HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Lee, S. I. Articles by Hahn, P. F. Search for Related Content PubMed PubMed Citation Articles by Lee, S. I. Articles by Hahn, P. F.

Does Radiologist Recommendation for Follow-up with the Same Imaging Modality Contribute Substantially to High-Cost Imaging Volume?¹

¹ From the Department of Radiology (S.I.L., A.S., K.J.D., J.H.T., P.F.H.) and Massachusetts General Physician's Organization (J.B.W.), Massachusetts General Hospital, Harvard Medical School, 55 Fruit St, Boston, MA 02114. From the 2004 RSNA Annual Meeting. Received October 27, 2005; revision requested December 22; revision received May 17, 2006; final version accepted July 14. Address correspondence to S.I.L.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Purpose: To retrospectively measure repeat rates for high-cost imaging studies, determining their causes and trends, and the impact of radiologist recommendations for a repeat examination on imaging volume.

Materials and Methods: This HIPAA-compliant study had institutional review board approval, with waiver of informed consent. Repeat examination was defined as a same-modality examination performed in the same patient within 0 days to 7 months of a first examination. From a database of all radiology examinations (>2.9 million) at one institution from May 1996 to June 2003, a computerized search identified head, spine, chest, and abdominal computed tomographic (CT), brain and spine magnetic resonance (MR) imaging, pelvic ultrasonography (US), and nuclear cardiology examinations with a prior examination of the same type within 7 months. Examination pairs were subdivided into studies repeated at less than 2 weeks, between 2 weeks and 2 months, or between 2 and 7 months. Automated classification of radiology reports revealed whether a repeat examination from June 2002 to June 2003 had been preceded by a radiologist recommendation on the prior report. Trends over time were analyzed with linear regression, and 95% confidence intervals were calculated.

Results: Between July 2002 and June 2003, 31 111 of 100 335 examinations (31%) were repeat examinations. Body CT (9057 of 20 177 [45%] chest and 8319 of 22 438 [37%] abdomen) and brain imaging (6823 of 18 378 [37%] CT and 3427 of 11 455 [30%] MR imaging) represented the highest repeat categories. Among five high-cost, high-volume imaging examinations, 6426 of 85 014 (8%) followed a report with a radiologist recommendation. Most common indications for examination repetition were neurologic surveillance within 2 weeks and cancer follow-up at 2–7 months. From 1997 to mid-2003, MR imaging and CT repeat rates increased (0.71% per year [P < .01] and 1.87% per year [P < .01], respectively).

Conclusion: Repeat examinations account for nearly one-third of high-cost radiology examinations and represent an increasing proportion of such examinations. Most repeat examinations are initiated clinically without a recommendation by a radiologist.

© RSNA, 2007

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Imaging represents the second most rapidly growing sector of the health care industry, with costs reported to be escalating at 20% per year among private insurers (1). Between 2002 and 2003, per capita volume of imaging services grew up to 16.5%, whereas overall physician services grew 4.9% (2). This rapid increase has been due almost exclusively to high-cost imaging modalities—magnetic resonance (MR) imaging, computed tomography (CT), ultrasonography (US), and nuclear medicine (3).

The question of which forces drive increased utilization is controversial. Certainly, imaging technologies contribute importantly to patient care. However, other less beneficial reasons, such as patient demand, fear of litigation, lack of clinician time for patient evaluation, and self-referral have also been proposed (4–7). The contribution of radiologist self-referral has repeatedly come under scrutiny (8,9). The need has become pressing to identify factors contributing most substantially to imaging growth and to determine whether radiologists need to modify their examination recommendation practices. Thus, the purpose of our study was to retrospectively measure repeat rates for high-cost imaging studies, determining their causes and trends, and the impact of radiologist recommendations for a repeat examination on imaging volume.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
The study was compliant with the Health Insurance Portability and Accountability Act and was approved by the hospital institutional review board, with waiver of informed consent.

Our radiology department receives referrals from more than 2500 hospital-based staff physicians (16% primary care physicians, 84% specialists) and approximately 1100 community-based physicians in six counties with a total population of more than 4.5 million. The hospital, together with its emergency room and level I trauma center and its affiliated health centers, receives more than 1.4 million outpatient visits per year. The hospital has 893 beds, with 44 500 admissions and 33 000 surgical procedures annually. From July 2002 to June 2003, approximately 75% of the high-cost imaging studies (MR imaging, CT, US, and nuclear medicine) were performed in outpatients. Eighty percent of all studies were performed at the hospital, and the remainder were performed at one of three free-standing imaging centers. Studies were interpreted by 86 staff radiologists, 75 (87%) of whom are subspecialized by organ system (eg, chest, abdomen, central nervous system).

