Impact of a standardized protocol for the Management of Prolonged Neonatal Jaundice in a regional setting: an interventional quasi-experimental study

Background Prolonged neonatal jaundice (PNNJ) is often caused by breast milk jaundice, but it could also point to other serious conditions (biliary atresia, congenital hypothyroidism). When babies with PNNJ receive a routine set of laboratory investigations to detect serious but uncommon conditions, there is always a tendency to over-investigate a large number of well, breastfed babies. A local unpublished survey in Perak state of Malaysia revealed that the diagnostic criteria and initial management of PNNJ were not standardized. This study aims to evaluate and improve the current management of PNNJ in the administrative region of Perak. Methods A 3-phase quasi-experimental community study was conducted from April 2012 to June 2013. Phase l was a cross-sectional study to review the current practice of PNNJ management. Phase ll was an interventional phase involving the implementation of a new protocol. Phase lll was a 6 months post-interventional audit. A registry of PNNJ was implemented to record the incidence rate. A self-reporting surveillance system was put in place to receive any reports of biliary atresia, urinary tract infection, or congenital hypothyroidism cases. Results In Phase I, 12 hospitals responded, and 199 case notes were reviewed. In Phase II, a new protocol was developed and implemented in all government health facilities in Perak. In Phase III, the 6-month post-intervention audit showed that there were significant improvements when comparing mean scores of pre- and post-intervention: history taking scores (p < 0.001), family history details (p < 0.05), physical examination documentation (p < 0.001), and total investigations done per patient (from 9.01 to 5.81, p < 0.001). The total number of patient visits reduced from 2.46 to 2.2 per patient. The incidence of PNNJ was found to be high (incidence rate of 158 per 1000 live births). Conclusions The new protocol standardized and improved the quality of care with better clinical assessment and a reduction in unnecessary laboratory investigations. Trial registration Research registration number: NMRR-12-105-11288. Electronic supplementary material The online version of this article (10.1186/s12887-019-1550-3) contains supplementary material, which is available to authorized users.


Background
Prolonged neonatal jaundice (PNNJ) is defined as visible jaundice with yellowish staining of the skin, mucous membrane and conjunctival icterus or serum bilirubin > 85 μmol/L that persists beyond 14 days of life in a term baby and 21 days in a preterm baby [1,2].
Breast milk jaundice is the most common cause of PNNJ. It is almost always benign [3], presenting as prolonged unconjugated hyperbilirubinaemia, and occurs in up to one-third of healthy breastfed newborns [4]. It develops as the result of poor calories intake associated with breast-feeding difficulties [5], liver immaturity and the inhibitory effect of mother's milk in the clearance of unconjugated bilirubin [6].
Prolonged neonatal jaundice could also be an early presentation of serious conditions such as biliary atresia [7,8]. The incidence of biliary atresia varies worldwide from 1 in 6000 live births in Taiwan, 1 in 12,000 in the United States, 1 in 17,000 in the United Kingdom [9], 1 in 18,000 in Europe and 1 in 19,000 in Canada [10]. It can rapidly lead to liver cirrhosis and liver failure if left untreated. However, outcomes of babies with biliary atresia benefit from early diagnosis and hence babies with prolonged neonatal jaundice, and specifically babies with prolonged conjugated hyperbilirubinaemia need to be investigated to rule out biliary atresia [11][12][13][14][15]. Other pathological causes of prolonged neonatal jaundice are urinary tract infection, sepsis, congenital hypothyroidism, metabolic and haemolytic disorders [16].
A local unpublished survey among medical officers and staff nurses in Perak region had shown that the knowledge of health care providers in managing babies with PNNJ was not satisfactory. Majority of the personnel from health clinics (71%) and hospital (83%) were unable to give accurate diagnostic criteria. The study also showed that the management of PNNJ was not standardised. Seventy percent of health staff would refer all babies straight to specialist hospitals without conducting a preliminary investigation while the remaining 30% would be conducting some investigations before referring. Over investigation among hospital staff occurred in 50%, with medical officers ordering 8 or more laboratory investigation in the initial workup [17].
The current Malaysian and many international guidelines (Additional file 2 Table S1) state that babies with PNNJ should receive a routine set of laboratory investigations to detect serious conditions such as biliary atresia, without much emphasis on clinical assessment. A large number of well breastfed babies will similarly undergo these tests. Of note, these tests include urine culture which sometimes leads on to unnecessary treatment because of the false positive result [18].
On the other hand, despite this standard list of laboratory investigations there were still missed cases and cases with late diagnosis of biliary atresia [19,20].
So how do we balance the need to detect serious conditions but not to over-investigate well babies? At present, the only effective method for early detection of biliary atresia is the universal stool colour screening and registry, which has been implemented in Taiwan in recent years [21]. In Bath (UK), the Paediatric Team in Royal United Hospital has adopted an approach where only babies with suspected abnormal clinical findings and history would be investigated further. Well, term babies will only require a simple laboratory investigation, which is total serum bilirubin with differential (direct and indirect levels) [22].
This successful strategy of using clinical assessment of risk factors in the evaluation and management of PNNJ has encouraged us to revise our pre-existing practices and develop a new protocol that standardises the management of prolonged neonatal jaundice.

