Skip to main content
NCBI home page
Kidney360 logoLink to Kidney360
. 2022 Dec 18;4(3):316–325. doi: 10.34067/KID.0005552022

Health Care Resource Utilization and Costs of Persistent Severe Acute Kidney Injury (PS-AKI) Among Hospitalized Stage 2/3 AKI Patients

Jay L Koyner 1, Rachel H Mackey 2,3,✉, Ning A Rosenthal 2, Leslie A Carabuena 2, J Patrick Kampf 4, Jorge Echeverri 5, Paul McPherson 4, Michael J Blackowicz 5, Toni Rodriguez 6, Aarti R Sanghani 6, Julien Textoris 7,8
PMCID: PMC10103312  PMID: 36996299

graphic file with name kidney360-4-316-g001.jpg

Open in a new tab

Keywords: Acute Kidney Injury and ICU Nephrology, costs, health care resource utilization, persistent severe acute kidney injury, severe acute kidney injury

Abstract

Key Points

  • Among hospitalized patients with stage 2/3 AKI, persistent severe acute kidney injury (PS-AKI) is associated with significantly longer length of stay (LOS) and higher costs during index hospitalization and 30 days postdischarge.

  • Relative differences in LOS and costs for PS-AKI versus NPS-AKI were similar for intensive care (ICU) and non-ICU patients.

  • Preventing PS-AKI among patients with stage 2/3 AKI may reduce hospital LOS and costs.

Background

Persistent severe acute kidney injury (PS-AKI) is associated with worse clinical outcomes, but there are no data on costs of PS-AKI. We compared costs and health care resource utilization for inpatients with PS-AKI versus not persistent severe AKI (NPS-AKI) overall and by ICU use.

Methods

This retrospective observational study included 126,528 adult US inpatients in the PINC AI Healthcare Database (PHD), discharged from January 1, 2017, to December 31, 2019, with KDIGO stage 2 or 3 AKI (by serum creatinine [SCr] criteria) during hospitalization, length of stay (LOS) ≥3 days, and ≥3 SCr measurements. Patients were categorized as PS-AKI (defined as stage 3 AKI lasting ≥3 days or with death within 3 days or stage 2/3 AKI (by SCr criteria) with dialysis within 3 days) or NPS-AKI. Generalized linear model regression compared LOS and costs during index hospitalization (total cohort) and 30 days postdischarge (survivors of index hospitalization), adjusted for patient, hospital, and clinical characteristics.

Results

Among 126,528 patients with stage 2/3 AKI, 30,916 developed PS-AKI. In adjusted models, compared with NPS-AKI, patients with PS-AKI had 32% longer total LOS (+3.3 days), 45% longer ICU LOS (+2.6 days), 46% higher total costs (+$13,143), 58% higher ICU costs (+$15,908), and during 30 days postdischarge 13% longer readmission LOS (+1.0 day), 22% higher readmission costs (+$4049), and 12% higher outpatient costs (+$206) (P<0.005 for all). Relative LOS and cost differences for PS-AKI versus NPS-AKI were similar for ICU (n=57,947) and non-ICU (n=68,581) patients.

Conclusions:

Among hospitalized patients with stage 2/3 AKI, PS-AKI was associated with significantly longer LOS and higher costs during index hospitalization and 30 days postdischarge, overall, and in ICU and non-ICU patients. Preventing PS-AKI among patients with stage 2/3 AKI may reduce hospital LOS and costs.

Introduction

AKI and higher severity (stage 2 and 3) AKI are associated with higher mortality, longer length of stay (LOS), and higher costs compared with no AKI.1–8 Longer AKI duration and persistent AKI (lasting ≥3 days) are also associated with adverse long-term outcomes compared with no AKI9-15 and compared with patients with not persistent severe (stage 2 or 3) AKI (NPS-AKI).16 Clinical trials and observational studies have demonstrated that early detection and prompt intervention can reduce AKI severity and duration17-19 and costs.20 However, adherence to best practices remains low for patients with21 or at risk for19 moderate or severe AKI. Quantifying the health care resource utilization (HRU) burden and costs associated with PS-AKI could provide powerful motivation to implement organization-level interventions to decrease AKI incidence, severity, and duration. However, few studies have evaluated HRU and costs related to AKI severity,2,8,22 and no studies have quantified the costs of persistent severe AKI (PS-AKI).

There are also no data on costs of PS-AKI among ICU and non-ICU patients, for whom we expect substantial difference in HRU and costs. And although few studies of PS-AKI have included non-ICU patients, two recent reports found that in-hospital mortality risk was multiplied similarly (four- to five-fold four-five-fold higher mortality) for ICU patients and non-ICU patients.16,23 Therefore, to fill these gaps in the literature, the objectives of this study were to evaluate PS-AKI–related HRU and incremental costs during index (initial) hospitalization and 30 days postdischarge among hospitalized patients with KDIGO stage 2/3 AKI, overall, and separately for patients with and without ICU use during index hospitalization.

Materials and Methods

Data Source, Study Timeline, Cohort, and Exposure

As previously described in detail,16 this retrospective observational study was conducted using the Premier PINC AI Healthcare Database (PHD), which contains inpatient and hospital-based outpatient discharge information from more than 1000 US hospitals and 20%–25% of all US inpatient admissions24-26 (Supplemental Methods). On the basis of US Title 45 Code of Federal Regulations, Part 46, this study of fully deidentified data was exempted from institutional review board approval, as for prior PHD studies.25-28

Qualifying hospitalizations included all hospitalizations of adult patients during the main study period, from January 1, 2017, through December 31, 2019, with a 30-day postdischarge follow-up for clinical outcomes and a 12-month look‐back to assess reference serum creatinine (SCr) and comorbidities (Supplemental Figure 1). Reference (baseline) SCr was defined as the lower of (1) the median of inpatient and outpatient SCr measurements during the prior 12 months and (2) the first SCr measurement during the qualifying hospitalization.29 Inclusion criteria included age ≥18 years, inpatient hospital LOS ≥3 days, and ≥3 SCr measurements during the hospitalization. Exclusions, defined using administrative codes (Supplemental Table 1), included receipt of extracorporeal membrane oxygenation, history of renal transplant or ESKD at admission or prior 12 months, ≥2 inpatient or outpatient visits with dialysis during prior 12 months, reference SCr ≥4 mg/dl, or reference eGFR <15 mL/min per 1.73 m2 (eGFR calculated using reference SCr) or stage 5 CKD present at admission.

