| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Health Policy and Practice |
1 From the Depts of Radiology (R.H.G., J.L.S., S.L.V., L.S., R.E., D.D., T.C.L., I.M.), Urology (E.N.E., J.V.J., E.M.M.), and Neurology (R.G.H.); Emergency Dept (F.L.Z.); and Clin Research Ctr (W.J.D.), Univ of Rochester Medical Ctr, 601 Elmwood Ave, Box 648, Rochester, NY 14642; and Dept of Radiology, Hosp Univ of Pennsylvania, Philadelphia (A.T.T.). Received Feb 13, 2002; revision requested Apr 5; revision received Apr 22; accepted May 29. Supported in part by Div of Research Resources-GCRC grant 5M01 RR-00044 and Medical Ctr Insurance of Vermont. Address correspondence to R.H.G. (e-mail: ronald_gottlieb@urmc.rochester.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Hospital clinical and radiology information systems were used to retrospectively identify patients presenting with UTC symptoms from January to December 1997 (before introduction of unenhanced CT) and from January to December 1999 (after introduction of unenhanced CT). Chart abstraction was used to confirm the identification of patients with presenting symptoms suggestive of UTC and assess patient outcomes. Two hundred sixty-five patients presented before (1997) and 602 after (1999) unenhanced CT was introduced. Distributions of dichotomous variables were compared between the 1997 and 1999 groups by using logistic regression. Means were compared between the groups by using analysis of variance and mean total numbers of imaging studies by using Poisson regression.
RESULTS: Total number of imaging studies increased by 26.7% per patient visit (P < .001). Rates of admission following the initial ED visit (13.7% in 1997 vs 13.4% in 1999), as well as percentage of patients who subsequently returned to the ED (12.0% in 1997 vs 13.7% in 1999) or subsequently were admitted to the hospital (4.5% in 1997 vs 5.3% in 1999) in the month following the initial ED visit, were similar between the two groups. Unsuspected unenhanced CT findings that could affect acute patient care were observed at 5.9% of examinations.
CONCLUSION: Use of imaging for suspected UTC has increased markedly since the introduction of unenhanced CT, with little effect on acute care of patients in the ED.
© RSNA, 2002
Index terms: Computed tomography (CT), clinical effectiveness • Computed tomography (CT), utilization • Kidney, calculi, 81.811 • Kidney, CT, 81.12111, 81.12112, 81.12115 • Radiology and radiologists, socioeconomic issues
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
To our knowledge, in the literature minimal attention has been devoted to the potential negative effects of using CT unrestrictedly to examine patients with suspected stone disease. The radiation dose per examination for unenhanced CT has been reported to be two to three times higher than that with IVU (6,7). This could be a concern if unenhanced CT was used substantially more in the general population, particularly for relatively young patients, including women of child-bearing age. As with all diagnostic examinations, false-positive diagnoses are possible and can lead to additional unnecessary procedures being performed and needless patient anxiety.
The goals of this study were to determine the change in examination patterns and effectiveness of care since the introduction of unenhanced CT for the examination of patients who present to the University of Rochester Medical Center emergency department with symptoms suggestive of urinary tract calculi (UTC).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The study was performed at the University of Rochester Medical Center, a 679-bed urban teaching hospital with an emergency department to which there were 55,356 patient visits during the 1997 calendar year and 68,086 visits during the 1999 calendar year. The radiology and emergency medicine departments have residency programs in which all final decisions regarding examination ordering and finding interpretation are made following attending physician review.
The study design was that of a retrospective evaluation in which data on the patients who presented to the emergency department with symptoms suggestive of UTC from January 1 to December 31, 1997, before the introduction of unenhanced CT at our institution, were compared with those on the patients who presented to the emergency department with symptoms suggestive of UTC from January 1 to December 31, 1999, after unenhanced CT had become firmly established as the diagnostic imaging examination of choice in this setting. The hospital information system was used to identify patients who presented to the emergency department with symptom codes for blood in the urine or kidney stone, as well as patients who were discharged from the emergency department with a diagnosis of UTC with one of the following International Classification of Disease codes: 592.0, 592.1, 592.9, 593.4, 594.9, 596.0, 599.6, or 788.0 (8). The hospital information system contains data regarding emergency department visits (eg, presenting symptoms and diagnosis at discharge), hospital admissions, surgical procedures, pathology reports, and outpatient clinic visits.
