Pharmaceutical intervention is the primary means of treatment for rheumatoid arthritis (RA), a chronic autoimmune disease characterized by inflammation of the synovial joints.1 Medications used for RA treatment generally function to reduce inflammation and immune-system activation. For many years, diseasemodifying antirheumatic drugs (DMARDs) have been the mainstay of RA treatment, reducing inflammation and preventing further joint deterioration and disease progression. In the past 15 years, biologic DMARDs have become available and have proved effective for reducing inflammation and managing the disease. Biologics have changed the treatment of patients with RA, providing an alternative course of treatment when nonbiologic DMARDs alone are no longer effective or in cases of especially severe disease. The American College of Rheumatology recommends intervention with biologic DMARDs for intermediate and long duration for patients with moderate disease activity, poor prognosis, and inadequate disease response to nonbiologic DMARDs (methotrexate), and for patients with high disease activity regardless of their prognosis.2
Because of the chronic nature of RA, maintenance therapy regimens are most effective when used continuously, particularly because disease flares, loss of remission status, increased disease activity, and disability are more common with medication nonadherence and long-term discontinuation.3,4 Therefore, medication adherence is a crucial component of RA management; however, a recent review showed inadequate adherence to DMARDs among patients with RA.5 In addition to increases in disease severity and decreases in patient quality of life, nonadherence can affect the total costs of care, because patients with more severe RA incur greater costs.6,7 Because adherence affects many aspects of RA care, the rates of adherence associated with different medications are of interest to various sectors of the medical community.
Medication adherence varies widely among patients with RA; review articles on the topic have cited adherence rates ranging from 30% to 107%.8,9 This variability is in part the result of a combination of measurement methods and certain patient factors that affect adherence. There is currently no gold standard for measuring medication adherence. Methods used include subjective assessments (eg, patient surveys), direct measurements (eg, patient observation, metabolite measurements), and indirect means (eg, administrative claims–based assessment of pharmacy refills).5,9 Furthermore, because measuring adherence is effectively the measurement of human behavior, it is innately variable.
Patient factors that reportedly affect adherence include medication side effects, socioeconomics, complexity of the medication regimen, healthcare self-efficacy, knowledge of the medication and disease, and the patient’s perception of medication efficacy.8,9 In addition, there is an unintentional component to adherence that includes factors such as physician and appointment availability, infusion-center capacity, prescription costs, language barriers, and pharmacy access.5 As a result, it can be difficult to compare adherence data across different medications, populations, and studies.
For RA, adherence calculations are further complicated by biologic DMARDs, whose routes of administration may not translate well to standard adherence metrics such as medication possession ratio (MPR) or proportion of days covered. MPR and proportion of days covered are assessments of the supply of a drug received during a defined period of time, expressed as a ratio. This period can be either fixed (proportion of days covered) or variable (MPR); the latter is generally based on the prescription fill range in a pharmacy database. Even when used in the same sample, these 2 methods can yield different results.10
One method for obviating the difficulties of calculating adherence is to report persistence instead of adherence. Persistence and adherence, although related, are different metrics. Adherence refers to a patient’s compliance with a prescribed medication regimen, including the dose, timing, and frequency of medication administration.11 Persistence simply reflects the length of time that patients continue to use their medication regimen.11 Shorter durations of therapy may be attributable to patient nonadherence, but medication discontinuation also can result from lack of efficacy or the emergence of adverse side effects. Therefore, studies that include only persistence (effectively, time until discontinuation) are limited for assessing true medication adherence.
MPR measures adherence to the treatment regimen while the patient continues to use it, whereas proportion of days covered combines the adherence concept with a persistence component, permitting the measurement of adherence and persistence throughout a fixed period. However, neither metric clearly delineates the magnitude of the breach in adherence, which may be more important when considering potential clinical implications of underutilization related to nonadherence, and when attempting to implement programs to correct nonadherent behavior.
