| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Radiology, Yale University School of Medicine, New
Haven, CT.
2 Present address: Gundersen Lutheran Hospital, La Crosse, WI.
3 Research Department, American College of Radiology, 1891 Preston White Dr.,
Reston, VA 20191.
4 Present address: Institute for Health Care Studies, Michigan State University,
East Lansing, MI.
5 Department of Economics, Yale University, New Haven, CT.
6 Yale School of Management, New Haven, CT.
Received October 4, 2006;
accepted after revision October 31, 2006.
Presented at the 2006 meeting of the Radiological Society of North America,
Chicago, IL.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We analyze non-individually identified data from the American College of Radiology's 2003 Survey of Radiologists, a stratified random sample mail survey that achieved a response rate of 63%, and data from the American College of Radiology's 1999 Survey of Practices. Responses were weighted to represent the distribution of individual radiologists and radiology practices nationwide. We present descriptive statistics and multivariable regression analysis results on the prevalence and uses of teleradiology in 2003 and comparisons with 1999.
RESULTS. Overall, 67% of all radiology practices in the United States, which included 78% of all U.S. radiologists, reported using teleradiology. A significant increase (p < 0.05) was seen in the prevalence of teleradiology or PACS, from 58% of practices in 1999 to 73% in 2003. Regression results indicate that, other practice characteristics being equal, in 2003, primarily academic practices were less likely to use teleradiology than private radiology practices, and medium-sized practices (5-14 radiologists) were more likely to have teleradiology than larger ones. In practices using teleradiology, home was the most frequent receiving site in both 1999 (81%) and 2003 (75%), the percentages being not significantly different.
CONCLUSION. Already a fixture of radiology practice in 1999, teleradiology increased in prevalence substantially by 2003. The primary use of teleradiology, transmission of images to home, did not change, suggesting that easing the burden of call remains the main use of teleradiology.
Keywords: digital images • Internet • practice of radiology • teleradiology
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The American College of Radiology (ACR) has published nationally representative data from its 1999 Survey of Practices describing the state of teleradiology in the United States at that time [13]. Although other data are available on a regional basis [14], the ACR survey is, to our knowledge, the only available source of data regarding the use of teleradiology in a clinical setting on a nationally representative scale.
We present both a comprehensive overview of the state of teleradiology in 2003 based on the ACR 2003 Survey of Radiologists, and an analysis of changes between 1999 and 2003.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The 2003 survey asked the respondents specifically whether their practice used teleradiology, and it asked a separate question about PACS. The 1999 survey did not ask respondents simply if teleradiology was used but if "teleradiology or PACS" was used. When data from 1999 and 2003 are compared, the 2003 results were adjusted to measure the use of "teleradiology or PACS" as opposed to just teleradiology in order to have meaningful comparisons. To be more specific, the prevalence compared between 1999 and 2003 was "use of teleradiology or PACS," coded 1 if the practice (or the radiologist) used teleradiology or PACS, and 0 otherwise.
Statistical and Data Analysis Methods
Responses were weighted so that the weighted statistics would be
representative of the answers that would have been received if all
radiologists in the United States had been surveyed and had responded. The
weighting process has been described previously
[18].
Univariate displays in Table 1 show the prevalence of teleradiology and its uses in radiology practices in 2003. Multiple logistic regression was used in Table 2 to identify the independent effect of each practice characteristic on the likelihood of radiology practices using teleradiology in 2003, statistically controlling for the effect of all other factors included in the analysis. Similar regressions were performed to analyze the destinations of teleradiology transmission (to home, within facility, between a practice's facilities, to outside the practice). Table 3 shows the prevalence of teleradiology and its patterns among individual radiologists in 2003. Table 4 shows the results of multiple logistic regression, which was used to statistically control for the effect of all other factors considered when examining the independent effect of each characteristic on the likelihood of individual radiologists being in a practice that used teleradiology in 2003.
|
|
|
|
We report the results of logistic regressions in terms of odds ratios (the ratio between the probability that an event does and the probability that it does not take place) to compare whether the probability of the analyzed event is similar for two groups. An odds ratio greater than 1 implies that an event is more likely in that group than in the reference category, and an odds ratio of less than 1 implies that an event is less likely in that group than in the reference category. The category with the largest n value was chosen as the reference category. For example, the category "private radiology practice" had more respondents than any other category and was therefore chosen as the reference category when comparing practice types.
