1031099 research-article20212021 TAN0010.1177/17562864211031099Therapeutic Advances in Neurological DisordersG Zanga, E Drzewiscki Therapeutic Advances in Neurological Disorders Original Research Predictors of adherence and persistence Ther Adv Neurol Disord 2021, Vol. 14: 1–11 to disease-modifying therapies in DOI: 10.1177/ https://doi.org/10.1177/17562864211031099 https://doi.org/10.1177/17562864211031099 17562864211031099 Multiple Sclerosis © The Author(s), 2021. Article reuse guidelines: sagepub.com/journals- permissions Gisela Zanga , Estefania Drzewiscki, Paula Tagliani, Maximiliano Smietniansky, Maria M. Esnaola y Rojas, and Diego Caruso Abstract Background and Aims: In multiple sclerosis (MS), non-adherence/non-persistence is related to suboptimal response to treatment, including disease relapses and the need for more expensive healthcare. The aim of this study was to identify predictors related to adherence to disease modifying therapies (DMTs) in a cohort of Argentinian MS patients. Methods: We conducted a cross-sectional study at the National Medical Care Program from Argentina. MS patients with at least one claim for a DMT from 1 January 2017 to 1 October 2017 were identified. A telephone survey was performed to assess clinical and demographic factors. The medication possession ratio (MPR) was used to estimate adherence; MPR <80% defined non-adherence. Associations were studied using a logistic regression model. Results: Our database included 648 MS patients. A total of 360 patients (60% females, mean age 55.3 years) accepted to participate. Of these, 308 (85.5%) patients were receiving DMT at the time of the survey. Some 198 (63.7%) were receiving injectable therapies. Optimal adherence was 47.7%. Adherence was associated with oral medication [odds ratio (OR) 1.83 Correspondence to: 95% confidence interval (CI) 1.13–3.00, p = 0.014]. A factor related to oral drugs was higher Gisela Zanga Department of Neurology, educational level (OR 2.86 95%CI 1.41–5.81, p = 0.004). César Milstein Hospital, Conclusion: This real-world study showed better adherence and persistence on treatment associated with University of Buenos Aires, La with oral therapies in MS patients in Argentina. Rioja 951, Buenos Aires, C1221ACI, Argentina dra.gzanga@gmail.com Keywords: multiple sclerosis, adherence, persistence, predictors, disease-modifying Estefania Drzewiscki Maria M. Esnaola y Rojas therapies Department of Neurology, César Milstein Hospital, associated with University Received: 26 November 2020; revised manuscript accepted: 21 June 2021. of Buenos Aires, Buenos Aires, Argentina Diego Caruso Department of Internal Medicine, César Milstein Introduction and improving quality of life in patients with MS Hospital, associated with University of Buenos Aires, Multiple sclerosis (MS) is a chronic, immune (PwMS). Buenos Aires, Argentina mediated, and neurodegenerative disease that Paula Tagliani generates physical, cognitive, and emotional dis- In general, only half of patients with chronic dis- Servei de Neurologia- Neuroinmunologia, Centre abilities. It affects young people with a female eases have optimal adherence to treatment.4 d’Esclerosi Multiple de predominance.1,2 Progression in MS varies from Adherence to treatment (also referred to taking Catalunya, (CEMCAT), Vall d’Hebron Hospital person to person. Furthermore, the various types medication according to the dosage of the pre- Universitari, Universitat of MS disease also progress in different ways. scribed program) and persistence to treatment Autónoma de Barcelona, Barcelona, Spain Most people with MS experience an increase in (i.e., continuation of treatment by the patient Maximiliano Smietniansky physical disability at some point and it worsens from the beginning to the interruption of treat- Department of Internal Medicine Interna, (progression) when disease activity causes neuro- ment for any reason) of DMTs are both critical Programa de Medicine degeneration.3 