Publication
Article
Author(s):
Higher cost sharing is associated with reduced branded antidepressant initiation among patients trying generic therapy. Dynamic benefit designs could enhance access to branded medications when appropriate.
ABSTRACT
Objectives: To determine the relationship between consumer cost sharing for branded antidepressants and the initiation of branded therapy among patients with major depressive disorder (MDD) filling a prescription for generic MDD medication.
Study Design: Retrospective cross-sectional analyses.
Methods: Patients aged 18 to 64 years with MDD who filled a generic antidepressant were identified in commercial claims data for 2012 to 2014. For each year-specific analysis, an average cost-sharing index for branded antidepressants at the level of the plan was computed. Multivariable models were used to estimate the relationship between plan-level cost sharing for branded antidepressant medications and the filling of branded prescriptions, with demographic and clinical variables as covariates.
Results: For patients with MDD filling a generic prescription, increases in branded cost sharing were associated with significant decreases in the likelihood of filling a branded antidepressant in each year (P <.001). Results in 2012 imply that a shift from the 0th to 90th percentile in the branded cost-sharing index corresponded with a 9.5% decrease in the relative likelihood of a branded fill among patients receiving a generic antidepressant. The corresponding figures for 2013 and 2014 were 9.3% and 3.5%, respectively.
Conclusions: In MDD, patients and clinicians who dutifully adhere to guidelines requiring a trial of first-line medication may ultimately require therapy with alternate agents to achieve adequate disease control. A “reward the good soldier” benefit design would lower cost sharing for higher-tier evidence-based therapies when clinically indicated. Results suggest that narrowing the gap in cost sharing between branded and generic medications following a trial of a generic agent might improve access to second-line treatment in MDD.
Am J Manag Care. 2018;24(4):180-186Takeaway Points
Cost sharing for any particular medication is generally fixed and rarely reflects patient-specific factors that might change over the course of a disease. These factors may have important implications for appropriateness.
Payers and purchasers increasingly rely on high cost sharing to reduce expenditures. The use of high-deductible health plans (HDHPs), in particular, is growing rapidly.1-3 In 2016, 39% of Americans younger than 65 years with private insurance were enrolled in an HDHP3; that same year, 57% of firms with 1000 or more workers offered an HDHP plan option.2 Apart from deductibles, high cost sharing for covered prescription medications may bring about cost-related nonadherence.4-8 Average co-payments in 2016 for third- and fourth-tier medications were $57 and $102, respectively, for employer-sponsored coverage,2 while beneficiaries in the most popular Medicare Part D prescription drug plans faced coinsurance of 30% to 50% for nonpreferred branded medications.9 Advocates and researchers have raised concerns that cost sharing at these levels may reduce receipt of needed high-value services, especially among poorer enrollees and those with chronic medical conditions.10,11
In addition to high consumer cost sharing, payers and pharmacy benefit managers employ utilization management techniques, such as step therapy and prior authorization, to limit spending. A 2016 Pharmacy Benefit Management Institute report found that 83% of employers surveyed applied step therapy requirements and another 9% were considering their use.12 More than one-third of employer plans used prior authorization and/or step therapy requirements for branded mental health medications in 2016, as did 60% of exchange-sold plans.13
These policies have serious consequences for patients and society, including for those living with major depressive disorder (MDD). MDD is the second leading medical cause of long-term disability and the fourth leading cause of global burden of disease; it is predicted to become the second highest cause of disability by 2020.14,15 MDD affects more than 15 million Americans, accounting for more than $95 billion in direct costs and $105 billion in lost workplace productivity each year.16 New approaches to reduce the burden of this illness merit consideration.
