Background: Numerous prescription products have become available over the counter (OTC) in recent years. Previous simulation models have shown the Rx-to-OTC switch of loratadine to be cost effective.

Objective: To empirically assess the overall effect of the Rx-to- OTC switch of loratadine and the specific effect of different pharmacy benefit structures on prescription drug utilization and cost among different plan sponsors.

Methods: Data from a national pharmacy benefit management organization covering lives throughout the United States were used. The analysis included a comparison of the before and after change in prescription utilization and cost for plan sponsors that instituted 1 of 3 second-generation antihistamine (SGA) benefit responses: made no change, moved SGAs to the third tier, or restricted SGA benefits through a requirement for prior authorization. Multivariate regression analysis was used to control for differences across the study groups.

Results: There was a substantial decrease in utilization and cost of all prescription drugs and combinations of drug classes. Patients with allergic rhinitis facing restricted prescription benefits for SGAs did not appear to increase utilization of other allergic rhinitis medications or other medications used to treat comorbid conditions such as asthma, sinusitis, and otitis media.

Conclusions: Utilization and cost decreased substantially for all types of medications and all pharmacy benefit structures. Future studies need to examine the effect of the Rx-to-OTC switch of loratadine and resultant prescription benefit policies on medical utilization and OTC antihistamine utilization.

(Am J Manag Care. 2005;11:374-382)

Numerous prescription products have become available over the counter (OTC) in recent years, and there are several reasons why this number may increase in the future.^1-3 After years of use, some prescription drugs have the proven safety record needed for OTC status.^3,4 Patients are also becoming more interested and involved in the management of their health. Survey estimates suggest that US adults are becoming more likely to self-treat their medical conditions than in the past.⁵ In addition, the escalating economic burden of providing insurance coverage for pharmaceutical products has prompted payers to shift cost to patients.⁶ For example, in 2001, BlueCross of California, a healthcare insurer, initiated a citizens' petition to the Food and Drug Administration to request the OTC sale of second-generation antihistamines (SGAs). This was the first time that a health insurer initiated a request to transfer prescription drugs to OTC status. Most payers do not provide coverage for OTC medications and see them as a means of shifting medication cost to the consumer. The cost shifting of Rx-to-OTC drugs typically takes the form of higher prescription copayments for remaining prescription substitutes, prior authorization, or other restrictions.⁶ Payers may increasingly encourage OTC conversions of popular medications as they face pressure to shift the cost burden of providing safe and effective pharmaceutical products to patients.

Whether the Rx-to-OTC transition of a particular drug is desirable, however, depends on an assessment of the costs and benefits from different perspectives, including consumers, payers, and society as a whole. It is also important to examine the clinician's perspective on the effectiveness of the transition. Previous research has attempted to assess the costs and benefits of Rx-to-OTC transitions from different perspectives using empirical data and simulation modeling approaches. In general, studies^7-16 of the effect of Rx-to-OTC switches have reported cost savings to society or to payers depending on the drug assessed. However, it has been argued that the effect on patients is less clear, and some analyses have argued that the costs outweigh the benefits for certain drugs.^17-19

To date, there have been simulation models but no empirical assessments of the effect of the Rx-to-OTC switch of loratadine. Sullivan et al¹⁵ used a simulation model to assess the cost effectiveness of the Rx-to-OTC switch of loratadine from the societal perspective in the United States. They found that the Rx-to-OTC switch of loratadine improved effectiveness and was cost saving for society as a whole. The authors also conducted a cost-effectiveness analysis of different SGA benefit policies in response to the OTC availability of loratadine from a plan sponsor perspective (managed care, Medicaid, and employer).²⁰ This analysis found that more generous prescription benefits for SGAs were more expensive but were cost effective compared with more restrictive policies.

