Measuring Adherence and Outcomes in the Treatment of Patients With Multiple Sclerosis
-
Jing Hao
Abstract
Context
Both adherence and outcomes are more difficult to measure in patients with multiple sclerosis (MS) than in patients with diseases such as hypertension, for which most medications are taken orally and surrogate outcomes (eg, blood pressure) are routinely collected.
Objectives
To characterize the adherence and outcomes of patients with MS within a large integrated health system and to assess the relationship between adherence and outcomes.
Study Design
Retrospective review of adherence and health care utilization outcomes via electronic health records and claims (2004-2013) combined with a prospective survey regarding adherence and functional outcomes (2012-2013).
Methods
Retrospectively, medication possession ratios were calculated, and adherence groups were compared regarding health care utilization and costs. Prospectively, patients were recruited to complete questionnaires to measure self-reported adherence (SRA) and MS-specific outcomes, including the Multiple Sclerosis Impact Scale (MSIS), the Kurtzke Expanded Disability Status Scale (EDSS), and the Treatment Satisfaction Questionnaire for Medication (TSQM). Regression was used to statistically test for differences in these outcomes among adherence groups.
Results
A total of 681 patients were studied. Most patients (307 of 375 [82%] in the retrospective cohort and 244 of 306 [89%] in the prospective cohort) had greater that 80% adherence to their MS medications. Mean inpatient days per patient for MS-related admissions was highest for high-adherence than for intermediate and low-adherence patients (0.52 vs 0.23 and 0.34, respectively; P<.001), but no other associations between adherence and health care utilization were found. Mean outpatient costs and total costs were lowest for the low-adherence group, suggesting that higher adherence may not guarantee cost savings overall. Patients with intermediate and high self-reported adherence showed significantly better mean scores than patients with low adherence on several MS outcomes, including EDSS (4.1 and 4.2 vs 4.8, P<.05), MSIS physical function (33 and 35 vs 41, P<.05), and TSQM Global Satisfaction (75 and 78 vs 70, P<.05).
Conclusions
The findings of this study indicate that patients with MS are mostly adherent to their existing treatments. Patients with greater medication adherence may have increased cost, but their physical outcomes are better. This finding may shed light on other chronic disease entities and how we view the success of treatments.
Multiple sclerosis (MS) is a chronic disease of the central nervous system characterized by inflammation, demyelination, and axonal degeneration.1 Multiple sclerosis is the most common cause of neurologic disability in young adults, with a mean age of onset of 33 years, and it is most prevalent among whites.2 The prevalence is estimated to be 135 per 100,000 in the United States, with approximately 12,000 newly diagnosed cases of MS reported annually.2,3 Twelve disease-modifying agents are approved by the US Food and Drug Administration.4 Many factors may affect adherence to MS medications, including fear of needles, frequency of administration, perceived lack of efficacy, mental illness or depressed mood, and adverse effects, such as injection-site reactions, infections, and flulike symptoms.5-7 Additionally, patients may encounter financial and logistical challenges to obtaining their MS medications, which vary based on a patient's health insurance provider, medical history, disease severity, and specific medication; these challenges affect medication adherence because adherence requires access.
The impact of adherence on outcomes, however, is more challenging to study for patients with MS than for patients with other conditions, such as diabetes or hypertension, which have been reported on extensively.8-10 First, unlike diabetes or hypertension, for which surrogate outcome measures such as glycated hemoglobin or blood pressure can be routinely measured and captured in electronic health records (EHRs), MS has no easily obtainable and comparable surrogate measures. Second, adverse events may take the form of relapses or new lesions identified in imaging studies, which are not routinely captured in a quantitative fashion. Third, some treatments for patients with MS are injections, as opposed to an oral medication for which both the prescription orders and pharmacy claims indicate clearly the number of days’ supply of medication being provided.
Because of these challenges, there are few literature reports related to treatment adherence in patients with MS.11 Several studies have assessed adherence using simple measures of patient self-reporting continuing, discontinuing, or switching a medication at the time of assessment,12-14 but such measures provide little insight into the way in which patients adhere to their prescribed medication regimens before a discontinuation or switch. Turner et al11 reported the results of a study in which medication adherence was assessed monthly in a small cohort of Veterans’ Administration patients using a single self-reported measure of the number of missed doses during the previous month. Although self-reported measures have limitations, self-reported adherence may, in the absence of robust claims or EHR data, be the optimal method for understanding MS treatment-related adherence on a day-to-day and week-to-week basis.
The purpose of this study was to use 2 approaches to characterize treatment adherence and outcomes of patients with MS within a large integrated health system and to assess the relationship between adherence and outcomes. We performed a retrospective examination of health care utilization and costs stratified by adherence level as measured from claims data. We also conducted a prospective survey-based assessment of self-reported adherence and MS-specific outcomes.