We also used the radiology information system (IDXrad; IDX Systems, Burlington, Vt) to identify all patients who presented to the emergency department and underwent either unenhanced CT (1999 only) or IVU (1997 and 1999). The medical records of all patients who were identified by using the hospital and radiology information systems were reviewed to confirm that the patients presented with symptoms suggestive of UTC. Three chart abstractors (J.L.S., R.E., A.T.T.) were specifically trained to collect relevant patient data and record them on case collection forms designed for this purpose and were actively supervised by the principal investigator (R.H.G.).
Four hundred thirty-three emergency department visits for the 1997 calendar year and 950 for the 1999 calendar year were identified by using the hospital and radiology information systems. Review of the medical records resulted in the exclusion of 107 visits from 1997 and 221 visits from 1999 owing to presenting clinical symptoms that were not suggestive of UTC. In addition, medical records could not be obtained for 35 1997 visits and 70 1999 visits. The resulting patient population included 265 patients (102 female; 163 male; mean age, 42.2 years; age range, 8–87 years) who made 291 visits during 1997 and 602 patients (259 female, 343 male; mean age, 39.9 years; age range, 8–87 years) who made 659 visits during 1999.
We used the medical records to document the presenting clinical symptoms and laboratory findings, which included abdominal or flank pain, elevated white blood cell count, and fever. We also determined the patients’ comorbid conditions, including whether they had experienced a prior UTC episode (Table 1).
|
Outcome Measures
Emergency department visits separated by at least 30 days were considered to be separate episodes of care, beginning with the first visit of each year. The following outcome measures were abstracted from the medical records as measures of the effectiveness of patient care to be compared between the 1997 and 1999 patient groups:
1. Admission to the hospital (with associated diagnoses at discharge) immediately following the initial emergency department visit and length of hospital stay.
2. Admission to the hospital (with associated diagnoses at discharge) within 30 days following the initial discharge from the emergency department, for a medical condition potentially related to the initial presenting symptoms.
3. Return visits to the emergency department (with associated diagnoses at discharge) within 30 days for a medical condition potentially related to the initial presenting symptoms.
4. Length of stay in the emergency department at the time the patient initially presented and length of stay for any immediate hospital admissions following the initial emergency department visit.
5. Urologic procedures performed within 30 days following the initial emergency department visit, as well as the findings and complications of these procedures.
Only data from our institution were available for review, so we were unable to assess the emergency department and hospital data for other regional and state institutions.
Imaging Patterns and Findings
We used the radiology information system to determine the numbers and types of imaging examinations that patients underwent at the time of their initial emergency department visit. These examinations included abdominal ultrasonography (US), unenhanced CT, conventional abdominal CT (including the unenhanced CT examinations that were changed to conventional CT examinations), IVU, and abdominal radiography. Unenhanced CT examinations were performed without orally or intravenously administered contrast material with helical scanners (CTi; GE Medical Systems, Milwaukee, Wis) by using 5-mm collimation and reconstruction interval with a 1:1 pitch. Imaging coverage was from the upper poles of the kidneys through the lower part of the pelvis. We considered conventional CT all other CT scans of the abdomen, pelvis, or both, for suspected abdominal disease other than those using our unenhanced CT protocol for suspected UTC. IVU was performed by using low-osmolar intravenous contrast material (iohexol, Omnipaque 300; Nycomed, Princeton, NJ) at routine nephrotomography of the kidneys.
The chart abstractors used the radiology information system to record the findings of all imaging examinations, including recommendations based on the incidental findings for follow-up imaging. The biopsies and surgeries performed as a result of findings at imaging performed at the time of the initial emergency department visit were documented. The chart abstractors used the hospital information system to evaluate the surgery reports and pathology results from these procedures. The subsequent clinical and procedural notes from the hospital information system also were reviewed by the chart abstractors. The chart abstractors used urology billing records, clinical notes, and the hospital information system to determine which patients had undergone subsequent urologic procedures and to assess the complications from these procedures. The chart abstractors identified all potential abnormalities and associated factors (eg, return visits for abdominal pain, surgical results, and diagnoses other than stone disease). The relevance of all findings with regard to patient care was determined at consensus between the principal investigator and the chart abstractor who documented the given finding. The effect of unenhanced CT findings on patient care in the emergency department subsequently was evaluated.
Statistical Analyses
The distributions of dichotomous variables—namely, return emergency department visits, hospital admissions, and treatment procedures—in the 1997 and 1999 treatment groups were compared by using logistic regression. An individual patient could have visited the emergency department more than once in any given year. To account for the correlations among the dependent responses within a patient, the model parameters and standard errors were estimated by using the generalized estimating equations technique (9). This technique enables one to calculate estimates of standard errors that vary little owing to incorrect specification of the within-subject correlation structure, which was assumed to be exchangeable in our analysis (ie, equal correlations among pairs of responses within a patient).
