Secondary Logo

Journal Logo

Brief Reports

Eighteen Months of “Matching Michigan” at a UK Neonatal Intensive Care Unit

Montgomery-Taylor, Sarah BA*; Emery, Faith MB BS; Anthony, Mark PhD

Author Information
The Pediatric Infectious Disease Journal: May 2013 - Volume 32 - Issue 5 - p 565-567
doi: 10.1097/INF.0b013e3182868389
  • Free


Matching Michigan (MM), initiated by the UK National Patient Safety Agency, is a care bundle focusing on reducing healthcare-associated infection. It is modeled on a program introduced to reduce central venous catheter–related blood stream infections (CR-BSIs) in adult intensive care units (ICUs) in Michigan.1 MM has since been implemented in 97% of adult and pediatric ICUs across the NHS where it has proven effective.2 The MM program was initiated in neonatal ICUs in January 2011 where infection reduction is of paramount importance. Neonatal infection increases mortality, lengthens hospital stays and worsens neurodevelopment outcome.3,4

We report here our first 18 months experience since January 2011 of MM in our neonatal unit (NNU) at the John Radcliffe Hospital, Oxford, United Kingdom. The procedures recommended from the adult Michigan care bundle included: training in hand washing, using full-barrier precautions during the insertion of central venous catheters, cleaning the skin with chlorhexidine, avoiding the femoral site if possible and removing unnecessary catheters.1 These and other strategies were incorporated into our MM program which had 3 branches: training; root cause analyses (RCAs) and feedback.



We implemented an induction training program consisting of:

  • Aseptic nontouch technique hub care to reduce late CR-BSIs—those occurring after 10 days of catheter duration;5
  • Improved hand hygiene to reduce spread of pathogens;
  • Central venous catheter insertion to reduce early CR-BSIs—those occurring before 7 days of catheter duration;
  • Aseptic nontouch technique blood culture (BC) technique to reduce BC contaminants;
  • Audit to reinforce the implementation of these techniques.6

Root Cause Analysis

For every positive BC, we undertake a bedside RCA with key members of the team: lead neonatal consultant for MM, pediatric infectious diseases doctor, infection control nurse and the attending neonatal consultant. This team classifies the BC growth as one of the following: (1) contaminant (ie, from poor BC technique), (2) a true infection in the absence of a central line, (3) a catheter-associated blood stream infection (CA-BSI; meaning the catheter is coincidentally in situ, and there is another source of the bacteremia, usually intestinal pathology) or (4) a CR-BSI (where the catheter itself is implicated).

A positive BC is designated as a contaminant where:

  • Coagulase-negative staphylococcus (CoNS) was cultured on the first day of life (CoNS do not cause early-onset sepsis);
  • A skin commensal was cultured from only 1 BC (or >2 different commensals in 1 BC) in late-onset sepsis;
  • No increase was measured in C-reactive protein on >2 occasions;
  • No change was measured in full blood count parameters;
  • The baby recovered before antibiotics took effect, antibiotics were stopped at 36 hours and infection did not recur.

For real infections, if a baby had a central line in situ at the time of BC positive, but there was another obvious source of infection (eg, new onset necrotizing enterocolitis, peritonitis or urinary tract infection), and the line tip culture showed no growth, the culture is classified as a CA-BSI. Conversely, if a baby had a normal abdomen, no urinary tract signs or other source of infection, for example, skin, bone or meningitis, and the central line tip grew the same organism as the BC, the culture is classified as CR-BSI.


We feedback action points from RCAs by e-mail to all doctors concerned and through weekly infection and monthly clinical governance meetings. Furthermore, we ensure action points are included in development of clinical guidelines. Some of the changes that we have made in the first 18 months of MM are:

  • Removing umbilical lines by days 5–7 of life;
  • Giving Candida prophylaxis to extremely premature babies;
  • Starting breast milk feeds on day 1;
  • Removing central lines when on 120 mL/kg/d instead of 2–3 days after reaching 150 mL/kg/d;
  • Stopping antibiotics at 36 hours in noninfected babies;
  • Reducing broad-spectrum antibiotic use;
  • Using 2% chlorhexidine, 70% alcohol for skin antisepsis (except periumbilical skin on day 1 of life in extremely premature babies).


