| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Musculoskeletal Imaging |
1 From the Program for the Assessment of Radiological Technology (ART Program), Departments of Radiology (J.J.N., E.H.G.O., A.Z.G., G.P.K., M.G.M.H.), Epidemiology and Biostatistics (J.J.N., E.H.G.O., M.G.M.H.), Orthopaedic Surgery (J.A.N.V.), and Traumatologic Surgery (A.B.v.V.), Erasmus MC, University Medical Center Rotterdam, Dr Molewaterplein 40, 3015 GD Rotterdam, the Netherlands. Received June 26, 2004; revision requested September 1; revision received October 20; accepted November 11. Address correspondence to M.G.M.H. (e-mail: m.hunink{at}erasmusmc.nl).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Institutional review board approval and informed consent were obtained. Patients with recent trauma of the wrist, knee, or ankle were randomized across two diagnostic strategies: radiography alone (reference group) or radiography followed by a short MR imaging examination (intervention group). Measures of effectiveness included the number of additional diagnostic procedures, time to last diagnostic procedure, and number of days absent from work. Measures of effectiveness were analyzed by using an exact Wilcoxon-Mann-Whitney test. Time to convalescence and quality of life were analyzed by using a t test. Cost analysis was performed from a societal perspective and analyzed by using a t test.
RESULTS: Five hundred patients (207 women, 293 men; mean age, 34.8 years) with acute injury of the wrist, knee, or ankle were randomized. In the intervention group, quality of life for patients with knee injuries was significantly higher during the first 6 weeks, and time to completion of diagnostic work-up was significantly shorter (mean, 3.5 days for intervention group vs 17.3 days for reference group). The number of additional diagnostic procedures was significantly lower in the intervention group versus the reference group (nine vs 35, respectively) for patients with knee injuries. Patients with knee injuries showed the largest difference in costs (intervention group, 
1820 [$1966]; reference group, 2231 [$2409]) owing to a reduction in productivity loss. Costs were higher in patients with wrist injuries and almost equal in patients with ankle injuries. All cost differences, however, were not significant.
CONCLUSION: Compared with radiography, MR imaging in patients with acute wrist or ankle injuries is neither cost saving nor effective in expediting diagnostic work-up or improving quality of life. In patients with knee injuries, a short MR imaging examination shortens the time to completion of diagnostic work-up, reduces the number of additional diagnostic procedures, improves quality of life in the first 6 weeks, and may reduce costs associated with lost productivity.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Although the addition of an MR imaging examination to the initial diagnostic work-up of patients with recent acute traumatic joint injury will obviously generate extra costs, it is expected that this extra examination may shorten the time to diagnosis. MR imaging at first presentation may also save costs if it obviates further follow-up, additional diagnostic procedures, and temporary treatment measures, such as the use of a plaster cast in cases of a suspected scaphoid fracture. It may also save costs to society if the early diagnosis leads to an earlier treatment and, hence, to an earlier recovery, thereby reducing a loss of productivity. Because the costs of lost productivity can be considerable, reducing the number of days off of work in a minority of patients may justify a short MR imaging examination in all patients. Thus, the purpose of our study was to assess prospectively if a short MR imaging examination performed with low-field-strength dedicated extremity MR imaging in addition to radiography is effective and cost saving compared with the current diagnostic imaging strategy (radiography alone) in patients with recent acute traumatic injury of the wrist, knee, or ankle.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Three parallel studies were conducted for the wrist, knee, and ankle. Patients were eligible if they had recent injury (within 7 days of trauma) of the wrist, knee, or ankle and if radiography of the affected joint had been requested by specialists or residents in traumatology, orthopedic surgery, or emergency medicine. If more than one joint was affected, inclusion was directed according to the joint that caused the most complaints. The patients were randomized across two diagnostic strategies: radiography alone (reference group) and radiography followed by a short MR imaging examination (intervention group). Randomization was achieved by drawing from consecutively numbered sealed envelopes containing computer-generated random assignments. Block randomization was applied, and a block size of 20 was used to achieve an equal number of patients in both study groups. Randomization was stratified according to joint, and the envelopes were labeled accordingly.
Patients were excluded if they had substantial injury to the head, back, thorax, or abdomen; if they had a compound fracture; if they were in need of urgent treatment (eg, if they had ankle luxation or an open fracture); if they had preexisting complaints in the same joint; or if they were intoxicated. Patients were enrolled from 8:00 AM to 11:00 PM, 7 days a week, from August 1999 to May 2001. Inclusion was carried out by research staff and by the radiology technologists on service.
