Over the next few installments of this series, we will be discussing “bounceback” studies, and answering the following questions, in sequence:
- What is the incidence of bouncebacks?
- What is the incidence of bounceback admissions?
- What is the incidence of deaths in patients recently discharged from the ED?
- What percent of bouncebacks occur because of medical errors?
How can we use this information to improve patient safety?
Our feeling is that if we can use these data to identify highrisk patients, we can assure that our urgent care evaluation was appropriate.
If we can identify patients who are more likely to bounce back, we can revisit their evaluation before they leave the urgent care center.
Question I: What is the incidence of bouncebacks?
Several studies have attempted to answer this question, using 72 hours as a bounceback “window” and producing strikingly similar results.
Though the data were gathered from emergency departments, they may also be applied to the urgent care setting.
- 1998, Annals of Emergency Medicine: Gordon, et al published a study of 52,553 ED visits during a 12-month period and found a return rate of 2.7%.
- 1992, Archives of Emergency Medicine: Wilkins and Beckett’s audit of 5,811 ED visits found 102 unscheduled returns, a rate of 1.9%.
- 1991, Archives of Emergency Medicine: O’Dwyer and Bodiwala published a study encompassing more than 8,000 ED visits; they found a bounceback rate of 2.9%.
- 1990, Annals of Emergency Medicine: Pierce, et al published a study of 17,214 visits and found a 3% bounceback rate.
So, the incidence of ED bouncebacks is felt to be roughly 3%; this translated into 3.3 million return visits in 2005 (of 115 million ED visits total).
Which patients are most likely to return?
Pierce found that 18% of bouncebacks were due to physician-related factors, and that 30% required hospitalization upon their return. Reasons for the bounceback visits included:
- treatment error
- admission indicated at initial visit
- psychiatric illness with admission indications
- radiology call-back
- no pain medication given
This month’s JUCM case reinforces several general risk management principles, primarily the kind of “misdiagnosis” cited by Pierce, above.
The patient is a 33-year-old man who presents with abdominal pain and is discharged without a definitive diagnosis. In this case, stronger documentation and timelier follow-up may have ensured a better outcome, decreased patient morbidity, and minimized the practitioner’s malpractice exposure in a clearly high-risk patient.
As with previous cases we have presented, this case illustrates the utility of our two-step approach:
- Identify high-risk patients (i.e., patients with a high-risk complaint and without a definitive diagnosis).
- Review the patients’ evaluations before they leave the urgent care clinic.
See how many “red flags” you can spot, and consider if you would have done anything differently.
A 33-Year-Old Male with Abdominal Pain
(Note: The following is the actual documentation of theproviders, including punctuation and spelling errors.)
CHIEF COMPLAINT (at 20:50): Abdominal pain
HISTORY OF PRESENT ILLNESS (at 21:06): He is a 33 year old male who states that at 7pm, after having normal BM, he developed gradual onset of RLQ and lower abdominal pain. He describes it as a bloating, spasm pain. After BM, pt noted urinary stream cut off and was no longer able to urinate. Gradually pain got worse, intermittent RLQ pain radiated into the right groin and testicles. Pain is 8/10 with spasms, mild nausea with pain. No dysuria, hematuria, testicular swelling, flank pain, chest pain, or fever/chills. Pt has no history of kidney stones but grandfather had many kidney stones.
PAST MEDICAL HISTORY/TRIAGE:
Medications: Claritin D
Allergies: No known allergies
EXAM (at 21:15):
General:Well-appearing; well-nourished; A&O X 3, in no apparent distress
Head: Normocephalic; atraumatic.
Nose: The nose is normal in appearance without rhinorrhea
Resp: Normal chest excursion with respiration; breath sounds clear and equal bilaterally; no wheezes, rhonchi, or rales
Card: Regular rhythm, without murmurs, rub or gallop
Abd: Non-distended; Tender RLQ but no rebound. Mild right flank/side pain. No rigidity, rebound or guarding
Skin: Normal for age and race; warm and dry; no apparent lesions
GU Exam: External genitalia normal, no urethral discharge, testes descended bilaterally. No lesions noted on penis or scrotum. Epididymus normal bilaterally.
ORDERS (at 21:25): Dilaudid 1 mg IVP, Toradol 30 mg IVP, Phenergan 12.5 mg IVP, .9NS-500cc bolus then 125cc/hr.
RESULTS (at 22:09):
Urine dip: WNL except: Bilirubin–1 mg/dL
Noncontrast helical CT of the abdomen/pelvis (at 22:29)–Unremarkable helical CT of the abdomen and pelvis.PROGRESS NOTE (at 23:23): Pt felt much better but still had pain into the lower abd. bilaterally with sitting up.
DIAGNOSIS: Abdominal pain, unspecified site, suspect bladder spasms.