Some of the difficultly in calculating adherence metrics is related to the dosing of biologics. Infusible medications present a greater measurement challenge than subcutaneous biologics when using retrospective data sources. Dosing schedules of subcutaneous biologics range from once daily to once monthly. Subcutaneous biologics are often obtained by the patient at the pharmacy for self-administration. In a pharmacy claims database, a “days’ supply” field is present, allowing for a similar assessment of adherence metrics designed for typical oral medications.
However, infusible biologics generally are administered not at the pharmacy but in the outpatient setting. Therefore, the corresponding claims for these treatments will not appear in the pharmacy claims database but will appear in the medical claims database, where a “days’ supply” field is not standard. In addition, although the recommended dose and infusion interval are provided, the dose (mg/kg) and the infusion interval can both be modified to personalize the therapy for the patient. As a result, adherence to infusible biologics is often measured using the infusion interval (ie, number of infusions over time) or simply the time to discontinuation.12-15
Given the inadequate16 and inconsistent8,9 reported levels of adherence to biologics, the purpose of the present study was to develop novel measures of medication adherence and to assess their performance using an administrative claims database. Measures were designed with infusible medications in mind and focused on elucidating underutilization, specifically the time outside the drug’s explicit therapeutic levels.17-19 The genesis of these measures was a RAND Health study report, which referenced a “cumulative medication gap,” and stated, “This measure of gaps in medication rather than total availability is viewed by some as an improved measure over MPR or PDC [proportion of days covered], since it does not allow a gap in medication in one time period to be erased by later stockpiling.”17 For the current study, new measures were developed for calculating adherence during the maintenance phase of treatment, where nonadherence may be a concern.20,21 The infusible products abatacept and infliximab were selected for pilot-testing these measures.
Methods
Sample Selection
Study data were obtained from the Optum Clinformatics Data Mart database (OptumInsight Life Sciences, Inc). Medical and pharmacy claims were studied for members with a rheumatic disease diagnosis during calendar years 2005 (Q4) through 2012 (Q1).
To be eligible for this study, members (aged ≥18 years) were required to start treatment with either abatacept or infliximab between January 1, 2009, and March 31, 2011. A 1-year treatment-naïve period preceding the index date was imposed. Therefore, all members were required to have at least 2 years of continuous eligibility—1 year before the index date through 1 year after it. Only members who reached the maintenance phase of treatment, defined as presenting a minimum of 4 treatment infusions, were retained, and no gaps of ≥90 days during induction were permitted. Members also were required to have at least 2 claims for RA (714.x), 1 in each of the 6 months before and after the index date, and no diagnoses of ankylosing spondylitis (720.x) or psoriatic arthritis (696.x) in their record; these are approved indications for infliximab and many tumor necrosis factor (TNF)-alpha biologics. Members who received subcutaneously administered abatacept were excluded. Females must not have become pregnant during the study.
The final sample size was 461 for the abatacept group and 449 for the infliximab group. The process for selecting the study sample is shown in Figure 1.
Measures
Seven new measures of medication adherence were constructed for the study. Each measure was designed for calculating adherence during the maintenance phase of treatment. For abatacept and infliximab, the fourth administration constitutes the commencement of maintenance therapy. Maintenance abatacept infusions are recommended every 4 weeks,22 and maintenance infliximab infusions are recommended every 8 weeks.23 Infusion gaps were defined as any amount of time between the observed and expected interval.
The measures of adherence are defined below. Because the majority of measures were calculated based on recommended maintenance intervals, the mean maintenance intervals were calculated for both groups, which served as a reference measure.
- Cumulative time with infusion gap ≥20% beyond expected interval (CG20): summation of all infusion gap days ≥20% beyond the recommended maintenance treatment interval (56 days for infliximab, 28 days for abatacept)
- Cumulative time off treatment (CToTx): summation of all infusion gaps; that is, any gap in infusions >0 days beyond the recommended maintenance treatment interval (56 days for infliximab, 28 days for abatacept)
- Days of uninterrupted use (DoUU): length of time from index date to the first infusion gap ≥10% beyond the recommended maintenance treatment interval (56 days for infliximab, 28 days for abatacept)
- Observed versus expected refill ratio (OvERR): actual number of maintenance infusions in measurement period divided by the expected number of maintenance infusions in a measurement period (5 for infliximab, 11 for abatacept based on recommended treatment intervals)
- Repeated observations of underuse (RoUU): total number of infusion events within 365 days for which the infusion gap was ≥10% beyond the recommended maintenance treatment interval (56 days for infliximab, 28 days for abatacept)
- Variance in time between refills (ViTBI): infusion gaps were categorized as 0-7 days, 8-14 days, 15-21 days, or >21 days; a member could have gaps in various categories
- Persistence or time to discontinuation (TTD): number of days between the index date and a gap in treatment ≥90 days beyond the recommended number of days in a maintenance interval (or until the end of measurement year).
