Pain over speed bumps in diagnosis of acute appendicitis: diagnostic accuracy study
Pain over speed bumps in diagnosis of acute appendicitis: diagnostic accuracy study by Helen F Ashdown and colleagues (BMJ 2012;345:e8012)
Objective—To assess the diagnostic accuracy of pain on travelling over speed bumps for the diagnosis of acute appendicitis.
Design—Prospective questionnaire based diagnostic accuracy study.
Setting—Secondary care surgical assessment unit at a district general hospital in the UK.
Participants—101 patients aged 17-76 years referred to the on-call surgical team for assessment of possible appendicitis.
Main outcome measures—Sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios for pain over speed bumps in diagnosing appendicitis, with histological diagnosis of appendicitis as the reference standard.
Results—The analysis included 64 participants who had travelled over speed bumps on their journey to hospital. Of these, 34 had a confirmed histological diagnosis of appendicitis, 33 of whom reported increased pain over speed bumps. The sensitivity was 97% (95% confidence interval 85% to 100%), and the specificity was 30% (15% to 49%). The positive predictive value was 61% (47% to 74%), and the negative predictive value was 90% (56% to 100%). The likelihood ratios were 1.4 (1.1 to 1.8) for a positive test result and 0.1 (0.0 to 0.7) for a negative result. Speed bumps had a better sensitivity and negative likelihood ratio than did other clinical features assessed, including migration of pain and rebound tenderness.
Conclusions—Presence of pain while travelling over speed bumps was associated with an increased likelihood of acute appendicitis. As a diagnostic variable, it compared favourably with other features commonly used in clinical assessment. Asking about speed bumps might contribute to clinical assessment and could be useful in telephone assessment of patients.
Why do the study?
As noted in a BMJ Easily Missed article published in 2011, acute appendicitis can be difficult to diagnose, particularly when patients don’t present with the classic set of symptoms and signs. In fact, the clinical pattern of vague abdominal pain followed by vomiting with migration of the pain to the right iliac fossa, which was first described by Murphy, is present only in about half of patients. Over-zealous surgeons have been known to remove normal appendices, particularly since laparoscopic appendicectomy made this easier to do. Such an approach is less than ideal because surgery is not without hazards. The ideal scenario would be to operate on patients with genuine appendicitis and to leave the rest well alone. Tools can aid diagnosis: scores that sum positive findings from the history, examination, and inflammatory markers. Of these, the appendicitis inflammatory response (AIR) score seems to be the best discriminator. But wouldn’t it be great to find one foolproof clinical symptom or sign that could rule in or rule out the diagnosis of appendicitis?
What did the authors do?
The authors of this paper had, during the course of conversations with other clinicians, heard that some doctors routinely asked patients with possible appendicitis about their experience of pain on being driven over speed bumps on the way to hospital. They decided to investigate the predictive potential of recalled speed-bump pain as a diagnostic sign for acute appendicitis. They prospectively recruited patients over the age of 16 who presented to a single emergency department in Buckinghamshire, England, with signs and symptoms suggestive of acute appendicitis between February and August 2012. They tried to recruit consecutive patients identified by the examining doctor as having a likely diagnosis of appendicitis. This would reduce recruitment bias—that is, reduce the possibility that recruiting researchers would influence the study by choosing to include only those patients that they had identified, for example, or by recruiting patients only at certain times of the day or night.
For patients who agreed to take part, staff administered a questionnaire about their symptoms including four particular questions about their journey into hospital on that occasion. How did they get there? Did they travel over speed bumps on the way? Did they have pain while being driven to hospital? And did they notice having pain when they went over speed bumps? The authors then collected data on outcomes, including whether the patient was operated on, and if their appendix was removed, what the histological findings were. Histological diagnosis of appendicitis was regarded as the gold standard “positive” outcome. Interestingly, they assumed that patients whose symptoms resolved without surgery were appendicitis negative.
The authors used pilot data to calculate that they would need to recruit 100-150 participants to be able to show a likelihood ratio greater than 1.8-2.0 for speed-bump-pain positivity as a diagnostic sign. That means that they wanted to be able to show that the sign was around twice as likely to be positive in people with appendicitis as in people without the condition. They considered this to be a clinically useful difference.
They recruited 101 patients into the study. They calculated the sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios, with 95% confidence intervals, for the outcome diagnosis of appendicitis (see box). If patients were unsure if they’d been driven over speed bumps, or were unsure if their pain had intensified if they had been, they were considered as speed-bump-pain negative.
What did they find?
Of 101 recruited, 61 patients went for surgery and 54 lost their appendix. Only 43 of those appendices were found, on histological assessment, to be in a state of bona fide “itis”; the rest were deemed to be faking. 43/54 = 0.796 so the negative appendicectomy rate was 20%.
64 of 68 patients who had travelled over speed bumps (58 in a car and six in an ambulance) were included in the main analysis; the remaining four were excluded because of incomplete data (see table). 54 of 64 said they had experienced pain when riding over speed bumps (speed-bump positive).
|Pain over speed bumps||Appendicitis||Total|
Thirty four participants had a (histologically rubber stamped) diagnosis of appendicitis, 33 of whom were speed-bump positive and one of whom was speed-bump negative.
From these data, the authors were able to make various calculations which describe how useful the symptom (speed-bump pain) is diagnostically. The sensitivity of the symptom was 33/34=97% (95% confidence interval 85% to 100%) and the specificity was only 9/30=30% (95% CI 15% to 49%).
