PEDIATRICS Vol. 114 No. 1 July 2004, pp. 323
Hyperbilirubinemia Benchmarking
Thomas B. Newman, MD, MPH
Petra J. Liljestrand, PhD
Gabriel J. Escobar, MD
University of California San Francisco
San Francisco, CA 94143-0560
To the Editor. —
Chou et al1 recently reported on their experience with neonatal hyperbilirubinemia in the Henry Ford Hospital System (HFHS). They suggested that rates of hyperbilirubinemia obtained with a "rigorous protocol for hyperbilirubinemia management" at HFHS could be used as a benchmark for other healthcare systems. They compared their results with those we reported from the Northern California Kaiser Permanente Medical Care Program (NC-KPMCP)2 and speculated that the significantly lower rates of hyperbilirubinemia at HFHS were due to their more rigorous management protocols.
Although we agree that it can be helpful to compare rates of hyperbilirubinemia across health systems and appreciate their attention to our work, we have some reservations about their methods and conclusions. First, the definition of hyperbilirubinemia they used was the total serum bilirubin (TSB) level at which the American Academy of Pediatrics (AAP) 1994 guideline3 recommends phototherapy be considered, ie, a TSB level of 17 mg/dL at 
72 hours of age. This level is too low to use as a benchmarking outcome, because it primarily reflects the underlying prevalence of risk factors for hyperbilirubinemia rather than perinatal management decisions. In both the HFHS and NC-KPMCP studies, the main risk factors for hyperbilirubinemia were race, gestational age, maternal age, gender, and breastfeeding. Of these, the only one that pediatric health care providers might be able to influence is breastfeeding. However, greater success at encouraging breastfeeding would lead to higher rates of hyperbilirubinemia, because breastfeeding is a risk factor for hyperbilirubinemia. It is noteworthy that only 30% of newborns in the HFHS study were exclusively breastfed, compared with 66% in the NC-KPMCP study.
The authors also compared rates of "severe hyperbilirubinemia," defined as TSB 20 mg/dL, between the HFHS and the NC-KPMCP studies. Although a little more clinically relevant, a TSB level of 20 mg/dL is still too low for benchmarking, because the proportion of newborns reaching that level would be minimally affected by adherence to treatment recommendations. For well, term newborns 72 hours old, phototherapy is recommended by the AAP above this level. If the HFHS protocol called for phototherapy at lower levels than those in the 1994 AAP guideline, then they could at least partially attribute a lower incidence of newborns with TSB 20 mg/dL to their more rigorous protocols. Surprisingly, however, the laminated pocket card with the HFHS bilirubin management guidelines, depicted in their Figure 1, calls for considerably less phototherapy treatment than recommended by the AAP. For example, for healthy term infants 120 hours old, the card suggests that phototherapy is optional for a TSB level of 22 to 25 mg/dL and required for aTSB 25 mg/dL. It is hard to understand how implementation of such an explicitly lenient guideline could be responsible for a low rate of hyperbilirubinemia.
There are at least 2 better explanations for the lower rate of TSB 20 mg/dL reported in the HFHS study, compared with the NC-KPMCP study. The first, as mentioned above, is case mix. Although the authors used Poisson regression to adjust for case-mix differences, they did so only 1 variable at a time. This does not work. For example, the difference in the rate of hyperbilirubinemia between white newborns from the NC-KPMCP (1.7%) and HFHS (1.2%) studies reported in Table 3 of the Chou et al article can be explained by the greater proportion of exclusive breastfeeding in the NC-KPMCP study (66% vs 30%).
A second explanation for differences in the incidence of hyperbilirubinemia is much more troublesome, not only for the comparison between the HFHS and NC-KPMCP, but for any benchmarking efforts at all. The problem, which for many years sat like an ignored elephant in the living room of neonatal jaundice researchers, is interlaboratory variability in measurements of TSB. This problem finally received much-needed attention at a recent National Institute of Child Health and Human Development-sponsored meeting. However, until more reliable bilirubin standards can be developed and implemented, any comparisons of hyperbilirubinemia rates across hospitals must consider the possibility that differences are caused by different laboratories.
REFERENCES
- Chou SC, Palmer RH, Ezhuthachan S, et al. Management of hyperbilirubinemia in newborns: measuring performance by using a benchmarking model.
Pediatrics. 2003;112
:1264
–1273[Abstract/Free Full Text]
- Newman TB, Escobar GJ, Gonzales V, Armstrong MA, Gardner M, Folck B. Frequency of neonatal bilirubin testing and hyperbilirubinemia in a large health maintenance organization.
Pediatrics. 1999;104
:1198
–1203[Abstract/Free Full Text]
- American Academy of Pediatrics, Provisional Committee for Quality Improvement and Subcommittee on Hyperbilirubinemia. Practice parameter: management of hyperbilirubinemia in the healthy term newborn [published correction appears in Pediatrics. 1995;95:458–461].
Pediatrics. 1994;94
:558
–565[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2004 by the American Academy of Pediatrics