Several disease-modifying thera- for the achievement of optimal outcomes in Geriatric Program, Hospital Italiano de pies (DMTs) have shown efficacy in reducing PwMS. Different studies on DMT have shown Buenos Aires, Buenos relapse frequency and progression of disability, that multiple factors may affect adherence and Aires, Argentina journals.sagepub.com/home/tan 1 Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 License (https://creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). Therapeutic Advances in Neurological Disorders 14 persistence, including patients who are tired of Patient-related. Patient-related factors were age, chronic treatment, perceived loss of efficacy, sex, family environment, educational level, associ- adverse reactions, injection device problems, and ated comorbidities, including cardiovascular dis- dosing frequency.5 eases (arrhythmias, congestive heart failure, valvular disease, myocardial infarction), liver dis- Adherence and persistence rates to DMTs in MS ease, chronic lung disease, diabetes, autoimmune range from 49% to 88%.1 Identification of these diseases (systemic lupus erythematosus, rheuma- factors might contribute to the planning of dis- toid arthritis, Hashimoto’s disease, myasthenia ease management programs, improving the prog- gravis), and cancerous tumor.7 Cognitive com- nosis and quality of life of patients.3 Few previous plaint was also included (using a 5-point Likert studies have evaluated adherence to treatment in scale). MS in Latin America. The aim of the present study is to identify predictors related to adher- Disease-related. We collected data on disease ence and persistence to DMTs in a cohort of duration (<10 years / >10 years), presence of relapse, Argentinian MS patients. steroid treatment and hospitalizations in the pre- vious year, fatigue quantified with the Fatigue Impact Scale of Daily Use (D-FIS),8 depression Methods assessed using the PHQ-2 scale,9 and disability A cross-sectional study was conducted in a cohort measured with the Patient Determined Disease of PwMS older than 18 years included in the drug Scale (PDDS).10 dispensing database from the National Medical Care Program (PAMI) between 1 January 2017 Drug-related. Information on drug selection and 1 October 2017, in Argentina. The pharmacy (chosen only by treating doctor / shared with database had only 9 months of follow up available patient), DMT forgotten (by 5-point Likert at the time we accessed it. PAMI is a government- scale), presence of adverse effects, polypharmacy funded social insurance program which serves (defined as ⩾5 drugs per day), and route of 5 million people older than 65 years, people with administration (injectable: subcutaneous/intra- disabilities unable to work, pensioners, and war muscular/intravenous versus oral) were collected. veterans.6 PwMS receiving DMT were included. Health-system-related. Information on neuro- From October 2019 to May 2020, a brief tele- logical appointments per year, delay in the deliv- phone survey (Supplemental Material) was con- ery of DMT, administrative requirements per ducted by neurologists and previously trained year, and person in charge of carrying out the medical students. Patients were contacted at least administrative paperwork were noted. twice at different times. The protocol was approved by the Research Ethics Committee of the César Milstein Hospital (Number 14-2019 / 27.06.2019). Statistical analyses Oral informed consent was given by telephone by Based on the results, descriptive statistics of cat- all participants (see supplemental material). egorical variables were reported as absolute fre- quency and percentages. Continuous variables were described as mean and standard deviation Outcome variables (SD) or median values and interquartile range The medication possession rate (MPR) was used (IQR) according to the data distribution. The to estimate adherence and was defined as the Shapiro–Wilk test was