PHN and pDPN expenditures (including pharmacy costs) accompanied by greater opioid utilization.
In another study, six health plans with (
n = 2084) and six plans without (
n = 1320) PA restriction policies for pregabalin (from 2005 to 2007) were evaluated for potential effects on medication use and healthcare costs for treatment of pDPN and PHN [
28]. Consistent with the previous study, pregabalin utilization with PA restriction policies did not increase as much as with non-restricted plans over the post-index period (+7.5% vs +12.8% net unadjusted differences, respectively for a difference in difference of −5.3%,
P < 0.001). The PA restriction (vs. no restriction) plans had a decreased probability of pregabalin use by 5.0% (adjusted estimated marginal effect
P < 0.001). Statistically significant relative changes were observed with PA restriction plans in the probability of using other antiepileptic drugs (aside from pregabalin, relative increase of 3.7%) and non-opioid analgesic medications (relative decrease of 5.2%) (both
P < 0.05). In spite of the relatively lower increase in pregabalin utilization between PA restricted compared with unrestricted plans, no statistically significant cost savings were found in total disease-related healthcare expenditures (
P = 0.40).
Bazalo et al. simulated scenarios with and without a PA requirement for pregabalin using IMS Health prescription data for patients with FM, pDPN, PHN, and partial onset seizures over a 1-year period [
26]. This Excel-based model calculated healthcare and administrative costs for each scenario with PA plans. In the baseline analysis, PA plans had cost differences of 0.4% less than plans without PA requirements (calculated drug acquisition cost of PA $885,564 vs. non-PA $888,822). Sensitivity analyses were conducted with various assumptions (e.g. no PA administration costs, lowering PA approval rate [50% to 10%], raising pregabalin market share without PA [10% to 20%], limiting substitutable therapies to valproate sodium and gabapentin). PA plans using any one of the assumptions ranged from 0.4% to 1.9% less than the cost of no PA plans. In addition, using the assumption of 50% of patients switching to pregabalin over 1 year, a sensitivity analysis found a 0.4% higher cost with PA plans relative to no PA plans. Thus, the investigators concluded that any significant cost savings of PA policies may be offset by administration costs.
Placzek et al. conducted a retrospective, observational analysis of HealthCore Integrated Research Environment claims data (2007–2012) to assess the HCU and economic impact of PA versus no PA policies for pregabalin in patients with FM or pDPN [
29]. After propensity score matching, patients were evaluated with (FM
n = 1852; pDPN
n = 1040) and without PA requirements (FM
n = 463; pDPN
n = 260). From the pre-index date (6 months prior to the index date, defined as the first pharmacy claim for an FM or pDPN medication) to the post-index date (6 months post index date), no significant differences were found between PA and no PA plans in prescription costs (or medical costs (all
P > 0.05). Moreover, with PA versus non-PA plans, disease-related use of opioids in FM patients was more commonly used (64.6% vs. 57.9%;
P = 0.0082) and serotonin norepinephrine reuptake inhibitors (SNRIs) in pDPN patients were less commonly used (1.1% vs. 3.1%;
P = 0.0152).
Payer restriction policies for pregabalin: Step therapy
To evaluate how ST protocols might affect HCU and costs, Udall et al. conducted a retrospective observational study of Humana claims data (including ST restrictions for pregabalin) and the Thomson Reuters MarketScan Commercial and Medicare Supplemental Database (no ST restrictions) [
30]. Patients were adults (aged 18–65 years) with a diagnosis of pDPN, PHN, or FM during 2008 or 2009, with the majority having FM. This study analyzed a 1-year period each before and after the index date (the date of ST protocol implementation). Patients in ST restricted plans and unrestricted plans were matched (
n = 3876 in each cohort) on diagnosis and geographic region. Before and after ST implementation, a reduction in claims for pregabalin per patient was observed from 3.6 (SD 3.9) to 2.3 (3.6), respectively. For unrestricted plans, claims went from 2.8 (3.1) pre-index to 2.6 (3.3) post-index. Overall, the patients in ST restricted (vs. unrestricted) plans showed significantly greater year-over-year reductions in use (−2.6%;
P = 0.008) and number of claims (−1.1;
P < 0.001) for pregabalin from before to after the ST policies were implemented. Using a multiple regression model of medication utilization, the patients in ST restricted (vs. unrestricted) plans were found to have net decreased odds of using pregabalin from 2008 to 2009 (odds ratio [OR] 0.04; 95% confidence interval [CI] 0.02, 0.08;
P < 0.001), while having increased odds of using gabapentin (OR 2.6; 95% CI 1.72, 3.94), as well as TCAs (OR 2.44; 95% CI 1.48, 4.00), lidocaine or other anesthetics (OR 2.26; 95% CI 1.18, 4.34), SNRIs (OR 1.51; 95% CI 1.06, 2.14), or selective serotonin reuptake inhibitors (SSRIs; OR 2.19; 95% CI 1.42, 3.37) (all
P < 0.05). When evaluating HCU, significant increases were found with ST restricted (vs. unrestricted) plans in the use of disease-related outpatient visits (difference of +3.6%,
P = 0.022). When adjusted for demographic and clinical characteristics, the ST restricted (vs. unrestricted) health plans had associated increases in disease-related total healthcare costs (+$859;
P = 0.002). None of the other costs were significantly different between health plans, suggesting that the pregabalin ST policies are unlikely to yield cost savings.