DISPOSITION (at 23:41): The patient was discharged to Home ambulatory. Follow-up with primary care physician in 2 days. Prescription for Vicodin 5mg. Aftercare instructions for abdominal pain and kidney stone/renal colic.
Discussion of Documentation and Risk Management Issues at Initial Visit
Error: Failure to maintain a thorough differential diagnosis.
Intervention: The history seems to have led the physician down a ureteral calculus/spasm pathway. This was appropriate, given the patient’s symptoms. Although it is not uncommon to see ureterolithiasis without hematuria, a CT that fails to show a ureteral stone or hydronephrosis combined with a normal urine should suggest another cause for the pain, and should prompt the provider to move further down the differential. RLQ tenderness and pain radiating to the testicles may be due to acute appendicitis, incarcerated hernia or testicular torsion, yet it does not appear the practitioner considered these diagnoses. Perform additional evaluation (H&P and/or further testing) when things just don’t add up.
Teaching point: Start with a broad differential diagnosis when evaluating undifferentiated abdominal pain, focusing on high-risk/surgical diagnosis and use ancillary testing to hone in on your diagnosis. Understand the limitations of your tests.
Error: Failure to perform serial abdominal examinations.
Intervention: There is only one abdominal exam documented on this chart. Abdominal pain is a highrisk complaint and serial exams may discover an acute appendicitis or another surgical process that was not evident on initial assessment. Some clinicians seem to feel the best use for labs is that the patient spends more time in the ED/urgent care clinic, which allows his disease to progress to the point where it is easier to make an accurate diagnosis. A more responsible course of action, if it is not possible for the patient to wait in the urgent care clinic, is to send him to an ED.
Teaching point: The workup for abdominal pain is often a time-intensive process; use this to your advantage and be sure to perform and document serial exams.
Error: Failure to document medical decision making in a potentially high-risk patient.
Intervention: Not all high-risk complaints require a “million dollar workup,” but good documentation is essential. It is not clear from this chart which diagnoses were considered, as the progress note simply states “…still had pain into the lower abd. Bilaterally with sitting up.” It is wise to involve the patient and family in this discussion.
Teaching point: Document a progress note regarding medical decision making when dealing with highrisk patients, such as undifferentiated, ongoing abdominal pain.
Error: Failure to provide appropriate time for follow-up.
Intervention: An unremarkable helical CT scan of the abdomen makes appendicitis less likely, but does not exclude the diagnosis. The patient was instructed to follow up with his primary care physician in two days—a timeframe in which the appendix would be likely to rupture. In light of the fact that the patient was documented to have ongoing lower pain, and because appendicitis was still a distinct possibility, the patient should have been reexamined within eight to 12 hours. In such a case, if the patient can’t get in to see his primary care physician, then tell him to return to the urgent care for a repeat exam or to go to an ED.
Teaching point: If you are concerned about acute appendicitis, prompt repeat examination within eight to 12 hours (not two days) will improve patient safety and minimize your malpractice exposure.
33-Year-Old Male with Abdominal Pain Return Visit—Less Than 24 Hours Later
The patient returned 21 hours later with ongoing abdominal pain, now with associated vomiting and fever.
In addition, he was now tachycardic, appeared quite ill, and had right lower quadrant tenderness with a (+) Rovsing’s sign and guarding.
White blood count was 16K.
The patient was given meperidine and cefotetan and taken to the operating room, where he was found to have a retrocecal appendicitis, with rupture. He went on to develop a post-operative ileus and went home five days later without further complications.
Summary of Case and Risk-Management Principles
Our patient described RLQ pain and difficulty urinating, which ultimately led the practitioner down the kidney stone pathway. However, when the imaging did not demonstrate a urinary obstruction, the workup stopped. The patient had unexplained, ongoing abdominal pain—clearly a high-risk patient, warranting early and aggressive follow-up.
Unfortunately, it does not appear that appendicitis was in the differential because the patient was told to follow up in two days, a time span in which it was likely that the appendix would rupture.
If appendicitis is a consideration, the patient should be reexamined within eight to 12 hours.
Practitioners must develop a broad differential diagnosis for any chief complaint and then use the history, physical exam, and ancillary testing to rule in or out a specific disease. While there was a strong workup here for renal colic, the workup was aborted when the scan was negative. The practitioner would have been best served to go back, obtain additional history, perform a repeat exam, discuss the possibility of appendicitis with the patient, and develop a time-appropriate follow-up.
Application of our two-step approach may have improved patient outcome. The practitioner would have recognized this to be a high-risk patient due to ongoing abdominal pain without a definitive diagnosis.
He/she would have then obtained additional pertinent history, re-examined the patient, and documented serial abdominal exam. This would have enabled the practitioner to consider additional diagnoses and then document a thorough progress note and discuss the time-appropriate follow-up with the patient and his family.