To illustrate changes in teleradiology or PACS prevalence from 1999 to 2003 among various categories of radiology practices, Table 5 presents prevalence and uses in both years, by practice characteristics. In Table 6, ordinary least squares (OLS) multiple regression analysis is used to identify the independent effect of each practice characteristic on the change in teleradiology or PACS prevalence between 1999 and 2003. (In addition, occasionally categories are omitted or further subdivided in Tables 5 and 6 when effective comparisons could not be made due to differences in the survey questions between 1999 and 2003 or the smaller number of respondents in the 1999 survey.) We used OLS, not logistic regression, because we are primarily interested in the interaction coefficients, characteristic multiplied by year (where year is 0 if the characteristic is observed in 1999 and 1 if observed in 2003), that will show the change over time in teleradiology or PACS prevalence in various practice categories, such as private practices. Inferring changes in probabilities from interaction terms in nonlinear estimations—such as the logistic and its odds ratios—is not straightforward; the interaction coefficients may indicate the wrong magnitude, sign, and statistical significance of the marginal effects [19, 20]. Therefore, OLS, rather than logistic regression, is often used when interest is in interaction terms.
|
|
The dependent variable in the OLS regression is whether a practice used teleradiology or PACS, coded 1 if it does and 0 if it does not. The explanatory variables of primary interest are interactions between practice characteristics and time (0 if in 1999, 1 if in 2003). These measure the 1999-2003 change in the probability of using teleradiology relative to the change experienced by a reference category.
For comparisons of means in the univariate analyses, statistical tests were
conducted to determine whether differences between groups were statistically
significant at p 
0.05. Regression coefficients were considered
significantly different from 0 when p 0.10. Statistical tests at
p 0.10 are the standard in the health services research
literature. More restrictive tests at p 0.05 were used in the
univariate analyses because the SDs are relatively small in these analyses.
Readers who prefer the p 0.05 significance level in regressions
will find it indicated in the tables.
When discussed in tables and text, the following terms are used: large metropolitan area (population > 1 million), smaller metropolitan area (population 50,000-1 million), and nonmetropolitan or rural area (population < 50,000).
Data analyses were conducted using Stata statistical software (release 8, StataCorp), taking into account the stratified nature of the survey data, and Excel software (release 2002, Microsoft).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
When controlling for other variables using regression analysis, many practice characteristics were found to have a significant and independent effect on the likelihood of a practice using teleradiology (Table 2). For example, other factors being held constant, primarily academic practices are less likely to use teleradiology than private radiology practices (the reference category), their odds being approximately half the odds for private radiology practices. Other practices relatively likely to use teleradiology (statistically controlling for other factors) include those with 5-7, 8-10, and 11-14 radiologists (relative to the reference category of practices with 15 or more radiologists) and the reference category of practices that serve both hospital and non-hospital settings (relative to those that serve only hospital or only nonhospital settings). Practices located in the main city of a large metropolitan area had relatively low odds of using teleradiology when compared with the reference category of practices located in the main city of a small metropolitan area.
When other factors are held constant, among practices with teleradiology, multispecialty private practices had twice the odds of private radiology practices of using teleradiology to transmit images to home; and small and mid-sized practices (2-4, 5-7, 8-10, and 11-14 radiologists) had at least twice the odds of transmitting images to home compared with practices of 15 or more radiologists.
Teleradiology Prevalence Among Individual Radiologists in 2003
Overall in 2003, 78% of professionally active, posttraining individual
radiologists (as opposed to radiology practices) reported their practices used
teleradiology (Table 3). Of
radiologists in practices that used teleradiology, 77% were in practices that
used teleradiology for transmission to home, 50% were in practices that used
teleradiology to transmit images within the facility, 60% were in practices
that used teleradiology to transmit images among the practice's facilities,
and 19% were in practices that used teleradiology for transmitting images
outside the facility.
Controlling for other factors, radiologists ages 60-64 years had an odds ratio of 0.6 relative to radiologists ages 45-54 years of working in practices that use teleradiology (Table 4). Also, part-time radiologists were less likely to work in practices that use teleradiology (Table 4). Among radiologists whose practices used teleradiology, female radiologists had lower odds than male radiologists of being in practices that used teleradiology to transmit images to home.