Methods
Study Population
The study was conducted at Geisinger Health System, an integrated delivery system primarily in central and northeastern Pennsylvania. Geisinger Health System includes Geisinger Clinic and an insurance provider (Geisinger Health Plan)—a hospital-insurance network with more than 40 primary care clinics, a large tertiary care teaching hospital, and 11 other hospitals. The initial study population was defined as all Geisinger Clinic patients identified in the EHR who were aged 18 years or older when they met either of the following criteria between January 1, 2004, and December 31, 2013: (1) an International Classification of Diseases, 9th Revision (ICD-9) diagnosis code for MS entered on their problem list and an order for an MS medication; or (2) had at least 2 separate encounters coded as MS and at least 1 medication order indicated for MS.
We examined medication adherence and outcomes in 2 samples of this population, using both retrospective and prospective approaches.
Retrospective Analysis
We used the subset of the MS population described above who were also covered by the Geisinger Health Plan for the entire period (approximately 30% of the population) to describe medication adherence, health care utilization, and costs as contained in the EHR and claims records. Patients with fewer than 12 months of health plan enrollment or who died during the follow-up period were excluded. We removed 5 patients who died during the follow-up period to have equal follow-up time (12 months) for all patients. We recognized that the high cost of end-of-life care may affect outcomes, but we reran cost analyses including the deceased patients’ data, and our results and conclusions did not change meaningfully (results not shown).
Data extracted from the EHRs included demographic variables (eg, age, sex, race); body mass index; and all encounters, diagnoses, and procedures. Data extracted from insurance claims included all diagnosis and procedure codes, prescriptions filled, total allowed amount paid for each claim, and location of services.
Prospective Analysis
We recruited a random sample of the MS population described above who were actively being seen by a neurologist and taking MS medications stratified by last known active medication. Medication adherence and outcomes were assessed in this patient group using a survey instrument. A self-reported adherence (SRA) measure was calculated as the patient's self-report of the number of doses taken divided by the number of doses prescribed in the previous month. Patients were then assigned into 1 of 3 SRA categories consistent with the medication possession ratio (MPR) categories in the retrospective analysis: low, <80%; intermediate, 80%-99%; or high, 100%. Surveys consisted of a combination of published, validated, newly developed questionnaires wherein patients reported their doses taken in the past month, number of relapses, Medical Outcome Study (MOS) Cognitive Function Score,15,16 Kurtzke Expanded Disability Status Scale (EDSS),17 Treatment Satisfaction Questionnaire for Medication (TSQM) satisfaction score,18 and Multiple Sclerosis Impact Scale (MSIS) physical and psychological functional score.19 Number of relapses and all self-reported outcome scores (eg, MOS, TSQM) were then compared among SRA groups using nonparametric bootstrapped 95% CIs based on 1000 replications and resampling at the patient (as opposed to score) level to account for repeated measures. The patients who consented were surveyed 3 times, 6 months apart (baseline, 6-month, and 12-month follow-up) and were also surveyed after any non–routine office visits to the neurology department. Baseline surveys were completed in mid-2012, with follow-up surveys completed in 2012 and 2013. Follow-up versions of the questionnaire (6-month, 12-month, and after office visits) were similar to the baseline version, but follow-up surveys did not ask the patients about the time since initial diagnosis, and after-visit follow-up surveys added a question specifically about the reason for the patient's most recent MS appointment. This study received approval from the Geisinger Institutional Review Board.
Analytic Approach
In the retrospective analysis, each patient's records were examined for 12 months before their most recent medical encounter (eg, office visit), and MS medication possession ratio (MPR)20 was calculated to assess medication adherence during this period. Our primary interest was adherence to the 6 major MS medications (interferon β-1a, interferon β-1b, high-dose interferon β-1a, glatiramer acetate, natalizumab, or fingolimod). Medication possession ratio can only be calculated when patients have a prescription filled at least 2 times for the medications of interest, so patients with fewer than 2 of these medications filled at least 2 times were excluded from further analysis. Patients were assigned to 1 of 3 adherence categories based on MPR: low (<80%), intermediate (80%-99%), and high (≥100%) adherence.
The primary outcomes of interest were inpatient admissions, emergency department (ED) visits, outpatient visits, and health care costs during the study follow-up period. The proportions of patients with inpatient admissions, outpatient visits, and ED visits were calculated both for all-cause utilization and MS-related health care utilization only, estimated using ICD-9 diagnosis codes associated with the encounter. Encounters related to MS were defined as those for which the primary ICD-9 diagnosis code associated with the claim was for multiple sclerosis (ICD-9 340). Health care costs were defined as the total cost of care (health insurance payment to health care providers plus patients’ copayments), and these costs were calculated per member per month. Per-member-per-month costs were stratified by inpatient, outpatient, ED, and prescription medication costs. All-cause vs MS-related costs were calculated separately, with all costs adjusted for inflation to 2013 US dollars.
