Study design, setting and population
A prospective, quasi-experimental study was conducted in a 30-bed acute neonatal ward in Cape Town, South Africa to assess the impact of a multimodal intervention on surface and equipment cleaning. Tygerberg Hospital is a 1384-bed public teaching hospital, including maternity and neonatal services with 8000 high-risk deliveries (37% low birth weight) and approximately 2500 neonatal admissions annually. Despite being an upper middle-income country, South Africa has a high Gini coefficient indicating inequality; most patients utilizing this public hospital are indigent and the public hospital resources are more typical of a low-middle income country (LMIC).21 The 124-bed neonatal unit includes a 12-bed NICU, three high-dependency wards, and one kangaroo mother care ward, with mean occupancy rates exceeding 100%. The neonatal unit provides medical and surgical care for sick, preterm (<37 weeks’ gestation) and/or low-birthweight (<2500 g) neonates; prematurity, perinatal asphyxia and infection are the most common indications for admission. Given the extreme shortage of NICU beds, non-invasive ventilation (nCPAP and high-flow oxygen therapy) is used extensively in the high-dependency neonatal wards. The hospital’s on-site Unit for Infection Prevention and Control (UIPC) has one infection prevention nurse practitioner responsible for the maternity, paediatric and neonatal departments.
Cleaning policies and procedures prior to intervention
The UIPC presents quarterly training sessions to cleaning personnel and requires attendance of at least one training session per annum. Ad-hoc training sessions are given during outbreaks and to orientate newly-appointed cleaning staff. For routine cleaning of surfaces and equipment, water, household detergent and disposable cloths are used, and sometimes re-used. Spot cleaning is done immediately for spillage or gross contamination. Routine practices for the neonatal ward cleaning staff include daily wiping down of horizontal surfaces (work surfaces, tables, chairs and floors) and sinks plus twice daily emptying of waste bins. The cleaning personnel are responsible for cleaning of the milk kitchen, medication room, ward kitchen, toilet areas, mothers’ rooms and passages. The housekeeping supervisor uses a daily cleaning checklist to ensure that all the areas are cleaned. A color-coding system is used with separate cloths for patient, toilet, isolation and kitchen areas. The nursing staff are responsible for cleaning items in the patient zone i.e. all items touched directly or indirectly by the neonate or touched by the staff/mother while delivering care. These items included incubators, cots, bedside cabinets and clinical equipment; there was no system of monitoring the cleaning adequacy for these items prior to this study. Terminal cleaning is performed by nursing and cleaning staff after discharge or transfer of patients who were isolated and/or under transmission-based precautions using a checklist, followed by disinfection with 70% alcohol. Prior to implementation of the neoCLEAN intervention, the routine ward cleaning process did not specifically focus on high-touch areas. Apart from the cleaners’ daily environmental cleaning checklist for shared non-clinical areas (e.g. toilets, kitchens), there was neither a process to prioritize high-touch surfaces/equipment nor record the frequency of cleaning in clinical areas.
Methods used to assess the adequacy of routine cleaning
The adequacy of routine cleaning was evaluated for 100 pre-specified items (58 surfaces and 42 equipment) using quantitative bacterial surface cultures, adenosine triphosphate (ATP) bioluminescence assays and fluorescent ultraviolet (UV) markers, performed at baseline (P1, October 2019), early intervention (P2, November 2019) and late intervention (P3, February 2020) phases. Surfaces and equipment (e.g. mattresses, cots, saturation monitors, infusion pumps) were swabbed using E-swabs in liquid transport medium (Sigma transwab); for flat surfaces a 10x10cm swabbing template was used and for complex surfaces/equipment, a custom standard collection procedure was developed for each item. The swab was applied to the surface using a sweeping action in close parallel streaks while rotating the swab using a constant pressure. ATP relative light unit (RLU) readings were collected from the 100 specified surfaces and equipment using 3M Clean-Trace swabs and a 3M ATP meter. Ultraviolet (UV) disclosing lotion (GlitterBug Potion, Brevis Corp, Salt Lake City, UT) was applied to flat surfaces using a cotton bud in a circular motion in a 2cm diameter; 100 UV marks were placed on the 100 specified surfaces and equipment items. The investigator returned after 24 hours with a UV torch to count the proportion of markers that remained after 2 staff shift changes (i.e. after 2 opportunities for removal by routine cleaning).
Metrics used to assess adequacy of cleaning
To assess the impact of the multi-modal intervention on cleaning adequacy we measured: (1) the proportion of the 100 pre-specified surfaces and equipment swabs with no bacterial growth and the proportion with growth of potential neonatal pathogens; (2) the proportion of ATP swab readings below an accepted threshold for cleanliness (< 200 relative light units [RLU])22 and (3) the proportion of UV marks removed following cleaning.