Sample Size Calculation
To demonstrate a mean difference in costs of 500 ($540), with a standard deviation of the mean difference equal to 800 ($864), a sample size of 80 patients (power = .8, 
= .05, two-sided test) was required per joint evaluated. Patients were enrolled until at least 80 patients per joint were included.
Radiography and MR Imaging
Radiographs of the wrist were obtained in the lateral and posteroanterior projection and were supplemented with a scaphoid series if a scaphoid fracture was suspected. Radiographs of the knee were obtained in the lateral and anteroposterior projection and were supplemented with patellar or tunnel views if pathologic abnormalities of the patellofemoral joint or intercondylar notch were suspected. Radiographs of the ankle were obtained in the lateral and anteroposterior projection, with 15° endorotation of the foot.
MR imaging was performed with a 0.2-T dedicated extremity MR imaging system (Artoscan M; Esaote Biomedica, Genoa, Italy) immediately after radiography. We used a short MR imaging protocol (Table 1) with one excitation for each sequence by using a rectangular field of view and a limited number of phase-encoding steps. After a software upgrade in February 2000, the T1-weighted spin-echo half-Fourier sequence was replaced with a T1-weighted spin-echo sequence, and, for the turbo multi-echo sequence, the first echo time was shortened from 38 to 28 seconds (on the Artoscan M system, echo time is fixed and cannot be changed by the user). The software upgrade included improved filtering and frequency sampling. The total acquisition time was 5–6 minutes for all joints, with a total examination time (including MR imaging system start-up and patient positioning) of approximately 15 minutes.
|
Radiographs were initially assessed by the treating physician in the emergency department. Radiographs were then reassessed the next day during a reading session with the treating physician by one of the two musculoskeletal radiologists mentioned earlier. If the reassessment differed from the initial assessment for either the MR images or the radiographs, treatment from that moment on was based on the assessment of the musculoskeletal radiologist. No second MR imaging or radiographic examinations were performed between the two readings.
Patient Follow-up
Clinical data were collected from patient hospital records and from the hospital computer system (J.J.N., E.H.G.O). The follow-up period was for as long as the patient had daily complaints, as recorded on questionnaires, with a maximum follow-up period of 6 months. We expected that this 6-month follow-up period would be long enough to capture all relevant differences in the outcomes between the two strategies. Questionnaires were sent to patients at 1 week, 6 weeks, 3 months, and 6 months after inclusion. These questionnaires included quality-of-life measuring instruments and questions about the number of days absent from work, the number of days to convalescence (ie, the number of days until the patient no longer had daily complaints), medical treatment at other institutions, treatment by a general practitioner, treatment by alternative medicine practitioners, and out-of-pocket expenses. If questionnaires were not returned, we interviewed the patients by telephone (J.J.N., E.H.G.O.) and urged them to return the questionnaire, provided that patients could be reached within two calling attempts. If the patient indicated on the questionnaire that he or she no longer had daily complaints, then follow-up was terminated because no additional effects of the injury were expected.
Measures of Effectiveness
Measures of effectiveness included quality of life, time to completion of the diagnostic work-up, the number of additional diagnostic procedures during follow-up (eg, radiologic imaging, diagnostic arthroscopy, or blood tests), the number of days absent from work, and the number of days to convalescence. We included the number of diagnostic procedures during follow-up to calculate costs and to be sure that the relatively lower quality of the short MR imaging examination did not lead to additional standard MR imaging examinations in the intervention groups. Quality of life was measured with a descriptive instrument (Short Form 36 [SF-36] health survey) (9) and with a valuative instrument (EuroQol), which provided general population values (10,11). The SF-36 health survey consists of 36 questions that are used to assign values to eight health domains (ie, physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health). Role physical and role emotional refer to the degree of interference with work or other daily activities as a result of impaired physical health or emotional problems. The EuroQol uses five questions to obtain a single index value for health status. The time to completion of the diagnostic work-up was defined as the time from initial presentation to the time of last diagnostic imaging or arthroscopic examination. The number of diagnostic procedures was assessed by reviewing the hospital computer system and by asking patients by questionnaire or by telephone inquiry if any diagnostic procedure had been performed at another institution. Likewise, the number of days absent from work and the number of days to convalescence were asked in the questionnaires or, if the questionnaire was not returned, by telephone inquiry.