The mean values for the groups were compared by using analysis of variance, and the mean total numbers of imaging studies were compared between the groups by using Poisson regression. Again, the generalized estimating equations technique was used to account for the correlation among the dependent responses within a patient. A P value of less than .05 was considered to indicate a significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
Regarding those patients who presented with symptoms suggestive of UTC but may not have been identified, our random search revealed that we missed only a small percentage (n = 12 [3.6%]) of these patients.
Imaging and Clinical Findings
Unenhanced CT depicted UTC in 317 (61.9%) of 512 examinations. Findings that corresponded to what were considered to be acute processes (other than UTC) that potentially were responsible for the patients’ symptoms were identified at 30 (5.9%) of 512 unenhanced CT examinations (Table 4). Acute inflammatory conditions were identified in 19 patients at unenhanced CT. These conditions included diverticulitis (n = 7), unspecified colitis (n = 4), appendicitis (n = 3), pelvic inflammatory disease (n = 3), cholecystitis (n = 1), and small-bowel enteritis (n = 1). A fever or elevated white blood cell count in 14 (73.7%) of these patients rendered them at risk of having an abnormality other than uncomplicated urinary tract stone disease.
|
In nine patients, masses in the abdomen or pelvis, which were considered to be responsible for the symptoms, were identified at unenhanced CT. These lesions included six complex adnexal masses, one renal mass, one vertebral body mass, and one adrenal mass. Two adnexal masses—a pelvic peritoneal inclusion cyst and a benign ovarian cyst—were resected at elective surgery. Four patients with adnexal masses were discharged from the emergency department with no follow-up treatment. One of these patients returned to undergo US, which depicted an ovarian cyst. In a patient with known lung cancer, unenhanced CT depicted a renal mass that was confirmed to be a metastasis at subsequent biopsy, a vertebral body mass that also was confirmed to be a metastatic lesion at biopsy, and an adrenal mass that was found to be a residual adrenal hematoma at surgical resection.
Other abnormalities that were considered at unenhanced CT to be responsible for the patients’ acute symptoms included a transverse process fracture and a choledochocele. The patient with the transverse process fracture was given a diagnosis of having recently passed a urinary tract calculus, despite unenhanced CT findings, and was discharged from the emergency department. Subsequent magnetic resonance imaging findings failed to confirm a transverse process fracture but showed a herniated lumbar disk to be responsible for the patient’s symptoms. The choledochocele in the other patient was identified at US that had been performed 1 week earlier at our institution.
Incidental findings were considered to be possible abnormalities depicted at unenhanced CT that were unrelated to the patient’s symptoms but were potential motivations to perform additional imaging or other procedures (Table 5). Incidental findings were observed at 93 (18.2%) of 512 unenhanced CT examinations. The majority of these were findings such as lung nodules, liver lesions, gallstones, small renal lesions, and pelvic lesions. The pelvic lesions included four dermoid tumors. Many of the patients with incidental findings had them in association with obstructing UTC, which were responsible for their symptoms. Most of these findings were not followed up with imaging, biopsy, or surgery at our institution with follow-up out to at least 18 months. Only eight (8.6%) of the 93 patients with incidental findings underwent subsequent imaging. None of these examinations revealed an abnormality that would prompt an alteration in patient care (Table 5).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The use of CT in the emergency department to diagnose intraabdominal abnormalities has increased substantially. Unenhanced CT is a relatively new CT technique (14) in the detection of UTC that can be performed instead of conventional IVU and thus spare patients the intravenous injection of contrast material. Numerous reasons for using unenhanced CT instead of IVU are cited in the radiology and urology literature (1–3,15–17). However, to our knowledge, no formal outcome studies have been performed to establish whether increased use of unenhanced CT is justified. There has been very little consideration of whether the clinical outcomes of patients with suspected UTC have improved with the advent of unenhanced CT for this purpose.
In our study, the use of imaging for detection of stone disease increased significantly per patient visit in 1999 (26.7% increase) when all diagnostic imaging examinations (ie, CT, US, IVU, and abdominal radiography) were considered. There was a clear shift to using CT as the predominant imaging modality in the evaluation of stone disease. There was a more than tenfold increase in the use of CT per patient visit in 1999 compared with that in 1997. IVU use per patient visit decreased by 95.1%. Far more patients underwent unenhanced CT in 1999 compared with the number of patients who underwent IVU in 1997 because unenhanced CT began to be used more liberally in the emergency department setting in 1999. The female patients who underwent unenhanced CT were slightly younger than their male counterparts; half were younger than 36 years.