To compare baseline patient characteristics among the 3 adherence groups, we used χ2 tests (for categorical variables, such as sex), Fisher exact tests (for categorical variables with rare occurrences, such as heart failure), or analysis of variance (for continuous variables assumed to be normal, such as age). To compare inpatient and outpatient health care utilization, we used generalized linear models with logistic (for binary outcomes) or Poisson (for discrete outcomes, such as inpatient days) regression. Because medical costs are not normally distributed, we used generalized linear models with a log-γ distribution to compare costs among adherence groups. We assumed that 40% of the retrospective cohort would have low adherence (MPR <80%), and our study was adequately powered to detect a ±12% difference in percentages of patients with ED visits between the low- and high-adherence groups with 72% power, or a ±13% difference with 81% power, at the P<.05 significance level. All analyses were performed using SAS version 9.4 software (SAS Institute Inc), with differences of P<.05 considered statistically significant.
Results
We identified an initial population of 2957 patients with any MS diagnosis in the EHR, 1046 of whom were also insurance plan members. After applying the remaining inclusion/exclusion criteria, 375 patients met all inclusion criteria for the retrospective analysis. For the prospective study, we contacted 643 patients to participate, with a response rate of 48% (n=306). These patients completed a total of 971 surveys.
Table 1 displays the baseline characteristics of the patients in the retrospective and prospective cohorts (N=681), stratified by adherence categories. Overall, patients in the retrospective cohort were predominantly female (283 [75%]) and white (371 [99%]), with 149 patients (40%) having a body mass index of 30 or greater. Depression (117 [31%]) and hypertension (114 [30%]) were the most common comorbidities. Patients in the prospective cohort had similar attributes: 259 (85%) were female, 302 (99%) were white, and 127 (42%) had a body mass index of 30 or greater. Depression (125 [41%]) and hypertension (62 [20%]) were common comorbidities.
Baseline Characteristics of Patients Stratified by Adherence to Multiple Sclerosis Treatment (N=681)
| Retrospective Cohort | Prospective Cohort | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic, No. (%)a | All | MPR <80% | MPR 80%-99% | MPR 100% | P Value | All | SRA <80% | SRA 80%-99% | SRA 100% | P Value |
| Total | 375 (100) | 68 (18) | 192 (51) | 115 (31) | 306 (100) | 62 (20) | 206 (67) | 38 (12) | ||
| Age, y, mean | 47 | 49 | 47 | 47 | .74b | 50 | 47 | 50 | 52 | .19b |
| Male | 92 (25) | 17 (25) | 42 (22) | 33 (39) | .40c | 47 (15) | 13 (21) | 27 (13) | 7 (18) | .28c |
| White | 371 (99) | 67 (99) | 189 (98) | 115 (100) | .29d | 302 (99) | 61 (98) | 203 (99) | 38 (100) | .89d |
| BMI | .69c | .73c | ||||||||
| <25 | 114 (30) | 18 (28) | 57 (31) | 39 (35) | 88 (29) | 20 (32) | 60 (29) | 8 (21) | ||
| 25-29.9 | 99 (26) | 22 (34) | 49 (27) | 28 (25) | 90 (29) | 21 (34) | 58 (28) | 11 (28) | ||
| ≥30 | 149 (40) | 25 (39) | 79 (43) | 45 (40) | 127 (42) | 22 (36) | 85 (42) | 20 (51) | ||
| Comorbidity | ||||||||||
| Asthma | 54 (14) | 13 (19) | 28 (15) | 13 (11) | .30c | 22 (7) | 6 (10) | 14 (7) | 2 (5) | .66c |
| Diabetes | 38 (10) | 7 (10) | 21 (11) | 10 (9) | .79c | 18 (6) | 4 (7) | 11 (5) | 3 (8) | .81c |
| Congestive heart failure | 2 (1) | 0 | 1 (<1) | 1 (<1) | .99d | 1 (<1) | 0 | 1 (<1) | 0 | .99d |
| COPD | 14 (4) | 5 (7) | 3 (2) | 6 (5) | .09d | 5 (2) | 0 | 2 (1) | 3 (8) | .02d |
| Chronic kidney disease | 12 (3) | 2 (3) | 3 (2) | 7 (6) | .10d | 3 (<1) | 1 (2) | 0 | 2 (5) | .02d |
| Hypertension | 114 (30) | 26 (38) | 49 (26) | 39 (34) | .08c | 62 (20) | 11 (18) | 38 (18) | 13 (34) | .07c |
| Depression | 117 (31) | 26 (38) | 65 (34) | 26 (23) | .04c | 125 (41) | 30 (48) | 79 (38) | 16 (42) | .37c |
| Cancer, any | 19 (5) | 2 (3) | 10 (5) | 7 (6) | .42c | 9 (3) | 2 (3) | 5 (2) | 2 (5) | .45c |
a Data reported as No. (%) unless otherwise noted.
b Analysis of variance.
c χ2 test.
d Fisher exact test.
Abbreviations: BMI, body mass index; COPD, chronic obstructive pulmonary disease; MPR, medication possession ratio; SRA, self-reported adherence.