The natural history of chronic conditions often necessitates multiple therapies to achieve desired clinical outcomes. In MDD, more than two-thirds of patients fail to achieve remission with first-line treatment.17 For many of these patients, more effective treatment options may be available, but only at higher drug tiers that require greater cost sharing. Given the heterogeneity of patients with MDD and that several patient factors impact therapeutic response, successful treatment requires a personalized approach. Research has shown, for example, that a patient’s failure to respond to one selective serotonin reuptake inhibitor (SSRI) does not mean that he or she will not respond to another SSRI or to a serotonin—norepinephrine reuptake inhibitor (SNRI).18,19 Guidelines therefore generally recommend consideration of another SSRI or SNRI after unsuccessful trial of a first-line SSRI/SNRI.19,20
It is well established in the literature that greater out-of-pocket (OOP) costs for indicated chronic disease medications can reduce initiation and adherence, lower the likelihood of achieving desired health outcomes, and, for some conditions, increase acute care utilization.4-8,21-26 Under today’s more traditional formulary arrangements, patients with MDD who have a suboptimal response to a first-tier, usually generic, drug are often subject to higher cost sharing for access to therapies that may be more effective, by virtue of these therapies’ placement on higher tiers. A “reward the good soldier” approach, or “step edit with co-pay relief,” has received attention as an alternative strategy.27-30 This “dynamic” model would lower consumer cost sharing (ie, co-payment or coinsurance) for higher-tier evidence-based therapies, but only when first-line lower-cost therapies prove ineffective in achieving desired clinical outcomes. This would mean, for instance, that a patient with MDD who has not achieved adequate disease control after having tried a generic antidepressant would be subject to low cost sharing if prescribed a branded antidepressant. This would ensure that individuals with chronic disease, such as MDD, would not be penalized in the form of higher cost sharing simply because of their biology.
Objective
Given that consumer cost-sharing levels for specific medications are fixed and do not reflect the varying nature of most chronic diseases, rigorous analyses are warranted to explore the extent to which current formulary designs facilitate access to alternate options when first-line therapies prove ineffective. Accordingly, the objective of this study is to determine whether higher consumer cost sharing for branded MDD medications is associated with reduced initiation of second-line therapies for individuals diagnosed with MDD filling a prescription for a generic MDD medication. The findings have implications for the merit of piloting new benefit designs.
METHODS
Data
We used the Truven Health Analytics MarketScan Commercial Claims and Encounters Database from 2012 to 2014, the most recent complete calendar years available at the time the analyses were commenced. The database contains the inpatient medical, outpatient medical, and outpatient prescription drug experiences of approximately 200 million employees and their dependents covered under a variety of fee-for-service and managed care health plans between 1995 and 2014, including approximately 30 million covered lives in 2014.
Patient Cohort
Patients were included in each year-specific analysis if they had at least 1 claim with a diagnosis of MDD and 1 claim for a generic antidepressant during the 12-month year or the 3 months preceding the calendar year of interest. The lists of generic and branded medications used to treat MDD were obtained from First Databank. International Classification of Diseases, Ninth Edition, Clinical Modification codes were used to identify patients with MDD. Patients with less than 12 months of continuous medical and prescription coverage in any given year, younger than 18 years, older than 65 years, and/or having any claims indicative of pregnancy were excluded. Also excluded were patients in firm/plan combinations with fewer than 30 prescriptions for generic or branded antidepressant medication to ensure that the firm/plan-specific cost-sharing estimation was not unduly influenced by a relatively small number of prescriptions.
Variables
Use of branded antidepressant medications. The measure of antidepressant utilization was a binary variable, set equal to 1 if a patient filled at least 1 branded antidepressant prescription in the 12 months of the observation year. The variable was otherwise set to 0.
Cost sharing for branded antidepressant medications. For each year-specific analysis, an average co-payment/coinsurance at the level of the firm/plan was computed for 30-day adjusted branded antidepressant prescriptions based on the aggregated experience of attributed patients that year. Firm/plan-level values were assigned to each attributed patient. The use of firm/plan-specific experience rather than patient-level experience can help reduce endogeneity associated with the drug-related decision making of employees and allows for estimation of how much a branded antidepressant would have cost for patients who did not receive a branded drug in a given year. The measure of cost sharing was ranked across firm/plan combinations to calculate an index of branded antidepressant cost sharing.