Study design and setting
This was a community quasi-experimental study conducted in 3 phases from April 2012 to June 2013, aimed at evaluating and improving the management of PNNJ in the Perak region, Malaysia. A flow chart of the methodology is illustrated in Fig. 1.
In Phase I, the pre-interventional phase, the current practices of PNNJ management at the selected sites were assessed. For each site, 20 most recent PNNJ case notes were reviewed using a pre-tested data collection form (Additional file 1). This form evaluated how babies with PNNJ were diagnosed, assessed and managed. The form was filled in by paediatricians or by a medical officer under the supervision of visiting paediatricians if in nonspecialist hospitals.
Evaluation of patient history taking, family history taking and physical examinations (Additional file 1) obtained from the case notes were summed up in the form of a score. This score was based on important points that were relevant to the assessment of PNNJ. The maximum score was 5 for history taking, 4 for family history taking, and 5 for physical examination. The higher the score, the more complete was the clinical assessment.
In Phase II, the interventional phase, a new protocol was developed based on analysis of data from Phase I, extensive literature review (see Table 1 and Additional file 2 Table S1, Additional file 3 Table S2,  Additional file 4 Table S3, Additional file 5 Table S4 and Additional file 6 Table S5), and consensus from the stakeholders (local paediatricians, public health professionals, and local policy stakeholders). This new protocol focused on the management of PNNJ according to risk stratification and on educating parents about warning signs (unwell baby/ pale stool dark yellow urine/ new onset of jaundice/ persistent jaundice > 2 months). It comprised a flow chart for PNNJ management (Fig. 2) and an assessment form (Fig. 3), which were distributed together and was implemented in all 322 health clinics and 12 hospitals in the Perak region on September 2012.
In Phase III, post-interventional phase, an assessment was conducted in the 12 hospitals six months later. The same mechanism of data collection as in Phase I was repeated to evaluate 20 most recent PNNJ cases.
A self-reporting surveillance system was also created for paediatricians in the Perak state to report on any biliary atresia, urinary tract infection, or congenital hypothyroidism cases. Information, which included the age of the infant, when and how the diagnoses of those conditions were made, and whether the new protocol was used --was emailed or faxed to the author. This reporting was more of a voluntary than a compulsory one and attempted to identify any cases missed after the implementation of the new protocol. A regional level PNNJ registry was also introduced and initiated by the Perak state health department to collect monthly returns of PNNJ cases from all government health facilities.

Meeting and consensus among stakeholders
A total of 32 members consisting of paediatricians, public health professionals and local policy stakeholders (2 hospital directors, 6 paediatricians, 9 representatives from health district office/ health clinics, 1 pathologist, 1 laboratory technician, 5 hospital medical officers, 3 Clinical Research Centre, CRC, members, 5 staff nurses) met on 10th August 2012 to review the results of Phase I (see Results) and were guided by the findings of the literature review (see Additional file 2 Table S1, Additional  file 3 Table S2, Additional file 4 Table S3, Additional file  5 Table S4 and Additional file 6 Table S5).
Several issues in the management of PNNJ were identified from the data analysis from Phase I when compared with the latest evidence. The differences between the current practice and new protocol are outlined in Table 1. A new protocol was developed to address these issues. The new protocol, which was agreed upon during the consensus meeting, consists of a management flow chart which is based on risk stratification (Fig. 2), assessment form (Fig. 3) and parental education on warning signs; to guide the medical officer at primary care clinics  Total serum bilirubin with differential is needed at 14 days and only repeated as necessary Main aim of total serum bilirubin with differential is to pick up conjugated hyperbilirubinaemia [2] Heel-prick capillary bilirubin is not useful in the management of PNNJ.

Urine sampling
Babies with PNNJ undergo urine culture, whereby sampling is done by clean catch, bladder catheterization or suprapubic aspiration.
Only urine dipstick & microscopy test and is needed. Sampling via urine bag is acceptable. Urine culture will be considered for suspected cases [18].