As was expected, the use of intravenous contrast material diminished substantially (76.5% reduction) from 1997 to 1999. This was a major advantage associated with the increased use of unenhanced CT compared with the use of IVU.
The rate of hospital admissions immediately following the initial emergency department visits was approximately the same in 1997 (13.7%) and 1999 (13.4%). The reassurance value of negative CT results in reducing the hospital admission rate for patients with suspected UTC was not supported by our findings. Although previous studies to evaluate the effect of CT on subsequent patient admissions have been prospective, they have been based on responses from referring physicians who were surveyed with regard to what they would do before and after having the CT results (4,18). Although our study was retrospective, it involved evaluations of how referring physicians actually treated the patients. One might believe that if the patients with no clinically important abnormalities had been admitted prior to the use of unenhanced CT more frequently, there would have been a larger number of patients admitted for shorter stays in 1997 than in 1999. We did not find this to be true. The lengths of stay in the hospital, including those of 2 days or less, were nearly identical between the two years.
One might also believe that if important diagnoses were missed, the patients would return to the emergency department with abdominal symptoms that were the same as or more severe than their initial presenting symptoms. We did not observe a significant difference in the number of return visits to the emergency department by patients who initially presented with symptoms suggestive of UTC between 1997 (12.0%) and 1999 (13.7%). In both years, most patients returned for UTC disease.
Similarly, there was no significant difference in the number of admissions to the hospital within 30 days following the initial discharge from the emergency department between 1997 (4.5%) and 1999 (5.3%). Most of these admissions were for UTC disease, as were the return visits to the emergency department.
Other clinical outcome measures, including the length of stay in the emergency department and the spectrum of diagnoses at discharge from the emergency department, were similar in 1997 and 1999. The same types of diagnoses were made in patients who presented with symptoms suggestive of stones in both years. Fewer patients were examined in 1997, however, and when they were, they usually were examined with conventional IVU.
In our study, unenhanced CT depicted UTC in patients at a rate (61.9%) that was slightly higher than that reported by Smith et al (14) (49.5%) but comparable to rates reported by Abramson et al (4) and Kornecki et al (19) (60% and 59%, respectively). Although performing unenhanced CT shortens the examination time for patients with obstructing stones, in the present study the shorter examination time did not lead to a shorter stay in the emergency department. The emergency department length of stay was comparable between the two years. This finding might have been related to other factors in the emergency department being rate-limiting steps for discharging patients rather than imaging.
Unsuspected unenhanced CT findings that could affect acute patient care in the emergency department were observed in a minority of patients. Most of these patients had preexamination characteristics that indicated that they could be at risk of having a clinically important abnormality: age older than 40 years, history of malignancy, fever, and/or elevated white blood cell count. Young women with pelvic masses were the exception to this pattern, but one would expect pelvic US rather than unenhanced CT to depict the pelvic abnormality in these patients and thus spare them the exposure to radiation.
A substantial number of patients had incidental unenhanced CT findings that were unrelated to their symptoms. Follow-up imaging was recommended for most of these findings, which could have resulted in a large burden in terms of economics and patient time; however, these findings were rarely followed up at our institution. We did not have access to community or state medical records to determine whether these patients sought care at other institutions. Therefore, we could not assess the proportion of patients who underwent clinical treatment or imaging outside of our institution. Possible reasons that follow-up imaging was not performed in most of the patients with suspicious incidental findings at unenhanced CT include (a) the information was never received by the patient’s primary care provider or the patient lacked a primary care provider and (b) the patient’s physician did not believe the findings (ie, small liver or renal lesions thought to be benign) were important enough to warrant follow-up imaging.
The number of urologic procedures (eg, lithotripsy, ureteric stone extraction) performed for stone disease in 1997 increased by 67.7% (normalized to patient visits) in 1999. Although one could argue that these procedures may have spared patients from complications such as urosepsis or renal failure from prolonged obstruction, we did not observe a greater number of these complications in 1997 than in 1999. In this study, we were not able to assess whether the stones extracted in these procedures never would have passed or performing these procedures shortened the time to stone passage. Performing unenhanced CT might result in a greater number of cases in which urologists obtain information regarding stone position and size that motivates them to intervene earlier in patients. Although the greater frequency of urologic interventions may be beneficial to patients, these procedures have potential complications, such as ureteral perforation and adverse reactions to spinal and general anesthesia. An increased frequency of urologic procedures could lead to an increased frequency of these complications.