The mean maintenance interval is the arithmetic mean length of the member’s maintenance intervals (ie, quotient of the summed amount of all maintenance intervals, divided by the total number of maintenance infusions received).
The measures were calculated in days, except for RoUU (number of infusion events), ViTBI (categorical variable), and OvERR (ratio).
Patient demographics for the 2 study groups were recorded from the membership table, which is a summary file that accompanies the medical and pharmacy claims tables and provides detailed information on each member’s benefit eligibility period(s) and plan type (eg, HMO, PPO), in addition to select patient characteristics. Specific demographics included age, sex, geographic region of residence, insurance line of business, and type of benefit plan.
The Charlson Comorbidity Index, a predictive measure of mortality,24 was calculated for the 1-year preindex period as a proxy of overall health, as were the rates for other comorbidities of interest. Previous use of a biologic agent during the 1-year preindex period also was calculated.
Data Analyses
Sample characteristics were calculated and presented as descriptive statistics of frequencies, percentages, means, and standard deviations. Pearson correlations were computed for all adherence measures, including the reference measure (mean maintenance interval). Bivariate group comparisons between the abatacept and infliximab groups were conducted for all adherence measures. Demographics and other patient characteristics also were compared. To assess statistical significance, chi-squared tests of equality of proportions were used for categorical variables, and independent t-tests were used for continuous variables. All data management and analyses were conducted using SPSS v.20 (SPSS Inc, Chicago, IL). The critical alpha level was set at 0.05.
Results
Table 1 shows the demographic characteristics of the study population. Members were predominantly female (80.6%), residing mainly in mid-America (46.5%) and the southeastern (30.6%) region of the United States. The mean age was 52.0 (standard deviation [SD], 11.28) years. Nearly all patients (99.6%) had commercial insurance. The mean Charlson Comorbidity Index was 0.73 (SD, 1.15), and the percentage of patients taking each drug was similar (51.5% abatacept, 48.5% infliximab).
Table 2 displays Pearson correlations for adherence metrics with scale levels of measurement (6 of the 7; ViTBI was excluded). The mean maintenance infusion interval served as the reference measure of adherence. Of these 6 measures, 5 correlated significantly with the mean maintenance interval, including positive correlations for CG20 (r = .258), DoUU (r = .212), and TTD (r = .081), and negative correlations for OvERR (r = –.072) and RoUU (r = –.189; P <.05). CToTx was the only metric that did not correlate significantly with the mean maintenance interval. Significant correlations between the new measures of adherence were achieved for all possible comparisons except for RoUU versus CToTx (r = –.043) and RoUU versus OvERR (r = –.064; P >.05). The greatest associations among measures were found for OvERR and CToTx (r = –.859), TTD and CToTx (r = –.938), and TTD and OvERR (r = .787; P <.001 for all 3 comparisons).
Table 3 shows demographic and adherence comparisons for the abatacept and infliximab groups. Abatacept recipients were significantly more likely than infliximab recipients to be female (83.7% vs 77.4%) and were almost twice as likely to have used a biologic previously (62.9% vs 32.1%; P <.05 for both comparisons). The groups did not differ in geographic region of residence, type of insurance, or prevalence of select comorbidities (P >.05 for all), but a nonsignificant trend toward increased prevalence of osteoarthritis in the infliximab group was found (72.3% vs 66.0%; P = .055).
