Monday, November 30, 2015

A curious trend in appendectomies by residents

Some experts are worried that laparoscopic cholecystectomy is so prevalent that future surgeons may have difficulty doing open cases. I was going to blog about the possibility that open appendectomy would become the next operation that next generation surgeon might have trouble with. But while looking at some data [link added 12/4/15] collected by the RRC for Surgery, I was struck by something else.

Since 1999, the total number of appendectomies (open and laparoscopic) performed by surgical residents who completed 5 years of training has risen by 65.1% compared to the total number of appendectomies done in the US, which has increased only 16.4%. Here are the numbers:

Except for the academic year ending in 2006, the average total appendectomy rate per resident has risen every year since 2000. The chart below displays that change and the changes in the numbers of open and laparoscopic appendectomies.

Click on chart to enlarge
The difference in the average combined number of appendectomies between the two academic years ending 2000 and 2014 is significant, p < 0.0001.

The population of the US rose from 279 million in 1999 to 318.9 million in 2014, an increase of 14.3%. The number of appendectomies done in the US for those years was about 281,000 vs. about 327,000 respectively, a 16.4% increase. The appendectomy rate increase does not significantly differ from the increase in population.

The difference between the total number of appendectomies done in the US and the total done by residents at teaching hospitals was significant, p < 0.0001.

What is going on here? Why is the increase in the rate of appendectomies being done by residents significantly higher than the rise in the rate of appendectomy for the entire country?

I posed this question on Twitter and got a number of replies.

Could the difference be due to more appendectomies being referred to teaching hospitals? It is possible, but except for a few anecdotes, a pattern of has not emerged. For example, the change in total numbers of cholecystectomies has been much more modest. In 1999-2000, graduating residents did 99.5 cholecystectomies and in 2013-14, they did 120.9, an increase of only 21.5% which is significantly lower than the increase in appendectomies, p < 0.0001.

Have academic centers increased their catchment areas? I don't think catchment areas have changed much in 15 years.

Are attending surgeons lazier than they were in 1999-2000? That's not likely. For most cases, they had to be present in the OR anyway in both eras.

Are surgeons doing more appendectomies to increase their incomes? If that were so, the number of normal appendices (that is, appendices that were judged to be not inflamed by pathologists) would be higher than it was in 99-00. In fact, the reverse is true. The rate of normals has fallen drastically over the years.

Residents might not have logged all their appendectomies in the past or maybe they are inflating their numbers now. Those are interesting theories both of which cannot be proven.

Could it be because of increased use of CT scans? In a recent New England Journal paper, Dr. David Flum suggests the CT scan that just about every patient with right lower abdominal pain now gets may be too sensitive. This could lead to the overdiagnosis of mild cases of appendicitis that might have resolved without any intervention.

However if the issues is CT scan overdiagnosis, the rate of appendectomy nationwide would be as high as the rate in teaching hospitals.

I'm out of ideas. Can you explain the appendectomy mystery?

By the way—in case you missed it in the table above—over the 15-year period, open appendectomies have decreased 68% and laparoscopic appendectomies have increased 546%.


Carolyn Thomas said...

I'm curious about what happened to account for that big slump in appendectomies during that academic year ending in 2006?

Skeptical Scalpel said...

I have no idea. No major changes in rules, programs or resident numbers took place. It's another mystery.

William Reichert said...

The "big slump" is probably de to "big data syndrome". This is where nothing new has happened but the coding for the data resulted in
the original thing being measured being assigned to another code. For example, appendectomies were assigned to the code for a sub category as a part of "large bowel surgery". Another possibility is that the computer started being used in 2006 and there was malfunction that took a year for the programmers to fix it. This is similar to the
subject of the book>"How to lie with statistics" although now we have
it changed to "How to be deceived with statistics"
As to the original issue of the residents doing more surgeries, I wonder if this is because there was a change in the assignment of the fee from the attending to the resident so that the insurance payout could be reduced.

Skeptical Scalpel said...

William, the coding was not changed. There has always been a separate code for appendectomy. That code was subdivided into open and laparoscopic sometime before 1999-2000. As far as I know, there was no computer malfunction. Fees have never been assigned to residents. Insurance companies, like patients, have no way of knowing if the resident or the attending surgeon did the case.

ffolliet said...

I'm sorry to point out the obvious but you are being mislead by the way the numbers are portrayed.

There has been a 46% increase in the number of appendicectomies. (I'm British)

Residents are operating on 8% more of the total number of cases.

More laparoscopic than open, as previously

And on ly an 11% increase in the number of residents.

In other words, more appendicectomies overall, more of which being done by trainees, the majority being done laparoscopically but only a small increase in number of trainees.

No conspiracy

Skeptical Scalpel said...

ffolliet, I will respond to your comments, which I have quoted, one by one.

“There has been a 46% increase in the number of appendicectomies. (I'm British).”

I do not know how you arrived at that 46% figure. Perhaps you could share your calculations.

“Residents are operating on 8% more of the total number of cases."

Again, I don’t know where your 8% number comes from.

Even if that 8% number is true, the number of residency programs was 252 in 1999-2002 and 250 in 2013-2014. The pie should have been cut into smaller pieces (in other words, more residents were available to operate on all of the patients with appendicitis who came into the same number of hospitals). This should have resulted in a lower number of appendectomies per resident rather than the number of cases markedly increasing.