applied to evaluate normal number of daily doses of drugs dispensed by the distribution of variables. A bivariate model and pharmacy to each patient divided by expected three logistic regression models were constructed drug collection during the study period, expressed to estimate odd ratios (ORs) to identify possible as a percentage. A MPR ⩾80% was considered factors independently associated with adherence adherent and < 80% as non-adherent. Adherence to DMT. was modeled as a categorical variable. The first model included age, sex, educational level, memory disorders, family environment, and Predictor variables comorbidities as demographic predictors. The Data were collected in the following four groups: second model included the predictors of the 2 journals.sagepub.com/home/tan G Zanga, E Drzewiscki et al. Figure 1. Flow chart. MS, multiple sclerosis. disease: duration time, relapses, hospitalizations, sample to a minimum of 288 participants. Data use of steroids, fatigue, depression, and current analysis was conducted using the STATA 13, disability. The third model included predictors of Stata Statistical Software: Release 13. StataCorp the drug used: drug choice, forgetting medica- LP, College Station, TX, USA. For all analyses, tion, medication supervision, adverse effects, p values < 0.05 were considered statistically polypharmacy, and route of administration. The significant. fourth model included predictors of the health system: neurological appointments, person in charge of administrative burden, administrative Results requirements, and delay in delivery of DMT. Of the 648 patients registered in our database, only 360 were contacted by phone, with a survey Adherence as an event of interest was coded as 1. response rate of 55.5% (Figure 1). Clinical and The Kruskal–Wallis nonparametric analysis test demographical data of the population included was used to estimate differences in the relevant are shown in Table 1. The mean age was 55.3 years subgroups of drugs used. The sample size was (SD 12) with a female predominance (60%). At calculated based on an event rate of 35.5% the time of the survey, 308 (85.5%) patients were according to our previous study11 with the inten- receiving DMT. The remaining 52 (14.5%) had tion of including at least 8 predictor variables in non-persistence to treatment: 30 patients (57.7%) the model. Taking into account the rule of due to a lack of authorization from their health thumb, 10 events are needed for each variable insurance provider and 22 (42.3%) because of included in the model.12 According to this calcu- medical decisions. lation, a minimum of 240 participants needs to be included. We allowed for 20% of possible As shown in Table 2, 156 (43%) patients had missing data or data errors and increased the comorbidities. A total of 103 PwMS reported journals.sagepub.com/home/tan 3 Therapeutic Advances in Neurological Disorders 14 Table 1. Demographic characteristics among non-adherent and adherent: bivariate model. Adherence MPR < 80% MPR ⩾ 80% p value n (%) 161 (52.3) 147 (47.7) – Age, median (IQR) 57 (49–65) 56 (46–64) 0.27 Sex, n (%) 0.17 Female 101 (63) 81 (55) Male 60 (37) 66 (45) Education level, n (%) 0.49 Primary 45 (28) 32 (22) Secondary 74 (46) 79 (54) Tertiary 30 (19) 24 (16) University 12 (7.5) 12 (8) Memory disorders, n (%) 0.85 Never 57 (35) 51 (35) Occasionally 66 (41) 54 (37) Frequently 28 (17) 32 (22) Usually 6 (4) 5 (3) Always 4 (2.5) 5 (3) Familiar surroundings, n (%) 0.61 Alone 25 (15.5) 26 (18) With others 136 (84.5) 121 (82) Comorbidities, n (%) 69 (43) 61 (41.5) 0.81 IQR, interquartile range; MPR, medication possession rate. Table 2. Comorbidities in MS patients. depression, resulting in a prevalence of depres- sion of 33.4%, and 201 (55.5%) had limitations Comorbidities n (%) in gait and mobility (Figure 2). Cardiovascular disease 65 (18) Of the PwMS under treatment, 147 (47.7%) Diabetes 52 (14) were adherent to their DMT. Median adherence was 