Members who received infliximab had significantly fewer maintenance treatments than those who received abatacept (5.0 vs 8.2; P <.001) and a greater mean maintenance infusion interval than those who received abatacept (52.7 days vs 32.6 days; P <.001), which coincides with each drug’s current labeling.22,23 There were significant differences between the study groups in all adherence measure outcomes. The infliximab group had significantly lower CG20 than the abatacept group (8.1 days vs 25.5 days, respectively), CToTx (41.2 days vs 79.6 days), and RoUU (0.42 intervals vs 1.93 intervals), and significantly greater DoUU (283.7 days vs 160.4 days) and TTD (333 days vs 312; P <.001 for all).
The percentage of infusion gaps within 7 days of recommended guidelines was significantly higher in the infliximab group than in the abatacept group (93.5% vs 88.4%), and the percentage of gaps that occurred between 15 and 21 days of recommended guidelines was significantly lower in the infliximab group (0.7% vs 1.7%; P <.01). Differences in adherence measures between infliximab and abatacept groups may be viewed graphically in Figure 2.
Discussion
Despite the increasing use of insurance claims databases to assess the relationship between medication adherence and outcomes,25 actual patient adherence to biologic agents remains unclear,8 because the inconsistent definitions of medication adherence persist. Further compounding the problem are the difficulties inherent in assessing adherence to infusible medications, which are usually not dispensed at the pharmacy but rather administered in physician offices or infusion centers. Consequently, the detailed information that typically appears in a pharmacy claims table, such as dose, quantity, and days’ supply, is not readily available for infusion treatments.
Therefore, the purpose of this study was to construct and assess various novel adherence measures designed specifically for infusible biologic agents. The measures focused on the maintenance period and the magnitude of nonadherence in situations where patients may have underutilized their medication (ie, looking for gaps in treatment beyond what is expected). Using a sample of claims with abatacept and infliximab data, a total of 7 new measures of adherence were examined, based on each drug’s labeling guidelines for maintenance infusions.
The new measures were shown to correlate significantly with the mean maintenance interval, which served as the comparator measure of adherence in the study. The maintenance interval dosage of intravenous abatacept is 500 mg, 750 mg, or 1000 mg (depending on patient weight) administered every 4 weeks.22 The maintenance dosage of infliximab is 3 mg/kg every 8 weeks, although some patients may require an increase in dosage (up to 10 mg/kg) or a decrease in the dosing interval (down to 4 weeks).23
Both groups generally adhered to their treatment. Between 88% (abatacept) and 94% (infliximab) of infusion intervals were within 7 days of the expected interval, and the persistence period for both groups averaged more than 300 days (maximum 365). Between-group comparisons of performance on adherence measures showed that the infliximab group was significantly more adherent than the abatacept group. This effect coincided with the difference in mean maintenance interval, which was 3 days less in the infliximab group than stated in the labeling guidelines (53 days vs 56 days),23 and 4 days greater in the abatacept group than stated in the labeling guidelines (32 days vs 28 days).22 As noted, drug-labeling guidelines were considered in the construct of the adherence measures, which focused on periods when patients may be clinically exposed because of underutilization of their treatment.
Of the 7 adherence measures, 5 correlated significantly with the mean maintenance interval, and many measures correlated positively with each other. Although the adherence metrics are in part related, each measure involves a slightly different facet of medication adherence. Therefore, some measures may be more appropriate for assessing a particular aspect of adherence than others. For example, CToTx provides a very stringent assessment of adherence: all days in which the patient is not covered by medication are summed and reported.
In some cases, this strict measure may indicate reduced adherence even though the patient may not be at fault. For example, using CToTx, a delay in maintenance infusion because of a scheduling issue or the short-term discontinuation of medication for clinical reasons would be considered nonadherence.
Inclusion of other adherence metrics, such as ViTBI, CG20, or RoUU with CToTx may provide insight as to whether the nonadherence is a single episode or recurrent, allowing for a more complete assessment of compliance. Moreover, biologics are not effective for all patients; 20% to 40% of patients do not respond to the first anti-TNF biologic prescribed.26 Therefore, including TTD and OvERR or DoUU metrics may provide a more comprehensive picture of medication adherence.
