Transcutaneous bilirubinometry versus total serum bilirubin measurement for newborns

Abstract

Background

Jaundice is a very common condition in newborns, affecting up to 60% of term newborns and 80% of preterm newborns in the first week of life. Jaundice is caused by increased bilirubin in the blood from the breakdown of red blood cells. The gold standard for measuring bilirubin levels is obtaining a blood sample and processing it in a laboratory. However, noninvasive transcutaneous bilirubin (TcB) measurement devices are widely available and used in many settings to estimate total serum bilirubin (TSB) levels.

Objectives

To determine the diagnostic accuracy of transcutaneous bilirubin measurement for detecting hyperbilirubinaemia in newborns.

Search methods

We searched CENTRAL, MEDLINE, Embase, CINAHL and trial registries up to 18 August 2022. We also checked the reference lists of all included studies and relevant systematic reviews for other potentially eligible studies.

Selection criteria

We included cross‐sectional and prospective cohort studies that evaluated the accuracy of any TcB device compared to TSB measurement in term or preterm newborn infants (0 to 28 days postnatal age). All included studies provided sufficient data and information to create a 2 × 2 table for the calculation of measures of diagnostic accuracy, including sensitivities and specificities. We excluded studies that only reported correlation coefficients.

Data collection and analysis

Two review authors independently applied the eligibility criteria to all citations from the search and extracted data from the included studies using a standard data extraction form. We summarised the available results narratively and, where possible, we combined study data in a meta‐analysis.

Main results

We included 23 studies, involving 5058 participants. All studies had low risk of bias as measured by the QUADAS 2 tool. The studies were conducted in different countries and settings, included newborns of different gestational and postnatal ages, compared various TcB devices (including the JM 101, JM 102, JM 103, BiliChek, Bilitest and JH20‐1C) and used different cutoff values for a positive result. In most studies, the TcB measurement was taken from the forehead, sternum, or both. The sensitivity of various TcB cutoff values to detect significant hyperbilirubinaemia ranged from 74% to 100%, and specificity ranged from 18% to 89%.

Authors' conclusions

The high sensitivity of TcB to detect hyperbilirubinaemia suggests that TcB devices are reliable screening tests for ruling out hyperbilirubinaemia in newborn infants. Positive test results would require confirmation through serum bilirubin measurement.

Plain language summary

Is measuring bilirubin levels through the skin a reliable alternative to measuring bilirubin levels in the blood of newborns?

Key messages

The studies included in this review suggest that measuring bilirubin levels through the skin without a needle can identify high bilirubin levels in newborns.

Why is it important to diagnose high bilirubin levels in newborns?

Bilirubin is a substance produced through the breakdown of red blood cells. Jaundice is a very common problem in the newborn period and results from high levels of bilirubin in the blood (hyperbilirubinaemia). It is important to detect hyperbilirubinaemia early to prevent unwanted consequences such as brain damage.

What is transcutaneous bilirubin measurement?

The usual procedure for measuring bilirubin in newborns is to collect a sample of blood (by making a small cut in the heel or inserting a needle into a vein, which can be painful for the infant) and test it in the laboratory (total serum bilirubin measurement). However, there are devices that measure bilirubin by sending a flash of light through the skin (transcutaneous bilirubin measurement). This method is painless and gives an almost immediate result.

What did we want to find out?

We wanted to find out whether transcutaneous bilirubin measurement devices could accurately diagnose hyperbilirubinaemia.

What did we do?

We searched for studies that had investigated the accuracy of transcutaneous bilirubin measurement compared with total serum bilirubin measurement. We had intended to combine the results across studies using statistical methods but were unable to; instead, we presented the results narratively.

What did we find?

We found 23 studies (5058 participants) that were conducted in different countries and settings, used different transcutaneous bilirubin measuring devices, and defined hyperbilirubinaemia with different bilirubin values. Some of the infants were premature and others were born at term (from 37 weeks' pregnancy); their ages ranged from birth to one month of life. Overall, the findings of the studies suggest that transcutaneous bilirubin measurement is a good screening tool for detecting hyperbilirubinaemia in newborns. The included studies found different degrees or levels of accuracy for the use of transcutaneous bilirubin measurement. However, due to the differences between studies, we could not provide an overall combined summary of the accuracy of the different tests. The differences in these studies included factors like the threshold values for hyperbilirubinaemia, the types of transcutaneous bilirubin measuring devices, and age and ethnicity/skin colour of the included infants.

What are the limitations of the evidence?

The included studies were of high methodological quality. However, we reported the results narratively and did not formally evaluate the quality of evidence using GRADE.

How up to date is this evidence?

The evidence is up to date to August 2022.

Summary of findings

Summary of findings 1. Summary of included studies.

Study ID	Ethnicity or race	GA and PNA	Sample size and number of paired results	TcB device and site of measurement	TSB measurement and time interval	Indication for bilirubin	Phototherapy use	Cutoff values	Accuracy outcomes
Bhutta 1991	Pakistani newborns, race not specified	GA: >34 weeks (term and late preterm) PNA: not reported	63 newborns; 100 paired comparisons	JM 101: forehead	Photometry; ≤ 2 hours	Clinical jaundice	No prior phototherapy	TcB index of 17 to detect TSB > 12.5 mg/dL	Sensitivity 87%, specificity 53%, NPV 86.5%, PPV 55.6%
Bilgen 1998	Turkish newborns, race not specified	GA: > 37 weeks (term) PNA: 1–5 days	96 newborns; 96 paired comparisons	JM 101; forehead	Direct spectrophotometry; ≤ 30 minutes	Clinical jaundice	No prior phototherapy	TcB 13 mg/dL to detect TSB 12.9 mg/dL	Sensitivity 100%, specificity 55%, NPV 100%, PPV 32%
Christo 1988	South Indian	GA: >37 weeks (term) and 32–37 weeks (preterm) PNA: not reported	138 newborns; 138 paired comparisons	JM 101; forehead and sternum (average of the 2 readings)	Spectrophotometry; 30 minutes	Infants from whom blood specimen was drawn for bilirubin or other reasons	No prior phototherapy	TcB 18 mg/dL to detect TSB > 13 mg/dL	Sensitivity 95%, specificity 83%, NPV 98%, PPV 69%
Karrar 1989	Saudi newborns (Riyadh, Saudi Arabia)	GA: > 37 weeks (term) PNA: 4–10 days	155 newborns; 155 paired comparisons	JM 101; forehead	American optical bilirubinometer; time interval not reported (probably 30 minutes)	Clinical jaundice	No prior phototherapy	TcB index of 21 to detect TSB ≥ 12.5 mg/dL	Sensitivity 74.0%, specificity 90%, NPV 88.0%, PPV 78.0%
Maisels 1982	White newborns	GA: > 37 weeks (term) PNA: not reported	157 newborns; 292 paired comparisons	JM 101; forehead and sternum	Diazo method; ≤ 30 minutes	Routine bilirubin measurement except in 11 infants (measurement obtained on clinical grounds)	No prior phototherapy	Forehead TcB index of 24 to detect TSB > 12.9 mg/dL. Sternum TcB index of 23 to detect TSB > 12.9 mg/dL	Forehead TcB: sensitivity 100%, specificity 97%, NPV 100%, PPV 58%. Sternum TcB: sensitivity 100%, specificity 96%, NPV 100%, PPV 44%
Schumacher 1985	"Caucasian" (presumed white)	GA: > 36 weeks; PNA: 1–5 days	106 newborns; 106 paired comparisons	JM 101; sternum	Direct spectrophotometry; ≤ 30 minutes	Clinical jaundice	No prior phototherapy	TcB 20 mg/dL to detect TSB > 12.9 mg/dL	Sensitivity 94%, specificity 77.5%, NPV 99%, PPV 44%
Taha 1984	Saudi newborns	GA: > 37 weeks (term) PNA: 3–12 days	68 newborns; 120 paired comparisons	JM 101; forehead	Direct spectrophotometry; near simultaneous measurements	Clinical jaundice	No prior phototherapy	TcB index of 22.5 to detect TSB > 12.9 mg/dL	Sensitivity 69%, specificity 92%, NPV 95%, PPV 58%
Wong 2002	Mostly white, 6 non‐white	GA > 31 weeks (term and preterm) PNA: not reported	64 newborns; 64 paired comparisons	JM 102 and BiliChek; forehead	Photometry; ≤ 30 minutes	Clinical jaundice	No prior phototherapy	JM 102: TcB > 170 μmol/L (10 mg/dL) to detect TSB > 250 μmol/L (14.6 mg/dL) Bilichek: TcB > 150 μmol/L (8.8 mg/dL) to detect TSB > 250 μmol/L (14.6 mg/dL)	JM 102: sensitivity 100%, specificity 32%, NPV 100%, PPV 35% Bilichek: sensitivity 100%, specificity 21.3%, NPV 100%, PPV 31.5%
Bental 2009	Mixed (Ashkenazi and Sephardic Ethiopian) at Natanya Hospital, Israel	GA: ≥ 35 weeks (late preterm) and > 37 weeks (term) PNA: ≤ 7 days	628 newborns; 1091 paired comparisons	JM 103; forehead and sternum (mean from 2 sites was calculated)	Colorimetric method; ≤ 90 minutes	Clinical jaundice	No prior phototherapy	TcB > P75 to predict TSB > P95 on the nomogram	At 24–48 hours: sensitivity 83.3%, specificity 77.5%, NPV 99.2%, PPV 12.2% At 48–72 hours: sensitivity 100%, specificity 75.9%, NPV 100%, PPV 4.5%
Kitsommat 2013	Asian newborns (not specified)	GA term and late preterm PNA: 5–14 days	405 newborns; 455 paired comparisons	JM‐103 forehead	Diazo method; ≤ 30 minutes	Clinical jaundice	No prior phototherapy, or at least 24 hours since phototherapy	TcB 204 μmol/L (11.9 mg/dL) to detect	Sensitivity 96.0%, specificity 58.0%, NPV 97.4%, PPV 45.3%
Lam 2008	Chinese newborns (Hong Kong)	GA: > 35 weeks (term and late preterm) PNA: 3–7 days	113 newborns; 113 paired comparisons	JM 103; forehead and sternum	Spectrophotometry; ≤ 5 minutes	Clinical jaundice	No prior phototherapy	TcB 230 μmol/L (13.5 mg/dL) and 298 μmol/L (17.7mg/dL) to predict TSB > 250 μmol/L (14.6 mg/dL)	Sensitivity 100%, specificity 100%, NPV 100%, PPV 100% (for both cutoff values)
Mendoza‐Chuctaya 2021	Cusco, Peru	GA: ≥ 37 weeks (term) PNA: not reported	123 newborns; 123 paired comparisons	JM 103; forehead and sternum	TSB measurement method not reported; samples obtained 30 minutes	Clinical jaundice	No prior phototherapy	TcB > P95 and > P75 to predict TSB > P95 on the nomogram	Sternum TcB ≥ P95: sensitivity 100%, specificity 80%, NPV 100%, PPV 81% Sternum TcB ≥ P75: sensitivity 100%, specificity 26%, NPV 100%, PPV 54% Forehead TcB ≥ P95: sensitivity 93%, specificity 89%, NPV 94%, PPV 88% Forehead TcB ≥ P75: sensitivity 100%, specificity 33%, NPV 100%, PPV 56%
Yaser 2014	South African newborns	GA: 24–34 weeks (preterm) PNA: < 8 days	122 newborns; 122 paired comparisons	JM 103; forehead, sternum, interscapular area	TSB measurement method not reported; time interval not reported	Clinical jaundice	No prior phototherapy	Single cutoff values not reported as phototherapy initiation was based on unit phototherapy guidelines ("photo line")	Forehead: sensitivity 80.6%, specificity 87.3%, NPV 78.7%, PPV 88.5% Chest: sensitivity 70.2%, specificity 96.4%, NPV 72.6%, PPV 95.9% Interscapular region: sensitivity 94.0%, specificity 72.7%, NPV 90.9%, PPV 80.8%
Boo 2007	Malay, Chinese, and Indian newborns	GA: 37 weeks (term) PNA: not reported	345 newborns; 345 paired comparisons	BiliChek; forehead and sternum	Direct spectrophotometry; ≤ 30 minutes	Clinical jaundice	No prior phototherapy	Forehead: TcB 250 μmol/L (14.9 mg/dL) to detect TSB of 300 μmol/L (17.5 mg/dL). Sternum: TcB 200 μmol/L (11.7 mg/dL) to detect TSB 300 μmol/L (17.5 mg/dL)	Forehead: sensitivity 100%, specificity 39.2%, NPV 100%, PPV 38.5%.  Sternum: sensitivity 100%, specificity 33.6%, NPV 100%, PPV 36.4%
Campbell 2011	White, Black Asian, Indian, Latino (Canada)	GA: ≥ 35 weeks (late preterm and term) PNA: not reported	430 newborns; 430 paired comparisons	BiliChek; forehead	Direct spectrophotometry; ≤ 30 minutes	Clinical jaundice	No prior phototherapy	TcB 180 μmol/L (10.5 mg/dL) to detect TSB > 200 μmol/L (11.7 mg/dL) TcB 200 μmol/L to detect TSB 250 μmol/L (14.6 mg/dL)	TcB 180 μmol/L to detect TSB 250 μmol/L: sensitivity 96%, specificity 55%, NPV 96%, PPV 64% TcB 200 μmol/L to detect TSB 250 μmol/L: sensitivity 96%, specificity 57%, NPV 97%, PPV 34%
Hemmati 2013	Iranian newborns	GA: ≥ 35 weeks (term and late preterm, most infants) and ≤ 35 weeks (preterm, 5 infants) PNA: 1–15 days	560 newborns; paired comparisons not reported	BiliChek; forehead	Spectrophotometry and Diazo method; time interval reported	Clinical jaundice	No prior phototherapy	TcB 15 mg/dL 256 μmol/L) to detect TSB > 15mg/dL (256 μmol/L)	Sensitivity 96.6%, specificity 99.0% NPV 99.8% PPV 95.7%
Karon 2008	146 "Caucasian" (presumed white) 19 Asian, 9 Hispanic, 3 African American	GA: > 32 weeks (term and late preterm) PNA: 1–4 days	177 newborns; 177 paired comparisons	BiliChek; forehead	Diazo and vitros method; 30 minutes	Clinical jaundice	No prior phototherapy	High or high‐intermediate TcB to predict high or high‐intermediate TSB	Diazo: sensitivity 98%, specificity 40%, NPV 98%, PPV 44% Vitros: sensitivity 94%, specificity 55%, NPV 90%, PPV 68%
Knupfer 2001	128 "Caucasian" (presumed white), 7 Asian	GA: 23‐36 weeks (preterm) PNA: 2–6 days	135 newborns; 400 paired comparisons	BiliChek; forehead	Diazo method; time interval not reported	Indication for bilirubin measurement not reported	No prior phototherapy	TcB to correctly identify various threshold for phototherapy; specific cutoff values not reported	Sensitivity 88%, specificity 89.1%, NPV 97.4%, PPV 62.2%
Kolman 2007	Hispanic newborns (Phoenix, AZ)	GA: > 35 weeks (term and late preterm) PNA: 0–144 hours	192 newborns; 192 paired comparisons	BiliChek; forehead	Diazo method; ≤ 30 minutes	Universal bilirubin screening	No prior phototherapy	TcB > P75 to predict TSB > P95 on the nomogram	Sensitivity 100%, specificity 66%, NPV 100%, PPV 16%
Neocleous 2014	White newborns	GA: ≥ 37 weeks (term) PNA: 24– 96 hours	222 newborns; 368 paired comparisons	BiliChek; forehead	Direct spectrophotometry; ≤ 20 minutes	Clinical jaundice	Prior use of phototherapy not reported	TcB 207 μmol/L (12 mg/dL) to detect TSB 205 μmol/L (12 mg/dL) TcB 265 μmol/L (15 mg/dL) to detect TSB 240 μmol/L (14 mg/dL)	TcB 207 μmol/L to detect TSB 205 μmol/L: sensitivity 95.4%, specificity 18.6%, NPV 55.2%, PPV 79.4% (at 24 hours) TcB 265 μmol/L to detect TSB 240 μmol/L: sensitivity 94.2%, specificity 22.5%. NPV 73.8%, PPV 62.5% (at 48 hours)
Samanta 2004	Race/population not reported	GA: ≥ 33 weeks (preterm) and ≥ 37 weeks (term) PNA: 1–11 days	300 newborns; 300 paired comparisons	BiliChek; site of measurement not reported	Diazo method; near simultaneous measurements with TcB	Clinical jaundice	No prior phototherapy	TcB 195 μmol/L (11.4 mg/dL) to detect TSB 250 μmol/L (14.6 mg/dL)	Sensitivity 91% specificity 66%, NPV 96% PPV 42%
Bertini 2008	Italian/white (Florence, Italy)	GA: > 32 weeks (preterm and term) PNA: not reported	241 newborns; 241 paired comparisons	Bilitest BB 77; forehead	Spectrophotometry; 10 minutes	Clinical jaundice	No prior phototherapy	TcB 13 mg/dL to detect TSB 15 mg/dL	Sensitivity 85.7%, specificity 97.8%, NPV 99.1%, PPV 70.6%
Ercan 2018	Turkish newborns, race not specified	GA: ≥ 35 weeks (term and late preterm) PNA: 2–30 days	218 newborns; 271 paired comparisons	JH20‐1C; forehead	Spectrophotometry ≤ 10 minutes	Clinical jaundice	No prior phototherapy	TcB 205 μmol/L (12 mg/dL) to detect TSB ≥ 222 μmol/L (13 mg/dL)	Sensitivity 95.9%, specificity 59.4%, NPV 90%, PPV 77.7%;

