Impact of Specialty Pharmacy on Treatment Costs for Rheumatoid Arthritis
For patients who use biologic treatments for rheumatoid arthritis, specialty pharmacy care yields higher medication adherence and lower medical costs than retail pharmacy care.
Rheumatoid arthritis (RA) is a chronic and progressive inflammatory disease that can produce increasing levels of joint pain and stiffness, seriously impairing a person’s mobility and quality of life.1 Rheumatoid arthritis affects an estimated 1.3 million Americans,2 and it is a significant driver of healthcare costs for treatment of the disease and its complications.3 It is also a signifi cant source of indirect healthcare costs associated with productivity loss, absenteeism, and disability.4,5 Self-injectable biologic diseasemodifying antirheumatic drugs (DMARDs) are frequently used in combination with synthetic DMARDs or corticosteroids to manage symptoms and control disease progression in patients with moderate to severe RA.1 Biologic treatments for RA are considerably more expensive than traditional oral treatments, but for many patients they offer improved efficacy, reduced disability, and improved quality of life.1
The full benefits of self-injectable biologic treatments for RA can only be achieved when patients are adherent to their medication regimens, but many patients have diffi culty taking their medications as often as directed. Barriers to adherence may include side effects, high out-of-pocket costs, perceived lack of efficacy, or the complexities of filling prescriptions and managing medications. Patients in a usual care environment, who get a prescription from their physician and fill it at a busy retail pharmacy, may not receive the support they need in managing the issues they have with their medications. Pharmacy-based programs that provide in-depth counseling, patient education, or case-team support have been shown to improve outcomes for RA patients,6,7 and programs focused on medication therapy management can improve adherence.7,8
Specialty pharmacies provide specialized dispensing, clinical support services, and patient education for biologic products, so they are well positioned to offer therapy management programs for patients with RA. Mail-order specialty pharmacies also provide expedited refill and delivery of medications to the patient’s home or doctor’s office. Specialty pharm acy services can improve medication adherence,7,8 but their impact on medical utilization and costs for RA patients has not been studied systematically. Improvements in RA medication adherence would be expected to moderate disease progression,1,9,10 reducing the utilization of medical services and the associated costs.
The objective of this study was to evaluate the impact of specialty pharmacy management on medication adherence, medical resource utilization, and healthcare costs for patients with RA. We evaluated outcomes and costs over a 3-year follow-up period for RA patients who were exposed to mail-order specialty pharmacy services, comparing them with patients who only received their RA medications through community-based retail pharmacies. Our analysis focused on the 2 most commonly used selfinjectable biologic DMARDs for RA treatment: etanercept and adalimumab.
We conducted a retrospective cohort analysis of RA patients who filled prescriptions for etanercept or adalimumab, comparing outcomes for a cohort exposed to specialty pharmacy management with outcomes for a cohort that only filled prescriptions through community-based retail pharmacies. Medication adherence, medical utilization, and costs were tracked over a 3-year period from January 1, 2006, through December 31, 2008, using de-identified, integrated pharmacy and medical claims data.
Patients were members of prescription benefit plans managed by Medco Health Solutions, Inc (now Express Scripts), for a broad range of plan sponsors in the United States. Study participants were drawn from a population of 31,676 patients who filled at least 1 prescription for adalimumab or etanercept in 2006 (Figure 1). The study included patients who had a diagnosis for RA in 2005 or 2006 (International Classifi cation of Diseases, Ninth Revision [ICD-9] code 714.0), who had continuous drug and medical benefit eligibility for the 3-year period from 2006 to 2008, and who were aged 18 to 62 years in 2006. The study excluded Medicare patients (because medical cost data were not available for their primary benefit plan), patients who filled their target prescriptions through a pharmacy benefit manager other than Medco or through a retail specialty pharmacy (because of possible exposure to similar interventions), and patients whose data were not available for research purposes.
