Page 273 - 16Neonatal Jaundice_compressed

Appendix H: Evidence tables

Bibliographic details Study type & Patient characteristics Test, Reference Standard, Results Reviewers Comments
Evidence level Threshold for a positive test
mean BW 3.3 ± 0.5 kg accuracy of three tests in predicting hyperbilirubinaemia (Univariate analysis) Methodology described
significant hyperbilirubinaemia by (n = 812) adequately
GA < 38 weeks: 13.4% the c-statistic (mathematically Blinding – not specified
Gender: males = 49.4% equal to area under ROC curve)
Ethnicity: Factors increasing risk
White = 33.5%
Black = 53.2% Test 1: Pre-discharge bilirubin –
Pre-discharge bilirubin measured
Asian = 9.8% either by TcB or TSB at < 52 hours high risk zone OR: 147 (95% CI 34–639)
Other = 3.4% of age, and expressed as risk-zone high-intermediate risk zone OR: 21 (95% CI
4.9–93.0)
Since the population in the area was on hour specific nomogram. GA < 38 weeks OR: 9.2 (95% CI 4.4–19.0)
predominantly black, stratified Daily TcB levels recorded using intended breastfeeding OR: 2.2 (95% CI
BiliChek, and TSB performed if
sampling scheme used to get a th 1.0–4.5)
representative sample. TcB above 75 centile on hour- intended breast + bottle feeds OR: 3.7
Group 1: Infants with significant specific nomogram or TcB (95% CI 1.6–8.6)
hyperbilirubinaemia (n = 48) reading = 205 micromol/litre. TSB Grade 4 or higher degree of clinical jaundice
value taken for analysis when both OR 6.0 (95% CI 2.1 to 17)
Group 2: Infants without significant TcB and TSB done.
hyperbilirubinaemia (n = 703) Factors decreasing risk
Test 2: Black race OR 0.43 )95% CI 0.23–0.80)
Exclusion: Clinical risk factors assessed by Maternal history of smoking OR: Not
babies transferred to the intensive care review of hospital charts for reported
nursery for any reason maternal race,
Babies who received intravenous intended method of feeding, Factors significant in multivariate analysis
antibiotics for concern for sepsis. GA, model (P < 0.05)
history of previous infant with
jaundice, GA< 38 weeks OR 19 (95% CI 6.3- 56)
clinical assessment of jaundice, Mother’s plan of exclusive breastfeeding:
G6PD deficiency.
OR 3.7 (95% CI 1.1- 13)
Test 3: Black race: OR 0.22 (95% CI 0.08- 0.61)
Combination of pre-discharge Grade 4 or higher jaundice observed
bilirubin risk zone and clinical risk clinically: OR 1.7 (95% CI 1.2–2.6)
factors. Female sex: OR 3.2 (95% CI 1.2–8.4)

Reference standard: 2) Predictive ability of the three tests in
Bilirubin levels (TcB or TSB) predicting significant hyperbilirubinaemia
measured on day 3–5 on both (multivariate regression)
hospitalised and discharged babies
(at home) using similar method as Test 1: Pre-discharge bilirubin risk zone
c-statistic 0.88 (95% 0.85 to 0.91)
in Test 1, and Significant
Hyperbilirubinaemia defined as
bilirubin levels exceeding or within Test 2: Clinical risk factors (final model had
17 micromol/litre of the hour- 5 factors – GA, intended method of feeding,
specific phototherapy treatment black race, extent of jaundice and gender)
thresholds. c-statistic 0.91 (95% 0.86 to 0.97)

Test 3: Combination model (pre-discharge
risk zone + clinical factors of GA and %

209