The standard of care for clinically-defined neonatal infections is diagnostic blood culture and antibiotic treatment [
14,
16]. This large, multi-country analysis of the gap between antibiotic prescribing and blood culture use for newborns, using routine data, revealed that for 144,146 newborns admitted across 61 hospitals, most were prescribed antibiotics (70%, 95%CI 65–75), but very few had blood cultures (6%, 95%CI 3–10). Neonatal unit blood culture use for newborns prescribed antibiotics varied (0–65%), and none met the quality-of-care target of ~ 100% [
16]. Positive outlier hospitals in each country demonstrate that performing more neonatal blood cultures is possible in LMIC hospital settings. However, there is a major missed opportunity for neonatal infection detection in two-thirds of hospitals where laboratories with blood culture services are available, but culture use is < 50%. Hospitals with laboratories have an opportunity to implement quality improvement strategies to increase blood culture usage and improve newborn outcomes.
Blood culture gap and hospital performance
Neonatal blood culture use was markedly low overall. Non-utilisation of culture in the diagnosis of neonatal sepsis can have significant implications on diagnostic accuracy, leading to delayed or missed treatment and potentially negative outcomes. Blood culture remains the diagnostic standard for neonatal sepsis, enabling identification of causative agents and selection of targeted antibiotics. Empiric treatment without blood culture can lead to unnecessary antibiotic exposure and increased AMR. Despite low culture use overall, there was some variation across hospitals, consistent with other multi-site studies [
19]. A lack of basic laboratory infrastructure did not account for low routine culture use in this study, as no hospitals were without laboratories (
Tier 1). Neonatal infection prevention and care bundles for LMICs contain very few detection elements, possibly exacerbating the detection gap in NEST360-associated neonatal units [
28].
Most hospitals with laboratories but no microbiology (
Tier 2) were in Malawi (19/22, 86%), despite the country’s efforts in recent years to improve institutional infrastructure and AMR policy [
29]. Almost all these hospitals were District hospitals (17/19, 89%), the level at which blood culture should ideally be available [
30]. The positive outlier hospital in Malawi has support from international research institutions to undertake onsite microbiological surveillance [
29]. This Tertiary/National hospital accounted for 91% (2,580/2,847) of all cultures across the 37 Malawian hospitals, indicating that national AMR surveillance is limited to a single sentinel site. It also highlights the gap between Malawi's Tertiary/National and District hospital laboratory capacity. Health system factors that might exacerbate this gap include understaffing, and unmet training and accreditation needs in District laboratories, especially as countries expand neonatal care at the district level to meet ENAP target 4 and SDG 3.2 [
31].
Laboratories with microbiology were available for most neonatal units across the countries included in this study (43/65, 66%). However, almost all performed < 50% cultures for newborns on antibiotics (41/43, 95%), supporting evidence that laboratory services remain underutilised even when diagnostic testing is available (
Tier 3) [
32]. Consumable and human resource constraints likely impact the blood culture service in these hospitals. For example, blood culture bottles and culture media were only available in 30% and 55% of hospitals, respectively, which would limit blood culture use. Additionally, in Africa, the availability of quality-assured consumables is often compromised by lack of onsite production and ineffective supply chains, which are incompatible with product shelf life and cold storage requirements [
33‐
35]. Lack of culture bottles might also reflect insufficient resources available to purchase blood culture bottles, as well as various other microbiology laboratory supplies. There is also limited commercial interest in developing new diagnostics adapted for low-resource settings because of low profit margins [
35]. Although, some manufacturers have launched research initiatives for low-cost innovations targeting simplified blood culture systems [
36]. Insufficient neonatal unit staff might impede culture use, particularly at night when nurse-to-newborn ratios in this study decreased. Shortages and inequitable distribution of competent, motivated health workers result in heavy workloads for existing staff, unacceptable staffing ratios, and gaps in effective coverage of newborn care, including infection detection and management [
37]. Human resource strategies to improve newborn care in health facilities in LMIC require prioritisation [
38].
Who pays for laboratory tests likely contributes to the infection detection gap. Most health expenditures in LMIC rely on out-of-pocket payments by families. Studies in Nigeria show that costs for paediatric laboratory services are borne by families and are often barriers to access. [
39]. Healthcare users tend to underinvest in diagnostic tests, perceiving more immediate value in treatment versus diagnosis [
40]. These factors could contribute to fewer requests for blood cultures.
The Unitaid Fever Diagnostic Landscape points out that short reagent shelf life, supply chain difficulties, and need for highly skilled labour drive up the cost of blood cultures in LMIC, resulting in few blood cultures performed [
41]. The low numbers further contribute to higher prices per test, creating a vicious cycle [
42]. Governments aiming for universal healthcare coverage are unlikely to be able to cover these costs with their current designated health expenditure. For example, in Kenya, the cost of processing a single blood culture specimen (US$5.07 in 2017) constitutes a substantial proportion of per capita public expenditure on health from domestic sources (US$27.96 in 2017) [
29,
43,
44].
