ABSTRACT
The medical system in the United States has been riddled with insurance restrictions used by insurance companies to limit health care costs. The effects of insurance restrictions on patients receiving disease-modifying therapies for multiple sclerosis (MS) have not been specifically studied.
A retrospective cross-sectional study of 52 individuals recently diagnosed with MS at a tertiary neurology clinic was conducted to measure the association between prior authorization (PA) duration and other variables of interest. The Cox proportional hazards model was used to determine likelihood of approval. Further analysis included multivariable logistic regression to assess the influence of variables of interest on the initial decision from the insurance company and the effect of the PA on disease activity.
Of 52 PAs, 50% were initially denied. An initial denial decreased the likelihood of approval by 98% (HR, 0.02; 95% CI, <0.01-0.09; P < .001). The odds of denial for oral medications (odds ratio [OR], 4.91; 95% CI, 1.33-21.52; P = .02) and infusions (OR, 8.35; 95% CI, 1.10-88.77; P = .05) were significantly higher than for injections. Medicaid had higher odds of denial compared with commercial insurance (OR, 4.51; 95% CI, 1.13-22.01; P = .04). An initial denial by insurance significantly increased the likelihood of disease activity (OR, 6.18; 95% CI, 1.33-44.86; P = .03).
Insurance restrictions delay necessary treatments, increase the likelihood of disease activity, and rarely change the approved disease-modifying therapy. Reducing PAs may lead to improved outcomes for patients with MS.
Multiple sclerosis (MS) is an incurable, chronic inflammatory disease of the central nervous system that affects nearly 1 million people in the United States.1 Timely initiation of disease-modifying therapies (DMTs) can reduce relapses, prevent new lesions, and delay progression of disability.2,3 Although numerous MS DMTs have been approved by the United States Food and Drug Administration, standard guidelines for initiation of specific medications have not been established.4 The DMT selection process involves a nuanced discussion between the provider and the patient regarding expected benefits and potential adverse events.
The United States perennially underachieves in health care outcomes and health coverage among industrialized nations while consistently spending more per capita.5 In the United States, the total estimated cost of MS care was $85.4 billion in 2019, with an average annual per-person cost greater than $65,000.6 As the source of the exorbitant costs,6 MS DMTs represent approximately 60% to 70% of the total cost of health care for patients with MS,6,7 with increases in price estimated to be 15% annually.8 The cost of health care for patients with MS approximately doubled between 2009 and 2015, with DMTs accounting for 82% of this rise.9 Medicare, presumed to be the largest purchaser of MS DMTs in the United States, provides insurance for approximately 25% to 30% of the nation’s MS population10 and spent more than $4 billion through Medicare Part D in 2017.11 For comparison, the United States spends more on MS DMTs than the gross domestic product of numerous countries, including Guyana and Somalia.12
With medication costs representing a substantial barrier to access, insurance companies have restricted coverage through a myriad of techniques, including prior authorizations (PAs), cost sharing, quantity limits, preferred drug lists, and step therapy policies.8,13 The need for PA, a restriction that requires specific clinical criteria to be met before coverage approval,14 has risen steadily from approximately 60% in 2007 to nearly 90% in 2016.8 In other diseases, PAs have been associated with increased costs to patients,15 reduced access to medications,16 and inferior outcomes.17 Estimates of time for medical staff to address difficulties in insurance coverage are approximately 20 to 30 hours per month.18 In a survey of physicians, 88% reported the PA burden as “high or extremely high.”19
Delays in care due to insurance coverage restrictions have been documented in dermatology,20,21 rheumatology,22 asthma,23 inflammatory bowel disease,24 and epilepsy.13 Medicare part D coverage restrictions for MS DMT were associated with reduced medication compliance and increased health care utilization.25 The purpose of this analysis was to assess the duration of PAs in MS DMT, investigate factors determining PA approval, and evaluate the relationship between insurance restrictions and disease activity in patients newly diagnosed as having MS.
METHODS
This was a retrospective cohort study approved by the institutional review board of Advent Health. A retrospective medical record review identified patients seen between August 1, 2020, and January 1, 2022, at a tertiary neurology clinic by a fellowship neurologist trained in MS care.
Inclusion Criteria
To be included in the study, participants met the following criteria: (1) diagnosis of clinically isolated syndrome (CIS), relapsing-remitting MS (RRMS), or primary progressive MS (PPMS) based on the 2017 revised McDonald criteria26 ; (2) diagnosed as having MS within 6 months of PA initiation; (3) not previously prescribed a medication for MS; (4) willing to start an MS DMT; (5) PA was required by insurance; (6) not pregnant; and (7) 18 years of age or older.
Variables of Interest
Once eligible participants were identified, demographic data (ie, age, race, MS type, risk factors for highly active disease,27 and payer type) were collected. The MS types were CIS, RRMS, and PPMS based on the 2017 McDonald criteria.26 Risk factors suggestive of an aggressive form of MS included demographics (male sex, age at onset >40 years, African American or Hispanic race), type of initial attack (motor, cerebellar, sphincteric, or cognitive, multifocal attack, or incomplete recovery), and MRI features (excess lesion burden on T2, more than 2 contrast-enhancing lesions present on initial MRI, presence of T1 black holes on initial MRI, infratentorial lesions).27 Other variables included DMT type (injection, oral, infusion), brand name or generic, and whether the DMT was on formulary for the respective insurance.