Laboratory methods for the quantitative bacterial cultures
Surface and equipment swab tubes were vortexed for 30 seconds; 25 uL of the fluid was plated on to blood- and MacConkey agar plates and incubated at 37oC for 48 hours. Manual aerobic colony counts (ACC) were recorded for each plate and each unique colony was Gram stained. Gram-positive cocci were identified as S. aureus, coagulase negative staphylococci, enterococci, or streptococci through catalase testing, pyrrolidonyl aminopeptidase activity, and/or latex agglutination (Pastorex Staph-Plus; Bio-Rad, Redmond,WA). Gram-negative isolates were identified using the automated Vitek-2 system (BioMerieux, Marcy-l’Étoile, France). Agar plates with non-pathogens (as defined by the Centers for Disease Control and Prevention)23 were classified as “environmental flora”; agar plates with neonatal pathogens e.g. S. aureus and Gram-negative bacilli, were classified as “potential pathogens.”
NeoCLEAN multimodal intervention
The focus of the NeoCLEAN intervention was to improve cleaning of frequently-touched surfaces in the patient zone and areas associated with direct clinical care (medication trolley, supply and emergency trolleys, milk kitchen and hand washbasins). Other frequently-touched surfaces (e.g. telephones, computer keyboards) were included only if present in the clinical areas; toilet areas, sluice rooms, administrative offices and door handles to non-clinical areas were not included. The enhanced cleaning strategy focused mainly on nursing staff and mothers and was implemented in addition to the routine surface cleaning performed by the environmental cleaning staff. The key elements of the multi-modal NeoCLEAN intervention were training of nursing staff regarding surface/equipment cleaning, introduction of customized cleaning checklists, cleaning audits with staff feedback, use of in-room disinfectant wipes and involvement of neonates’ mothers with cleaning of the patient zone.
Training of staff and mothers to clean the patient zone
Baseline training of nursing staff was conducted for approximately 1 hour with 30 neonatal ward nursing staff (in 4 sessions to include day and night shifts). The training reinforced why enhanced cleaning is needed for neonatal wards to maximize “buy-in” and cooperation from staff with the intervention. New cleaning standards were set and cleaning procedures were clarified by discussion e.g. what needed to be cleaned, how, when, how often, and by whom. Clearly delineating the roles and cleaning responsibilities of cleaning staff, nursing staff and mothers was a key task prior to implementing the intervention. Each nurse working in the clinical rooms took responsibility for teaching the mothers of newly admitted neonates to perform daily cleaning of their babies’ cots, incubators and bedside cabinets.
Development of cleaning checklists
Clinical observation was conducted in each area to determine the most frequently touched surfaces. For each area targeted (6 patient cohort rooms, a medication preparation area and the milk kitchen), a specific checklist was designed based on the surfaces/equipment present. The checklist included color photographs of the items to be cleaned with space for day and night shift staff to sign after completing cleaning (Figure 1). The checklists were prominently displayed in each room, acting as a visual reminder to prompt cleaning. A new blank checklist was placed in each room every week. Frequent reminders had to be given initially to ensure staff compliance with completing the checklist.
Introduction of disinfectant wipes
Commercially available pre-packaged hospital disinfectant wipes (Clinell TM Universal wipes) were placed in an easily accessible position in each clinical room, the medication preparation area and the milk kitchen. The Clinell wipes (active ingredients benzalkonium chloride, didecyldimonium chloride and phenoxyethanol) utilize a one-step cleaning and disinfection process which reduces the cleaning time. When Clinell wipes were out of stock, staff were trained to use alternative cleaning agents e.g. cleaning with liquid detergent and water and drying, followed by disinfection with 70% alcohol spray.
Implementation of cleaning audits with staff feedback
Three formal assessments of neonatal ward cleaning adequacy were performed at baseline (P1, October 2019), early intervention (P2, November 2019) and late intervention (P3, February 2020). Following each phase, verbal feedback and encouragement was provided to the neonatal nurses and the assessment results were prominently displayed on the ward’s infection prevention noticeboard. Intermittent, informal cleaning audits with feedback were conducted over the 5-month study period. These informal audits used UV marks to provide visual feedback to staff on frequently missed surfaces when cleaning. The audits were also an opportunity to hear how the nursing staff perceived the cleaning intervention, identify obstacles to cleaning and record suggestions for programme improvement.
Data collection, statistical analysis and study approvals
Data for each assessment phase were collected on a report form for the 100 surfaces/equipment items for bacterial cultures, UV gel markers and ATP swabs then entered into a REDCap database. Proportions were calculated for each of the test metrics at each study phase. Continuous and categorical variables were compared using the Kruskal-Wallis test and the X2 test, respectively. A p-value of <0.05 was considered statistically significant. Stata Statistical Software version 13.0 IC (College Station, TX: StataCorp LP) was used for analysis. At the informal ward cleaning audits, the investigators collected and manually recorded verbatim feedback from nursing staff and mothers. The Stellenbosch University Health Research Ethics Committee and the Tygerberg Hospital management approved the study protocol (N18/07/068).