Patients were stratified into specialty or retail cohorts based on the pharmacy channel through which they filled their prescriptions during the 3-year period from 2006 to 2008. Patients in the specialty group filled at least 1 prescription for adalimumab or etanercept through the Accredo specialty pharmacy in 2006 and in each subsequent year that they filled a prescription for these drugs. Patients in the retail group filled all prescriptions for these drugs through retail pharmacies in 2006 and in each subsequent year that they filled a prescription. In this observational study, the pharmacy channel used to fill a prescription was influenced by plan design and patient choice; it was not constrained or assigned by the study sponsor. The inclusion criteria for the 2 study cohorts were designed to ensure consistent use of a pharmacy channel across the 3 years of the study.
Specialty Pharmacy Services
For plan members whose prescription benefi ts are managed by Medco, specialty pharmacy fulfillment and therapy management services are provided by Accredo Health Group, Inc. The pharmacy coordinates and expedites delivery of biologic RA medications to the patient’s home or doctor’s offi ce on an overnight basis. Nursing staff contact patients on a scheduled basis during the first year of treatment to discuss management of side effects (such as injection site pain), storage and administration of the drugs, and any barriers to adherence with their medications. Patients are also contacted in advance of scheduled refi lls to coordinate the refi ll process and review any questions they may have about their treatment. Pharmacists and nurses with specialized knowledge of RA medications are available by telephone to provide consultation for patients on a 24-hour basis. These services are generally not available to patients who fill their prescriptions through community-based retail pharmacies.
Medical data were taken from administrative claims databases maintained by the benefi t plan sponsors. Medical claims data included service date, place of service, ICD-9 diagnosis code, Current Procedural Terminology code, plan sponsor paid amount, and patient copayment. Charlson Comorbidity Index score11 was calculated using ICD-9 codes in each patient’s medical claims for 2006.
Drug utilization data were taken from a prescription claims database maintained by Medco. The data included prescription fill date, drug name, national drug code number, fill pharmacy name, days of supply, fill quantity, and average wholesale price. Chronic Disease Score (CDS),12,13 an overall measure of comorbidity based on drug utilization and cost, was calculated from each patient’s pharmacy claims for 2006. Pharmacy and medical claims were integrated at the patient level, and all personal identifying information was removed.
Patient eligibility and demographic data were taken from an eligibility database maintained by Medco on behalf of plan sponsors. The data included age, sex, state, and eligibility start and end dates. States of residence were classified into 4 geographic regions—Midwest, Northwest, South, and West—as defi ned by the US Census Bureau.14
Medication adherence was measured for each year of the 3-year study period using a medication possession ratio (MPR). The MPR was defined as the proportion of days covered by medication supply (adalimumab or etanercept) during the 1-year fixed interval. Patients who did not fill a prescription in 2007 or 2008 were retained in the analysis, but their MPR for that year was 0.
All-cause medical resource utilization was measured for each year of the study. Utilization measures included the percentage of patients who had 1 or more office visits, the percentage of patients who were hospitalized, the percentage of patients who had an emergency department (ED) visit, and the number of offi ce visits per patient. An ED visit was defi ned as a visit that did not exceed 24 hours; if the visit extended to more than 24 hours, it was counted as a hospitalization.
Cost outcomes included RA prescription drug costs (etanercept or adalimumab), total prescription drug costs (all drugs), and total medical costs for each year of the study. Prescription drug costs were calculated based on average wholesale price. Medical cost was the total billed cost for medical services, combining plan costs and patient copayments. Service types were outpatient visit, ED visit, hospitalization, surgery, physical therapy, and laboratory. Drug costs and medical costs were measured per patient per year. To minimize the impact of medical cost outliers, the top 1% cases were excluded from the sample.
Two-tailed t tests and χ2 tests were used to evaluate differences in baseline characteristics between the 2 study groups. For continuous outcome measures (adherence, drug cost, and medical cost), generalized linear regression models were used to compare outcomes for the 2 groups duringeach year of the study. Negative binomial models were used to test betweengroup differences in the number of offi ce visits per patient per year. Logistic regression models were used to evaluate differences in the likelihood of an office visit, hospitalization, or ED visit.