For the first time in our NNU, we now have a detailed analysis of our BC positives. As a tertiary surgical and medical NNU, in 2011 we had 877 admissions, of whom 151 were <1500 g, with a total of 2955 catheter days for all babies. We have now been monitoring results for 18 months (January 2011–June 2012), and in this period there were 99 BC positives, of which 54 were deemed to be true BSIs. Of these, 14 were present at birth (early-onset sepsis) and 40 were late-onset sepsis. Of the late-onset sepsis, 9 occurred in babies without a central line, 11 were classified as CA-BSI and 20 as CR-BSI. Before January 2011, we did not have a detailed breakdown of the number of lines in use or the designation of BSI. We are now able to calculate infections per 1000 line days for the last 18 months: the total infection rate is 9.0, CA-BSI or CR-BSI infection rate is 7.0 and CR-BSI infection rate is 4.5. This is in keeping with the published literature from adult ICUs.


An interesting aspect of our analysis is that almost all CR-BSIs are caused by CoNS. Most contaminants are also CoNS, and the most significant impact of MM our NNU has seen is a reduction in BC contaminants (Fig. 1A). Our procedure was for the lead MM to write an e-mail to the doctor who took the BC whenever the RCA deems that there is a contaminant, requesting to rediscuss BC technique. This simple measure by a consultant neonatologist, directed at the junior staff, has led to a dramatic reduction in CoNS-contaminated BCs. The reduction in all CoNS infections, including contaminants, comparing before January 2011, to after, is statistically significant (P = 0.006, 2-tailed Mann–Whitney U test).

A, Bar chart showing CoNS BC positives. Before January 2011, the contaminants are included in the total but since this date the data shows only the real infections. Even when the contaminants are included in the total, after January 2011, there is a significant reduction in BC positives with CoNS (P = 0.006, 2-tailed Mann–Whitney U test). B, Bar chart showing non-CoNS BSIs, mainly Gram-negative bacilli, but also Staphylococcus aureus, enterococci, group B streptococcus and Candida albicans. BSI numbers have reminded static, but at a time of a near doubling in NNU activity.

Our detailed analysis also shows that the non-CoNS infections, Gram-negative bacilli, Staphylococcus aureus, enterococci and Candida albicans, are virtually all unrelated to the presence of a central catheter. This suggests that reduction of the non-CoNS infections in newborn babies in neonatal ICU may be less amenable to MM than BSIs in adults in ICUs. Furthermore, at first glance, our figures indicate that the MM program has not had a major impact on the true non-CoNS infections (Fig. 1B). The high prevalence of non-CoNS BSIs in babies reflects the pathological process, different in adults, in which preterm babies (and some surgical babies) are particularly predisposed to seeding of bacteria from the intestine, such as in necrotizing enterocolitis. This makes it unlikely that a program aimed at reducing catheter-related sepsis could have a major impact on neonatal BSIs. A more in-depth analysis of our figures, however, suggests that the non-CoNS infections have also been reduced. During the MM period, NNU bed days per month have almost doubled, yet the non-CoNS infections have not increased in line with the enhanced activity. Thus, it appears that MM can reduce non-CR-BSIs, as well as CR-BSIs.

Prevention of non-CoNS BSIs is also likely to be facilitated by using conventional infection control measures to prevent transmission of Gram-negative bacilli and multiresistant organisms between babies. Rational use of broad-spectrum antibiotics to limit selection of resistant organisms and early breast milk feeding to encourage normal bowel flora are further beneficial strategies.


1. Pronovost P, Needham D, Berenholtz S, et al. An intervention to decrease catheter-related bloodstream infections in the ICU. N Engl J Med. 2006;355:2725–2732
2. McKee C, Berkowitz I, Cosgrove SE, et al. Reduction of catheter-associated bloodstream infections in pediatric patients: experimentation and reality. Pediatr Crit Care Med. 2008;9:40–46
3. Adams-Chapman I, Stoll BJ. Neonatal infection and long-term neurodevelopmental outcome in the preterm infant. Curr Opin Infect Dis. 2006;19:290–297
4. Schlapbach LJ, Aebischer M, Adams M, et al. Impact of sepsis on neurodevelopmental outcome in a swiss national cohort of extremely premature infants. Pediatrics. 2011;128:e348–e357
5. Rowley S, Clare S, Macqueen S, et al. ANTT v2: an updated practice framework for aseptic technique. Br J Nurs. 2010;19:S5–S11
6. Harron K, Ramachandra G, Mok Q, et al. Consistency between guidelines and reported practice for reducing the risk of catheter related infection in British paediatric intensive care units. Intensive Care Med. 2011;37:1641–1647

care bundle; neonatal; infection; central venous catheter

© 2013 Lippincott Williams & Wilkins, Inc.