GA: gestational age; NPV: negative predictive value; PNA: postnatal age; PPV: positive predictive value; P_x: xth percentile TcB: transcutaneous bilirubin; TSB: total serum bilirubin.

Background

Target condition being diagnosed

Hyperbilirubinaemia is a term used to describe excess of bilirubin in the blood. In newborns, hyperbilirubinaemia becomes clinically apparent as jaundice, a yellow coloration of the skin and sclera (Woodgate 2015). This condition is very common in both term and preterm newborns (affecting around 60% of newborns in the first week of life) and is caused by a predisposition to the production of bilirubin and limited ability to excrete it (Lauer 2011). Most cases of newborn jaundice are mild and resolve spontaneously (Srgo 2006). However, in rare cases, babies have very high levels of bilirubin that can lead to bilirubin encephalopathy and kernicterus (Ebbesen 2005; Srgo 2006). The acute‐phase signs of kernicterus are poor feeding, lethargy, high‐pitched cry, hypertonia or hypotonia, opisthotonos and seizures (Johnson 2002). Chronic manifestations include athetoid cerebral palsy, motor delay, gaze palsy, dental dysplasia, mental retardation and sensorineural hearing loss (AAP 2004). Studies from high‐income countries have reported incidence rates of kernicterus ranging from 0.4 to 2 per 100,000 live births (Burke 2009; Mannig 2007; Srgo 2006). However, studies from low‐income countries suggest that the incidence may be much higher (Nair 2003; Owa 2009). Following guidelines issued by the American Academy of Pediatrics for the management of jaundice in neonates, clinicians are moving away from the long‐established critical cutoff value of 20 mg/dL (342 µmol/L) total serum bilirubin (TSB) in favour of a plot of TSB against time (hours) when deciding which infants require therapy (Kemper 2022). The newer method involves comparing each infant's plot to the age‐specific nomogram to determine the line of management (Higgins 2012). Current treatments for hyperbilirubinaemia include phototherapy and, in severe cases, exchange transfusion (Woodgate 2015).

Index test(s)

Transcutaneous bilirubin (TcB) measurement is a non‐invasive method for measuring serum bilirubin level (Dai 1997). Transcutaneous bilirubinometry works by directing light into the skin and measuring the intensity of the wavelength of light that is returned (Boo 2007). This method is based on optical spectroscopy, which relates the amount of light absorption by bilirubin to the concentration of bilirubin in the skin. The technology was first introduced in 1980 (Yamanouchi 1980). In most cases, the healthcare provider takes the measurement by gently pressing the TcB meter against the sternum or forehead. The result appears in under one minute (Dai 1997). Using this point‐of‐care device saves time and may reduce costs compared to TSB measurement (Maisels 1997). However, gestational age, bodyweight and skin colour may affect the accuracy of TcB results (Knüpfer 2001). For example, TcB tends to underestimate TSB in light and medium skin colours and overestimate TSB in dark skin colours (Samiee‐Zafarghandy 2014). Many TcB devices are available, including the BiliChek device, JM 103 and JM 105 (Grohmann 2006).

Clinical pathway

Nursing staff and physicians routinely monitor newborns for the development of jaundice in the first few hours of life and before discharge from the newborn nursery. This usually involves visual inspection and skin blanching to assess for yellowish discolouration. Visual estimation of bilirubin level is unreliable (Barrington 2007). Therefore, it is necessary to obtain an objective TcB or TSB measurement. Some centres measure TcB or TSB in all infants before hospital discharge, or in all infants with risk factors for severe hyperbilirubinaemia, which include breastfeeding, ABO/Rhesus incompatibility, glucose‐6‐phosphate dehydrogenase deficiency, use of oxytocin during delivery, vacuum‐assisted delivery, prematurity and history of jaundice in a sibling (AAP 2004; Bhutani 2010; Keren 2005). Other centres only perform TcB or TSB measurement in newborns that are visibly jaundiced; the value is plotted on a nomogram to assess the need for treatment (AAP 2004). Healthcare providers can also take measurements in newborns undergoing phototherapy to decide when to stop treatment. The bilirubin levels are interpreted based on the infant's gestational age and postnatal age (Kemper 2022).

Role of index test(s)

TcB assay is a non‐invasive method for measuring bilirubin levels that may help to reduce the risk of anaemia and trauma associated with blood sampling for TSB measurement (Dai 1997). Studies have shown that TcB measurement works well in both hospital and outpatient settings, and is better than visual inspection for estimation of hyperbilirubinaemia (De Luca 2008a; Wainer 2012). Additionally, TcB measurement ensures an immediate result for rapid clinical decision‐making while reducing the risk of infection associated with all invasive procedures (Jangaard 2006). Healthcare professionals can use the TcB meter as a screening tool to estimate the serum bilirubin level in newborns who are not clinically jaundiced, or as a diagnostic tool in jaundiced newborns to assess the need for treatment (Kemper 2022).

Alternative test(s)

There are various methods for determining bilirubin levels in newborns. These include visual assessment, direct spectrophotometric methods (requiring capillary blood) and use of an icterometer (Higgins 2012). Visual assessment for jaundice is common in newborn nurseries and outpatient settings, such as physicians' offices (Harrison 1989). However, studies have shown that this method is ineffective for assessing the severity of jaundice. The icterometer is a specialised ruler marked with different shades of yellow; the operator presses it against the newborn's skin to estimate their bilirubin level (Akman 2000).

Rationale

Bilirubin measurement is one of the most frequently performed tests in newborn infants (Donzelli 2000; Madsen 2000). The current reference standard involves chemical processing. This requires repeated blood sampling, which can be painful for the newborn, costly and time consuming. TcB measurement may be a more cost‐effective and less traumatic method of measuring bilirubin levels in newborns (Dai 1996). However, to justify routine use of TcB devices, we need to systematically review all available evidence from well‐designed studies on the accuracy of TcB measurements in newborn infants. A clear understanding of the diagnostic test accuracy (DTA) of transcutaneous bilirubinometry using a variety of instruments in a variety of populations (including preterm and term infants as well as infants with various racial backgrounds) would be invaluable for understanding the usefulness of TcB measurement in newborns.

Objectives

To determine the diagnostic accuracy of transcutaneous bilirubin measurement for detecting hyperbilirubinaemia in newborns.

Methods

Criteria for considering studies for this review

Types of studies

We included DTA studies with cross‐sectional or prospective cohort design comparing TcB and TSB measurement for hyperbilirubinaemia in newborns.