In another retrospective observational study by Suehs et al., claims data from Humana and Thomson Reuters MarketScan databases were evaluated in older subjects (aged 65 to 89 years) with FM, pDPN, or PHN who were continuously enrolled in a Medicare Advantage Prescription Drug plan from 2008 through 2009 with at least one medical claim or pain intervention within 60 days of diagnosis. The objective of this study was to determine the HCU and expenditures associated with ST requirements for pregabalin compared with unrestricted plans over a 24-month period [
31]. From the pre- to post-index periods, the number of members who had a claim for pregabalin numerically decreased (no restrictions 2274 to 2035; ST restriction 1029 to 518) and for gabapentin increased (no restrictions 2828 to 3182; ST restriction 2414 to 3082). Relative to patients in unrestricted plans, patients under ST restrictions had a relative 2.0% decrease in pregabalin utilization and a relative 2.3% increase in the use gabapentin (both
P ≤ 0.001). A generalized linear mixed model (controlling for age and comorbidities) estimated the OR of patients in ST restricted (vs. unrestricted) plans using pregabalin to be 0.013 (95% CI 0.009, 0.019;
P < 0.001), while the odds of using gabapentin were higher at 1.908 (95% CI 1.598, 2.276;
P < 0.001). The odds were also increased in restricted plans for patients using non-opioid analgesics and SSRIs and decreased for other antiepileptic drugs and SNRIs (all
P < 0.01). With ST restricted (vs. unrestricted) plans (controlling for covariates), annual disease-related pharmacy costs were significantly higher (+$12) (both
P < 0.001). However, no significant differences were found associated with ST health plans in disease-related healthcare or medical costs (all
P > 0.05).
In the most recent ST study, Null et al. conducted a retrospective, interrupted time series analysis using data from the Humana Research Database on a ST policy for pregabalin that was implemented and later lifted (in the Medicare plan only) for patients with DPN, FM, and PHN who had a filed a pharmaceutical claim (Medicare Advantage and Pharmacy Benefits, Pharmacy Drug Plan for Medicare, or a commercial plan) [
32]. Monthly time series data were evaluated for changes in utilization of pregabalin or of therapeutic alternatives (reported as prescriptions per 100,000 members per month unless otherwise noted), as well as medical and total costs. These data were analyzed during the following time periods: 1) January 2007 to 2009 (prior to the implementation of ST), 2) January 2009 to April 2013 (ST policy in commercial and Medicare plans), 3) May 2013 to April 2014 (after ST policy was no longer in effect in the Medicare plan). Overall, prior to ST implementation, the number of prescriptions increased numerically (+11.9 prescriptions; 95% CI –4.2 to 28.1,
P = 0.148) in Medicare plans and significantly (+3.5 prescriptions, 95% CI 2.7 to 4.3,
P < 0.001) in commercial plans. After ST policies went into effect, the trend decreased in the number of pregabalin prescriptions in the Medicare plan (−16.2, 95% CI –37.8 to 5.5,
P = 0.143), and had an increasing trend after the ST policy was lifted (+8.2 prescriptions, 95% CI –18.4 to 34.7,
P = 0.546). In the commercial plan, there was a significant decrease in pregabalin prescriptions by −3.6 (95% CI –4.7 to −2.5,
P < 0.001) after the ST policy was initiated. During the period when the ST policies were lifted in the Medicare but not commercial plans, the commercial plans continued to show a non-significant decrease in pregabalin utilization by −1.3 prescriptions (95% CI –5.9 to 3.2,
P = 0.568).