Definition of Repeat Examination
A repeat examination was defined as the same-modality examination performed with or without contrast material in the same patient within 0 days to 7 months after a first examination. This time interval was chosen to enrich the data set for examinations that were repeated during a single episode of a patient's care. The goal was to try to exclude examinations performed for annual screening or for a second episode of illness.

Three time intervals were chosen to approximate various patient care episodes. Zero days to 2 weeks encompasses most hospital admissions (our mean length of inpatient stay is 5.8 days ± 9.7 [standard deviation]), 2 weeks to 2 months encompasses care for an acute complaint requiring outpatient evaluation and treatment (eg, renal colic, pneumonia), and 2–7 months encompasses follow-up of lengthy treatment (eg, chemotherapy, physical therapy). This last interval captures commonly used 3- and 6-month follow-up time points, recognizing that 6-month follow-up might not occur at exactly 6 months.

Repeat Examination Identification
Finalized radiology reports in our radiology information system (IDX Systems, Burlington, Vt) are archived by patient medical record number and examination accession number in the database (Folio Views, version 4.2; the Fien Group, Encino, Calif). For each year from 1996 through 2002 and the first 6 months of 2003 containing a total of more than 2.9 million reports, we (P.F.H.) queried the database. The queries used a combination of examination codes and descriptions to identify all examinations of the following types: CT of the head, cervical spine, lumbar spine, chest, abdomen, and pelvis; MR imaging of the brain, cervical spine, and lumbar spine; US of the pelvis (nonobstetrical); and nuclear cardiology.

Lists for each examination type (eg, head CT, lumbar spine MR imaging) exported from database searches included accession number, medical record number, and date of service (DOS). A computer program identified pairs of examinations with identical medical record numbers but different accession numbers, calculating the number of days (DOS) between them. For each examination pair, the later examination was considered a repeat examination. For each repeat examination, only the most recent prior examination was retained. For examination pairs with DOS values equal to zero, we systematically chose the examination with the higher accession number as the repeat examination. The algorithm allows for multiple repeat pairs in the same patient. Thus, if a patient underwent three head CT examinations within 7 months, both the second and the third examinations would have been counted as repeat examinations.

Repeat examinations from 1996 were discarded. The data for 1996 had been carried along to account for repeat examinations in 1997 for which the preceding examination had been performed in 1996.

From the list of repeat examinations, we prepared the following three time frame sublists: (a) examinations repeated within 2 weeks of each other, (b) examinations repeated only between 2 weeks and 2 months, and (c) examinations repeated only between 2 and 7 months. From these lists, the number of repeat examinations in each month from January 1997 through June 2003 was compared with the monthly total number of examinations of the same type in the original lists exported from the database to yield the monthly percentage of all cases of each examination type in each of the repeat examination time frame categories.

Repeat Examination Database Verification
One percent of the repeat examinations pairs were randomly selected for manual verification. Working independently, two radiologists (S.I.L. and A.S.) verified that the same examination (accession number) was performed in the same patient (medical record number) on the dates specified (ie, DOS).

Examinations Performed after a Radiologist Recommendation
We separated the repeat cases with DOS during the full year between July 2002 and June 2003 for further analysis to determine the relationship between repeat examinations and a radiologist's recommendation for a repeat examination. Examinations with low repeat volume were not included in this analysis.

To identify reports with radiologist recommendations, we used LEXIMER, a computer algorithm for automated classification of unstructured radiology reports that has recently been described and verified (10). Briefly, this natural language report interpreter identifies keywords or phrases to determine that a follow-up examination has been recommended, with a reported sensitivity and specificity of 98.2% and 99.9%, respectively. The program exported the list of all examinations for which a radiologist recommendation was made.

The earlier examination of each examination pair for which the repeat examination was performed between July 2002 and June 2003 was then compared with the list of examinations after which follow-up was recommended. For each examination type and each time frame, percentages of total examinations that were repeated after a radiologist recommendation were computed.

The accuracy of LEXIMER was verified by two radiologists (S.I.L. and A.S.) independently. For the recommended examination pairs, the fact that the earlier examination contained a recommendation and the fact that the recommendation was for the same-modality imaging examination was verified in a random 10% sample. For the nonrecommended examination pairs, the fact that the earlier examination report did not contain a recommendation was verified in a random 1% sample.

