Background/Purpose:

Little is known about the healthcare costs of systemic lupus erythematosus (SLE) patients in the years leading up to SLE diagnosis.  We estimated the incremental (extra) direct medical costs of a general population-based cohort of incident SLE for five years before diagnosis, and examined the impact of sex on these costs. 

Methods:

Data Source: Our administrative data captured all provincially-funded outpatient encounters and hospitalisations (1990-2013), and all dispensed medications, for ALL residents of the province of British Columbia, Canada.

Sample: We assembled a population-based cohort of incident SLE: those with a new diagnosis of SLE from at least one hospitalisation or rheumatologist visit, or two non-rheumatologist visits, between Jan 2001 and Dec 2010, and no prior SLE diagnosis between Jan 1990 and Dec 2000.  For each SLE case, we matched up to 5 non-SLE from the general population on age (±2 years), sex and calendar year of diagnosis.  All persons had ≥ 5 years’ follow-up in the databases before index date (first SLE-coded encounter; random date for non-SLE). 

Cost Calculation: Outpatient and prescription costs were summed directly from billing data. Case-mix methodology was used for hospitalisations.

Analysis:

We estimated the unadjusted incremental costs of SLE (difference in per-person costs between SLE and non-SLE) for the five pre-diagnosis/pre-index years (Y-5 to Y-1) and index year (Y0). 

Generalised linear models were used to further adjust for socioeconomic status (SES), urban/rural and comorbidities between SLE and non-SLE, and evaluate the impact of sex on costs.

Results:

We included 3,632 incident SLE (86% female, mean age 49.6 years) and 18,152 non-SLE.  Index-year (Y0) costs for SLE averaged $12,019 per-person (2013 CDN): 59% from hospitalisations, 24% outpatient, and 18% medications.  Costs increased by 35% per year, on average, with the biggest increases in the two years before diagnosis (see Table).  Adjusted cost ratios between SLE and non-SLE rose from 1.7 in Year -3 (Y-3) to 1.9 (Y-2), 2.4 (Y-1), and 4.0 in Y0. 

Among non-SLE, adjusted costs were higher for females than males, but among SLE, males had higher costs, controlling for age, SES and previous year’s comorbidity score.  SLE males had higher odds of hospitalisation than SLE females in Y-1 (OR=1.44, 95% CI=1.18-1.76), while non-SLE males had 15% lower odds. In Years -2 & -1, encounters with primary diagnosis of diabetes, or renal or cardiovascular disease accounted for 11% of costs for SLE males, vs. 5% for SLE females.

Adjusted M/F cost ratios in SLE were 1.15 (95% CI 1.03-1.29) in Y-2, 1.22 (1.11-1.35) in Y-1, and 1.44 (1.29-1.61) in Y0.  When comparing costs of SLE and non-SLE, the Male*SLE interaction term was significant in all years.

Conclusion:

The incremental costs of SLE are considerable, even in the years prior to diagnosis.  Unlike the general population, SLE males had higher costs than females, potentially from early comorbidities.