Guest Skeptics: Dr. Anthony Crocco is a Pediatric Emergency Physician and is the Medical Director & Division Head of the Division of Pediatric Emergency at McMaster’s Children’s Hospital. He is the creator of SketchyEBM.
Case: A 15-year-old female patient presents to your emergency department with a chief compliant of abdominal pain. A medical student picks up the chart and comes back to tell you about the case. His presentation includes that the abdominal pain has been going on for two days, has worsened over this time period, and is worse in the right lower quadrant from previously being around her umbilicus. She had a temperature this morning of 101 degrees Fahrenheit, vomited twice this morning (non-billious, non-bloody), says the bumps in the road travelling to your hospital hurt badly, and she doesn’t want to eat anything. The medical student also tells you that she doesn’t have dysuria, discharge, diarrhea, or constipation. She just finished her menstrual cycle and is not sexually active. The medical student tells you that he noticed pain over McBurney’s point. He wants to CT scan her. You start to tell him about trying to limit radiation in pediatric patients and the Pediatric Appendicitis Score (PAS) but can’t help to wonder, is this clinical prediction rule as helpful in this female adolescent as other pediatric patients?
Background: Abdominal pain is a common complaint in any emergency department, as is true in a pediatric emergency department. Appendicitis is usually on the differential in a pediatric patients presenting with abdominal pain and is a common surgical emergency.
Female adolescents pose a unique diagnostic dilemma due to having competing gynecologic diagnoses, such as urinary tract infections, sexually transmitted infections or ovarian cysts or torsion. Being accurate about diagnosis of appendicitis is important to avoid complications of missed appendicitis as well as complications of negative appendectomies.
The PAS is a clinical prediction tool which uses elements of the history and physical examination such as symptoms of right lower quadrant pain and fever and combines this with laboratory findings such as the white blood cell count (WBC) to predict the risk of acute appendicitis.
The PAS score was initially developed with a single cutoff, but validation studies showed better performance with two cutoffs: one at the low end to identify patients with a low risk of appendicitis who may not need further evaluation, and one at the high end to identify patients with a high risk of acute appendicitis on clinical grounds alone.
With a possible score of ten, scores at the lower cutoff, such as less than or equal to two are used to discharge a patient home without further work up due to low suspicion for appendicitis, thus making sensitivity important at this cutoff. Scores at the higher end, such as greater than or equal to seven or eight are used to predict high suspicion for appendicitis, thus making specificity important at this cutoff.
There are not any studies to our knowledge that compare the PAS performance at both different ages and genders which is an important consideration in caring for pediatric patients.
Clinical Question: How well does the Pediatric Appendicitis Score (PAS) perform for adolescent female patients?
Reference: Scheller et al. Utility of Pediatric Appendicitis Score in Female Adolescent Patients. This is Hot off the Press in AEM May 2016
Population: Female patients 13-21 years old presenting with symptoms suggestive of appendicitis
Intervention: Use of the Pediatric Appendicitis Score
Comparison: Pathology proven appendicitis
Outcome: Diagnostic metrics (Sensitivity, Specificity, PPV, NPV) including comparison of these metrics to All other patients presenting with symptoms of appendicitis
Author’s Conclusions:“Our study demonstrates that the PAS score, as commonly used clinically (i.e. with cutoffs of >=3 and >=8), showed better specificity and equivalent sensitivity for female adolescent patients compared to all other patients, as well as a good NPV in both groups.”
Quality Checklist for Randomized Clinical Trials:
The clinical problem is well defined. Yes
The study population represents the target population that would normally be tested for the condition including (i.e. no spectrum bias). Yes
The study population included or focused on those in the ED. Yes
The study patients were recruited consecutively (i.e. no selection bias). Unsure
The diagnostic evaluation was sufficiently comprehensive and applied equally to all patients (i.e. no evidence of verification bias). Yes
All diagnostic criteria were explicit, valid and reproducible (i.e. no incorporation bias). Yes
The reference standard was appropriate (i.e. no imperfect gold-standard bias). Yes
All undiagnosed patients underwent sufficiently long and comprehensive follow-up (i.e. non double gold-standard bias). Unsure
The likelihood ratio(s) of the test(s) in question is presented or can be calculated from the information provided. Yes. Not presented but can be calculated.
The precision of the measure of diagnostic performance is satisfactory. Yes
Key Results: n=901 with 28% pathology-proven appendicitis. Of the total population enrolled in the study, 30% were adolescent females (age 13-21).
There was no significant difference in the sensitivity for the cutoff value of three or the specificity for the cutoff value of seven.
The specificity of the cutoff value of eight was significantly better in the adolescent female group compared to the pediatric non-adolescent female group.
Dr. RoseAnn Scheller
Lead author of this study was Dr. RoseAnn Scheller. RoseAnn is a pediatric emergency physician and assistant professor at the Children’s Mercy Hospital and Clinics in Kansas City, MO. Her career goals are to improve the quality of emergency care of pediatric patients through improvements on resuscitation and trauma, and research in pediatric emergency management, resuscitation, and violence prevention. Her study was done at the Cincinnati Children’s Hospital Medical Centre. We asked RoseAnn five questions and her responses are in italic.