Demographic and clinical variables. The analyses included the following covariates: age, sex, geographic region, urban/rural residence, insurance plan type, patient relationship to the primary subscriber, and median household income in the zip code of residence. Age was also categorized into intervals of 18 to 34, 35 to 44, 45 to 54, and 55 to 64 years. For each patient, a Charlson Comorbidity Index score31 and the number of psychiatric diagnostic groupings were characterized.32 All covariates were measured as of January 1 of the calendar year of interest.
Analysis
Standard descriptive statistics were calculated for all demographic and clinical variables for each year-specific analysis. Descriptive statistics were also calculated to describe prescribing patterns, use, and cost sharing for branded antidepressant medications among included patients for each year. Cost-sharing indices for each year were calculated as described above. Multivariable models were used to assess the relationship between cost sharing and use of branded antidepressant medications at the patient level. The main covariate of interest was the co-payment/coinsurance index for branded medications. The models were adjusted for the demographic and clinical variables described in the previous paragraph.
Because the primary end point, any use of branded antidepressant medication, was binary, logistic regression models were used to estimate the probability of success (any use). The probability of branded use in varying cost-sharing percentile ranges was assessed using margins. Adjusted standard errors accounted for clustering.
Sensitivity analyses were conducted by varying the measure of total OOP responsibility (co-payment, coinsurance, and deduc­tible) for co-payment/coinsurance only in calculating the branded cost-sharing indices.
RESULTS
Descriptive Characteristics
Characteristics of the more than half-million patients meeting the inclusion criteria for each year-specific analysis are presented in Table 1. Consistent with other MDD research, patients were predominately female33; the majority of patients lived in urban areas, and nearly two-thirds had employer-sponsored preferred provider organization coverage.
Table 2 describes the cost-sharing index for patients meeting criteria for each year-specific analysis. eAppendix Tables 1-3 (eAppendices available at ajmc.com) show the variation in demographic characteristics across patients associated with firms at different cost-sharing index quartiles for the 2012, 2013, and 2014 cost-sharing indices, respectively.
Impact of Cost Sharing for Branded Antidepressants
Increases in branded cost sharing were associated with significant decreases in the predicted likelihood of filling a branded antidepressant in each of the year-specific analyses (P <.001) (Figure). For example, 2012 results imply that a shift from the 0th to 90th percentile in the branded cost-sharing index ($3.96 to $47.00) corresponded with a 9.5% decrease in the relative likelihood of a branded fill among patients who had filled a prescription for a generic antidepressant that year (or in the last 3 months of 2011). The corresponding figures for 2013 and 2014 were 9.3% and 3.5%, respectively. The shift from the 0th to the 90th percentile implied a shift in cost sharing from $7.10 to $43.65 in 2013 and from $5.42 to $36.25 in 2014.
In the sensitivity analyses using OOP expenditures, rather than co-payment/coinsurance only, results remained statistically significant (P <.001) in 2012 and 2013; however, the simulated probabilities were reduced in magnitude compared with the co-payment/coinsurance model. Results using the OOP cost-sharing index were not statistically significant in 2014.
DISCUSSION
High OOP costs can inhibit access to medications necessary to achieve satisfactory patient-centered outcomes. For commercially insured patients using generic antidepressants, the odds of filling a branded drug decreased significantly as cost sharing increased. The association was significant but somewhat attenuated over time. This decrease across years may be due to the greater availability of patient assistance programs in recent years34,35 or to the narrowing of the gap in absolute dollars between the high and low ends of the branded cost-sharing index (Table 2).