Although many difficulties in measuring adherence to infusible biologics using claims data have been identified, a noteworthy advantage of these medications over their subcutaneous and oral counterparts is that no assumptions are required concerning the administration of medication. In contrast, the prescription fill data in pharmacy databases only provide evidence that a medication was filled. It cannot be assumed that the patient has been taking the medication as prescribed, even if the claims data suggest that adherence is 100%. However, in a medical claims database, the appearance of the relevant billing code for an infusion (eg, J0129 for abatacept, J1745 for infliximab), coupled with the allowed amount of medication, provides evidence that a specific amount of the drug was administered. This creates less uncertainty when interpreting the results of analyses in which the benefits of medication adherence are examined in relation to costs, proxies of disease status, or other outcomes.
Studies have demonstrated an inverse relationship between dosing frequency and medication adherence among patients with chronic diseases.27,28 A recent survey of patients with RA who previously used biologic therapy and were currently receiving an infusible agent showed that patients prefer infusions with long intervals between treatments to subcutaneous injections every other week.29 Therefore, attention to patient preference and scheduling may improve adherence to biologic medications, especially for infusible medications.
As healthcare policies and insurance programs evolve, adherence is becoming more important. Medication adherence is essential to quality care. An inverse relationship between adherence and reduced total healthcare costs has been reported for a variety of diseases.30-32 Evidence suggests that nonadherence to biologic medication can lead to escalating medical costs by increasing inpatient, outpatient, and laboratory services.33 There is also evidence that adherence to biologics among patients with Crohn’s disease (for which some RA-prescribed biologics are also indicated) is associated with lower overall costs and fewer disease-specific hospitalizations.13,34,35 Measuring true lifetime cost-effectiveness of any therapy for a chronic disease is a challenging task, and longitudinal studies will be needed to estimate this relationship more accurately.
Although the cost of treating RA is significantly higher in this era of biologics,36 treatment with these agents (as opposed to DMARDs) may lead to an increase in quality-adjusted life-years37 and reductions in disease activity and joint damage.36 The return on investment associated with medication adherence is particularly important for patients with chronic diseases such as RA, for which previous disease management can affect the severity of symptoms later in life.
Moreover, as the development and use of expensive biologic agents expand, the importance of adherence to these specialty pharmaceuticals is amplified. Otherwise, the efficacy of these medications in real-world settings are unlikely to ever reflect the efficacy demonstrated in clinical trials. Adherence has been shown to improve health and may potentially offset high-cost service utilization such as hospitalization and admissions to the emergency department.31,38 In addition, there is an increased push for measuring adherence in relation to improving the quality of care.39,40
The Affordable Care Act contains provisions aimed at improving quality through the launch of qualityimprovement initiatives in the forefront of healthcare. Adherence is a major component of many quality improvement programs, and various quality-monitoring groups, including the National Quality Forum, have endorsed specific measures related to adherence to medications for cardiovascular disease, diabetes, and behavioral health disorders.41 Medicare Star Ratings, which can translate to performance-based financial bonuses for providers, also are focused on medication adherence; metrics related to adherence account for 11% of the total Star Rating.
In 2013, the Pharmacy Quality Alliance (PQA) produced an adherence measure for Biologic Medications for Rheumatoid Arthritis and Other Inflammatory Conditions, but it excluded infusion products.42 The PQA has drafted a measure for health plans on primary nonadherence (ie, nonadherence to the first prescribed dose of treatment).43 Looking to the future with accountable care organizations and value-based insurance design, it is reasonable to predict that medication adherence will become an increasingly relevant factor in these initiatives. Therefore, the development and modification of adherence metrics to fully analyze and understand patient adherence to medication and the underlying factors affecting it are relevant to payers, who will soon be at the forefront of healthcare quality improvement.
Limitations
This study has several limitations. The newly determined adherence measures should be examined in relation to patient disease severity, medical cost offset, and overall quality, so that their utility in health outcomes research may become known. Moreover, adherence to other infusible biologics should be examined, such as golimumab, rituximab, and tocilizumab; each agent presents different issues with respect to measuring adherence.
