| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Departments of Radiology (S.T., K.H.V.), Internal Medicine (D.L.R.), and Pulmonary Medicine (D.A.S.), Emory Clinic, Bldg A, 1365 Clifton Rd NE, Atlanta, GA 30322. Received November 11, 2003; revision requested February 3, 2004; final revision received April 18; accepted April 21. Address correspondence to S.T. (e-mail: stefan_tigges@emoryhealthcare.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Radiography reports on all patients who underwent routine or screening posteroanterior and lateral chest radiography at a university-affiliated primary care clinic in 2001 were reviewed. Radiographic results were coded as normal or minor findings or as major abnormalities, such as pulmonary nodules, requiring further diagnostic evaluation. Outcomes of patients with major abnormalities were established by using chart reviews or reviewing additional radiographs. Costs were estimated by using 2002 Medicare reimbursement rates. The main measures assessed were frequency, costs, and rate of false-positive results of routine chest radiography.
RESULTS: Of 3812 radiographs obtained at the primary care clinic, 1282 (34%) were ordered for routine or screening purposes by the referring physician. Nine hundred twenty-two radiographs were obtained in male patients and 360 were obtained in female patients; their mean and median age was 49 years (age range, 4–87 years). Fifteen chest radiographs showed major abnormalities. No patient younger than 40 years had a major abnormality. Fourteen of the 15 findings of major abnormalities proved to be false-positive. No disease requiring treatment was diagnosed as a result of radiographic findings. The total cost for follow-up radiography and computed tomography was $46 609.49.
CONCLUSION: Routine chest radiography has low diagnostic yield in asymptomatic primary care patients.
© RSNA, 2004
Index terms: Cancer screening • Cost-effectiveness • Radiology and radiologists, socioeconomic issues • Thorax, radiography, 68.11
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The frequency of chest radiography continues to increase: An estimated 112 million examinations were performed in the United States in 2000 (Bhargavan M, unpublished data, 2002). The proportion of these examinations that were performed in asymptomatic patients in the primary care setting is unknown.
Previous studies (2–16) have been restricted to hospitalized patients or narrowly defined groups such as pregnant women, individuals with asthma, and patients undergoing psychiatric treatment. The purpose of our study, however, was to determine the frequency, diagnostic yield, costs, and rate of false-positive results of routine chest radiography performed in asymptomatic primary care patients.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Radiographs were considered routine only if the referring physician indicated screening or routine examination as the reason for performing radiography. Radiographic results were coded as normal or minor findings or as major abnormalities, such as pulmonary nodules, requiring further diagnostic evaluation. Examples of minor findings include healed rib fractures, pleural or parenchymal scarring, granulomas, and age-related changes such as osteophytes and aortic calcification. Findings such as an enlarged cardiac silhouette that had been documented at previously performed examinations were not considered major abnormalities for the purposes of this study because they required no further diagnostic testing. No further data on those patients who had normal, known, or minor findings were gathered.
The charts of patients with major abnormalities were retrieved and separately reviewed by two authors (S.T. and D.L.R.), who determined whether additional diagnostic examinations had been performed and gleaned other pertinent patient information, such as history of known radiographic abnormalities. If no further work-up had been performed and no follow-up images of any type were available, the putatively abnormal chest radiograph and any available previously obtained corresponding images were retrieved and four thoracic radiologists reevaluated the finding described in the radiographic report. A group consensus as to whether the finding actually represented a major abnormality was then reached. Two of the four radiologists (S.T. and K.H.V.) are authors, whereas the other two radiologists are on staff at our institution. These authors have 15 to more than 40 years experience in interpreting chest radiographs.
Patients with major abnormalities were assigned to one of two groups: those with true-positive results (positive radiograph in the presence of disease) and those with false-positive results (positive radiograph in the absence of disease).
Costs
Cost analysis was restricted to that of the costs of conventional radiography and computed tomography (CT), if performed. These costs were estimated by using 2002 Medicare reimbursement rates that were provided by our radiology billing office. We did not estimate the costs associated with nonradiologic procedures, follow-up office visits, wages lost by patients, or future expenses.
Statistical Analyses
Data were entered onto an Excel (Microsoft, Redmond, Wash) spreadsheet, and descriptive statistical values were generated by using Excel and SAS, version 8.2 (SAS Institute, Cary, NC) software. StatsDirect (Iain Buchan, Manchester, England) was used to calculate 95% confidence limits for the proportions of radiographs that demonstrated major abnormalities.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
Costs
Costs were estimated by using 2002 Medicare reimbursement rates. The 2002 Medicare global reimbursement rates for two-view chest radiography, chest CT without contrast material, and chest CT with intravenous contrast material are $35.05, $283.51, and $331.05, respectively. With the follow-up radiographs obtained in the patients with major abnormalities included, a total of 1288 posteroanterior and lateral chest radiographs were obtained, with a resulting cost for radiography of $45 144.40. One contrast material–enhanced chest CT scan and four nonenhanced CT scans were obtained at a cost of $1465.09. Thus, the total cost for radiography and CT was $46 609.49.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
No treatable abnormalities were detected on any of the more than 1000 screening radiographs obtained. Our results are consistent with those of investigators who evaluated the diagnostic yield of chest radiography in hospitalized patients (2–16). Sagel et al (11) found that the majority (68%) of hospitalized patients referred for screening chest radiography had normal findings. The low diagnostic yield of routine radiography has been demonstrated in pediatric patients, healthy pregnant women, hospitalized patients with asthma, preoperative patients, veterans, and other groups (2–16). Our study differed from previous investigations in that we evaluated primary care patients, who are likely to be healthier than hospitalized patients. In this patient population, the prevalence of occult disease should be low.
