Economic and Clinical Burden of Herpes Zoster Among Patients With Inflammatory Bowel Disease in the United States

Abstract Background Patients with ulcerative colitis (UC) or Crohn’s disease (CD) are at increased risk of herpes zoster (HZ); however, relevant cost and healthcare resource utilization (HCRU) data are limited. Methods We estimated HCRU (hospitalization, emergency department [ED], and outpatient visits) and costs in patients with UC or CD, with and without HZ, using administrative claims data (October 2015–February 2020). HCRU and costs (2020 US dollars) were compared at 1 month, 1 quarter, and 1 year after the index date, using propensity score adjustment and generalized linear models. Results In total, 20 948 patients were included: UC+/HZ+ (n = 431), UC+/HZ– (n = 10 285), CD+/HZ+ (n = 435), and CD+/HZ– (n = 9797). Patients with HZ had higher all-cause HCRU rates and all-cause total healthcare costs relative to those without HZ. In the first month, adjusted incidence rate ratios (aIRRs) for hospitalizations and ED visits for patients with UC and HZ compared with UC alone were 2.87 (95% confidence interval [CI], 1.93–4.27) and 2.66 (95% CI,1.74–4.05), respectively; for those with CD and HZ, aIRRs were 3.34 (95% CI, 2.38–4.70) and 3.31 (95% CI, 2.32–4.71), respectively, compared with CD alone (all P < .001). Adjusted cost differences in UC and CD cohorts with HZ over the first month were $2189 and $3774, respectively, chiefly driven by higher inpatient costs. The incremental impact on HCRU and costs in cohorts with HZ predominantly occurred during the first quarter following diagnosis. Conclusions HZ is associated with increased HCRU and costs in patients with UC and CD, especially shortly after diagnosis.

Eligible individuals were patients ≥18 years of age at the index date, with a diagnosis of UC or CD identified using relevant ICD-10 CM codes. Based on previously validated claims algorithms patients with UC (K51) or CD (K50) were grouped into two mutually exclusive patient cohorts (UC and CD). For individuals with claims for both UC and CD, a majority-based algorithm of claims in the 12-month baseline period was used to allocate patients into the UC or CD cohorts, as used in previous studies. 1,2 Patients were classified as having CD if (a) the number of CD-related inpatient (IP) admissions was higher than the number of UC-related IP admissions; (b) there was an equal number of CD-and UC-related IP admissions but more CDrelated outpatient (OP) visits than UC-related OP visits; or (c) there was an equal number of CD-and UC-related IP admissions and OP visits but the most recent claim prior to the HZ index date was for CD.

Comparison of Baseline Characteristics
Standardized differences were used to compare baseline characteristics between cohorts. For continuous variables, the standardized difference was calculated by dividing the absolute difference in means of each cohort comparison by the pooled standard deviation of both groups (where the pooled standard deviation was the square root of the average of the squared standard deviations). For categorical variables with two levels, the standardized difference was calculated using the following equation: abs(P1-P2)/√p (1-p)), where p = (P1+P2)/2; P1 was the respective proportion of participants in the UC+/HZ+ and CD+/HZ+ cohorts; and P2 was the respective proportion of participants in the UC+/HZ-and CD+/HZ-cohorts. Standardized differences of 20%, 50%, and 80% suggest small, medium, and large differences between cohorts, respectively. 3

Incidence Rate Ratio and Propensity Score Adjustment
Adjusted incidence rate ratios (aIRRs) and 95% confidence intervals (CIs) were estimated using generalized linear models (assuming a negative binomial distribution and log link, adjusting for potential confounders using the patients' propensity score as a covariate in addition to other baseline variables to adjust for baseline differences between cohorts). P-values were calculated using the negative binomial distribution.
In this approach, logistic regression was used to estimate a patient's propensity (or probability) of being in their cohort, based on their demographics on the index date (i.e., age, gender, geographic region, and insurance type), clinical characteristics (including baseline comorbidities and modified Charlson-Quan comorbidity index scores, baseline inflammatory bowel disease medication use, and prior clinical procedures) and healthcare costs incurred in the 12 months prior to the index date.

Cost Estimations and Adjusted Cost Differences
Healthcare costs for each of the study cohorts were estimated using a two-part model, an established method used in health economics to accommodate any skewed distribution for positive costs and any significant proportion of zero values, 4 as used in prior studies. 5,6 In the first part, logistic regression, incorporating propensity scores and other baseline variables as covariates in the model, was used to predict the probability of incurring any positive costs. In the second part, a generalized linear model with a gamma distribution and log link and incorporating propensity scores and other baseline variables as covariates in the model was used to estimate costs among patients with positive (non-zero) costs; combining predictions from both models generated the mean cost estimates for each of the study cohorts. Each model was fitted separately, and their predictions were combined to derive mean estimated costs.
A recycled predictions approach was then used to estimate absolute cost differences between groups with and without HZ (i.e., between the UC+/HZ+ and UC+/HZ-and between the CD+/HZ+ and CD+/HZ-cohorts), with 95% CIs estimated using non-parametric bootstrap procedures with 499 replications. For adjusted cost differences, 95% CIs were estimated using non-parametric bootstrap procedures with 499 replications.   a Adjusted incidence rate ratios (aIRRs) and 95% confidence intervals (CIs) were estimated using generalized linear models, assuming a negative binomial distribution and log link, adjusting for potential confounders using the patients' propensity scores to account for potential baseline differences between cohorts. P-values were calculated using the negative binomial distribution. b Incidence rates were calculated as the average number of HCRU events per patient. c Incidence rates were calculated by dividing the number of encounters over the observation period by the patient-time observed; the incidence rates were then reported on a PPPY basis.
Other resource use or ancillary care visits include home/hospice health visits, skilled nursing facilities, transportation services, and durable medical equipment. Abbreviations: aIRR, adjusted incidence rate ratio; CD, Crohn's disease; CI, confidence interval; ED, emergency department; HCRU, healthcare resource utilization; HZ, herpes zoster; PPPY, per person per year; UC: ulcerative colitis UC+/HZ+ and UC+/HZ-, patients with UC with and without HZ, respectively; CD+/HZ+ and CD+/HZ-, patients with CD with and without HZ, respectively a Healthcare costs for each of the study cohorts were estimated using a two-part model. (1) In the first part, the probability of observing a positive cost was modeled using logistic regression; (2) in the second part, a generalized linear model with a gamma distribution and log link was used to predict costs among patients with positive costs. Both models accounted for the patients' propensity scores of being in the UC+/HZ+ or CD+/HZ+ cohorts and relevant baseline characteristics. Mean adjusted cost differences between the cohorts with and without HZ were then estimated using a recycled predictions approach. b 95% CIs for adjusted cost difference comparisons were estimated from non-parametric bootstrap procedures with 499 replications. Other resource use costs include home health/hospice visits, skilled nursing facilities, and use of transportation services and durable medical equipment. Cost data were adjusted to 2020 US$ values using the medical care component of the US Consumer Price Index.