Common Indications for Repeat Examinations
The indications for repeat examinations from July 2002 to June 2003 were manually determined from the radiology reports. Five percent of the repeat examination pairs were chosen by creating a sublist of cases with the numeral 6 in the 100s column of the seven-digit accession number (xxxx6xx) and any numeral between 0 and 4 in the 10s column. The following categories were analyzed: chest and abdominal CT and brain MR imaging repeated between 2 and 7 months and head CT repeated between 0 days and 2 weeks. For the examinations that were repeated between 2 and 7 months, indications were subdivided into one of the following categories: new clinical presentation, persistent or worsening symptoms without diagnosis, cancer follow-up, lung nodule follow-up, other lesion follow-up, or "miscellaneous." For head CT, the indications were subdivided into stroke, bleeding, or trauma; after intracranial surgery; hydrocephalus; mental status change (not included in the above categories); worsening headache; and "other." The indications were reviewed by two radiologists (S.I.L. and A.S.) independently, and differences were resolved by consensus.

Statistical Analysis
Using regression analysis, we modeled the monthly percentage of repeat cases in each category with a straight line, with month 1 (January 1997) through month 78 (June 2003) as the x-axis. When the 95% confidence interval of the slope excluded zero, the percentage of repeat cases was considered to be changing over time. Regression analysis was performed by using software (Statview II, version 1.03, 1998; Abacus Concepts, Berkeley, Calif). No correction for repeated comparisons was applied. Because about half of the duplicate pairs were contributed by patients with multiple duplicate pairs (data not shown), we performed a separate analysis of five key categories (all CT, all MR imaging, abdominal CT, chest CT, and brain MR imaging examinations repeated within 7 months) by selecting one duplicate pair at random for each patient with more than one duplicate pair.

Confidence intervals for the common indications for repeat examinations were calculated by using the formula 95% confidence interval = 1.96 ± [x · (1 – x)/n], where n is the sample size and x is the fraction of total examinations. Analysis to measure the level of agreement of the manual verification process was performed by using web-based software (www.chestx-ray.com/Statistics/kappa.html).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 
Repeat Examination Verification
Manual verification of 1% of the repeat pairs demonstrated 100% accuracy of the computer algorithm, with 100% concordance between the two radiologists.

Repeat Examination Rate
Among 100 335 high-cost studies, 31 111 (31%) were repeat examinations (Table 1). The highest repeat examination rates were for CT (24 825 [38%] of 65 793 examinations) and the lowest were for cardiac nuclear imaging (594 [8%] of 7831 examinations). When subdivided by examination type, body CT (9057 [45%] of 20 177 chest CT examinations and 8319 [37%] of 22 438 abdominal CT examinations) and brain imaging (6823 [37%] of 18 378 CT examinations and 3427 [30%] of 11 455 MR imaging examinations) represented the highest repeat categories.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

Beyond describing the role of the radiologist, to our knowledge, this is the first study to evaluate and quantify the impact of repeat examination ordering in high-cost imaging. In our practice, 31% of the high-cost radiology studies are repeat examinations. Over the 6.5 years surveyed, repeat examinations represented an increasing proportion of high-cost imaging examinations.

The most commonly repeated examinations are those in outpatient oncology patients who are undergoing chemotherapy or radiation therapy. These include body CT and brain MR imaging examinations repeated between 2 and 7 months. This reflects the increasing reliance on cross-sectional imaging in oncology follow-up (14–16). As treatment options multiply and prove increasingly expensive, oncologists have found that close surveillance, often with imaging, expedites the most effective and least complicated course of care. Annual costs for individual chemotherapeutics can exceed $100 000 (17). From this perspective, an additional $350 for a CT examination (18) billed to Medicare to confirm that the drugs are working seems fiscally reasonable.

Head CT examinations in inpatient neurosurgical patients that are repeated between 0 days and 2 weeks comprise the other major group of commonly repeated examinations. These are performed in patients in the intensive care unit after intracranial surgery who require close monitoring and multiple interventions. Because physical examination does not provide a reliable assessment in these patients, there is a very low threshold for performing a head CT examination to assess for intracranial hemorrhage or hydrocephalus. Again, the goal is to expedite treatment decisions, minimizing time in a unit bed. A recent analysis of inpatient hospital costs demonstrates that, for every $100 spent on inpatient diagnostic imaging, hospitalization was reduced by about 6 hours (19).