In the models for adherence and medical utilization, results were adjusted for age, sex, geographic region, CDS, Charlson Comorbidity Index score, and concomitant use of corticosteroids, nonsteroidal anti-infl ammatory drugs (NSAIDs), methotrexate, or other DMARDs during the year of the analysis. Models for 2006 included a binary independent variable (new to therapy vs existing); new to therapy was defi ned as no prescriptions for etanercept or adalimumab during the 180 days prior to the index claim date. For models evaluating costs, the covariates were age, sex, Charlson Comorbidity Index score, and region; medical cost models also included CDS. All analyses were performed using SAS statistical software version 9.2 (SAS Institute Inc, Cary, North Carolina).
A total of 4388 patients met the inclusion and exclusion criteria for the fi nal study sample—3054 in the specialty pharmacy group and 1334 in the retail pharmacy group (Figure 1). Baseline characteristics are summarized in the
. Patients who filled their RA prescriptions through a specialty pharmacy were slightly older on average than patients who fi lled their RA prescriptions through a retail pharmacy alone. Approximately 75% of patients in each group were female. Comorbidity levels in the 2 groups were similar overall. Approximately one-third of patients in each group were new to adalimumab or etanercept therapy in the first year of the study. A larger percentage of retail patients took corticosteroids, but concomitant use of methotrexate, other DMARDs, or NSAIDs did not differ significantly between the 2 groups. Differences in regional distribution were also observed.
Patients who filled their prescriptions through the specialty pharmacy had signifi cantly higher levels of medication adherence in each year of the study (
). Specialty patients showed mean adjusted adherence rates of 62.6% in 2006, 67.9% in 2007, and 60.5% in 2008. By contrast, mean adherence for retail patients averaged 49.8%, 50.6%, and 44.0% for the 3 years, respectively. These differences were statistically significant for each year (P <.0001).
Medical Resource Use
Specialty patients were significantly less likely than retail patients to have an office visit in years 2 and 3 (
, available at www.ajpblive.com). The mean number of offi ce visits per patient did not differ significantly between the 2 groups for any of the 3 years (Appendix A). Office visits averaged approximately 1 per month for each group.
Hospitalization risk was similar for the 2 groups in each year of the study (Figure 3). Specialty patients appeared to have a lower percentage of hospitalizations compared with retail patients in years 2 and 3, but the differences were not statistically signifi cant. Risk of an ED visit for specialty patients was signifi cantly lower than that for retail patients in year 3.
Prescription drug costs were signifi cantly higher for specialty patients than for retail patients in all 3 years of the study (
). Adjusted costs for RA medications averaged $2733 higher for specialty than for retail patients in 2006, $3676 higher in 2007, and $3912 higher in 2008 (P <.0001 for each comparison). Adjusted costs for all medications averaged $3107 higher for specialty than for retail patients in 2006, $4049 higher in 2007, and $4059 higher in 2008 (P <.0001 for each comparison). Rheumatoid arthritis medications accounted for the majority of prescription drug costs (approximately 70%-75%) for each group.
Average medical costs (excluding prescription drug costs) were signifi cantly lower for specialty pharmacy patients than for retail patients in all 3 years of the study (Figure 4). Adjusted medical costs for specialty patients averaged $1099 lower than those for retail patients in 2006 (P <.01), $2137 lower in 2007 (P <.0001), and $1324 lower in 2008 (P = .01).
The overall cost difference between specialty and retail pharmacy can be estimated by combining the cost differences for medical costs and total prescription drug costs (Figure 4). The total adjusted healthcare costs for specialty patients appear to be higher than those for retail patients in each year of the study, because the savings in medical costs are more than outweighed by the increased medication costs associated with higher adherence. This is a rough measure of cost offsets, because medical cost differences are based on billed cost, while drug cost differences are based on listed wholesale price.