Analysis of Response
To test for response bias, the baseline characteristics (ie, sex, age, and randomization result [MR imaging vs no MR imaging]) of patients who returned the questionnaires were compared with the baseline characteristics of patients who did not return the questionnaires. Baseline characteristics were compared by using a t test, and a P value of .05 was used as a threshold for statistical significance. To assess if the response rate depended on the severity of disease, we used the variables "treatment" and "number of days absent from work" as proxy measures of the severity of disease and assessed whether the outcome of these variables differed significantly between responders and nonresponders. Although we did not record the ethnicity of the patients, we analyzed if patients with a nonnative name showed a response rate that was different from those with a native name. This analysis allowed us to determine if culturally determined differences in response rate were present.
Statistical Analysis of Effectiveness
The data were analyzed on an intent-to-diagnose and intent-to-treat basis. For both strategies, the number of days to convalescence was compared by using a t test. The number of days to completion of the diagnostic work-up, the number of diagnostic procedures performed during follow-up, the number of days absent from work, and the evaluation of response bias were analyzed by using an exact Wilcoxon-Mann-Whitney test and a 
2 test. Because of multiple testing, we adjusted the P value significance threshold from .05 to .01.
If a patient indicated on one of the four questionnaires that he or she was free of complaints, we expected no further change in the quality of life for the patient, and the EuroQol and SF-36 health survey data measured at that point in time were extrapolated to the remaining questionnaires. If, at this stage, less than 20 responses were available in either the intervention or reference groups, which were stratified according to joint, then these responses were regarded as too few for meaningful analysis. If the EuroQol and SF-36 health survey data for patients without complaints were not available for extrapolation, then the mean values for the EuroQol and for the eight SF-36 health survey domains were used instead. These mean values were derived from the completed questionnaires in all complaint-free patients. Linear regression analysis was used to determine if these mean values had to be adjusted for age and sex. Imputed values were used in the analysis only if they constituted less than 20% of the data. The preference-based EuroQol score was calculated by using the regression equation, as described by Dolan (12). Differences between the randomized groups for EuroQol and SF-36 health survey domain scores were analyzed by using a t test.
Cost Analysis and Related Statistical Analysis
The cost analysis included measures of all medical and nonmedical costs associated with the initial injury during the 6-month follow-up period that were relevant from a societal perspective. Direct medical costs included the cost of hospital visits, diagnostic and therapeutic procedures, physical therapy inside or outside the hospital, visits to general practitioners, and visits to alternative medical practitioners, as well as the travel costs for each visit. Costs of all diagnostic procedures were calculated by using a bottom-up approach (13) that took into account the initial investment of equipment, additional costs during use, maintenance costs, years of use, discounting and annuitization (14), the number of procedures performed per year, personnel costs, materials used, room rent, housekeeping, administration, and overhead costs. Costs were discounted at a rate of 3% per annum (15). Costs of initial treatment (eg, the costs of a bandage or plaster cast) were calculated likewise. The estimated actual cost of hospital visits and hospital admissions were obtained from the Dutch Council for Care Insurances (13). For costs of surgical procedures, reimbursement tariffs, as established by the Dutch Central Organ for Tariffs in Healthcare, were used. Indirect medical costs (ie, medical costs in life-years gained) were not considered because we did not expect a difference in life-years.
For measurement of direct nonmedical costs, we took into account out-of-pocket expenses and patient time costs. Information on out-of-pocket expenses was derived from the questionnaires or from telephone inquiry. Patient time costs were determined by multiplying the time spent on medical procedures (eg, hospital admissions, outpatient hospital visits, visits to the general practitioner, travel time, and waiting time) by the average net income of patients (both working and not working) who were comparable in age for the year 2000 (Dutch Central Bureau for Statistics, www.cbs.nl) (15).
For indirect nonmedical costs, we assessed costs associated with production losses. Production losses were estimated by using the friction cost method described by Koopmanschap et al (16) and recommended in guidelines for cost-effectiveness analysis (15). This method assumes that production losses are generated only during the time it takes to replace a sick employee–that is, during the friction period. Short-term absence from work may cause limited loss of production because work may be postponed and performed on return, or work may be performed by colleagues. Long-term absence from work generates costs owing to actual production losses, extra costs to maintain production, and the costs of filling a vacancy and, if a permanent replacement is needed, of training a new worker. These costs are called friction costs and are influenced by the unemployment rate and the extent of mobility in the labor market. Friction costs were calculated by using the friction cost data estimated for workers in the Netherlands in 1998, which are based on age and sex and are described by Oostenbrink et al (13). These data were adjusted for a general increase in income in 2000 compared with 1998. The friction period for the year 1998 was estimated at 4 months (13), but because of increased shortage in the labor market, we estimated the friction period to be 6 months for the year 2000 (M. A. Koopmanschap, oral communication, 2002).
