You are here
Adherence and persistence to drug therapies for multiple sclerosis: A population-based study
Multiple Sclerosis and Related Disorders, Volume 8, July 2016, Pages 78 - 85
We aimed to estimate the prevalence and predictors of optimal adherence and persistence to the disease-modifying therapies (DMT) for multiple sclerosis (MS) in 3 Canadian provinces.
We used population-based administrative databases in British Columbia (BC), Saskatchewan, and Manitoba. All individuals receiving DMT (interferon-B-1b, interferon-B-1a, and glatiramer acetate) between 1-January-1996 and 31-December-2011 (BC), 31-March-2014 (Saskatchewan), or 31-March-2012 (Manitoba) were included. One-year adherence was estimated using the proportion of days covered (PDC). Persistence was defined as time to DMT discontinuation. Regression models were used to assess predictors of adherence and persistence; results were pooled using random effects meta-analysis.
4830 individuals were included. When results were combined, an estimated 76.4% (95% CI: 69.1–82.4%) of subjects exhibited optimal adherence (PDC ≥80%). Median time to discontinuation of the initial DMT was 1.9 years (95% CI: 1.6–2.1) in Manitoba, 2.8 years (95% CI: 2.5–3.0) in BC, and 4.0 years (95% CI: 3.5–4.6) in Saskatchewan. Age, sex and socioeconomic status were not associated with adherence or persistence. Individuals who had ≥4 physician visits during the year prior to the first DMT dispensation were more likely to exhibit optimal adherence compared to those with fewer (0–3) physician visits.
We observed adherence that is higher than what has been reported for other chronic diseases, and other non-population-based MS cohorts. Closer examination as to why adherence appears to be relatively better in MS and how adherence influences disease outcomes could contribute to our understanding of MS, and prove useful in the management of other chronic diseases.
- Adherence and persistence were examined in a large population-based MS cohort.
- Optimal adherence was observed in 76% of patients.
- Adherence was higher than previously reported results from other chronic diseases.
- Adherence was higher than previously reported in non-population-based MS cohorts.
- Sociodemographic factors were not consistently associated with adherence.
Keywords: Multiple sclerosis, Disease-modifying therapies, Medication adherence, Medication persistence.
The therapeutic options for managing multiple sclerosis (MS) have grown considerably over the past 20 years, with several disease-modifying therapies (DMT) now available. While efficacy rates vary between agents, the DMTs all reduce disease activity to some extent (Filippini et al, 2003, PRISMS, 1998, Jacobs et al, 1996, and The IFNB Multiple Sclerosis Study Group, 1993). However, even the most efficacious DMT will have limited effectiveness if adherence to the prescribed treatment is poor.
Adherence to drug therapies for many chronic conditions is estimated at only 50% (World Health Organization, 2003 and Osterberg and Blaschke, 2005), and non-adherence has been associated with increased morbidity, mortality, and health care costs (World Health Organization, 2003, Osterberg and Blaschke, 2005, and Simpson et al, 2006). While slightly higher adherence rates have been reported in MS, most studies have evaluated small, non-population-based cohorts, with varying definitions for adherence, and relied on self-reported outcomes over short observation periods, which limits the applicability of the results (Menzin et al., 2013). Findings regarding the factors associated with levels of adherence, including age, sex, socioeconomic status, and comorbidity have been inconsistent, therefore it is unknown whether these factors play a role in adherence to the DMTs for MS. Better knowledge of factors influencing adherence is important for clinicians and patients.
We aimed to estimate the level of adherence to first-line DMTs, and to identify factors associated with optimal adherence using population-based data from three Canadian provinces.
2. Materials and methods
2.1. Data sources
We used population-based health administrative data from British Columbia (BC), Saskatchewan (SK), and Manitoba (MB). Each province maintains its own health services databases and population registry that are linkable using anonymized unique identifiers. Although maintained separately, these databases all contain information on hospital separations (Canadian Institute for Health Information creator, 2012), physician services (British Columbia Ministry of Health creator, 2012), prescription drug dispensations (BC Ministry of Health creator, 2012), vital status, and residency (British Columbia Ministry of Health creator, 2012). Almost the entire population in each province (BC, 4.6 million; SK, 1.1 million; MB, 1.3 million) is captured, with the exception of member of the armed forces, Royal Canadian Mounted Police, and federal penitentiary inmates. Prescription costs for registered First Nations patients (<10% of the population) are paid for by another government agency, and were only captured by the prescription databases for part of the study period. These databases have been successfully accessed to investigate medication adherence and persistence in other chronic diseases (Ediger et al, 2009, Evans et al, 2012, and Dormuth et al, 2009). Ethics approval was granted by the governing Research Ethics Boards in each province, and approval for administrative data access was provided by the relevant provincial data steward.
2.2. Study cohort
The study period varied between provinces due to data availability. It began on January 1, 1996 in all provinces and ended either December 31, 2011 (British Columbia), March 31, 2014 (Saskatchewan) or March 31, 2012 (Manitoba). All subjects who received a prescription dispensation for a first-line DMT (beta-interferon-1b, beta-interferon-1a, glatiramer acetate) during their respective study period were eligible for inclusion. Natalizumab is second-line therapy in Canada, and due to small numbers was not included in the analyses. Oral therapies were also not included as the study preceded their approval in Canada.