Both cohorts were extremely compliant with their MS medications, with 307 of 375 retrospective patients (82%) having an MPR adherence measure of 80% or greater, and 244 of 306 prospective patients (80%) having an SRA of 80% or above. In the retrospective cohort, patients with low adherence (MPR <80%) were statistically significantly more likely to have depression (38% vs 26% and 34% in the intermediate- and high-adherence groups), though a similar trend in the prospective cohort was not statistically significant (48% vs 38% and 42%). In the prospective cohort, significantly more patients in the high-adherence group (SRA 100%) had had chronic obstructive pulmonary disease (8%) and chronic kidney disease (5%) (P=.02).
Table 2 presents the health care utilization of the patients in the retrospective cohort. Patients with high adherence were the least likely to have any all-cause ED visits (22% vs 25% and 31% for the other adherence groups), but the differences among the 3 groups were not statistically significant, nor were any of the other comparisons for all-cause inpatient and ED utilization. Differences in the mean number of inpatient days per patient for MS-related admissions were significantly different among the 3 groups (P<.001), and a follow-up pairwise test between the high-adherence group and intermediate-adherence group showed a significant difference, even after adjusting for multiple comparisons (0.52 vs 0.23; P<.001). No other MS-related health care utilization measures showed a statistically significant relationship with adherence.
Heath Care Utilization by Patients With MS Stratified by MPR Adherence Categorya (n=375)
| Heath Care Utilization | MPR <80% (n=68) | MPR 80%-99% (n=192) | MPR 100% (n=115) | P Value |
|---|---|---|---|---|
| All-Cause | ||||
| Patients with any ED visits, No. (%) | 21 (31) | 47 (25) | 25 (22) | .55a |
| Patients with any inpatient admissions, No. (%) | 8 (12) | 22 (12) | 14 (12) | .97a |
| Patient with any outpatient visits, No. (%) | 68 (100) | 190 (99) | 115 (100) | .90a |
| Inpatient d per patient, mean (SD) | 1.28 (4.16) | 1.16 (5.14) | 1.20 (5.24) | .74b |
| No. of ED visits per patient, mean (SD) | 0.50 (0.91) | 0.45 (0.94) | 0.47 (1.59) | .89b |
| MS-Related | ||||
| Patients with any ED visits, No. (%) | 4 (6) | 9 (5) | 3 (3) | .55a |
| Patients with any inpatient admissions, No. (%) | 6 (9) | 8 (4) | 5 (4) | .33a |
| Patients with any outpatient visits, No. (%) | 56 (82) | 161 (84) | 98 (85) | .89a |
| Inpatient d per patient, mean (SD) | 0.34 (2.11) | 0.23 (2.00) | 0.52 (3.44) | <.001b |
| No. of ED visits per patient, mean (SD) | 0.12 (0.53) | 0.10 (0.54) | 0.17 (1.45) | .26b |
a Logistic regression.
b Poisson regression.
Abbreviations: ED, emergency department; MPR, medication possession ratio; MS, multiple sclerosis.
Table 3 presents the health care costs of patients in the retrospective study, both all-cause and MS-related. The mean total health care costs for patients in the low-adherence group was significantly lower than for patients in the intermediate- or high-adherence groups. Although the patients in the high-adherence group had lower inpatient costs on average, those costs seem to have been offset by higher outpatient and pharmacy-related costs. Whether examining all-cause or MS-related health care utilization, the group with intermediate adherence consistently had the highest mean outpatient, pharmacy, and total health care costs per person. Consistent with the health care utilization results in Table 2, costs related to ED or inpatient care did not show statistically significant trends with adherence.
Mean Heath Care Utilization Costs Incurred by Patients With MS Stratified by MPR Adherence Categorya (N=375)
| Heath Care Utilization | MPR <80%a (n=68) | MPR 80%-99%a (n=192) | MPR 100%a (n=115) |
|---|---|---|---|
| All-Cause PMPM Costs (USD) | |||
| ED | |||
| Mean (median) | 23 (0) | 20 (0) | 87 (0) |
| Maximum | 415 | 428 | 7315 |
| Inpatient | |||
| Mean (median) | 168 (0) | 152 (0) | 72 (0) |
| Maximum | 2814 | 12,325 | 1971 |
| Outpatient | |||
| Mean (median) | 256 (189) | 536 (137)a | 352 (194) |
| Maximum | 1823 | 30,948 | 8420 |
| Pharmacy | |||
| Mean (median) | 1633 (1660) | 2430 (2723) | 1973 (1118) |
| Maximum | 4548 | 6018 | 6332 |
| Total | |||
| Mean (median) | 2081 (2081) | 3142 (3176)b | 2490 (1637)c |
| Maximum | 6846 | 31,022 | 13,720 |
| MS-Related PMPM Costs (USD) | |||
| ED | |||
| Mean (median) | 2.5 (0) | 1.8 (0) | 11.5 (0) |
| Maximum | 79 | 86 | 1265 |
| Inpatient | |||
| Mean (median) | 40 (0) | 15 (0) | 5 (0) |
| Maximum | 2025 | 1130 | 545 |
| Outpatient | |||
| Mean (median) | 70 (27) | 263 (24)b | 167 (27)d |
| Maximum | 660 | 30,808 | 8259 |
| Pharmacy | |||
| Mean (median) | 1405 (1237) | 2235 (2323) | 1790 (966) |
| Maximum | 4040 | 4975 | 5057 |
| Total | |||
| Mean (median) | 1518 (1354) | 2516 (2553)b | 1976 (1092)c |
| Maximum | 4086 | 30,808 | 8259 |
a P<.05
b P<.01
c P<.001
d P<.0001
Abbreviations: ED, emergency department; MPR, medication possession ratio; MS, multiple sclerosis; PMPM, per-member-per month; USD, US dollar.
