An Audit of Computed Tomography Usage for Abdominal Pain in a Singapore Emergency Department.

INTRODUCTION Computed tomography (CT) is increasingly used in the management of patients with nontraumatic abdominal pain at the emergency departments (EDs) in Singapore and Asia. While this is an established practice in many Western countries, the value of this approach remains unknown in most Asian public hospitals. As healthcare is often heavily subsidised in these hospitals, there are valid concerns about whether CT scan of abdomen and pelvis (CTAP) is a viable and financially sustainable strategy in the ED. We conducted a retrospective review of CTAP for nontraumatic abdominal pain performed in our ED with the following objectives: To examine the indications for ordering the ED CTAP To determine the rate of positive and negative ED CTAP To quantify the cost and impact on the length of stay (LOS), using acute appendicitis as a proxy, for ED CTAP To ascertain the utility of ED CTAP in assessing geriatric patients with abdominal pain. METHODS Our hospital is a public teaching hospital in Eastern Singapore with a daily average ED attendance of approximately 430 patients. Our ED charges a fixed attendance fee of SGD120 (USD1 = SGD1.33). In addition, the Ministry of Health reimburses the ED another SGD186 for each patient (subvention). The total amount of SGD306 per ED attendance is meant to cover the costs of laboratory tests, simple plain film radiological investigations, medications and treatment, but not CT scans in the ED. Instead, the costs of ED CT scans are to be borne by the patients. Patients who do not wish or are unable to bear the costs of ED CT scans may choose to be admitted, thus allowing their payments to be covered by private hospitalisation insurance or MediSave. Medisave is a government-mandated national medical savings scheme, in which individuals set aside between 8% and 10.5% of their income in a medical-focused savings account, which can then be used to pay for their personal or approved dependents’ hospitalisation, day surgery and certain outpatient expenses, and their healthcare needs in old age. An anonymised de-identified electronic database of all ED patients who had ED CTAPs done in 2020 was created. We excluded CTAP for trauma, CT aortograms and CT intravenous urograms, as these are dedicated scans indicated for specific conditions. Relevant data, including demographic information, indications for CTAP, CT scan reports, ED diagnoses, disposition from ED, costs and LOS in both ED and wards, were captured. A specialist emergency physician (EP) must approve the CT scans ordered in the ED. ED CTAPs were reviewed by radiologists on duty and their conclusions were considered the final diagnoses. For comparison of the costs and time taken to surgery for patients who had their CTAP done in the ED versus in the wards, we reviewed patients with a diagnosis of acute appendicitis, as this is a common indication for CTAP. The time to surgery (starting from ward admission), gross bill sizes and the LOS were recorded for both populations. We used the Mann–Whitney U test for comparison analysis. We also compared our data with previous overseas papers evaluating similar ED utilisation of CTAPs. These included reviews on the appropriateness of utilisation, negative rates of scanning and usage in the elderly population. A waiver of consent for this study was approved by the institutional ethics review committee. RESULTS In 2020, 1,860 patients (56% male) underwent CTAP in the ED for nontraumatic abdominal pain. The mean age of the patients was 50 years (range 14–99), with 365 (19.6%) patients aged >70 years and 9 (0.5%) patients aged 14–16 years. The mean LOS in the ED was 5.2 h (range 1.37–26.83 h). When we analysed the indications for ordering a CTAP [Table 1], the top six indicators were right upper and lower quadrant pains, flank pains, left lower quadrant pain, persistent abdominal pain despite observation after analgesia, and suspicion for intestinal obstruction (clinical diagnosis with or without plain abdominal radiographs).Table 1: Indications for CTAP and the common diagnoses and disposition outcomes in ED patients. (N=1,860)We had a positive CT scan rate of 83.8% (1,558 out of 1,860 cases) with abnormal findings to explain the symptoms. The top ten ED diagnoses after CTAP for all ages are listed in Table 1. Acute appendicitis, biliary tract disease, renal stones, ovarian disease and bowel disease were the common conditions in the ED. About 16% of the CT scans revealed no abnormality to account for the symptoms. Indications for these cases were either nonspecific abdominal pain or sepsis of unidentified source with abdominal symptoms. When we analysed the geriatric patient group (≥70 years), the spectrum was similar, but 15 unsuspected basal pneumonia cases were diagnosed after being seen on the upper sections of the CTAP. Malignancies were less common diagnoses at 3% [Table 1], with bowel, gynaecological and hepatobiliary malignancies being the most prevalent. Almost 70% of the patients were admitted to inpatient wards. We referred 6.5% to a partner maternity hospital emergency clinic for urgent gynaecological conditions, thus avoiding 121 unnecessary admissions to our surgery department. We discharged 3.4% of patients who had a negative CTAP, and they neither reattended the ED within 72 h nor suffered any adverse outcomes. We could confirm some medical causes of abdominal pain only after a negative CTAP