Individual subjects were followed from their index date (first dispensation of a DMT) until their exit date of death, loss of beneficiary status, or the end of the study period, whichever came first. Because women are advised to discontinue DMT three months before conception, women with a delivery diagnosis code (International Classification of Disease [ICD]-9/10 codes 630-679 and O00-O99) were censored one year before their delivery date or at their exit date if it fell within this one year period. All subjects had to have at least one year of follow-up before and after their index date. No other formal exclusion criteria were applied. However, as only those subjects with relapsing-onset MS are eligible to receive coverage for a DMT in these provinces, subjects with a primary progressive disease course would, by definition, be excluded. The first DMT was approved for use in Canada in mid-1995, but did not receive provincial formulary coverage for another 1–2 years (depending on the province). This means that virtually all included subjects were incident users (i.e. had not received a DMT prior to their index date). This, combined with similar prescribing criteria across the three provinces (Appendix e-A), allows for baseline comparability of the subjects.
2.3. Study outcomes
The primary outcome was the proportion of patients achieving optimal adherence at one year. Adherence was estimated using the proportion of days covered (PDC), one of the most commonly used measures employed with administrative data (Ho et al., 2009). It provides a more conservative measure of adherence, and has been recommended when measuring adherence to a class of medications (Horne et al., 2013). The PDC was calculated by dividing the number of days of DMT supplied by the number of days of observation (Ho et al, 2009 and Martin et al, 2009). A PDC ≥80% was considered optimal. The 80% threshold is widely used in adherence research, and has been associated with fewer hospitalizations and decreased mortality (Osterberg and Blaschke, 2005, Simpson et al, 2006, and Karve et al, 2009). To be comparable with other studies, we estimated the proportion of subjects with optimal adherence over their entire study period (i.e. index to exit date) as a sensitivity analysis. We also examined patterns of DMT persistence, including the median time to discontinuation of the initial and any DMT, and the proportion of subjects who discontinued their DMT within the first 6, and 12 months of therapy. A discontinuation of any DMT was defined as a >90-day interruption in treatment. A discontinuation of the initial DMT was defined as more than 90 days of treatment interruption or a switch to another first- or second-line (natalizumab, fingolimod) DMT.
2.4. Statistical analysis
We described the demographic and clinical characteristics of subjects at the index date using frequencies, means, and standard deviations. We used multivariable logistic regression to examine the association between optimal adherence and the following covariates: age (continuous variable), sex, initial DMT dispensed, index year (1996–1998, 1999–2000, 2001–2002, 2003–2005, 2006–2008, 2009-study end), and socioeconomic status in quintiles. Socioeconomic status was estimated by linking the first 3 postal code digits to Canadian census data to determine median household income for the residential area. We also included the number of physician visits, hospitalizations and non-MS medication classes (using the Anatomical Therapeutic Chemical classification at the fourth level) in the one year before the index date as a proxy measure of health care utilization, comorbidity, and “pill burden”. Covariates were selected based on statistical significance, clinical relevance or both. The time to DMT discontinuation was estimated using Kaplan-Meier survival analysis; Cox regression was used to identify any associations between the above covariates and DMT discontinuation. Multivariable logistic regression was also used to examine the association between the covariates and DMT discontinuation within 6, and 12 months of DMT initiation.
Analyses were conducted separately in each province and combined using random effects meta-analysis. Random effects models were chosen, because tests for heterogeneity (I2 test) suggested moderate (25–50%) to high (>75%) levels of heterogeneity between the outcome data from the 3 provinces (Higgins et al., 2003). Statistical analyses were performed using SAS (Enterprise Guide 4.3), and R (Version 3.1.0).
A total of 4830 subjects were included in the analyses (BC, n=2323; SK, n=1297; MB, n=1210). At the index date, the mean age was 40.4 years (SD 10.0), and 75.8% were female. Most subjects visited a physician at least 4 times in the year before their index date, and were receiving at least one non-MS medication. The use of DMTs varied by province; beta-interferon-1a (subcutaneous) was most often initiated in British Columbia, glatiramer acetate in Saskatchewan, and beta-interferon-1b in Manitoba (Table 1).
Characteristics of the multiple sclerosis subjects overall and by each Canadian province.
|Characteristic||Overall Cohort||British Columbia||Saskatchewan||Manitoba|
|Age at index date (years), mean (SD)||40.4 (10.0)||41.1 (10.0)||39.0 (9.9)||40.6 (10.2)|
|Female||3661 (75.8)||1800 (77.5)||945 (72.9)||916 (75.7)|
|Male||1169 (24.2)||523 (22.5)||352 (27.1)||294 (24.3)|
|Neighbourhood income quintile at index date (%)|
|1 (Lowest)||821 (17.0)||390 (18.8)||176 (13.6)||255 (21.1)|
|2||890 (18.4)||397 (17.1)||257 (19.8)||236 (19.5)|
|3||1015 (21.0)||477 (20.5)||274 (21.1)||264 (21.8)|
|4||1022 (21.1)||493 (21.2)||296 (22.8)||233 (19.3)|
|5 (Highest)||1001 (20.7)||485 (20.9)||294 (22.7)||222 (18.3)|
|Missing||81 (1.7)||81 (3.5)|
|Index year (%)|
|1996–1998||586 (12.1)||297 (12.8)||192 (14.8)||97 (8.0)|
|1999–2000||944 (19.5)||441 (19.0)||245 (18.9)||258 (21.3)|
|2001–2002||861 (17.8)||430 (18.5)||205 (15.8)||226 (18.7)|
|2003–2005||1040 (21.5)||493 (21.2)||246 (19.0)||301 (24.9)|
|2006–2008||785 (16.3)||391 (16.8)||187 (14.4)||207 (17.1)|
|2009 – study end||614 (12.7)||271 (11.7)||222 (17.1)||121 (10.0)|
|Physician visits in year prior to index date (%)|
|0–3||298 (6.2)||164 (7.1)||52 (4.0)||82 (6.8)|
|4–11||1228 (23.3)||516 (22.2)||283 (21.8)||429 (35.4)|
|≥12||3304 (68.4)||1643 (70.7)||962 (74.2)||699 (57.8)|
|Hospitalizations in year prior to index date (%)|
|0||3747 (77.6)||1912 (82.3)||893 (68.9)||942 (88.9)|
|≥1||1083 (22.4)||411 (17.7)||404 (31.1)||268 (22.1)|
|Number of medications (ATC classifications) in year prior to index date (%)|
|0||576 (11.9)||261 (11.2)||163 (12.6)||152 (12.6)|
|1–2||1261 (26.1)||564 (24.3)||380 (29.3)||317 (26.2)|
|3–4||1148 (23.8)||532 (22.9)||333 (25.7)||283 (23.4)|
|≥5||1845 (38.2)||966 (41.6)||421 (32.4)||458 (37.9)|
|DMT at index date (%)|
|IFNB-1b||1519 (31.4)||720 (31.0)||299 (23.1)||500 (41.3)|
|IFNB-1a subcutaneous||1378 (28.5)||855 (36.8)||265 (20.4)||258 (21.3)|
|IFNB-1a intramuscular||688 (14.2)||337 (14.5)||118 (9.1)||233 (19.3)|
|Glatiramer acetate||1245 (25.8)||411 (17.7)||615 (47.4)||219 (18.1)|
|Follow-up time (years), mean (SD)||7.8 (4.0)||8.6 (4.5)||8.0 (3.8)||8.1 (4.1)|
ATC, anatomical therapeutic chemical classification system; DMT, disease modifying therapy; IFNB, beta-interferon.