Costs were analyzed on an intent-to-diagnose and intent-to-treat basis and were compared by using a t test, with a P value of .05 as the threshold for statistical significance. We performed cost analysis for the three joints separately. Costs are expressed in both euros and in U.S. dollars and were calculated by using the mean exchange rate during the inclusion period (1.08 = $1.00).
Sensitivity analysis was used to determine the robustness and generalizability of the study results. One-way sensitivity analysis was conducted by exploring 50%–200% of the baseline value for each single cost and by analyzing the time assigned to each medical procedure, as well as the travel time, friction costs, and friction period. Three-way sensitivity analysis was conducted on the parameters that were most sensitive in the one-way analysis.
All analyses were performed by using statistical software packages (Excel 97 SR-2, Microsoft, Redmond, Wash; SPSS, version 10.0.0, SPSS, Chicago, Ill; and StatXact-4, Cytel Software, Cambridge, Mass).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Of the remaining 472 patients, 237 were allocated to the MR imaging strategy (MR imaging plus radiography) and 235 were allocated to the reference strategy (radiography alone). In eight patients who were allocated to the MR imaging strategy, either MR imaging was not performed or the results were not interpretable. In one patient, a technical failure occurred, and in another patient, the examination was interrupted because of pain. In one obese patient, the image quality was too poor for interpretation, and in two patients, the MR imaging system was not available. In three patients, the joint of interest could not be positioned in the center of the magnet bore: One patient was not able to extend the elbow, one patient had a locked knee, and one patient's knee was too big to fit into the magnet. For analysis, these eight patients remained in the strategy they were originally assigned to.
Response Characteristics
We found no significant difference in sex, age, or the number of days absent from work between patients who had returned questionnaires and patients who had not returned questionnaires. Among patients with ankle injuries, however, the number of patients who returned one or more questionnaires was slightly higher for those who underwent MR imaging than for those who did not (84% vs 72%, respectively; P = .05). Furthermore, in patients with knee or ankle injuries, the response rate differed significantly between patients who had undergone additional treatment and those who had not undergone additional treatment (93% vs 61% for patients with knee injury [P < .01] and 85% vs 68% for patients with ankle injury [P = .01]). An estimated 20% of our patients were nonnatives who had a varying ability to speak the national language. Patients with a nonnative name showed a response rate that was significantly lower than that of patients with a native name (P = .001, 2 test). The patients with nonnative names were evenly distributed between the two randomized groups (P = .23, 2 test).
Effectiveness
After extrapolation of EuroQol and SF-36 health survey data in censored patients (ie, those in whom follow-up was ceased because they no longer had complaints), data were used for analysis only if the number of completed questionnaires exceeded 20 in both the intervention and reference groups for each of the joints. More than 20 questionnaires were returned in all groups except for the third and fourth questionnaire in patients with wrist injuries. Linear regression analysis showed that mean EuroQol and SF-36 health survey domain scores of patients without complaints were not significantly influenced by age or sex, and, therefore, the total mean values of EuroQol and SF-36 health survey domain scores were used for imputation. In all four questionnaires, which were stratified according to joint and randomization result, less than 20% of the questionnaires were imputed.
In patients with wrist or ankle injuries, no significant difference was found between the intervention and reference group for EuroQol scores measured at 1 week, 6 weeks, 3 months, and 6 months after injury (Fig 2). In patients with knee injuries, the EuroQol score at 1 week and at 6 weeks was significantly higher in the intervention group (P = .003 and P = .01, respectively; t test). Three months after injury, this difference was no longer significant.
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
We expected that the main benefit of early diagnosis of acute trauma in the wrist, knee, or ankle by using a dedicated extremity MR imaging system would be twofold—that is, a decrease in the need for follow-up and a reduction in the time to convalescence. A decreased need for follow-up could lead to potential cost savings by reducing follow-up visits, additional imaging examinations, diagnostic arthroscopy, and temporary treatment measures. This, in turn, would also reduce patient time, travel, and friction costs. A reduction in the time to convalescence would result from a reduction in the time to diagnosis, which is associated with earlier treatment. This could lead to cost savings for society owing to an earlier resumption of work (ie, a reduction in friction costs). The aforementioned effects could lead to a total cost reduction provided that the cost savings outweigh the additional costs of performing a short dedicated extremity MR imaging examination in every patient.