The results of this study are consistent with published research findings showing that cost-sharing differentials among formulary tiers between generic and branded medications discourage the use of branded agents.22-26 Previous studies have examined prescribing and utilization patterns across entire populations, but this is the first to explicitly assess a patient population using a generic agent in the same therapeutic class. More generally, these findings are consistent with the experience of payers and purchasers who have implemented value-based insurance design (VBID) (ie, benefit designs that seek to align patients’ OOP cost sharing with the underlying value of the service or medication).36 A 2012 systematic review of VBID implementations found modest but meaningful improvements in adherence to chronic disease management medications associated with reductions in cost sharing, including, but not limited to, medications for MDD37; subsequent work has reported similar results.38,39 The findings of the present study—that lower cost sharing is an enabler of access to therapy for chronic disease—are consistent with the broader evidence on the impact of VBID. Approaches, such as VBID, that reduce barriers to care and improve patient-centered outcomes merit consideration, especially at a time when significant changes to the US healthcare system are being contemplated.
Even for conditions commonly managed with generic medications, patients and clinicians who dutifully adhere to guidelines requiring a trial of first-line medication (or medications) may require therapy with alternate agents to achieve adequate disease control. In MDD, for example, Rush et al reported that only 37% of patients responded adequately to a first-line antidepressant, with many patients deriving clinical benefit from second- and third-line therapies.17 A “reward the good soldier” approach commits to established policies that encourage the use of lower-cost first-line therapies and would lower consumer cost sharing for second-line medications only when clinically indicated (eg, contraindication, disease progression, unsatisfactory response). Dynamic VBID benefit designs of this kind might also provide for lower cost sharing for targeted therapies (eg, based on specific biomarkers).29 Unlike currently implemented step-edit programs that do not account for patient- or condition-specific variations in safety, efficacy, and tolerability within drug classes, the “reward the good soldier” VBID approach would enhance access to effective therapies when clinically appropriate.
Instances of selective cost-sharing relief already exist. Medicare Part D plans are required to maintain cost-sharing tiering exception processes for beneficiaries seeking coverage in a higher tier, given contraindications, adverse effects that have occurred with more preferred agents, or other pertinent patient-specific circumstances.40 Drugs on the plan’s specialty tier are excluded from these processes, and exceptions are available only on a case-by-case basis.40 Similar exception processes appear to be uncommon outside Part D.
Limitations
This study had several limitations. Analyses did not account for the timing of branded treatments or reasons for medication changes. Patients were counted as using a branded medication if they had a branded prescription at any point in the year, even if the claim for the branded medication preceded the claim for the generic. However, as shown in Table 1, very few patients received a branded medication prior to a generic (between 3% and 9%, depending on the year). This means the results reflect 2 scenarios. The first scenario, which is true 91% to 97% of the time, is that patients on generics are less likely to switch to branded drugs if branded cost sharing is higher. The second scenario, which is true only 3% to 9% of the time, is that patients on branded antidepressants are more likely to switch to generic antidepressants if branded cost sharing is higher. The existence of patients in this second scenario presents a low risk of bias because these patients are rare and will only cause bias if their sensitivity to branded cost sharing is very different than that of patients who start on generic medications. Relatedly, the analyses could not account for the reasons for medication switches, given the limitations inherent in claims data. This would only bias the main findings if the reasons for switching medications tend to vary across cost-sharing quartiles, which appears unlikely.
Another limitation pertains to patients’ choice of health plan. It is possible that individuals preferring branded medications differentially selected more generous plans with lower cost sharing for branded medications. Although some firms within the MarketScan data offer identical formularies across plan offerings, thereby mitigating the risk of this type of endogeneity, others do not. The regression models controlled for the number of patient comorbidities, which would reduce the impact of differential plan selection if patients with greater illness burdens regularly chose more generous plans when available.
In addition, this study was unable to ascertain the effect of patient assistance programs, such as charitable grants, co-payment cards, and coupons, which reduce the likelihood of cost-related nonadherence by assuming a substantial share of patients’ cost-sharing liability.34,35 However, the observed year-over-year decreases in cost sharing (Table 2) suggest that these programs may be working as intended (ie, helping patients access medications that they otherwise could not). The inability to measure the use of assistance programs likely biased findings toward underestimating the impact of higher cost sharing on antidepressant utilization in MDD.
