Reducing MRSA Infection in a New NICU During the COVID-19 Pandemic

Yale New Haven Children’s Hospital is a nonfreestanding children’s hospital in New Haven, Connecticut, with a 68-bed level IV NICU supporting a high-risk delivery service and is a regional referral center for infants with complex conditions. Approximately 1000 infants are admitted annually, with 15% transferred from outside hospitals. In January 2018, the NICU moved from a 54-bed, open-bay layout occupying 1 area of a single floor to 32 single and 18 double rooms, spanning the entirety of 2 floors. The MRSA prevention program, in place since March 2003, encompassed screening all infants on admission and weekly, then cohorting and isolating those colonized with contact precautions.²⁸  Contact precautions were discontinued after 3 months with 3 consecutive negative weekly cultures.

In the Fall of 2019, active surveillance in the NICU detected an increase in MRSA colonization (10 neonates in July–September 2019 compared with 4 colonized neonates the previous 3 months) and invasive MRSA infections (3 neonates from February–September 2019 compared with only 1 in all of 2018). There were no changes in surveillance methods or case definitions of colonization or active infection during this period.

In October 2019, a multidisciplinary group convened consisting of registered nurses (RNs), nursing leadership, nurse educators, neonatologists, advanced practice providers, infection preventionists, a pediatric infectious diseases pharmacist, and pediatric infectious diseases physicians. Other groups consulted throughout the process included the NICU family advisory council, respiratory therapists, radiology technicians, environmental service (EVS) leadership, and front-line workers. Meetings occurred every 2 to 4 weeks to review outcome and process measures, assess plan–do–study–act (PDSA) cycles (summarized in Table 1), and direct future interventions.

We developed a strategy to reduce MRSA infection and colonization by targeting 3 key drivers: (1) optimization of environmental infection prevention, (2) identification and management of high-risk patients, and (3) prevention of person-to-person transmission (Fig 1). We then studied baseline infection prevention practices and identified differences between the NICU layouts that may have affected MRSA colonization and/or infection rates. These included: (1) increased reliance on cell phone communication for patient care, (2) increased environmental high-touch points in the new unit, (3) limited sight lines in the new NICU to observe hand hygiene and practice 200% accountability, and (4) lack of standard procedures for environmental disinfection. These efforts identified opportunities for improvement displayed on our key driver diagram (Fig 1).

After forming our team, exploring our problem, and planning our strategy, we focused on transitioning to patient-specific stethoscopes. At baseline, stethoscopes were provider-specific, with nonstandardized disinfection methods, and were considered potential vectors for transmission.^29–31  The transition to patient-specific stethoscopes required several PDSA cycles. The first, in November 2019, assessed the usability of disposable stethoscopes; however, these were abandoned for poor sound quality. Next, various reusable stethoscopes were trialed, and a favorite was selected. We then aligned protocols for stethoscope disinfection with manufacturer’s instructions using 70% isopropyl alcohol wipes and developed processes for storage and terminal cleaning. Stethoscopes were inoculated with SA in the laboratory and demonstrated no growth after applying our disinfection protocol. NICU staff were then educated on the various new processes.

Our next endeavor was to optimize environmental infection prevention. Baseline fluorescent marking of 3 rooms combined with staff interviews determined that the groups responsible for environmental cleaning, bedside RNs and EVS staff, lacked delineated responsibilities. A series of PDSA cycles were performed to formalize a process for cleaning high touchpoint surfaces. From December 2019 to January 2020, a survey of RNs and meetings with EVS leadership led to creation of role-specific environmental cleaning checklists. Partnering with EVS allowed us to: (1) provide education about MRSA in the NICU, (2) share baseline fluorescent marking data to support the need for improvement, (3) seek checklist feedback, and (4) identify barriers experienced by EVS staff. For example, high activity by NICU staff in work areas during scheduled cleaning times hindered EVS staff. In response, cleaning times were adjusted. The final environmental cleaning checklists were implemented in January 2020. Compliance was tracked monthly via fluorescent markings of ∼10 rooms per month, with direct performance feedback provided to nursing and EVS leadership. This included focusing on problematic touchpoints, such as the monitor pause button and broad education to reinforce practices.