Our primary study cohort included all patients with stage 2 or 3 AKI (by KDIGO SCr criteria), classified as PS-AKI or NPS-AKI (Supplemental Table 2), on the basis of prior studies of PS-AKI30,31 and AKI persistence.14,32,33 In brief, patients with PS-AKI were identified using the first qualifying hospitalization with (1) KDIGO stage 3 AKI by SCr criteria (SCr ≥3.0 times reference SCr or SCr ≥4.0 mg/dl) persisting for ≥3 days or with death or dialysis in ≤3 days or (2) KDIGO stage 2 AKI by SCr criteria (SCr ≥2.0 times reference SCr) with dialysis in ≤3 days. For sensitivity analysis, we also defined PS-AKISens to only include patients with stage 3 AKI (by SCr) lasting 3 or more days. By excluding patients who met PS-AKI criteria by death or dialysis within 3 days of AKI by SCr criteria, PS-AKISens allows sensitivity analysis to evaluate potential effects of including those patients in the definition of PS-AKI. Among patients without PS-AKI hospitalizations, patients with NPS-AKI were identified using the first qualifying hospitalization with SCr-based KDIGO stage 2 or 3 AKI (SCr ≥2.0 times reference SCr or SCr ≥4.0 mg/dl) that did not meet PS-AKI criteria.

Patient, Visit, Clinical, And Hospital Characteristics

Patient age, sex, race, ethnicity, primary insurance payer, admission type (elective, emergency, urgent, or information unavailable), point of origin (clinic, non-health care, etc.), and hospital bed size, geographic region, population served (urban versus rural), and teaching status were from the index hospitalization. MS-DRG codes classified patients as medical or surgical. International Classification of Diseases (ICD)-10-CM codes were used to assess comorbidities during index hospitalization and the Deyo-modified Charlson Comorbidity Index (CCI) score34,35 using characteristics from the index visit and prior 12 months (Supplemental Tables 3–4).

Outcomes

Primary outcomes during index hospitalization were total LOS and costs and ICU LOS and costs. Total LOS was hospital-reported LOS for the index hospitalization. As detailed (Supplemental Methods), total costs summed all direct hospital costs, including ICU costs, incurred by the health care provider during the index hospitalization which have been reconciled by Premier, as previously described.7,25 All costs were adjusted to 2019 US dollars using the Consumer Price Index. ICU LOS was calculated as number of days with ICU room and board charges during index hospitalization. ICU costs were calculated as the sum of all costs incurred on the days with ICU room and board charges. Secondary outcomes (readmission LOS, readmission costs, and outpatient costs) were ascertained during the 30 days after index discharge among survivors of index hospitalization. We also assessed ICU stay during index admission and 30-day postdischarge readmissions and outpatient visits.

Statistical Analyses

Baseline characteristics were compared using descriptive statistics. Continuous variables were expressed as means and standard deviations and compared by the t-test, and categorical variables were expressed as counts and percentages and compared by the χ2 test. Multivariable regression models were adjusted for the following potential confounders: age, sex, race-ethnicity, CCI score, hospital characteristics (number of beds, teaching status, region, and urban/rural), admission point of origin, admission type, medical versus surgical (categorized by MS-DRG codes), primary payer, ICU use, and the presence of CKD or sepsis. Logistic regression was used to calculate unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (95% CI) for ICU stay, readmissions, and outpatient visits. Generalized linear model (GLM) regression with a log link and negative binomial distribution was used to quantify unadjusted and adjusted LOS ratio (the ratio of PS-AKI LOS to NPS-AKI LOS) with 95% CIs.36 Unadjusted and adjusted mean cost ratios were quantified using a GLM with a log link and gamma distribution.36 For outcomes such as ICU costs, which are zero for non-ICU patients, we used a two-part model where, first, a logistic regression model estimated the probability of having an ICU stay (or readmission or outpatient visit), and second, for patients with nonzero costs, a GLM model was used to compare costs, as described above.36 As in prior studies,37 the adjusted absolute incremental difference in mean LOS and costs was calculated using the recycled prediction method with bootstrapped 95% confidence intervals36,38 (Supplemental Methods). Finally, we conducted sensitivity analysis to calculate relative LOS and relative costs for patients with PS-AKISens versus patients with NPS-AKI to evaluate potential effects of including patients with early death or dialysis after AKI in the definition of PS-AKI.

Results

Study Cohort

From January 1, 2017, through December 31, 2019, 15,586,984 unique patients aged ≥18 years had 23,075,687 hospital inpatient discharges (Supplemental Figure 2). After applying all inclusion and exclusion criteria, 1,951,240 discharges and 1,409,446 unique patients remained. In this index population, 126,528 patients had stage 2/3 AKI, forming the primary study cohort, among whom 30,916 patients (24%) developed PS-AKI and 95,612 patients (76%) developed NPS-AKI. In the primary study cohort, 57,947 (46%) of patients had an ICU stay during index hospitalization (ICU patients), among whom 18,858 (33%) developed PS-AKI. Of 68,581 patients with no index ICU stay (non-ICU patients), 12,058 (18%) developed PS-AKI. During index hospitalization, dialysis occurred in 40% of patients with PS-AKI overall, in 49.3% of patients with PS-AKI with an ICU stay and in 25.9% of patients with PS-AKI with no ICU stay.

Baseline Characteristics

Compared with NPS-AKI, patients with PS-AKI had a slightly lower mean age (65 versus 67 years), a lower proportion of women and White race, and a higher proportion with hospitalization in large teaching hospitals, urgent admission, transfer from an acute care facility, surgical (versus medical) diagnosis, sepsis, and in-hospital death during index admission (31.4% versus 8.4%) (Table 1). Patients with PS-AKI also had lower eGFR, higher mean CCI score (5.0 versus 4.3), and a higher prevalence of myocardial infarction, congestive heart failure, peripheral vascular disease, CKD, diabetes, and anemia than patients with NPS-AKI (Table 1).

Table 1.

Characteristics of patients with persistent severe acute kidney injury (PS-AKI) versus not persistent severe AKI (NPS-AKI), overall and for ICU patients and non-ICU patients