The proportion of subjects with optimal adherence at one year differed between provinces (p<0.001). Saskatchewan had the highest proportion of adherent subjects at 80.3% (1041/1297), followed by Manitoba with 78.3% (948/1210), and British Columbia with 70.0% (1625/2323). When all 3 provinces were combined, an estimated 76.4% (95% CI: 69.1–82.4%) of subjects exhibited optimal adherence at one year. In contrast, the combined proportion of subjects with optimal adherence over the entire study period was 42.4% (95% CI: 32.5–52.9%), but also varied between provinces (Table 2). Within each individual province there were factors that were associated with optimal adherence (Table 3). However, they were not consistent across provinces. For example increasing age was associated with optimal adherence in Saskatchewan, but not in the other provinces. In the combined cohort, demographics, such as age, sex, and socioeconomic status were not associated with optimal adherence, nor was the type of DMT first used (Table 3). Subjects who visited a physician 4–11 times in the year prior to the index date were more likely to have optimal adherence compared to those who had fewer visits, and although not always statistically significant, the frequency of optimal adherence appeared to be lower in the more recent years (Table 3).
Adherence and persistence to the DMTs, by individual province and overall (combined).
|British Columbia||Saskatchewan||Manitoba||p-valuea||All provinces combinedb|
|n=2323||n=1297||n=1210||(%; 95% CI)|
|Adherence ≥80% one year after index date (%)||1625 (70.0)||1041 (80.3)||948 (78.3)||<0.001||76.4 (69.1–82.4)|
|Adherence ≥80% over entire study period (%)||771 (33.2)||586 (45.2)||598 (49.4)||<0.001||42.4 (32.5–52.9)|
|Subjects who discontinued initial DMT within first 6 months (%)||329 (14.2)||134 (10.3)||280 (23.1)||<0.001||15.2 (9.7–23.1)|
|Subjects who discontinued initial DMT within first 12 months (%)||621 (26.7)||259 (20.0)||437 (36.1)||<0.001||27.1 (19.7–36.2)|
|Subjects who discontinued any DMT within first 6 months (%)||250 (10.8)||99 (7.6)||114 (9.4)||0.01||9.3 (7.6–11.3)|
|Subjects who discontinued any DMT within first 12 months (%)||467 (20.1)||198 (15.3)||213 (17.6)||0.001||17.6 (15.0–20.7)|
b Combined using random effects meta-analysis; DMT, disease-modifying therapy; CI, confidence interval.
Predictors of optimal adherence (≥80%) at one-year after index date.
|British Columbia||Saskatchewan||Manitoba||All provinces combineda|
|Variable||Odds ratio (95% CI)b|
|Age (years)||1.005 (0.996–1.015)||1.04 (1.025–1.056)||1.000 (0.986–1.015)||1.015 (0.991–1.040)|
|Male||1.128 (0.903–1.413)||1.090 (0.786–1.511)||0.999 (0.715–1.395)||1.072 (0.900–1.277)|
|Neighbourhood income quintile|
|2||1.212 (0.895–1.643)||0.950 (0.577–1.563)||1.418 (0.920–2.185)||1.180 (0.864–1.611)|
|3||1.233 (0.921–1.651)||1.090 (0.613–1.660)||1.695 (1.100–2.610)||1.282 (0.898–1.830)|
|4||1.500 (1.115–2.019)||0.960 (0.5900–1.563)||1.102 (0.725–1.675)||1.171 (0.839–1.634)|
|5 (Highest)||1.194 (0.889–1.604)||0.826 (0.510–1.338)||1.352 (0.872–2.096)||1.103 (0.775–1.569)|
|1999–2000||1.066 (0.738–1.537)||1.208 (0.707–2.064)||1.268 (0.711–2.264)||1.171 (0.883–1.553)|
|2001–2002||0.978 (0.6742–1.418)||1.442 (0.804–2.584)||1.658 (0.895–3.073)||1.317 (0.874–1.984)|
|2003–2005||0.813 (0.562–1.172)||1.324 (0.761–2.304)||1.401 (0.798–2.458)||1.137 (0.744–1.736)|
|2006–2008||0.665 (0.454–0.971)||0.793 (0.455–1.384)||0.769 (0.429–1.373)||0.737 (0.551–0.985)|
|2009 – study end||0.627 (0.417–0.941)||0.682 (0.398–1.171)||1.078 (0.542–2.145)||0.768 (0.500–1.183)|
|Physician visits in year prior to index date|
|4–11||2.141 (1.426–3.208)||1.642 (0.781–3.452)||1.336 (0.742–2.405)||1.677 (1.121–2.057)|
|≥12||1.519 (1.025–2.242)||1.403 (0.680–2.893)||1.138 (0.628–2.060)||1.345 (0.958–1.890)|
|Hospitalizations in year prior to index date|
|≥1||0.695 (0.550–0.880)||0.789 (0.578–1.077)||1.070 (0.757–1.512)||0.835 (0.625–1.118)|
|Number of ATC medication classifications in year prior to index date|
|1–2||0.826 (0.581–1.164)||1.022 (0.616–1.695)||1.213 (0.742–1.986)||1.006 (0.725–1.396)|
|3–4||0.894 (0.621–1.279)||0.926 (0.546–1.569)||1.006 (0.605–1.672)||0.940 (0.722–1.225)|
|≥5||0.744 (0.521–1.052)||0.898 (0.528–1.526)||0.981 (0.592–1.626)||0.865 (0.642–1.166)|
|DMT at index date|
|IFNB-1a subcutaneous||1.079 (0.838–1.387)||1.186 (0.712–1.975)||1.288 (0.861–1.926)||1.184 (0.951–1.475)|
|IFNB-1a intramuscular||1.251 (0.909–1.729)||0.641 (0.357–1.152)||0.981 (0.661–1.456)||0.929 (0.600–1.438)|
|Glatiramer acetate||0.879 (0.657–1.178)||0.704 (0.467–1.06)||1.478 (0.956–2.286)||0.970 (0.618–1.525)|
a Random effects meta-analysis.