The duration of PA was defined as the difference between the submission date of the PA and the date of acceptance or abandonment. If the first response from the assessing organization was an approval, this was defined as an initial approval. If the first response from the assessing organization was a denial, this was defined as an initial denial. After an initial denial, 2 more attempts, at most, were made to obtain approval. After a total of 3 denials for a PA (ie, the initial denial and 2 additional denials), the PA was abandoned. The date on the notice of the third denial was the end of the duration for an abandoned PA. Abandoned PAs occurred only if the initial response was a denial. The duration of PA was measured in days. All patients whose initial PA was abandoned had a subsequent PA submitted for approval, although only the initial PA was included in the analysis. All PAs were for United States Food and Drug Administration–approved medications for the treatment of MS.
Reasons for denial were obtained from the notification letter and included step therapy, nonformulary, and criteria not met. Step therapy is a denial due to the requirement to try other medications before approval. A nonformulary DMT is one that is not a part of the insurance plan’s preferred medication list. Criteria not met is a denial based on the insurance company requesting further information. No medication was put into more than 1 of these categories by the insurance company. Disease activity was defined as a relapse or a new or enhancing lesion on imaging that occurred after the PA was initiated and within 6 months of approval.
Statistical Analysis
The Cox proportional hazards model was used to assess the risk of PA approval. Variables of interest were assessed by univariate log-rank analysis for inclusion in the Cox proportional hazards model (P < .20) using stepwise forward selection.28 Censuring did not occur because all PAs were approved or abandoned within the given collection time. An exploratory analysis using a multivariable logistic regression approach was used to assess the relationship of variables of interest to the initial decision by the insurance company and to assess the relationship of insurance restrictions on MS disease activity. Before inclusion in the logistic regression, variables were checked for collinearity via the variance inflation factor and considered acceptable if less than 2.5.29 A forward stepwise multivariable logistic regression was used for modeling outcome variables selected during the previous step. We considered variables with P < .20 on univariate analysis as candidates for multivariable modeling. Data are reported with adjusted odds ratios (ORs) and 95% CIs. Overall model evaluation was performed using the likelihood ratio test. Discrimination assessment for predicted probabilities was performed via area under the curve (AUC) of the receiver operating characteristic (ROC) curve and were considered adequate at 0.7 to 0.8, good at 0.8 to 0.9, and excellent at 0.9 or greater.30 All statistical analyses were performed using RStudio version 4.1.3 (Posit). A P ≤ .05 was considered significant.
RESULTS
Of the 58 patients with MS who were eligible to be included in the analysis, 3 were not willing to start a DMT, 2 became pregnant during the PA process, and 1 prescription was approved without requiring a PA (FIGURE S1, available online at IJMSC.org). Due to these exclusions, the final sample size was 52 patients. The analyzed sample was composed of predominantly female participants (65%) with a median age of 35.0 years (range, 18-57 years). The sample was 38% White, 15% African American, 35% Latin American, and 12% Asian and Pacific Islander. The sample had 15 patients with CIS (29%), 35 with RRMS (67%), and 2 with PPMS (4%). The patients had a median of 5 (range, 0-10) risk factors suggestive of aggressive disease. In the period analyzed for activity, 12 patients (23%) experienced disease activity; 10 (83%) of these patients had disease activity after an initial denial. Demographic information is included in TABLE 1.
Of the 52 patients who required a PA, 26 were initially approved and 26 were initially denied. The overall median approval time was 14 days (range, 1-79 days). The median time until approval was 4 days (range, 1-15 days) among those with an initial PA approval and 37 days (range, 8-79 days) among those with an initial PA denial (P < .01). Ultimately, 47 of the PAs (90%) were eventually approved, and 5 (10%) were abandoned. Brand name DMTs were prescribed 38 times (73%), and in-formulary medication accounted for 47 of the PAs (90%). Injections accounted for 24 of the PAs (46%), followed by 22 oral medications (42%) and six infusions (12%). Specific numbers of DMTs and percentages are included in Table 1. The reasons for denial were step therapy (14; 54%), criteria not met (7; 27%), and nonformulary (5; 19%). The payer types were commercial insurance (35; 67%), Medicaid (15; 29%), and Medicare (2; 4%).
For the Cox proportional hazards model that assessed the time to PA approval, only the initial decision by the insurance company met the inclusion criteria. An initially denied PA decreased the likelihood of approval by 98% (HR, 0.02; 95% CI, <0.01-0.09; P < .001). See FIGURE S2 for cumulative incidence curve.
With a multivariable logistic regression analysis with DMT type and payer type as the independent variables and the initial decision by the insurance company as the dependent variable, the odds of having PA initially denied were increased for oral medications (OR, 4.91; 95% CI, 1.33-21.52; P = .02) and infusions (OR, 8.35; 95% CI, 1.10-88.77; P = .05) compared with injections and holding payer type constant. While holding type of medication constant, the odds of having PA initially denied were increased for Medicaid (OR, 4.51; 95% CI, 1.13-22.01; P = .04) over commercial insurance (TABLE S1). The AUC of ROC curve was 0.73. The odds of denial for Medicare were not statistically significant.