Trends, 1999-2003
In 2003, 73% of radiology practices used teleradiology or PACS, an increase
of one fourth from the 58% level of 1999 (p 0.05)
(Table 5). Little of this
increase is due to the changing characteristics of practices over time.
Unreported regression results indicate that even controlling for changes in
practice characteristics, teleradiology or PACS prevalence increased by one
fifth from 1999 to 2003. In practices having teleradiology or PACS,
transmission to home was the most common destination in both 1999 (81%) and
2003 (75%), with no significant difference between the two years. Other
destinations also did not show any statistically significant changes from 1999
to 2003.
The use of teleradiology or PACS among private multispecialty practices increased from 53% in 1999 to 81% in 2003 (p < 0.05), and from 74% to 88% among private radiology-only practices (p < 0.05). Primarily academic practices did not experience a statistically significant change in teleradiology or PACS use.
OLS regression showed that, statistically controlling for the effects of other practice characteristics, all practices under the size of 15 radiologists had greater increases in teleradiology or PACS prevalence from 1999 to 2003 than did practices with 15 or more radiologists (Table 6). No significant differences were seen in rates of adoption of teleradiology or PACS among private versus academic practices or among practices in different census regions or in locations having different degrees of urbanness.
OLS regressions also indicated that in practices with teleradiology or PACS, no significant changes occurred from 1999 to 2003 in the patterns of transmitting images from practice to home when comparing practices across types, sizes, regions, or locations.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The high prevalence in practices located in nonmetropolitan areas speaks to the important role that teleradiology plays in allowing radiologists to serve a more geographically dispersed patient base. This is in keeping with the general concept that perhaps the most valuable role for telemedicine will be in making high-level skills available to patients in thinly populated areas.
Between 1999 and 2003, a significant increase occurred in the percentage of practices having teleradiology or PACS. Possible causes for this increase include advances in the technology of teleradiology, such as the decreasing cost of transmission bandwidth, or that a greater percentage of practices came to think that teleradiology would be useful. This latter factor, in turn, might be due to changing practice needs or to a growing appreciation of the potential value of teleradiology.
Against the background of these possible explanations, it is interesting to note that the period of 1999-2003 was largely one of a major easing of the radiologist shortage [21, 22]. This would suggest that the increased prevalence of teleradiology was not due to a growth in need, but to some combination of technologic advance—that is, to teleradiology systems becoming easier to use, more capable, and less expensive—and growing recognition of the value of teleradiology.
The easing of the radiology shortage over the period 1999-2003 is paradoxical, for the average radiologist's workload increased substantially during this period [17], with this increase driven by a continuation of the long-standing pattern in the United States of rapid increase in per capita utilization of medical imaging [23]. Teleradiology may explain much of the paradox. To be specific, by making call easier through transmission of images to homes or by eliminating call through transmission of images to nighthawk services, teleradiology may have made handling their increasing workload a good deal easier for radiologists. In addition, teleradiology may have allowed radiologists to become more efficient in completing work that was previously difficult to access because of geographic separation of the multiple facilities that a practice staffs.
The percentage of practices having teleradiology increased substantially—by one fourth—between 1999 and 2003, but among practices having teleradiology no significant change was seen in the percentage with transmission to home, within the facility, or beyond the facility but within the practice. In contrast, apparently a large increase occurred in the use of transmission to outside the practice. This probably reflects the growth of nighthawk services. The 2003 survey addressed this issue directly: We consider the percentage of all practices using teleradiology to transmit to out-of-practice radiologists (15%) a good approximation of nighthawk use in 2003 (Table 1). Unfortunately, the 1999 survey did not address directly the use of teleradiology to send images to out-of-practice locations, but inferring from some of the details of responses suggests the 1999 percentage was only a fraction as large as the 2003 percentage.
Certainly the fact that the most prevalent transmission route is to home indicates that reducing the burden of call is the main reason practices use teleradiology.
Many of our findings can be explained by the nature of practices. For example, call responsibilities tend to be shared by residents in academic practices, so they have a relatively smaller need for teleradiology. Small and medium-sized practices have fewer radiologists available to share call duties than large practices and so probably have more use for teleradiology for call.