We excluded randomised controlled trials, retrospective studies, case‐control studies, case reports and any studies that did not provide data for true positives, true negatives, false positives and false negatives, or studies that only reported correlation coefficients between TcB and TSB.

Participants

We included studies that evaluated TcB measurement, performed as part of universal screening or due to visible jaundice, in term or preterm infants aged 0 to 28 days with or without prior phototherapy. We included studies conducted in different clinical settings, such as neonatal intensive care units, paediatric emergency units and paediatric wards; and studies that recruited infants from home or in the community.

Index tests

The index test was TcB measurement in newborns with the use of any TcB device. Normal and abnormal TcB results were as defined by the study authors.

Target conditions

The target condition was hyperbilirubinaemia requiring phototherapy or exchange transfusion.

Reference standards

The reference standard was TSB measured in the laboratory, which requires blood sampling from the newborn. Various methods are available for measuring TSB, including high‐performance liquid chromatography (HPLC), diazo‐based methods, direct spectrophotometry and capillary electrophoresis (Higgins 2012; Kazmierczak 2002). TSB measurement by HPLC is not subject to interference from haemoglobin or lipaemia. However, this method is costly, labour‐intensive and not practical for routine use (Kazmierczak 2004). The diazo‐based methods are the most frequently used laboratory assays, but they may be affected by haemolysis (el‐Beshbishi 2009). The blood sampling required for TSB measurement causes pain and trauma to the neonate, and repeated blood sampling can cause anaemia, especially in preterm infants. There is evidence showing inter‐ and intralaboratory variability with this technique (Lo 2011). For this review, we used the threshold values for hyperbilirubinaemia as defined by the study authors.

Search methods for identification of studies

Electronic searches

The Neonatal Information Specialist (MF) designed the search strategies, which we ran without date, language, or publication type restrictions in August 2022. We searched the following databases.

• Cochrane Central Register of Controlled Trials (CENTRAL) via CRS Web (2022, Issue 8)

• Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions (1946 to 18 August 2022)

• Embase (1974 to 18 August 2022)

• CINAHL Complete via EBSCOhost (18 August 2022)

We searched the following clinical trial registries for ongoing or recently completed trials.

• World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP; who.int/ictrp/search/en)

• U.S. National Library of Medicine (NIH) ongoing trials register ClinicalTrials.gov (clinicaltrials.gov)

• ISRCTN Registry (isrctn.com)

Search strategies are available in Appendix 1, Appendix 2, Appendix 3, Appendix 4 and Appendix 5. We managed search results using Endnote. We identified and removed duplicates using Endnote and Covidence.

Searching other resources

 We checked the reference lists of all included studies and relevant systematic reviews for other potentially eligible studies.

Data collection and analysis

Selection of studies

Two review authors (CO and AO) independently screened the titles and abstracts of the records recovered by the searches, excluding those they considered clearly ineligible. We retrieved the full‐text reports of all remaining records, and the same two review authors assessed them against our eligibility criteria. We resolved any disagreement through discussion or by involving a third review author, if necessary. We used Covidence for the screening process.

Data extraction and management

Two review authors (CO and AO) independently extracted data on study characteristics using a standard piloted data extraction form. Where possible, we computed 2 × 2 tables of true positives, false positives, true negatives, and false negatives for the index tests at the thresholds reported in the primary study. For each included study, we extracted the following information.

• First author's last name and year of publication

• Country of study and race/ethnicity of participants

• Gestational age of included newborns

• Number of enrolled infants and number of paired TcB and TSB results

• Type of TcB device and site of measurement

• Time interval between TcB and TSB measurement

• Prior use of phototherapy

• Accuracy outcomes (sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)).

Table 2 presents the information we extracted from each study.

1. Data extraction form.

[Study ID]	First author, year of publication
Type of study	Journal article, unpublished data
Participants	Sample size Country of study Baseline characteristics (gestational age, postnatal age, race/ethnicity, bodyweight)
Study design	Retrospective/prospective design Sample (consecutive, random, or unclear)
Study criteria	Inclusion criteria Exclusion criteria
Reference standard	Name of assay/manufacturer Name of instrument Cutoff values Time between reference standard performance and TcB measurement Blinding of operator to TcB result
Index test	Name of device Cutoff values Operator training Site of measurement (forehead, sternum)
Target condition	Universal screening of newborns for hyperbilirubinaemia Diagnostic determination of hyperbilirubinaemia in visibly jaundiced newborns
Data	Number of true positives and false positives Number of true negatives and false negatives Number of undetermined/uninterpretable results Sensitivity and specificity of index test Number of missing results for index test Number of missing results for reference standard
Notes	Source of funding (whether any author is affiliated with the manufacturer of the index test; the study was directly funded by the manufacturer; whether authors reported conflicts of interests related to the manufacturer or other funding sources)

TcB: transcutaneous bilirubin.

Assessment of methodological quality

Two review authors (CO and AO) independently assessed the methodological quality of each study using the Quality Assessment of Diagnostic Accuracy Studies‐2 (QUADAS‐2) tool, which consists of four domains: patient selection, index tests, reference standard and flow and timing (Whiting 2011). We developed a rating guideline to ensure consistency when assessing each domain. We assessed risk of bias in all four domains, and we assessed applicability in the first three domains. We piloted our review‐specific QUADAS‐2 tool on five studies to identify possible areas of discrepancy between review authors. We resolved any discrepancies by consensus. Appendix 6 presents the items of the QUADAS‐2 tool and our scoring interpretations for each item.

Statistical analysis and data synthesis

We performed statistical analysis according to Cochrane guidelines for DTA reviews (Macaskill 2010). We included studies that reported sufficient data to construct a 2 × 2 table. For all included studies, we used the data in the 2 x 2 tables to calculate the sensitivity of TcB measurement for detecting hyperbilirubinaemia, need for phototherapy or categorization into various hyperbilirubinaemia risk categories on the TSB nomograms. To calculate specificity, we divided the number of true negatives by the number of participants without the target condition (true negatives plus false positives). Higher specificity at a particular cutoff value represents greater accuracy of the test to correctly identify individuals who do not have the target condition. We had planned to use the 2 × 2 tables to calculate sensitivity and specificity for each study and to meta‐analyse sensitivities and specificities, where appropriate, using the bivariate model (if studies used same threshold for positivity). However, because the studies used different index tests and cutoff values, we were unable to perform a meta‐analysis and have reported the results narratively.

Investigations of heterogeneity

We had planned to investigate heterogeneity by visual examination of both the ROC plot of raw data and the forest plots of sensitivities and specificities. However, we were unable to do this because meta‐analysis was not appropriate.

Sensitivity analyses

We did not perform any sensitivity analyses because we did not combine data from different studies in a meta‐analysis.

Assessment of reporting bias

We had not planned to use funnel plots to evaluate the impact of publication bias or other biases associated with small studies.

Results

Results of the search

Database searches identified 2056 references. After removing 876 duplicates, we screened the titles/abstract of 1180 records, of which we excluded 960. We retrieved and evaluated 220 full‐text articles and excluded 197 (Characteristics of excluded studies). We included 23 studies (Characteristics of included studies). Figure 1 presents the study selection process in a flow diagram.

1.

Flow diagram

Methodological quality of included studies

We assessed the risk of bias of each included study using the QUADAS‐2 tool. We judged most studies at low risk of bias across the four domains (Figure 2; Figure 3). The Characteristics of included studies table provides a detailed description of the risk of bias judgements. Applicability concerns were low in all studies across the domains of patient selection, index test and reference standard.

2.

Risk of bias and applicability concerns graph: review authors' judgements about each domain presented as percentages across included studies.

3.

Risk of bias and applicability concerns summary: review authors' judgements about each domain for each included study.

Findings

The 23 included studies have a total sample size of 5058 newborns. We were unable to ascertain the total number of paired comparisons of TcB and TSB because one study did not provide this information (Hemmati 2013). In the remaining 22 studies, there were 5750 paired comparisons of TcB and TSB. The number of comparisons is higher than the number of newborns because some studies obtained more than one paired TcB and TSB result (Bental 2009; Bhutta 1991; Ercan 2018; Kitsommart 2013; Knupfer 2001; Maisels 1982; Neocleous 2014). The studies were published between 1982 and 2021 and varied in terms of population characteristics/ethnicity, gestational age and postnatal age of the included infants, sample size and number of paired measurements of TcB and TSB, the TcB device used, site of TcB measurement, type of method used for serum bilirubin assay in the laboratory, time interval between TcB and TSB measurement, indication for measurement of bilirubin, and whether the participants received phototherapy or not prior to measurement of TcB and TSB. Most studies took place in well‐baby nurseries; a few were conducted in outpatient settings or the emergency department (Kitsommart 2013; Lam 2008; Maisels 1982). Different race groups were represented across the studies. Some studies included a mixture of infants from different ethnic backgrounds, while others focused on newborns from a specific ethic group. The different ethnic groups/races included, but were not limited to, white infants (Bertini 2008: Maisels 1982; Schumacher 1985), Chinese infants (Lam 2008), African infants (Yaser 2014), and Hispanic infants (Kolman 2007). The gestational age of the newborns also differed across the included studies. Most studies included late preterm newborns (from 35 weeks' gestational age) and term newborns (Bental 2009; Bhutta 1991; Campbell 2011; Ercan 2018; Hemmati 2013; Kitsommart 2013; Kolman 2007; Lam 2008; Schumacher 1985), while other studies included preterm infants (under 35 weeks' gestation; Knupfer 2001; Yaser 2014), or only term newborns (Bilgen 1998; Boo 2007; Karrar 1989; Maisels 1982; Mendoza‐Chuctaya 2021; Neocleous 2014; Taha 1984).

Postnatal age varied significantly across the studies and ranged from birth to 28 days of life. Some studies did not report postnatal age of the infants. Sample sizes ranged from 63 infants (Bhutta 1991) to 628 infants (Bental 2009), and the number of paired measurements of TcB and TSB ranged from 63 (Bhutta 1991) to 1091 (Bental 2009). Some studies included multiple paired measurements of TcB and TSB from the same participants. The number of paired TcB and TSB measurements corresponded to the number of newborns in eight studies (Bertini 2008; Bilgen 1998; Boo 2007; Campbell 2011; Christo 1988; Kolman 2007; Lam 2008; Schumacher 1985).

The studies assessed TcB measurement using various devices (JM 101, JM 102, JM 103, BiliChek, Bilitest and JH20‐1C) compared with TSB measurement in the laboratory by different methods (including HPLC, direct spectrophotometry and diazo method). All studies reported measures of accuracy such as sensitivity or specificity using a specific TSB threshold as diagnostic for hyperbilirubinaemia. The Pearson correlation coefficient is a measure of association and not a valid measure for diagnostic accuracy. Therefore, we did not report the findings of the Pearson correlation coefficient from the included studies. We focused on studies that provided estimates of sensitivity and specificity at specified thresholds, even if the studies did not report association between TcB and TSB using the Pearson correlation coefficient. The remainder of this section provides a narrative summary of the study findings, including details of the TcB devices used, study populations and the DTA results for various reported threshold values. The studies reported the TcB or TSB result either in mg/dL or in μmol/L.

JM 101 transcutaneous device

Seven studies measured TcB with the JM 101 jaundice meter (Bhutta 1991; Bilgen 1998; Christo 1988; Karrar 1989; Maisels 1982; Schumacher 1985; Taha 1984). See Figure 4 and Figure 5.

4.

Forest plot of Analysis 1 (JM 101).

5.

Summary ROC Plot of Analysis 1 (JM 101).

Bhutta 1991 included 63 late preterm and term Pakistani newborns infants with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the forehead to predict a serum bilirubin of 12.5 mg/dL or greater. TSB measurement was by spectrophotometry using an autoanalyser, and the blood sample was obtained at the time of TcB measurement. There were 100 paired comparisons of TcB and TSB from the 63 newborns. At TSB 12.5 mg/dL, the TcB index was 17, with the following accuracy outcomes.

• Sensitivity 88%

• Specificity 53%

• NPV 85.6%

• PPV 55.6%

The study authors did not undertake ROC curve analysis to generate a cutoff with maximal sensitivity and specificity.