An unexpected barrier to implementation occurred in April 2020 when the COVID-19 pandemic precipitated severe supply chain disruption of PDI wipes in our unit. This necessitated substitution with products of differing compositions (eg, bleach) over the next 13 months. Disinfectant wipe containers were removed from individual patient rooms to prevent theft and ration inventory, forcing staff to first locate and then carry wipes over extended distances. Some dried out during transport, so sealed plastic bags were deployed. The variability in product type, location, and means of transport resulted in highly variable cleaning practices.

Our next focus was to modify unit-based MRSA screening. Before February 2020, all infants were screened weekly via a nares swab, and colonized infants were placed under contact precautions. Decolonization was not performed. In February 2020, a series of PDSA cycles were performed to optimize MRSA screening, and targeted mupirocin decolonization was implemented. Swabs of the nares, umbilicus, and groin were performed to increase culture sensitivity. We next reviewed risk factors for invasive infection from colonized infants to determine eligibility criteria for decolonization (Supplement 2). In the final implementation phase, an infectious diseases pharmacist performed cross-checks to ensure infants met criteria. Inappropriate mupirocin utilization was assessed as a balancing measure to ensure only those meeting criteria were decolonized.

We reviewed all cases of MRSA colonization to identify common risk factors, but abandoned this when no modifiable factors were identified. Although staff were more aware of personal protective equipment requirements because of the pandemic, a shortage of gowns resulted in the need to reuse. Despite the shortage, we did not observe an increase in MRSA colonization rates during this time.

We next addressed hand hygiene. Baseline NICU-specific hand hygiene compliance was assessed at 79.6%. In November 2019, a PDSA cycle focused on staff education reemphasized optimal hand hygiene practices. Educational hand hygiene signs were also displayed throughout the unit. Next, QI team members performed covert monthly audits and provided just-in-time feedback to noncompliant staff. Compliance data were presented, stratified by role, and shared with applicable leadership as a mechanism for comparison, feedback, and intervention. The COVID-19 pandemic led to supply shortages of Purell, with subsequent adoption of less conveniently located products from various manufacturers in small containers rather than wall-mounted. Additionally, the viscosity varied such that a uniform volume of sanitizer was not consistently dispensed.

Our final intervention focused on staff infection prevention beyond hand hygiene. Patient care in private rooms spread out over a large geographic footprint required increased mobile device communication. We hypothesized that contaminated phones were fomites for MRSA transmission. In January 2020, after feedback from NICU and infection prevention leadership, a new disinfection policy (Supplement 3) was implemented that included:

1. cell phone disinfection with PDI wipes at the beginning and end of each shift and after each use in a patient room;

2. disinfection of identification badges;

3. nothing worn below the elbow except a single smooth ring; and

4. no artificial nails or gel manicures.

Implementation was initiated with multiple staff meetings where NICU leadership explained the rationale for the new policy and answered questions. To further improve compliance, QI team members performed covert, direct observation audits each month and delivered real-time feedback. Again, COVID-19 resulted in supply chain shortages of PDI wipes. Additionally, the lack of access in patient rooms presented a challenge for compliance with mobile device and badge disinfection after in-room use.

The primary outcome measure of invasive MRSA infection rate was calculated as MRSA infections per 1000 patient days and defined as any positive blood culture or deep wound culture with clinical signs of infection (eg, erythema, purulence). Superficial skin cultures obtained without documented signs of infection were excluded. The infection prevention team that reviews all hospital-related infections reviewed suspected cases to determine if they met criteria for invasive infection. Days between invasive MRSA infections was also tracked. The secondary outcome measure was new MRSA colonization rate, defined as a previously negative or untested infant with a new positive surveillance culture. For infants who repeatedly tested positive, only the initial culture was included. To enhance our understanding of this outcome, we compared median MRSA colonization days per patient before and after implementation of decolonization using the Kruskal-Wallis test. The definitions of invasive infection or new colonization were consistent during the baseline and intervention periods.

Multiple process measures were tracked. Environmental disinfection was assessed through compliance with the cleaning checklist as monitored by fluorescent marking. Areas were grouped by nursing and EVS responsibilities, with percentage compliance for each displayed on individual control charts. QI team members performed covert hand hygiene audits, mobile device disinfection, and “bare below the elbow” policy compliance. Because of initial low numbers of audits, compliance data for mobile device disinfection and bare below the elbow were combined as a composite percentage compliance with person-to-person transmission prevention policies.