n Total Cohort (n=126,528) ICU Patients (n=57,947) Non-ICU Patients (n=68,581)
Not Persistent Severe AKI Persistent Severe AKI Not Persistent Severe AKI Persistent Severe AKI Not Persistent Severe AKI Persistent Severe AKI
 95,612 30,916 39,089 18,858 56,523 12,058
Age, yr, mean±SD 66.9±15.7 64.9±15.5 65.6±15.4 64.2±15.1 67.8±15.9 65.9±15.9
Age, n (%)            
 18–34 3878 (4.1) 1505 (4.9) 1714 (4.4) 960 (5.1) 2164 (3.8) 545 (4.5)
 34–49 8859 (9.3) 3364 (10.9) 3857 (9.9) 2065 (11.0) 5002 (8.8) 1299 (10.8)
 50–64 26,072 (27.3) 8936 (28.9) 11,312 (28.9) 5666 (30.0) 14,760 (26.1) 3270 (27.1)
 65–74 23,974 (25.1) 8116 (26.3) 10,290 (26.3) 5077 (26.9) 13,684 (24.2) 3039 (25.2)
 75+ 32,829 (34.3) 8995 (29.1) 11,916 (30.5) 5090 (27.0) 20,913 (37.0) 3905 (32.4)
Female sex, n (%) 50,047 (52.3) 12,873 (41.6) 18,891 (48.3) 7560 (40.1) 31,156 (55.1) 5313 (44.1)
Race-ethnicity, n (%)            
 White 69,996 (73.2) 21,457 (69.4) 28,454 (72.8) 13,104 (69.5) 41,542 (73.5) 8353 (69.3)
 Black 13,425 (14.0) 4939 (16.0) 5019 (12.8) 2816 (14.9) 8406 (14.9) 2123 (17.6)
 Hispanic 4820 (5.0) 1852 (6.0) 2275 (5.8) 1193 (6.3) 2545 (4.5) 659 (5.5)
 Other or unknown 7371 (7.7) 2668 (8.6) 3341 (8.5) 1745 (9.3) 4030 (7.1) 923 (7.7)
Payer, n (%)            
 Medicaid 12,168 (12.7) 4405 (14.2) 5336 (13.7) 2880 (15.3) 6832 (12.1) 1525 (12.6)
 Medicare 63,652 (66.6) 19,168 (62.0) 25,274 (64.7) 11,386 (60.4) 38,378 (67.9) 7782 (64.5)
 Private insurance 14,389 (15.0) 5383 (17.4) 6006 (15.4) 3359 (17.8) 8383 (14.8) 2024 (16.8)
 Other/unknown 5403 (5.7) 1960 (6.3) 2473 (6.3) 1233 (6.5) 2930 (5.2) 727 (6.0)
Hospital size (number of beds), n (%)            
 1–299 31,726 (33.2) 8509 (27.5) 12,404 (31.7) 4830 (25.6) 19,322 (34.2) 3679 (30.5)
 300–499 27,884 (29.2) 8946 (28.9) 10,855 (27.8) 5335 (28.3) 17,029 (30.1) 3611 (29.9)
 500+ 36,002 (37.7) 13,461 (43.5) 15,830 (40.5) 8693 (46.1) 20,172 (35.7) 4768 (39.5)
Teaching status, n (%)             
 Nonteaching 49,696 (52.0) 14,677 (47.5) 19,834 (50.7) 8696 (46.1) 29,862 (52.8) 5981 (49.6)
 Teaching 45,916 (48.0) 16,239 (52.5) 19,255 (49.3) 10,162 (53.9) 26,661 (47.2) 6077 (50.4)
Population served, n (%)             
 Rural 13,934 (14.6) 4070 (13.2) 6008 (15.4) 2446 (13.0) 7926 (14.0) 1624 (13.5)
 Urban 81,678 (85.4) 26,846 (86.8) 33,081 (84.6) 16,412 (87.0) 48,597 (86.0) 10,434 (86.5)
Geographic location, n (%)             
 Midwest 20,700 (21.7) 6902 (22.3) 8749 (22.4) 4511 (23.9) 11,951 (21.1) 2391 (19.8)
 Northeast 16,735 (17.5) 5185 (16.8) 5746 (14.7) 2750 (14.6) 10,989 (19.4) 2435 (20.2)
 South 54,596 (57.1) 17,527 (56.7) 23,117 (59.1) 10,805 (57.3) 31,479 (55.7) 6722 (55.7)
 West 3581 (3.7) 1302 (4.2) 1477 (3.8) 792 (4.2) 2104 (3.7) 510 (4.2)
Admission point of origin, n (%)            
 Clinic 5765 (6.0) 1754 (5.7) 2166 (5.5) 988 (5.2) 3599 (6.4) 766 (6.4)
 Non-health care facility 77,428 (81.0) 23,501 (76.0) 30,537 (78.1) 13,945 (73.9) 46,891 (83.0) 9556 (79.3)
 Transfer from a long-term care facility 3083 (3.2) 799 (2.6) 1251 (3.2) 424 (2.2) 1832 (3.2) 375 (3.1)
 Transfer from an acute care facility 9065 (9.5) 4780 (15.5) 5007 (12.8) 3443 (18.3) 4058 (7.2) 1337 (11.1)
 Other 271 (0.3) 82 (0.3) 128 (0.3) 58 (0.3) 143 (0.3) 24 (0.2)
Admission type, n (%)            
 Emergent 74,942 (78.4) 23,555 (76.2) 29,726 (76.0) 14,087 (74.7) 45,216 (80.0) 9468 (78.5)
 Trauma 670 (0.7) 231 (0.7) 415 (1.1) 204 (1.1) 255 (0.5) 27 (0.2)
 Urgent 10,489 (11.0) 4521 (14.6) 4675 (12.0) 2839 (15.1) 5814 (10.3) 1682 (13.9)
 Elective 8944 (9.4) 2407 (7.8) 4087 (10.5) 1633 (8.7) 4857 (8.6) 774 (6.4)
 Other/unknown 567 (0.6) 202 (0.7) 186 (0.5) 95 (0.5) 381 (0.7) 107 (0.9)
MS-DRG categorization, n (%)            
 Medical 68,572 (71.7) 19,795 (64.0) 24,430 (62.5) 10,809 (57.3) 44,142 (78.1) 8986 (74.5)
 Surgical 27,039 (28.3) 11,121 (36.0) 14,658 (37.5) 8049 (42.7) 12,381 (21.9) 3072 (25.5)
Discharge status, n (%)            
 Expired 8071 (8.4) 9711 (31.4) 5854 (15.0) 7910 (41.9) 2217 (3.9) 1801 (14.9)
 Home 28,559 (29.9) 5477 (17.7) 8686 (22.2) 2011 (10.7) 19,873 (35.2) 3466 (28.7)
 Home health 17,810 (18.6) 3907 (12.6) 6148 (15.7) 1648 (8.7) 11,662 (20.6) 2259 (18.7)
 Hospice 7965 (8.3) 2444 (7.9) 3638 (9.3) 1395 (7.4) 4327 (7.7) 1049 (8.7)
 Nursing or rehabilitation facility 28,896 (30.2) 8041 (26.0) 12,598 (32.2) 5039 (26.7) 16,298 (28.8) 3002 (24.9)
 Transferred to an acute care facility 557 (0.6) 141 (0.5) 257 (0.7) 99 (0.5) 300 (0.5) 42 (0.3)
 Other/unknown 3754 (3.9) 1195 (3.9) 1908 (4.9) 756 (4.0) 1846 (3.3) 439 (3.6)
Sepsis 28,821 (30.1) 13,711 (44.3) 17,355 (44.4) 10,821 (57.4) 11,466 (20.3) 2890 (24.0)
Index serum creatinine,a mg/dl 2.0±1.5 2.8±2.4 1.8±1.4 2.5±2.2 2.1±1.5 3.4±2.6
Reference serum creatinine, mg/dl 1.0±0.6 1.5±0.9 1.0±0.5 1.4±0.8 1.1±0.6 1.6±1.0
Index eGFRb, ml/min per 1.73 m2 52.5±42.9 43.9±41.2 56.2±43.1 47.9±40.6 49.9±42.6 37.8±41.5
Reference eGFRc, ml/min per 1.73 m2 82.0±43.1 65.3±43.3 83.9±43.2 67.0±42.9 80.6±43.0 62.6±43.7
Dialysis, n (%) 383 (0.4) 12,422 (40.2) 306 (0.8) 9295 (49.3) 77 (0.1) 3127 (25.9)
History of comorbidities, n (%)            
 Charlson comorbidity score 4.3±3.0 5.0±3.1 4.2±2.9 4.8±3.0 4.3±3.1 5.2±3.2
 Myocardial infarction 19,800 (20.7) 7674 (24.9) 9777 (25.0) 5341 (28.3) 10,023 (17.7) 2333 (19.5)
 Congestive heart failure 37,921 (39.7) 14,027 (45.5) 16,956 (43.4) 8989 (47.7) 20,965 (37.1) 5038 (42.1)
 Peripheral vascular disease 12,756 (13.3) 4632 (15.0) 5617 (14.4) 2905 (15.4) 7139 (12.6) 1727 (14.4)
 Diabetes 44,704 (46.8) 15,658 (50.6) 18,718 (47.9) 9265 (49.1) 25,986 (46.0) 6393 (53.0)
 Hypertension 80,836 (84.5) 26,111 (84.5) 32,720 (83.7) 15,668 (83.1) 48,116 (85.1) 10,443 (86.6)
 Chronic kidney disease 36,669 (38.4) 17,196 (55.6) 13,970 (35.7) 9658 (51.2) 22,699 (40.2) 7538 (62.5)
 Anemia 50,545 (52.9) 20,268 (65.6) 20,298 (51.9) 11,649 (61.8) 30,247 (53.5) 8619 (71.5)
Open in a new tab