b Logistic regression models simultaneously adjusted for all listed variables, fitted for each individual province.
CI, confidence interval; ATC, anatomical therapeutic chemical; DMT, disease modifying therapy: IFNB, beta-interferon.
The median time to discontinuation of the initial DMT was 1.9 years (95% CI: 1.6–2.1) in Manitoba, 2.8 years (95% CI: 2.5–3.0) in British Columbia, and 4.0 years (95% CI: 3.5–4.6) in Saskatchewan. Median time to discontinuation of any DMT was 4.6 years (95% CI: 4.1–5.0) in Manitoba, 4.1 years (95% CI: 3.9–4.5) in British Columbia, and 5.9 years (95% CI: 5.3–6.2) in Saskatchewan. Fewer subjects in Saskatchewan discontinued their initial DMT (i.e. switching between DMTs was allowed) within the first 6 and 12 months of therapy when compared to those in British Columbia and Manitoba (Table 2). When combined, nearly 1 in 5 (17.6%, 95% CI: 15.0–20.7) subjects had discontinued any DMTs within the first year of therapy (Table 2). As the number of medications in the year prior to the index date increased the risk of discontinuing a DMT increased (Table 4). An association was also observed with index year: subjects who first received a DMT after 2005 were more likely to discontinue both the initial and any DMT compared to those in earlier years (Table 4). Findings were similar for discontinuations at 6 and 12 months, although generally not statistically significant (Table e-1).
Predictors of DMT discontinuation for all provinces combined.a
|Initial DMT||Any DMT|
|Variable||Hazard ratio (95% CI)||Hazard ratio (95% CI)|
|Age (years)||0.983 (0.971–0.995)||0.985 (0.972–0.999)|
|Male||1.053 (0.944–1.172)||1.076 (0.975–1.187)|
|Neighbourhood income quintile|
|2||1.067 (0.946–1.203)||1.046 (0.924–1.185)|
|3||1.023 (0.865–1.213)||0.950 (0.821–1.100)|
|4||1.050 (0.898–1.229)||0.976 (0.861–1.108)|
|5 (Highest)||1.008 (0.899–1.131)||0.946 (0.839–1.066)|
|1999–2000||1.128 (0.823–1.547)||1.122 (0.954–1.319)|
|2001–2002||1.144 (0.768–1.705)||1.140 (0.906–1.435)|
|2003–2005||1.197 (0.930–1.541)||1.438 (1.041–1.986)|
|2006–2008||1.370 (1.186–1.581)||1.624 (1.359–1.941)|
|2009 – study end||1.352 (1.041–1.757)||1.615 (1.267–2.059)|
|Physician visits in year prior to index date|
|4–11||0.818 (0.627–1.068)||0.827 (0.663–1.032)|
|≥12||0.812 (0.593–1.113)||0.816 (0.642–1.037)|
|Hospitalizations in year prior to index date|
|≥1||1.023 (0.874–1.194)||1.045 (0.927–1.178)|
|Number of ATC medication classifications in year prior to index date|
|1–2||1.090 (0.919–1.292)||1.060 (0.920–1.221)|
|3–4||1.276 (1.060–1.537)||1.221 (1.042–1.431)|
|≥5||1.409 (1.091–1.820)||1.270 (1.053–1.533)|
|DMT at index date|
|IFNB-1a subcutaneous||1.474 (0.678–3.204)||0.981 (0.813–1.183)|
|IFNB-1a intramuscular||1.142 (0.919–1.418)||0.940 (0.721–1.226)|
|Glatiramer acetate||1.003 (0.863–1.165)||0.929 (0.686–1.259)|
a Cox models simultaneously adjusted for all listed variables, fitted for each individual province, and combined using random effects meta-analysis.
DMT, disease modifying therapy; CI, confidence interval; ATC, anatomical therapeutic chemical; IFNB, beta-interferon.
In this population-based cohort study, optimal adherence (≥80%) was observed in 76% of subjects after one year of therapy. Differences in adherence and persistence were observed between provinces. Those subjects who initiated therapy in recent years were more likely to have suboptimal adherence and to discontinue their DMT within the first 12 months than those who started treatment in earlier years.
