Without evidence of benefit, an intervention should not be presumed to be beneficial or safe.

- Rogue Medic

Definitely Certain


ResearchBlogging.org

Also posted over at Research Blogging. Go check out the rest of the research blogging there.

This study was looking at the use of MDCT (MultiDetector Computed Tomography) scanners to assess for the presence of appendicitis in patients with a physical exam/history suggestive of appendicitis (as the study’s title suggests). It provides an interesting look at how good the self-assessments were of the accuracy of the scans.

Definitely positive = 556/567 or 98% positive.

Claiming to be definitely right and being right 98% of the time is pretty accurate. Still, it is not 100%, which is what definitely suggests.

Probably positive = 85/103 or 83% positive.

Equivocally positive = 24/38 or 63% positive.

Equivocally negative = 25/27 or 93% negative.

Probably negative = 170/174 or 98% negative.

Definitely negative = 1958/1962 or 99.8% negative.

Too often we will look at just the rare, but well publicized error. We ignore what may be an otherwise overwhelming history of accuracy. We are frequently far less accurate in our assessment than those we condemn.

If we do not measure what we are doing, we probably will be wildly inaccurate in our estimation of our performance.

Patient selection was performed without knowledge of subsequent clinical, surgical, or pathology findings after MDCT. No patient was excluded for suboptimal MDCT evaluation.[1]

The total number of equivocal (positive and negative) interpretations was 65.

That is 65/2871 or just over 2% of the total were equivocal interpretations and none were excluded due to low quality. In other words, there was no weasel factor.

It is noteworthy that the preoperative MDCT interpretation was true-negative in nearly one half of the clinically suspected cases in our series for which pathology findings were negative at appendectomy, outnumbering false-negative MDCT findings by 26 to 10. If surgery had been avoided in all 26 cases, the rate of negative findings at appendectomy would have decreased from about 8% to 4%.[1]

Paying attention to the interpretation would have cut the unnecessary surgery rate in half, but we do not know what the outcome would have been for the 10 patients with appendicitis, but negative interpretations.

In examining EMS interventions, we need to take assess our ability to identify what we are treating in a similarly thorough manner.

Do we use smaller doses when we are less confident in our assessments?

Do we reassess more frequently when we are less confident in our assessments?

Do we look for more definite indications when we are less confident in our assessments?

If not, why not?

Footnotes:

[1] Diagnostic performance of multidetector computed tomography for suspected acute appendicitis.
Pickhardt PJ, Lawrence EM, Pooler BD, Bruce RJ.
Ann Intern Med. 2011 Jun 21;154(12):789-96.
PMID: 21690593 [PubMed – in process]

Pickhardt PJ, Lawrence EM, Pooler BD, & Bruce RJ (2011). Diagnostic performance of multidetector computed tomography for suspected acute appendicitis. Annals of internal medicine, 154 (12), 789-96 PMID: 21690593

.