A hospital-wide study from Hong Kong found that targeted decolonisation was associated with a 23% lower adjusted hazard of MRSA appearing in a later clinical specimen. The result is encouraging, but the operational lesson is just as important: identifying an eligible patient is only the beginning. Hospitals also need reliable systems to start treatment, track completion and follow up when the patient has already left.
Lee ALH et al.,
At a glance
Researchers at the 1,900-bed Prince of Wales Hospital evaluated a hospital-wide programme offering decolonisation to adults with meticillin-resistant Staphylococcus aureus (MRSA).
MRSA is a strain of Staphylococcus aureus that is resistant to meticillin and related antibiotics. It can live on the skin or in the nose without causing symptoms—known as colonisation—but carriers remain at greater risk of later infection and can contribute to transmission.
Compared with MRSA-positive patients from the preceding year, those who received decolonisation had:
A 23% lower adjusted hazard of MRSA being identified in a later clinical specimen.
A 48% lower adjusted hazard of subsequent MRSA bacteraemia, although this narrowly missed statistical significance and cannot be treated as a confirmed effect.
A programme supported by daily case identification, earlier laboratory reporting and follow-up after discharge; 89% of patients who had already left before their result was available were subsequently reached and treated.
What the hospital tested
From 27 August 2024, adult patients with MRSA identified through screening or a clinical specimen were offered:
2% mupirocin nasal ointment three times daily for five days.
A daily 4% chlorhexidine shower for five days.
Chlorhexidine washcloths when patients could not shower independently.
The researchers compared 932 patients who received decolonisation with 996 patients from the preceding control period. Patients were followed for two outcomes: the time to MRSA next appearing in any clinical specimen and the time to the next MRSA-positive blood culture.
After adjustment, decolonisation was associated with a lower hazard of subsequent MRSA in a clinical specimen (adjusted hazard ratio 0.77; P=0.015). The adjusted hazard ratio for later MRSA bacteraemia was 0.52, but the result did not reach statistical significance (P=0.054).
How this fits with the wider evidence
The findings sit within a larger evidence base showing that decolonisation can work, but that its effect depends heavily on the patient group, setting and delivery model.
In 2013, the landmark 43-hospital REDUCE MRSA trial → found that universal decolonisation in adult intensive-care units was more effective than either screening and isolation or targeted decolonisation. It reduced MRSA-positive clinical cultures by 37% and bloodstream infections from any pathogen by 44% from baseline.
The 2019 CLEAR trial → showed that the prevention opportunity continues after patients leave hospital. Among 2,121 MRSA carriers, repeated post-discharge decolonisation reduced the risk of MRSA infection by 30% compared with education alone. Participants who fully followed the regimen had 44% fewer MRSA infections in an as-treated analysis.
However, benefit should not be assumed across every hospital population. The 53-hospital ABATE Infection trial → found no significant overall benefit from universal chlorhexidine bathing plus targeted mupirocin across general medical and surgical units. A post-hoc analysis suggested larger reductions among the smaller, higher-risk group of patients with medical devices.
Taken together, these studies support a more precise conclusion than “decolonisation works”. The strongest approach depends on who is being treated, when they are at risk and whether the hospital can deliver the intervention reliably.
Delivery was as important as the protocol
The Hong Kong programme required more than publishing a policy. The Infection Control Team generated a daily list of newly identified MRSA cases and contacted wards to start treatment. Ward staff coordinated with a central bathing team, while treatment records were collected to track delivery.
The laboratory also began issuing a preliminary positive result one day earlier. Among the first 150 MRSA isolates reported this way, no discrepancies were found between the preliminary and final results.
The clearest workflow gap appeared at discharge. Among 1,073 MRSA episodes that initially missed inpatient decolonisation, 840 patients—78%—had already left before the result was available. A tracing process subsequently enabled 750 of those 840 discharged patients, or 89%, to receive decolonisation in the community or another healthcare institution.
The practical lesson is straightforward: an effective treatment creates little value if the result arrives after the patient has left or if no one can see that the required action remains incomplete.
From a result to a completed action
To run a targeted programme reliably, teams need a live view of:
Newly identified MRSA carriers.
Whether each patient remains in hospital or has been discharged.
Whether decolonisation has been prescribed and started.
Whether the full course has been completed.
Which follow-up actions remain unresolved.
Which patients face the greatest risk of subsequent infection.
In the study, the Infection Control Team created lists, contacted wards, collected treatment records and traced patients after discharge. These steps improved delivery, but required substantial coordination and manual work.
Connected Infection Intelligence can help bring laboratory results, patient movement, infection status and follow-up activity into the same workflow. The goal is not another MRSA alert. It is a clear route from detection to ownership, action and completion.
Targeted prevention depends on timely intelligence
This study adds useful real-world evidence for targeted MRSA decolonisation across a general hospital population. It also shows why treatment choice is only one part of prevention.
Hospitals must identify the right patient, communicate the result quickly, coordinate treatment and confirm that the action was completed—even when the patient has already moved beyond the ward.
Targeted prevention works when surveillance is connected to action.