78% (IQR 56–100). As shown in Figure 3, Chronic lung disease 20 (5) the median distribution of adherence was greater Cancerous tumor 10 (3) among oral drugs compared with injectables. The parenteral route of administration was the Inmune-mediated diseases 7 (2) most used in 197 (64%) PwMS, as shown in Table 3. We detected that 86 (27%) patients had Liver disease 2 (1) adverse effects that could affect the persistence of Total 156 (43) treatment, among them: 36 (11%) were intoler- ant to injectables, 26 (8.2%) experienced diges- MS, multiple sclerosis. tive intolerance, 12 (3.8%) had lymphopenia, 9 4 journals.sagepub.com/home/tan G Zanga, E Drzewiscki et al. Figure 2. PDDS Score. PDDS, patient determined disease steps. Figure 3. Median adherence to injectable versus oral therapies. Expressed CI 5–95% and IQR 25–75%. Self- injectables: IFN, AG. Injectables: NTZ, ALZ. Oral: DMF, FG, TF, CLD. GA, glatiramer acetate; ALZ, alemtuzumab; CLD, cladribine; DMF, dimethyl fumarate; FG, fingolimod; IFN, interferon beta 1a-1b; NTZ, natalizumab; TF, teriflunomide. CI, confidence interval; IQR, interquartile range. (2.8%) showed infections, 2 (0.63%) suffered In the multivariate model, we found an associa- uncontrolled hypertension, and 1 (0.31%) suf- tion between the oral route of administration and fered thrombocytopenia. adherence (OR 2.02, 95% CI 1.24–3.29, journals.sagepub.com/home/tan 5 Therapeutic Advances in Neurological Disorders 14 Table 3. DMTs used. Different variables that could affect adherence to treatment were evaluated, as summarized in DMT n (%) Table 4. Interferon Beta 128 (41.5) No statistically significant association in demo- Glatiramer Acetate 58 (19) graphic, disease, or health system variables was Fingolimod 67 (22) found. We detected a decrease in adherence, which did not reach statistical significance, among Dimethyl fumarate 34 (11) the number of neurological appointments per year and the person in charge of carrying out the Teriflunomide 11 (3.5) administrative burden. Cladribine 2 (1) Natalizumab 5 (2) Discussion Alemtuzumab 3 (1) The results of this cross-sectional study showed that approximately 5 out of 10 PwMS were non-adher- DMT, disease-modifying therapy. ent/non-persistence to treatment. Adherence has been defined by the World Health Organization (WHO) as “the degree to which the person’s behav- p = 0.005). Figure 4 shows the median adherence ior corresponds with the agreed recommendations of each drug. In a secondary post hoc analysis, we from a health care provider” (WHO, 2003),13 while found that the main demographic predictor for persistency refers to the amount of time that a patient those who received oral treatment was tertiary or remains on chronic drug therapy.14 Adherence to universitary educational level (OR 2.86, 95% CI treatment in MS varies according to the quantifica- 1.41–5.81, p = 0.004). tion method used.15 Our results were superior to Figure 4. Median of adherence according to DMTs. Expressed confidence interval 5-95% and interquartile range 25-75% IFN, interferon beta 1a-1b; GA, glatiramer acetate; DMF, dimethyl fumarate; FG, fingolimod; TF, teriflunomide; CLD, cladribine; NTZ, natalizumab; ALZ, alemtuzumab. 6 journals.sagepub.com/home/tan G Zanga, E Drzewiscki et al. Table 4. Characteristics of the disease, drug used, and health system associated with adherence and persistence to DMTs: logistic regression model. n = 308 OR p value 95% CI Disease Disease duration time/years, mean (IQR) 14.5 (13) 0.99 0.51 0.96–1.01 Relapses/last year, n (%) 117 (38) 0.80 0.49 0.42–1.51 Hospitalization for MS/last year, n (%) 49 (16) 0.65 0.26 0.30–1.37 Use steroid/last year, n (%) 90 (29) 1.33 0.38 0.70–2.55 Fatigue (D-FIS), n (%) 297 (96) 1.28 0.19 0.88–1.87 Depression (PHQ-2), n (%) 103 (33) 0.64 1.40 0.36–1.14 Disability Score (PDDS), mean (DS) 2.25 (2) 1.00 0.96 0.52–1.60 Drug Share drug election, n (%) 85 (27.5) 1.04 0.86 0.63–1.74 Forget medication 107 (35) 1.01 0.93 0.70–1.46 