Table 4 compares the self-reported MS outcomes and related measures among the 3 SRA groups. Patients with intermediate and high adherence showed statistically significantly better scores on EDSS and MSIS physical scales than patients with low adherence. Patients in the high-adherence group also had statistically significantly better mean MSIS psychological and TSQM Global Satisfaction scores than patients in the low-adherence group. Although the other TSQM scores did show improvements in mean with increased adherence, the trends were not statistically significant. No statistically significant relationships were seen between adherence and the number of relapses or MOS cognitive function.
Survey Results of Patients With MS Compared Across Adherence Categoriesa (n=306)
| SRA Measuresa | SRA <80% (n=250) | SRA 80%-99% (n=274) | SRA 100% (n=447) |
|---|---|---|---|
| No. of Relapses | |||
| Mean (95% CI) | 1.1 (0.6-1.5) | 1.1 (0.6-1.4) | 1.2 (0.7-1.5) |
| Median (IQR) | 0 (0-2) | 0 (0-1) | 0 (0-2) |
| Range | (0-31) | (0-30) | (0-35) |
| MOS Cognitive Functioning | |||
| Mean (95% CI) | 46 (44-48) | 47 (46-48) | 47 (46-48) |
| Median (IQR) | 48 (37-57) | 50 (40-57) | 48 (40-57) |
| Range | (20-59) | (18-59) | (20-59) |
| EDSS | |||
| Mean (95% CI) | 4.8 (4.6-5.1) | 4.1 (3.8-4.3)b | 4.2 (4.0-4.5)b |
| Median (IQR) | 5.5 (3.5-6.0) | 4.0 (2.0-5.5) | 4.5 (2.0-6.0) |
| Range | (0-8.5) | (0-9) | (0-8.5) |
| MSIS Physical | |||
| Mean (95% CI) | 41 (39-46) | 33 (30-36)b | 35 (33-39)b |
| Median (IQR) | 39 (18-64) | 28 (10-55) | 35 (14-54) |
| Range | (0-100) | (0-96) | (0-95) |
| MSIS Psychological | |||
| Mean (95% CI) | 40 (39-45) | 36 (34-39) | 33 (31-36)b |
| Median (IQR) | 33 (14-63) | 31 (14-50) | 28 (14-47) |
| Range | (0-100) | (0-100) | (0-100) |
| TSQM Effectiveness | |||
| Mean (95% CI) | 67 (61-70) | 70 (66-74) | 72 (69-74) |
| Median (IQR) | 67 (50-83) | 67 (50-92) | 75 (58-92) |
| Range | (0-100) | (0-100) | (0-100) |
| TSQM Side Effects | |||
| Mean (95% CI) | 77 (71-80) | 82 (79-85) | 82 (79-84) |
| Median (IQR) | 83 (67-100) | 92 (75-100) | 92 (67-100) |
| Range | (0-100) | (0-100) | (17-100) |
| TSQM Convenience | |||
| Mean (95% CI) | 70 (66-73) | 67 (65-70) | 74 (71-75) |
| Median (IQR) | 67 (56-89) | 67 (56-83) | 72 (61-89) |
| Range | (0-100) | (11-100) | (0-100) |
| TSQM Global Satisfaction | |||
| Mean (95% CI) | 70 (63-71) | 75 (72-78) | 78 (75-79)b |
| Median (IQR) | 67 (50-92) | 75 (67-92) | 83 (67-100) |
| Range | (0-100) | (8-100) | (0-100) |
a Medical Outcomes Study (MOS) Cognitive Functioning Scale of 0-60, with higher scores reflecting better function. Treatment Satisfaction Questionnaire for Medication (TSQM) scores range from 0-100, with 100 being the best score. Kurtzke Expanded Disability Status Scale (EDSS) is scored on a 0-10 scale, with 10 being the worst. Multiple Sclerosis Impact Scale (MSIS) includes both physical and psychological subscales that are scored on a scale of 0-100, with 100 being the worst score. Because outcomes were not expected to be normally distributed, all variables were compared among 3 self-reported adherence (SRA) groups and 95% CIs were generated using nonparametric bootstrapping as described in the text. Significant differences in mean from the SRA <80% group were determined.
b P<.05.
