Abstract
Background: Acute kidney injury (AKI) increases patient morbidity and mortality. In value-based care, the documented and coded diagnoses during hospitalization influences an encounter’s relative weight (RW), including severity of illness (SOI), and risk of mortality, which ultimately determines reimbursement for care. The impact of a secondary diagnosis of AKI on RW in pediatric patients has not been evaluated.
Methods: A single-center, retrospective observational study was conducted over six months. The institutional coding database was queried for secondary diagnoses signifying AKI. The RW for each case was determined with and without an AKI secondary diagnosis. Patients were further stratified by their SOI score to evaluate change in RW and SOI.
Results: Over a six-month period, 372 patients had a secondary AKI diagnosis, with a mean RW 2.14 decreasing to a mean RW 1.83 without an AKI diagnosis (p = 2.2e-16). When stratified by SOI, one patient had SOI 1 with RW change -0.286; six patients had SOI 2 with mean RW change -0.0669; 189 patients had SOI 3 with mean RW change -1.862 (p=2.23E-16); and 176 patients had SOI 4 with mean RW change -0.452 (p=9.46E-14), when the AKI secondary diagnosis was removed.
Conclusions: Significant negative changes in RW were observed when AKI was removed, suggesting diagnostic omission may result in inaccurately lesser representation of patient medical complexity and severity of illness upon hospitalization coding, which may lower reimbursement.
Introduction
Acute kidney injury (AKI) is the sudden onset of decreased kidney function characterized by a rise in serum creatinine and/or decreased urine output. It may be induced by a variety of etiologies, including low blood pressure, medications that injure kidneys, urinary tract blockages, or inflammatory kidney diseases. The sudden onset of AKI is potentially life threating when one is rendered unable to excrete metabolic waste products and excess volume from the body. AKI is a significant and increasing cause of hospital comorbidity in the United States, annually affecting thousands of hospitalized patients, and more than 50 percent of intensive care unit patients.1-3 AKI is associated with both poor health outcomes and significant financial burden due to prolonged lengths of stay, prolonged need for mechanical ventilation, and increased mortality.4-6 In the United States, AKI affects up to one in five hospitalized patients, and AKI-related hospital costs in adults are estimated to be $5 billion a year.7 These numbers may be underestimated, as identification depends on proper recognition and documentation of AKI, and billing codes have been shown to lack sensitivity in truly capturing the condition.8
In pediatric patients, recognition and understanding of AKI is also a significant yet poorly understood issue, with studies demonstrating coded diagnostic data under represents AKI when compared to validated clinical criterion data.9,10 In the United States, based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9) codes, used prior to 2015 when the Centers for Medicaid and Medicare Services (CMS) switched to ICD-10, nearly four in 1,000 of all hospitalized children develop AKI.11 One in every three to four pediatric patients in intensive care units develop some degree of AKI, which has been associated with both increases in ICU and overall hospital lengths of stay.12-14
In practicing value-based care, outcomes are closely linked to the payments for care.15 In the United States, CMS uses diagnosis-related group (DRG) based coding to determine a relative weight (RW) score for each admitted patient’s encounter, along with associated severity of illness (SOI) and risk of mortality (ROM) scores. Relative weight (RW) thus reflects severity of illness and the degree of complication for a patient’s illness encounter, as well as the overall cost of hospitalization. The RW for a hospitalization is multiplied by a hospital’s base payment rate to determine facility reimbursement for the encounter. Both SOI and ROM are measured on a scale from 1-4, with 4 being the most severe, to describe the extent of disease manifestation. The indices are based on several dimensions, incorporating the level of principal diagnosis, complications, interacting conditions, dependence, procedures, response and recovery to treatment, and impairment.15
Around 44 million children in the United States have healthcare coverage via Medicaid or the Children’s Health Insurance Program (CHIP), which is coded in many states using All Patients Refined (APR)-DRG methodology, which was developed for pediatric patients.16, 17 Using APR-DRG coding, the RW for a patient encounter is determined as a compilation of all diagnoses incurred during a hospitalization, leading with a primary diagnosis and including all secondary diagnoses. While a diagnosis of AKI inherently carries additional weighting, the final value of RW is determined within the coding algorithm; thus, when considered case by case, an increased RW may not always be frankly evident.
As an under-recognized diagnosis, not only must practitioners recognize and treat the condition, they must understand the impact proper documentation imparts for improved reimbursement in the face of the condition’s significant health impact.