These analyses were unable to measure benefit design or cost-sharing requirements directly. Instead, cost sharing was inferred based on co-payment/coinsurance spending per prescription. Changes in progress against the plan deductible and OOP maximum limits may also have complicated measurement. Nevertheless, a substantial body of work has relied on similar approaches.41-43 In addition, relatively few patients in the study sample were enrolled in HDHPs compared with national averages,2 potentially limiting generalizability. Finally, this study employed cross-sectional analyses. Future work should consider use of cohort-based designs, which might reduce the risk of bias due to confounding from unobserved variables. Such designs might take as their study population those patients discontinuing generic medication and proceed to follow their experience over multiple years.
CONCLUSIONS
Higher levels of cost sharing for branded antidepressants discourage patients using generic antidepressants from trying alternative therapy options. This is particularly germane because the best therapy for one patient may be suboptimal for another patient; prescribing decisions must be tailored to the patient’s individual circumstances. Failure to optimize therapy places patients with MDD at higher risk for relapse or recurrence, greater impairment, and, potentially, additional acute care utilization. By reducing cost-related barriers, dynamic benefit designs could enhance access to effective therapies that increase the likelihood of achieving the patient-centered outcomes that matter most to those with MDD: better functioning at home and at work, improved quality of life, and higher levels of overall well-being.
Acknowledgments
Dr Mucha, who was employed with Takeda Pharmaceuticals USA, Inc, when this work was completed, is now employed with EMD Serono (Billerica, MA). Author Affiliations: Center for Value-Based Insurance Design, Departments of Internal Medicine and Health Management and Policy, University of Michigan (JDB, AMF), Ann Arbor, MI; Harvard Medical School (MEC), Boston, MA; IBM Watson Health (MB), Cambridge, MA; IBM Watson Health (AV), St. Louis, MO; Takeda Pharmaceuticals USA, Inc (DW, LM), Deerfield, IL.
Source of Funding: Financial support for this research was provided by Takeda Pharmaceuticals USA, Inc, and Lundbeck.
Author Disclosures: Mr Buxbaum is a research associate with the Center for Value-Based Insurance Design and a consultant with VBID Health, LLC, and has received honoraria for participating in a Merck Expert Input Forum. Dr Chernew is a co-editor-in-chief of The American Journal of Managed Care® (AJMC®); due to length, his other disclosures are available in the eAppendix. Dr Bonafede is an employee of Truven Health LLC, an IBM company, which received a research contract to conduct this analysis. Ms Vlahiotis is an employee of Truven Health Analytics. Ms Walter is an employee and stock owner of Takeda. Dr Mucha was an employee and stock owner of Takeda at the time the work was completed. Dr Fendrick is a co-editor-in-chief of AJMC®; due to length, his other disclosures are available in the eAppendix.
Authorship Information: Concept and design (MEC, DW, LM, AMF); acquisition of data (MB, AV, LM); analysis and interpretation of data (MEC, JDB, MB, AV, DW, LM, AMF); drafting of the manuscript (JDB, AV, DW, LM, AMF); critical revision of the manuscript for important intellectual content (JDB, MEC, MB, AV, DW, LM, AMF); statistical analysis (MEC, MB, AV); provision of patients or study materials (MB, AV); obtaining funding (DW); administrative, technical, or logistic support (JDB, MEC, AMF); and supervision (MEC, AMF).
Address Correspondence to: Jason D. Buxbaum, MHSA, Center for Value-Based Insurance Design, University of Michigan, North Campus Research Complex, 2800 Plymouth Rd, Ann Arbor, MI 48109. Email: buxbaum@umich.edu.REFERENCES
1. Reynolds C, Rudder L, Westrich K. Consumer-directed health plans: pharmacy benefits & “better practices.” National Pharmaceutical Council website. npcnow.org/publication/consumer-directed-health-plans-pharmacy-benefits-better-practices. Published September 30, 2014. Accessed March 14, 2017.
