Tuesday, 25 June 2013

Hospital Weekend Mortality and the FFS analysis technique.

Where did all this nonsense about "excessive deaths at the weekend in hospitals is because they don't run a 24/7 service like supermarkets" come from?

'The NHS is not Tesco', say Doctors. Telegraph June 2013

BMA calls round-the-clock 'Tesco NHS' plan ridiculous. Independent June 2013

Tesco-style NHS plan 'ridiculous'. BBC June 2013

Sunday hospital admissions 'a bigger risk'. NHS Choices February 2012

Press Release: New report finds higher death rates at hospitals with fewer doctors at evenings and weekends. Dr Foster November 2011

The rhetoric in the UK seems to assume some miraculous recent discovery of data analysis in this area and somehow peculiar to the NHS and explained by factors such as European Working Time Directive, patient choice, out-of-hours GP contract, MRSA, unions, lack of regulators, too much regulation, weak colleges, the wrong type of teabag. However, 12 years ago in Canada they looked at nearly 4 million hospital admissions and found:
"Patients with some serious medical conditions [ruptured abdominal aortic aneurysms, acute epiglottitis, pulmonary embolism] are more likely to die [compared with 3 control conditions MI, intracerebral haemorrhage, acute hip fracture] in the hospital if they are admitted on a weekend than if they are admitted on a weekday." [1]
So, some conditions - that can present in a challenging way and require senior input into their correct diagnosis and management - seem to be associated with excess mortality. This seems to make sense, especially as it is controlled against conditions that, on the whole, present in an unequivocal way and require management as per an easily recognised protocol.

This is the medical evidence dialogue that has been lost in the low rigour, non-peer reviewed, middle-management led, headline grabbing, trawling of evidence - what I'd refer to as the Fifth Form Spreadsheet (FFS) technique. The FFS technique is where you gather whatever data you can, commenting on but largely ignoring any form of potential bias, go fishing for differences and correlations, and release a press release saying there is a 'signal of evidence' leaving an open goal for whatever political commentator wants to score their own prejudice.

Careful consideration of the case mix and cautious interpretation of the results may show otherwise. [2]

Can we have some more rigorous analysis of what the possible case for causation is between our spreadsheet 'big' data before diverting resources into 24 hour care? Those who present at weekends are likely to be different to those who present in the week because of the different social norms and interpretation of vague symptoms. More staff and services in a hospital won't necessarily help with challenging presentations but could be very beneficial for certain types of admission.

How about focussing resources where there is evidence that it would help and not over-interpret whatever data we can analyse FFS.

1. Bell CM, Redelmeier DA. Mortality among patients admitted to hospitals on weekends as compared with weekdays. N. Engl. J. Med. 2001 Aug;345(9):663–668. Available from: http://www.ncbi.nlm.nih.gov/pubmed/11547721

2. McShane P, Draper ES, McKinney PA, McFadzean J, Parslow RC, Paediatric Intensive Care Audit Network (PICANet). Effects of Out-of-Hours and Winter Admissions and Number of Patients per Unit on Mortality in Pediatric Intensive Care. J. Pediatr. 2013 Apr; Available from: http://www.ncbi.nlm.nih.gov/pubmed/23623513

No comments:

Post a Comment