“And only an 11% increase in the number of residents.”

Yes, only an 11% increase in the number of residents but a 65% increase in the *average* number of appendectomies being done.

“In other words, more appendicectomies overall, more of which being done by trainees, the majority being done laparoscopically but only a small increase in number of trainees.”

You did not consider that the US population increased 14.3% and the appendectomy rate increased 16.4% -- nearly the same percentage increase. Even if your 46% increase in the number of appendectomies is correct, that is a still significant and unexplained difference. You also ignored the fact that the average rate of cholecystectomies being done by residents did not increase nearly as much as the average rate for appendectomies.

“No conspiracy.”

Nowhere in the post did I use the word *conspiracy.* I said it was a *mystery.* I posted the article because I was hoping someone could explain it. I’m afraid your explanation needs more work.

ffolliet said...

"The number of appendectomies done in the US for those years was about 281,000 vs. about 327,000 respectively".

From the table of number of appendicectomies being done by residents 38769 and 71,494 respectively. The rise from 38769 to 71,494 is 46% of the 2014 total.

The appendicectomy rate for the population (279 million in 1999 to 318.9 million in 2014) is 0.001 for each year, effectively unchanged.

As a proportion of all appendicectomies performed by residents is (38769/281000) 13.8% and (71494/327000) 21.9% respectively, thus a rise of 8% in the number of appendicectomies being done by residents. I would suggest this may reflects the fall off in Chiefs doing laparoscopic appendicectomies or a view that teaching needed to be improved.

I'm not sure why you feel this is "a still significant and unexplained difference." There were 32725 more appendicectomies to be done but by only 116 extra residents.

The dramatic increase in the laparoscopic appendicectomy rate for residents surely is a reflection of the fact that nowadays the vast majority of procedures are done laparoscopically, as opposed to the number of trainees performing that surgery in 1999.

I'm not sure the value of considering cholecystectomies performed in this analysis and the figures aren't available to comment specifically but isn't it likely that the experience of the chiefs in cholecystectomy and willingness to allow the residents to perform has not increased at the same rate as their willingness to allow the resident to perform an appendicectomy?

No mystery. I think it is just the way the numbers are being viewed. More people, more appendicectomies, a greater percentage being done by residents but not very many more residents. A good thing for training?

Skeptical Scalpel said...

I appreciate your trying to clarify this. You have made me and my mathematician consultant think about everything in different ways. Here are my responses to your comments, which are in quotes.

"The rise from 38769 to 71,494 is 46% of the 2014 total."

On this side of the pond, we calculate change from the initial to the final value using this formula: [FINAL - INITIAL]/INITIAL X 100 = % change, which in this instance is 84%. Using the average values instead of the total values in the formula gives you a 65% increase which is what we displayed. It turns out some British use my method as well. Here are two of many UK websites that do it the way I did [ and].

"I would suggest this may reflects the fall off in Chiefs doing laparoscopic appendicectomies or a view that teaching needed to be improved."

As I stated in the post, the figures for cases performed was for the cumulative total of the 5 years of training for each graduating resident. Therefore, the difference in cases done could not have been solely due to chief residents passing cases along.

"There were 32725 more appendicectomies to be done but by only 116 extra residents."

The additional cases were not only done by the 116 additional residents. They were distributed among the 1105 residents over the aggregate years of their training. The 64.7 figure was the *average* number of cases for the 1105 residents. The whole point of my post was that there has been an unexplained and excessive increase in the number of appendectomies done by graduating residents. Despite yours and my best efforts, it remains unexplained.

Regarding your comment about the dramatic increase in appendectomies being done laparoscopically—those of us here in the US didn't need the data to confirm this. We have seen it with our own eyes.

Vamsi Aribindi said...

I think a lot of the issue here is that recording cases has become a game. As medical students start looking at case numbers to decide where to go, everyone started being more cagey about them. A few places I interviewed at had their interns put in ports for a case and call it a diagnostic laparoscopy- which to be fair is actually what they are doing. In some places, Chief residents would often just stop recording cases after they hit their ABS minimums (apparently the programs would get stern letters if they logged too many cases, the assumption being that the Chiefs weren't spending enough time in clinic). Another place put in central lines in the OR, and called it a case.

I would suspect that this trend is the result of residents being given an edict to start recording cases they previously hadn't so that their programs would look better, or some data manipulation of that sort.

Skeptical Scalpel said...

Vamsi, thanks for the interesting comments. I alluded to data manipulation in the post. You came right out and said it. It's a bigger problem than most people think. There is no way for the RRC to verify any of the log entries submitted.

Anonymous said...

Another possibility:

Increase in surgery residents double scrubbing cases and both logging the same case (one as surgeon junior, one as surgeon senior). This became ever more frequent over my residency as the trend went from an attending scrubbed with a single resident to an attending who is not scrubbed watching the laparoscopic monitors and observing while a chief resident led an intern or second year resident through the case. Does your data set include cases logged by PGY level?

Skeptical Scalpel said...

The data PDFs are remarkably reticent about the teaching assistant numbers, but a look at what is there seems to indicate that there were more teaching assistant cases in 1999-2000 than in 2013-2014.

I have updated the post with a link to all the data. Here it is

Post a Comment

Note: Only a member of this blog may post a comment.