2. Claxton G, Rae M, Long M, et al. Employer Health Benefits 2016. Menlo Park, CA, and Chicago, IL: Kaiser Family Foundation, Health Research & Educational Trust, NORC at the University of Chicago; 2016. files.kff.org/attachment/Report-Employer-Health-Benefits-2016-Annual-Survey. Accessed March 14, 2017.
3. Martinez ME, Zammitti EP, Cohen RA. Health insurance coverage: early release of estimates from the National Health Interview Survey, January-September 2016. CDC National Center for Health Statistics website. cdc.gov/nchs/data/nhis/earlyrelease/insur201702.pdf. Published February 2017. Accessed March 14, 2017.
4. Goldman DP, Joyce GF, Zheng Y. Prescription drug cost sharing: associations with medication and medical utilization and spending and health. JAMA. 2007;298(1):61-69. doi: 10.1001/jama.298.1.61.
5. Iuga AO, McGuire MJ. Adherence and health care costs. Risk Manag Healthc Policy. 2014;7:35-44. doi: 10.2147/RMHP.S19801.
6. Gibson TB, Ozminkowski RJ, Goetzel RZ. The effects of prescription drug cost sharing: a review of the evidence. Am J Manag Care. 2005;11(11):730-740.
7. Goldman DP, Joyce GF, Escarce JJ, et al. Pharmacy benefits and the use of drugs by the chronically ill. JAMA. 2004;291(19):2344-2350. doi: 10.1001/jama.291.19.2344.
8. Luiza VL, Chaves LA, Silva RM, et al. Pharmaceutical policies: effects of cap and co-payment on rational use of medicines. Cochrane Database Syst Rev. 2015;(5):CD007017. doi: 10.1002/14651858.CD007017.pub2.
9. Hoadley J, Cubanski J, Neuman T. Medicare Part D: a first look at prescription drug plans in 2017. Kaiser Family Foundation website. kff.org/report-section/medicare-part-d-a-first-look-at-prescription-drug-plans-in-2017-findings. Published October 17, 2016. Accessed January 26, 2017.
10. Wharam JF, Zhang F, Eggleston EM, Lu CY, Soumerai S, Ross-Degnan D. Diabetes outpatient care and acute complications before and after high-deductible insurance enrollment: a Natural Experiment for Translation in Diabetes (NEXT-D) Study. JAMA Intern Med. 2017;177(3):358-368. doi: 10.1001/jamainternmed.2016.8411.
11. Rabin DL, Jetty A, Petterson S, Saqr Z, Froehlich A. Among low-income respondents with diabetes, high-deductible versus no-deductible insurance sharply reduces medical service use. Diabetes Care. 2017;40(2):239-245. doi: 10.2337/dc16-1579.
12. 2016 trends in drug benefit design report. Pharmacy Benefit Management Institute website. pbmi.com/ItemDetail?iProductCode=BDR_2016&Category=BDR. Published December 12, 2016. Accessed March 14, 2017.
13. Avalere Health. PlanScape review of patient access to medicines in exchange plans, by condition. PhRMA website. phrma-docs.phrma.org/sites/default/files/pdf/avalere-planscape-2016-by-condition.pdf. Published April 2016. Accessed March 14, 2017.
14. Santaguida P, MacQueen G, Keshavarz H, Levine M, Beyene J, Raina P. Treatment for depression after unsatisfactory response to SSRIs [comparative effectiveness review #62]. Agency for Healthcare Research and Quality website. effectivehealthcare.ahrq.gov/sites/default/files/pdf/depression-treatment-ssri_research.pdf. Published April 26, 2012. Accessed March 29, 2017.
15. Mental Health: A Call for Action by World Health Ministers. Geneva, Switzerland: World Health Organization; 2001. who.int/entity/mental_health/media/en/249.pdf. Accessed March 28, 2017.