A multivariable logistic regression was performed to evaluate the relationship between PA denial and disease activity; an initial PA denial or approval and the total number of risk factors for severe disease27 were the independent variables, and disease activity was the dependent variable. After adjusting for potential confounding variables, the odds of having disease activity increased if PA was initially denied (OR, 6.18; 95% CI, 1.33-44.86; P = .03) compared with an initial approval (TABLE 2). The AUC of ROC curve was 0.78.
DISCUSSION
In this study, MS DMT insurance restrictions were associated with significant delays in the care of newly diagnosed patients with MS. Previously, coverage restrictions in Medicare part D have been associated with reduced medication compliance and increased utilization of health care in patients with MS.25 Although the overall median approval time was similar to other diseases, an initial denial led to a delay in approval nearly 8 times longer than an initial approval in patients with MS. This lengthy delay occurred in the setting of an overall approval rate of 90%, which was commensurate with similar studies in other fields.22,31
With the ultimate purpose of starting newly diagnosed patients with MS on a DMT as early as possible, substantial strides have been made in refining the diagnostic criteria and decreasing the time to MS diagnosis.32 However, insurance restrictions are associated with delays in approval of MS medications. Variables other than the initial response from the insurance company, including patient demographics, DMT characteristics, and insurance factors, did not significantly affect time to approval because the only variable that affected PA duration was the initial decision of the payer.
As the only variable that significantly affected DMT approval time, further assessment was undertaken to ascertain which factors influence the initial response from the insurance company. Compared with commercial insurance, PAs for Medicaid had almost 5 times greater odds of an initial denial compared with holding the medication type constant. Of similar studies in different diseases, only in adult epilepsy was a delay attributed to an insurance type, for which Medicare had significantly longer approval delays.21-24 Treatment delays originating from social determinants of health and the potential impact throughout the life span of patients with MS should be explored in future studies.
The type of DMT prescribed impacted the PA process. Compared with injectables, a PA for an oral medication had almost 5 times greater odds of receiving an initial denial, and PAs for infusions had 8 times greater odds of receiving an initial denial. The PA process was implemented by insurance companies to help curb the cost of medications, and the high cost of MS medications is well documented.14 The total cost of prescribed medications for MS was estimated to be nearly $38 billion in 2019.6 According to Hartung,14 “DMT prices have accelerated upwards in a coordinated pattern where each successive increase is seemingly matched by other products.”
Coverage restrictions in Medicare part D for patients with MS have been associated with reduced medication compliance and increased clinic visits,25 although this is not a direct measure of MS disease activity. In this study, compared with an initial approval, an initial PA denial from the payer was more than 5 times more likely to result in disease activity for the patient. Disease activity occurred in 38% of patients whose PA was initially denied compared with 7% of patients whose PA was initially approved. To our knowledge, this is the first evidence of an association between insurance restrictions and disease activity in MS. With a delayed medication effect,33 the introduction of further delays by the insurance approval process is deleterious to the health of patients with MS, particularly when only 10% of PAs were ultimately changed. Relapses contribute to long-term disability,34 and future studies should investigate the long-term effects of PA delays on disease activity and progression.
This study has several limitations. The analysis was retrospective, and bias may have been introduced via unassessed variables. Only the duration of the PA was assessed, and obtaining a PA is only a portion of the process. Receiving the medication from the specialty pharmacy represents another barrier to treatment, as greater delays were seen when a specialty pharmacy filled the medication compared with the PA process.23 An assessment of specialty pharmacies’ contributions to MS DMT delay is warranted in future studies. Other barriers to treatment, such as wait times for neurologists and acquisition of required safety laboratory tests or imaging before prescription, were not assessed and could contribute to delays. Another limitation is that this is an analysis of a single center, which may not reflect the PA experience at other centers. Other centers may have policies in place or dedicated employees who aid in obtaining PA approval. This is indicative of the bureaucratic arms race, where more costs are incurred by the practice to better navigate the complexities of this restrictive process. A multicenter analysis composed of diverse centers and regions is warranted to assess the applicability of these results to a larger population. Similarly, the selected DMTs were reflective of the proclivities of a single provider and may not be representative of a larger population.
CONCLUSIONS
Insurance restrictions resulted in significant delays in MS DMT approval and were associated with an increased likelihood of disease activity. With an alteration in MS DMT occurring in merely 10% of cases, the utility of this harmful process is disputable and warrants further evaluation.
Due to Medicaid insurance restrictions, patients with multiple sclerosis receiving oral or infusion disease-modifying therapies faced considerable treatment delays.
Insurance restrictions on disease-modifying therapies for multiple sclerosis are a barrier to receiving care and are associated with higher rates of disease activity.
REFERENCES
FINANCIAL DISCLOSURES: The author declares no conflicts of interest.
FUNDING/SUPPORT: None.
Author notes
Note: Supplementary material for this article is available online at IJMSC.org.