Bilgen 1998 included 96 term Turkish newborns between one and five days of life with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the forehead to detect significant hyperbilirubinaemia (TSB above 12.9 mg/dL). TSB measurement was by spectrophotometry, and the blood sample was obtained within 30 minutes of TcB measurement. There were 96 paired comparisons of TcB and TSB. A TcB cutoff of 13 mg/dL could detect TSB levels greater than 12.9 mg/dL with the following accuracy outcomes.

• Sensitivity 100% 

• Specificity 56%

• NPV 100%

• PPV 32%

The study authors did not undertake ROC curve analysis to evaluate different cutoff values.

Christo 1988 included 138 preterm (32 to 37 weeks' gestation) and term South Indian newborns with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the forehead and sternum. The investigators took the average of two readings, one each from the sternum and forehead. TSB measurement was by spectrophotometry using the Technicon RA1000 Autoanalyser, and the blood sample was obtained within 30 minutes of TcB measurement. There were 138 paired comparisons of TcB and TSB. A TcB cutoff of 18 mg/dL to detect a TSB level above 13 mg/dL had the following accuracy outcomes.

• Sensitivity 95%

• Specificity 83%

• NPV 98%

• PPV 69%

Karrar 1989 included 155 Saudi term newborns between four and 10 days of life with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the forehead. TSB measurement was by optical bilirubinometry, and the blood sample was obtained within 30 minutes of TcB measurement. There were 155 paired comparisons of TcB and TSB. A TcB index of 21 to detect a TSB level of 12.5 mg/dL or greater had the following accuracy outcomes.

• Sensitivity 74%

• Specificity 90%

• NPV 78%

• PPV 88%

Maisels 1982 included 157 white term newborns with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the forehead and sternum. TSB measurement was by the diazo method, and the blood sample was obtained within 30 minutes of TcB measurement. There were 292 paired comparisons of TcB and TSB. There were different reported sensitivities and specificities for different TcB cutoffs. For TcB measured at the forehead, a TcB index of 24 to detect a TSB level above 12.9 mg/dL had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 97%

• NPV 100%

• PPV 58%

For TcB measured at the sternum, a TcB index of 23 to detect a TSB level above 12.9 mg/dL had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 96%

• NPV 100%

• PPV 44%

Schumacher 1985 included 106 white late preterm and term newborns (more than 36 weeks' gestation) between one and five days of life with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the sternum. TSB measurement was by direct spectrophotometry, and the blood sample was obtained within 30 minutes of TcB measurement. There were 106 paired comparisons of TcB and TSB. A TcB cutoff of 20 mg/dL to detect a TSB level above 12.9 mg/dL had the following accuracy outcomes.

• Sensitivity 94%

• Specificity 77.5%

• NPV 99%

• PPV 44%

Taha 1984 included 68 Saudi term newborns between three and 12 days of life with no prior phototherapy, and assessed the accuracy of the JM 101 applied to the forehead. TSB measurement was by spectrophotometry using the AO bilirubinometer, and the blood sample was obtained at the time of TcB measurement. There were 120 paired comparisons of TcB and TSB. A TcB index of 22.5 mg/dL to detect a TSB level above 12.9 mg/dL had the following accuracy outcomes.

• Sensitivity 69%

• Specificity 92%

• NPV 95%

• PPV 58%

JM 102 transcutaneous device

Only Wong 2002 evaluated the JM 102 transcutaneous device. Wong 2002 included 64 mostly white preterm (less than 31 weeks' gestation) and term newborns between one and 17 days old with no prior phototherapy, and assessed the accuracy of the JM 102 applied to the forehead. TSB measurement was by spectrophotometry, and the blood sample was obtained within 30 minutes of TcB measurement. There were 64 paired comparisons of TcB and TSB. A TcB cutoff of 170 μmol/L to detect a TSB level above 250 μmol/L (14.6 mg/dL) had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 32%

• NPV 100%

• PPV 35%

JM 103 transcutaneous device

Five studies evaluated the JM 103 TcB measurement device (Bental 2009; Kitsommart 2013; Lam 2008; Mendoza‐Chuctaya 2021; Yaser 2014).

Bental 2009 included 628 late preterm (more than 35 weeks' gestation) and term newborns in the first week of life with no prior phototherapy, and assessed the accuracy of the JM 103 applied to the forehead and sternum. TSB measurement was by colorimetric assay, and the blood sample was obtained within 90 minutes of TcB measurement. There were 1091 paired comparisons of TcB and TSB. The study authors assessed the ability of TcB above the 75th percentile for age to predict significant hyperbilirubinaemia (TSB level above the 95th percentile for age) between 24 and 48 hours of life and between 48 and 72 hours of life. A TcB level above the 75th percentile between 24 and 48 hours of life had the following accuracy outcomes.

• Sensitivity 83.3%

• Specificity 77.5%

• NPV 99.2%

• PPV 12.2%

Between 48 and 72 hours of life, the accuracy outcomes of a TcB level above the 75th percentile were as follows.

• Sensitivity 100% 

• Specificity 75.9%

• NPV 100%

• PPV 4.5%

Kitsommart 2013 included 405 Asian late preterm (from 35 weeks' gestation) and term newborns between five and 14 days of life with no prior phototherapy, and assessed the accuracy of the JM 103 applied to the forehead. TSB measurement was by the diazo method, and the blood sample was obtained within 30 minutes of TcB measurement. There were 455 paired comparisons of TcB and TSB. The study authors undertook a ROC curve analysis, finding an area under the curve for TcB (average of three measurements) of 0.86, and an area under the curve for a single TcB measurement of 0.85. They did not assess the optimal TcB cutoff with the greatest sensitivity and specificity to detect TSB levels of 255 μmol/L (14.9 mg/dL) or greater. The JM 103 device was found to have the following accuracy outcomes.

• Sensitivity 96%

• Specificity 58%

• NPV 97.4%

• PPV 45.3%

Lam 2008 included 133 Chinese late preterm (from 35 weeks' gestation) and term newborns between three and seven days of life with no prior phototherapy, and assessed the accuracy of the JM 103 applied to the forehead and sternum. TSB measurement was by spectrophotometry, and the blood sample was obtained within five minutes of TcB measurement. There were 113 paired comparisons of TcB and TSB. TcB cutoff values of 230 μmol/L (13.5 mg/dL) and 298 μmol/L (17.7mg/dL) to detect a TSB level of 250 μmol/L (14.6 mg/dL) or greater had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 100%

• NPV 100%

• PPV 100%

Mendoza‐Chuctaya 2021 included 123 term newborns without prior phototherapy; it did not report the postnatal age or ethnicity of the infants. The study assessed the accuracy of JM 103 applied to the forehead and sternum. The TSB measurement method was not specified. There were 123 paired comparisons of TcB and TSB. A sternum TcB cutoff value of 95th percentile to predict a TSB level of the 95th percentile or higher had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 80%

• NPV 100%

• PPV 81%

A sternum TcB cutoff value of 75th percentile to predict a TSB level of the 95th percentile or higher had the following accuracy outcomes:

• Sensitivity 100%

• Specificity 26%

• NPV 100

• PPV 54%

A forehead TcB cutoff value of 95th percentile to predict a TSB level of the 95th percentile or higher had the following accuracy outcomes

• Sensitivity 93%

• Specificity 89%

• NPV 94%

• PPV 88%

A forehead TcB cutoff value of 75th percentile to predict a TSB level of the 95th percentile or higher had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 33%

• NPV 100%

• PPV 56%

Yaser 2014 included 122 preterm neonates (less than 35 weeks' gestation) aged under 8 days in whom serum bilirubin had been requested and phototherapy had not been initiated. The study assessed the accuracy of JM 103 applied to different body sites of TcB to detect the need for phototherapy. The study did not report single cutoff values, as phototherapy initiation was based on unit phototherapy guidelines ("photo line"). The forehead TcB had the following accuracy outcomes.

• Sensitivity 80.6 %

• Specificity 87.3%

• NPV 78.7%

• PPV 88.5%

TcB measured at the chest had the following accuracy outcomes.

• Sensitivity 70.2%

• Specificity 96.4%

• NPV 72.6%

• PPV 95.9%

TcB measured at the interscapular region had the following accuracy outcomes.

• Sensitivity 94.0%

• Specificity 72.7%

• NPV 90.9%

• PPV 80.8%

BiliChek transcutaneous device

Nine studies evaluated BiliChek (Boo 2007; Campbell 2011; Hemmati 2013; Karon 2008;og Knupfer 2001; Kolman 2007; Neocleous 2014; Samanta 2004; Wong 2002).

Boo 2007 included 345 term neonates (from 37 weeks' gestation) and preterm neonates (33 to 37 weeks' gestation) with no prior use of phototherapy. The study assessed the accuracy of the Bilichek applied to the sternum and the forehead. There were 345 paired comparisons of TcB and TSB. TcB measured at the forehead with a cutoff value of 250 μmol/L to detect TSB of 300 μmol/L or greater had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 39.2%

• NPV 100%

• PPV 38.5%

TcB measured at the sternum with cutoff value of 200 μmol/L to detect TSB of 300 μmol/L or more had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 33.6%

• NPV 100%

• PPV 36.4%

Campbell 2011 included 430 late preterm infants (up to 35 weeks' gestation) and term neonates (from 37 weeks' gestation) with no prior use of phototherapy. This study did not report the postnatal age of the infants. It assessed the accuracy of BiliChek applied to the forehead. A TcB cutoff value of 180 μmol/L had the following accuracy outcomes to detect a TSB value of 200 μmol/L or greater.

• Sensitivity 96%

• Specificity 55%

• NPV 96%

• PPV 64%

A TcB cutoff value of 200 μmol/L had the following accuracy outcomes to detect a TSB value of 250 μmol/L or greater.

• Sensitivity 96%

• Specificity 57%

• NPV 97%

• PPV 34%

Hemmati 2013 included 560 preterm neonates (up to 35 weeks' gestation), late preterm and term newborns (from 35 weeks' gestation) aged between 1 and 15 days with no prior use of phototherapy. The study assessed the accuracy of Bilichek applied to the forehead. The number of paired comparisons between TcB and TSB was not reported. A TcB cutoff of 256 μmol/L to detect a TSB value of 256 μmol/L had the following accuracy outcomes.

• Sensitivity 96.6%

• Specificity 99.0%

• NPV 99.8%

• PPV 95.7%

Karon 2008 included 177 preterm and term newborns (from 32 weeks' gestation) between one and four days of life with no prior use of phototherapy. The study assessed the accuracy of TcB measured at the forehead. TcB measured from the forehead in the high or high‐intermediate range to predict TSB in the high or high‐intermediate range according to the diazo method had the following accuracy outcomes.

• Sensitivity 98%

• Specificity 40%

• NPV 98%

• PPV 44%

TcB measured from the forehead in the high or high‐intermediate range to predict TSB in the high or high‐intermediate range according to the vitros method had the following accuracy outcomes.

• Sensitivity 94%

• Specificity 55%

• NPV 90%

• PPV 68%

Knupfer 2001 included 135 preterm newborns (23 to 36 weeks' gestation) between two and six days of life with no prior use of phototherapy. The study assessed the accuracy of Bilichek applied to the forehead, and did not report any specific cutoff values. There were 400 paired comparisons of TcB and TSB. TcB measured from the forehead had the following accuracy measures to detect TSB level at or above the threshold for phototherapy.

• Sensitivity 88%

• Specificity 89.1%

• NPV 97.4%

• PPV 62.2%

Kolman 2007 included 192 late preterm and term newborns (from 35 weeks' gestation) between birth and six days of life with no prior use of phototherapy. The study assessed the accuracy of Bilichek applied to the forehead. There were 192 paired comparisons of TcB and TSB. A TcB cutoff of 75th percentile on the nomogram had the following accuracy measures to predict a TSB above the 95th percentile.

• Sensitivity 100%

• Specificity 66%

• NPV 100%

• PPV 16%

Neocleous 2014 included 222 term newborns (from 37 weeks' gestation) between one and four days of life with no prior use of phototherapy. The study assessed the accuracy of Bilichek applied to the forehead. There were 368 paired comparisons of TcB and TSB. A TcB cutoff of 207 μmol/L had the following accuracy outcomes to detect a TSB level of 205 μmol/L or greater.