All cases of mupirocin decolonization were reviewed for appropriateness, and inappropriate use was tracked as a balancing measure. Appropriateness was defined as meeting criteria per the decolonization guideline (Supplement 2) or another appropriate medical reason documented by the attending physician.

Statistical process control with control charts for outcome metrics used the application of published rules for evaluating special cause variation.³²  The “prebaseline period” of 15 monthly data points, from October 2016 to December 2017, came from the previous open-bay NICU. The “baseline period” of 20 monthly data points was from February 2018 (first full month in the new NICU) to September 2019 (the month before the start of this initiative). The intervention period spanned from October 2019 to September 2021. Sustainability of our effort was assessed during the maintenance period (October 2021 through July 2022). During the intervention and maintenance periods, means were adjusted when 8 or more consecutive data points were on the same side of the center line after an intervention. Interventions/PDSA cycles (Table 1) were superimposed on each outcome measure control chart to assess their impact. Invasive MRSA infections were also displayed on a control chart depicting days between events, with the mean calculated using all data points. Process measures were displayed as control charts with monthly percentage compliance. For the control chart depicting monthly hand hygiene compliance, the initial centerline was calculated using 15 data points from the period before the start of this initiative. The centerline was adjusted as previously described. For new processes, the centerline was calculated using all displayed data points.

This project met the Yale School of Medicine institutional review board criteria for exemption as a clinical QI project.

The rate of invasive MRSA infections in the open-bay NICU was 0.083 per 1000 patient days (Fig 2A) and increased to 0.131 per 1000 patient days after moving to private rooms. In February 2020, we achieved our goal mean rate of 0 infections per 1000 patient days and sustained this for 30 months with only 1 MRSA infection (Fig 2A) (Supplement 4). There were no statistically significant differences (via Mann-Whitney U test) in the median gestational age (37 weeks [22–44 weeks] versus 37 weeks [22–43 weeks]; P = .853) or birth weight (2870 g [390–6580 g] versus 2880 g [360–5640 g]; P = .709) of neonates admitted during the baseline period (n = 1563) compared with those admitted during the intervention period (n = 2501). We did note a longer median length of stay (7 days [1–366 days] versus 5 days [1–367 days]; P < .001) in the baseline period. This was adjusted for in the outcome measures by utilizing patient days as the denominator. Days between invasive MRSA infections in the new NICU showed a mean of 149.93 (Fig 2B).

The prebaseline rate of new MRSA colonization in the open-bay NICU was 0.498 per 1000 patient days (Fig 2C). The baseline rate in the new configuration increased to 1.177 per 1000 patient days (Fig 2C). A centerline shift in MRSA colonization was appreciated in November 2021, during the maintenance period, to 0.384 per 1000 patient days. Additionally, although not statistically significant, there was a reduction in median days of MRSA colonization per patient after introduction of targeted decolonization. Before this intervention, the median days of MRSA colonization was 21 per patient (range 1–112 days, n = 26) compared with 8 days per patient (range 1–141 days, n = 31 patients) after implementation of targeted decolonization (P = .258, Kruskal-Wallis).

Mean monthly staff hand hygiene compliance from April 2018 to September 2019 was 79.6%, and with our interventions, improved to 89.0% from January 2021 to April 2021, and improved further to 94.1% from May 2021 to July 2022 (Fig 3A). For new processes, mean monthly RN and EVS cleaning compliance was 70.1% and 64.0%, respectively (Fig 3B, 3C). Mean monthly compliance with person- to-person transmission prevention interventions was 86.0% (Fig 3D). Of note, the number of audits for process measures increased starting in June 2021 with additional QI-dedicated nursing time (Fig 3).

Fifteen MRSA-colonized infants received 25 courses of mupirocin, 17 of which were appropriately administered. An additional 7 were administered to MRSA-colonized infants that did not meet criteria but were deemed high risk by the medical team and therefore categorized as appropriate. One decolonization was deemed inappropriate (1 of 25, 4%). Discussion with the medical team revealed that they had awareness of the guideline but recalled incorrect criteria for decolonization and had not physically consulted the document.