In this same study, effects of the ST policy on utilization of other pain drugs were also evaluated [
32]. After the ST policy went into effect in Medicare plans, significant changes were found in the utilization per 100,000 members per month of antiepileptic drugs other than pregabalin (gabapentin) (+44.1, 95% CI 27.4 to 60.9), opioids (−52.7, 95% CI –81.5 to −24.0), and SSRIs (−34.0, 95% CI –51.8 to −16.1) (all
P < 0.001). After ST started in commercial plans, utilization of some of these drugs significantly changed, including opioids (−9.1, 95% CI –14.1 to −4.2), SNRIs (−3.9, 95% CI –6.6 to −1.3), and SSRIs (+10.1, 95% CI 6.8 to 13.5) (all
P < 0.01). After the ST policy was lifted in Medicare plans, significant decreases occurred in the utilization of antiepileptic gabapentin (−62.6, 95% CI –100.1 to −25.1), opioids (−83.4, 95% CI –161.6 to −5.2), and SSRIs (−62.8, 95% CI –106.1 to −19.6) (all
P < 0.05). During this same period when ST was lifted in the Medicare plan, the commercial plans (which still had ST policies) had significant decreases in the utilization of SSRI prescriptions (−19.4, 95% CI –31.4 to −7.4,
P = 0.002).
No statistically significant differences were found in the medical costs after the pregabalin ST policy was implemented in either Medicare or commercial plans or was lifted in the Medicare plan (all
P > 0.05) [
32]. Following ST policy initiation, total healthcare costs (per 1000 members per month) were significantly lower in the Medicare plan (−20,483.2, 95% CI –30,728.6 to −10,237.9,
P < 0.001), but the decrease was non-significant in the commercial plan (
P = 0.086). However, the overall trend showed a steady rise in total healthcare cost over the entire analysis period (implementation of ST in both plans and lift of ST in the Medicare plan) with costs in the Medicare plan rising at a steeper slope than the commercial period.
Payer restriction policies for pregabalin: Mail order requirement
Martin et al. conducted a retrospective analysis (February 1, 2010 through February 28, 2011) of the drug utilization and cost impacts of a pharmacy program that required either switching to mail order of pregabalin to avoid higher member cost sharing or changing to a lower cost alternative medication (brand or generic) that could be filled either at a retail location or via mail order [
33]. Patients (≥19 years of age) filed at least one claim for pregabalin for FM, pDPN, PHN, or partial onset seizures during the analysis period. The majority of subjects (77.6%) had FM and were female (76.7%). A logistic regression model was used to propensity score match 1218 patients in each cohort (program cohort had a mail order requirement, non-program did not) based upon demographic and other characteristics (i.e. mail order and pregabalin use, comorbidities, healthcare costs, and HCU prior to the index date [defined as the first pregabalin claim during the identification period]). Prior to the index date, no significant differences were found in the percentage of retail or mail order claims or the use of any alternative medications (all
P > 0.5). After the program start, the total number of claims for pregabalin (retail and mail order combined) decreased in the program cohort (4.66 pre-index to 3.80 post-index), but increased in the non-program cohort (4.68 to 6.16) (difference in difference
P < 0.001). In addition, a significantly larger increase was observed in the mail order claims for pregabalin in the program cohort (3.1% pre-index to 48.1% post-index) than in the non-program cohort (2.8% to 9.4%) (difference in difference
P < 0.001). There was also a greater increase in the percentage of patients who switched to gabapentin in the program cohort (21.1% pre-index to 31.0% post-index) than in the non-program cohort (16.7% to 15.9%) (difference in difference
P < 0.001), as well as a relative decrease in the use of SSRIs in the program cohort (30.1% to 27.6%) versus non-program cohort (30.0% to 30.7%) (difference in difference
P = 0.026). Those program members who switched to gabapentin were significant more likely to have high pre-index Charlson Comorbidity Index scores than those who did not switch (1.25 vs. 0.94,
P = 0.001). Moreover, during the post-index period, patients who switched to gabapentin (vs. those who did not switch) were more likely to have at least one claim for opioids (63.5% vs. 57.3%), TCAs (17.2% vs. 12.1%), or SSRIs (31.8% vs. 25.7%), as well as utilization of at least one healthcare resource, including outpatient visits (83.9% vs. 78.1%), emergency room visits (51.1% vs. 43.5%), and inpatient stays (23.0% vs. 14.8%) (all
P < 0.05).
The mean total healthcare and medical costs did not significantly increase in either cohort (all
P > 0.05) [
33]. However, the pharmacy costs were significantly higher in both cohorts during the post-index period (pre-index to post-index, respectively: program cohort $7033 to $7853; non-program cohort $7064 to $7854; both
P < 0.001), but the relative increases were comparable between cohorts (difference in difference
P = 0.888).