Continuous variables were compared using the Student t-test. Categorical variables were compared using the chi-square test. P<.001 for all within-group comparisons versus NPS-AKI, except P = 0.005 for diabetes in the ICU subcohort, and hypertension is not significantly different in overall cohort or in ICU subcohorts. sCr, serum creatinine.

a

Index SCr is the first SCr measure during the index hospitalization while reference is the creatinine used as the baseline to calculate AKI status.

b

eGFR calculated using index SCr.

c

eGFR calculated using reference SCr.

Comparisons between PS-AKI and NPS-AKI were similar among ICU patients and non-ICU patients (Table 1). However, overall, ICU patients seemed to have slightly lower mean age and proportion of women, similar or lower CCI score, and higher proportion of hospitalization in large, teaching hospitals, urgent or elective admissions, transfer from an acute care facility, surgical patients, and sepsis, and substantially higher in-hospital death during index admission than non-ICU patients (Table 1).

Health Care Resource Utilization during Index Hospitalization and 30 Days Postdischarge

In adjusted models, the odds of ICU use during index hospitalization were higher for patients with PS-AKI than NPS-AKI (OR, 2.15; 95% CI, 2.09 to 2.21; P<0.001) (Supplemental Table 5). Among survivors of index hospitalization, odds of 30-day postdischarge readmission were slightly higher for (former) patients with PS-AKI than NPS-AKI in the overall cohort (OR, 1.07; 95% CI, 1.02 to 1.11; P = 0.002) and in non-ICU patients but not in ICU patients. Finally, the odds of 30-day postdischarge outpatient visits were slightly lower for (former) patients with PS-AKI than NPS-AKI in the overall cohort (OR, 0.95; 95% CI, 0.91 to 0.98; P = 0.003) and in ICU patients but not in non-ICU patients.

Unadjusted mean total LOS, ICU LOS, and readmission LOS were longer for patients with PS-AKI than NPS-AKI (Figure 1, P<0.0001 for all). In adjusted models, the mean total LOS was 32% longer (+3.3 days), ICU LOS was 45% longer (+2.6 days), and 30-day postdischarge readmission LOS was 13% longer (+1.0 days) for patients with PS-AKI than NPS-AKI. For both ICU and non-ICU patients, patients with PS-AKI had longer unadjusted mean LOS for index hospitalization and 30-day postdischarge readmission (Figure 2, P<0.0001 for all). For both ICU and non-ICU patients, adjusted absolute and relative differences (LOS ratios) for PS-AKI versus NPS-AKI were similar to each other and to the overall cohort (Figure 2). Specifically, for patients with PS-AKI versus patients with NPS-AKI, index hospitalization LOS was 24% (+3.2 days) and 41% (+3.4 days) longer, and 30-day postdischarge readmission LOS was 14% (+1.2 days) and 11% (+0.9 days) longer among ICU and non-ICU patients, respectively.

Figure 1.

Figure 1

Open in a new tab

Mean length of stay (LOS) and adjusted absolute and relative LOS differences for patients with PS-AKI versus NPS-AKI. *P<0.001 for all comparisons versus NPS-AKI (reference). Covariates for adjusted model are age, sex, race-ethnicity, Charlson Comorbidity Index, hospital characteristics (number of beds, teaching status, region, urban/rural) and admission point of origin, admission type, medical versus surgical patient (categorized by MS_DRG), primary payer and CKD, sepsis, and ICU use.

Figure 2.

Figure 2

Open in a new tab

Mean length of stay (LOS) and adjusted absolute and relative LOS differences for PS-AKI versus NPS-AKI for patients with and without ICU stay during index hospitalization. *P<0.001 for all comparisons versus NPS-AKI (reference). Covariates for adjusted model are age, sex, race-ethnicity, Charlson Comorbidity Index, hospital characteristics (number of beds, teaching status, region, urban/rural) and admission point of origin, admission type, medical versus surgical patient (categorized by MS_DRG), primary payer and CKD, sepsis, and ICU use.

Patients with PS-AKI had higher unadjusted mean total and ICU costs during index admission and higher 30-day readmission and outpatient costs than patients with NPS-AKI (Figure 3, P<0.0001 for all). In adjusted models, patients with PS-AKI had 46% higher total costs (+$13,143), 58% higher ICU costs (+$15,908), 22% higher 30-day readmission costs (+$4049), and 12% higher 30-day outpatient costs (+$206) than patients with NPS-AKI. Stratified by ICU use, PS-AKI was associated with higher adjusted total costs and 30-day readmission costs among ICU and non-ICU patients but higher 30-day outpatient costs only among non-ICU patients (Figure 4, P<0.0001 for all). Furthermore, relative differences (cost ratios) were similar for ICU and non-ICU patients, although mean absolute cost differences between PS-AKI and NPS-AKI were larger for ICU patients. For example, PS-AKI was associated with 44% (+$18,687) and 46% (+$7944) higher total costs among ICU patients and non-ICU patients, respectively. Finally, in our sensitivity analyses (Supplemental Table 6), unadjusted and adjusted relative LOS and relative cost differences for patients with PS-AKISens (n=17,285) versus NPS-AKI were very similar to those for PS-AKI versus NPS-AKI.

Figure 3.

Figure 3

Open in a new tab

Mean costs and adjusted absolute and relative cost differences for PS-AKI versus NPS-AKI. *P<0.005 for all comparisons versus NPS-AKI (reference). US dollars (USD). Covariates for adjusted model are age, sex, race-ethnicity, Charlson Comorbidity Index, hospital characteristics (number of beds, teaching status, region, urban/rural) and admission point of origin, admission type, medical versus surgical patient (categorized by MS_DRG), primary payer and CKD, sepsis, and ICU use. Models for ICU costs do not adjust for ICU use or sepsis.