Repeat ordering of high-cost examinations is not primarily or even substantially caused by radiologists' recommendations but seems to be driven by changes in medical practice that take advantage of the information offered by imaging studies to achieve optimal care at minimal cost. An effective approach to control the ordering of high-cost imaging examinations must take into account the pressures driving utilization. Limiting high-cost imaging across the board without regard to these factors is unlikely to achieve the intended economic benefits and will adversely affect patient care. Protocols for imaging outpatient oncology and inpatient neurosurgical patients should be reviewed with a goal of minimizing imaging utilization without jeopardizing patient outcomes. Thus, the unit measured should not be the cost of a CT or MR imaging examination but the cost of the chemotherapy cycle or neurosurgical inpatient stay.

In light of this analysis, steps have been taken by the oncologists and by the neurologists and neurosurgeons at our institution to review their examination-ordering practices. Their goal is to arrive at recommendations for imaging by evaluating the entire clinical scenario. This will take into account not only the necessity to decrease unnecessary imaging but also the mandate that we not compromise patient outcomes. Once these guidelines are in place, it will be of interest to repeat our analysis to determine whether they have affected repeat imaging rates. On a more national scale, authors of chemotherapy regimens and algorithms for management of intracranial injury should specify the clinical guidelines for and the frequency of imaging, giving thought to the contribution to overall health care costs. Radiologists can then take note of these guidelines and include appropriate recommendations for follow-up imaging in their reports.

A limitation of our study was that it was performed at an academic medical center. Approximately 10% of radiology examinations nationally are performed in academic centers (3). Whether recommendation patterns for additional imaging differ between academic and nonacademic radiologists is unknown, but imaging utilization does differ between academic and community hospitals (20). Tertiary centers utilize high-cost examinations disproportionately and would be the centers most likely to respond homogeneously to centralized multidisciplinary consensus recommendations for examination ordering. Understanding and modifying practice patterns at the academic centers could impact community practice through training of future nonacademic physicians.

Another limitation was that examinations performed outside our institution would not have been captured. We could be under- or overestimating, by a similar percentage, the repeat examination rates and the volume of high-cost imaging resulting from radiologist recommendation. If repeat studies tend to stay in the system, then "leakage" of high-cost examinations to other imaging facilities inflates the percentage of radiologist-recommended studies by diminishing the denominator of all studies. Conversely, leakage can lead to underestimation if there are radiologist-recommended studies that are not captured as repeat examinations in the analysis.

Another limitation was that our analysis of the impact of radiologist self-referral on imaging volume did not include the impact of radiologist recommendations for different-modality but anatomically analogous examinations (eg, brain MR imaging after head CT). For complete assessment of the contribution of radiologist recommendations to the volume of high-cost imaging, this must also be included. However, this type of analysis requires that the reason for a radiologist recommendation be examined in greater detail. It is possible that another modality is being recommended appropriately—because an incorrect examination was requested for the question to be answered or because the examination revealed an abnormality that is most appropriately evaluated with a second, different imaging study. For more accurate measurement of inappropriately recommended examinations originating from use of another modality, methods for measuring study appropriateness and yield need to be developed.

In conclusion, nearly one-third of high-cost radiology examinations are repeat examinations in the same patient within 7 months. Repeat examinations represent an increasing proportion of high-cost radiology studies. Radiologist recommendations for a repeat examination account for only 8% of high-cost imaging studies. Thus, changing the practice among radiologists of making recommendations for follow-up imaging alone is unlikely to have a large impact on utilization rates. Rather, our results argue that repeat examinations are performed in response to medical indications and economic pressures within the health care system (eg, expensive drugs or inpatient beds) to minimize the overall cost of the patient's episode of care. Patient demand or fear of litigation are other possible factors not directly measured in our study. An effective approach toward establishing the optimal use of high-cost radiologic examinations will require evaluating the entire episode of a patient's care, defining scenarios in which high-cost imaging studies have maximal medical effect, and developing consensus multidisciplinary recommendations for high-cost examination utilization in those clinical situations.

	ADVANCES IN KNOWLEDGE

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 

• Nearly one-third of high-cost radiology examinations are repeat examinations in the same patient.
  
• Only 8% of high-cost examinations occur in cases in which a radiologist made a recommendation for a repeat examination on the preceding examination report.
  
• The majority of repeat examinations are in outpatients undergoing cancer treatment and neurosurgical inpatients.
  

	ACKNOWLEDGMENTS

 
The authors acknowledge the useful recommendations provided by Elkan Halpern, PhD, concerning statistical analysis of some of their data.

