Abstract
Purpose:
Several investigators have suggested the cost-effectiveness of earlier screening, management of risk factors, and early treatment for diabetic retinopathy (DR). We aimed to evaluate the extent of health care utilization and cost of delayed care by insurance type in a vulnerable patient population.
Methods:
A retrospective analysis of patients with DR was conducted using electronic medical record (EMR) data from January 2014 to December 2020 at Denver Health Medical Center, a safety net institution. Patients were classified by disease severity and insurance status. DR-specific costs were assessed via Current Procedural Terminology (CPT) codes over a 24-month follow-up period.
Results:
Among the 313 patients, a higher proportion of non-English speaking patients were uninsured. Rates of proliferative DR at presentation differed across insurance groups (62% of uninsured, 42% of discount plan, and 33% of Medicare/Medicaid, P = 0.016). There was a significant difference in the total median cost between discount plan patients ($1258, interquartile range [IQR] = $0 – $5901) and both Medicare patients ($751, IQR = $0, $7148, P = 0.037) and Medicaid patients ($593, IQR = $0 – $6299, P = 0.025).
Conclusions:
There were higher rates of proliferative DR at presentation among the uninsured and discount plan patients and greater total median cost in discount plan patients compared to Medicare or Medicaid. These findings prioritize mitigating gaps in insurance coverage and barriers to preventative care among vulnerable populations.
Translational Relevance:
Advanced diabetic disease and increased downstream health care utilization and cost vary across insurance type, suggesting improved access to preventative care is needed in these specific at-risk populations.
The following demographic characteristics were collected: age, sex, race, primary language, and health plan type. The study population was classified into five patient groups based on health insurance status. Health insurance groups included Medicare, Medicaid, private, uninsured, and two discounted healthcare programs specific to low-income patients in Colorado. These two programs are the Colorado Indigent Care Program (CICP), providing discounted health care at participating hospitals and clinics for patients who do not qualify for Medicaid, and the Denver Financial Assistance Program (DFAP), which provides discounted health care at DHMC for patients ineligible for Medicaid or CICP. Whereas patients with CICP and DFAP are by definition uninsured, these groups were combined into their own insurance group for the purpose of statistical comparison. In this study, the uninsured group refers to patients without medical insurance who did not receive discounted care through CICP or DFAP.
Clinical characteristics determined at the initial visit were the presence of hypertension (HTN), HbA1c, duration of diabetes, and DR severity. Complications of DR assessed within the study timeframe included vitreous hemorrhage (VH), diabetic macular edema (DME), and neovascular glaucoma (NVG). DR severity was evaluated by an optometrist or ophthalmologist at the initial visit and stratified into five categories according to the International Clinical Disease Severity Scale for DR
20: (1) no retinopathy; (2) mild non-proliferative retinopathy (NPDR); (3) moderate NPDR; (4) severe NPDR; and (5) proliferative diabetic retinopathy (PDR).
Standard summary descriptive statistics were used to assess differences in demographic and clinical characteristics across patient insurance groups. Frequency and percentage of healthcare services and associated costs are provided by insurance group. Comparisons across groups were performed using the Chi-square or Fisher's exact test for categorical variables and ANOVA or Kruskal-Wallis test for continuous variables. The Kruskal-Wallis test, a rank-based nonparametric test, was used to compare costs across groups as cost was not normally distributed. The Wilcoxon rank sum test was used for pairwise comparisons between insurance groups to assess differences for our primary outcome of total cost. SAS version 9.4 was used for all statistical analyses (SAS Institute, Cary, NC, USA).
This retrospective study evaluated the relationship between health insurance status and the economic burden of DR in an underserved population. The findings described herein demonstrate that advanced diabetic disease and increased downstream health care utilization and cost vary by insurance type. Improved access to preventative care is needed in these specific at-risk populations.
In this study, health insurance was categorized into five groups, Medicaid, Medicare, private insurance, uninsured, and discount healthcare. Per standard Medicaid eligibility requirements, individuals in this group had limited annual incomes and were either disabled, pregnant, or financially responsible for children, elderly, or disabled family members.
22 The Medicare group was predominantly comprised of individuals 65 years and older in addition to younger individuals with certain health conditions, such as end-stage renal disease (ESRD) and amyotrophic lateral sclerosis (ALS), in accordance with standard Medicare enrollment criteria.
22,23 The private insurance group consisted of individuals in a safety net population, who typically cannot afford the cost of follow-up care. Individuals in the DFAP/CICP groups met eligibility criteria based on household income and proof of Denver residence. They tended to present with advanced DR and enrolled in a discount health care plan soon after presentation. Patients in the uninsured group did not qualify for discounted health care and were typically unable to return for further treatments due to high self-pay costs, despite high rates of presentation with advanced DR.
