By Anais Cottage-stone · 14 September 2026
NeoNET AFRICA closed its Neonatal Outbreaks in Africa series on Thursday 24th July 2025, working with the African Neonatal Association and KlebNET-GSP. Dr Felicity Fitzgerald (NeoNET AFRICA / Imperial College London) opened the session, and the focus this time was different from the first two seminars: not just finding outbreaks, but figuring out which everyday practices in a neonatal unit actually prevent infection, and whether they're realistic given the staff and space most units have.
Watch the full seminar recording here.
Professor Andrew Whitelaw (Stellenbosch University and NHLS Tygerberg Hospital) opened with five years of data on carbapenem-resistant Klebsiella pneumoniae at Tygerberg Hospital, a 1,384-bed tertiary hospital where the paediatric and neonatal service alone runs about 300 beds (132 of them neonatal) at 93% occupancy and roughly 4,000 neonatal admissions a year.
Between 2016 and 2020, the lab collected 413 resistant K. pneumoniae isolates. Across the hospital as a whole, 45% were clinical and 55% came from carriage screening. But look at neonates specifically and that flips hard: of 120 neonatal isolates, 99 (82.5%) turned up through screening, and only 21 came from a sick baby. In other words, most of what was circulating in the unit was invisible if you were only watching for symptoms, which is exactly the point Andrew's talk was named for. And the stakes are real: across the Western Cape, Klebsiella made up 72% of 2,242 resistant Enterobacterales episodes, and mortality was 27% for clinical cases versus 6% for carriage.
Sequencing added detail rather than clarity. 97% of isolates carried a carbapenemase gene, split almost evenly between blaOXA-181 and blaNDM-1, with two sequence types, ST39 and ST2621, doing most of the damage. They behaved differently: ST2621 stayed mostly in adult wards, while ST39 hit neonatal and paediatric wards hard, with cases traceable back to 2016. Resistance genes were also jumping between bacterial species, not just between patients, turning up in E. coli, Serratia, Enterobacter, Providencia and K. oxytoca as well as Klebsiella. An environmental sampling study is now underway across the neonatal ward and NICU, checking everything from incubator buttons to the milk kitchen fridge; so far it's picked up 134 swabs from 16 colonised neonates and their surroundings.
Andrew was upfront about the limits of the work. It's all retrospective, so the sequencing never actually changed how an outbreak was handled at the time. Nobody yet knows how much of the transmission comes from the environment, staff or shared equipment, and gaps in the isolate collection make it hard to fully trace the chains. His conclusion: the basics still matter most, hand hygiene, education, cleaning, and the weak response in adult wards was, in his words, a missed opportunity.
Professor Julia Bielicki (City St George's, University of London) pushed on where outbreak control even sits within everything else a unit does. Her argument: the line between "an outbreak" and ordinary healthcare-associated infection is mostly a matter of how closely you're looking. Once you sequence isolates, cases that looked unconnected often aren't, and small overlapping clusters of infection are probably just normal life in a neonatal unit rather than a crisis. Which means preventing everyday healthcare-associated infection is the real job.
The strongest evidence is for bundled interventions around devices, catheters and ventilators especially. But a few single interventions also cut mortality in preterm babies: probiotics (RR 0.75), topical emollients (RR 0.81), synbiotics (RR 0.53). For probiotics specifically, a meta-analysis of 25 trials and almost 11,000 babies puts the mortality reduction at RR 0.80, and sepsis at RR 0.78, both rated high-certainty evidence, though almost none of those trials were run in Africa.
Kangaroo care tells a similar story. A meta-analysis covering 25 trials and over 17,000 low-birthweight babies, mostly in low- and middle-income countries, found it lowered the odds of death, sepsis and invasive infection, which is why Julia argued it belongs in standard infection prevention, not as a nice extra. Part of the mechanism may be the microbiome: nearly 60% of an infant's gut bacteria can be traced back to the mother, and in one trial, skin-to-skin contact more than doubled how often MRSA and MRSE were cleared from a baby's nose (52.8% versus 22.4%).
Not every intervention panned out. The NeoCHG trial of chlorhexidine skin care, run at Tygerberg and in Dhaka, found no safety issues but also no real effect on colonisation, whatever concentration or schedule was used, and resistant bacteria colonised babies quickly regardless.
The most striking part of the talk was practical. A survey of African neonatal units found kangaroo care being delivered in crowded conditions in over half of units, and in properly uncrowded conditions in only a quarter. The barriers units named were mostly about family and culture (89%) and space (60%), while what they said they needed most was training (78%). It's a real tension: the same intervention that prevents infection can also crowd a ward if there's nowhere to do it properly. That's exactly what the NeoDeco trial, running across 24 units in Europe until April 2026, is trying to measure alongside infection outcomes.
Dr Lauren Hookham (Brighton and Sussex Medical School, and MRC/UVRI & LSHTM Uganda) closed the day with the intervention everyone assumes is simple. Hospital cleaning isn't like cleaning a house. It has to target bacteria you can't see, follow standardised methods, and use the right product for the right surface and the right patient. Get any of four things wrong, technique, contact time, product or training, and it doesn't work.
The barriers Lauren has seen aren't about attitude, they're structural: understaffed teams with high turnover, patchy supply chains (including outright counterfeit bleach), no way to check whether cleaning actually worked, and cleaning staff who are rarely given any infection control training at all. Photos from her fieldwork made the point better than any slide could: disinfectant decanted into old drink bottles, instructions taped up that didn't match the products actually being used, shared equipment left with nowhere clean to dry.
For units wanting to improve, she pointed to existing resources rather than reinventing anything: TEACH Clean, WASH & CLEAN, and the WHO's own training package. She was honest that every method for checking whether cleaning worked, from watching staff to fluorescent markers to actual lab cultures, measures something slightly different, and only culture tells you directly whether the bioburden is gone. CDC guidance gives a reasonable baseline: disinfect high-touch surfaces at least twice a day, only clean the outside of an occupied incubator, and audit at least 5% of beds (or 15 care areas) regularly.
Her closing message is worth carrying forward: cleaning is cheap and effective, but only when it's done properly, and cleaners deserve to be treated as part of the infection control team, not as an afterthought.
This was the last session in the Neonatal Outbreaks in Africa series, which started with outbreak epidemiology in May 2025 and moved through detection and investigation in June. NeoNET AFRICA picked things up again in November 2025 with KlebNET-GSP for a seminar on Klebsiella genomics.