Our study did not identify any specific characteristics associated with adherence to the DMTs. While some demographic factors in individual provinces suggested an association with adherence, they were not observed in the combined results. We observed a trend towards lower adherence and persistence with an increased number of medications used in the year before the index date. Previous research both supports and contradicts our findings, as the literature reports conflicting results related to predictors of adherence in MS and other chronic diseases (Ho et al, 2009, Reynolds et al, 2010, Halpern et al, 2011, Lafata et al, 2008, and Wheeler et al, 2013). These findings highlight the challenges in identifying individuals that may be at risk for poor adherence or persistence to the DMTs for MS.
Although nearly 1 in 4 individuals had suboptimal adherence, we observed a higher proportion of optimal adherence than previously reported in the MS literature. A recent review identified 24 studies that examined DMT adherence in MS (Menzin et al., 2013). Many of the identified studies were short-term, had small sample sizes, and relied on self-reported measures of adherence; however, three studies used methodology (administrative data examining incident users of DMTs) and outcomes (proportion of subjects with optimal adherence) similar to ours. All were conducted in the United States and included commercially insured subjects. The first study reported that 60% of their cohort was adherent (≥80%) after one year of follow-up (Tan et al., 2011). Another examined adherence in 224 patients from a multispecialty clinic over a 2-year period and reported adherence (≥80%) was 68% (Lafata et al., 2008). The third study used an adherence threshold of 85% but only examined the beta-interferons, and found the proportion defined as adherent varied between 27% and 41% (Steinberg et al., 2010).
These observed differences may reflect differences in health care systems (Blackwell et al., 2009). Unlike the United States, health care in Canada is publicly funded, and if coverage criteria are met, much of the cost of DMTs are paid for by the provincial governments. While third party health (drug) insurance is common in Canada, if the provincial criteria are not met, it is rare for insurance companies to cover DMTs. As such, our cohort was population-based and included virtually all residents in the 3 provinces, regardless of socioeconomic or insurance status. A recent study from Germany examined DMT adherence in a population-based cohort, in a health care system comparable to that in Canada (Hansen et al., 2015). After the first two years of therapy, 40% of their cohort had adherence ≥80%. While this is lower than our adherence results after one year, it is similar to the 42% optimal adherence we observed over our entire study period. Although socioeconomic status has not been consistently associated with adherence (Menzin et al, 2013 and Alsabbagh et al, 2014), there may be unmeasured differences that exist between those subjects who are privately insured versus those who are publicly insured.
Adherence was also higher than typically observed in other chronic diseases. Studies examining adherence in cardiovascular and psychiatric diseases have consistently found that optimal adherence is around 50%, with many medications being discontinued within the first year of therapy (World Health Organization, 2003, Osterberg and Blaschke, 2005, and Blackburn et al, 2013). MS may be perceived by some as a more “serious” disease or the DMTs as more “serious” drugs, therefore resulting in better adherence. However, when comparing our findings to other neurological conditions such as Parkinson's disease and epilepsy (Kulkarni et al, 2008, Davis et al, 2010, and Liu et al, 2013), or rheumatoid arthritis, which is arguably more similar to MS in terms of disease and drug profiles and costs, we still observed differences. For instance, the proportion of adherent subjects (≥80%) with rheumatoid arthritis to tumor necrosis factor-α inhibitors after the first year of therapy was 63–65% in one American study (Borah et al., 2009), and much lower at 11–43% for a second study (Li et al., 2010). However, none of these studies were population-based, which may, contribute to the observed differences.
Higher rates of adherence for our cohorts might also be due to the active MS patient support programs that are provided by the individual DMT manufacturers, specialty MS clinics, or both. Unlike many other chronic medications, when a patient is prescribed a DMT they typically receive education from a specialist nurse on injection training, dose titration (if applicable), management of potential adverse effects, and are provided ongoing support and follow-up during their time on drug. Indeed, several recent studies suggest that adherence and persistence to DMT is better in individuals participating in patient support programs compared to those who do not (Menzin et al, 2013, Stockl et al, 2010, Tan et al, 2010, and Roche et al, 2014). However, as this is one of the first large population-based studies using administrative data to examine DMT adherence in MS, we are limited in what we can compare our findings to, and further exploration of these differences is warranted.
Less than half of our cohort had adherence ≥80% over the entire study period, and 18% discontinued all DMTs within the first year. However, unlike many other chronic diseases, there is no consensus on how long an individual with MS should be treated with DMTs (Butler et al., 2015). The current DMTs appear to be less effective in secondary progressive MS, and as inflammation becomes less prominent. Indeed, some health care plans require the DMTs to be discontinued once a certain level of progression or disability has been reached. This uncertainty makes it challenging to assess long-term adherence and persistence in MS, and is one of the reasons why we focused on adherence in the first year of therapy. We also chose this time period as it has been shown that the risk for non-adherence and discontinuation is usually highest in the first year of therapy for chronic medications (World Health Organization, 2003 and Osterberg and Blaschke, 2005).
Interestingly, adherence and persistence to the DMTs varied across provinces, despite similar access and coverage criteria. British Columbia and Manitoba both have well-established MS clinics and MS-specialist neurologists who assess almost all MS patients treated with DMTs in the province. In Saskatchewan, most MS patients are managed by community-based neurologists, none of which have a practice dedicated to MS, and the provincial MS clinic is primarily focused on rehabilitative care. This difference in practice focus may account for these variations. For example, the proportion of subjects who discontinued their initial DMT was higher in British Columbia and Manitoba than in Saskatchewan, yet the proportion that discontinued any DMT was similar between all 3 provinces. This suggests that subjects in British Columbia and Manitoba are more likely to switch DMTs within the first year of therapy compared to those in Saskatchewan. MS-specific neurologists may be more aware of current treatment recommendations (Freedman et al., 2013) and have more experience with, or willingness to switch DMTs earlier than neurologists who do not specialize in MS. It is also possible that geographic variation may have contributed to some of the observed differences, as a study from Ontario, Canada found that only 59–63% of individuals remained on their initial DMT after the first year of therapy (Wong et al., 2011).