The positive predictive value is the probability that the patient who has a positive test really does have the condition tested for. The negative predictive value refers to the probability that the patient who tests negative doesn’t have the condition tested for. In this study, the positive predictive value of speed-bump pain was 61% (47% to 74%), and the negative predictive value was 90% (56% to 100%)—that is, a patient who recalled pain over speed bumps could perhaps turn out to have appendicitis (but then again they might not), whereas a patient who said they didn’t have pain over speed bumps was unlikely to turn out to have been a stoic who had raging appendicitis all along.
The likelihood ratio (LR), as defined by the Centre for Evidence Based Medicine, is the likelihood that a given test result would be expected in a patient with the target disorder compared to the likelihood that that same result would be expected in a patient without the target disorder. The likelihood ratios were 1.4 (1.1 to 1.8) for a positive test result and 0.1 (0.0 to 0.7) for a negative result—that is, the speed-bump sign was 1.4 times more likely to be positive in a patient with confirmed appendicitis than in a patient without appendicitis.
Calculations for investigating data in table
Sensitivity=(number of true positives)/(number of true positives + number of false negatives)
Specificity=(number of true negatives)/(number of true negatives + number of false positives)
Positive predictive value (PPV)=(number of true positives)/(number of true positives + number of false positives)
Negative predictive value (NPV)=(number of true negatives)/(number of true negatives + number of false negatives)
Positive test likelihood ratio=(sensitivity)/(1- specificity)
Negative test likelihood ratio=(1 – sensitivity)/(specificity)
These calculations represent three different ways of looking at the data in the two by two table. Each tells us something different about the usefulness of the test.
Because some patients were diagnosed with other serious abdominal pathology that required surgical treatment the authors were interested in seeing if the test was good for diagnosing any abdominal pathology that was important. So they performed a post hoc secondary analysis of the diagnostic accuracy of pain over speed bumps for the diagnosis of important abdominal pathology (including appendicitis) requiring treatment, which increased the sensitivity to 98% (87% to 100%) and the specificity to 39% (20% to 61%).
Is the test any good as a diagnostic marker for appendicitis?
It’s hard to say. Although the authors recruited more than 100 patients and therefore are just able to say they “met” their prespecified recruitment target, only 64 of 68 patients who remembered that they had travelled over a speed bump were included in the main analysis, which means that it could be argued that the study didn’t have enough participants to render the findings clinically meaningful. At the BMJ’s request the authors performed a sensitivity analysis in which all the 33 patients who did not remember having travelled over speed bumps were classed as being speed-bump negative, which allowed analysis using data from all 101 recruited patients. In this analysis the sensitivity decreased to 77% (61% to 88%) and the specificity increased to 61% (47% to 74%). The positive likelihood ratio was 2.0 (1.4 to 2.9) and the negative likelihood ratio 0.4 (0.2 to 0.7).
A simple way around this problem for future studies bent on replicating these methods would be to conduct the study in a hospital where all roads that approached the emergency department had speed bumps, so that patient recall of whether or not they had travelled over speed bumps would not be an issue. But then if asked specifically about being tossed by speed bumps some patients might report having experienced an increase in pain just because they were asked about it. One might need to ask a more neutral question like “Was your journey to hospital very uncomfortable?” which might allow patients to volunteer pain related to a bumpy journey without direct prompting. This might make the test more transferrable to other settings too.
Another limitation to this study was that patients who did not go to surgery but were treated conservatively with antibiotics were classed as not having appendicitis because they, of course, could not offer an appendix, inflamed or otherwise, for histological examination. As the authors note, we now know from meta-analyses of randomised trials that conservative management with antibiotics is an acceptable treatment of uncomplicated appendicitis. Future studies might need to consider using another gold standard for the diagnosis of appendicitis.
Further studies are probably needed before we can say for sure that experiencing pain while travelling over speed bumps is a truly reliable diagnostic symptom for acute appendicitis. However, the authors discuss how their findings compare with the diagnostic utility of other clinical symptoms and signs of the condition and find that it compares well with, for example, nausea and vomiting or rebound tenderness. Incidentally, this author has wondered, from her first reading of this study, if being driven over speed bumps isn’t just another way of eliciting good old rebound tenderness. Would I ask patients about speed-bump pain in future? I probably would ask about whether the journey to hospital had been uncomfortable and allow them to volunteer the information first, perhaps using the question as a way of distracting them as I test for rebound.Kirsten Patrick, assistant editor, BMJ
Correspondence to: email@example.com
Conflict of interest: KP declares no conflict of interest other than that she once failed (initially) to diagnose appendicitis in her mother (despite classic signs) because she had just come off a busy accident and emergency shift and expected to find Sunday lunch and not another patient waiting at her childhood home. The author’s husband, a non-medic, contributed to the error by reassuring his mother in law that the pain of appendicitis is always on the left side, which she believed.
Provenance and peer review: Commissioned; not externally peer reviewed.
- Lewis SR, Mahoney PJ, Simpson J. Easily missed? Appendicitis. BMJ 2011;343:d5976.
- Andersson M, Andersson RE. The appendicitis inflammatory response score: a tool for the diagnosis of acute appendicitis that outperforms the Alvarado score. World J Surg 2008;32:1843-9
- Centre for Evidence Based Medicine. Likelihood rations. www.cebm.net/index.aspx?o=1043.
- Varadhan KK, Neal KR, Lobo DN. Safety and efficacy of antibiotics compared with appendicectomy for treatment of uncomplicated acute appendicitis: meta-analysis of randomised controlled trials. BMJ 2012;344:e2156.
Cite this as: Student BMJ 2012;20:e8347