	FOOTNOTES

 

Abbreviations: DOS = date of service

Authors stated no financial relationship to disclose.

Author contributions: Guarantor of integrity of entire study, S.I.L.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, S.I.L., J.H.T.; clinical studies, all authors; statistical analysis, P.F.H.; and manuscript editing, all authors

	References

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ADVANCES IN KNOWLEDGE References

 

1. Farnsworth C. House Ways and Means Subcommittee on Health. March 2005. http://waysandmeans.house.gov/hearings.asp?formmode=detail&hearing=392. Accessed July 7, 2005.
2. 2005 MedPAC Report to Congress. March 2005. http://www.medpac.gov/publications/congressionalreports/Mar05TOC.pdf. Accessed July 7, 2005.
3. Bhargavan M, Sunshine JH. Utilization of radiology services in the United States: levels and trends in modalities, regions, and populations. Radiology 2005;234:824–832.[Abstract/Free Full Text]
4. Levin DC, Rao VM. Turf wars in radiology: other causes of overutilization and what can be done about it. J Am Coll Radiol 2004;1:317–321.[CrossRef][Medline]
5. Studdert DM, Mello MM, Sage WM, et al. Defensive medicine among high-risk specialist physicians in a volatile malpractice environment. JAMA 2005;293:2609–2617.[Abstract/Free Full Text]
6. Weisman CS, Morlock LL, Teitelbaum MA, Klassen AC, Celentano DD. Practice changes in response to the malpractice litigation climate: results of a Maryland physician survey. Med Care 1989;27:16–24.[CrossRef][Medline]
7. Childs AW, Hunter ED. Non-medical factors influencing use of diagnostic x-ray by physicians. Med Care 1972;10:323–335.[CrossRef][Medline]
8. Dehn T. A parallax view of diagnostic imaging. Decisions in Imaging Economics. 2003. http://www.imagingeconomics.com/library/200312-01.asp. Accessed July 7, 2005.
9. Baumgarten DA, Nelson RC. Outcome of examinations self-referred as a result of spiral CT of the abdomen. Acad Radiol 1997;4:802–805.[CrossRef][Medline]
10. Dreyer KJ, Kalra MK, Maher MM, et al. Application of recently developed computer algorithm for automatic classification of unstructured radiology reports: validation study. Radiology 2005;234:323–329.[Abstract/Free Full Text]
11. Hartman TE. Radiologic evaluation of the solitary pulmonary nodule. Radiol Clin North Am 2005;43:459–465.[CrossRef][Medline]
12. Conway C, Zalud I, Dilena M, et al. Simple cyst in the postmenopausal patient: detection and management. J Ultrasound Med 1998;17:369–372.[Abstract]
13. Kroon E, Andolf E. Diagnosis and follow-up of simple ovarian cysts detected by ultrasound in postmenopausal women. Obstet Gynecol 1995;85:211–214.[CrossRef][Medline]
14. Kahi CJ, Rex DK. Screening and surveillance of colorectal cancer. Gastrointest Endosc Clin N Am 2005;15:533–547.[CrossRef][Medline]
15. Reinhardt MJ, Herkel C, Altehoefer C, Finke J, Moser E. Computed tomography and 18F-FDG positron emission tomography for therapy control of Hodgkin's and non-Hodgkin's lymphoma patients: when do we really need FDG-PET? Ann Oncol 2005;16:1524–1529.[Abstract/Free Full Text]
16. Parsa AT, Wachhorst S, Lamborn KR, et al. Prognostic significance of intracranial dissemination of glioblastoma multiforme in adults. J Neurosurg 2005;102:622–628.[CrossRef][Medline]
17. Berenson A. Cancer drugs offer hope, but at a huge expense. New York Times Business Section. July 12, 2005.
18. Center for Medicare and Medicaid Services. 2005 Hospital Outpatient Prospective Payment System. http://www.cms.hhs.gov/providers/hopps/2005fc/1427fc.asp. Accessed July 20, 2005.
19. Beinfeld MT, Gazelle GS. Diagnostic imaging costs: are they driving up the costs of hospital care? Radiology 2005;235:934–939.[Abstract/Free Full Text]
20. Bansal S, Sunshine JH. Hospital activities of radiology groups in the United States: results of a 1992 ACR survey. AJR Am J Roentgenol 1995;165:453–465.[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Lee, S. I. Articles by Hahn, P. F. Search for Related Content PubMed PubMed Citation Articles by Lee, S. I. Articles by Hahn, P. F.