Because empirical data were not available, the authors modeled the potential changes in utilization of prescription SGAs and substitute medications based on assumptions about the elasticity of demand within specific patient populations. Specifically, the authors hypothesized the following: (1) The overall demand for prescription SGAs would decrease and shift to OTC loratadine because of improved access, ease of purchase, and lower opportunity cost of OTC medications. (2) Managed care payers would benefit from reduced prescription utilization and cost. (3) Managed care members would benefit because of the lower cost of OTC loratadine compared with the opportunity cost of obtaining an SGA prescription. (4) Restrictive SGA policies may encourage patients to seek substitute therapies and increase utilization of intranasal corticosteroids (INS) and leukotriene receptor antagonists after the OTC switch. Although not assumed in either simulation model, it has been suggested that the potential exists for exacerbation of comorbid illnesses such as asthma, sinusitis, and otitis media if patients are not capable of appropriately self-treating their allergic rhinitis (AR) with OTC loratadine.²¹

As is generally the case in simulated models, these hypotheses were based on assumptions because actual estimates were not available. It is important to examine the actual changes in utilization after the Rx-to-OTC switch of loratadine to assess their accuracy and provide estimates for future simulation models in similar disease categories. In addition to assessing the general effect of the Rx-to-OTC switch of loratadine, it is also important to examine the effect of different pharmacy benefit structures instituted in response to the switch. Previous research has shown that increasing cost sharing can have a negative effect on utilization of prescription drugs,²² and any assessment of the costs and benefits of an Rx-to-OTC switch needs to incorporate these payer responses.²⁰

Simulation models are helpful in the absence of existing data, but empirical evaluation is essential to inform policy decisions and provide actual estimates for future research. The objective of this research was to assess the effect of implementing different pharmacy benefit structures in response to the Rx-to-OTC switch of loratadine on prescription drug utilization and cost using empirical data.

METHODS

Data Source

The data source was MedImpact Healthcare Systems, Inc, a private, national pharmacy benefit management organization based in California. The organization covers members across the United States and coordinates pharmacy benefits and specialty pharmacy services. Plan sponsors include employers, unions, insurance and managed care organizations, third-party administrators, and state and federal employee programs. The plans used in this research were a subset of MedImpact Healthcare Systems, Inc, members in which clear SGA benefit changes were initiated.

Sample Selection and Study Groups

The study sample consisted of individuals with AR who were continuously enrolled from 2002 to the study end. Allergic rhinitis was identified based on utilization of at least 2 prescriptions in 2002 for INS or SGAs. Loratadine was available OTC starting in December 2002, but most health plans did not change benefits for SGAs until months later. The study period spanned 12 months before and after the index date, which was the date of the change in SGA benefits. The index date was different for most plan sponsors, generally falling between January and April of 2003. Although some plans in the United States instituted coverage of OTC loratadine, the present research did not examine this benefit change. The classification of study groups was based on 3 common responses to the OTC availability of loratadine:

Plan Sponsors That Made no Change in the Formulary Status of Second-Generation Antihistamines (No-Change Group). This group was the largest among the 3 study groups and consisted of 7 plan sponsors that were a combination of 2-and 3-tier plans. Membership in these plan sponsors ranged from 35 000 to 300 000 covered lives, with a total plan membership of approximately 1.2 million covered lives for all plans combined. Before 2003, all but 1 plan in this group had at least 1 SGA available in the second tier.

Plan Sponsors in 3-Tier Plans That Moved All Second-Generation Antihistamines to the Third Tier (Third-Tier Group). Seven plan sponsors were categorized in the third-tier group, with sample sizes ranging from 8000 to 90 000 covered lives, with a total plan membership of approximately 260 000 covered lives for all plans combined. All but 1 plan sponsor were health maintenance organization-based plans. The remaining plan was a state-based plan sponsor. Before 2003, all plans in this group had at least 1 SGA available in the second tier. The change in formulary status of all SGAs (index date) typically occurred between January 1, 2003, and February 28, 2003, while 1 plan made a change on April 1, 2003.

Plan Sponsors That Imposed a Prior Authorization Restriction for Second-Generation Antihistamines (Restricted Group). Three large Medicaid plan sponsors (2-tier plans with approximately 170 000 covered lives combined) and 1 small health maintenance organization-based plan sponsor (2-and 3-tier plans with approximately 10 000 covered lives) represented the restricted group. The restricted group was unique in comparison with the others because of the high proportion of Medicaid plan sponsors. Before 2003, all plans in this group had at least 1 SGA available on formulary without restrictions. Prior authorization restrictions went into effect between January and June of 2003.

