The Journal of Infectious Diseases has just published a special issue on norovirus, which is well worth reading. When norovirus strikes, there is an inclination to close the ward to new admissions at the earliest available opportunity in order to protect incoming patients. But when should the ward closure trigger be pulled? Not at all, as recommended by latest UK guidelines (risking continuation of the outbreak, fed by a steady stream of new victims…I mean admissions), when you get a single case of vomiting or diarrhoea (lots of unnecessary ward closure) or only when you have a lab confirmed outbreak on your hands (by which time the horse has already bolted and galloped through your hospital). The special issue included a useful modelling study providing some idea of the impact of various approaches to ward closure in response to noro outbreaks.
The authors developed an epidemic simulation model to test the impacts of various approaches to ward closure, including an assessment of cost-effectiveness. The scenarios tested were ward closure 1, 3 or 5 days after the detection of the first case, combined with 3 estimates of the effectiveness of ward closure: 0%, 25% and 50%. The cost-effectiveness analysis was driven by the cost of lost bed days, and the cost of lost staff productivity.
The study used a well-characterised dataset from southwest England between 2002 and 2003. Worth noting at the outset that this dataset is now almost 15 years old; I suspect that the epidemiology of noro will have changed over that period.
The dataset included 232 new outbreaks leading to 2494 outbreak-days over 1 year. Perhaps unsurprisingly, the additional risk of a new outbreak was greatest from an admission ward, and lowest from a long-stay ward, due to the pathways of the patients involved. Surprisingly, the level of activity had very little impact on the risk of in-hospital norovirus outbreaks.
In the epidemic simulation, ward closure always reduced the number of new cases, even when it was assumed to have no direct impact on reducing transmission. This is because ward closure reduces the duration of the outbreak (effectively by cutting off the supply of fuel (patients) to the fire (norovirus). Without any intervention, there would have been 34 outbreaks across 29 wards in your average acute care hospital (imagine that!); ward closure would reduce this to between 15 (for ward closure on day 1 with 50% efficacy) or 32 (for ward closure on day 5 with 0% efficacy).
In terms of cost-effectiveness, in none of the scenarios was ward closure cost saving. This is not surprising, since closing a ward is always going to result in lost bed days, which are expensive. However, if we apply a fairly low “willingness to pay” threshold of around £2500 (how much you are prepared to pay to prevent a case or an outbreak), ward closure is cost-effective to prevent a case assuming a 25% efficacy, but less so to prevent an outbreak (>£35,000, requiring a 50% efficacy).
So where does this analysis leave us in deciding whether or not to close a ward due to norovirus? If you are not interested in the cost-effectiveness equation then it makes sense to close wards – and to close them early. But, since the cost-effectiveness is an increasingly important driver, this study suggests that the costs required to close wards in an effective way don’t yielded the degree of benefit to justify them. Although not evaluated in this study, it seems to me that bay closure would be a better approach (and in line with current UK guidelines).
Image: Public Domain Images.