Figure 4.

Figure 4

Open in a new tab

Mean costs and adjusted absolute and relative cost differences for PS-AKI versus NPS-AKI for patients with and without ICU stay during index hospitalization. *P<0.001 for all comparisons versus NPS-AKI (reference), except for 30-day outpatient costs among patients with an ICU stay. US dollars (USD). Covariates for adjusted model are age, sex, race-ethnicity, Charlson Comorbidity Index, hospital characteristics (number of beds, teaching status, region, urban/rural) and admission point of origin, admission type, medical versus surgical patient (categorized by MS_DRG), primary payer and CKD, sepsis, and ICU use. Models for ICU costs do not adjust for ICU use or sepsis.

Discussion

This is the first study to evaluate the incremental HRU burden and costs of PS-AKI. Among 126,528 hospitalized US adults with KDIGO stage 2/3 AKI and LOS ≥3 days, nearly one in four developed PS-AKI. PS-AKI was associated with significantly longer LOS and higher costs during index hospitalization and during 30-day postdischarge follow-up. Longer relative LOS and higher relative costs were only modestly attenuated when adjusted for many potential confounders including ICU stay, comorbidity score, sepsis, and CKD. Furthermore, relative differences in LOS and costs for PS-AKISens, which excluded patients who met PS-AKI criteria by early death or dialysis after AKI, were similar to results comparing PS-AKI versus NPS-AKI. Notably, although ICU patients had longer LOS and higher total costs and larger absolute differences in PS-AKI–related total costs, relative (ratio) cost differences and both absolute and relative LOS differences were broadly similar for patients with and without an ICU stay during index hospitalization.

In our study, the average patient with PS-AKI had 32% longer (+3.3 days) LOS and 46% higher (+$13,143) total costs and 45% longer ICU LOS (+2.6 days) and 58% higher ($15,908) ICU costs compared with similar patients with NPS-AKI. The similarity between relative differences in LOS and costs suggests that longer LOS is a major contributor to higher costs. The relative increase in LOS and costs for PS-AKI versus NPS-AKI in our study is smaller than the roughly doubled LOS and total hospitalization costs reported by other studies for patients with AKI compared with patients with no AKI.6,8 However, the reference group in our study is patients with stage 2/3 AKI (by SCr criteria) not meeting criteria for PS-AKI, not (lower risk) patients with no AKI. Furthermore, the incremental absolute difference is closer to studies of HRU and costs related to higher AKI stage versus no AKI.2,8,39-41 Among 31,970 hospitalizations at a single US academic medical center in 2010, hospitalizations with KDIGO stage 2 or 3 AKI had +4.2 and +6.4 days longer LOS (+2.2 days longer for stage 3 versus 2), and +$15,200 and +$27,300 higher incremental costs (+$12,100 higher for stage 3 versus 2), respectively, compared with no AKI.2 A large study from China reported that KDIGO AKI stages 1, 2, and 3 were associated with 22%, 25%, and 32% longer LOS and 6%, 15%, and 33% higher daily cost, respectively, versus no AKI.5 Although no prior studies have evaluated costs for PS-AKI versus NPS-AKI, the higher 30-day readmission and outpatient costs and longer LOS in our study are consistent with a study of Canadian adults, showing higher 3–12 month postdischarge costs (in 2015 Canadian dollars): for patients with any stage AKI with recovery (+$2912 to $3231) and for any stage AKI without kidney recovery (+$6035 to $8563) compared with patients with no AKI.22

In the current study, during 30 days postdischarge among survivors of index hospitalization, patients with PS-AKI had 13% (+1 day) longer mean 30-day readmission LOS, 22% (+$4049) higher 30-day readmission costs, and 12% (+$206) higher 30-day outpatient costs than patients with NPS-AKI. Relative differences for patients with PS-AKI versus NPS-AKI were similar for ICU and non-ICU patients except for 30-day outpatient costs, which were higher for ICU patients (+20%, +$338). The small differences in 30-day postdischarge readmissions and outpatient visits may be due to chance but also to differences in index hospitalization discharge status. Patients with PS-AKI were much less likely than patients with NPS-AKI to be discharged to home/home health (30.3% versus 48.5%), which was much lower for ICU patients (19.4% versus 37.9%) than non-ICU patients (47.4% versus 55.8%). Therefore, among patients with ICU stay and PS-AKI during index hospitalization only 19.4% were available for 30-day postdischarge outpatient visits, and this group was likely affected by survival bias (i.e., only the healthiest of the ICU patients with PS-AKI would be included). Finally, since 30-day postdischarge outcomes are captured only for visits back to the same hospital, some outcomes may have been missed. However, assuming that underascertainment of readmissions or outpatient visits is nondifferential between groups, comparisons of mean 30-day postdischarge LOS and costs should be unaffected.

Our study has several strengths including the very large, diverse sample of recent hospitalizations available from the PHD, which as noted, represents approximately 20%–25% of US inpatient admissions and also includes outpatient visits to emergency departments, ambulatory surgery centers, and other hospital-based care. Although hospital administrative data contain less clinical detail than electronic health records, its use for research is well established.42-44 Another study strength was the availability of SCr laboratory data, which allowed us to define PS-AKI and NPS-AKI using the KDIGO SCr-based criteria, rather than ICD-10 codes as in prior AKI studies using administrative data.42-44 Finally, a novel strength of our study is our sensitivity definition of PS-AKISens, which shows that relative differences in LOS and costs for PS-AKI versus NPS-AKI were not due to the inclusion of patients with PS-AKI with early death or dialysis.

Study limitations include the inability to track postdischarge outcomes outside of the hospital system of the index hospitalization, noted above, but also potential misclassification of PS-AKI because of limitations of our database for identifying reference SCr. We defined reference SCr on the basis of prior research and recommendations,29 and because most patients did not have outpatient SCr measures during the 12 months before index hospitalization, we used the median of inpatient and outpatient SCr during the prior 12 months to minimize the effects of potential outliers from prior hospitalizations. Finally, to avoid bias caused by excluding patients with no SCr measures during the prior 12 months,29 we used the lower of (1) the median of historical inpatient and outpatient SCr and (2) the first SCr during index admission. Another limitation is that as in other large AKI studies,2,11,12,33 we did not have urine output (UO) data available to define AKI. However, because clinical outcomes are worse for patients with AKI identified by both UO and SCr,45 any misclassification of patients with PS-AKI (because of absence of UO) as patients with NPS-AKI should have reduced differences in clinical outcomes which should also reduce differences in LOS and costs.