Our observation of a trend towards increased discontinuations of DMT in the more recent years is likely a reflection of greater choice and drug access. When beta-interferon-1b was approved in Canada in 1995, it was the first time a ‘disease modifying’ drug was available for MS; prior to this MS treatment was predominantly symptomatic with short courses of high dose corticosteroids for relapses. With the novelty of these new drugs, individuals may have been more likely to remain adherent to their DMT (Halpern et al., 2011). Enthusiasm may have waned over time, as more drugs became available, providing opportunity to switch between products. Many individuals are now being treated earlier, when their disease may be less severe than those who were treated when the DMTs first became available. This may result in a perceived lack of need or benefit of drug therapy, which is known to predict poor medication adherence and persistence (Horne et al., 2013). Finally, there has been an increase in the role and availability of MRI in monitoring disease activity and facilitating treatment decisions, and some guidelines incorporate MRI results into recommendations for when a change in DMT should be considered (Freedman et al., 2013).
This study had some limitations. Inherent to all observational studies, we were unable to control for all potential confounders. With administrative data we lacked MS-specific clinical information that may have influenced adherence and persistence, such as disease severity. While we accounted for DMT discontinuations related to pregnancy, we were unable to determine whether suboptimal adherence or discontinuations were due to monitoring recommendations for withholding doses for temporary issues like increased liver enzymes or hematologic abnormalities. Finally, in the absence of a gold standard for estimating comorbidity burden (Quail et al., 2011), we included prior health care and medication utilization as proxy measures, which predict important outcomes such as hospitalizations and mortality (Quail et al, 2011 and Schneeweiss et al, 2001).
This is one of the first large population-based studies to examine adherence and persistence to the DMTs in MS, and the first to do so in a North American population. We observed adherence to the DMTs that is higher than what has been reported for other chronic diseases, as well as for other non-population-based MS cohorts. Medication adherence is complex and multifactorial, and in the absence of a consensus on appropriate DMT treatment duration, it is difficult to conclude whether our observed levels were adequate. However, closer examination as to why adherence appears to be better in MS, and how adherence influences disease outcomes will not only help increase our understanding of the disease, but may also prove useful in advancing the management of other chronic diseases.
The corresponding author (CE) and the analysts (FZ, SL, XL) had access to the data in the study and take responsibility for the integrity of the data and accuracy of the data analyses. CE designed the study and CE, HT, RAM and YZ obtained funding. CE drafted the manuscript. All authors were involved with the interpretation of the data, critically revising the manuscript, and have approved the final version to be published.
Supported by the National Multiple Sclerosis Society (RG-4757-A-3). The study sponsors had no role in the study design, data collection, data analysis, interpretation of results, writing of this manuscript, or decision to submit. All inferences, opinions, and conclusions drawn in this publication are those of the authors and do not reflect the opinions or policies of the Data Stewards. No official endorsement by Manitoba Health, Population Data BC, Pharmanet, BC Ministry of Health, BC Data Stewardship Committee, the Saskatchewan Government or the Saskatchewan Ministry of Health is intended or should be inferred.
This study was funded by the National Multiple Sclerosis Society (RG-4757-A-3). Dr Evans declares no conflicts. Dr Marrie receives research funding from: Canadian Institutes of Health Research, Research Manitoba, Multiple Sclerosis Society of Canada, Multiple Sclerosis Scientific Foundation, National Multiple Sclerosis Society, Rx & D Health Research Foundation, and has conducted clinical trials funded by Sanofi-Aventis. Mr Zhu declares no conflicts. Ms Leung declares no conflicts. Dr Lu declares no conflicts. Mr Melesse declares no conflicts. Dr Kingwell declares no conflicts. Dr Zhao declares no conflicts. Dr Tremlett has been funded by the Multiple Sclerosis Society of Canada (Don Paty Career Development Award); is a Michael Smith Foundation for Health Research Scholar and the Canada Research Chair for Neuroepidemiology and Multiple Sclerosis. She has received: research support from the National Multiple Sclerosis Society, the Canadian Institutes of Health Research, and the UK MS Trust; speaker honoraria and/or travel expenses to attend conferences from the Consortium of MS Centres (2013), the National MS Society (2012, 2014), Bayer Pharmaceuticals (2010), Teva Pharmaceuticals (2011), ECTRIMS (2011, 2012, 2013, 2014), UK MS Trust (2011), the Chesapeake Health Education Program, US Veterans Affairs (2012), Novartis Canada (2012), Biogen Idec (2014), American Academy of Neurology (2013, 2014, 2015). Unless otherwise stated, all speaker honoraria are either donated to an MS charity or to an unrestricted grant for use by her research group. All other authors have nothing to declare. No other relationships or activities have influenced the submitted work.
We thank the BC Ministry of Health, BC Vital Statistics Agency and BC PharmaNet for approval and support with accessing provincial data; and Population Data BC for facilitating approval and use of the data. We also thank Dr John Petkau (Department of Statistics, University of British Columbia) for his review of the manuscript.
- Alsabbagh et al., 2014 M.W. Alsabbagh, M. Lemstra, D. Eurich, et al. Socioeconomic status and nonadherence to antihypertensive drugs: a systematic review and meta-analysis. Value Health. 2014;17:288-296 Crossref
- BC Ministry of Health [creator], 2012 BC Ministry of Health [creator], 2012. PharmaNet. V2. BC Ministry of Health [publisher]. (Data Extract. Data Stewardship Committee) (2012). 〈http://www.popdata.bc.ca/data〉.
