30th May 2017 - Journal Club - Effect of day of the week on short‐ and long‐term mortality after emergency general surgery
Link to paper: http://onlinelibrary.wiley.com/doi/10.1002/bjs.10507/pdf
Click on the link above labelled "Already logged in? Click here to join in with the latest journal club discussion!" to join the chat 🙂
Angus Hotchkies 5/30 07:34PM Welcome to SCALPEL's first online journal club of the year! Thanks for coming along, hopefully it will be a useful discussion! In a few moments we will be joined by Dr Nick Boxall, a urology registrar based at Salford Royal Hospital, 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!
Gauri Vithlani 5/30 07:35PM Hi everyone, I'm Gauri (firstname.lastname@example.org) 3rd year medical student from MRI
Nick Kowal 5/30 07:35PM Hi everyone, I'm Nick Kowal, 3rd year at UHSM. My email is Nick.Kowal8@gmail.com, thanks Angus.
Mark Woodward 5/30 07:36PM Hi, I'm Mark a 4th year student at Preston (email@example.com)
haphuongdole 5/30 07:36PM Hi, I'm Ha Phuong, 4th year based at Salford (firstname.lastname@example.org)
Dilraj Bhullar 5/30 07:36 PM Hi guys, I'm Dil, 3rd year uhsm - email@example.com
Moradeke Orekoya 5/30 07:36PM Hey everyone, I'm Keke a 3rd year based at Salford (firstname.lastname@example.org)
nboxall 5/30 07:36PM Hi all, great to see 10 participants. I'm Nick, urology SpR at Salford Royal and chairing this evening
Bradley Storey 5/30 07:36PM Hi, Bradley, 4th year at Salford (email@example.com)
Sarah Michael 5/30 07:36PM I'm Sarah a 2nd year going to Wythenshawe (sarah.Michael@student.Manchester.ac.uk)
Sam 5/30 07:36PM Hey, I'm Sam Wade, a 3rd year from Salford (firstname.lastname@example.org)
Angus Hotchkies 5/30 07:37PM Welcome Nick! Thanks for all the emails guysReady when you are. Anyone else that joins midway feel free to introduce yourself and you can send your email over
nboxall 5/30 07:39PM Looks like that's everyone? Great to see a mix from years and sites. Remember this is a very informal journal club so any contribution is welcome or if you have any questions then feel free to askI'll kick with what did people think about the topic?*kick off
haphuongdole 5/30 07:41PM Obviously it's a very sensitive topic that was the root of Hunt's argument for a 7-day week
Angus Hotchkies 5/30 07:41PM I thought it was an interesting paper with a fairly big impact factor
Bradley Storey 5/30 07:41PM Topic was quite relevant to recent politics etc, interesting read too
Sam 5/30 07:41PM Quite an interesting one following all the talk from the political world yeah
Nick Kowal 5/30 07:42PM I agree, was good to see the surgical side of the topic that was popular in the news.
Moradeke Orekoya 5/30 07:42PM The comment section under the article's abstract shows how highly debated the topic still is
Dilraj Bhullar 5/30 07:42PM Pretty much the same, interesting to see that there was no difference in mortality despite claims by politicians, but I guess its only limited to emergency surgery
haphuongdole 5/30 07:43PM It's possible that if the authors are against a 7-day week, they may have been biased during analysing of their data
Mark Woodward 5/30 07:43PM Yes interesting to see there was no difference but it does reference another study which found an increase in mortality at the weekend
Bradley Storey 5/30 07:44PM I did think that considering publication date, after all the strikes and fuss last year. Being doctors/surgeons there must clearly be some vested interest in the results they got, though not saying they purposely would try to skew data
nboxall 5/30 07:45PM I think bias could be a possible issue here but let's not forget it was north of the border where strikes didn't happen
Angus Hotchkies 5/30 07:46PM I think the lead author is an emergency medicine consultant - how you decide if there's any bias I'm not sure.
Gauri Vithlani 5/30 07:46PMI think the fact that it is limited to emergency surgery is a strength as this is carried out throughout the week, whereas elective surgery is mostly Mon-Fri and would skew data.
Moradeke Orekoya 5/30 07:46PM But does they type of emergency surgery performed on weekends differ?