For all plan sponsors in each of the 3 study groups, the "after" period lasted for 12 months after the index date of formulary change for SGAs. (Plan sponsor confidentiality policies prevent us from revealing individual plans and the exact date of formulary changes for the SGA drug class.) Because there was no change in SGA benefits for the no-change group, the index date used to separate the before and after periods was based on the mean index date for the third-tier and restricted groups. The number of days between January 1, 2003, and the index date was calculated for each plan sponsor in the third-tier and restricted groups. This number of days was averaged over all of the plans to derive the mean index date and was applied to each of the plan sponsors in the no-change group.

Outcome Measures and Statistical Analysis

The main dependent variables were pharmacy utilization (measured as the number of prescriptions filled per member per month [PMPM] and days supplied PMPM), the cost of the prescription to the plan sponsor PMPM, and the member prescription copayment PMPM. Second-generation antihistamines, INS, montelukast sodium, all AR medications combined (SGAs, INS, and montelukast), all asthma medications combined, all antibiotics, and all pharmaceutical products were analyzed separately.

Comorbid conditions, including asthma, sinusitis, and otitis media, are prevalent among patients with AR.²³ It has been suggested that reduced utilization of prescription SGAs, if not offset by simultaneous and appropriate self-treatment with OTC antihistamines, could exacerbate existing comorbid conditions.²¹ To measure this potential deleterious "spillover" effect of reduced utilization of prescription SGAs, utilization and cost of asthma medications, antibiotics, and all pharmaceutical products combined were examined. Asthma classes included β-adrenergic agents, β-adrenergic agent and glucocorticoid combinations, leukotriene receptor antagonists, mast cell stabilizers, oral inhaled corticosteroids, and xanthines. Antibiotic classes included absorbable sulfonamides, betalactams, carbapenems, first-through fourth-generation cephalosporins, macrolides, nitrofuran derivatives, oxazolidinones, penicillins, quinolones, and tetracyclines. The Kruskal-Wallis test was used to test for significant differences between groups in the unadjusted analyses of demographic characteristics in Table 1. Ordinary least squares regression analysis was used for the multivariate analyses of the before and after index date change in utilization and cost of prescriptions PMPM.

Independent Variables

The independent variables of primary interest were the plan sponsor policy as represented by 3 mutually exclusive dichotomous variables of the postindex pharmacy benefit structure for SGAs, including made no change, moved SGAs to the third tier, or restricted SGA benefits, as described previously (the no-change variable was used as the reference category in the multivariate regression). To control for possible confounding differences across the 3 groups that could affect utilization and cost of prescriptions, several covariates were included in the multivariate analyses. Asthma, chronic AR, and severe AR were included as dichotomous indicator variables to control for the presence of significant clinical differences in the 3 groups at baseline. Individuals were defined as having chronic AR based on more than 6 months' utilization of at least 1 AR medication in 2002. Severe AR was defined by more than 60 days' simultaneous use of 2 or more AR medications in 2002. The presence of asthma in 2002 was captured by the RxRisk asthma category.²⁴ Because the incentive structure (lower copayments, etc) and patient type (more severe and lower income) are significantly different in Medicaid plans than in commercial managed care plans, receiving health insurance through Medicaid was included as a dichotomous variable. In addition, generic burden of chronic disease in 2002 was measured by the RxRisk algorithm to control for differences across groups in the burden of chronic illness.²⁴ Age and sex were also included as covariates.

RESULTS

Baseline Descriptive Statistics

The no-change group was the largest in size, accounting for more than 75% of the study population (Table 1). The mean ages ranged from 40 to 44 years, and the sample was predominantly female (range, 60%-66%). The restricted group had the oldest individuals and a higher proportion of women. The types of plan sponsors represented in each group varied. While the no-change and third-tier groups primarily consisted of commercial plans, the restricted group was largely made up of Medicaid managed care plans. Members of the restricted plans appeared to have a higher chronic disease burden (highest RxRisk score) and greater comorbid asthma, but fewer had chronic or severe AR. A greater percentage of the restricted group's members took INS, and fewer took SGAs.