- Blackburn et al., 2013 D. Blackburn, J. Swidrovich, M. Lemstra. Non-adherence in type 2 Diabetes: Practical Considerations for interpreting the Literature. Patient Preference Adherence. 2013;7:183-189 Crossref
- Blackwell et al., 2009 D. Blackwell, M. Martinez, J. Gentleman, C. Sanmartin, J.-M. Berthelot. Socioeconomic status and utilization of health care services in Canada and the United States - findings from a binational health survey. Med Care. 2009;47:1136-1146 Crossref
- Borah et al., 2009 B. Borah, X. Huang, V. Zarotsky, D. Globe. Trends in RA patients' adherence to subcutaneous anti-TNF therapies and costs. Curr. Med Res. Opin.. 2009;25:1365-1377 Crossref
- British Columbia Ministry of Health [creator], 2012 British Columbia Ministry of Health [creator], 2012. Consolidation File (MSP Registration & Premium Billing). V2. Population Data BC [publisher]. (Data Extract. MOH) (2012). 〈http://www.popdata.bc.ca/data〉.
- British Columbia Ministry of Health [creator], 20122012 British Columbia Ministry of Health [creator], 2012. Medical Services (MSP) Payment Information File. V2. Population Data BC [publisher]. (Data Extract. MOH) (2012). 〈http://www.popdata.bc.ca/data〉.
- Butler et al., 2015 M. Butler, M. Forte, N. Schwehr, A. Carpenter, R. Kane. Decisional dilemmas in discontinuing prolonged disease-modifying treatment for multiple sclerosis. Minn. Evid.-Based Pract. Cent.: Agency Healthc. Res. Qual.. 2015;
- Canadian Institute for Health Information [creator], 20122012 Canadian Institute for Health Information [creator], 2012. Discharge Abstract Database (Hospital Separations). V2. Population Data BC [publisher]. (Data Extract. MOH) (2012). 〈http://www.popdata.bc.ca/data〉.
- Davis et al., 2010 K. Davis, H. Edin, J. Allen. Prevalence and cost of medication nonadherence in Parkinson's disease: evidence from administrative claims data. Mov. Disord.. 2010;25:474-480 Crossref
- Dormuth et al., 2009 C. Dormuth, A. Patrick, W. Shrank, et al. Statin adherence and risk of accidents - a cautionary tale. Circulation. 2009;119:2051-2057 Crossref
- Ediger et al., 2009 J. Ediger, J. Walker, L. Graff, et al. Predictors of medication adherence in inflammatory bowel disease. Am. J. Gastroenterol.. 2009;102:1414-1426
- Evans et al., 2012 C. Evans, D. Eurich, A. Remillard, Y. Shevchuk, D. Blackburn. First-fill medication discontinuations and non-adherence to antihypertensive therapy: an observational study. Am. J. Hypertens.. 2012;25:195-203 Crossref
- Filippini et al., 2003 G. Filippini, L. Munari, B. Incorvaia, et al. Interferons in relapsing remitting multiple sclerosis: a systematic review. Lancet. 2003;361:545-552 Crossref
- Freedman et al., 2013 M. Freedman, D. Selchen, D. Arnold, et al. Treatment optimization in MS: Canadian MS Working Group updated recommendations. Can. J. Neurol. Sci.. 2013;40:307-323 Crossref
- Halpern et al., 2011 R. Halpern, S. Agarwal, C. Dembek, L. Borton, M. Lopez-Bresnahan. Comparison of adherence and persistence among multiple sclerosis patients treated with disease-modifying therapies: a retrospective administrative claims analysis. Patient Preference Adherence. 2011;5:73-84 Crossref
- Hansen et al., 2015 K. Hansen, K. Schussel, M. Kieble, et al. Adherence to disease modifying drugs among patients with multiple sclerosis in Germany: a Retrospective Cohort Study. PLoS One. 2015; 10.1371/journal.pone.0133279
- Higgins et al., 2003 J. Higgins, S. Thompson, J. Deeks, D. Altman. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557-560 Crossref
- Ho et al., 2009 P.M. Ho, C. Bryson, J. Rumsfeld. Medication adherence - its importance in cardiovascular outcomes. Circulation. 2009;119:3028-3035 Crossref
- Horne et al., 2013 R. Horne, S. Chapman, R. Parham, N. Freemantle, A. Forbes, V. Cooper. Understanding patients' adherence-related beliefs about medicines prescribed for long-term conditions: a meta-analytic review of necessity-concerns framework. PLoS One. 2013;8:e80633 Crossref
- Jacobs et al., 1996 L. Jacobs, D. Cookfair, R. Rudick. Intramuscular interferon beta-1a for disease progression in relapsing multiple sclerosis. The Multiple Sclerosis Collaborative Research Group (MSCRG). Ann. Neurol.. 1996;39:285-294 Crossref
- Karve et al., 2009 S. Karve, M. Cleves, M. Helm, T. Hudson, D. West, B. Martin. Good and poor adherence: optimal cut-point for adherence measures using administrative claims data. Curr. Med Res. Opin.. 2009;25:2303-2310 Crossref
- Kulkarni et al., 2008 A. Kulkarni, R. Balkrishnan, R. ANderson, H. Edin, J. Kirsch, M. Stacy. Medication adherence and associated outcomes in medicare health maintenance organization-enrolled older adults with Parkinson's disease. Mov. Disord.. 2008;23:359-365 Crossref
- Lafata et al., 2008 J. Lafata, M. Cerghet, E. Dobie, et al. Measuring adherence and persistence to disease-modifying agents among patients with relapsing remitting multiple sclerosis. J. Am. Pharm. Assoc.. 2008;48:752-757