16. Greenberg PE, Fournier A-A, Sisitsky T, Pike CT, Kessler RC. The economic burden of adults with major depressive disorder in the United States (2005 and 2010). J Clin Psychiatry. 2015;76(2):155-162. doi: 10.4088/JCP.14m09298.
17. Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry. 2006;163(11):1905-1917. doi: 10.1176/ajp.2006.163.11.1905.
18. Huskamp HA. Managing psychotropic drug costs: will formularies work? Health Aff (Millwood). 2003;22(5):84-96. doi: 10.1377/hlthaff.22.5.84.
19. Qaseem A, Snow V, Denberg TD, Forciea MA, Owens DK; Clinical Efficacy Assessment Subcommittee of American College of Physicians. Using second-generation antidepressants to treat depressive disorders: a clinical practice guideline from the American College of Physicians [erratum inAnn Intern Med. 2009;150(2):148]. Ann Intern Med. 2008;149(10):725-733. doi: 10.7326/0003-4819-149-10-200811180-00007.
20. Gelenberg AJ, Freeman MP, Markowitz JC, et al. Practice guideline for the treatment of patients with major depressive disorder, third edition. Am J Psychiatry. 2010;167(10):1-3,9-11,13-118. psychiatryonline.org/pb/assets/raw/sitewide/practice_guidelines/guidelines/mdd.pdf. Accessed March 28, 2017.
21. Seabury SA, Lakdawalla DN, Walter D, et al. Patient outcomes and cost effects of Medicaid formulary restrictions on antidepressants. Forum Health Econ Policy. 2014;17(2):153-168. degruyter.com/view/j/fhep.2014.17.issue-2/fhep-2014-0016/fhep-2014-0016.xml. Accessed March 14, 2017.
22. Gibson TB, Fendrick AM, Gatwood J, Chernew ME. Gaps in treatment, treatment resumption, and cost sharing. Am J Pharm Benefits. 2012;4(6):e159-e165. ajmc.com/journals/issue/2012/ajpb_novdec2012/gaps-in-treatment-treatment-resumption-and-cost-sharing. Accessed January 27, 2017.
23. Shrank WH, Hoang T, Ettner SL, et al. The implications of choice: prescribing generic or preferred pharmaceuticals improves medication adherence for chronic conditions. Arch Intern Med. 2006;166(3):332-337. doi: 10.1001/archinte.166.3.332.
24. Briesacher BA, Andrade SE, Fouayzi H, Chan KA. Medication adherence and use of generic drug therapies. Am J Manag Care. 2009;15(7):450-456.
25. Gagne JJ, Choudhry NK, Kesselheim AS, et al. Comparative effectiveness of generic and brand-name statins on patient outcomes: a cohort study. Ann Intern Med. 2014;161(6):400-407. doi: 10.7326/M13-2942.
26. Chen S-Y, Shah SN, Lee Y-C, Boulanger L, Mardekian J, Kuznik A. Moving branded statins to lowest copay tier improves patient adherence. J Manag Care Pharm. 2014;20(1):34-42. doi: 10.18553/jmcp.2014.20.1.34.
27. Chernew M, Gibson TB, Fendrick AM. Trends in patient cost sharing for clinical services used as quality indicators. J Gen Intern Med. 2010;25(3):243-248. doi: 10.1007/s11606-009-1219-y.
28. Fendrick AM; V-BID Center Staff. Reward the good soldier: a dynamic approach to consumer cost-sharing for prescription drugs. The American Journal of Managed Care website. ajmc.com/contributor/vbid-center/​2016/09/reward-the-good-soldier-a-dynamic-approach-to-consumer-cost-sharing-for-prescription-drugs. Published September 26, 2016. Accessed March 8, 2017.