Another study limitation is the lack of etiology of AKI. Although we have reported information on comorbidities and procedures such as sepsis and cardiac surgery and adjusted for multiple confounders, administrative codes (e.g., ICD-10 diagnosis codes) may underestimate disease states such as sepsis. However, the magnitude of the unadjusted and adjusted differences in our study makes it unlikely that imperfectly measured or unmeasured confounders would fully explain the differences in LOS and costs for patients with PS-AKI versus NPS-AKI. To fairly compare patients with the opportunity to develop PS-AKI, our inclusion criteria required LOS ≥3 days and ≥3 SCr measurements. Requiring ≥3 SCr measurements excluded a high proportion of patients because only a subset of hospitals submit laboratory data to PHD. However, hospital reporting of laboratory data should be unrelated to PS-AKI status and, therefore, should not bias our study cohort or affect the comparisons of PS-AKI versus NPS-AKI among the more than 125,000 cases of stage 2 or 3 AKI in our study. Finally, as with any observational study, this retrospective observational study design cannot prove causality.

In conclusion, among 126,528 US inpatients with stage 2/3 AKI, SCR measures and LOS ≥3 days, PS-AKI was associated with significantly longer total and ICU LOS and significantly higher total and ICU costs during index hospitalization and during 30-day postdischarge follow-up, with higher readmissions, longer readmission LOS, and higher readmission costs, before and after adjusting for a wide range of potential confounders. Furthermore, although ICU patients had higher total costs and mean cost difference during index admission than non-ICU patients, ratios of costs and LOS for PS-AKI versus NPS-AKI were generally similar for ICU patients and non-ICU patients. These study results suggest that preventing PS-AKI among hospitalized patients who develop stage 2/3 AKI, both in and out of the ICU, could reduce HRU, LOS, and costs and provide support for interventions to reduce AKI duration and severity.

Supplementary Material

SUPPLEMENTARY MATERIAL
kidney360-4-316-s001.pdf (549.5KB, pdf)

Acknowledgments

Cate Polacek, MLIS, Senior Medical Writer employed by Premier Inc., provided manuscript editing support.

The findings of this study were presented in part at the 40th ISICEM (International Symposium on Intensive Care & Emergency Medicine) conference (Brussels, Belgium) 8/31/21.

Footnotes

J.L.K. and R.H.M. contributed equally to this work.

Disclosures

M.J. Blackowicz reports the following: Employer: Baxter; Alexion Pharmaceuticals; and Ownership Interest: Baxter. L.A. Carabuena reports the following: Employer: Premier Inc; Ownership Interest: United Health; and Research Funding: As an employee of Premier, Inc, I have contributed to research projects funded by BioMerieux, Inc, Baxter Healthcare, Inc, AstraZeneca, Amgen, and Alexion. Premier, Inc. receives and manages the funds. J. Echeverri reports the following: Employer: Baxter; Ownership Interest: Baxter; and Honoraria: Baxter. J. Patrick Kampf reports the following: Employer: Astute Medical, Inc. (a bioMerieux company); Ownership Interest: bioMerieux; and Patents or Royalties: Astute Medical, Inc. (a bioMerieux company). J.L. Koyner reports the following: Employer: University of Chicago; Consultancy: Astute Medical/Biomerieux Baxter, Novartis, Mallinckrodt, SeaStar,; Research Funding: Astute Medical; Nxstage medical; Fresenius Medical; NIH; Honoraria: American Society of Nephrology; ISICEM; CSCTR, Acute Disease Quality Initiative (ADQI); Patents or Royalties: Listed on a patent for Pi GST to detect severe AKI following cardiac surgery—with Argutus Medical; Advisory or Leadership Role: Editorial Board of Clinical Journal of American Society of Nephrology (CJASN), American Journal of Nephrology and Kidney360, Scientific Ad Board for the NKF of Illinois; Guard Therapuetics, Novartis; and Speakers Bureau: NxStage Medical.J.L. Koyner reports receiving research fees from bioMerieux and consulting fees from Baxter and bioMerieux. RHM, NAR, and LAC are full-time employees of Premier, Inc., which received payment from bioMerieux to conduct the study, and have no competing interests with respect to the study. R.H. Mackey reports the following: Employer: Premier, Inc.; Research Funding: As an employee of Premier, Inc., I have led research projects funded by bioMerieux, Inc., Baxter Healthcare, Inc., AstraZeneca, Amgen, and Alexion. Premier, Inc. receives and manages the funds.; and Other Interests or Relationships: I have the following unpaid activities: I am a Fellow and volunteer member of the American Heart Association, a volunteer member of the National Lipid Association and ISPOR, and an Associate Editor of the Journal of Clinical Lipidology. P. McPherson reports the following: Employer: Astute Medical Inc (a bioMerieux company); Ownership Interest: bioMerieux; Patents or Royalties: Astute Medical Inc (a bioMerieux company); and Advisory or Leadership Role: Astute Medical Inc (a bioMerieux company). T. Rodriguez reports the following: Employer: bioMerieux; and Research Funding: bioMerieux. N.A. Rosenthal reports the following: Employer: PINC AI Applied Sciences, Premier Inc.; Scopely; Ownership Interest: PINC AI Applied Sciences, Premier Inc.; and Research Funding: All of our research projects are funded by life science companies. A.R. Sanghani reports the following: Employer: Grifols Shared Services North America, Inc.; bioMerieux, Inc. (2020–2022). J. Textoris reports the following: Employer: bioMérieux; Ownership Interest: bioMérieux; Research Funding: bioMérieux; Patents or Royalties: the company I work for holds patent related to diagnostics, in particular in the field of the host immune response. I am an inventor listed on some of them.; and Advisory or Leadership Role: I am employee of a diagnostic company, bioMérieux, and part of the leadership team (as VP, EME Medical Affairs). P. McPherson, J. Patrick Kampf, A.R. Sanghani, J. Textoris, and T. Rodriguez are full-time employees of bioMerieux. M.J. Blackowicz and J. Echeverri are full-time employees of Baxter International with ownership interests. The authors report no other conflicts of interest with this work.

Funding

This study was funded by bioMerieux, Inc. Design and conduct of the study were led by Premier coauthors, with collaboration from the entire study team consisting of investigators from Premier, Inc., the University of Chicago, Baxter Healthcare Corporation, and bioMerieux, Inc.

Author Contributions

Conceptualization: Michael J. Blackowicz, Jorge Echeverri, J. Patrick Kampf, Jay L. Koyner, Rachel H. Mackey, Paul McPherson, Toni Rodriguez, Ning A. Rosenthal, Aarti R. Sanghani, Julien Textoris.

Data curation: Leslie A. Carabuena, Rachel H. Mackey.

Formal analysis: Leslie A. Carabuena, Rachel H. Mackey.

Funding acquisition: Rachel H. Mackey, Ning A. Rosenthal.

Investigation: Michael J. Blackowicz, Leslie A. Carabuena, Jorge Echeverri, J. Patrick Kampf, Jay L. Koyner, Rachel H. Mackey, Paul McPherson, Toni Rodriguez, Ning A. Rosenthal, Aarti R. Sanghani, Julien Textoris.

Methodology: Leslie A. Carabuena, Rachel H. Mackey, Ning A. Rosenthal.

Project administration: Rachel H. Mackey, Ning A. Rosenthal.

Resources: Ning A. Rosenthal.

Software: Leslie A. Carabuena, Rachel H. Mackey.

Supervision: Rachel H. Mackey, Ning A. Rosenthal.