Medication supervision 163 (53) 1.12 0.63 0.69–1.80 Adverse effects, n (%) 86 (28) 0.95 0.47 0.83–1.08 Polypharmacy, n (%) 128 (42) 1.17 0.52 0 .72–1.91 Oral route administration n (%) 111 (36) 1.83 0.014* 1.13–3.00 Health system Neurological appointments per year, median (IQR) 3 (2) 0.92 0.09 0.84–1.01 Another person in charge of administrative burden, n (%): 158 (52) 0.63 0.06 0.39–1.02 Administrative requirements per year, median (IQR) 2 (1) 1.09 0.45 0.87–1.36 Delay in delivery of DMT, median weeks (IQR) 2 (1) 1.01 0.89 0.85 - 1.21 *P <0.05, statistically significant. DMTs, Disease-modifying therapies; OR, Odds Ratio; CI, confidence interval; IQR, interquartile range; D-FIS, Fatigue Impact Scale for Daily Use; PDDS, Patient Determined Disease Scale. those described by Hansen et al., who showed an absence of injection-related issues.20 Nonetheless, adherence of 30–40%, although lower than those other studies have not found d ­ ifferences between reported by other researchers, which varied between oral and injectable DMTs.3,23–25 Patients with 59.6% and 76.4% in different studies.16–19 comorbid diseases have shown difficulties in adhering to the proposed medication.26 Of the 21 variables analyzed, only the route of drug administration had significant implications One third of our PwMS reported symptoms sug- for non-adherence/non-persistence. We identified gestive of depression. The prevalence data ranges that patients who used the injectable route were from 19 to 54% in PwMS.27 Depression showed more likely to be non-adherent. Several previous an association with non-adherence in several studies obtained similar findings to the present studies.3,28 Tarrants et al. showed that patients study.20–22 Better adherence to oral drugs is prob- with depression were half as likely to be adherent ably due to an easier administration and the to their treatment.32 Memory disorders also journals.sagepub.com/home/tan 7 Therapeutic Advances in Neurological Disorders 14 negatively affect adherence.28 About 68.5% of the et al., we did not find an association between cur- PwMS reported worsening memory symptoms. rent disability and adherence.25 Similar results were obtained by Tremlett et al.27 In our study, we did not find cognitive disorders Our patients had 3.8 neurological visits on average are a predictor of adherence. However, our study per year. This association was previously studied did not use any specific test to evaluate this and it evidenced a greater number of outpatient variable. medical visits (mean 5.6/year) in adherent patients.23 Missing more than 20% of scheduled Only 27.5% of the PwMS expressed shared deci- appointments was identified as the most important sions between doctor and patient regarding drug factor associated with suboptimal adherence.18 election. A previous study from our country revealed that the frequency of shared decisions In our country, 32.3% of the PwMS expressed ranges between 76% and 79%.30 The disparity problems obtaining the medication from the medi- of values is probably due to differences in the cal insurance provider.37 Those with a state-run population included. Ben-Zacharia et al.1 health insurance reported a longer waiting time described that patients obtained better levels of (4.5–6 weeks) for first prescription and mainte- adherence when the therapeutic decision was nance of therapies and inappropriate delivery of shared. DMT compared with other types of medical insur- ance.38 As a requirement of health insurance policy, More than one-third of the patients reported hav- medication for MS is requested through chronic ing forgotten to take the medication. Between medication forms, which must be updated every 16% and 51% missed at least 1 dose of their med- 6 months. Our PwMS reported a waiting time of ication in the short and long term. Unlike other 2 weeks for administrative procedures, which could authors, we did not find an association with delay the start and/or continuity