In light of the chronic nature of RA, it may be appropriate to replicate the current analysis for a follow-up period of more than 1 year. It is well-known that dose escalation may occur among patients who receive infliximab,44 which may take the form of shorter intervals between infusions. Adherence to the new maintenance schedule should be identifiable in claims data by the pattern of infusion dates. However, it is not feasible to identify patients who have been assigned a reduced maintenance interval who do not comply, because their infusion record may continue to indicate 8-week intervals, which coincides with general prescribing guidelines.
Claims data lend themselves well to real-world assessment of medication use; medication fills, service utilization, and diagnostic and procedural codes are available for a large population of patients for a number of years. However, the limited patient variables and the finite periods of claims generally available (in relation to the length of a chronic disease) can make it difficult to adjust patient cohorts for baseline differences.
Our results can be viewed as real-world adherence data, and the differences between cohorts suggest the degree of accuracy that can be measured between these 2 patient populations. Finally, medical claims data may contain administrative coding errors, and do not allow for assessment of clinical outcomes related to adherence.45 That nearly all patients in this analysis participated in the same commercial health plan may limit the generalizability of the findings.
Conclusion
The difficulties in measuring adherence to infusibles, coupled with the multitude of definitions and calculation methods of existing measures, makes the assessment of adherence an unnecessarily challenging endeavor. Publications should include explicit and detailed explanations of all assumptions used for calculating or comparing adherence rates for biologics in immunology, and for treatments reimbursed under the medical benefit specifically. The 7 measures we used are straightforward, easy to calculate from an administrative database, and address different aspects of treatment adherence. These methods of reporting adherence may have greater clinical significance than MPR or proportion of days covered and may provide more detail related to “gaps” in treatment. Overall, the patients in our study population were generally adherent to their medication.
Given the evolving nature of quality assessment in healthcare and the continuing rise in overall healthcare spending, it is important to investigate whether these new measures may be associated with improved overall healthcare quality, reduced costs, or both. Combining traditional and more clinically relevant measures may result in an improved overall assessment of adherence, particularly to infusible medications, which may better elucidate the relationship between adherence and quality.
Funding Source
This study was funded by Janssen Scientific Affairs.
Author Disclosure Statement
Mr Tkacz is an employee of Health Analytics, which does research for Janssen; Mr Ingham is an employee and shareholder of Johnson & Johnson; Dr Brady is an employee of Health Analytics; Dr Meyer is an employee of and owns stocks of Janssen Scientific Affairs; Dr Ruetsch received research support from Janssen.
Mr Tkacz is Director of Analytics, Health Analytics, Columbia, MD; Mr Ingham is Director, HECOR, Janssen Scientific Affairs, Horsham, PA; Dr Brady is Project Director, Health Analytics, Columbia, MD; Dr Meyer is Manager, HECOR, Janssen Scientific Affairs, Horsham, PA; Dr Ruetsch is President and Chief Executive Officer, Health Analytics, Columbia, MD.
References
1. Singh JA, Furst DE, Bharat A, et al. 2012 update of the 2008 American College of Rheumatology recommendations for the use of disease-modifying antirheumatic drugs and biologic agents in the treatment of rheumatoid arthritis. Arthritis Care Res (Hoboken). 2012;64:625-639.
2. Saag KG, Teng GG, Patkar NM, et al; for the American College of Rheumatology. American College of Rheumatology 2008 recommendations for the use of nonbiologic and biologic disease-modifying antirheumatic drugs in rheumatoid arthritis. Arthritis Rheum. 2008;59:762-784.
3. Contreras-Yáñez I, Ponce De León S, Cabiedes J, et al. Inadequate therapy behavior is associated to disease flares in patients with rheumatoid arthritis who have achieved remission with disease-modifying antirheumatic drugs. Am J Med Sci. 2010;340:282-290.
4. Pascual-Ramos V, Contreras-Yáñez I, Villa AR, et al. Medication persistence over 2 years of follow-up in a cohort of early rheumatoid arthritis patients: associated factors and relationship with disease activity and with disability. Arthritis Res Ther. 2009;11:R26.