Validation: Michael J. Blackowicz, Leslie A. Carabuena, Jorge Echeverri, J. Patrick Kampf, Jay L. Koyner, Rachel H. Mackey, Paul McPherson, Toni Rodriguez, Ning A. Rosenthal, Aarti R. Sanghani, Julien Textoris.

Visualization: Leslie A. Carabuena, Rachel H. Mackey.

Writing – original draft: Rachel H. Mackey.

Writing – review & editing: Michael J. Blackowicz, Leslie A. Carabuena, Jorge Echeverri, J. Patrick Kampf, Jay L. Koyner, Paul McPherson, Toni Rodriguez, Ning A. Rosenthal, Aarti R. Sanghani, Julien Textoris.

Data Sharing Statement

The data that support the findings of this study are available from Premier Inc., but restrictions apply to the availability of these data, which were used under license for the current study and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Premier Inc (https://www.pinc-ai.com/applied-sciences).

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/KN9/A253.

Supplemental Methods.

Supplemental Figure 1. Study timeline Supplemental.

Supplemental Figure 2. Patient flow diagram.

Supplemental Table 1. ICD-10 diagnosis and procedure and CPT codes for conditions to exclude, by condition and time period.

Supplemental Table 2. Definitions of PS-AKI and NPS-AKI.

Supplemental Table 3. ICD-10 diagnosis codes for sepsis, chronic kidney disease (CKD), hypertension, and anemia.

Supplemental Table 4. Charlson comorbidity index diagnosis and procedure codes.

Supplemental Table 5. Unadjusted and adjusted comparison of clinical outcomes for persistent severe AKI* vs. not persistent severe AKI, overall and for ICU and non-ICU patients separately.

Supplemental Table 6. Comparison of unadjusted and multivariable-adjusted LOS ratios and cost ratios for PS-AKI or PS-AKISens vs. NPS-AKI.