The mean copayment per SGA prescription remained unchanged for the no-change group, increased by $11.40 for the third-tier group, and decreased by $1.70 for the restricted group (Table 2). Overall, there was a substantial decrease in utilization and cost of all prescription drugs and combinations of drug classes except montelukast in this AR patient population. Second-generation antihistamine use decreased in a similar manner across all 3 groups by 3.4 to 4.8 days PMPM (range, 43%-50% reduction). The mean number of prescriptions decreased in a similar manner (range, 46%-51% reduction); however, cost to the plan sponsor was significantly lower for the third-tier and restricted groups compared with the no-change group (68% and 69% reductions, respectively, compared with a 45% reduction). The mean prescription copayment PMPM for SGAs also decreased for all groups, but members of the third-tier group experienced the lowest decrease because of the higher copayments per prescription (a 23% reduction vs a 46%-64% reduction). The decline and trends were similar for INS, asthma, antibiotics, all AR medications, and all pharmaceutical products across all study groups. The total cost of all pharmaceutical products to the plan sponsor decreased for all groups but was 2 times lower for the third-tier and restricted groups compared with the no-change group (16% and 12% reductions, respectively, compared with a 7% reduction).

Multivariate Regression Analysis

Table 3 presents a summary derived from the full regression results (utilization PMPM is shown in Table 4). After controlling for all covariates in the multivariate regression analyses, there was a larger and statistically significant decrease in mean prescription fills PMPM for SGAs in all 3 groups compared with the results in Table 3. There was a decrease of -0.184 prescriptions PMPM in the no-change group. To provide perspective, this would equate to a 65% reduction from the unadjusted baseline mean of 0.281 for the no-change group. The changes in prescription fills PMPM for the third-tier (change, -0.197) and restricted (change, -0.217) groups were statistically significantly lower than for the no-change group and would equate to 65% and 88% reductions from the unadjusted baseline means of 0.303 and 0.246 for the third-tier and the restricted groups, respectively. The trends were similar for the cost to the plan sponsor, but the magnitude percentage change compared with the unadjusted baseline was significantly greater: the decrease in mean cost PMPM was -$17.56 (-144%) for the no-change group, -$20.25 (-136%) for the third-tier group, and -$21.57 (-147%) for the restricted group.

Contrary to expectations of "substitution" and spillover effects, there was a significant decrease in utilization and cost for all other categories of prescription drugs across all 3 groups (Table 3). There was a decrease of -0.134 mean prescriptions PMPM for INS in the noc-hange (change, -122%) and third-tier (change, -103%) groups and a statistically significantly lower magnitude decrease of -0.125 (-86%) for the restricted group. Similarly, there was a decrease in the cost of INS to the plan sponsor for the no-change (-$6.70 [-150%]), third-tier (-$6.90 [-127%]), and restricted (-$6.11 [-79%]) groups. The number of prescription fills PMPM for all AR medications combined decreased for the no-change (-0.367 [-88%]), third-tier (-0.377 [-82%]), and restricted (-0.392 [-95%]) groups. Likewise, there was a more significant decrease in the cost to the plan sponsor for all AR medications for the no-change (-$26.65 [-146%]), third-tier (-$29.47 [-133%]), and restricted (-$30.19 [-125%]) groups. For all 3 groups, there was a decrease in the number and the cost to the health plan PMPM for montelukast and all asthma prescriptions. Antibiotic utilization decreased for all 3 groups, but there was no statistically significant difference between the groups and no significant change in cost. The number of prescriptions PMPM of all pharmaceutical products combined decreased for the no-change (-0.586 [-26%]), third-tier (-0.586 [-28%]), and restricted (-0.639 [-19%]) groups, and the cost to the health plan for all pharmaceutical products declined in a similar fashion for all 3 groups (-34%, -33%, and -20%, respectively).

The full regression results examining the change in the number of prescriptions PMPM for the 7 different drug categories are displayed in Table 4. Controlling for drug benefit design and other covariates, chronic AR patients decreased utilization of SGAs by twice as much as nonchronic AR patients, but there was no significant difference for severe AR patients. Asthmatic patients and Medicaid enrollees had less decreased utilization of SGAs compared with nonasthmatic patients and non-Medicaid enrollees.