Angus Hotchkies 5/30 07:46PM Yeah I agree
Mark Woodward 5/30 07:47PM Yeah and if so are staffing levels less prepared for complications that might arise
Sarah Michael 5/30 07:47PM but in the paper they also say the weekend surgeries were more complex and higher risk
Angus Hotchkies 5/30 07:48PM The big table on p939 of the paper breaks down the types of surgery performed quite nicely
Sam 5/30 07:48PM They stated that younger patients have been operated on at weekends, but yes, more complex also
nboxall 5/30 07:48PM given that elective surgery rarely happens at weekends, I don't think you can fairly compare the effect of the weekend as opposed to emergency surgery
Sam 5/30 07:48PM And both of those finding were significant, I don't know what could explain that though?
Bradley Storey 5/30 07:50PM yeah in terms of politics does looking into emergency surgery not really meet the argument that hunt etc put forward? emergency treatment always available but other care is lacking etc
nboxall 5/30 07:50PM so, we all agree interesting and relevant topic with possible bias? I have a few issues with the methodology so will dive into the meat shortly. What did people think about the design?
Nick Kowal 5/30 07:51PM I thought that selection of all specialty emergency procedures was good. Although selecting patients with one admission and omitting those with several could have skewed the data, but then this was highlighted in the weaknesses.
Sam 5/30 07:52PM Nice large population (does that fall under design?)
Angus Hotchkies 5/30 07:53 PM I also thought it was a large population but the paper actually highlighted that the population was quite small compared to other similar studies
Gauri Vithlani 5/30 07:53PM Had the same thought.
haphuongdole 5/30 07:53PM They also mentioned the difficulties with coding during data collection
Bradley Storey 5/30 07:54PM i didn't think codes were ever that consistent but I'm not sure
nboxall 5/30 07:54PM 50,000 sounds like a lot, but can anyone pick out its massive downfall in design? I can identify two significant ones.
Gauri Vithlani 5/30 07:55PM I didn't understand the flow chart where they exclude non-emergency surgery, then exclude non emergency general surgery. Surely non emergency general surgery is included in non emergency surgery. Not sure whether this comes under design or methods.
Nick Kowal 5/30 07:57PM Is it the fact that we are comparing all emergency procedure regardless of the type? Would comparing individual procedures done on weekends or weekdays give a better idea?
nboxall 5/30 07:58PM look at table 1 and the breakdown of procedure categories. Now does anyone know what small emergency op litters most emergency op lists?
Mark Woodward 5/30 08:00PM Cholecystectomy?
Angus Hotchkies 5/30 08:00PM Appendectomy?
Moradeke Orekoya 5/30 08:00PM Appendectomy?
Angus Hotchkies 5/30 08:01PM And they have a cutoff is 16 for the study?*of
Nick Kowal 5/30 08:01PM Yeah I couldn't think of a reason for the 16 cut off
nboxall 5/30 08:02PM So about 40% of them are 'skin and soft tissue' ops. This is going to include abscesses. I can't see anywhere these are excluded. I've never known a mortality from an abscess. This dilutes the results significantly and this, in my opinion, is where the study falls down.
Sam 5/30 08:04PM Oh that's interesting, so would it be more appropriate to exclude clearly safe surgeries?
Angus Hotchkies 5/30 08:04PM So in their inclusion criteria, do they need to only use emergency surgeries where is a relatively significant mortality rate?Same question as Sam basically
haphuongdole 5/30 08:05PM Would it be better if they had compared the results for each risk category?
Bradley Storey 5/30 08:05PM but if they have a similar proportion on weekdays and weekends in a similar ratio wouldn't their effect not matter? its looking at a relative difference rather than absolute mortality?
Sam 5/30 08:05PM I suppose when the study is designed to compare mortality rates that would make sense, however with these surgeries being involved in both weekday and weekend groups, would it affect the data that significantly?
nboxall 5/30 08:09PM Whilst there was risk adjustment, these cases should have been completely excluded
haphuongdole 5/30 08:10PM And what is the second downfall in the study design that you mentioned?
nboxall 5/30 08:12PM 1st line of the methods in the abstract- all cases between 2005-12. Working practices and surgical techniques have changed. Even since 2012 I've seen many changes in practice in emergency surgery. 2005 things were very different without EU working directives, FY1s were still PRHOs etc
Bradley Storey 5/30 08:13PM oh
haphuongdole 5/30 08:14PM I think it's only 3 years from the beginning of 2005 to end of 2007
Sam 5/30 08:14PM What would a more suitable time frame be for such a study?
nboxall 5/30 08:14PM Sorry, yes but followed through to 2012. Even so, my point stands
Gauri Vithlani 5/30 08:15PM So may not be representative of current rotas anyway?