• Sensitivity 95.4%

• Specificity 18.6%

• NPV 55.2%

• PPV 79.4%

A TcB cutoff of 265 μmol/L had the following accuracy outcomes to detect a TSB level of 240 μmol/L or greater.

• Sensitivity 94.2%

• Specificity 22.5%

• NPV 73.8%

• PPV 62.5%

Samanta 2004 included 300 preterm newborns (from 33 weeks' gestation) and term newborns (from 37 weeks' gestation) aged between one and 11 days with no prior use of phototherapy. The study assessed the accuracy of Bilichek (measurement site not specified). There were 300 paired comparisons between TcB and TSB. A TcB cutoff of 195 μmol/L to detect a TSB level of 250 μmol/L or greater had the following accuracy outcomes.

• Sensitivity 91%

• Specificity 66%

• NPV 96%

• PPV 42%

Wong 2002 included 64 preterm and term newborns (from 31 weeks' gestation) with no prior use of phototherapy. The study did not report the postnatal age of the infants. The study assessed the accuracy of Bilichek applied to the forehead. There were 64 paired comparisons between TcB and TSB. A TcB cutoff of 150 μmol/L to detect a TSB value of 250 μmol/L or greater had the following accuracy outcomes.

• Sensitivity 100%

• Specificity 21.3%

• NPV 100%

• PPV 31.5%

See Figure 6 and Figure 7.

6.

Summary ROC plot of Analysis 4: BiliChek (cutoff > 250 μmol/L).

7.

Summary ROC plot of Analysis 5: BiliChek (cutoff > 200 μmol/L)

Bilitest transcutaneous device

Bertini 2008 included 241 preterm and term newborns (from 32 weeks' gestation) with no prior phototherapy, and assessed the accuracy of Bilitest BB 77 applied to the forehead. TSB measurement was by spectrophotometry, and the blood sample was obtained within ten minutes of TcB measurement. There were different reported sensitivities and specificities for different TcB cutoffs. A TcB cutoff of 222 μmol/L (13 mg/dL) to detect a TSB level of 256 μmol/L (15 mg/dL) or greater had the following accuracy outcomes.

• Sensitivity 85.7%

• Specificity 97.8%

• NPV 99.1%

• PPV 70.6%

JH20‐1C transcutaneous device

Ercan 2018 included 218 late preterm and term newborns between two and 30 days of life. The study authors reported multiple cutoffs in their paper. The main result of interest is that a TcB cutoff of 205 μmol/L (12 mg/dL) determined a TSB level of at least 222 μmol/L (13 mg/dL) with the following accuracy outcomes.

• Sensitivity 95.9%

• Specificity 59.4%

• NPV 90%

• PPV 77.7%

Discussion

The American Academy of Pediatrics and the Canadian Paediatric Society recommend screening newborns for hyperbilirubinaemia with either TcB or TSB before discharge from hospital. Findings from our review support the AAP guideline that recommends TcB as a suitable alternative measure in newborns (AAP 2004). Unlike authors of previous reviews, we did not restrict our search to newborns of a particular gestational age or from a specific ethnic background or race. Another difference between out review and previous reviews is that we reported valid DTA measures (sensitivity, specificity, NPV and PPV) and not correlation between TcB and TSB (Moey 2016; Nagar 2013; Nagar 2016).

Summary of main results

We tabulated and summarised the main results from this review of studies on the accuracy of various TcB devices in different populations of newborns (Table 1). The studies compared TcB with TSB for measuring bilirubin in both term and preterm newborns. The studies evaluated the accuracy of TcB measurement performed for different reasons, including to predict significant hyperbilirubinaemia (TSB above 95th percentile on the Bhutani nomogram) or to predict TSB levels above specific thresholds for hyperbilirubinaemia that require phototherapy. Findings from this review demonstrate that there is extensive evidence on the use of various TcB devices in newborns from different ethnic groups and in different settings. The JM‐101, JM‐102, JM 103 and BiliChek all showed high sensitivity and specificity compared with the different methods for TSB measurement in the laboratory. We thus conclude that TcB measurement is an accurate method for detecting clinically significant hyperbilirubinaemia in newborns.

Strengths and weaknesses of the review

To the best of our knowledge, this is the first systematic review to evaluate the accuracy of TcB measurement in newborn infants. We conducted a comprehensive systematic review to evaluate the use of TcB measurement in newborns of varying gestational and postnatal age and from different ethnic backgrounds and races. Our search strategies identified over 200 studies evaluating the accuracy of TcB devices. However, most of these studies did not meet the inclusion criteria and were excluded. Unlike many previous systematic reviews, we summarised results from studies that reported measures of accuracy such as sensitivity and specificity and excluded studies that only reported the correlation between TcB and TSB, which is not a clinically meaningful measure to recommend the use of TcB for estimating serum bilirubin levels (Hynes 2020; Nagar 2013).

Applicability of findings to the review question

Diagnosis of hyperbilirubinaemia in newborns is dependent on gestational and chronological age. There is no set bilirubin cutoff level for diagnosis of hyperbilirubinaemia in all newborns. The threshold serum bilirubin for phototherapy or exchange blood transfusion also varies with the gestational and chronological age of the newborn. Various nomograms have been developed to assess risk classification for subsequent severe hyperbilirubinaemia in newborns. Given the clinical heterogeneity of the populations of the included studies, it is not surprising that they evaluated different thresholds for hyperbilirubinaemia. We presented the results narratively because this clinical heterogeneity precluded meta‐analysis.

Authors' conclusions

Implications for practice.

An increasing number of newborn nurseries are using transcutaneous bilirubin (TcB) devices. It is important to determine the accuracy of TcB devices in relation to total serum bilirubin (TSB) measurement. There is substantial published evidence on the accuracy of TcB devices in different settings and in different populations of newborns. Guidelines published by the American Academy of Pediatrics and the Canadian Paediatric Society suggest that TcB measurement is a suitable substitute for TSB for evaluating risk of significant neonatal hyperbilirubinaemia in healthy newborns (AAP 2004; CPS 2007). A variety of TcB devices are being used in many countries and settings. TcB measurement is a non‐invasive, painless method of measuring bilirubin, whereas TSB requires blood sampling through heel stick or venipuncture. Furthermore, TcB results are available almost instantaneously. Previous guidelines suggested that there was insufficient evidence to recommend one device over another (Kazmierczak 2006).

Implications for research.

We screened over 1100 studies and included 23 studies with 4836 participants. Detailed assessment of the included studies enabled us to identify specific implications for future studies or research on TcB accuracy. Future studies should assess the impact of various factors, including gestational age, chronological age, and race or skin tone, on the accuracy of TcB measurement. It is important to evaluate true measures of accuracy, such as sensitivity and specificity, using prespecified threshold values for hyperbilirubinaemia in specific population of newborns. Several studies of the JM devices suggest that sternum measurements correlate better with TSB compared to forehead measurements (Yamauchi 1991). However, the device manufacturers suggests that either location is acceptable. Evidence from future studies could help to clarify the best site for TcB measurement and the reasons for any differences between forehead and sternum measurements. In addition, there is a need to ascertain the impact of phototherapy on TcB measurement accuracy, and to evaluate the optimal timing after phototherapy when TcB readings are comparable to those obtained in newborns without prior phototherapy exposure.

History

Protocol first published: Issue 5, 2017

Appendices

Appendix 1. Cochrane CRS strategy

	Cochrane CRS 19 August 2022
#	Searches	Results
1	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj4 bilirubin*) AND CENTRAL:TARGET	131
2	MESH DESCRIPTOR Bilirubin EXPLODE ALL AND CENTRAL:TARGET	924
3	(bilirubin* or bliverdin*) AND CENTRAL:TARGET	10431
4	MESH DESCRIPTOR Administration, Cutaneous EXPLODE ALL AND CENTRAL:TARGET	3992
5	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj3 (administ* or device or devices or measur* or meter* or metre* or monitor? or test*)) AND CENTRAL:TARGET	31790
6	(#2 OR #3) AND (#4 OR #5)	292
7	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj3 bilirubin* adj3 (device or devices or measur* or meter* or metre* or monitor? or test*)) AND CENTRAL:TARGET	43
8	(TcB or bilicare? or bilicheck* or bilichek or Bilicheck* or BiliMed or bilmeter* or JM‐103* or JM103* or JM‐105 or bilirubinometre* or bilirubinometer* or Bilitest or jaundice meter* or jaundicemeter* or icterometer*) AND CENTRAL:TARGET	113
9	#8 OR #7	138
10	MESH DESCRIPTOR Infant, Newborn EXPLODE ALL AND CENTRAL:TARGET	17947
11	MESH DESCRIPTOR Intensive Care, Neonatal EXPLODE ALL AND CENTRAL:TARGET	361
12	MESH DESCRIPTOR Intensive Care units, Neonatal EXPLODE ALL AND CENTRAL:TARGET	905
13	MESH DESCRIPTOR Gestational Age EXPLODE ALL AND CENTRAL:TARGET	2938
14	MESH DESCRIPTOR Infant, Premature EXPLODE ALL AND CENTRAL:TARGET	4263
15	MESH DESCRIPTOR Hyperbilirubinemia, Neonatal EXPLODE ALL AND CENTRAL:TARGET	377
16	(babe or babes or baby* or babies or gestational age? or infant? or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born? or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term? or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs) AND CENTRAL:TARGET	102544
17	#15 OR #14 OR #13 OR #12 OR #11 OR #10 OR #16	102544
18	#1 AND #17	111
19	#6 AND #17	64
20	#9 AND #17	85
21	#18 OR #19 OR #20	144

Appendix 2. MEDLINE strategy

	Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations, Daily and Versions 1946 to 18 August 2022
#	Searches	Results
1	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj4 bilirubin*).ti,ab,kw,kf. [Keyword Transcutaneous Bilirubinometry and Synonyms]	610
2	exp bilirubin/ [no Mesh for bilirubin blood level]	25906
3	(bilirubin* or bliverdin?).ti,ab,kw,kf.	43249
4	Administration, Cutaneous/ or ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj3 (administ* or device or devices or measur* or meter* or metre* or monitor? or test*)).ti,ab,kw,kf.	101466
5	(or/2‐3) and 4 [Bilirubin and Cutaneous Administration]	369
6	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj3 bilirubin* adj3 (device or devices or measur* or meter* or metre* or monitor? or test*)).ti,ab,kw,kf.	243
7	(TcB or bilicare? or bilicheck* or bilichek or Bilicheck* or BiliMed or bilmeter* or JM‐103* or JM103* or JM‐105 or bilirubinometre* or bilirubinometer* or Bilitest or jaundice meter* or jaundicemeter* or icterometer*).ti,ab,kw,kf.	1463
8	or/6‐7 [Cutaneous Bilirubin OR devices]	1524
9	exp Infant, Newborn/ or Intensive Care, Neonatal/ or Intensive Care Units, Neonatal/ or Gestational Age/ or exp Infant, Premature, Diseases/ or exp Hyperbilirubinemia, Neonatal/	710340
10	(babe or babes or baby* or babies or gestational age? or infant? or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born? or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term? or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs).ti,ab,kw,kf.	1006385
11	or/9‐10 [Filter: Neonatal Population 04‐2022‐MEDLINE]	1317137
12	exp animals/ not humans/	5038073
13	(1 and 11) not 12 [Cutaneous Bilirubinometry Keyword]	543
14	(5 and 11) not 12 [Bilirubin and Cutaneous Administration and Neonate]	273
15	(8 and 11) not 12 [Cutaneous Bilirubin OR devices AND Neonate]	479
16	or/13‐15	635

 

