Colorectal surgery in northern Europe: survey of practice

Rhona MacDonald takes you through a short report and encourages you to question everything

This month's paper is Lassen K, Hannermann P, Ljungqvist O, Fearon K, Dejong CHC, Von Meyenfeldt MF, et al. Patterns in current perioperative practice: survey of colorectal surgeons in five northern European countries. BMJ 2005;330:1420-1. You can read it by going to studentbmj.com and clicking on the link. There is no abstract as it is a short report.

Please don't be put off by the title of this paper or the excessive number of contributors. If you take the time and effort to read the paper, you will find that it is not difficult to follow and that the results are displayed in a simple format (percentages of responses), with no statistics in sight (see table).

This paper is an example of what the BMJ describes as a short report, which means that it has to be no more than 600 words and only have one figure or table. It was published with an editorial.¹ Following published papers is really just common sense. As with many other things in life, you want to know the who, why, what, how, and, finally, where does it leave us?

Who?

Before you read any paper published anywhere, read who the authors are, how the study was funded, and whether they have any competing interests, and if so, what these are. You should consider these when evaluating the study more than anything. For example, before reading a study about how effective a certain drug was when used in a double blind randomised controlled trial-hopefully, a term you now understand-against a placebo, you should routinely study the funding statement and competing interests. If the funding statement said that the drug company who made the drug funded it, and in the competing interest statement it said that some of the authors (or even one) regularly received money from this drug company to talk at symposiums, this may make you suspicious about the credibility of the results.

But not all journals publish funding statements. Also, not all journals have the same rigorous process as the BMJ for getting authors to make competing interest statements and publishing them. This means that sometimes you could be reading a study and not know that the authors had received funding from drug companies. And competing interest statements are largely taken on trust-there are no competing interest police as such.

Responses (percentages) to questionnaire on perioperative care in colonic resections in five northern European countries

Responses	Scotland	Netherlands	Sweden	Norway	Denmark
Response rate	72.4 (n=55)	83.7 (n=36)	68.3 (n=43)	92.7 (n=38)	70.0 (n=28)
For an elective left sided* hemicolectomy for cancer, would bowel preparation be administered?
No	4	18	3	5	19
Yes, oral purgative	85	52	95	89	62
Yes, enema	4	12	0	3	15
Yes, both	8	18	3	3	4
Is nasogastric decompression tube routinely left in place for more than four hours after surgery?
No	75	22	83	82	85
Until next morning	12	36	17	18	11
Two days or more	12	17	0	0	0
Until bowel movement	2	25	0	0	4
Is epidural analgesia used routinely postoperatively after transfer to general ward? (Not high dependency ward/intensive care unit)
Yes	11	83	93	89	96
No	89	17	7	11	4
Are there any restrictions on routine intravenous clear fluid administration in the first 72 hours after surgery?
Yes	24†	14	5	8	4
No	76	86	95	92	96
Are there any restrictions on routine intravenous sodium administration in the first 72 hours after surgery?
Yes	26†	14	0	8	7
No	74	86	100	92	93
How long would a patient be "nil by mouth" (less than 450 ml clear fluid) postoperatively?
0 days	38	58	71	82	96
1-2 days	46	39	26	18	4
3-4 days	17	3	2	0	0
When would patients be allowed to resume oral intake at will for solids (eat freely)?
Operating day	10	3	12	8	41
Postoperative day one	17	43	32	45	44
At bowel sounds	13	11	23	5	0
At passage of gas	44	17	27	32	15
Bowel movement	17	26	7	11	0

Results are percentages after exclusion of missing or ambiguous responses (overall 2.55% excluded). Percentage is of the total numbers of responses to that question. *The only question in which colonic resection was further specified.†Scottish centres declaring a fluid/sodium restricting routine allowed maximum values of 3000 ml water and 154 mmol sodium per 24 hours (median).

ERAS who?

All of the authors and all of the participants are part of what they describe as the ERAS Group (the Enhanced Recovery After Surgery Group; see box). You might want to know what this is as it may have direct relevance to the study. For example, if participants had not been part of this group, would that have made any difference to the results? I imagine it would as it implies they must have some knowledge, probably more so than surgical departments who were not part of the group. This might bias the results.

Why?

Why did the authors want to do this study in the first place? It is much harder to show best evidence based practice in surgery than in medicine and this is further discussed in the accompanying editorial.¹ So currently a substantial amount of surgery is not based on best evidence.

Bowel surgery
AJ PHOTO/HOP AMERICAIN/SPL

AJ PHOTO/HOP AMERICAIN/SPL

Bowel Surgery

There have been at least five studies published that show best practice in managing patients undergoing colorectal surgery, perioperatively (before, after, and during surgery) that are cited in the paper. So the authors wanted to know if this new evidence had changed the practice of colorectal surgeons.

But is it really realistic to expect surgeons to have read all the studies as each of them are published in different journals? Also, you might want to know more about these studies and journals so that we can trust the results rather than take it for granted that they are evidence based. Were the studies conducted properly and what type of study were they? Are the journals peer reviewed, are they international, would the results have realistically got into the consciousness of colorectal surgeons?

What?

They chose to ask colorectal surgeons this: "We presented a hypothetical case of elective laparotomy with colonic resection for cancer in an otherwise healthy 70 year old man. We asked respondents to answer according to the practice most widely used in their department at that time."

Bowel resections for cancer in older men is very common. But why not a woman, as the female incidence of bowel cancer is greatly increasing? Why did the man have to be "otherwise healthy," as many patients with cancer have either related or other medical conditions? If they had chosen any other scenario would it have influenced the results?

How?

The contributors chose to ask the surgeons the above question by sending them a questionnaire. We know some facts about what they did but there are still things that we don't know:

We know that they sent the questionnaire to 200 centres, but we don't know why these centres were chosen (other than being in the ERAS Group, which could bias the results).

We know the five countries involved, but why just them? Are they representative of Europe, or is it just because they are in the ERAS Group? And if so, is this just because they were easier to contact or because it had some relevance to the outcome of the study?

We can work out how many questionnaires were sent to each country (because we have each country's response rate and the actual number of responses) and they are unequal. Why? Again, is this related to who was in the ERAS group (in which case Scotland has substantially more but why?) and again, could this influence the results.

We know the hypothetical question and from the response table we know the follow-on questions that the respondents were asked in the questionnaire. But read them again. They are leading questions. For example, instead of asking, "How would you prepare for a left sided hemicolectomy?" They asked, "Would bowel preparation be administered?" You have to be careful about having leading questions as it could bias the results.

Then you have all the usual criticism of surveys in which respondents are asked to report what they do. How can we be sure that what they say is true? They might be reporting what they think they should be doing, rather than what they actually do.

Where does this leave us?

There seems to be an awful lot we don't know about this study and how it was conducted. The main conclusion is that perioperative routines in colorectal cancer treatment in northern Europe differ substantially from evidence based practice.

Rhona MacDonald, Oxdocs lead and associate director and the Make Poverty History representative of the Asha Foundations
Email: rhona_macdonald2005@yahoo.co.uk


studentBMJ 2005;13:265-308 July ISSN 0966-6494

1. Urbach DR, Baxter NN. Reducing variation in surgical care. BMJ 2005;330:1402-3.