Background: Neonatal abstinence syndrome (NAS) occurs when an infant is exposed to licit or illicit chemical substances in utero and manifests symptoms of withdrawal from the substance(s) after birth. NAS is a clinical diagnosis, but is generally confirmed with testing of infant urine and/or meconium to identify specific causative substances. Appropriate identification, documentation, and communication of NAS diagnosis is needed to qualify for federally funded therapies aimed at supporting development through early intervention (EI) programs. When NAS is not diagnosed in hospital settings, children must either exhibit developmental delays or families must meet a minimum of 4 social risk criteria to qualify for EI. Our objective was to assess accuracy of administrative billing codes in identifying NAS and whether this differed by discharge setting, newborn nursery (NBN) vs NICU.
Methods: From January to March 2017, all discharges of neonates born at Baystate Medical Center were identified utilizing a billing database. 922 discharges were identified, 139 from NICU and 783 from NBN. Hospital documentation of all discharged neonates was reviewed manually to identify babies exposed to substances associated with NAS. Babies with ICD-10 codes commonly utilized to designate neonates born to drug-exposed mothers (table 1) were identified by financial analyst review of billing databases.
Results: Of 22 babies confirmed as opioid-exposed through chart review, 5 of 7 discharged from NBN and 15 of 15 discharged from NICU were assigned a billing code associated with in utero drug exposure, resulting in sensitivities of 71.4% and 100% respectively. Of 63 babies confirmed as exposed to non-opioid substances (including tobacco, alcohol, cannabinoids, antidepressants, cocaine), through chart review, 12 of 54 discharged from NBN and 9 of 9 discharged from NICU were assigned a billing code associated with in utero drug exposure, resulting in sensitives of 22.2% and 100% respectively. Two babies with no identified NAS-associated substance exposure (1 from NICU and NBN each) were assigned billing codes associated with in utero drug exposure, resulting in specificities of 99.8% and 99.1% respectively (table 2).
Conclusions: Analysis of administrative billing data is not reliable in identifying neonates exposed to substances associated with NAS. This was due primarily to missed designation of ICD-10 codes associated with NAS in NBN, whereas babies discharged from NICU were uniformly identified utilizing ICD-10 code review. Non-opioid substance exposed babies were less likely to be identified than opioid exposed babies.
.png)
.png)