In this analysis of RA patients who use self-injectable biologics, we found that specialty pharmacy management was associated with significantly higher levels of medication adherence compared with that of patients who only filled prescriptions at retail pharmacies. These differences were large and consistent across all 3 years of the study. Specialty pharmacy patients also showed higher rates of treatment persistency; 80% of patients in the specialty cohort refi lled a target prescription during the third year of the study, compared with only 67% of the retail patients. Our results are consistent with an extensive research literature that reports significant impacts of pharmacy-based initiatives on medication adherence and therapeutic outcomes.6-8
Specialty pharmacy management was associated with some reductions in medical resource utilization, but the results by service type were not consistent. The likelihoodof an office visit was significantly lower for specialty patients compared with retail patients in the second and third years of the study, but the average number of visits per patient was not signifi cantly different between the groups. Hospitalization rates in the second and third years of the study suggested a trend toward a lower hospitalization risk for specialty patients, but the differences were not signifi cant. Rates of ED use were significantly lower for specialty patients in the third year of the study. Overall, the linkage between specialty pharmacy management and medical utilization was not as strong and consistent as the linkage to medication adherence.
Specialty pharmacy management had a strong and consistent impact on medical costs. Excluding prescription drug costs, medical costs were significantly lower for specialty patients than for retail patients in all 3 years of the study. Although the differences in medical utilization were variable by service type, the composite impact of these differences was an overall reduction in the costs of medical services for patients who filled their prescriptions through the specialty pharmacy.
Not surprisingly, drug costs for RA medications were significantly higher for specialty patients than for retail patients in all 3 years of the study—a direct reflection of the higher medication adherence for the specialty patients. Total drug costs were also higher for specialty patients in proportion to the increased costs of their RA medications.
In this analysis, we did not observe an overall healthcare cost reduction associated with improved medication adherence. The increased cost of medications appeared to outweigh the reductions in costs for medical services. For a condition like RA, for which high-cost, brand-name biologics are becoming the standard of care, a net cost offset is less likely to be found than for conditions that are treated with lower-cost (often generic) medications.15-18 The primary cost offset for RA patients may be reductions in indirect healthcare costs such as the costs associated with productivity loss, disability leave, and absenteeism. By some estimates, the indirect costs associated with RA can exceed the direct medical costs.19,20 Improved medication adherence could have a signifi cant impact on these indirect costs by improving disease control and reducing physical disability.21 Estimating these impacts is beyond the scope of this study, but it is a fruitful line of inquiry for future research. Rheumatoid arthritis is a progressive condition, so an analysis of cost offsets by disease stage and over longer observation periods would be informative. A more comprehensive cost-benefit analysis of specialty pharmacy will also need to evaluate the intervention costs of the value-added services provided, as well as the quality-of-life benefi ts associated with improved disease control.
Conducting an analysis of this kind in a real-world setting has both benefits and limitations. In contrast to a clinical trial, this study looks at outcomes in an unconstrained “field” setting, where patients receive usual care from a mail-order specialty pharmacy and from a wide range of community-based retail pharmacies. This design gives the results broad external validity when evaluating the relative impact of these 2 pharmacy types. However, it limits our ability to identify which features of specialty pharmacy are most influential in improving adherence and reducing medical costs, because the features cannot be varied independently. Also, in this setting, the level of exposure to specialty pharmacy services varied across specialty patients during the 3-year study period. Although this reflects the realities of patients’ refi ll behavior, it increases the heterogeneity of the group and reduces precision in identifying the impacts of specialty pharmacy. In spite of this variability, the observed impacts of specialty pharmacy on medication adherence were remarkably large and consistent. As in any claims-based analysis of this type, it is possible that unmeasured attributes of the 2 study groups were confounded with the observed differences.
In this study, the specialty pharmacy offered several therapy management services. It provided in-depth education for patients at the onset of therapy, it consulted with patients on a scheduled basis to discuss barriers to medication adherence, it provided 24-hour access to pharmacists and nurses with specialized knowledge of RA treatment, and it expedited refi ll and delivery of the patient’s medications. These pharmacy-based services were associated with higher sustained rates of adherence to RA treatment, reductions in the use of medical resources, and lower costs for medical services. In an era of increasing pressure to improve quality and reduce costs, specialty pharmacies offer an effective solution for managing the costs and benefi ts of biologic treatments.