- Li et al., 2010 P. Li, M. Blum, J. Von Feldt, S. Hennessy, J. Doshi. Adherence, discontinuation, and switching of biologic therapies in medicaid enrollees with Rheumatoid Arthritis. Value Health. 2010;13:805-812 Crossref
- Liu et al., 2013 J. Liu, Z. Liu, H. Ding, X. Yang. Adherence to treatment and influencing factors in a sample of Chinese epilepsy patients. Epileptic Disord.. 2013;15:289-294
- Martin et al., 2009 B. Martin, E. Wiley-Exley, S. Richards, M. Domino, T. Carey, B.L. Sleath. Contrasting measures of adherence with simple drug use, medication switching, and therapeutic duplication. Ann. Pharmacother.. 2009;43:36-44 Crossref
- Menzin et al., 2013 J. Menzin, C. Caon, C. Nichols, L.A. White, M. Friedman, M. Pill. Narrative review of the literature on adherence to disease-modifying therapies among patients with Multiple Sclerosis. J. Manag. Care Pharm.. 2013;19(1-a):S24-S40
- Osterberg and Blaschke, 2005 L. Osterberg, T. Blaschke. Adherence to medication. New Engl. J. Med. 2005;353:487-497 Crossref
- PRISMS, 1998 PRISMS. (Prevention of Relapses and Disability by Interferon beta-1a Subcutaneously in Multiple Sclerosis) Study Group. Randomised double-blind placebo-controlled study of interferon beta-1a in relapsing-remitting multiple sclerosis. Lancet. 1998;352:1498-1504
- Quail et al., 2011 J. Quail, L. Lix, B.A. Osman, G. Teare. Comparing comorbidity measures for predicting mortality and hospitalization in three population-based cohorts. BMC Health Serv. Res.. 2011;11:146 Crossref
- Reynolds et al., 2010 M. Reynolds, R. Stephen, C. Seaman, K. Rajagopalan. Persistence and adherence to disease modifying drugs among patients with multiple sclerosis. Curr. Med Res. Opin.. 2010;26:663-674 Crossref
- Roche et al., 2014 J. Roche, Y. MCCarry, K. Mellors. Enhanced patient support services improve patient persistence with multiple sclerosis treatment. Patient Preference Adherence. 2014;8:805-811 Crossref
- Schneeweiss et al., 2001 S. Schneeweiss, J. Seeger, M. Maclure, P. Wang, J. Avorn, R. Glynn. Performance of comorbidity scores to control for confounding in epidemiological studies using claims data. Am. J. Epidemiol.. 2001;154:854-864 Crossref
- Simpson et al., 2006 S. Simpson, D. Eurich, S. Majumdar, et al. A meta-analysis of the association between adherence to drug therapy and mortality. BMJ. 2006; 10.1136/bmj.38875.675486.55
- Steinberg et al., 2010 S. Steinberg, R. Faris, C. Chang, A. Chan, M. Tankersley. Impact of adherence to interferons in the treatment of multiple sclerosis - a non-experimental, retrospective, cohort study. Clin. Drug Invest.. 2010;30:89-100 Crossref
- Stockl et al., 2010 K. Stockl, J. Shing, S. Gong, A. Harada, B. Solow, H. Lew. Improving patient self-managment in multiple sclerosis through a disease therapy management program. Am. J. Manag Care. 2010;16:139-144
- Tan et al., 2010 H. Tan, J. Yu, D. Tabby, A. Devries, J. Singer. Clinical and economic impact of a specialty care management program among patients with multiple sclerosis: a cohort study. Mult. Scler.. 2010;16:956-963 Crossref
- Tan et al., 2011 H. Tan, Q. Cai, S. Agarwal, J. Stephenson, S. Kamat. Impact of adherence to disease-modifying therapies on clinical and economic outcomes among patients with multiple sclerosis. Adv. Ther.. 2011;28:51-61 Crossref
- The IFNB Multiple Sclerosis Study Group, 1993 The IFNB Multiple Sclerosis Study Group. Interferon beta-1b) is effective in relapsing-remitting multiple sclerosis. I. Clinical results of a multicenter, randomized, double-blind, placebo-controlled trial. Neurology. 1993;43:655-661
- Wheeler et al., 2013 K. Wheeler, M.E. Roberts, M. Neiheisel. Medication adherence part two: predictors of nonadherence and adherence. J. Am. Assoc. Nurse Pract.. 2013;26:225-232
- Wong et al., 2011 J. Wong, T. Gomes, M. Mamdani, M. Manno, P. O’Connor. Adherence to multiple sclerosis disease-modifying therapies in Ontario is low. Can. J. Neurol. Sci.. 2011;38:429-433 Crossref
- World Health Organization, 2003 World Health Organization, 2003. Adherence to long-term therapies - evidence for action. 2003; 〈http://www.who.int/chp/knowledge/publications/adherence_introduction.pdf〉. Accessed September 1, 2009.
a College of Pharmacy & Nutrition, University of Saskatchewan, 104 Clinic Place, Saskatoon, SK S7N 5E5, Canada
b Departments of Internal Medicine and Community Health Sciences, University of Manitoba, Health Sciences Centre, GF 543-820 Sherbrook Street, Winnipeg, MB, Canada R3A 1R9
c Department of Medicine (Neurology), University of British Columbia, UBC Hospital, 2211 Wesbrook Mall, Vancouver, BC, Canada V6T 2B5
d Department of Community Health Sciences, University of Manitoba, Health Sciences Centre, GF 543-820 Sherbrook Street, Winnipeg, MB, Canada R3A 1R9
e Saskatchewan Health Quality Council, 241-111 Research Drive, Saskatoon, SK, Canada S7N 3R2
∗ Corresponding author.
© 2016 Published by Elsevier B.V.