The aim of this study was to estimate of the actual working RW impact an AKI secondary diagnosis has on RW in a pediatric patient population coded via APR-DRG. We hypothesized that documentation and coding including a secondary diagnosis of AKI significantly increases RW for a pediatric patient’s hospitalization, reflective of the increased resource consumption required to manage a child with AKI.
Methods
An institutional board review approved retrospective observational study was conducted at Dell Children’s Medical Center in Austin, Texas. The institutional database was queried over a six-month period, November 1, 2018, through April 30, 2019, for ICD-10 codes N17.0 (Acute kidney failure with tubular necrosis), N17.1 (Acute kidney failure with acute cortical necrosis), N17.2 (Acute kidney failure with medullary necrosis), N17.8 (Other acute kidney failure), and N17.9 (Acute kidney failure, unspecified) for all patients less than 18 years of age who had inpatient hospital stays. All cases were coded using APR-DRG coding within the 3M 360 Encompass System platform. APR-DRG relative weight, SOI, and ROM were extracted, and demographic data, including age, sex, and race or ethnic group was compiled for each patient. Each patient’s encounter coding data was reviewed within the 3M 360 Encompass System platform to determine the patient’s RW for their respective hospitalization without the AKI diagnosis.
Cases were excluded from the study if, upon review, coding data was no longer located in the 3M 360 Encoder software, most likely indicating a hospitalization status downgrade from an inpatient to an observation stay had occurred. Such cases are institutionally coded via a different system.
Data analysis was performed using RStudio Version 1.4 to assess statistical difference between the presence and absence of an AKI secondary diagnosis for the patient encounters. Wilcoxon signed-rank test was used because the data was determined to be non-parametric, suggesting no significant difference between the groups. Data was further stratified by SOI categories 1 through 4 to compare mean RW with and without an AKI secondary diagnosis.
Results
Four hundred fourteen patients with an AKI diagnosis were identified. Seven were excluded due to a lack of coding software data. Thirty-five patients were excluded who had a primary diagnosis of AKI (N17.0, n = 33; N17.9, n = 2). Three hundred seventy-two had a secondary diagnosis of AKI (N17.0, n = 370; N17.1, n = 1; N17.8, n = 2) and were included in further analysis.
Of the 372 patients with a secondary diagnosis of AKI, 52.5 percent were male and 47.5 percent were female, with a mean age of 7.3±6.9 years. The race or ethnic groups were parent or guardian reported as follows: white/non-Hispanic n=156 (41.9 percent), white/Hispanic n=150 (40.3 percent), black n=47 (12.6 percent), Asian/Pacific Islander n=14 (3.8 percent), and ‘decline to specify/unknown’ n=6 (1.6 percent). Patients with a secondary diagnosis of AKI had a RW mean 2.14 and median 1.02 (SD±2.91), and with removal of the AKI secondary diagnosis, had RW mean 1.83 and median 0.72 (SD±2.91). The group had a RW mean change of -0.31 (p = 2.2e-16) when the secondary diagnosis of AKI was removed (Table 1).
Patients with a secondary AKI diagnosis were further stratified by SOI categories 1-4 prior to removing the AKI diagnosis (Table 2).
One patient with a secondary diagnosis of AKI had had an SOI=1 (female, age 2 months, RW with AKI 0.6206, RW without AKI 0.3355).
Six patients with a secondary diagnosis of AKI had an SOI=2, 66.7 percent were male and 33.3 percent were female, with a mean age of 14±3.6 years. The group had a mean RW 0.452 with the AKI diagnosis and a mean RW 0.385 without an AKI diagnosis, for a change in mean RW -0.067, with a range of change of 0 to 0.102. Four patients’ SOI decreased from 2 to 1 when the secondary AKI diagnosis code was removed.
One hundred eighty-nine patients with a secondary diagnosis of AKI had an SOI=3, 52.1 percent being male and 47.9 percent being female, with a mean age of 8±6.7 years. The group had a mean RW 0.754 and median RW 0.667 (SD±0.396), and with removal of the AKI diagnosis, mean RW 0.567 and median RW 0.453 (SD±0.39). The change in mean RW was -0.186 (p=2.23-16). The range of RW change was 0 to 1.138. 137 patient’s SOI decreased from 3 to 2 and 16 patients’ SOI decreased to 1 when the secondary AKI diagnosis code was removed.