References

  • 1.Selby NM, Crowley L, Fluck RJ, et al. Use of electronic results reporting to diagnose and monitor AKI in hospitalized patients. Clin J Am Soc Nephrol. 2012;7:533–540. doi: 10.2215/CJN.08970911 [DOI] [PubMed] [Google Scholar]
  • 2.Zeng X, McMahon GM, Brunelli SM, Bates DW, Waikar SS. Incidence, outcomes, and comparisons across definitions of AKI in hospitalized individuals. Clin J Am Soc Nephrol. 2014;9:12–20. doi: 10.2215/CJN.02730313 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Ülger F, Pehlivanlar Küçük M, Küçük AO, et al. Evaluation of acute kidney injury (AKI) with RIFLE, AKIN, CK, and KDIGO in critically ill trauma patients. Eur J Trauma Emerg Surg. 2018;44:597–605. doi: 10.1007/s00068-017-0820-8 [DOI] [PubMed] [Google Scholar]
  • 4.Luo X, Jiang L, Du B, Wen Y, Wang M, Xi X. A comparison of different diagnostic criteria of acute kidney injury in critically ill patients. Crit Care. 2014;18:R144. doi: 10.1186/cc13977 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Xu X, Nie S, Liu Z, et al. Epidemiology and clinical correlates of AKI in Chinese hospitalized adults. Clin J Am Soc Nephrol. 2015;10:1510–1518. doi: 10.2215/CJN.02140215 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Silver SA, Chertow GM. The economic consequences of acute kidney injury. Nephron. 2017;137:297–301. doi: 10.1159/000475607 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Prasad A, Rosenthal NA, Kartashov A, Knish K, Dreyfus J. Contemporary trend of acute kidney injury incidence and incremental costs among US patients undergoing percutaneous coronary procedures. Catheter Cardiovasc Interv. 2020;96:1184-1197. doi: 10.1002/ccd.28824 [DOI] [PubMed] [Google Scholar]
  • 8.Dasta JF, Kane-Gill S. Review of the literature on the costs associated with acute kidney injury. J Pharm Pract. 2019;32:292–302. doi: 10.1177/0897190019852556 [DOI] [PubMed] [Google Scholar]
  • 9.Mehta S, Chauhan K, Patel A, et al. The prognostic importance of duration of AKI: a systematic review and meta-analysis. BMC Nephrol. 2018;19:91. doi: 10.1186/s12882-018-0876-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Brown JR, Kramer RS, Coca SG, Parikh CR. Duration of acute kidney injury impacts long-term survival after cardiac surgery. Ann Thorac Surg. 2010;90:1142–1148. doi: 10.1016/j.athoracsur.2010.04.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Goldberg A, Kogan E, Hammerman H, Markiewicz W, Aronson D. The impact of transient and persistent acute kidney injury on long-term outcomes after acute myocardial infarction. Kidney Int. 2009;76:900–906. doi: 10.1038/ki.2009.295 [DOI] [PubMed] [Google Scholar]
  • 12.Pannu N, James M, Hemmelgarn B, Klarenbach S; Alberta Kidney Disease Network. Association between AKI, recovery of renal function, and long-term outcomes after hospital discharge. Clin J Am Soc Nephrol. 2013;8:194–202. doi: 10.2215/CJN.06480612 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Coca SG, King JT, Jr, Rosenthal RA, Perkal MF, Parikh CR. The duration of postoperative acute kidney injury is an additional parameter predicting long-term survival in diabetic veterans. Kidney Int. 2010;78:926–933. doi: 10.1038/ki.2010.259 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Perinel S, Vincent F, Lautrette A, et al. Transient and persistent acute kidney injury and the risk of hospital mortality in critically ill patients: results of a multicenter cohort study. Crit Care Med. 2015;43:e269–e275. doi: 10.1097/CCM.0000000000001077 [DOI] [PubMed] [Google Scholar]
  • 15.Heung M, Steffick DE, Zivin K, et al. ; Centers for Disease Control and Prevention CKD Surveillance Team. Acute kidney injury recovery pattern and subsequent risk of CKD: an analysis of veterans health administration data. Am J Kidney Dis. 2016;67:742–752. doi: 10.1053/j.ajkd.2015.10.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Koyner JL, Mackey RH, Rosenthal NA, et al. Clinical outcomes of persistent severe acute kidney injury among patients with KDIGO stage 2 or 3 AKI [published online ahead of print]. Am J Nephrol. 2022. doi: 10.1159/000528158 [DOI] [PubMed] [Google Scholar]
  • 17.Kolhe NV, Reilly T, Leung J, et al. A simple care bundle for use in acute kidney injury: a propensity score-matched cohort study. Nephrol Dial Transpl. 2016;31:1846–1854. doi: 10.1093/ndt/gfw087 [DOI] [PubMed] [Google Scholar]
  • 18.Selby NM, Casula A, Lamming L, et al. An organizational-level program of intervention for AKI: a pragmatic stepped Wedge cluster randomized trial. J Am Soc Nephrol. 2019;30:505–515. doi: 10.1681/ASN.2018090886 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Zarbock A, Küllmar M, Ostermann M, et al. Prevention of cardiac surgery-associated acute kidney injury by implementing the KDIGO guidelines in high-risk patients identified by biomarkers: the PrevAKI-multicenter randomized controlled trial. Anesth Analg. 2021;133:292–302. doi: 10.1213/ane.0000000000005458 [DOI] [PubMed] [Google Scholar]
  • 20.Selby NM, Korrodi-Gregorio L, Casula A, et al. Randomized controlled trial evidence of cost-effectiveness of a multifaceted AKI intervention approach. Kidney Int Rep. 2021;6:636–644. doi: 10.1016/j.ekir.2020.12.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Moledina DG, Belliveau O, Yamamoto Y, et al. Variation in best practice measures in patients with severe hospital-acquired acute kidney injury: a multicenter study. Am J Kidney Dis. 2021;77:547–549. doi: 10.1053/j.ajkd.2020.08.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Collister D, Pannu N, Ye F, et al. Alberta Kidney Disease Network. Health care costs associated with AKI. Clin J Am Soc Nephrol. 2017;12:1733–1743. doi: 10.2215/CJN.00950117 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ozrazgat-Baslanti T, Loftus TJ, Ren Y, et al. Association of persistent acute kidney injury and renal recovery with mortality in hospitalised patients. BMJ Health Care Inform. 2021;28. doi: 10.1136/bmjhci-2021-100458 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Premier Healthcare Database White Paper. Data that informs and performs March 2, 2020. Accessed February 25, 2023. https://learn.premierinc.com/white-papers/premier-healthcare-database-whitepaper
  • 25.Jeong IG, Khandwala YS, Kim JH, et al. Association of robotic-assisted vs laparoscopic radical nephrectomy with perioperative outcomes and health care costs, 2003 to 2015. JAMA. 2017;318:1561–1568. doi: 10.1001/jama.2017.14586 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Schneeweiss S, Seeger JD, Landon J, Walker AM. Aprotinin during coronary-artery bypass grafting and risk of death. N Engl J Med. 2008;358:771–783. doi: 10.1056/NEJMoa0707571 [DOI] [PubMed] [Google Scholar]
  • 27.Kadri SS, Gundrum J, Warner S, et al. Uptake and accuracy of the diagnosis code for COVID-19 among US hospitalizations. JAMA. 2020;324:2553–2554. doi: 10.1001/jama.2020.20323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Moon RC, Mackey RH, Cao Z, et al. Is COVID-19 less deadly now?—Trends of in-hospital mortality among hospitalized COVID-19 patients in the United States. Clin Infect Dis. 2022;74:2238–2242. doi: 10.1093/cid/ciab830 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Siew ED, Matheny ME. Choice of reference serum creatinine in defining acute kidney injury. Nephron. 2015;131:107–112. doi: 10.1159/000439144 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Hoste E, Bihorac A, Al-Khafaji A, et al. Identification and validation of biomarkers of persistent acute kidney injury: the RUBY study. Intensive Care Med. 2020;46:943–953. doi: 10.1007/s00134-019-05919-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Bagshaw SM, Al-Khafaji A, Artigas A, et al. External validation of urinary C-C motif chemokine ligand 14 (CCL14) for prediction of persistent acute kidney injury. Crit Care. 2021;25:185. doi: 10.1186/s13054-021-03618-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Kellum JA, Sileanu FE, Bihorac A, Hoste EA, Chawla LS. Recovery after acute kidney injury. Am J Respir Crit Care Med. 2017;195:784–791. doi: 10.1164/rccm.201604-0799OC [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Uchino S, Bellomo R, Bagshaw SM, Goldsmith D. Transient azotaemia is associated with a high risk of death in hospitalized patients. Nephrol Dial Transplant. 2010;25:1833–1839. doi: 10.1093/ndt/gfp624 [DOI] [PubMed] [Google Scholar]
  • 34.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383. doi: 10.1016/0021-9681(87)90171-8 [DOI] [PubMed] [Google Scholar]
  • 35.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619. doi: 10.1016/0895-4356(92)90133-8 [DOI] [PubMed] [Google Scholar]
  • 36.Mihaylova B, Briggs A, O'Hagan A, Thompson SG. Review of statistical methods for analysing healthcare resources and costs. Health Econ. 2011;20:897–916. doi: 10.1002/hec.1653 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Cao Z, Villa KF, Lipkin CB, Robinson SB, Nejadnik B, Dvorak CC. Burden of illness associated with sinusoidal obstruction syndrome/veno-occlusive disease in patients with hematopoietic stem cell transplantation. J Med Econ. 2017;20:871–883. doi: 10.1080/13696998.2017.1336623 [DOI] [PubMed] [Google Scholar]
  • 38.Basu A, Rathouz PJ. Estimating marginal and incremental effects on health outcomes using flexible link and variance function models. Biostatistics. 2005;6:93–109. doi: 10.1093/biostatistics/kxh020 [DOI] [PubMed] [Google Scholar]
  • 39.Ortiz-Soriano V, Donaldson K, Du G, et al. Incidence and cost of acute kidney injury in hospitalized patients with infective endocarditis. J Clin Med. 2019;8:927. doi: 10.3390/jcm8070927 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Hobson C, Ozrazgat-Baslanti T, Kuxhausen A, et al. Cost and mortality associated with postoperative acute kidney injury. Ann Surg. 2015;261:1207–1214. doi: 10.1097/SLA.0000000000000732 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Kolhe NV, Eldehni MT, Selby NM, McIntyre CW. The reimbursement and cost of acute kidney injury: a UK hospital perspective. Nephron Clin Pract. 2014;126:51–56. doi: 10.1159/000358435 [DOI] [PubMed] [Google Scholar]
  • 42.Silver SA, Long J, Zheng Y, Chertow GM. Cost of acute kidney injury in hospitalized patients. J Hosp Med. 2017;12:70–76. doi: 10.12788/jhm.2683 [DOI] [PubMed] [Google Scholar]
  • 43.Alshaikh HN, Katz NM, Gani F, et al. Financial impact of acute kidney injury after cardiac operations in the United States. Ann Thorac Surg. 2018;105:469–475. doi: 10.1016/j.athoracsur.2017.10.053 [DOI] [PubMed] [Google Scholar]
  • 44.Amin AP, McNeely C, Spertus JA, et al. Incremental cost of acute kidney injury after percutaneous coronary intervention in the United States. Am J Cardiol. 2020;125:29–33. doi: 10.1016/j.amjcard.2019.09.042 [DOI] [PubMed] [Google Scholar]
  • 45.Kellum JA, Sileanu FE, Murugan R, Lucko N, Shaw AD, Clermont G. Classifying AKI by urine output versus serum creatinine level. J Am Soc Nephrol. 2015;26:2231–2238. doi: 10.1681/ASN.2014070724 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SUPPLEMENTARY MATERIAL
kidney360-4-316-s001.pdf (549.5KB, pdf)

Data Availability Statement

The data that support the findings of this study are available from Premier Inc., but restrictions apply to the availability of these data, which were used under license for the current study and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Premier Inc (https://www.pinc-ai.com/applied-sciences).

Articles from Kidney360 are provided here courtesy of American Society of Nephrology

RESOURCES