We conducted a retrospective cohort study using United States (US) health insurance claims, demographic, and enrollment data. Two data sources were used. IQVIA’s PharMetrics Plus database (1/1/2006-6/30/2021) was used to determine patients with AMN demographic and clinical characteristics, all-cause HRU, and the actual amounts paid by payers for all-cause medical and prescription medication services. The PharMetrics Plus data is representative of the US commercially insured population for individuals < 65 years and contains demographic, enrollment, and fully adjudicated medical and prescription drug claims data for approximately 150 million deidentified individuals enrolled in US commercial health insurance plans, with an annual capture of approximately 40 million individuals [10].

PharMetrics Plus data do not, however, include mortality. Consequently, we used the Medicare Limited Data Set (LDS) (1/1/2016-12/31/2020) to examine mortality among patients with AMN. The LDS contains enrollment, demographic and claims data for deidentified Medicare beneficiaries, with approximately 60 million Medicare beneficiaries enrolled in or entitled to Medicare, within a given calendar year. These beneficiaries include age-eligible individuals (≥ 65 years) and individuals with Medicare coverage who are disability-eligible or eligible related to end-stage renal disease (< 65 years) [11]. Medicare claims data from the Medicare Common Working File, online date of death edits submitted by the deceased’s family members, and benefit information collected from the Railroad Retirement Board and the Social Security Administration are the main sources used to develop LDS mortality data – the Master Beneficiary Summary File base segment within LDS contain date of death but not cause of death [12].

Study patients were male adults with ≥ 1 inpatient or ≥ 2 outpatient claims with a diagnosis of AMN (International Classification of Diseases, 9th or 10th revision, Clinical Modification (ICD-9-CM/ICD-10-CM) claim (Table 1). Individuals with evidence of Zellweger syndrome (ICD-10-CM: E71.510), rhizomelic chondrodysplasia punctata (ICD-10-CM: E71.540), and/or childhood cerebral X-linked adrenoleukodystrophy (ICD-10-CM: E71.520) were excluded. There is no ICD diagnosis code for adult cerebral adrenoleukodystrophy, which may be associated with very high costs. Consequently, we examined the distribution of outpatient utilization and costs relative to inpatient utilization and costs. However, we did not observe any association that might lead to stratifying patients with adult cerebral adrenoleukodystrophy (Figs. 1 and 2). Commercially insured individuals were limited to those age 18–64, while those with Medicare coverage were age 18 and above, with disability-eligible beneficiaries age < 65, and age-eligible age ≥ 65. Inclusion and exclusion criteria were applied to study patients with commercial and Medicare insurance coverage.

Average in-patient and out-patient costs per month for patients meeting study criteria

Average in-patient and out-patient healthcare encounters per month for patients meeting study criteria

Commercially insured study patients were 1:4 propensity score matched (PSM) to individuals with no evidence of AMN on the basis of sex, age, geographic region, and continuous enrollment time period. To permit assessment of differences in coexisting clinical conditions, these non-AMN controls were not matched to AMN cases based on comorbid conditions.

All patients were followed from their index date, which was defined as the date of the first claim with a qualifying diagnosis in the study data. Commercially insured patients were observed for as long as they were enrolled and observable and censored when they were lost to follow-up. For measuring mortality, study patients with Medicare coverage were followed until they were no longer enrolled in Medicare fee-for-service coverage or died.

Demographic characteristics (gender, age, and geographic region) were identified using demographic information in the PharMetrics Plus and LDS data. Patient demographic characteristics were measured as of the index date. Patient comorbidities and clinical characteristics, as well as HRU and costs, were measured in the post-index observation window.

Individual’s comorbid conditions were identified using clinical codes recorded on claims (Tables 2 and 3). In addition, the Charlson Comorbidity Index [CCI], a commonly used measure of health status was calculated. The CCI examines and sums comorbid conditions associated with an increased likelihood of mortality. Higher CCI scores are associated with poorer health status and an increased risk of death [13, 14].

Healthcare resource use was identified by examining claims for inpatient admissions, outpatient encounters, and prescription medications for each patient. Setting-level information was identified using the place-of-service recorded on each patient’s medical claims. Inpatient admissions were measured overall and stratified by those with and those without an intensive care unit (ICU) admission. Outpatient encounters included emergency department (ED), outpatient hospital, office, laboratory and imaging testing, home healthcare, and physical therapy visits, durable medical equipment, and other outpatient encounters. ED visits were those where a patient was not hospitalized, as ED visits that result in a hospitalization are rolled into the claim for the hospital admission. Prescription drug information was evaluated using information on claims for filled prescriptions and prescription medications for AMN symptomatic treatment (including adrenal insufficiency, mood disorders, neuropathy, incontinence, and sexual dysfunction) identified using Generic Product Identifier (GPI) codes.

Costs (pharmacy, inpatient, and outpatient costs) were identified using the actual amounts paid by health plans to providers, which were recorded on each claim.

Analyses of mortality were conducted using Medicare LDS data. Patients with AMN were compared to the overall population of Medicare enrollees during the Medicare study window (1/1/2016-12/31/2020). Age-eligible (age ≥ 65) and disability-eligible (age < 65) beneficiaries were evaluated separately.

Univariate analyses were conducted and unadjusted counts for study patients reported. Because adults with AMN might move from commercial to other commercial insurance plans or public insurance coverage, or die, we followed each patient as long as they were observable in the data and calculated HRU rates and costs over the time they were observable [15]. HRU and costs, were then estimated as averages per patient per year (PPPY), included individuals without utilization (0 encounters) or with no costs ($0 USD) recorded for a given healthcare encounter.

Bivariate analyses, Mann-Whitney U/Student’s t-test for continuous variables and Fisher’s Exact/Chi-squared tests for discrete variables, were used to compare unadjusted counts to assess differences between cases and controls.

Mortality outcomes, based on Medicare data, were calculated as the crude death rate and average over the 5-year study window.

Statistical analyses were conducted using SAS 9.3 (SAS Institute, Cary, NC) and R 2022.02. Statistical significance was defined as a two-tailed p < 0.05.

This study was exempt from Institutional Review Board approval, as it involved secondary data analyses of fully deidentified data.