5. de Achaval S, Suarez-Almazor ME. Treatment adherence to disease-modifying antirheumatic drugs in patients with rheumatoid arthritis and systemic lupus erythematosus. Int J Clin Rheumatol. 2010;5:313-326.
6. Michaud K, Messer J, Choi HK, Wolfe F. Direct medical costs and their predictors in patients with rheumatoid arthritis: a three-year study of 7,527 patients. Arthritis Rheum. 2003;48:2750-2762.
7. Yelin E, Wanke LA. An assessment of the annual and long-term direct costs of rheumatoid arthritis: the impact of poor function and functional decline. Arthritis Rheum. 1999;42:1209-1218.
8. Salt E, Frazier SK. Adherence to disease-modifying antirheumatic drugs in patients with rheumatoid arthritis: a narrative review of the literature. Orthop Nurs. 2010;29:260-275.
9. van den Bemt BJF, Zwikker HE, van den Ende CHM. Medication adherence in patients with rheumatoid arthritis: a critical appraisal of the existing literature. Expert Rev Clin Immunol. 2012;8:337-351.
10. Kozma CM, Dickson M, Phillips AL, Meletiche DM. Medication possession ratio: implications of using fixed and variable observation periods in assessing adherence with disease-modifying drugs in patients with multiple sclerosis. Patient Prefer Adherence. 2013;7:509-516.
11. Cramer JA, Roy A, Burrell A, et al. Medication compliance and persistence: terminology and definitions. Value Health. 2008;11:44-47.
12. Agarwal SK, Maier AL, Chibnik LB, et al. Pattern of infliximab utilization in rheumatoid arthritis patients at an academic medical center. Arthritis Rheum. 2005;53:872-878.
13. Kane SV, Chao J, Mulani PM. Adherence to infliximab maintenance therapy and health care utilization and costs by Crohn’s disease patients. Adv Ther. 2009;26:936-946.
14. Neovius M, Arkema EV, Olsson H, et al; for the ARTIS Study Group. Drug survival on TNF inhibitors in patients with rheumatoid arthritis comparison of adalimumab, etanercept and infliximab. Ann Rheum Dis. 2015;74:354-360.
15. Wendling D, Materne G-E, Michel F, et al. Infliximab continuation rates in patients with rheumatoid arthritis in everyday practice. Joint Bone Spine. 2005;72:309-312.
16. Koncz T, Pentek M, Brodszky V, et al. Adherence to biologic DMARD therapies in rheumatoid arthritis. Expert Opin Biol Ther. 2010;10:1367-1378.
17. Gellad WF, Grenard J, McGlynn EA. A review of barriers to medication adherence: a framework for driving policy options. 2009. www.rand.org/content/dam/rand/pubs/technical_reports/2009/RAND_TR765.pdf. Accessed February 6, 2015.
18. Cooper LA. A 41-year-old African American man with poorly controlled hypertension: review of patient and physician factors related to hypertension treatment adherence. JAMA. 2009;301:1260-1272.
19. Krueger KP, Berger BA, Felkey B. Medication adherence and persistence: a comprehensive review. Adv Ther. 2005;22:313-356.
20. Ghods AJ, Nasrollahzadeh D. Noncompliance with immunnosuppressive medications after renal transplantation. Exp Clin Transplant. 2003;1:39-47.
21. Menzies R, Rocher I, Vissandjee B. Factors associated with compliance in treatment of tuberculosis. Tuber Lung Dis. 1993;74:32-37. Erratum in: Tuber Lung Dis. 1993;74:217.
22. Orencia (abatacept) [prescribing information]. Princeton, NJ: Bristol-Myers Squibb Company; June 2015.
23. Remicade (infliximab) [prescribing information]. Horsham, PA: Janssen Biotech, Inc; October 2015.
24. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383.
25. Peterson AM, Nau DP, Cramer JA, et al. A checklist for medication compliance and persistence studies using retrospective databases. Value Health. 2007;10:3-12.
26. Rubbert-Roth A, Finckh A. Treatment options in patients with rheumatoid arthritis failing initial TNF inhibitor therapy: a critical review. Arthritis Res Ther. 2009;11(suppl 1):S1.