Sub-group Analysis: This is a retrospective cohort study, effectively a sub-group analysis of a previously done study on the PAS. Was this study adequately powered for the diagnostic metrics you were investigating?
As we stated in our limitation,s it was possible that the sample size in regards to age and sex may have not been sufficient. We did do a prospective power calculation for this study before complete enrolment of the parent study. Using the expected prevalence estimates we had at the time, we determined that 222 female adolescents would allow us to detect a difference in the areas under the ROC curves for the two groups. Our study included 250 adolescent females.
Consecutive or Convenience: The manuscript says the clinical research coordinators identified a convenience sample of patients. Does this mean it was not consecutive patients presenting with abdominal pain suspected of appendicitis? If not consecutive, how do you think this may have impacted the results?
Research assistants were present 12-18 hours/day and enrolled study patients consecutively during these times. Because of this convenience sample, our results may not have clear generalizability. There may be sample bias and sampling error due to the nature of convenience sampling. It is possible that patients that present during times that the research assistants were not present may be different than patients that present other times of day.
Clinical Appendicitis Pathway: Physicians were given a new clinical appendicitis pathway prior to the start of the study. This diagnostic pathway, which included the PAS, was added to your electronic medical record halfway through the study period. What impact, if any, do you think that had on your results?
I think that having the clinical pathway and computerized decision support tool introduced may have increased awareness and may have influenced the physicians consideration and identification of symptoms as well as the overall score.
Incomplete Follow-up: You defined “negative appendicitis” to include all patients who did not undergo appendectomy one month after emergency department presentation. What about patients who sought care at another hospital or after one month?
We felt that it would be likely that a pediatric patient with appendicitis would follow up at our facility because it is the only hospital with pediatric surgeons in our catchment area. One month was chosen due to the probability that a case of appendicitis from an ED visit would likely present within one month of initial presentation allowing for full capture of data even with a delayed presentation or antibiotic therapy. However this is a potential limitation that we identified in the article.
PPV and NPV: These are dependent on prevalence of disease. The pathology proven appendicitis in the female adolescent cohort was 16% while in the comparison group it was 33%. While this difference does not impact the PPV and NPV that much, did you think about calculating the likelihood ratios, which are not dependent on prevalence of disease?
We had considered likelihood ratios however we wanted our study to be comparable to other studies of the PAS, which reported PPVs and NPVs more often. Additionally, we do report the primary outcomes of sensitivity and specificity, which are not dependent on prevalence and can be used to calculate the likelihood ratios
Is there anything else about the project you would like to mention?
I do think this difference in prevalence between the groups is interesting and would like to see research in the future address this discrepancy. One possibility is that adolescent females are more likely to have other abdominal pathology, which can mimic appendicitis, such as urinary tract infections, sexually transmitted infections or ovarian pathology, so clinicians suspect it, and test for it more. Since our study population was based on clinician concern for appendicitis, it is possible that we had high number of adolescent females with other final diagnoses, thus decreasing the prevalence in this group compared to the “all patients” group.
Comment on author’s conclusion compared to SGEM Conclusion: We agree with the authors’ conclusions. It appears that the PAS is at least as good a tool in adolescent female patients as it is for the rest of the pediatric population.
Clinicians have the option of using the PAS or the Alvarado score to help risk-stratify children with abdominal pain with regards to appendicitis risk. Both scores require laboratory investigations, specifically a CBC. Neither of these tools is perfect, however, and both should be used with caution. For patients with low risk who are being sent home, adequate discharge planning is paramount. For patients with intermediate risk, clinicians should adopt a strategy of investigation that minimizes radiation exposure.
SGEM Bottom Line: The PAS has similar utility in adolescent females patients compared to other pediatric patients.
Case Resolution: You decide to use the PAS and find that your patient’s score is eight. You quickly evaluate the patient and agree that her exam is concerning for appendicitis. According to your hospital’s pathway, you call surgery, they also evaluate the patient and have a concern for appendicitis as well. They decide to admit the patient and take her to the operating room without imaging. Her appendix is taken out and is found to be consistent with appendicitis on the pathology report.
Clinically Application: In children presenting to the emergency department with abdominal pain, where there is clinician concern for acute appendicitis, the PAS can be used, even if the patient is an adolescent female. Care must be taken, as the PAS, as well as any other scoring systems, are imperfect decision tools.
What do I tell my patient? We are going to use a score called the Pediatric Appendicitis Score to help us understand your risk of having appendicitis, to guide our testing and treatment.
Keener Kontest: The last winner was Marc Sonntag who knew the FDA approval for tamsolusin is for the treatment of benign prostatic hypertrophy.
Listen to the podcast for this weeks’ question. If you know the answer send it to TheSGEM@gmail.com with “keener” in the subject line. The first correct answer will receive a cool skeptical prize.
Dr. Jeff Kline
SGEM Hot Off the Press: Hello SGEMers, my name is Dr. Jeff Kline and I am the Editor-In-Chief of AEM. We are really interested in engaging EM community and find out what you think about this SGEMHOP episode? What questions do you have for Dr. Scheller and her team on the Pediatric Appendicitis Score? Join the conversation on Twitter (#SGEMHOP), Facebook or the SGEM blog. The best social media feedback will be published in a future edition of AEM.
Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.