Discussion
In this population of patients with MS in a large integrated health system, we observed very high levels of MS medication adherence overall and no strong associations between the degree of adherence and hospital admissions or ED use. Patients with low adherence had higher hospital-related costs but significantly lower mean total costs than patients with intermediate or high adherence, presumably due to more outpatient visits and pharmacologic utilization by patients in the higher-adherence groups. Patients with high adherence to MS medications fared better on MS-specific measures: on average, they had significantly better EDSS scores, MS impact scores (physical and psychological), and treatment satisfaction scores than patients with low adherence.
Previous research has shown positive relationships between adherence and outcomes and negative relationships between adherence and health care utilization in patients with MS or other diseases.21,22 The current study findings show positive effects of adherence on MS-specific functional outcomes. Patients in the low-adherence group had statistically significantly worse EDSS and MSIS scores than patients in the higher-adherence groups, supporting the idea that medication adherence in this population improves MS-specific outcomes that are not easily captured in the EHR or claims databases. We recognize, however, that the differences in EDSS scores observed, though statistically significant, may not be clinically significant. Another important finding was that highly adherent patients had the highest scores on the TSQM Global Satisfaction scale, showing patients’ satisfaction with their medication. The causality of this relationship is unknown from the data—eg, the patients could have been more satisfied because their medication was working, or the patients could have been more adherent because they felt satisfied with their treatment—but the association is strong.
In contrast to the MS disease-specific outcome measures, we did not see stronger evidence of a positive impact of adherence on health care utilization or cost that had been observed in previous literature on medication adherence.21,23 Previous studies of antihypertensive medication adherence, for example, have shown significantly lower health care costs and lower risk of hospitalization with increased adherence.10 Yermakov et al22 examined 1510 patients in a large claims database and found 9% to 19% reductions in the risk of hospitalization or ED visit and a very modest (3%-5%) reduction in costs associated with a 10% point increase in adherence; they did not define a threshold for desirable adherence or assess patient-reported outcomes, however. We expected to see similar trends but did not, which may be owing to our smaller sample size or differences in population. Our data did show a negative association between adherence and MS-related hospitalizations and ED visits, as well as higher hospital-related costs in patients with low adherence, but these associations were not statistically significant. Patients with low adherence had significantly lower mean total costs than patients with higher adherence, presumably because of more outpatient visits and medication costs. This finding is consistent with previous literature suggesting that the cost savings from reduced hospitalizations in higher-adherence patients may be balanced by an increased pharmacy expenditure.24 Unexpected findings in the intermediate-adherence group were that these patients had higher mean outpatient and total costs—both for all-cause and MS-related care—than either the low- or high-adherence groups. This finding suggests that the relationship between medication adherence and health care utilization by patients with MS may be more complex than a simple positive or negative association. This information may help the medical community by providing a slightly different spin on how we view the management of chronic disease in disciplines that prescribe medications similar to those used for MS, such as rheumatology.
The vast majority of patients in both the retrospective and prospective portions of the study had very high adherence that exceeded the 80% benchmark traditionally used for other diseases. We hypothesize that this high adherence is not unique to an integrated health system population but instead reflects the difference between MS (where consequences of nonadherence may directly manifest as relapses) and other diseases for which medication adherence has been studied (eg, blood pressure or lipid-lowering medications where the effects of nonadherence may take longer to observe).2 These data would suggest that it may be important to define different thresholds for adherence when studying different diseases, as 80% adherence may be clinically important or remarkable for one disease state but not for another.
The primary strength of our study was the 2-pronged approach of the design, which leveraged both administrative records and direct patient contact to examine relationships between medication adherence, outcomes, and health care utilization. Traditional data sources combined with self-reported outcomes allowed us to examine adherence-related clinical manifestations of MS as well as to estimate the quantitative impact on health care utilization and costs.
Limitations of the study included the relatively small sample size in a single health system. Ironically, the large number of patients with high adherence was detrimental to the statistical power of the study; because 18% of the retrospective cohort patients had poor adherence instead of the 40% we anticipated, our power to conclude statistical significance from the differences observed was greatly diminished. We acknowledge that MS is a long-term, progressive disease, and most patients were observed for 12 months, which limits the conclusions we can draw about long-term outcomes. Our regional population is also predominantly white, and so caveats regarding generalizability to other more diverse populations apply. We recognize that the patient-reported relapse rate in the prospective survey study could have incorrectly estimated the true rate. Also, during the study period (2010-2014), most therapy options were injectable, and adherence trends may have evolved now that more oral medications have become available.
Conclusion
Adherence to MS medications was very high overall in our samples. High adherence was reflected both in claims-based measures and SRAs, suggesting that either method of assessing adherence yields reasonably consistent information. Medications that are designed to promote adherence by virtue of their route (eg, oral over injectable), dosing regimen (eg, single dose over multiple doses), and price (affordable) should help to significantly improve symptoms and functioning for patients with MS. Decision-makers in the health care system and the health insurance industry should be aware that although patients with MS are mostly adherent to their existing treatments, adherence affects symptoms but does not necessarily reduce health care utilization or costs in the short term. However, there was some evidence of lower hospitalization costs related to higher adherence, but it did not reach statistical significance. Future research following the methods outlined in the current study but using claims from multiple centers, and thus a larger population, could examine these relationships further.