29. Fendrick AM, Buxbaum J, Westrich K; National Pharmaceutical Council; University of Michigan School of Public Health Center for Value-Based Insurance Design. Supporting consumer access to specialty medications through value-based insurance design. Center for Value-Based Insurance Design website. vbidcenter.org/wp-content/uploads/​2014/10/vbid-specialty-medications-npc2014-final-web.pdf. Published October 2014. Accessed March 7, 2017.
30. A “dynamic” approach to consumer cost-sharing for prescription drugs. Center for Value-Based Insurance Design website. vbidcenter.org/a-dynamic-approach-to-consumer-cost-sharing-for-prescription-drugs. Published September 7, 2016. Accessed January 27, 2017.
31. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CMadministrative databases. J Clin Epidemiol. 1992;45(6):613-619. doi: 10.1016/0895-4356(92)90133-8.
32. Ashcraft ML, Fries BE, Nerenz DR, et al. A psychiatric patient classification system: an alternative to diagnosis-related groups. Med Care. 1989;27(5):543-557.
33. Citrome L. Vortioxetine for major depressive disorder: an indirect comparison with duloxetine, escitalopram, levomilnacipran, sertraline, venlafaxine, and vilazodone, using number needed to treat, number needed to harm, and likelihood to be helped or harmed. J Affect Disord. 2016;196:225-233. doi: 10.1016/j.jad.2016.02.042.
34. Starner CI, Alexander GC, Bowen K, Qiu Y, Wickersham PJ, Gleason PP. Specialty drug coupons lower out-of-pocket costs and may improve adherence at the risk of increasing premiums. Health Aff (Millwood). 2014;33(10):1761-1769. doi: 10.1377/hlthaff.2014.0497.
35. Daubresse M, Andersen M, Riggs KR, Alexander GC. Effect of prescription drug coupons on statin utilization and expenditures: a retrospective cohort study. Pharmacotherapy. 2017;37(1):12-24. doi: 10.1002/phar.1802.
36. Chernew ME, Rosen AB, Fendrick AM. Value-based insurance design. Health Aff (Millwood). 2007;26(2):w195-w203. doi: 10.1377/hlthaff.26.2.w195.
37. Lee JL, Maciejewski ML, Raju SS, Shrank WH, Choudhry NK. Value-based insurance design: quality improvement but no cost savings. Health Aff (Millwood). 2013;32(7):1251-1257. doi: 10.1377/hlthaff.2012.0902.
38. Yeung K, Basu A, Hansen RN, Watkins JB, Sullivan SD. Impact of a value-based formulary on medication utilization, health services utilization, and expenditures. Med Care. 2017;55(2):191-198. doi: 10.1097/MLR.0000000000000630.
39. Hirth RA, Cliff EQ, Gibson TB, McKellar MR, Fendrick AM. Connecticut’s value-based insurance plan increased the use of targeted services and medication adherence. Health Aff (Millwood). 2016;35(4):637-646. doi: 10.1377/hlthaff.2015.1371.
40. Prescription drug benefit manual: chapter 18 - Part D enrollee grievances, coverage determinations, and appeals. CMS website. cms.gov/Medicare/Appeals-and-Grievances/MedPrescriptDrugApplGriev/Downloads/Chapter18.zip. Published 2006. Updated May 12, 2014. Accessed March 7, 2017.
41. Gatwood J, Gibson TB, Chernew ME, Farr AM, Vogtmann E, Fendrick AM. Price elasticity and medication use: cost sharing across multiple clinical conditions. J Manag Care Spec Pharm. 2014;20(11):1102-1107. doi: 10.18553/jmcp.2014.20.11.1102.
42. Chernew M, Gibson TB, Yu-Isenberg K, Sokol MC, Rosen AB, Fendrick AM. Effects of increased patient cost sharing on socioeconomic disparities in health care. J Gen Intern Med. 2008;23(8):1131-1136. doi: 10.1007/s11606-008-0614-0.
43. Chernew M, Gibson TB. Cost sharing and HEDIS performance. Med Care Res Rev. 2008;65(6):713-728. doi: 10.1177/1077558708319683.