Appendix 3. Embase strategy

	Embase 1974 to 2022 August 18
#	Searches	Results
1	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj4 bilirubin*).ti,ab,kw,kf. [Keyword Transcutaneous Bilirubinometry and Synonyms]	865
2	exp bilirubin/ or exp bilirubin blood level/	106013
3	cutaneous drug administration/ or ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj3 (administ* or device or devices or measur* or meter* or metre* or monitor? or test*)).ti,ab,kw,kf.	113275
4	2 and 3 [Bilirubin and Cutaneous Administration]	600
5	((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) adj3 bilirubin* adj3 (device or devices or measur* or meter* or metre* or monitor? or test*)).ti,ab,kw,kf.	342
6	(TcB or bilicheck* or bilichek or Bilicheck* or BiliMed or bilmeter* or JM‐103* or JM103* or JM‐105 or bilirubinometre* or bilirubinometer* or Bilitest or jaundice meter* or jaundicemeter* or icterometer*).ti,ab,kw,kf.	1929
7	or/4‐6 [Biirubin, cutaneous OR devices]	2286
8	newborn/ or prematurity/ or newborn intensive care/ or newborn care/ or gestational age/	753504
9	(babe or babes or baby* or babies or gestational age? or infant? or infantile or infancy or low birth weight or low birthweight or neonat* or neo‐nat* or newborn* or new born? or newly born or premature or pre‐mature or pre‐matures or prematures or prematurity or pre‐maturity or preterm or preterms or pre term? or preemie or preemies or premies or premie or VLBW or VLBWI or VLBW‐I or VLBWs or LBW or LBWI or LBWs or ELBW or ELBWI or ELBWs or NICU or NICUs).ti,ab,kw,kf.	1187262
10	neonatal hyperbilirubinemia/ or newborn jaundice/	7762
11	or/8‐10 [Filter: Neonatal Population 03‐2022‐OVID EMBASE]	1446934
12	(exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/ or nonhuman/) and (human/ or normal human/ or human cell/)	23985429
13	(exp animals/ or exp invertebrate/ or animal experiment/ or animal model/ or animal tissue/ or animal cell/ or nonhuman/) not 12	6960458
14	(1 and 11) not 13 [ Keyword Transcutaneous Bilirubinometry AND Neonate]	739
15	(4 and 11) not 13 [Bilirubin and Cutaneous Administration AND Neonate]	353
16	(7 and 11) not 13 [Biirubin, cutaneous OR devices AND Neonate]	680
17	or/14‐16 [Results‐Embase]	849

 

Appendix 4. CINAHL strategy

	CINAHL Complete EBSCOhost 18 August 2022
#	Searches	Results
1	TI ( ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) N4 bilirubin*) ) OR AB ( ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) N4 bilirubin*) )	257
2	(MH "Bilirubin")	3263
3	TI ( (bilirubin* or bliverdin*) ) OR AB ( (bilirubin* or bliverdin*) )	5841
4	TI ( ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) N3 (administ* or device or devices or measur* or meter* or metre* or monitor? or test*)) ) OR AB ( ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) N3 (administ* or device or devices or measur* or meter* or metre* or monitor? or test*)) )	12693
5	(S2 OR S3) AND S4	142
6	TI ( ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) N4 bilirubin* N3 (device or devices or measur* or meter* or metre* or monitor? or test*)) ) OR AB ( ((cutaneous* or dermal* or epicutaneous* or percutaneous* or skin or transcutaneous* or trans‐cutaneous or subcutaneous* or sub‐cutaneous*) N4 bilirubin* N3 (device or devices or measur* or meter* or metre* or monitor? or test*)) )	118
7	TI ( (TcB or bilicare? or bilicheck* or bilichek or Bilicheck* or BiliMed or bilmeter* or JM‐103* or JM103* or JM‐105 or bilirubinometre* or bilirubinometer* or Bilitest or jaundice meter* or jaundicemeter* or icterometer*) ) OR AB ( (TcB or bilicare? or bilicheck* or bilichek or Bilicheck* or BiliMed or bilmeter* or JM‐103* or JM103* or JM‐105 or bilirubinometre* or bilirubinometer* or Bilitest or jaundice meter* or jaundicemeter* or icterometer*) )	235
8	S6 OR S7	277
9	S1 OR S5 OR S8	362

Appendix 5. Trial registry strategies

Date	Site	Terms	Results
August‐18‐2022	clinicaltrials.gov	transcutaneous bilirubin* [Other terms]	34
August‐18‐2022	clinicaltrials.gov	cutaneous bilirubin* [Other terms] AND Child brith‐17	11
August‐18‐2022	clinicaltrials.gov	transcutaneous bilirubinometry [Other terms]	11
August‐18‐2022	clinicaltrials.gov	bilicheck [Intervention]	0
August‐18‐2022	clinicaltrials.gov	bilicare [Intervention]	0
August‐18‐2022	clinicaltrials.gov	bilicheck [other terms]	2
August‐18‐2022	clinicaltrials.gov	BiliMed [Other terms]	0
August‐18‐2022	ICTRP	transcutaneous bilirubin [Intervention]	6
August‐18‐2022	ICTRP	transcutaneous bilirubinometry [Intervention]	1
August‐18‐2022	ICTRP	bilichek [Intervention]	0
August‐18‐2022	ICTRP	jaundice meter [Intervention]	0
August‐18‐2022	ICTRP	bilicare [Intervention]	1
August‐18‐2022	ICTRP	bilicheck [Intervention]	0
August‐18‐2022	ISRCTN	transcutaneous bilirubinometry [Text search]	0
August‐18‐2022	ISRCTN	transcutaneous bilirubinometry [Intervention	0
August‐18‐2022	ISRCTN	bilicheck [Text]	0
August‐18‐2022	ISRCTN	bilichek [Text]	0
August‐18‐2022	ISRCTN	bilicare [Text or Intervention]	0
Total			66

Appendix 6. QUADAS‐2 tool

QUADAS‐2 tool: risk of bias and applicability judgements

Domain 1: Patient selection
A. Risk of bias
Describe methods of patient selection:
1. Was a consecutive or random sample of patients enrolled?	Yes/No/Unclear "Yes" if it is clearly stated in the paper that a consecutive or a random sample of patients was enrolled. "No" if the patients were not recruited consecutively or the sample was not random. "Unclear" if there is insufficient information to answer "yes" or "no".
2. Was a case‐control design avoided?	Yes/No/Unclear The answer will always be "yes", as this review will exclude case‐control studies.
3. Did the study avoid inappropriate exclusions?	Yes/No/Unclear "Yes" if the stated inclusion and exclusion criteria are clear and appropriate. "No" if the stated inclusion and exclusion criteria include inappropriate subjects. "Unclear" if insufficient information is available to answer "yes" or "no".
Could the selection of patients have introduced bias?	RISK: Yes/No/Unclear "No" if questions 1 and 3 are answered "yes" (Low risk) "Yes" if questions 1 or 3 is answered "no" (High risk) "Unclear" if insufficient information is available to answer questions 1 or 3.
B. Concerns regarding applicability
Describe included patients (prior testing, presentation, intended use of index test and setting):
Is there concern that the included patients do not match the review question?	CONCERN: Yes/No/Unclear "No" when the study population represents an unselected sample of newborns expected to receive TcB assessment for hyperbilirubinaemia (Low). "Yes" if included patients are inherently different from those expected to receive TcB assessment for hyperbilirubinaemia (High). "Unclear" if there is insufficient information to make a judgement on the patient inclusion (Unclear).
Domain 2: Index test(s) (if more than 1 index test was used, please complete for each test)
A. Risk of bias
Describe the index test and how it was conducted and interpreted:
1. Were the index test results interpreted without knowledge of the results of the reference standard?	Yes/No/Unclear "Yes" if the paper states that the index test is interpreted by individual(s) who were unaware of the results of the reference test(s). "No" if the results of the index test were known by the individuals performing the reference test, or if the same individual performed both tests. "Unclear" if there is insufficient information to answer "yes" or "no".
2. If a threshold was used, was it pre‐specified?	Yes/No/Unclear "Yes" if a prespecified positivity threshold was stated. "No" if a threshold was not prespecified. "Unclear" if there is insufficient information to answer "yes" or "no".
Could the conduct or interpretation of the index test have introduced bias?	RISK: Low/High/Unclear "No" if questions 1 and 2 are answered "yes" (Low risk). "Yes" if questions 1 or 2 is answered "no" (High risk). "Unclear" if there is insufficient information available to answer "yes" or "no" (Unclear risk).
B. Concerns regarding applicability
Is there concern that the index test, its conduct or interpretation differ from the review question?	CONCERN: Low/High/Unclear "No" if there are no concerns based on the information available (Low). "Yes" if the index test is not TcB measurement for hyperbilirubinaemia in newborns or if the conduct of the test or its interpretation is not applicable to the review question (High). "Unclear" if there is insufficient information to answer "yes" or "no".
Domain 3: Reference standard
A. Risk of bias
Describe the reference standard and how it was conducted and interpreted:
1. Is the reference standard likely to correctly classify the target condition?	Yes/No/Unclear "Yes" if the reference standard is TSB measured by one of the laboratory methods mentioned in this protocol. "No" if the above condition is not met. "Unclear" if there is insufficient information to answer "yes" or "no".
2. Were the reference standard results interpreted without knowledge of the results of the index test?	Yes/No/Unclear "Yes" if it is stated that the individual performing/interpreting the results of the reference standard was kept unaware of the results of the index test. "No" if the results of the TcB measurement were known to the individual performing/interpreting the reference standard. "Unclear" if there is insufficient information to answer "yes" or "no".
Could the reference standard, its conduct or its interpretation have introduced bias?	RISK: Low/High/Unclear "No" if questions 1 and 2 are answered "yes" (Low risk). "Yes" if question 1 or 2 is answered "no" (High risk). "Unclear" if there is insufficient information to answer "yes" or "no".
B. Concerns regarding applicability
Is there concern that the target condition as defined by the reference standard does not match the review question?	CONCERN: Low/High/Unclear "No" if the target condition is hyperbilirubinaemia in newborns (Low). "Yes" if the target condition is not hyperbilirubinaemia in newborns, or it is not clearly stated (High). "Unclear" if there is insufficient information available to answer "yes" or "no" (Unclear).
Domain 4: Flow and timing
A. Risk of bias
Describe any patients who did not receive the index test(s) and/or reference standard or who were excluded from the 2 × 2 table (refer to flow diagram):     Describe the time interval and any interventions between index test(s) and reference standard:
1. Was there an appropriate interval between index test(s) and reference standard?	Yes/No/Unclear "Yes" if the time between the index test and the reference standard is less than 30 minutes. "No" if the time between the index test and the reference standard is longer than 30 minutes for a significant proportion of the patients. "Unclear" if insufficient information is available to answer "yes" or "no".
2. Did all patients receive a reference standard?	Yes/No/Unclear "Yes" if all the patients who received the index test received a reference standard. "No" if not all the patients who received the index test received a reference standard. "Unclear" if there is insufficient information to answer "yes" or "no".
3. Did patients receive the same reference standard?	Yes/No/Unclear "Yes" if the same reference standard was used for all patients. "No" if different reference standards were used. "Unclear" if there is insufficient information to answer "yes" or "no".
4. Were all patients included in the analysis?	Yes/No/Unclear "Yes" if all patients were included in the analysis, or if any withdrawals or exclusions are adequately explained with a flow chart. "No" if withdrawals or exclusions are not explained or accounted for. "Unclear" if there is insufficient information to answer "yes" or "no".
Could the patient flow have introduced bias?	RISK: Low/High/Unclear "No" if questions 1, 2, 3 and 4 are answered "yes" (Low risk). "Yes" if any of the four questions is answered "no" (High risk). "Unclear" if there is insufficient information to answer "yes" or "no" (Unclear risk).

TcB: transcutaneous bilirubin; TSB: total serum bilirubin.

Data

Presented below are all the data for all of the tests entered into the review.

Tests. Data tables by test.

