The trends depicted are among children hospitalized at children’s hospitals across the transition from International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) to International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM). The vertical dotted line indicates the date of the transition on October 1, 2015.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Puls HT, Anderst JD, Davidson A, Hall M. Trends in the Use of Administrative Codes for Physical Abuse Hospitalizations. JAMA Pediatr. 2020;174(1):91–93. doi:10.1001/jamapediatrics.2019.3992
Administrative data from health care systems are important resources in the research of physical child abuse, such as tracking its incidence.1 Validation studies of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) for the identification of abuse cases among hospitalized children demonstrated reasonably good sensitivity (73.5%) and specificity (92.4%), despite that ICD-9-CM codes captured some cases in which abuse was only suspected.2,3 The ICD-9-CM abuse codes (all ICD codes prefixed with 995.5) made no declaration for diagnostic certainty.4 However, the US transition to International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM), on October 1, 2015,5 allowed for new designations of suspected abuse (all ICD codes prefixed with T76) and confirmed abuse (all ICD codes prefixed with T74), which may have consequences in coding practices and subsequently the ascertainment of abuse hospitalizations. Demonstrating stability in the use of administrative coding for abuse hospitalizations would support research efforts to continue with ICD-10-CM.6 Conversely, a finding of instability would suggest a need for new validation studies. Our objective was to investigate trends in the use of administrative codes for abuse hospitalizations across the transition from use of ICD-9-CM codes to use of ICD-10-CM codes.
We used data from 49 children’s hospitals participating in the Pediatric Health Information System (Children's Hospital Association, Lenexa, KS) administrative database with continuous inpatient and observation data from January 1, 2013, to December 31, 2018. Data were analyzed from March 1, 2019, to June 1, 2019. Consistent with prior studies, hospitalizations qualified as cases of abuse if they had a code for abuse or assault and a code for an injury (or anoxic brain injury or retinal hemorrhages).1-3 We cross-walked ICD-9-CM codes to their ICD-10-CM equivalent codes using General Equivalence Mapping (GEM).7 This study was deemed exempt from institutional board review with a waiver of informed consent by Children’s Mercy Kansas City because of a nonhuman participant determination.
Use of administrative codes for abuse was measured as the percentage of all hospitalizations that met coding criteria for abuse. We used an interrupted time series to determine if the use of all ICD-10-CM codes for abuse hospitalizations (suspected abuse, confirmed abuse, and assault) changed at the time of the transition or during the months after the transition. Each ICD-10-CM code was analyzed for suspected and confirmed abuse and assault for changes. All statistical analyses used SAS, version 9.4 (SAS Institute). A 2-sided P value was significant at less than .05.
There were 16 051 (0.3%) abuse hospitalizations identified of 4 947 906 discharges. There was no change in the use of ICD-9-CM codes for abuse hospitalizations before the transition (β coefficient, −0.000479; 95% CI, −0.001663 to 0.0007037; P = .41) (Figure). When using all ICD-10-CM codes for abuse hospitalizations (ie, suspected and confirmed abuse and assault) there was no discontinuity at the transition to ICD-10-CM (β coefficient, 0.00859; 95% CI, −0.03159 to 0.048769; P = .68), but monthly rates after the transition significantly increased (β coefficient, 0.002807; 95% CI, 0.0.001837-0.003776; P < .001) (Figure). These significant after-transition increases were observed at 10 hospitals (20.4%). When including only confirmed and assault ICD-10-CM codes (ie, excluding suspected), there was an immediate 36.3% reduction at the transition from 0.29% of hospitalizations being coded as abuse at the end of ICD-9-CM to 0.18% of hospitalizations being coded as abuse at the beginning of ICD-10-CM (β coefficient, −0.1090; 95% CI, −0.14271 to −0.07529; P < .001) with monthly rates increasing significantly after the transition (β coefficient, 0.0006670; 95% CI, −0.00001077 to 0.001345; P = .054 [not depicted]). After the transition, the monthly rates of individual ICD-10-CM codes for suspected abuse (β coefficient, 0.002140; 95% CI, 0.001611-0.002669; P < .001) and confirmed abuse (β coefficient, 0.0007167; 95% CI, 0.0003011-0.001132; P = .001) increased significantly (Figure).
Findings from this analysis suggest instability in the use of administrative codes for the identification of abuse hospitalizations after the transition to ICD-10-CM. Initially, inclusion of new ICD-10-CM codes for suspected abuse appeared to identify abuse hospitalizations in a manner consistent with ICD-9-CM. However, because of increasing use of suspected abuse ICD-10-CM codes, their inclusion did not appear to allow for a consistent ascertainment of abuse hospitalizations between ICD-9-CM and ICD-10-CM. Ultimately, coding limitations may leave the tracking of physical abuse hospitalizations discontinuous near the ICD-10-CM transition.
Our results are limited in that they cannot comment on the accuracy or appropriateness of the observed trends. However, the association of observed trends with the ICD-10-CM transition appears to reflect changes in administrative coding practices rather than true changes in the clinical identification or occurrence of abuse hospitalizations. Future validation studies of ICD-10-CM for abuse hospitalizations are needed.
Accepted for Publication: June 11, 2019.
Corresponding Author: Henry T. Puls, MD, Division of Hospital Medicine, Department of Pediatrics, Children’s Mercy Hospitals, 2401 Gillham Rd, Kansas City, MO 64108 (email@example.com).
Published Online: November 4, 2019. doi:10.1001/jamapediatrics.2019.3992
Author Contributions: Matthew Hall, PhD, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Puls, Anderst, Hall.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Puls, Hall.
Critical revision of the manuscript for important intellectual content: Puls, Anderst, Davidson.
Statistical analysis: Puls, Hall.
Administrative, technical, or material support: Davidson.
Supervision: Puls, Anderst.
Conflict of Interest Disclosures: None reported.
Create a personal account or sign in to: