Angus Hotchkies 7/05 07:10PM
Ok great we'll make a start
Welcome to SCALPEL's second online journal club of the year! Thanks for coming along, hopefully it will be a useful discussion! We are joined by Daniel Lewis, a neurosurgical registrar who will facilitate and input to the discussion. I’m Angus Hotchkies, a 3rd year based at Salford. Firstly if everyone wants to introduce themselves and include an email (so we can send certificates out later) that would be great!
Sarah Michael 7/05 07:11PM
Hi I'm sarah, I'm a second year
Sashavuc 7/05 07:12PM
Hi guys I'm in third year
Sianjuniper 7/05 07:12PM
Siân Juniper, third year at UHSM
Danni Wilkinson 7/05 07:12PM
Danielle, third year at Salford
Angus Hotchkies 7/05 07:12PM
Danlewis 7/05 07:13PM
What were your thoughts on the paper everyone?
Sashavuc 7/05 07:14PM
Are these devices commonly used in practice ?
Angus Hotchkies 7/05 07:14PM
I thought it has quite a big impact factor since presumably there is quite a lot of time and money being invested into wound edge protection devices. And this paper seems to show there is no significant benefit to using them.
Sianjuniper 7/05 07:14PM
I thought it was interesting when I read it because you don't often hear about the studies that prove new technologies are NOT useful
Danlewis 7/05 07:16PM
Your right Sian I think it goes against the trend in reporting negative results but as you have said the shear size of the study makes its findings more relevant
Sianjuniper 7/05 07:17PM
however it is quite an old paper (2013) so i don't know if there have been newer things introduced now
Sarah Michael 7/05 07:17PM
I agree especially as it goes against the two other systematic reviews
Sianjuniper 7/05 07:17PM
from a clinical point of view as Angus said, will save a lot of resources if they are deemed unbeneficial
Danni Wilkinson 7/05 07:18PM
I was honestly quite surprised at the figures re. wound infections post surgery, had no idea they were so high, especially seeing so much surgery recently and seeing the protocols to avoid this
Sianjuniper 7/05 07:19PM
ikr shocking isn't it! seen so many referred to ID after surgery in recent weeks
danlewis 7/05 07:19PM
I think we all agree the clinical question was a valid one. Do you think the study design and randomisation was appropriate?
Danni Wilkinson 7/05 07:20PM
I think blind assessment of the site infection was necessary
Sianjuniper 7/05 07:20PM
It included so many patients over such a long time which i think is good because it will give clearer results
Angus Hotchkies 7/05 07:21PM
designed seemed pretty good: assessors were blinded etc. but they were properly trained
danlewis 7/05 07:21PM
The authors do acknowledge the high infection rate and mention in the discussion that their prospective follow up of patients in clinic likely counteracted previous under reporting of infections
Danni Wilkinson 7/05 07:23PM
Are there different results from clinicians and the self reported infections? (sorry just trying to find now)
Ignore, found it sorry!
Sianjuniper 7/05 07:25PM
there was no difference, have i read that right?
Danlewis 7/05 07:26PM
Do we think the study was adequately powered to detect a difference?
Sianjuniper 7/05 07:27PM
as in between the surgeon reports and self-reports?
Danni Wilkinson 7/05 07:28PM
I think that using data such as if the patient required further antibiotic treatment would be more reliable than self reported infections
Sianjuniper 7/05 07:28PM
yeah maybe clinical notes would be more reliable, but then again not everything is documented
Danni Wilkinson 7/05 07:29PM
Also, did patients have any information given in aiding them to recognise an infection?
Danlewis 7/05 07:29PM
Between device and no device. The authors include a power calculation in the statistical analysis. I highlight it as it's not something you always see even in studies of this size.
Sashavuc 7/05 07:29PM
I presume the questionnaire consisted of questions aimed at uncovering an eventual infection
Danni Wilkinson 7/05 07:30PM
Ah, I see apologies I'm still having a read through
Sianjuniper 7/05 07:30PM
oh okay! thats interesting - would you mind explaining that please? the stats confuse me slightly
Angus Hotchkies 7/05 07:31PM
So does it main that there's an 80% there is going to be a statistically significant result?
**80% chance sorry
danlewis 7/05 07:34PM
So if anything there study is overpowered to detect a difference. They initially went on the assumption that they were looking for a 12% difference in infection rates so calculated how many patients they would need to be sure of detecting that difference which was 710 having an 80% power. The 80% means an 80% chance of correctly rejecting the null hypothesis, the null hypothesis being that there is no difference in infection rates between device and no device
Sianjuniper 7/05 07:35PM
ohhh i see
danlewis 7/05 07:35PM
Put another way using 80% power the chance of a type II error (false negative) is 20%.
Sianjuniper 7/05 07:36PM
thats rather big
danlewis 7/05 07:36PM
The alpha value was 5% so they wanted the chance of producing a type I error (false negative) less than 5%
Sashavuc 7/05 07:37PM
what are the implications of overpowering it ?
danlewis 7/05 07:38PM
It does seem big but an issue with power calculations is that there is always a trade off between achieving an acceptable study power whilst reducing th risk of a type I error (false positives). To achieve a very high power and a low type I error rate you often need very high numbers indeed
Sianjuniper 7/05 07:39PM
so even though this study had large numbers its still unusual?
Danlewis 7/05 07:39PM
Because the actual infection rate was much higher, they probably needed less patient numbers to be 80% sure of seeing a true effect.
Sianjuniper 7/05 07:40PM
sashavuc 7/05 07:41PM
okay but wouldn't a larger population sample make the eventual statistical analysis more significant ?
Danni Wilkinson 7/05 07:42PM
That's what I was thinking
Danlewis 7/05 07:42PM
That's why many drug trials which are looking for very small differences in outcome (<5%) require very large patient numbers. Indeed a larger patient allows you to either increase the power of your study or reduce the type 1 error rate (false positive), it depends what you are after
Sashavuc 7/05 07:44PM
ahh I see thanks for that
Angus Hotchkies 7/05 07:44PM
Ok great thanks
Sashavuc 7/05 07:45PM
The graph showing the odds ratios between different factors affecting the outcomes are really interesting especially how it seems to be displaying a significant association between lower patient BMI and eventual infection
Angus Hotchkies 7/05 07:47PM
Did they even discuss that result?
Danlewis 7/05 07:48PM
That's a good observation. The device appeaed to be better at preventing infections in patients with a BMI of over 25
They didnt discuss the result nor did they randomise patients on the basis of BMI
Do you think BMI >25 is a helpful cutoff though?
Sashavuc 7/05 07:50PM
I don't think they discussed it but theres a graph showing the odds ratios between different factors. I think it summarises them quite well but none of them seem to stand out individually
Angus Hotchkies 7/05 07:50PM
I guess it's standard practice for defining a patient as overweight - it's not always an accurate figure though
Sianjuniper 7/05 07:50PM
people with an overweight BMI are at an increased risk of infection as it is
Angus Hotchkies 7/05 07:51PM
Yeah its a good table!
Sashavuc 7/05 07:51PM
yeah and I would presume higher BMI's as having a signficantly higher post-surgical complication rate
but this study shows that its actually the other way around
danlewis 7/05 07:51PM
It is standard practice but if you look at the literature in other areas such as lumbar back surgery, the complicatio and infection rate really goes up once BMI is greater than 30
Does higher BMI have a lower infection rate though
Sashavuc 7/05 07:53PM
oh no sorry
interpreted that wrong
just has a lower odds ratio
danlewis 7/05 07:54PM
If you look at the numbers the infection rate in under BMI 25 is about 13% whereas its 54/206 in the BMI>25
A slight concern with BMI data is that they have quoted median values. Whilst the median values are comparable it suggests that the BMI distribution is non-normal so they likley had some very obese or very thin people in either group
Danni Wilkinson 7/05 07:57PM
In the study you just mentioned, how much did the infection rate go up after BMI of 30?
/was it even more than 50% here?