Data regarding SA colonization and infection before and after moving from open-bay NICUs to individual rooms are variable, with some hospitals showing a reduction in SA colonization and infection rates and others an increase.^20–26  We noticed an increase in MRSA colonization and infection after moving from an open NICU to a unit with individual and double rooms. Several factors that likely contributed include:

1. increased environmental touchpoints given a larger geographic footprint, including wall-mounted touch pads to call for assistance and a heavy reliance on mobile communication devices;

2. lack of standardization and clear delineation of environmental cleaning in the new NICU;

3. decreased sight lines in the new environment, making observation of infection prevention practices, such as hand hygiene and equipment disinfection, more difficult; and

4. increased presence of family members at the bedside who could now room in.

Like other studies targeting SA infection,^17,18  our initiative effectively decreased the invasive MRSA infection rate from 0.131 to 0 infections per 1000 patient days. Days between invasive MRSA infections increased to a peak of 405 days between the most recent 2 infections. However, despite a reduction in invasive infection, new MRSA colonization rates were not significantly impacted by targeted interventions for the first 2 years. This outcome may be partially explained by increased sensitivity of the swabbing procedure during the intervention period because we moved beyond the nares to include the groin and umbilicus. Previous studies have shown inconsistent success rates in reducing MRSA colonization when utilizing decolonization efforts in the NICU. Although some found notable recolonization rates after decolonization (up to 50%), with little impact on invasive infection rates,^33,34  others have shown significant reductions in both colonization and infection rates.^17,35,36  MRSA colonization rates did decrease in November 2021 during the maintenance period and likely contributed to the continued absence of invasive infections. Additionally, the number of MRSA patient days decreased because of decolonization. Our emphasis on other infection prevention strategies, including staff hand hygiene and environmental disinfection, likely contributed to the sustained reduction in invasive infection rates. Other interventions targeting family members’ hand hygiene rates were not as successful, with <50% compliance observed (data not shown). Lack of improvement in these areas may also explain why MRSA colonization did not initially improve. It has been shown that visitor hand hygiene and decolonization of family members can reduce NICU SA colonization and infection.³⁷  Future PDSA cycles dedicated to parental/visitor infection, prevention-based interventions are top priorities. In tracking our balancing measure, rate of inappropriate mupirocin administration, we did not appreciate any untoward effects of our efforts. Additional balancing measures were considered, such as stethoscope attrition rate, but were difficult to measure. Ultimately, data collection resources were directed toward the tracking of process measures.

COVID-19 was an unexpected variable that affected our interventions (Table 2). We are unaware of literature demonstrating the impact of the pandemic on similar NICU-based infection prevention efforts. Several infection prevention policies implemented during the pandemic, like universal masking, reduced visitation, and heightened awareness of infection prevention practices by staff, may have contributed to reduced MRSA infections. A less restrictive prepandemic visitation policy, with families able to stay overnight in private rooms, may have contributed to the initial observed increase in MRSA colonization and infection. Infection prevention efforts were hampered by the pandemic because of extreme supply shortages of disinfectant wipes, gowns, and hand sanitizer. Despite these barriers, our primary outcome measure was not negatively impacted, possibly because of alternative products and novel strategies tested by PDSA cycles (ie, introducing plastic bags to keep wipes moist during transit).

Strengths of this project include a multidisciplinary QI team-based approach with robust assessment of process measures. Its success was driven by unit leadership’s sustained commitment to addressing the problem. It also offers a unique opportunity to examine how the COVID-19 pandemic directly altered infection prevention efforts. Weaknesses include the fact that it was a single-center intervention. Additionally, because of the many variables introduced by the pandemic, it is unclear which interventions were most responsible for the observed reduction in MRSA infection because we were unable to track or quantify these data objectively. Ideally, data on the numbers of supply shortages, visitors per day, and staff members properly donning personal protective equipment would be available for analysis. These measures were not tracked because of restrictions regarding personnel in the NICU during this time. Finally, the temporal proximity of different interventions makes it difficult to discern how much each contributed to our primary outcome.

Despite the challenges of the COVID-19 pandemic, implementation of a multidisciplinary QI initiative focused on environmental and personnel infection prevention strategies, coupled with increased efforts to identify and decolonize high-risk infants, led to 0 MRSA invasive infections per 000 patient days in our NICU.

COVID-19

coronavirus disease 2019

EVS

environmental service

MRSA

methicillin-resistant Staphylococcus aureus

PDSA

plan–do–study–act

QI

quality improvement

RN

registered nurse

SA

Staphylococcus aureus