Nick Kowal 5/30 08:16PM Is it possible that some of these changes have somehow contributed to a weekend effect that is talked about in Politics today? I presume all new working practises were implemented to improve healthcare but just wondering.
nboxall 5/30 08:17PM more recent data! The days of NELA and the midnight cut off as well as change of practice and the way junior doctors have gone to from being on a firm to just being on the wards means data from when I was starting medical school is old and outdated
haphuongdole 5/30 08:20PM In order to include more recent data, future studies probably will have to just look at the short term mortality rather than long term mortality. Otherwise these studies will always lagging far behind
Angus Hotchkies 5/30 08:22PM Agreed - I think for a greater impact factor more recent data would be required. If there was a more recent study, say 2014-2016, would a criticism then be that there wasn’t a long enough time period for any meaningful follow up?
nboxall 5/30 08:22PM I'm not sure long term mortality is really going to be reflective of the weekend effect
Sam 5/30 08:22PM I don't really understand where long-term mortality fits in anyway, obviously peri-operative mortality could have a link to a weekend effect if there is one
nboxall 5/30 08:22PM agreed, Sam
Sam 5/30 08:22PM but long-term mortality suggests that the patient has been subjected to care both during the midweek and weekends anyway
Bradley Storey 5/30 08:23PM not necessarily?
nboxall 5/30 08:24PM The chances if you're having a laparotomy is that you're going to be in hospital for both a weekend and during the weekappendix, less soChances of perioperative mortality from an appendix will be low, laparotomy will be differentfyi we're starting to run out of time, probably another ten mins or so
haphuongdole 5/30 08:27PM Probably a minor point now, would calculating 'time to surgery' in hours something feasible and more accurate?
Angus Hotchkies 5/30 08:27PM Ok thanks Nick
nboxall 5/30 08:28PM More accurate but a bit more difficult to calculate Any final thoughts from people?
haphuongdole 5/30 08:32PM No more from me. Thank you very much for chairing, Nick and Angus. I've learnt a few new things today!
Mark Woodward 5/30 08:32PM Thanks, I have certainly learnt to take extra time to look carefully at the methods
Gauri Vithlani 5/30 08:33PM I thought the percentage differences seemed really small initially before I looked at the p values. Thanks everyone defo useful thanks Angus thanks Nick
Angus Hotchkies 5/30 08:33PM If no one has any other questions, I had one quick question on the statistics side - they used non-parametric tests (chi squared and Mann Whitney) but could the population be considered a normal distribution hence a parametric test could be used?
nboxall 5/30 08:34PM Yeah I thought that was strange when they actually stated that the data was normally distributed
Angus Hotchkies 5/30 08:35PM Yeah I wasn't sure why they talked about that and then used non-parametric tests - I don't have a great understanding of complicated statistics as it is!But I couldn't see anywhere where they justified the use of them
nboxall 5/30 08:35PM You can still use non-parametric tests with parametric data but it's not as accurate
Angus Hotchkies 5/30 08:36PM Yeah
nboxall 5/30 08:38PM my thoughts are that this study has too many confounding variables, using out of date data, uses irrelevant data and I'm not sure how they're able to draw conclusions from this. Also, whether or not there is bias? Finally I think this is where we're seeing the rise of prospective audits such as NELA and PQUIP (look the, up if you don't know about them)
Sam 5/30 08:39PM Okay yeah I'm out, that's given me something to read up on, thanks Nick, thanks Angus!
Angus Hotchkies 5/30 08:39PM As we are approaching an hour, I think that's a good time to end unless anyone has any final questions for Nick!Thank you Nick for all the help!
Sarah Michael 5/30 08:40PM thank you!
Moradeke Orekoya 5/30 08:40PM Thanks Nick and Angus
Bradley Storey 5/30 08:40PM Thanks
Nick Kowal 5/30 08:40PM No questions from me, thank you both for the session!
nboxall 5/30 08:41PM Thanks all for your contributions tonight
Angus Hotchkies 5/30 08:43PM Thanks again everyone for contributing - will send out certificates before the end of the week!