Author Contributions
All authors provided substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; Drs Hao, Hoegerl, and Graham and Mr Pitcavage drafted the article or revised it critically for important intellectual content; all authors gave final approval of the version of the article to be published; and all authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
References
1. Kornek B , StorchMK, WeissertR, et al. Multiple sclerosis and chronic autoimmune encephalomyelitis: a comparative quantitative study of axonal injury in active, inactive, and remyelinated lesions. Am J Pathol. 2000;157(1):267-276. doi:10.1016/S0002-9440(10)64537-3Search in Google Scholar
2. Prescott JD , FactorS, PillM, LeviGW. Descriptive analysis of the direct medical costs of multiple sclerosis in 2004 using administrative claims in a large nationwide database.J Manag Care Pharm.2007;13(1):44-52. doi:10.18553/jmcp.2007.13.1.44Search in Google Scholar PubMed
3. Browne P , ChandraratnaD, AngoodC, et al. Atlas of Multiple Sclerosis 2013: a growing global problem with widespread inequity. Neurology. 2014;83(11):1022-1024. doi:10.1212/WNL.0000000000000768Search in Google Scholar PubMed PubMed Central
4. Multiple Sclerosis Coalition. The Use of Disease-Modifying Therapies in Multiple Sclerosis: principles and Current Evidence.Hackensack, NJ: Multiple Sclerosis Coalition; 2014. Updated March 2017. http://www.nationalmssociety.org/getmedia/5ca284d3-fc7c-4ba5-b005-ab537d495c3c/DMT_Consensus_MS_Coalition_color. Accessed December 17, 2015.Search in Google Scholar
5. Bischoff C , SchreiberH, BergmannA. Background information on multiple sclerosis patients stopping ongoing immunomodulatory therapy: a multicenter study in a community-based environment. J Neurol. 2012;259(11):2347-2353. doi:10.1007/s00415-012-6499-1Search in Google Scholar PubMed
6. Saunders C , CaonC, SmrtkaJ, ShoemakerJ. Factors that influence adherence and strategies to maintain adherence to injected therapies for patients with multiple sclerosis. J Neurosci Nurs. 2010;42(5 suppl):S10-S18.10.1097/JNN.0b013e3181ee122bSearch in Google Scholar PubMed
7. Costello K , KennedyP, ScanzilloJ. Recognizing nonadherence in patients with multiple sclerosis and maintaining treatment adherence in the long term. Medscape J Med. 2008;10(9):225.Search in Google Scholar
8. Sikka R , XiaF, AubertRE. Estimating medication persistency using administrative claims data.Am J Manag Care.2005;11(7):449-457.Search in Google Scholar
9. Vegter S , NguyenNH, VisserST, de Jong-van den BergLT, PostmaMJ, BoersmaC. Compliance, persistence, and switching patterns for ACE inhibitors and ARBs. Am J Manag Care. 2011;17(9):609-616.Search in Google Scholar
10. Pittman DG , TaoZ, ChenW, StettinGD. Antihypertensive medication adherence and subsequent healthcare utilization and costs.Am J Manag Care.2010;16(8):568-576.Search in Google Scholar
11. Turner AP , KivlahanDR, SloanAP, HaselkornJK. Predicting ongoing adherence to disease modifying therapies in multiple sclerosis: utility of the health beliefs model.Mult Scler.2007;13(9):1146-1152. doi:10.1177/1352458507078911Search in Google Scholar PubMed
12. Mohr DC , BoudewynAC, LikoskyW, LevineE, GoodkinDE. Injectable medication for the treatment of multiple sclerosis: the influence of self-efficacy expectations and injection anxiety on adherence and ability to self-inject.Ann Behav Med.2001;23(2):125-132.10.1207/S15324796ABM2302_7Search in Google Scholar PubMed
13. Rio J , PorcelJ, TellezN, et al. Factors related with treatment adherence to interferon beta and glatiramer acetate therapy in multiple sclerosis. Mult Scler. 2005;11(3):306-309. doi:10.1191/1352458505ms1173oaSearch in Google Scholar PubMed
14. Portaccio E , ZipoliV, SiracusaG, SorbiS, AmatoMP. Long-term adherence to interferon beta therapy in relapsing-remitting multiple sclerosis.Eur Neurol.2008;59(3-4):131-135.10.1159/000111875Search in Google Scholar PubMed
15. Stewart A , WareJ, eds. Measuring Functioning and Well-Being: The Medical Outcomes Study Approach.Durham, NC: Duke University Press; 1992.10.7249/CB361Search in Google Scholar
16. Stephenson JJ , KernDM, AgarwalSS, et al. Impact of natalizumab on patient-reported outcomes in multiple sclerosis: a longitudinal study. Health Qual Life Outcomes. 2012;10:155. doi:10.1186/1477-7525-10-155Search in Google Scholar PubMed PubMed Central