Danlewis 7/05 08:01PM
Some studies report inceases in infection rate from <5% to over 10% amongst obese patients
Danni Wilkinson 7/05 08:01PM
Danlewis 7/05 08:02PM
Whilst the spine is a cleaner surgical field, the same mechanisms apply- more fat means more tissue retraction, more tissue necrosis, more pain, less ambulatory etc. Never mind the likley imunosuppresive effects of the metabolic syndrome
Did we all think their conclusions were valid?
Angus Hotchkies 7/05 08:03PM
What I thought was interesting was the fact that these protection devices should work in theory and the problem might lie in the fact that the infections are just caused when the device is removed and the surgeons are closing up.
Danni Wilkinson 7/05 08:04PM
Or even further after that
Angus Hotchkies 7/05 08:04PM
Danlewis 7/05 08:04PM
Exactly and if you look at their secondary analysis its the one thing they didnt control for
Angus Hotchkies 7/05 08:05PM
So that seems to be a major weakness but they acknowledge that
I think the conclusions seem valid as they suggest more large scale studies to evaluate efficacy
Sashavuc 7/05 08:07PM
Yeah that's a really good point. Does anyone know if these devices are used at all in Manchester ?
Sianjuniper 7/05 08:07PM
ill google it now
okay mayne not that simple haha
Sarah Michael 7/05 08:08PM
they also say in some of the other studies that showed it worked they had double ring devices and they only tested single rings
danlewis 7/05 08:12PM
That may have been a factor but personally I was not surprised by the findings of this study. It sis difficult to say how efficacious it is without comparing wound closure. the skin edges can be beautifully protected throughout but if the closure is inadequate then the wound will get infected irrespective
Sianjuniper 7/05 08:13PM
it is good they acknowledged that though
Angus Hotchkies 7/05 08:14PM
Would a study on wound infection and changing gloves/re-sterilising before closure be useful then?
Danlewis 7/05 08:17PM
It is an interesting idea, and in some surgeries routinely done. When inserting VP shunts for example, gloves get changed before handling the shunt tubing
Angus Hotchkies 7/05 08:18PM
Ah ok that's interesting
Danlewis 7/05 08:18PM
It may be the start of a slippery slope though as the question then arises and what point do the gloves become 'dirty'
Angus Hotchkies 7/05 08:18PM
danlewis 7/05 08:21PM
Further to the point I made before here's an article on obesity in lumbar spine surgery if anyone is interested https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4868585/
Danni Wilkinson 7/05 08:22PM
Angus Hotchkies 7/05 08:22PM
Sianjuniper 7/05 08:22PM
danlewis 7/05 08:24PM
Does anyone have any further comments or questions on the paper?
Sianjuniper 7/05 08:24PM
no but thank you so much for the stats advice
i struggle sometimes when i read papers because the stats confuse me and therefore it makes reading the study so much harder and harder to have an opinion on
Angus Hotchkies 7/05 08:25PM
No the big one for me was the fact that the devices likely do have an impact - it's just the procedure before/after that might be causing infection
Danlewis 7/05 08:27PM
No problem, power calculations can be a confusing area at first but once you appreciate the power is essentially 1-false negative rate it becomes much clearer. Its an important thing to consider, so many studies especially pre-clinical animal ones have been done without a power calculation and they were likely underpowered to find a difference
Danni Wilkinson 7/05 08:28PM
Thanks for clearing things up
Danlewis 7/05 08:28PM
I will need to go but thank you for all your insightful comments and enthusiasm
Sianjuniper 7/05 08:29PM
thank you so much for your time!
Sarah Michael 7/05 08:29PM
Sashavuc 7/05 08:29PM
Thanks for everything Dan
Angus Hotchkies 7/05 08:30PM
Thanks very much Dr. Lewis and thank you everyone for coming
Danlewis 7/05 08:30PM
No problem, speak to you all at the next one hopefully and thanks to Angus for organising
Angus Hotchkies 7/05 08:30PM
Sending certificates now - thanks again guys!
Sianjuniper 7/05 08:30PM
amazing - thanks!