Posters at ASN Kidney Week 2023 explored the economic burden brought on by chronic kidney disease (CKD) and further validated the Klinrisk model’s ability to predict progressive forms of CKD.
Two posters presented at ASN Kidney Week 2023 analyzed the financial burden chronic kidney disease (CKD) can have on US patients and health systems, as well as assessed the efficacy of the Klinsk predictive model in commercial, Medicare Advantage, and Medicaid populations.
Economic Burden of CKD in the US
Over 35 million people are affected by CKD in the US. The economic impact of managing this disease is substantial as Medicare costs for these patients climbed to over $87 billion in 2019. In order to fully understand the economic burden of CKD on the health system and payers in the US—as well as what drives this burden—researchers conducted a systemic literature review to analyze the degree of direct medical costs experienced at the hands of this disease.
MEDLINE and Embase were used to identify relevant studies investigating the direct medical costs of CKD throughout the US. Characterizing these reports, researchers also sought to assess financial burden associated with CKD stage, comorbidities, and insurance type. Conference proceedings from 2020-2022 alongside published articles from January 2017-July 2022 were considered for this analysis. These studies were categorized as either CKD-broad (including overall CKD data not restricted to disease stage or present comorbidities) and CKD-specific (indicating CKD statistics restricted to disease stage or present comorbidities). The summarized costs were all converted to 2022 US dollar (USD) rates and presented as average annual costs per person.
In total, 52 citations across 39 district studies were included. Their samples sizes ranged from 52 individuals to over 7 million. Additionally, direct medical costs ranged from $6592-$280,727, with higher costs being linked to more severe CKD or the presence of cardiovascular disease (CVD) or diabetes.
The number one component affecting the CKD-broad population was inpatient costs (ranging from $2331-$$116,30), followed by outpatient costs. Additionally, dialysis for these patients could be as high as $13,248 while prescription drug costs could climb to $10,066.
For CKD-specific studies, outpatient costs constituted the highest financial burden. Annual costs for those with stage 3 CKD were reported at $6593 while those with end stage kidney disease could be as high as $143,745. Comorbid type 2 diabetes (T2D) and cardiovascular disease (CVD) led to costs as high as $280,727 and $70,742.
Utilizing the Klinrisk Model
Kidney function is measured with estimated glomerular filtration rate (eGFR), and until these levels indicate the majority loss of function, CKD can often go undiagnosed. The Klinrisk machine learning model has proven its ability to gather routine laboratory data to accurately predict CKD progression in patients. Early detection of CKD can help clinicians identify higher risk individuals, begin treatment interventions sooner, and dramatically improve patient outcomes. With this in mind, researchers set out to validate the efficacy of the Klinrisk model throughout commercial, Medicare Advantage, and Medicaid populations in the US.
The Klinrisk model analyzes factors such as age, sex, complete blood cell counts, metabolic panels, urinalysis, chemistry panels, among others to predict progressive forms of CKD (identified by kidney failure or a 40% decline in eGFR). Authors of the present study evaluated the model’s performance at 2 and 5 years post-index (defined as the patient’s first available serum creatinine result) for patients with or without urinalysis results (including albumin-t-creatinine ratio, protein-to-creatinine ratio, and semi-quantitative dipstick.
In total, 4,410,131 patients were included in the analysis. Populations from Medicare Advantage and Medicaid were made up of 341,666 and 93,056 individuals, respectively. Klinrisk performance was measured with discrimination (area under the receiver operating characteristic curve [AUC]), calibration plots, and Brier scores. The authors noted that discrimination results were very good overall regardless of a payor having a urinalysis result. Across all cohorts, AUC ranged from 0.80-0.83 at 2 years and 0.78-0.83 at 5 years. These ranges at 2 and 5 years were 0.81-0.87 and 0.80-0.87 if urinalysis data was available. For each combination of insurer type and urinalysis inclusion, Brier scores ranged from 0.068-0.075.
For these populations, the Klinsk model’s efficacy for predicting progressive forms of CKD was further validated. Being able to identify adults at risk for CKD—or who have it already—can provide a plethora of benefits patients. Among these benefits, authors conclude by emphasizing the potential to implement earlier intervention, delay disease progression, and even reduce health care costs utilizing the Klinsk model.
1. Rochon H, Osenenko KM, Chatterjee S, Donato BM. Economic burden of CKD in the United States a systematic literature review. Poster presented at ASN Kidney Week 2023; November 2-5; Philadelphia, PA.
2. Tangri N, Ferguson TW, Bamforth RJ. Eternal validation of the Klinrisk model in US commercial, Medicare Advantage, and Medicaid populations. Poster presented at ASN Kidney Week 2023; November 2-5; Philadelphia, PA.