17. Thompson A , LampingD, FreemanJ, HobartJ. Kurtzke expanded disability status scale (EDSS): how best should it be used? Neurology. 1997;48(3):4045.Search in Google Scholar
18. Ting J , LiuY, PetrilloJ, GiannattasioG, SabatellaG. Treatment satisfaction with disease modifying therapies in multiple sclerosis: a systematic review of studies using the Treatment Satisfaction Questionnaire For Medication (TSQM). Value Health. 2015;18(7):A760-A761. doi:10.1016/j.jval.2015.09.2484Search in Google Scholar
19. Hobart J , LampingD, FitzpatrickR, RiaziA, ThompsonA. The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure. Brain. 2001;124(pt 5):962-973.10.1093/brain/124.5.962Search in Google Scholar PubMed
20. Kozma CM , DicksonM, PhillipsAL, MeleticheDM. Medication possession ratio: implications of using fixed and variable observation periods in assessing adherence with disease-modifying drugs in patients with multiple sclerosis.Patient Prefer Adherence.2013;7:509-516. doi:10.2147/PPA.S40736Search in Google Scholar PubMed PubMed Central
21. Lizan L , ComellasM, PazS, PovedaJL, MeleticheDM, PolancoC. Treatment adherence and other patient-reported outcomes as cost determinants in multiple sclerosis: a review of the literature. Patient Prefer Adherence. 2014;8:1653-1664. doi:10.2147/PPA.S67253Search in Google Scholar PubMed PubMed Central
22. Yermakov S , DavisM, CalnanM, et al. Impact of increasing adherence to disease-modifying therapies on healthcare resource utilization and direct medical and indirect work loss costs for patients with multiple sclerosis. J Med Econ. 2015;18(9):711-720. doi:10.3111/13696998.2015.1044276Search in Google Scholar PubMed
23. Thomas NP , CurkendallS, FarrAM, YuE, HurleyD. The impact of persistence with therapy on inpatient admissions and emergency room visits in the US among patients with multiple sclerosis. J Med Econ. 2016;19(5):497-505. doi:10.3111/13696998.2015Search in Google Scholar
24. Tan H , YuJ, TabbyD, DevriesA, SingerJ. Clinical and economic impact of a specialty care management program among patients with multiple sclerosis: a cohort study. Mult Scler. 2010;16(8):956-963. doi:10.1177/1352458510373487Search in Google Scholar PubMed PubMed Central
© 2017 American Osteopathic Association
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Articles in the same Issue
- SURF
- Challenging Case of Parotitis: A Comprehensive Approach
- OMT MINUTE
- OMT for Patients With Multiple Sclerosis
- IN MY VIEW
- Defining Osteopathic Continuing Medical Education
- LETTERS TO THE EDITOR
- Landmark Article Transforms Traditional View of the Autonomic Nervous System
- ORIGINAL CONTRIBUTION
- Measuring Adherence and Outcomes in the Treatment of Patients With Multiple Sclerosis
- BRIEF REPORT
- Musculoskeletal Disorders in Ophthalmologists After Simulated Cataract Operation: A Pilot Study
- REVIEW
- Benign Breast Conditions
- Role of Antiplatelet Therapy in Stroke Prevention in Patients With Atrial Fibrillation
- JAOA/AACOM MEDICAL EDUCATION
- Predictors of Osteopathic Medical Students’ Readiness to Use Health Information Technology
- CASE REPORT
- Osteopathic Manipulative Treatment for the Management of Adjacent Segment Pathology
- Inguinal Herniation of Perinephric Tissue: Case Report and Review of the Literature
- CLINICAL IMAGES
- Vesicovaginal Fistula
Articles in the same Issue
- SURF
- Challenging Case of Parotitis: A Comprehensive Approach
- OMT MINUTE
- OMT for Patients With Multiple Sclerosis
- IN MY VIEW
- Defining Osteopathic Continuing Medical Education
- LETTERS TO THE EDITOR
- Landmark Article Transforms Traditional View of the Autonomic Nervous System
- ORIGINAL CONTRIBUTION
- Measuring Adherence and Outcomes in the Treatment of Patients With Multiple Sclerosis
- BRIEF REPORT
- Musculoskeletal Disorders in Ophthalmologists After Simulated Cataract Operation: A Pilot Study
- REVIEW
- Benign Breast Conditions
- Role of Antiplatelet Therapy in Stroke Prevention in Patients With Atrial Fibrillation
- JAOA/AACOM MEDICAL EDUCATION
- Predictors of Osteopathic Medical Students’ Readiness to Use Health Information Technology
- CASE REPORT
- Osteopathic Manipulative Treatment for the Management of Adjacent Segment Pathology
- Inguinal Herniation of Perinephric Tissue: Case Report and Review of the Literature
- CLINICAL IMAGES
- Vesicovaginal Fistula