27. Coleman CI, Limone B, Sobieraj DM, et al. Dosing frequency and medication adherence in chronic disease. J Manag Care Pharm. 2012;18:527-539.
28. Saini SD, Schoenfeld P, Kaulback K, Dubinsky MC. Effect of medication dosing frequency on adherence in chronic diseases. Am J Manag Care. 2009;15:e22-e33.
29. Willeke P, Becker H, Wassenberg S, et al. Patient/rheumatologist evaluation of infusion treatment for rheumatoid arthritis [in German]. Z Rheumatol. 2011;70:232-234, 236-238.
30. Buysman EK, Liu F, Hammer M, Langer J. Impact of medication adherence and persistence on clinical and economic outcomes in patients with type 2 diabetes treated with liraglutide: a retrospective cohort study. Adv Ther. 2015;32:341-355.
31. Roebuck MC, Liberman JN, Gemmill-Toyama M, Brennan TA. Medication adherence leads to lower health care use and costs despite increased drug spending. Health Aff (Millwood). 2011;30:91-99.
32. Sokol MC, McGuigan KA, Verbrugge RR, Epstein RS. Impact of medication adherence on hospitalization risk and healthcare cost. Med Care. 2005;43:521-530.
33. De Vera MA, Mailman J, Galo JS. Economics of non-adherence to biologic therapies in rheumatoid arthritis. Curr Rheumatol Rep. 2014;16:460.
34. Carter CT, Waters HC, Smith DB. Impact of infliximab adherence on Crohn’s disease-related healthcare utilization and inpatient costs. Adv Ther. 2011;28:671-683.
35. Carter CT, Waters HC, Smith DB. Effect of a continuous measure of adherence with infliximab maintenance treatment on inpatient outcomes in Crohn’s disease. Patient Prefer Adherence. 2012;6:417-426.
36. Furneri G, Mantovani LG, Belisari A, et al. Systematic literature review on economic implications and pharmacoeconomic issues of rheumatoid arthritis. Clin Exp Rheumatol. 2012;30(suppl 73):S72-S84.
37. Doan QV, Chiou C-F, Dubois RW. Review of eight pharmacoeconomic studies of the value of biologic DMARDs (adalimumab, etanercept, and infliximab) in the management of rheumatoid arthritis. J Manag Care Pharm. 2006;12:555-569.
38. DiMatteo MR, Giordani PJ, Lepper HS, Croghan TW. Patient adherence and medical treatment outcomes: a meta-analysis. Med Care. 2002;40:794-811.
39. Centers for Medicare & Medicaid Services. Quality measures. April 17, 2015. www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment-Instruments/QualityMeasures/index.html?redirect=/QualityMeasures/03_ElectronicSpecifications.asp. Accessed April 24, 2015.
40. Pharmacy Quality Alliance. PQA medication quality measures in the health insurance marketplace: Quality Rating System (QRS) Beta Test measure set. http://pqaalliance.org/measures/qrs.asp. Accessed April 24, 2015.
41. National Committee for Quality Assurance. HEDIS 2016 summary table of measures, product lines and changes. Volume 2. www.ncqa.org/Portals/0/HEDISQM/HEDIS2016/HEDIS%202016%20List%20of%20Measures.pdf. Accessed November 20, 2015.
42. Pharmacy Quality Alliance. Measure concept inventory. 2015. www.pqaalliance.org/images/uploads/files/Measure%20Concept%20Inventory%202015Q2.pdf. Accessed November 9, 2015.
43. Pharmacy Quality Alliance. Measure concepts under discussion: PQA measures. http://pqaalliance.org/measures/discussion.asp. Accessed May 26, 2015.
44. Ariza-Ariza R, Navarro-Sarabia F, Hernández-Cruz B, et al. Dose escalation of the anti-TNF-alpha agents in patients with rheumatoid arthritis. A systematic review. Rheumatology (Oxford). 2007;46:529-532.
45. Tyree PT, Lind BK, Lafferty WE. Challenges of using medical insurance claims data for utilization analysis. Am J Med Qual. 2006;21:269-275.