Radiologists between the ages of 60 and 64 years were less likely to use teleradiology than radiologists 45-54 years old. This may simply indicate less comfort with digitized images or a previous investment in other types of radiology-related infrastructure, making a transition to a newer teleradiology infrastructure more difficult.
The Future of Teleradiology
It is difficult to determine whether the prevalence of teleradiology has
reached a plateau or whether it will continue to increase. Some predict that,
given the continued increase in demand for diagnostic imaging, the use of
teleradiology will also increase
[24]. Also, with the cost of
any given system capability likely to continue to decrease, teleradiology will
probably be installed in practices that currently judge it too costly for what
it offers.
Increasing teleradiology usage brings to the forefront issues of patient safety, cost, accreditation, quality control, radiologist satisfaction, and international teleradiology [25-27]. European practices are also struggling to find the best solutions to these issues [28, 29]. Additionally, teleradiology may soon branch into newer technologies such as wireless image transfer and grid computing, a technology that uses multiple computers connected by a network to perform computations on large data sets. These new technologies will raise important issues of image quality and patient privacy [30].
Some concern exists among both radiologists and the general public, as highlighted by the lay press [11, 12], that as a result of the ease with which images are currently transmitted, the jobs of U.S. radiologists will be outsourced to radiologists who have been trained in other countries because some receive a much lower fee. A number of factors currently make this unlikely [31]. First, the ACR recommends that any radiologist interpreting an image be licensed in the state from which the image originated [32]. Nighthawk practices currently claim to conform to this guideline, whereas radiologists trained in other countries cannot. Second, Medicare will not reimburse for medical services provided outside the United States [33]. This does somewhat limit the growth of nighthawk companies; however, they can circumvent this regulation by receiving reimbursement directly from the originating hospital. However, were this to occur with radiologists trained in other countries, it would likely shift the burden of malpractice coverage to the hospital, thereby off-setting any potential saving in radiologist fees [34].
The ACR Task Force on International Teleradiology [32] and repeated statements by the ACR affirm that patient safety and the quality of image interpretation should remain the highest priorities.
Strengths and Limitations of the Study
The principal strengths and limitations of the 2003 survey, including its
improvements over previous surveys, have been discussed elsewhere
[15,
17]. Strengths include a large
sample size, a high response rate, sophisticated weighting to make the sample
nationally representative, and careful cleaning of the data to improve
quality.
As with all surveys, the 2003 survey has important limitations. All surveys have some level of sampling variability, as measured by the standard errors, and these are large for categories with relatively few respondents. Nonresponse bias may exist with respect to characteristics not considered in developing the weights. Despite data cleaning and quality assurance measures, some errors remain in the data.
Survey questions may also contain some ambiguity. The term "teleradiology" was not defined in the survey; respondents may have had varying ideas as to what it includes.
The survey also did not allow complete comparisons between 1999 and 2003 because some of the questions in the two surveys were different.
Conclusions
These limitations do not undermine the essential findings of this study.
Already a fixture of practice in 1999, teleradiology increased in prevalence
by one fourth by 2003 so that it became available in two thirds of radiology
practices and to almost four fifths of all radiologists in the United States.
Its most commonly used application remained sending images to radiologists'
homes, presumably to ease the burden of call. In addition, the use of
nighthawk services increased greatly from 1999 to 2003. The increase in
teleradiology use also correlated with an easing of the radiologist shortage
despite an increase in overall workload.
The 2003 survey provides a point of comparison with the past and will serve as a detailed benchmark for future observation as teleradiology continues to evolve, possibly branching into new technologies and breaking old geographic boundaries. Teleradiology solves important problems for both radiologists and their patients but brings with it new challenges in maintaining responsibility, confidentiality, and quality.
Acknowledgments
We thank David Larson for his contributions to the content of this
article.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1395y (2006)
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. P. Borgstede Radiology: Commodity or Specialty Radiology, June 1, 2008; 247(3): 613 - 616. [Full Text] [PDF]
|
|
|
|
 |
|
 S. N. Singh and R. M. Wachter Perspectives on Medical Outsourcing and Telemedicine -- Rough Edges in a Flat World? N. Engl. J. Med., April 10, 2008; 358(15): 1622 - 1627. [Full Text] [PDF]
|
|
|
|
|
|
R. Steinbrook The Age of Teleradiology N. Engl. J. Med., July 5, 2007; 357(1): 5 - 7. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