scan. Confirmed examples included 13 patients with basal pneumonia, eight patients with diabetic ketoacidosis, eight patients with mesenteric lymphadenitis and one patient with dengue [Table 1]. In Table 2, the time to CTAP, surgery, LOS (calculated from the time of ward admission) and the gross bill sizes for patients admitted with appendicitis are shown. Patients who underwent ED CTAPs had significantly faster time to a scan and surgery, lower bill sizes and shorter LOS. This was after adjusting (by exclusion) for 18 patients with confounding comorbidities (dementia, myocardial infarction, heart failure, renal failure, chronic pulmonary disease) that led to prolonged stays.Table 2: Time to computed tomography (CT) and surgery, bill size and length of stay for appendicitis patients.DISCUSSION Our ED installed a CT scanner initially in the late 2000s for assessing polytrauma and stroke patients. By the early 2010s, CTAPs were performed for patients with nontraumatic abdominal pain. This strategy was in response to access block from increasing demands for beds in our hospital. It was also to arrive at a faster diagnosis to avoid complications from acute surgical conditions. Initially, the costs of performing CTAPs in the ED were absorbed by the hospital if the patient was not admitted. Under a fixed fee of SGD306 for each emergency attendance, this meant a substantial loss for the hospital as each CTAP costs about SGD700 (standard nonsubvention rate). After 2015, patients had to pay for ED CTAPs. It was decided that the cost would be at a cost recovery price of SGD350 (priced at 50% less than the standard SGD700). Should the CTAP show a condition that necessitated emergency admission, the cost would be included in the inpatient bill instead of the ED charges. Patients could then choose to pay part or all of their hospitalisation bills with their own savings, private insurance or from MediSave. The average bill size for a one-day stay under surgery for subsidised wards is SGD1,215. This includes all cost drivers, procedure fees, operating theatre costs, medications, ward charges and others presubsidy. As such, this is still more expensive than the total cost of SGD656 for an ED consultation and CTAP scan. Our high positive CT scan rate of 83.8% allowed for timely definitive treatment after confirmation of diagnoses based on CT findings. Publications from Western and Japanese settings[1-4] have shown that CT for abdominal pain changes the leading diagnosis, increases diagnostic certainty and facilitates management decisions. This is so even for unstable patients after adequate resuscitation,[5,6] and the old description of the CT scan as being the ‘tunnel of death’ no longer applies as modern machines are faster with higher resolutions. Using appendicitis as a proxy, our results showed benefits for patients who underwent CT scans in the ED. Our department uses the Alvarado score to help define indications for ED CTAP in the right lower quadrant as part of a suspected appendicitis protocol. We had a shorter median LOS of 0.3 days, which was statistically significant when analysed by the Mann–Whitney test. For comparison, Sala et al.[3] found in their study that the average hospital stay was almost one day (22 h) shorter for patients in the CT group than for those in the control group, but this was not statistically significant. There are valid arguments that performing CT scans in the ED could prolong transit time through the ED, causing choke points in the ED. The counterargument is that with more access block, usage of CT scans could reduce unnecessary admissions and free up hospital beds that are in high demand. In our hospital, where the daily average bed occupancy often hovers above 95%, we favour the latter argument. Before the ED was equipped with the CT scanner, our EPs would base the decision for admission almost entirely on clinical grounds. For example, if a patient presents to the ED with fever, vomiting, anorexia and right lower quadrant rebound tenderness, with blood investigations showing an elevated white cell count that results in an Alvarado score of 8, most senior EPs would decide to admit the patient based on the rationale that even if the final diagnosis is not appendicitis, the condition will still likely warrant admission (e.g., diverticulitis, perforated colonic cancer, pyelonephritis, etc.). With the advent of ED CT scanning, the aforementioned approach to admit the patient based on the clinical status of the patient largely remains, even before the CT scan is fully reported or performed, so as to expedite the admission process. An urgent consultation or admission can then be obtained if the CT scan demonstrates a condition requiring urgent surgical intervention. A negative scan, on the other hand, may allow the patient to be discharged safely. This helps to mitigate the CTAP becoming a choke point for patients flowing through the ED. A literature review revealed that we were not alone in using this approach. In Toronto, Wang et al.