438 - When do Accidents Become Neglect? Factors Associated with Child Protective Services Referrals in Patients Presenting to a Pediatric Emergency Department for Unintentional Ingestions or Exposures

Publication Number: 4429.438

Catherine Cheng, University of Michigan Medical School, Ann Arbor, MI, United States; Allison Cator, University of Michigan Medical School, Ann Arbor, MI, United States; Cheyney Dobson, University of Michigan Medical School, Ann Arbor, MI, United States; Hope Coven, University of Michigan Medical School, New York, NY, United States; Miao Yu, University of Michigan, Ann Arbor, MI, United States; Harlan McCaffery, University of Michigan Medical School, Ann Arbor, MI, United States; Niko Kaciroti, University of Michigan Medical School, Ann Arbor, MI, United States; Chiu-Mei Chen, University of Michigan Medical School, Ypsilanti, MI, United States; Susan Woolford, University of Michigan Medical School, Ann Arbor, MI, United States

Background: There are no guidelines for Child Protective Services (CPS) referrals for unintentional ingestions or chemical exposures (IOCE), leaving clinicians to make judgements as to whether neglect is present. This may increase risk for inequities and variations in care.

Objective: We aimed to assess whether patient demographics, characteristics of unintentional IOCE were associated with CPS referrals in a pediatric emergency department (PED) setting.

Design/Methods: Retrospective study examining electronic health records of patients < 12 years old presenting to a PED with chief complaints of IOCE from January 1, 2020 to December 31, 2024. Unintentional IOCE was determined by manual chart review. We excluded visits for intentional IOCE, where a CPS referral was filed prior to PED presentation or for reasons other than the IOCE and repeat visits by patients during the study period. Predictors included demographics (age, race, ethnicity, insurance type, area deprivation index (ADI) score), IOCE type (Table 1) and ED disposition (proxy for clinical severity). Primary outcome was CPS referral. Bivariate and multivariate analyses were used to determine odds of CPS referral by demographics, IOCE type and clinical severity.

Results: Of 635 records meeting inclusion criteria, 132 were excluded after chart review, leaving 503 PED visits by unique patients. Of these, 109 (22%) had CPS referrals (Table 2). Multivariable logistic regression showed female gender (OR 2.04; 95% CI 1.08-4.00), alcohol/cannabinoid/nicotine ingestion (OR 185.2; 95% CI 52.04-930.47), illegal or controlled substance ingestion (OR 17.67; 95% CI 4.63-90.46), unknown ingestion (OR 7.65; 95% CI 1.43-47.24), and hospital admission (OR 8.65; 95% CI 3.89-19.69) were associated with greater odds of CPS referral. Having Medicaid was marginally associated with higher odds of being referred to CPS, but this did not reach statistical significance (OR 1.96; P=0.057). Age, race, ethnicity, and ADI score were not associated with CPS referrals (Table 3).

Conclusion(s): In patients presenting to a PED for unintentional IOCE, referrals to CPS were associated with IOCE type and patient disposition, but not with patient demographics other than female gender. Ingestions of legal recreational substances had higher likelihood of CPS referral than controlled medications and illegal drugs which may reflect bias against presence of cannabinoids in homes. IOCE type and severity may not indicate supervisory neglect which should consider the child's risk of ongoing harm. Guideline development around CPS referrals for unintentional IOCE could help to alleviate these differences.

Table 1. IOCE group based on level of regulation and risk to children
*IOCE types were grouped into subtypes due to sparsity of categories inhibiting model performance.

Bivariate analysis of patient demographics and IOCE characteristics associated with CPS filing
*ADI is a validated measure ranking neighborhoods based on level of socioeconomic disadvantage. A larger ADI indicates higher level of disadvantage.

Table 3. Likelihood of PED referral to CPS by patient demographics and IOCE characteristics