of treatment. adherence.24,27,31, Besides that, between 45% to 62% of patients discontinued the treatment due Given the heterogeneity in the data currently to adverse events.30,32–34 A quarter of our PwMS available, we consider it essential to continue this presented adverse events that justified discontin- line of research to establish more predictor varia- uation of treatment. Previous studies have inves- bles, facilitating the correct selection of treat- tigated polypharmacy and its consequences in ment, and taking into account the possible lower adherence patients.26,30 We obtained a barriers that could affect adequate adherence polypharmacy prevalence rate of 35.5%, while, in when starting a new therapy. a recent review, the reported rate varies between 15% and 59%.32 Several limitations in this study should be noted. Firstly, this was a cross-sectional study and it is Several studies showed that short disease dura- not possible to assess changes over time to explore tion and the absence of clinical relapses deter- causality. Secondly, our response rate was 55%, as mined better adherence.26,31,35 A minority of is the case in most telephone surveys,39,40 which patients reported hospitalizations for MS in the could introduce selection bias because the charac- previous months. Two studies showed that adher- teristics of the participants who respond and those ent patients were less likely to have MS-related who do not can differ. Third, this cohort is limited hospitalizations in the previous 6 months and in to patients with PAMI medical insurance. This the subsequent 12 months. This suggests a 50% population has distinctive characteristics, such as decrease in the hospitalization rate in adherent advanced age, long duration of disease, and high patients.17,25 use of injectable therapies. Since only this particu- lar health system was evaluated, we suggest being The patients who reported symptoms of fatigue cautious when drawing conclusions and extrapo- did not show an impact on therapeutic adherence, lating the findings to other populations. similar to that reported by Van Gaudecker.36 In contrast, other authors showed a relationship Fourthly, we do not have direct clinical data between fatigue and non-adherence.33,37 The related to adherence provided by physicians (e.g., assessment of disability with adherence showed medical reasons, therapeutic failure, progressive mixed results.21,35 More than half of our PwMS forms, adverse effects). This study cannot distin- required assistance to walk. Unlike Pengxiang guish between non-persistent patients (who could 8 journals.sagepub.com/home/tan G Zanga, E Drzewiscki et al. restart a new treatment in the future) from those References who definitively discontinued treatment.14 1. Ben-Zacharia A, Adamson M, Boyd A, et al. Finally, there is no data on PwMS medicated Impact of shared decision making on multiple with peginterferon beta 1a and Ocrelizumab, sclerosis. Int J MS Care 2018; 6: 287–297. both recently approved in our country at the time 2. Kołtuniuk A and Rosińczuk J. Adherence to of the study. Despite these limitations, this study disease-modifying therapies in patients with provides an initial approach to the potential vari- multiple sclerosis. Patient Prefer Adherence 2018; ables that could affect adherence to treatment in 12: 1557–1566. PwMS not previously evaluated in Argentina and 3. Munsell M, Frean M, Menzin J, et al. An this warrants further study. evaluation of adherence in patients with multiple sclerosis newly initiating treatment with a self- In conclusion, adherence is a key factor when injectable or an oral disease-modifying drug. optimizing management of PwMS. In this real- Patient Prefer Adherence 2017; 11: 55–62. world study, we showed better adherence and persistence to oral therapies in MS patients in 4. Menzin J, Caon C, Nichols C, et al. Narrative review of the literature on adherence to disease- Argentina. This result should be considered by modifying therapies among patients with multiple the healthcare professional when evaluating the sclerosis. J Manag Care Pharm 2013; 1: 19. risk of non-adherence/non-persistence with ther- apies. It is essential to implement efforts to 5. Halpern R, Agarwal S, Dembek C, et al. improve treatment adherence, disease care, and Comparison of adherence and persistence patients’ quality of life and to reduce public health among multiple sclerosis patients treated with costs for MS patients. disease-modifying therapies: a retrospective administrative claims analysis. Patient Prefer Adherence 2011; 5: 73–84. Acknowledgements We would like to thank Maria Celica Ysrraelit for 6. INSSJP - Instituto Nacional de Servicios Sociales her valuable recommendations on the manu- para Jubilados y Pensionados. https://www.pami. script, as well as the following people who have org.ar/historia (accessed 22 Mar 2020). put the time and effort in collecting the data pre- 7. Marrie RA, Cohen J, Stuve O, et al. A systematic sented in this work: Gabriela Costa, Federico review of the incidence and prevalence of Ramos, Gonzalo Lage, Carlos Turzio, Florencia comorbidity in multiple sclerosis: overview. Mult Tucillo, and Martin Arnedillo. Scler J 2015; 21: 263–281. 8. Fiest KM, Fisk JD, Patten SB, et al. Fatigue and Conflict of interest statement comorbidities in multiple sclerosis. Int J MS Care The authors declare that there is no conflict of 2016; 2: 96–104. interest. 9. Kroenke K, Spitzer RL and Williams JBW. The Funding patient health questionnaire-2: validity of a two- The authors disclosed receipt of the following item depression screener. Med Care 2003; 41: financial support for the research, authorship, and/ 1284–1292. or publication of this article: GZ disclosed receipt 10. Learmonth YC, Motl RW, Sandroff BM, et al. of the following financial support for the research, Validation of patient determined disease steps authorship and/or publication of this article: this (PDDS) scale scores in persons with multiple study was supported by a research grant awarded sclerosis. BMC Neurol 2013; 13: 37. by the Florencio Fiorini Foundation. The article 11. Zanga G, Caruso D, Drzewiscki E, et al. processing charges was funded by Roche. Adherence to treatment of multiple sclerosis in a health care program. Med 2021; 81: 311–317. ORCID iD Gisela Zanga https://orcid.org/0000-0002- 12. Peduzzi P, Concato J, Feinstein AR, et al. 2730-6734 Importance of events per independent variable in proportional hazards regression analysis II. J Clin Epidemiol 1995; 48: 1503–1510. Supplemental material Supplemental material for this article is available 13. Dobbels F, Van Damme-Lombaert R, Vanhaecke J, online. et al. Growing pains: non-adherence with the journals.sagepub.com/home/tan 9 Therapeutic Advances in Neurological Disorders 14 immunosuppressive regimen in adolescent factors in a national sample of medicare patients transplant recipients. Pediatr Transplant 2005; 9(3): with multiple sclerosis. Value Heal 2019; 23: 381–390. 328–334. 14. Sikka R, Xia F and Aubert R. Estimating 26. Bayliss EA, Steiner JF, Fernald DH, et al. medication persistency using administrative claims Descriptions of barriers to self-care by persons data. Am J Manag Care 2005; 11: 449–457. with comorbid chronic diseases. Ann Fam Med 2003; 1: 15–21. 15. Pagés NVM. Methods to assess medication adherence. Ars Pharm 2018; 59: 163–172. 27. Tremlett H, Van Der Mei I, Pittas F, et al. Adherence to the immunomodulatory drugs 16. Hansen K, Schüssel K, Kieble M, et al. for multiple sclerosis : contrasting factors Adherence to disease modifying drugs among affect stopping drug and missing doses. patients with multiple sclerosis in Germany: a Pharmacoepidemiol Drug Saf 2008; 17(6): retrospective cohort study. PLoS One 2015; 10: 565–576. 1–12. 