Data analysis
Data collected were entered into Microsoft Excel 2010 and SPSS version 15.0 was used to analyse the data collected. Numerical variables were calculated as mean and standard deviation, and independent t-test was employed to compare mean generated from data collected from pre-and post-intervention. Categorical data collected was presented in the form of frequency and percentages. A p-value less than 0.05 was deemed to be statistically significant.
All data collected was kept confidential, and no unique identifiers were collected for PNNJ cases. Data presented did not specifically identify any clinic or hospital that responded to this study. If day-21-tests were normal, the baby could be discharged with warning signs and reviewed during routine medical examination at 1 and 2 months old.
This will reduce unnecessary investigations, clinic visits and improve compliance to follow up.
Abbreviations: T4 Thyroxine, TSH Thyroid-Stimulating Hormone, PNNJ Prolonged Neonatal Jaundice, UTI Urinary Tract Infection  The main ethical issue encountered before implementing the new protocol was the risk of not detecting serious conditions because total serum bilirubin with differentials was the only test screened in well, term breastfed babies. This risk was minimised with improved clinical assessment and warning signs education to all parents.

Results
The changes seen in practice after implementing the new protocol A total of 240 cases were collected from 12 hospitals of Perak state; however, only data from 199 case notes preintervention and 145 case notes post-intervention with the new protocol were included for analysis. Cases with incomplete data or not fulfilling the study inclusion criteria were excluded. The data gathered were compared and shown in Table 2.

a. Postnatal age upon referral
The mean postnatal age of babies upon referral has significantly increased from 16.54 days to 20.01 days (p = 0.001).

b. Clinical assessment of PNNJ
The average score for patient-related history taking, where the maximum achievable score was 5, had improved from 3.26 to 4.44 (p < 0.001). Mean score of family history taking, with a maximum score of 4, had increased from 0.53 to 2.14 (p < 0.001). The average physical examination score, with a maximum achievable score of 5, had risen from 3.8 to 4.5 (p < 0.001).

c. Number of laboratory Investigations done before referral to hospital
There was a significant reduction in the total number of investigations done before referral to the hospital (from 2.2 to 1.7, p = 0.020).

d. Number of laboratory investigations done at the hospital level
A total of 1758 tests were done in 199 patients preintervention compared to 811 tests in 145 patients postintervention. The new protocol had significantly reduced the total number of laboratory investigations done per patient at the hospital level, from 9.01 laboratory investigations per patient to 5.81 per patient (p < 0.001).

e. Type of laboratory investigations done at the hospital level
The type and number of laboratory investigations done at the hospital level was more reasonable after the implementation of the new protocol (see Table 3). Preintervention, out of a total of 1758 tests, total serum bilirubin without differential was the most frequently done (249 tests), followed by urine culture and sensitivity test (237 tests), liver function test (198 tests), and total serum bilirubin with differentials (194 tests). Postintervention, out of a total of 811 tests, the top four laboratory investigations done were total serum bilirubin with differentials (203 tests), urine dipstick & microscopy test (107 tests), free T4/TSH (105 tests), and full blood count (98 tests).

f. Total number of visits to the hospital
There was only a slight reduction of the mean number of total visits to the hospitals for PNNJ (from 2.46 to 2.20, p = 0.046).

Self-reporting surveillance system
As of September 2013, the self-reporting system recorded 13 cases of urinary tract infection, 4 cases of congenital hypothyroidism, and 3 cases of suspected biliary atresia, which on further investigations were found to be all detected by the new protocol.

Prolonged neonatal jaundice registry
Out of 9967 total live births, 1576 cases of prolonged neonatal jaundice were reported to the Prolonged Neonatal Jaundice Registry (Table 4). This gives a PNNJ incidence rate of 158 per 1000 live births. The majority of the cases were detected in Health Clinics (78.7%) and of all the cases detected, 92% were stratified as low risk, 4% as intermediate and 4% high risk (see Table 4).