One hundred seventy-six patients with a secondary diagnosis of AKI had an SOI of 4, 52.3 of which were male and 47.7 percent of which were female patients, with a mean age of 6.4±7.1 years. The group had a mean RW 3.698 and median RW 1.617 (SD±3.626), and with removal of the AKI diagnosis, mean RW 3.246 and median RW 1.617 (SD±3.741). The change in mean RW was -0.452 (p= 9.46e-14). The range of RW change was 0 to 4.324. 65 patient’s SOI decreased from 4 to 3 and four patients’ SOI decreased to 2 when the secondary AKI diagnosis code was removed (Table 3).
Discussion
Our study demonstrated a significant change in the relative weight of cases when an AKI diagnosis was removed from a patient’s list of codeable secondary diagnoses. While less likely to be overlooked when AKI is a primary diagnosis, omitting an AKI secondary diagnosis may lead to a lesser, inaccurate reflection of the patient’s overall complexity and healthcare resource consumption and lower reimbursement for the care provided.
Patients with AKI have prolonged lengths of stay and increased hospital costs.7 A diagnosis of AKI that is both documented, then coded using the APR-DRG system, is recognized via both increased severity of illness and risk of mortality in the final relative weight value for an inpatient encounter. Proper identification and documentation of AKI, including accurate staging, is necessary to prevent revenue loss while providing care for patients with AKI.
The Assessment of Worldwide Acute Kidney Injury, Renal Angina and Epidemiology in critically ill children (AWARE) study conducted a large multicenter prospective study to evaluate the increased risk of morbidity and mortality associated with AKI in critically ill children. The AWARE study found that AKI was diagnosed in 27 percent of all the patients and that severe AKI was associated with increased 28-day mortality. It also highlighted that plasma creatinine levels by themselves were not enough to identify AKI in 67.2 percent of those with low urine output.12 This further emphasizes the high prevalence of AKI in hospitalized children, solidifying the need for proper documentation and coding that reflects a patient’s truer resource consumption and severity of illness.
There are potential limitations in our study. First, our population study period was over six months, a brief period more prone to situational changes. Additionally, there may be inconsistencies in the electronic charting system used in our facility; for example, not every patient has a nephrology consultation, and our system does not provide electronic validation for an AKI diagnosis, and it was beyond the scope of this study to interrogate the etiologies for AKI and validate the accuracy of each AKI diagnosis. Furthermore, in coding these cases, none of the patients carried a diagnosis code of N99.0, which would require coding both N99.0 as well as a diagnosis code of N17.1-N17.9; thus, we were unable to aspect the potential for this aspect of coding. Finally, coding is an inexact science, depending on both the coding professional’s skills and the accuracy of the documentation.
We present work demonstrating the value impact of a secondary diagnosis of AKI, where increased RW carries an impact more tangible for a clinician than complex coding algorithms. Through this, we seek to highlight the ongoing need for accuracy and thoroughness in clinical documentation and coding to truly reflect the additional patient care resource consumption reflected in the diagnosis. In the ever-changing healthcare landscape, the advent of value-based care has imparted an urgency to properly recognize and document this key diagnosis for pediatric patients.
Conflict of Interest Disclosures: The other authors have conflicts of interest to disclose.
Funding/Support: No funding was secured for this study.
Notes
1. Pavkov, Meda E., Jessica L. Harding, and Nilka R. Burrows. 2018. “Trends in Hospitalizations for Acute Kidney Injury — United States, 2000–2014.” MMWR. Morbidity and Mortality Weekly Report 67 (10): 289–93.
2. Sawhney, Simon, and Simon D. Fraser. 2017. “Epidemiology of AKI: Utilizing Large Databases to Determine the Burden of AKI.” Advances in Chronic Kidney Disease 24 (4): 194–204.
3. Hoste, Eric AJ, Gilles Clermont, Alexander Kersten, Ramesh Venkataraman, Derek C Angus, Dirk De Bacquer, and John A Kellum. 2006. “RIFLE Criteria for Acute Kidney Injury Are Associated with Hospital Mortality in Critically Ill Patients: A Cohort Analysis.” Critical Care 10 (3): R73.
4. Uchino, Shigehiko. 2005. “Acute Renal Failure in Critically Ill Patientsa Multinational, Multicenter Study.” JAMA 294 (7): 813.