Compared with patients with nonchronic and nonsevere AR, patients with chronic and severe AR decreased utilization of INS by less but had a statistically significant greater decrease in utilization PMPM of AR medications, asthma medications, and all pharmaceutical products combined (Table 4). Chronic AR patients decreased antibiotic utilization PMPM by less, but severe AR patients decreased antibiotic utilization PMPM by more than their counterparts. Utilization of asthma medications and all pharmaceutical products combined PMPM was lower for patients with comorbid asthma. The decrease in utilization of SGAs, INS, all AR medications, and all pharmaceutical products PMPM was lower for Medicaid enrollees. The trends in the full regression results were similar for the cost to the health plan and the days supplied (results not shown). For example, chronic and severe AR patients saved the plan sponsor more for SGAs (-$31.13 and -$18.86, respectively) than nonchronic and nonsevere AR patients, while Medicaid enrollees and asthmatic patients saved the plan sponsor less for SGAs (-$16.40 and -$16.25, respectively) than their comparators.

DISCUSSION

This analysis was an empirical assessment of the effect of the Rx-to-OTC switch of loratadine among different plan sponsors for the 3 most common pharmacy benefit structures in the face of a new Rx-to-OTC switch. The potential for substitution and spillover effects resulting from restricted prescription benefits for loratadine was also examined.

The results suggest that there was a substantial decrease in utilization and cost for prescription SGAs regardless of the pharmacy benefit change. Annual utilization PMPM of prescription SGAs decreased by more than 65%, and cost to the payer decreased by more than 136% for all 3 groups. Although there was a statistically significantly greater decrease in SGA utilization and cost for the third-tier and restricted groups compared with the no-change group, even individuals experiencing no change in SGA benefits after the Rx-to-OTC switch of loratadine substantially reduced utilization (change, -65%) and cost to the plan sponsor (change, -144%) in the postindex period. Similarly, members paid less to the health plan for SGAs. The mean member copayments PMPM for SGAs decreased by 23% to 64%. These results are consistent with but show a greater decrease in utilization and cost compared with the assumptions used in previous simulation models.^15,20

The results presented in Tables 2, 3, and 4 may be used by researchers involved in assessing the costs and benefits of an Rx-to-OTC switch. The percentages given may be helpful for analysts conducting simulation models to determine approximate changes in overall utilization, as well as those associated with different prescription drug benefit designs implemented as a result of a new OTC entrant.

Given the restriction on SGA benefits among some plan sponsors, it has been hypothesized that there may be a substitution effect in the prescription market (ie, patients and providers would attempt to seek alternatives to prescription SGAs). The results did not support this hypothesis. On the contrary, there was a significant reduction in utilization and cost to the plan sponsor for INS and montelukast for all groups. As a measure of the overall outcome of the hypothesized substitution effect, utilization and cost to the heath plan for all AR medications combined decreased substantially and mirrored those of the SGAs. In addition, overall utilization and cost for all AR medications combined decreased more for chronic and severe AR patients. It appears from these results that the Rx-to-OTC switch of loratadine resulted in a decrease of all AR-related prescription utilization and cost regardless of the pharmacy benefit structure of prescription SGAs after the OTC switch.

It has also been hypothesized that reduced utilization of prescription SGAs may lead to exacerbations of comorbid conditions such as asthma, sinusitis, and otitis media due to potential inappropriate self-treatment with OTC antihistamines.²¹ The results of this study suggest that reduced utilization of prescription SGAs was not associated with increases in utilization of prescription drugs used to treat these comorbid conditions. On the contrary, the results showed a significant reduction in utilization and cost for asthma, antibiotics, and all prescription drugs combined (although the decrease in cost PMPM to the health plan for antibiotics was not statistically significant). However, the use of medications is not necessarily a surrogate for episodes of exacerbation or worsening of symptoms. It is possible that the decreased utilization of medications for comorbid conditions such as asthma is a marker for lower quality of care. However, in this case, the study group experiencing the lowest quality of care would be the no-change group (because they had the greatest decrease in utilization of asthma medications). In addition, the inclusion of "all antibiotics" as an outcome may have low specificity for AR complications, given the other potential indications for their use.