Principal findings
There was no standardized management of PNNJ among the government hospitals and health clinics in Perak state. In addition to this, clinical assessments were incomplete.
A new protocol for the management of PNNJ was developed, which focused on risk stratification by good clinical assessment. The flow chart and the assessment form aided the health care providers to perform better history taking and examination; to have a clear guide on the necessary laboratory investigations, referral indications, and follow-up plans; and to give warning signs to the parents. From the literature review, the four most useful investigations in the initial management of PNNJ were found to be total serum bilirubin with differentials, Free T4/TSH, urine dipstick and microscopy, and full blood count plus reticulocyte counts. These tests were adapted into the new protocol.
There were significant improvements in certain aspects of PNNJ management after implementation of the new protocol. Skills in patient history and family history taking as well as physical examination showed significant improvement after implementing the new protocol. The number of laboratory investigations done before referring to the hospitals and at the hospital level was significantly reduced.
With a clear guidance from the new protocol, many well breastfed babies with PNNJ were currently managed at the local health clinics, and only babies with concerns or risk factors were referred to the hospitals --thus explaining the older postnatal age (by approximately four days) upon referral to hospitals post-intervention.
The choice of laboratory investigations for the initial management of PNNJ was also more rational postintervention. Total serum bilirubin without differential, which is known to be unhelpful to determine the causes of PNNJ, was frequently done (294 times in 199 babies) before implementation of the new protocol. Postintervention, this test was only done 39 times in 145 babies, a remarkable feat in diminishing the old practice. Of note, total serum bilirubin with differential was done at least twice the frequency of other tests post-intervention, reflecting the possibility that PNNJ in many babies resolved by the third week without warranting further tests.   We note that there was minimal reduction of the mean number of total visits to the hospitals. This small effect size could be explained by the fact that 9 out of 12 hospitals were non-specialist hospitals and postintervention they still served as a primary care centres for nearby patients.
Follow-up plans were unclear in the previous system. Most babies underwent repeated heel-prick capillary bilirubin till jaundice resolved. The new protocol enabled well, low-risk babies with normal tests results at day 21 to be discharged with warning signs and to be reviewed only at one-month and two-month old routine medical examination (RME). The requirement for less visits reduced the burden to parents and health care facilities while being sufficiently safe and effective.

Strengths and weaknesses of the study
This was the first study locally that looked into the burden and management of PNNJ, at both the primary care and tertiary care levels. In local terms, the strengths of the new protocol were standardising the approach to managing babies with PNNJ, empowering the health clinics to manage well babies, emphasising good clinical assessment, rationalizing the choice of investigations, giving guidance on how to follow-up well babies and creating a safety net by recommending warning signs for parents. Additionally, the state regional registry of prolonged neonatal jaundice had provided the incidence rate of PNNJ, which was not available nationally. The findings showed that prolonged neonatal jaundice is common and indirectly signifies a major workload.
The sampling method was the main limitation of this study. Case notes were conveniently selected by paediatricians during the pre-intervention phase, and this could result in sampling bias.

Implications for clinicians and policymakers
Clinical assessment, including the inspection of stool colour, remains the most important aspect in the management of prolonged neonatal jaundice. The essential laboratory investigation needed at two weeks of life is the total serum bilirubin with differentials, and further workup is required: a) in well, low risk babies who remain jaundice at 3 weeks of life (total serum bilirubin with differentials, urine dipstick plus microscopy, Free T4/ TSH, full blood count plus reticulocyte counts), or b) in the presence of positive clinical findings, at any age (the laboratory investigations above plus other relevant tests eg: full blood picture, liver function test etc.).
The reduction of the number of blood investigations and clinic visits by risk stratification approach has potential in improving quality of care for babies, parents' satisfaction and costs. Special thanks and appreciation to 1) Dr. Ranjit Kaur, State Health Deputy Director and Dr. Elya Zetti of the Family Health Development Unit in Perak who had actively supported the work and ensured the smooth implementation of the new protocol while spearheading the state PNNJ registry.
2) The staff nurses from the Paediatric Department of Hospital Slim River who had helped in the initial phase of proposal development and data collection.
3) All the Paediatricians and Family Medicine Specialists in Perak state who had helped in the data collection, implementation of the new protocol and had given valuable feedback regularly. 4) CRC member, Ooi Qing Xi for her involvement in data analysis and initial report writing. 5) Director General of the Ministry of Health Malaysia for his permission to publish the article.
Authors' contributions HS was the principle investigator for this study; the main person in reviewing all the latest evidence and creating the new protocol; and the main contributor in manuscript writing. IS first initiated the idea of this study and was in-charge during the proposal development phase and data collection. AS remained the main adviser during all the phases of the study, and also the second contributor in manuscript writing. ML was responsible for data collection, data entry and analysis. RK oversaw the implementation of the protocol in the region and started a PNNJ registry. CC was the third contributor for this manuscript writing. AY had contributed in the proposal development phase. FT and SV were responsible during the proposal development and data collection phase. All authors read and approved the final manuscript.