The prevalence of cardiovascular disease by vascular bed and impact on healthcare costs in a large, real‐world population with type 2 diabetes

Abstract Introduction The purpose of this study was to assess prevalence of atherosclerotic cardiovascular disease (ASCVD) according to number of affected vascular beds and the impact on healthcare utilization and costs in persons with type 2 diabetes mellitus (type 2 DM) and established ASCVD. Methods In this retrospective, cross‐sectional analysis, adults with type 2 DM and ASCVD in a large US administrative claims database were categorized by number of ASCVD‐affected vascular beds (brain, heart, peripheral vasculature). Annual healthcare utilization and costs for 2015 were determined, including subgroup analyses by age group (18‐44, 45‐64, ≥65 years). Results Among 539 089 individuals with type 2 DM and ASCVD, 47.0% had ASCVD affecting >1 vascular bed. The most prevalent ASCVD diagnoses were acute coronary syndrome (26.6%), peripheral arterial disease (24.5%) and stroke (18.6%). Mean annual total healthcare costs per person increased with increasing number of vascular beds, from 1 ($17 741) to 2 ($25 877) to 3 ($33 412). A similar pattern of increased healthcare utilization with increasing number of vascular beds was observed. Among individuals with 1 affected vascular bed, mean total healthcare costs per person were comparable across age subgroups; however, if >1 vascular bed was affected, the mean total healthcare costs were highest in the youngest age cohort. Conclusions These real‐world data showed that almost half of individuals with type 2 DM and ASCVD had ASCVD affecting >1 vascular bed. A higher number of affected vascular beds were associated with higher mean total healthcare costs and utilization, with a disproportionate increase noted in younger relative to older people.


| INTRODUC TI ON
Type 2 diabetes mellitus (type 2 DM) is a highly prevalent disease and associated with a large economic burden for individuals and healthcare systems. 1,2 In the United States (US), the annual cost of diabetes is approximately $327 billion (2017 estimate). 3 Atherosclerotic cardiovascular disease (ASCVD) is a major cause of morbidity and mortality in type 2 DM 4,5 and contributes to the economic burden of type 2 DM. 6 ASCVD affecting multiple vascular beds (ie, brain, heart, peripheral vascular) has been associated with higher vascular event rates than single-bed disease. 7 In addition, people with diabetes have been reported to have increased plaque burden and more diffuse ASCVD, possibly due to mechanisms inherent to diabetes such as hyperglycaemia-induced increases in oxidative stress. 8,9 There are limited data regarding the healthcare cost impact of ASCVD based on the number of affected vascular beds, particularly in people with type 2 DM. In the large, international REduction of Atherothrombosis for Continued Health (REACH) Registry, data for 23 974 participants in the United States revealed a significant trend of higher hospitalization costs as the number of affected vascular beds increased. 10 To our knowledge, there have been no studies of this kind in a population with diabetes.
We have reported previously, using data from a large US claims database, that the presence of ASCVD was associated with significantly higher healthcare costs in persons with type 2 DM. 11 The current analysis, using the same real-world claims database, focused on adults with type 2 DM and established ASCVD and assessed the prevalence of ASCVD according to the number of affected vascular beds, as well as the impact on healthcare utilization and costs.

| Data source and study population
This was a retrospective, cross-sectional analysis using data from a large, nationwide US administrative claims database (IBM ® Family of The claims database population used in the current study has been described previously 11,12 Briefly, eligible individuals were aged ≥18 years on 1 January 2015 and had an established diagnosis of type 2 DM before 1 January 2015, defined as ≥2 diagnoses for type 2 DM, based on international classification of diseases, ninth revision (ICD-9) codes of 250.×0 or 250.×2 or ICD-10 codes of E11.xx or ≥1 type 2 DM diagnosis with ≥1 oral antidiabetes drug claim, and no more than 1 type 1 diabetes diagnosis according to ICD-9 (250.×1, 250.×3) or ICD-10 (E10.×) codes. Continuous health plan enrolment with an insurance plan containing both medical and pharmacy benefits between 1 January 2014 and 31 December 2015 was required.
The baseline period was defined as 1 January 2014 to 31 December 2014, and the study period was defined as 1 January 2015 to 31 December 2015.
This analysis included adults with type 2 DM and at least one ICD-9 diagnosis code for ASCVD prior to 1 January 2015 (see Supplemental Table 1  Each ASVCD diagnosis was classified as one of three vascular beds: brain (stroke, TIA), heart (ACS, angina, MI) or peripheral vasculature (PAD). Patients were then categorized by the number of vascular beds affected by their ASCVD diagnoses (ie, 1, 2 or 3 vascular beds).

| Data analysis
This was a descriptive analysis. Population characteristics were measured using counts with percentages for the categorical variables and means with standard deviation for continuous variables.

| Study population
Of 1 202 596 individuals with type 2 DM who met study eligibility criteria, 539 089 (44.8%) had one or more ASCVD-related diagnoses and were included in the analysis ( Table 1). The study cohort

| Healthcare costs and utilization
Mean total healthcare costs per person increased with increasing number of affected vascular beds, ranging from $17 741 in people with 1 affected vascular bed to $33 412 in those with 3 affected vascular beds, representing an increase of 88.3% (Table 3). Cost differences were impacted primarily by increasing medical costs.
Relative to people with 1 affected vascular bed only, the mean total annual medical costs per person were 59.7% and 117.4% higher for those with 2 or 3 affected vascular beds, respectively.
Pharmacy costs were also higher but to a much smaller degree Note: Because of rounding, some categories may differ by 0.1 from the total for that category.

TA B L E 3 Per person annual (2015) healthcare costs and utilization by number of vascular beds affected
Mean healthcare utilization was markedly higher in people with higher numbers of affected vascular beds (

| Subgroup analysis by age category
Mean healthcare cost comparisons across age groups (18-44, 45-64 and ≥65) revealed different patterns as the number of affected vascular beds increased ( Figure 2). Among individuals with ASCVD affecting only 1 vascular bed, mean total healthcare costs per person were similar across the 3 age subgroups; however, if 2 or 3 vascular beds were affected, the mean total healthcare cost was negatively associated with age (ie, costs were higher in younger age groups).
With regard to healthcare utilization by age group and number of affected vascular beds, mean annual inpatient and outpatient utilization rates were generally similar between age groups within the same vascular bed category (Figure 3). The most notable difference between age groups was observed for ER outpatient visit frequencies (ER visits not resulting in inpatient admission); regardless of number of affected vascular beds, the youngest age group had approximately double the number of mean annual ER outpatient visits compared with the other age groups in the same vascular bed category. In conclusion, this descriptive analysis of a large, nationwide, real-world sample of over a half-million adults with type 2 DM and ASCVD found that almost half of such individuals had ASCVD in multiple vascular beds. Further, mean total healthcare costs and utilization increased with increasing number of affected vascular beds. Despite the limitations of the study, these findings provide real-world insight regarding the clinical and economic burden of multivascular bed ASCVD in people with type 2 DM, which we hope stimulates additional research into this phenomenon.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from IBM/Truven. Restrictions apply to the availability of these data, which were used under licence for this study. Data are available Wayne Weng with the permission of IBM/Truven.