Allergic rhinitis medications are likely the primary and most significant source of utilization for this study population. The sample is young and healthy and was identified based on prior utilization of SGAs or INS. It appears that OTC availability of SGAs resulted in a significant decrease in the utilization of all prescription drugs in this AR sample. It is possible that AR patients were able to obtain adequate treatment with OTC loratadine and had less need to contact the health plan for the treatment of their primary condition (AR) and, hence, were less likely to seek treatment for any reason. It is also possible, however, that secular trends such as weather and region of the country could have contributed to reduced postindex utilization for the classes of drugs examined.

The scope of this analysis is limited to the utilization of prescription drugs. It is unclear whether individuals who decreased their utilization of prescription SGAs self-treated appropriately with OTC loratadine. Previous research has documented the deleterious effects of first-generation antihistamine (FGA) use.²⁵ First-generation-associated sedation has been shown to result in increased motor vehicle crashes and occupational injuries, reduced productivity and learning, and death. It is unclear whether patients who decreased utilization of prescription SGAs simultaneously increased use of OTC FGAs. This effect would be detrimental to society and should be examined in future research. However, an examination of the price of OTC FGAs (diphenhydramine hydrocholoride [Rite Aid Corporation, Harrisburg, Pa], 50 mg, 4 times daily; cost, $15.60/mo) compared with OTC loratadine (loratadine [Rite Aid Corporation], 10 mg/d; cost, $8.50/mo) shows that OTC SGAs may be less expensive now.²⁶ This price comparison suggests that it is unlikely that utilization of OTC FGAs increased because there is no longer an economic incentive to choose OTC FGAs over OTC loratadine. (However, because of generic competition, the price of OTC loratadine has decreased from approximately $30/mo when it was first made available.)

In addition, it is unclear if the reduction in utilization of prescription SGAs found in this study was associated with an increase in medical utilization. Although utilization of prescription drugs for common comorbid conditions and all pharmaceutical products combined decreased substantially, the effect of the Rx-to-OTC switch of loratadine on medical utilization is uncertain. Future research needs to incorporate measures of medical utilization (such as physician visits and asthma-related emergency department visits).

Numerous economic studies^{7-14,18,27-29} of the effect of Rx-to-OTC switches have reported cost savings to society or payers. One study¹⁷ of consumer cost data and medical service utilization argues that 4 recent Rx-to-OTC switches did not benefit patients financially or medically. A microeconomic analysis³⁰ argues that the economic effect of the OTC switch of SGAs may not be cost saving because of the uncertain price and consumer demand response. Sullivan et al¹⁵ and Sullivan and Nichol,²⁰ however, suggested that the Rx-to-OTC switch of loratadine was cost effective for consumers, payers, and society as a whole. The present study demonstrates that the Rx-to-OTC switch of loratadine resulted in cost savings for payers (in terms of prescription utilization and cost only). In addition, given that mean copayments for prescription SGAs in commercial managed care before the OTC switch of loratadine were substantially higher than the current price of OTC loratadine, consumers appear to have benefited as well. The results of the present study support this notion because consumers facing no change in their prescription benefits significantly decreased utilization of prescription SGAs after the OTC availability of loratadine.

However, the effect of the Rx-to-OTC switch of loratadine and concomitant pharmacy benefit changes on the effectiveness of treatment or health outcomes is unclear from the results of this pharmacy-based analysis. Although patients may have been able to sustain or improve the effectiveness of treatment after the Rx-to-OTC switch, it is also possible that the reduced utilization found in this research was accompanied by worse health outcomes or lower quality of symptom relief. This is an important area of future research. It is likely that the effect of an Rx-to-OTC switch on health outcomes will differ depending on the disease associated with the switch (eg, AR vs hypercholesterolemia).

CONCLUSIONS

The results of this study show that utilization and cost to the plan sponsor decreased substantially for all types of medications after the OTC introduction of loratadine, even when there were no pharmacy benefit changes for SGAs. In addition, AR patients facing restricted prescription benefits for SGAs did not appear to increase utilization of other AR medications. Increased utilization of medications for comorbid conditions such as asthma, sinusitis, and otitis media was not seen in this pharmacy claims analysis regardless of the type of prescription benefit change for SGAs. The scope of the present analysis was limited to the utilization and cost of prescription drugs. Future studies need to examine the effect of the Rx-to-OTC switch of loratadine and resultant prescription benefit policies on medical utilization and OTC antihistamine utilization.