The marginal effects of chest radiography on the detection of unsuspected disease and on treatment decisions were evaluated at the Mayo Clinic (17). In a group of 61 primary care patients without thoracic complaints who were seen at routine office visits that included assessments of anatomic systems and physical examinations, only one patient had an important diagnosis that was made on the basis of chest radiography findings and that was not suspected on the basis of medical history and physical examination results. In that case, a mass that proved to be round atelectasis secondary to asbestos exposure was discovered. No treatment was necessary.
Chest radiographs are obtained in otherwise healthy patients to detect life-threatening diseases such as lung cancer early so that cure is possible. Results of the Mayo Clinic screening study (18,19) demonstrated no benefit in screening smokers for lung cancer by using radiography. This result is due in part to the poor sensitivity of chest radiography for the detection of early lung cancer. In one study, the sensitivity of chest radiography for the detection of lung nodules that proved to be cancer was only 36% (20). Results of another study (21) showed that radiographically missed lung cancers may be quite large, with a median diameter of 1.9 cm.
Even if a cancer is detected when it is 1 cm in diameter, it may be associated with the same prognosis as a lesion three times as large. Patz et al (22) compared the survivals among patients with different-sized stage IA non–small cell lung cancers and found no significant difference (P = .701) in survival between patients with smaller lesions and those with larger cancers. These authors pointed out that even a 1-cm lesion is composed of 1 billion cells and probably has been present for many years.
The possible harmful effects of routine chest radiography were emphasized in review articles soon after it became clear that the benefits of routine radiography are minimal (23–25). In a program carried out for many years, even a low false-positive rate can result in unnecessary diagnostic procedures being performed. In our study, 1.4% of the patients 40 years of age or older had false-positive results. If one assumes a binomial distribution and that the probability of having false-positive findings remains 1.4% throughout the course of such a program, then a patient older than 40 years who undergoes chest radiography annually for 10 years will have a 13% probability (1 – .98610) of having at least one false-positive result. One who undergoes radiography annually for 20 years will have a 25% probability (1 – .98620) of having at least one false-positive finding.
Not all investigators have observed a low diagnostic yield with routine radiography (26–28). For example, a 1981 study of 113 chest radiographs obtained at admission at a veteran’s hospital revealed abnormal findings in 46% of cases, and the authors concluded that obtaining such radiographs is necessary in patient groups with a high prevalence of cardiovascular disease (26). However, data regarding whether the abnormalities were acute or chronic and whether the findings led to changes in patient care were not reported. In addition, some of the findings, such as pneumonia, may have led to patient admission to the hospital.
Our study was subject to the usual limitations of a single-site retrospective investigation. It is possible that the asymptomatic patients with major abnormalities were incorrectly judged to have symptoms by the primary care clinicians, although this seems unlikely. In addition, our follow-up methods were imperfect and not uniform: Some patients were followed up with CT, whereas others were followed up with repeat radiography, and some outcomes were determined by using medical records, findings on prior radiographs, and/or consensus reading.
The radiology database that we used to obtain the bulk of our data did not include information on the patients’ smoking status. Although it is always desirable to have as much information about the study subjects as possible, the lack of this information should not have materially affected the generalizability of our results. If the proportion of patients in our study group who smoked was equal to or larger than the population as a whole, then our results should be broadly applicable. Even if our study group had included an unusually low proportion of smokers, our results would still probably be applicable to the general population because screening thoracic radiography is very insensitive for detecting smoking-related diseases such as emphysema, coronary atherosclerosis, and, as discussed earlier, lung cancer.
Our cost estimates are conservative because we considered only those costs related directly to outpatient radiography and used Medicare reimbursement rates to establish the dollar value of chest radiographs and CT scans. A study performed by using rigorous methods to establish the costs of conventional cervical spine radiography revealed that Medicare reimbursement rates are substantially lower than the costs of the resources actually used (29). The same may also be true for chest radiography and chest CT.
In summary, our study results indicate that a large proportion of asymptomatic primary care patients at our institution undergo routine chest radiography. The diagnostic yield of chest radiography in this population was low, and most findings were false-positive. Although 1282 patients were screened at a total cost of $46 609.49, performing routine chest radiography resulted in no treatable diagnoses and multiple unnecessary examinations.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.T.; study concepts and design, all authors; literature research, S.T.; clinical studies, S.T., D.L.R., K.H.V.; data acquisition and analysis/interpretation, all authors; statistical analysis, S.T., D.A.S.; manuscript preparation, definition of intellectual content, revision/review, and final version approval, all authors; manuscript editing, S.T., D.A.S.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. R. WILKINS and R. L. WILKINS Clinical and Radiographic Evidence of Pneumonia Radiol. Technol., November 1, 2005; 77(2): 106 - 110. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. O. Stolberg and S. Tigges Comments on Routine Chest Radiography * Dr Tigges responds: Radiology, July 1, 2005; 236(1): 368 - 368. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