5. Chertow, Glenn M., Elisabeth Burdick, Melissa Honour, Joseph V. Bonventre, and David W. Bates. 2005. “Acute Kidney Injury, Mortality, Length of Stay, and Costs in Hospitalized Patients.” Journal of the American Society of Nephrology 16 (11): 3365–70.
6. Hoste, Eric A. J., Sean M. Bagshaw, Rinaldo Bellomo, Cynthia M. Cely, Roos Colman, Dinna N. Cruz, Kyriakos Edipidis, et al. 2015. “Epidemiology of Acute Kidney Injury in Critically Ill Patients: The Multinational AKI-EPI Study.” Intensive Care Medicine 41 (8): 1411–23.
7. Silver, Samuel A., and Glenn M. Chertow. 2017. “The Economic Consequences of Acute Kidney Injury.” Nephron 137 (4): 297–301.
8. Grams, Morgan E., Sushrut S. Waikar, Blaithin MacMahon, Seamus Whelton, Shoshana H. Ballew, and Josef Coresh. 2014. “Performance and Limitations of Administrative Data in the Identification of AKI.” Clinical Journal of the American Society of Nephrology 9 (4): 682–89.
9. Schaffzin, Joshua K., Caitlin N. Dodd, Hovi Nguyen, Amanda Schondelmeyer, Suzanne Campanella, and Stuart L. Goldstein. 2014. “Administrative Data Misclassifies and Fails to Identify Nephrotoxin-Associated Acute Kidney Injury in Hospitalized Children.” Hospital Pediatrics 4 (3): 159–66.
10. Jones, Katherine, Alicia Neu, and Jeffrey Fadrowski. 2022. “AKI in Hospitalized Children: Poorly Documented (and Underrecognized).” Frontiers in Pediatrics 9 (January).
11. Sutherland, Scott M., Jun Ji, Farnoosh H. Sheikhi, Eric Widen, Lu Tian, Steven R. Alexander, and Xuefeng B. Ling. 2013. “AKI in Hospitalized Children: Epidemiology and Clinical Associations in a National Cohort.” Clinical Journal of the American Society of Nephrology 8 (10): 1661–69.
12. Kaddourah, Ahmad, Rajit K. Basu, Sean M. Bagshaw, and Stuart L. Goldstein. 2017. “Epidemiology of Acute Kidney Injury in Critically Ill Children and Young Adults.” New England Journal of Medicine 376 (1): 11–20.
13. Jetton, Jennifer G, Louis J Boohaker, Sidharth K Sethi, Sanjay Wazir, Smriti Rohatgi, Danielle E Soranno, Aftab S Chishti, et al. 2017. “Incidence and Outcomes of Neonatal Acute Kidney Injury (AWAKEN): A Multicentre, Multinational, Observational Cohort Study.” The Lancet Child & Adolescent Health 1 (3): 184–94.
14. Schneider, James, Robinder Khemani, Carl Grushkin, and Robert Bart. 2010. “Serum Creatinine as Stratified in the RIFLE Score for Acute Kidney Injury Is Associated with Mortality and Length of Stay for Children in the Pediatric Intensive Care Unit.” Critical Care Medicine 38 (3): 933–39.
15. Horn SD, Horn RA, Sharkey PD. “The Severity of Illness Index as a severity adjustment to diagnosis-related groups.” Health Care Financ Rev. 1984;Suppl(Suppl):33-45.
16. Medicaid.gov Children’s Health Insurance Program. “Reports & Evaluations.” Available online at https://www.medicaid.gov/chip/reports-evaluations/index.html.
17. 3M Health Information Systems. “All Patient Refined Diagnosis Related Groups (APR-DRGs).” Wallingford, CT, 2003. Available online at https://www.hcup-us.ahrq.gov/db/nation/nis/APR-DRGsV20MethodologyOverviewandBibliography.pdf.
Author Biographies
Ella Tierney (ellatierney0@gmail.com) is a student researcher at Dell Children’s Medical Center.
Ayesha Irani (ayesha.irani@utexas.edu) is a general pediatrician at the Kelsey Seybold Clinic in Houston, Texas.
Meena Iyer (miyer@ascension.org) is the chief medical officer and a pediatric hospitalist at Dell Children’s Medical Center.
Alyssa A. Riley (aariley@ascension.org) is a physician advisor and a pediatric nephrologist at Dell Children’s Medical Center.