Clinicians may deprioritise diagnostic testing for newborns with suspected infection, particularly if unreliability and prolonged turnaround times limit the utility of blood cultures in clinical decision-making. Studies have shown that clinicians in Kenyan referral hospitals often perceive diagnostic tests as unreliable or unhelpful, leading to the underutilisation of laboratory services [
45,
46]. Blood culture utility may be hampered by limited communication between the laboratory and neonatal unit [
35,
47,
48]. To address this issue, there is a need to increase demand for culture testing among clinicians and to evaluate the laboratory-ward interface, which is poorly studied in low-resource settings [
30]. The interface could be strengthened by educating clinicians on laboratory principles [
49‐
51] and improving laboratory management information systems, which facilitate communication of test results and improve AMR surveillance [
52]. Our study showed that availability of protocols for ‘reporting results back to the neonatal unit’ and ‘adding laboratory results to patient records’ was low (36% and 49%, respectively). Effective and timely communication of test results is critical to ensure diagnostic culture is utilised to inform patient care.
Antibiotic prescribing
Antibiotic prescribing was high in each hospital, with some neonatal units prescribing antibiotics to all admissions. Elevated antibiotic use is well-documented in neonatal inpatient settings, particularly in LMIC [
19,
53,
54]. Such findings are unsurprising, as hospitalised newborns are especially vulnerable to infection and can rapidly deteriorate if infected, so clinical infection algorithms prioritise sensitivity over specificity [
55]. The result is a tendency for clinicians to prescribe antibiotics as a safeguard against severe infection, implicitly calculating that the benefits outweigh the risks for the baby. In high-resourced settings, these patients would have a blood culture performed to inform antibiotic de-escalation or cessation. With limited availability or use of cultures, such decisions rely on clinical acumen alone, which can result in prolonged antibiotic therapy and hospitalisation, driving higher healthcare costs [
56]. Excess antibiotic exposure can also have consequences for the newborn (including increased risk of necrotising enterocolitis [
57], medication-related adverse events, infant-mother separation and an altered microbiome [
58,
59]) and society (through the propagation of AMR) [
60]. Improved antibiotic stewardship is imperative in every country and for every neonate but is improbable without increased diagnostic use in hospitals [
61].
Strengths and limitations
Our study has strengths, notably the large number of neonatal units and newborn records, standardised data collection tools across sites, and trained data collectors. Generalisability in Malawi is good since implementation with NEST360 is occuring nationally (i.e., in all public district, regional, and mission hospitals). Fewer hospitals are assiocated with NEST360 in Kenya, Tanzania, and especially a smaller proportion in Nigeria. Nonetheless, the results still provide insight into the extent of detection failure for newborn infections in the sub-Saharan African region and advocate strongly for improvement. The NID used for this study was systematically designed as a parsimonious tool focusing on critical indicators for impact. The tool supports frequent data quality assessments, ensuring consistent data use across sites to drive action. The opportunity for linkage with HFA data provided hospital-level service readiness context to individual-level data.
There were several limitations. First, these data are primarily from 2020–2021, and some factors may have changed after NEST360 implementation, such as infrastructure or staffing [
62]. Second, certain data were not available, including the timing of blood sampling versus antimicrobial administration, antibiotic dose and duration (limiting quality and appropriateness assessment of antibiotic prescriptions), antibiogram availability, and laboratory access to backup power. NEST360 plans to analyse the association between antibiotic choice, bacterial pathogen identified by blood culture, and outcomes such as neonatal mortality and extended hospitalisation in future studies. Finally, results were only documented for half of the blood cultures performed across all hospitals. This lack of documentation may reflect poor case note documentation practices, a widely recognised issue even in well-resourced settings [
63], or quality gaps in laboratory information management systems. Shifting toward electronic systems could improve result documentation, but managing extensive electronic databases is challenging [
64,
65], and more complex systems are required to collect individually-linked microbiology data. Implementation research is underway to assess if case-based antibiotic resistance surveillance is feasible in LMIC [
66].
The neonatal infection detection gap is actionable. Strategies to embed blood culture as routine standard of care for neonatal infection management require prioritisation by country governments, policymakers, clinicians, and laboratory staff. Short-term, efforts to improve culture use could focus on hospitals where laboratory and microbiology services are available but underutilised (
Tier 3). This could be achieved through iterative performance feedback, training, and strong leadership [
47,
53]. Longer-term, high-quality microbiology data needs to be collected, analysed, and disseminated to improve culture uptake, antimicrobial stewardship policies, and newborn outcomes. Availability of antibiograms at the hospital, regional, and country levels are required to appropriately target therapy. Using clinical audit data to stimulate quality improvement across hospitals with unexplained variation is a topical issue in high-income countries [
67]. Hospitals implementing with NEST360 could be provided with NID-linked quality dashboards to aid quality improvement efforts [
68]. Development of low-cost, point-of-care sepsis diagnostics could be transformative, but blood culture remains the current best practice [
69,
70]. Additionally, even when available, there is substantial variability in use of POCTs for management of childhood infections across countries [
71]. Newborns with infections, but who do not access hospital care, must not be forgotten (Fig.
1). Intersectoral efforts are necessary to address access and cost barriers to quality healthcare.
Future research could focus on why hospitals with similar laboratory and ward capacities vary in their routine use of laboratory diagnostics for newborns. A qualitative exploration of local barriers and enablers to blood culture is underway in hospitals implementing with NEST360. The clinical pathway encompassing the decision to perform blood culture, sample collection, specimen processing, and result communication involves various activities and stakeholders. Implementation research is required to identify and overcome local rate-limiting steps in these processes, especially for LMIC.