Test	No. of studies	No. of participants
1 JM 101	7	859
2 JM 102	1	64
3 BiliChek (cutoff > 250 μmol/L)	4	1263
4 BiliChek (cutoff > 200 μmol/L)	3	1143
5 JH20‐1C	1	217
6 BiliChek HR; diazo	2	369
7 BiliChek HR; vitros	1	131
8 JM 103 (TcB > 204 μmol/L)	1	455
9 JM 103 (TcB > 230 μmol/L)	2	568
10 JM 103 (sternum TcB P95)	1	123
11 JM 103 (sternum TcB P75)	2	455
12 JM 103 (forehead TcB P95)	1	123
13 JM 103 (forehead TcB P75)	1	123
14 BiliChek (need for phototherapy)	1	400
15 Bilitest (TSB > 15 mg/dL)	1	241

1. Test.

JM 101

2. Test.

JM 102

3. Test.

BiliChek (cutoff > 250 μmol/L)

4. Test.

BiliChek (cutoff > 200 μmol/L)

5. Test.

JH20‐1C

6. Test.

BiliChek HR; diazo

7. Test.

BiliChek HR; vitros

8. Test.

JM 103 (TcB > 204 μmol/L)

9. Test.

JM 103 (TcB > 230 μmol/L)

10. Test.

JM 103 (sternum TcB P₉₅)

11. Test.

JM 103 (sternum TcB P₇₅)

12. Test.

JM 103 (forehead TcB P₉₅)

13. Test.

JM 103 (forehead TcB P₇₅)

14. Test.

BiliChek (need for phototherapy)

15. Test.

Bilitest (TSB > 15 mg/dL)

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bental 2009.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 628 Gestational age: term (> 37 weeks) and late preterm (> 35 weeks) Race/ethnicity: mixed Prior phototherapy: no Setting: Laniado Hospital, Natanya, Israel
Index tests	TcB device: JM 103 Site of TcB measurement: forehead and sternum
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: colorimetric assay
Flow and timing	Time interval between tests: ≤ 90 minutes Number of paired results: 1091
Comparative
Notes	Participants were healthy newborns.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Bertini 2008.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 241 Gestational age: > 32 weeks Race/ethnicity: "Caucasian" (presumed white; 100%) Prior phototherapy: no Setting: neonatal nursery of Careggi University Hospital, Florence, Italy
Index tests	TcB device: Bilitest BB 77 Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 10 minutes Number of paired results: not reported
Comparative
Notes	Participants were healthy newborns. TSB was determined according to the judgement of the physician in charge.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Bhutta 1991.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 63 Gestational age: term and preterm (> 34 weeks) Race/ethnicity: Pakistani Prior phototherapy: not reported Setting: newborn nursery of Aga Khan University Hospital, Karachi, Pakistan
Index tests	TcB device: JM 101 Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: photometry
Flow and timing	Time interval between tests: ≤ 2 hours Number of paired results: 100
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	No
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Bilgen 1998.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 96 Gestational age: term Race/ethnicity: not specified Prior phototherapy: no Setting: Bakirköy Maternity and Children's Hospital, Marmara, Türkiye
Index tests	TcB device: not specified, probably JM 101 Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 96
Comparative
Notes	Participants included term infants from 1 to 5 days of life who were clinically jaundiced and required blood draw for serum bilirubin measurement.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Boo 2007.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 345 Gestational age: > 37 weeks Race/ethnicity: Malay, Chinese, Indian Prior phototherapy: no Setting: postnatal ward and NICU at the Hospital Universitii Kebangsaan, Malaysia
Index tests	TcB device: BiliChek Site of TcB measurement: sternum and forehead
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: diazo method
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 345
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Campbell 2011.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 430 Gestational age: term and preterm (35–37 weeks) Race/ethnicity: diverse population (white, Black, Asian, Indian, Latino) Prior phototherapy: no Setting: Academic Hospital in Toronto, Ontario, Canada
Index tests	TcB device: BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 430
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Christo 1988.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 138 Gestational age: term and preterm Race/ethnicity: South Indian Prior phototherapy: in some infants Setting: Kasturba Hospital, India
Index tests	TcB device: JM101 Site of TcB measurement: forehead and sternum
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 138
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Ercan 2018.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 218 newborns  Gestational age: ≥ 35 weeks  Race/ethnicity: not reported  Prior phototherapy: no Setting: not reported
Index tests	TcB device: JH20‐1C Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 10 minutes Number of paired results: 271
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Hemmati 2013.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 560 Gestational age: mostly term and late preterm (≥ 35 weeks), only 5 infants ≤ 35 weeks Race/ethnicity: Iranian Prior phototherapy: no Setting: neonatal ward of Nemazee Hospital, Shiraz, Iran
Index tests	TcB device: BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: spectrophotometry and diazo method
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: not reported
Comparative
Notes	Participants were healthy newborns.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	No
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	No
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Karon 2008.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 177 Gestational age: preterm (> 32 weeks) and term (> 37 weeks) Race/ethnicity: "Caucasian" (presumed white), Asian, Hispanic, African American Prior phototherapy use: no Setting: well‐infant nursery at the Mayo Clinic (Rochester, MN)
Index tests	TcB device: BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: diazo and vitro method
Flow and timing	Time intervalbetween tests: ≤ 30 minutes Number of paired results: 177
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	No
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Karrar 1989.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 155 Gestational age: term (> 37 weeks) Race/ethnicity: Saudi newborns Prior phototherapy: no Setting: King Abdulaziz and King Khaled teaching hospitals, Riyadh, Saudi Arabia
Index tests	TcB device: JM 101 Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal jaundice TSB measurement method: American optical bilirubinometer
Flow and timing	Time interval between tests: not reported (probably ≤ 30 minutes) Number of paired results: 155
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Kitsommart 2013.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 405 Gestational age: preterm (≥ 34 weeks) and term (≥ 37 weeks) Race/ethnicity: Asian newborns Prior phototherapy: in some infants Setting: outpatient paediatric department at Siriraj Hospital, Bangkok, Thailand
Index tests	TcB Device: JM 103 Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: diazo method
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 455
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Knupfer 2001.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 135 Gestational age: preterm (23–36 weeks) Race/ethnicity: "Caucasians" (presumed white) and Asians Prior phototherapy: in some infants Setting: obstetrics department and NICU at the University of Leipzig, Leipzig, Germany
Index tests	TcB Device: BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: diazo method (unclear if reference standard interpreted without knowledge of results of index test)
Flow and timing	Time interval between tests: not reported Number of paired results: 400
Comparative
Notes	.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Kolman 2007.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 192 Gestational age: preterm (≥ 35 weeks) and term Race/ethnicity: Hispanic newborns Prior phototherapy: no Setting: newborn nursery at Maricopa Medical Center in Phoenix, AZ
Index tests	TcB device: BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: diazo method
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 192
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	No
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Lam 2008.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 113 Gestational age: term and late preterm (> 35 weeks) Race/ethnicity: Asian (Chinese) Prior phototherapy: no Setting: accident and emergency department of a regional hospital in Hong Kong
Index tests	TcB device: JM 103 Site of TcB measurement: mid‐sternum and forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 5 minutes Number of paired results: 113
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Maisels 1982.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 157 Gestational age: term (≥ 37 weeks) Race/ethnicity: white Prior phototherapy: no Setting: well‐baby nursery of Milton S. Hershey Medical Center, Hershey, PA
Index tests	TcB device: JM 101 Site of TcB measurement: forehead and sternum
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: diazo method
Flow and timing	Time interval between tests: not reported (probably near simultaneous measurement) Number of paired results: 292
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	No
Could the patient flow have introduced bias?		Low risk

Mendoza‐Chuctaya 2021.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 123 Gestational age: term Race/ethnicity: not reported Prior phototherapy: no Setting: Hospital Regional in Cusco‐Peru
Index tests	TcB Device: JM 103 Site of TcB measurement: sternum and forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: not specified
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 123
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	No
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Unclear
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Neocleous 2014.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 222 Gestational age: term (> 37 weeks) Race/ethnicity: Greek; "Caucasian" (presumed white) Prior phototherapy use: not reported Setting: department of medical paediatrics, General Hospital, Drama, Greece
Index tests	TcB device: BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 20 minutes Number of paired results: 368
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Unclear risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Samanta 2004.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 300 Gestational age: preterm (≥ 33 weeks) and term (≥ 37 weeks) Race/ethnicity: not reported Prior phototherapy: no Setting: postnatal wards at Liverpool Women’s Hospital, Liverpool, UK
Index tests	TcB device: BiliChek Site of TcB measurement: not reported
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: diazo method (unclear if reference standard interpreted without knowledge of results of index test)
Flow and timing	Time interval between tests: near simultaneous measurements Number of paired results: 300
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Schumacher 1985.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 106 Gestational age: > 36 weeks Race/ethnicity: white Prior phototherapy: no Setting: not stated
Index tests	TcB device: JM 101 Site of TcB measurement: sternum
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: direct spectrophotometry
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 106
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Unclear risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Taha 1984.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 68 Gestational age: term (> 37 weeks) Race/ethnicity: Saudi newborns Prior phototherapy: no Setting: newborn nursery at King Abdul Aziz University Hospital Riyadh, Saudi Arabia
Index tests	TcB device: JM 101 Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: direct spectrophotometry
Flow and timing	Time interval between tests: near simultaneous measurements Number of paired results: 120
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Unclear risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Wong 2002.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 64 Gestational age: term and preterm (31–42 weeks) Race/ethnicity: mostly white, 6 non‐white Prior phototherapy: no Setting: tertiary neonatal service in South‐East Scotland
Index tests	TcB device: JM 102 and BiliChek Site of TcB measurement: forehead
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: spectrophotometry
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 64
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Unclear risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

Yaser 2014.

Study characteristics
Patient Sampling	Cross‐sectional study, consecutive sampling
Patient characteristics and setting	Sample size: 122 Gestational age: preterm (< 35 weeks) Race/ethnicity: not reported Phototherapy use: no prior phototherapy use Setting: level III neonatal care unit (Groote Schuur Hospital) in Cape Town, South Africa
Index tests	TcB device: JM 103 Site of TcB measurement: forehead, sternum, interscapular area
Target condition and reference standard(s)	Target condition: neonatal hyperbilirubinaemia TSB measurement method: not reported
Flow and timing	Time interval between tests: ≤ 30 minutes Number of paired results: 122 for each TcB site
Comparative
Notes
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient Selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	Yes
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Low risk
Are there concerns that the included patients and setting do not match the review question?			Low concern
DOMAIN 3: Reference Standard
Is the reference standards likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Unclear
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and Timing
Was there an appropriate interval between index test and reference standard?	Yes
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk

HPLC: high‐performance liquid chromatography; NICU: neonatal intensive care unit; TcB: transcutaneous bilirubin; TSB: total serum bilirubin.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Adriaansen 2020	Insufficient information to generate 2 × 2 table; correlation coefficients reported.
Afanetti 2014	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Agrawal 2019	Insufficient information to generate 2 × 2 table.
Ahmed 2010	Study used a series of cut‐offs points to generate ROC curve; but there was insufficient information for any specific cut‐off point to generate a 2 × 2 table
Ahn 2003	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Akahira‐Azuma 2013	Insufficient information to generate 2 × 2 table.
Alsaedi 2016	Insufficient information to generate 2 × 2 table.
Alsaedi 2018	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Amato 1985	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Amato 1990	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Arman 2020	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Babaei 2012	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Badiee 2012	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Beck 2003	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Berget 1984	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Bhat 1987	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Bhutani 2000	Study evaluated BiliChek for predicting risk of subsequent excessive hyperbilirubinaemia (not a focus of this review).
Boo 1984	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Bosschaart 2012	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Bourchier 1987	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Briscoe 2002	TcB device and unit of measurement for TcB not reported.
Carbonell 1999	Duplicate (See Carbonell 2001).
Carbonell 2001	Study evaluated predictive ability of TcB measurement for subsequent hyperbilirubinaemia (not a focus of this review).
Carceller 2006	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Casnocha 2021	Insufficient information to generate 2 × 2 table.
Castro 2019	Wrong study design.
Cat 2021	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Chang 2006	We were unable to find the full text article.
Chawla 2014	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Chimhini 2018	Insufficient information to generate 2 × 2 table.
Chokemungmeepisarn 2020	Insufficient information to generate 2 × 2 table.
Conceicao 2014	Study only reported correlation coefficient, not sensitivity and specificity.
Dagli 2018	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Dai 1996	Retrospective study.
De Luca 2007	Insufficient information to generate 2 × 2 table.
De Luca 2008b	Insufficient information to generate 2 × 2 table.
Dianova 2021	Study only reported correlation coefficient, not sensitivity and specificity.
Dominguez 1993	Study only reported correlation coefficient, not sensitivity and specificity.
Donzelli 2000	Insufficient information to generate 2 × 2 table.
Ebbesen 2002	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Ebbesen 2012	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
El‐Kabbany 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Engle 2002	Multiple cut‐off values for TSB; insufficient information to generate 2 × 2 table.
Engle 2005	Multiple cut‐off values for TSB; insufficient information to generate 2 × 2 table.
Fabris 1984	Study only reported correlation coefficient, not sensitivity and specificity.
Faridi 1998	No full‐text article available.
Felc 2005	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Fisher 2015	No full‐text article available at time of writing of this review.
Foged 1983	Study only reported correlation coefficient, not sensitivity and specificity.
Fok 1986	Insufficient information to generate 2 × 2 table.
Fonseca 2012	Wrong study design.
Goldman 1982	TcB used to identify which infants required TSB determinations rather than to predict actual bilirubin values.
Gondale 2013	TcB device not reported.
Gothwal 2021	No accuracy outcomes such as sensitivity and specificity.
Grabenhenrich 2014	Study only reported correlation coefficient, not sensitivity and specificity.
Greco 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Gunaseelan 2017	Insufficient information to generate 2 × 2 table.
Harish 1998	Multiple cutoff points; insufficient information to generate 2 × 2 table.
Hasani 2019	No full‐text article available.
Hegyi 1981	Sensitivity and specificity not reported.
Heick 1982	Study only reported correlation coefficient, not sensitivity and specificity.
Hemmati 2021	Study only reported correlation coefficient, not sensitivity and specificity.
Hill 2016	We were unable to find the full‐text article.
Ho 2006	Insufficient information to generate 2 × 2 table.
Ho 2006a	Retrospective study.
Holland 2009	Study used multiple cutoff points and 3 different gold standards, making it difficult to choose a specific cutoff for comparison of TcB for any of the gold standards.
Hoppenot 2012	Retrospective study.
Hulzebos 2019	Multiple threshold values for TSB for multiple different categories of participants based on the birth weight: difficult to choose a specific threshold for comparison.
Itoh 2001	Study only reported correlation coefficient, not sensitivity and specificity.
Jahadi 2013	Study only reported correlation coefficient, not sensitivity and specificity.
Jalalodini 2019	Insufficient information to generate 2 × 2 table; participants were already receiving phototherapy.
Jangaard 2006	Study only reported correlation coefficient, not sensitivity and specificity.
Janjindamai 2005	Study only reported correlation coefficient, not sensitivity and specificity.
Jegathesan 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Jegathesan 2021	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Jeon 2020	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Jnah 2018	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Johnson 2020	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Jones 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Joseph 2014	No full‐text article available.
Juster‐Reicher 2014	Study only reported correlation coefficient, not sensitivity and specificity.
Karen 2009	Study only reported correlation coefficient, not sensitivity and specificity.
Karolyi 2004	Insufficient information to generate 2 × 2 table.
Kaynak‐Turkmen 2011	Multiple cutoff values; insufficient information to generate 2 × 2 table.
Kazmierczak 2004	Study only reported correlation coefficient, not sensitivity and specificity.
Keshini 2018	No full‐text article available
Keshishian 1994	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Khajehei 2021	Insufficient information to generate 2 × 2 table.
Knudsen 1996	Study only reported correlation coefficient, not sensitivity and specificity.
Konana 2021	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Kosarat 2013	Study only reported correlation coefficient, not sensitivity and specificity.
Kumar 1994	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Kumar 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Kumar 2020	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Kumra 2017	Insufficient information to generate 2 × 2 table.
Laeeq 1993	Insufficient information to generate 2 × 2 table.
Lebrun 1982	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Leite 2007	Insufficient information to generate 2 × 2 table.
Lin 1993	Insufficient information to generate 2 × 2 table.
López‐Garrido 2015	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Mahajan 2005	Wide range of participants (28–42 weeks' gestational age); number of paired comparisons for TcB and TSB not reported; method of TSB measurement not reported.
Mahram 2015	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Maisels 2004	Multiple TcB threshold values; insufficient information to generate 2 × 2 table.
Maisels 2006	Study only reported correlation coefficient, not sensitivity and specificity.
Maisels 2011	Multiple target points and cutoff values for TSB and TcB.
Marco 2009	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Maya‐Enero 2021	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Mercanti 2007	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Mohamed 2022	Insufficient information to generate 2 × 2 table.
Moscicka 1994	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Murli 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Nahar 2017	Multiple cutoff values; insufficient information to generate 2 × 2 table.
Namba 2007	Study only reported correlation coefficient, not sensitivity and specificity.
Nanjundaswamy 2004	Study only reported correlation coefficient, not sensitivity and specificity.
Nanjundaswamy 2005	Wrong study design; infants were already receiving phototherapy.
Narang 1983	No full‐text article available.
Ochoa 2000	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Oyapero 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Pallas 1993	We were unable to find the full‐text article.
Palmer 1982	We were unable to find the full‐text article.
Panburana 2010	Multiple cutoff values; insufficient information to generate 2 × 2 table.
Pendse 2017	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Poland 2004	Study only reported correlation coefficient, not sensitivity and specificity.
Povaluk 2011	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Pratesi 2016	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Qandalji 2012	Poster presentation, no full‐text article available.
Qualter 2011	Insufficient information to generate 2 × 2 table.
Quist 2016	Study only reported correlation coefficient, not sensitivity and specificity.
Raba 2020	Study only reported correlation coefficient, not sensitivity and specificity.
Radfar 2016	Study only reported correlation coefficient, not sensitivity and specificity.
Raimondi 2012	Insufficient information to generate 2 × 2 table.
Reyes 2008	Study only reported correlation coefficient, not sensitivity and specificity.
Robertson 2002	Study only reported correlation coefficient, not sensitivity and specificity.
Rodriguez‐Capote 2009	Insufficient information to generate 2 × 2 table.
Rohsiswatmo 2018	Study only reported correlation coefficient, not sensitivity and specificity.
Romagnoli 2012	Study evaluated the use of TcB measurements plotted on bilirubin nomogram to predict subsequent significant hyperbilirubinaemia or need for phototherapy.
Romagnoli 2013	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Romo‐Pinales 1984	Study only reported correlation coefficient, not sensitivity and specificity.
Roy 2010	Study only reported correlation coefficient, not sensitivity and specificity.
Rubaltelli 2001	Insufficient information to generate 2 × 2 table.
Rubio 2017	Insufficient information to generate 2 × 2 table.
Rush 2011	Study only reported correlation coefficient, not sensitivity and specificity.
Rylance 2014	Insufficient information to generate 2 × 2 table.
Sajjadian 2012	Multiple cutoffs precluded extraction of data for 2 × 2 table.
Samiee‐Zafarghandy 2014	Insufficient information to generate 2 × 2 table.
Sanpavat 2004	Multiple cutoffs precluded extraction of data for 2 × 2 table.
Sanpavat 2005	Insufficient information to generate 2 × 2 table.
Sanpavat 2007	Insufficient information to generate 2 × 2 table.
Sardar 2021	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Sarici 2014	Study only reported correlation coefficient, not sensitivity and specificity.
Schmidt 2009	Insufficient information to generate 2 × 2 table.
Sellitto 2010	Study only reported correlation coefficient, not sensitivity and specificity.
Sharma 1988	Study only reported correlation coefficient, not sensitivity and specificity.
Sheridan‐Pereira 1982	Study only reported correlation coefficient, not sensitivity and specificity.
Shihadeh 2016	Study only reported correlation coefficient, not sensitivity and specificity.
Siu 2010	Study only reported correlation coefficient, not sensitivity and specificity.
Slusher 2004	Study only reported correlation coefficient, not sensitivity and specificity.
Soffer 2018	Study only reported correlation coefficient, not sensitivity and specificity.
Spence 2016	No accuracy outcomes like sensitivity and specificity.
Srinivas 2016	Insufficient information to generate 2 × 2 table.
Stillova 2007	Study only reported correlation coefficient, not sensitivity and specificity.
Stillova 2009	Study only reported correlation coefficient, not sensitivity and specificity.
Suckling 1995	Study only reported correlation coefficient, not sensitivity and specificity.
Szabo 2004a	Inappropriate exclusions (newborns with visible jaundice in first 36 hours of life).
Szabo 2004b	Insufficient information to generate 2 × 2 table.
Tan 1982	Study only reported correlation coefficient, not sensitivity and specificity.
Tan 1985	Study only reported correlation coefficient, not sensitivity and specificity.
Tan 1988	Study only reported correlation coefficient, not sensitivity and specificity.
Tan 1996	Study only reported correlation coefficient, not sensitivity and specificity.
Tan 2003	Study only reported correlation coefficient, not sensitivity and specificity.
Tayaba 1998	Study only reported correlation coefficient, not sensitivity and specificity.
Taylor 2015	Study only reported correlation coefficient, not sensitivity and specificity.
Teran 2011	We were unable to find a full‐text article.
Timalsina 2017	We were unable to find a full‐text article.
Vocel 1985	Article published in Czech and not in English.
Vocel 1987	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Vocel 1988	Article published in Czech and not in English.
Wainer 2009	Study only reported correlation coefficient, not sensitivity and specificity.
Weber 2021	Multiple cutoff value; insufficient information to generate 2 × 2 table.
Wickremasinghe 2011	Study assessed the use of inpatient TcB measurement to predict subsequent bilirubin levels of high and high‐intermediate risk on Bhutani nomogram.
Willems 2004	Study only reported correlation coefficient, not sensitivity and specificity.
Willemsen 2007	Study not published in English.
Williams 2011	Study only reported correlation coefficient, not sensitivity and specificity.
Yamanouchi 1980	Study only reported correlation coefficient, not sensitivity and specificity.
Yamauchi 1988	Study only reported correlation coefficient, not sensitivity and specificity.
Yamauchi 1989a	Study only reported correlation coefficient, not sensitivity and specificity.
Yamauchi 1989b	Study only reported correlation coefficient, not sensitivity and specificity.
Yamauchi 1989c	Study only reported correlation coefficient, not sensitivity and specificity.
Yamauchi 1990	This study evaluated use of TcB measurement for prediction of subsequent hyperbilirubinaemia > 15 mg/dL.
Yamauchi 1991	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Yang 2019	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Yap 2002	Study only reported correlation coefficient, not sensitivity and specificity.
Yasuda 2003	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Ying 2020	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Zecca 2009	Study reported correlation coefficients without reporting sensitivity and specificity of TcB measurement.
Zhan 2016	Study only reported correlation coefficient, not sensitivity and specificity.

TcB: transcutaneous bilirubin; TSB: total serum bilirubin.

Differences between protocol and review

In the protocol, we stated we would search DARE and MEDION databases; however, we were unable to access these databases, and we omitted them from the Search methods for identification of studies section of the review (Okwundu 2017).

Contributions of authors

Conception of review: CO
Screening and data extraction: AO, CO, and OU
Statistical analysis: CO
Drafting of protocol and review: CO
Writing and approval of review: AO, CO, OU, KG, JS, CW, VB
Search strategies; results of search; PRISMA: MF

Sources of support

Internal sources

• No sources of support provided

External sources

• Vermont Oxford Network, USA

  Cochrane Neonatal Reviews are produced with support from Vermont Oxford Network, a worldwide collaboration of health professionals dedicated to providing evidence‐based care of the highest quality for newborn infants and their families.

• The South African Medical Research Council, South Africa

  The South African Medical Research Council provided funding to conduct this review.

• The Cochrane Incentive Fund, UK

  An annual scheme whereby small incentive payments are offered to Cochrane Review Groups (CRGs) for preparing key new or updated Cochrane reviews.

Declarations of interest

AO: none known
CO: none known
OU: none known
KG has declared he has no conflict of interest; but he is an Editor with Cochrane Neonatal Group. He was not be involved in the editorial process of this manuscript.
JS: none known
CW: none know
VB: none known
MF is an Information Specialist and Managing Editor with the Cochrane Neonatal Group. She was not involved in the editorial process of this manuscript.

New