A high percentage of patients in the uninsured group were Spanish-speaking and Hispanic minorities. This group also had the greatest levels of PDR at the initial visit and VH during the study timeframe. Studies have shown that socioeconomic deprivation
24,25 and lack of insurance
26 are associated with increased prevalence of PDR
24,25 and more severe DR at initial presentation to care.
24,26 The uninsured group also had the lowest mean number of eye clinic visits in our study. Financial concerns,
27,28 lack of insurance,
28–30 decreased access to care,
28,29 and language and communication difficulties
29 are established barriers to receiving eye examinations and likely contributed to decreased utilization of eyecare in our uninsured group. Furthermore, Chan et al. described an increased risk of PDR and related complications in patients with fewer eyecare visits.
27 These findings may explain the higher prevalence of PDR and VH in our uninsured patients. Additionally, the presence of PDR is greater in patients with a longer duration of diabetes, elevated HbA1c, high blood pressure, and Hispanic race.
31–33 We did not find significantly elevated HbA1c, duration of diabetes, or prevalence of HTN in the uninsured compared to the other insurance groups in the present study.
The private insurance group had the second highest levels of PDR and VH and low rates of follow-up patient appointments during the study timeframe. There is limited research on the prevalence of PDR and VH in patients with DR with private insurance. Regarding follow-up eye care, a study on individuals with PDR showed patients with private and government insurance were more likely to be nonadherent to clinical recommendations for follow-up compared to self-pay patients.
34 In our study, we did not examine rates of follow-up eye care stratified by PDR status, thus it is difficult to compare our results with these findings. Moreover, given the small size of our private insurance group and specific safety net study population, our results may not be representative of most patients with private health insurance.
With respect to health care utilization and costs, individuals in the DFAP/CICP discount health care group had the highest number and total cost of eye clinic visits. Frequency and cost of injections were highest in the private insurance and DFAP/CICP groups. When stratified by the presence of PDR, costs to the healthcare system were markedly increased for all insurance groups compared to NPDR. The total cost for patients with PDR was greatest in DFAP/CICP, followed by the private insurance group, with similar costs for Medicaid and Medicare and the lowest cost in the uninsured. Individuals in the DFAP/CICP and private insurance groups cost the system the most given their initial poor glycemic control and ability to pay for discounted or covered health care, respectively. Costs for Medicaid and Medicare patients were lower than the discount and private insurance groups, presumably a result of receiving covered health care and ocular care and low rates of PDR relative to the other insurance groups. Costs for uninsured patients were limited as they were likely less able to follow-up due to the financial burden of health care.
There are few studies examining the economic burden of DR in the United States. Other studies investigating the cost of DR for Medicare
16,35,36 and privately insured patients
16,37 included the costs of inpatient care,
16,36,37 outpatient care,
16,35–37 emergency services,
37 prescription medications,
37 and other services attributable to visual impairment.
16,35–37 These findings are not comparable with our cost values as we only included DR specific costs of care at an eye clinic and did not include the cost of medications, inpatient care, emergency care, or other services. We also specifically examined a safety net population in a local community hospital in Denver, Colorado, which may differ from study populations in other research.
In 2019, the American Diabetes Association (ADA) recommended a comprehensive, dilated eye examination every 1 to 2 years if retinopathy is not present at the initial diagnosis of DM.
38 However, as of 2020, only 58.3% of adults with DM completed annual eye examinations.
39 Within the past 10 years, the use of tele-retinal screening for conditions such as DR has gained momentum.
40 Recent studies of tele-retinal DR screening have also demonstrated reduced costs for patients and the health care system.
41–43 Savings are attributed to reduced burden on eye clinics as patients without ocular findings do not need an in-person examination.
41,42,44,45 Furthermore, identifying pathology early in the disease course is expected to lead to decreased downstream health care costs and fewer individuals with DR related loss of vision, saving an estimated hundreds of millions of dollars for the federal government in health care and social security spending for the disabled and unemployed.
46 These findings demonstrate the cost-effectiveness of tele-retinal screening for DR and point to the necessity of earlier insurance coverage for screening and preventative treatment to reduce gaps in care for underserved populations and decrease financial burden on the healthcare system.
Supported in part by a challenge grant from Research to Prevent Blindness to the Department of Ophthalmology at the University of Colorado School of Medicine.
Disclosure: V. Rajeswaren, None; V. Lu, None; H. Chen, None; J.L. Patnaik, None; N. Manoharan, None