[7] described three similar patterns of ED dispositions to improve admission efficiency: Disposition after the initial imaging report: the most common pattern where CTAP is performed and interpreted before the disposition decision (83% of their patients) Disposition before report: where a disposition decision is made before the availability of the first radiology report but after the scan has been performed Disposition before CT: where the disposition decision is made before the start of the CTAP (e.g., when the plain films show obstructed bowel, resulting in a decision to admit before a CTAP is ordered) With adoption of patterns B (7%) and C (6%), the Toronto team found that the ED LOS for pattern A (mean 10.4 h) was statistically significantly longer than those for pattern B (mean 8.1 h) and pattern C (mean 6.9 h). Unfortunately, our anonymised database did not capture a corresponding set of data based on the three patterns. As such, we could not make a similar comparison. Nevertheless, the overall mean ED LOS for our patients was 5.2 h. In a survey by Kirsch et al.[8] in an American setting, the adult CTAP utilisation rate ranged from 11.3% at age 20–29 years to 24.6% for those over 65 years. This was similar to our study findings — 305 (16.4%) scans were performed in patients aged 20–29 years and 478 (25.6%) in those over 65 years. In an evidence-based guideline developed by Gans et al.[9] (in a multispecialty Dutch collaboration), the decision for CT scans had to be complemented with good history taking and physical findings, supplemented with relevant laboratory investigations and ultrasound, to have high CTAP sensitivity and specificity. With a positive CT scan rate of 83.8% in our study population, we may now use this as a reference to monitor the trends in the usage of CT scans and avoid misuse. At the Mayo Clinic (an academic tertiary care setting), Bellolio et al.[10] performed a cross-sectional study of all ED visits that included CT scans between 2003 and 2012. Overall, CTAP use per 1,000 ED visits was between 55.4 in 2003 and 75.62 in 2012 (P < 0.001). In comparison, we had 15.4 CTAPs per 1,000 ED visits for 2020. Conversely, Choy and Yoon[11] conducted a study to predict negative scans in 300 ED patients aged ≤60 years who underwent ED CTAP. Their retrospective model predicted that a patient with normal white blood cell count and normal abdominal exam is 70.8% less likely to have a significant positive finding. Older patients often present with vague symptoms, unreliable physical findings and laboratory values may be altered by chronic organ disease. The morbidity and mortality associated with elderly abdominal pathological conditions are more significant.[12,13] We should, therefore, have a lower threshold for scanning. Our positive rate was 19.6% in this age group, which was lower than in American and European studies. Gardner et al.[12], in an American retrospective study of 464 patients (>80 years), found that CTAPs were positive in 55% of the patients, even with a disease spectrum similar to ours in the top ten conditions. Also, 43% of their diagnoses were clinically unsuspected before CT and had a significant influence on the clinical management and disposition, such as appropriate surgical admission and surgery. Millet et al.,[13] reported 30.3% acute unsuspected pathologies in a European cohort. In our study, 16.2% of CTAPs revealed no abnormality to account for the symptoms (259 patients with normal findings and 43 patients with sepsis and abdominal symptoms but normal CTAP). A negative CTAP is reassuring to both patients and EPs. Medical causes of acute abdominal pain (e.g., diabetic ketoacidosis, dengue) were also confirmed after a surgical abdomen was safely excluded. When we looked at the literature, we could not find many suitable similar retrospective studies for comparing the negative scan rates for CTAPs of the entire abdomen. One comparable study by Laméris et al.[14] had 183 out of 1,021 patients (18%) with a negative scan. Pickhardt and Nelson[15] had a negative CT diagnosis in 52.7% of all patients. As such, we opine that our rate was acceptable compared to international data. We wish to emphasise that clinical judgement is guided by validated scoring systems, in conjunction with testing and/or preliminary imaging. For instance, Stengel et al.[16] reported a high negative CT rate of 77% for acute appendicitis in 2,283 patients in a retrospective study, without mention of Alvarado scoring or basis for clinical decision. This study is not without limitations. Working with a retrospective de-identified database without reviewing patients’ medical records, we were not able to evaluate other factors that could affect time to CTAP or ED LOS. These factors could include delayed development of physical signs, atypical presentations in the early stages, and patients’ consideration of financial options, possible contrast allergy or the risk of contrast-induced nephropathy, and others. The database also did not capture the various time stamps required for the analysis of patients’ transit time through the ED and inpatient wards. These include time of registration, time of triage, time and duration of consultation with the EP, time of admission and duration to initial assessment by the inpatient physician, and waiting time between the patient’s consent for surgery and the actual time of surgery. In conclusion, we should allow the funding model for ED patients to accommodate the usage of ED CTAP scans in the diagnostic workup for nontraumatic abdominal pain, and encourage the availability of other resources necessary for ED CTAPs to remain accessible. This would facilitate timely definitive management and appropriate disposition of patients from the ED. Utilisation of CTAPs should be guided by clinical suspicion and protocols with oversight by EPs to improve their accuracy and avoid indiscriminate ordering. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.