28. Bruce JM, Hancock LM, Arnett P, et al. 17. Tan H, Cai Q, Agarwal S, et al. Impact of Treatment adherence in multiple sclerosis: adherence to disease-modifying therapies on association with emotional status, personality, clinical and economic outcomes among patients and cognition. J Behav Med 2010; 3(33): with multiple sclerosis. Adv Ther 2011; 28: 219–227. 51–61. 29. Carnero Contentti E, Giachello S and Correale 18. Gromisch ES, Turner AP, Leipertz SL, et al. Risk J. Barriers to access and utilization of multiple factors for suboptimal medication adherence in sclerosis care services in a large cohort of Latin persons with multiple sclerosis: development of American patients. Mult Scler J 2020; 27(1): an electronic health record-based explanatory 117–129. model for disease-modifying therapy use. Arch Phys Med Rehabil 2020; 101(5): 807–814. 30. Bazargan M, Smith J, Yazdanshenas H, et al. Non-adherence to medication regimens among 19. Evans C, Marrie RA, Zhu F, et al. Adherence older African-American adults. BMC Geriatr and persistence to drug therapies for multiple 2017; 17: 1–12. sclerosis: a population-based study. Mult Scler Relat Disord 2016; 8: 78–85. 31. Devonshire V, Lapierre Y, Macdonell R, 20. Fernández O, Agüera J, Izquierdo G, et al. et al. The global adherence project (GAP): a Adherence to interferon β-1b treatment in multicenter observational study on adherence patients with multiple sclerosis in Spain. PLoS to disease-modifying therapies in patients with One 2012; 7: 1–7. relapsing-remitting multiple sclerosis. Eur J Neurol 2011; 18: 69–77. 21. Thach AV, Brown CM, Herrera V, et al. Associations between treatment satisfaction, 32. Frahm N, Hecker M and Zettl UK. medication beliefs, and adherence to disease- Polypharmacy among patients with multiple modifying therapies in patients with multiple sclerosis: a qualitative systematic review. Expert sclerosis. Int J MS Care 2018; 20: 251–259. Opin Drug Saf 2020; 19: 139–145. 22. Turner AP, Williams RM, Sloan AP, et al. 33. Treadaway K, Cutter G, Salter A, et al. Injection anxiety remains a long-term barrier Factors that influence adherence with disease- to medication adherence in multiple sclerosis. modifying therapy in MS. J Neurol 2009; 256: Rehabil Psychol 2009; 54: 116–121. 568–576. 23. Burks J, Marshall TS and Ye X. Adherence to 34. Río J, Porcel J, Téllez N, et al. Factors related disease-modifying therapies and its impact on with treatment adherence to interferon β and relapse, health resource utilization, and costs glatiramer acetate therapy in multiple sclerosis. among patients with multiple sclerosis. Clin Mult Scler 2005; 11: 306–309. Outcomes Res 2017; 9: 251–260. 35. McKay KA, Evans C, Fisk JD, et al. Disease- 24. Morillo Verdugo R, Herráiz ER, Olmo RF, et al. modifying therapies and adherence in multiple Adherence to disease-modifying treatments in sclerosis: comparing patient self-report with patients with multiple sclerosis in Spain. Patient pharmacy records. Neuroepidemiology 2017; 1: Prefer Adherence 2019; 13: 261–272. 124–130. 25. Li P, Ladage VP, Berger J, et al. Disease- 36. Von Gaudecker JR. Factors affecting the modifying therapy adherence and associated adherence to disease-modifying therapy in 10 journals.sagepub.com/home/tan G Zanga, E Drzewiscki et al. patients with multiple sclerosis. J Neurosci Nurs. patients. Mult Scler Relat Disord 2019; 33: 2018; 50: 302. 88–93. 37. Tarrants M, Oleen-Burkey M, Castelli-Haley 39. O’Toole J, Sinclair M and Leder K. Maximising J, et al. The impact of comorbid depression on response rates in household telephone surveys. adherence to therapy for multiple sclerosis. Mult BMC Med Res Methodol 2008; 8: 1–5. Scler Int 2011; 2011: 1–10. 40. Glass DC, Kelsall HL, Slegers C, et al. A Visit SAGE journals online 38. Carnero Contentti E, Pettinicchi JP, López telephone survey of factors affecting willingness to journals.sagepub.com/ home/tan PA, et al. Access and unmet needs to multiple participate in health research surveys. BMC Public sclerosis care in a cohort of Argentinean Health 2015; 15: 1. SAGE journals journals.sagepub.com/home/tan 11