Our study was retrospective and subject to the potential biases related to this type of analysis. Chart abstraction was performed by several authors, who may have incorrectly included or excluded cases. However, the principal investigator trained these individuals and reviewed the majority of their work to reduce this limitation to a minimum. The principal investigator and chart abstractors were not blinded to the dates in the medical records, and this also could have been a source of bias, although an unintentional one. The effects of the unenhanced CT findings on patient care were determined in consensus between the individual chart abstractors and the principal investigator, and this could have been another source of bias regarding the clinical importance of the acute and incidental unenhanced CT findings.
We relied on records of diagnoses at admission to and discharge from the emergency department to identify the patient populations. We searched both years by using the same symptom codes to make the study populations as similar as possible and because we wanted to include only those patients in whom obstructing UTC was a potential diagnosis. Although we acknowledge that in a percentage of patients, the symptoms might have been miscoded, we did not find this to be true for the majority of patients at review of the medical records.
The retrospective study design prevented us from assessing changes in either physician confidence regarding diagnoses or patient perceptions after unenhanced CT results were revealed. In this study, we measured actual changes in patient outcomes rather than subjective physician or patient responses. Owing to similar factors, we were unable to measure the presumed increase in patient satisfaction that resulted from undergoing unenhanced CT and thus being spared the intravenous contrast material injection that is required for IVU.
There were substantially more patients in our 1999 (ie, post–unenhanced CT) population sample than in our 1997 (pre–unenhanced CT) population sample. We believe that there are several possible reasons for this difference: More patients visited our emergency department in 1999, and there was a marked increase in the number of patients who presented with the admission code of abdominal or flank pain. The increased use of unenhanced CT also helped us to identify more patients with urinary tract stone disease than were coded with this abnormality before this examination began to be used for this purpose. The use of new diagnostic imaging techniques has been found to lead to an increased prevalence of disease (20).
Despite a general increase in the number of visits to our emergency department in 1999, we believe that the study populations and patterns of medical care (other than radiologic procedures) in the 1997 and 1999 patient groups were very similar. The chronologic difference in the groups was relatively small (2 years), care was received in the same emergency department setting, and the majority of the physicians were the same in the two years. We believe that our search methods enabled us to identify the majority of patients with suspected UTC in both 1997 and 1999, and, therefore, our findings seem to be generalizable to other institutions.
In summary, there has been a large increase in the use of unenhanced CT, compared with the use of IVU, to examine patients with suspected stone disease in our emergency department. Increased use of unenhanced CT has not led to significant differences in short-term clinical outcome, but it has the potential to result in physicians ordering additional procedures. Findings other than UTC that could affect acute care were observed in a small percentage of patients, most of whom had pre-examination characteristics that suggested the presence of an abnormality. Young female patients might be evaluated by means of pelvic US instead of unenhanced CT, as most unsuspected findings (ie, diagnoses other than stone disease) were pelvic masses that were easily identifiable at US. Additional prospective studies are needed to confirm the results of our evaluation.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, R.H.G., T.C.L., J.L.S., S.L.V., L.S., W.J.D.; study concepts and design, all authors; literature research, T.C.L., L.S., R.H.G.; clinical studies, R.H.G., A.T.T., R.E., J.L.S.; data acquisition, J.L.S., R.E., T.C.L., L.S.; data analysis/interpretation, all authors; statistical analysis, R.H.G., D.D.; manuscript preparation, R.H.G., D.D.; manuscript definition of intellectual content, R.H.G.; manuscript editing, R.H.G., E.M.M.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. K. Paulson, C. Weaver, L. M. Ho, L. Martin, J. Li, J. Darsie, and D. P. Frush Conventional and Reduced Radiation Dose of 16-MDCT for Detection of Nephrolithiasis and Ureterolithiasis Am. J. Roentgenol., January 1, 2008; 190(1): 151 - 157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Gurel, D. Akata, K. Gurel, M. N. Ozmen, and O. Akhan Correlation Between the Renal Resistive Index (RI) and Nonenhanced Computed Tomography in Acute Renal Colic: How Reliable Is the RI in Distinguishing Obstruction? J. Ultrasound Med., September 1, 2006; 25(9): 1113 - 1120. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. H. Gottlieb Imaging for Whom: Patient or Physician? Am. J. Roentgenol., December 1, 2005; 185(6): 1399 - 1403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kirpalani, K. Khalili, S. Lee, and M. A. Haider Renal Colic: Comparison of Use and Outcomes of Unenhanced Helical CT for Emergency Investigation in 1998 and 2002 Radiology, August 1, 2005; 236(2): 554 - 558. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. S. Katz, M. Jain, M. J. Lane, E. M. Meiner, S. Bhalla, C. O. Menias, and C. M. Rucker Invited Commentary * Authors' Response RadioGraphics, October 1, 2004; 24(suppl_1): S28 - S33. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