In our patient population, we could not demonstrate that a short MR imaging examination in addition to radiography provided cost savings from a societal perspective. In patients with knee injuries, however, the mean total costs were substantially lower in the intervention group because of the reduced costs of lost productivity, but this difference was not statistically significant. The influence of a short MR imaging examination on other outcomes was demonstrated only in patients with knee injuries; quality of life was improved in the first 6 weeks, the number of examinations during follow-up was lower in the intervention group, and the time to diagnosis was reduced by 12 days. Both the duration of absence from work and the time to convalescence did not differ significantly between the intervention group and the reference group for all three joints.
The addition of MR imaging to the initial diagnostic evaluation had no influence on the quality of life 3–6 months after injury, as measured with the EuroQol and SF-36 instruments. The short-term (1–6 weeks) EuroQol outcome in patients with knee injuries was, however, significantly higher in the intervention group compared with the reference group. This may reflect a more adequate initial treatment in the intervention group. Another possible explanation would be that, despite randomization, the intervention group had less severe injuries than the reference group. There were, however, no indications of a difference in the severity of injury between the two groups; there was no significant difference in number of patients receiving treatment, the number of hospital admissions, or the length of hospital stay between the two groups.
In the SF-36 health survey, the valuation of general health was strikingly similar in all groups and did not change over time, whereas all other domains improved over time. A traumatic injury of the wrist, knee, or ankle apparently does not affect a patient's appreciation of his or her general health. This might be caused by the fact that, for most injuries, total recovery is expected, and, therefore, the patient does not consider the injured joint as a sign of impaired general health.
A limitation of this study is the setting in which it was performed. The Dutch health care system is faced with the problem of long and variable waiting lists. The variation in waiting lists for outpatient visits and surgical procedures may have influenced the time to convalescence as an outcome measure. At our hospital, the waiting list for arthroscopy at the time of the trial was 5 months. We therefore used the time to the last diagnostic procedure as a more representative measure. Likewise, it is plausible that the friction costs were also influenced by the waiting lists. Although the extra friction costs caused by waiting lists were incurred in both the intervention group and the reference group, the increase of these costs in both groups could have masked a difference between the groups. Furthermore, friction costs were the most sensitive single variable influencing the cost difference between the randomized groups. Because these costs are influenced by income and by the state of the labor market, the results of the study are probably applicable only to societies comparable to the Dutch society.
We recognize that early detection and surgical treatment of a condition is not always necessary for recovery. Some meniscal tears, triangular fibrocartilage complex lesions, or ligament injuries may heal without specific treatment (20). In this perspective, the presence of waiting lists not only is unfavorable for the patient but also may allow time for the natural recovery of a condition.
The fact that not all eligible patients were asked to participate may have introduced a selection bias if either more or fewer patients with obvious injuries were preferentially not asked to participate. In the more seriously injured patients, MR imaging is likely to have had more additional diagnostic value than in patients with minor injury; therefore, this bias would cause the effects and potential cost reduction of a short MR imaging examination to be underestimated.
The response rate on the questionnaires was low for all groups. At the moment of enrollment, the importance of completing the questionnaires was stressed, and patients were urged to return the questionnaires. If the questionnaires were not returned in time, the patients were called by phone. Despite our efforts, the response rate was less than optimal. Typical response rates for postal questionnaires are about 60%–65% (21–30). A reason for the lower response rate in our study may be that patients are less willing to take the effort to respond if the disease does not have a major effect on their health and if the disease is not likely to last long.
Another factor influencing the response rate is the fact that nonnative participants may have had difficulty understanding the national language. The lower response rate for patients with a nonnative name may be attributable to their limited knowledge of the language or possibly to cultural differences. Patients with nonnative names were evenly distributed between the two randomized groups, so their relatively large contribution to the high nonresponse rate is not likely to have caused response bias. For the analysis of costs, we were much less dependent on the low response rate because we could retrieve most data from the hospital computer system and patient records, which were complemented by telephone inquiry.
In this pragmatic randomized controlled trial, no blinding was used. Patients knew whether an MR imaging examination was performed or not. This influenced the response rate of patients with ankle injury in that patients who had undergone MR imaging were more likely to return their questionnaires. It also potentially influenced the time to convalescence because patients may have been reassured by a negative MR imaging result. Reassurance is, however, part of performing MR imaging and should not be regarded as a bias in this context. Blinding could have been achieved by performing a placebo MR imaging examination or by not providing the information obtained at MR imaging (31). Because of the expected confusion and opposition of patients with an acute joint injury to such a study design, as well as ethical considerations of not providing information, we chose a nonblinded pragmatic study design.
Part of the pragmatic design of the study was the MR imaging system software upgrade that occurred during the study, which may have improved image quality in the latter part of the study. Because MR imaging software upgrades can be expected in daily practice, we considered the upgrade acceptable in this pragmatic study. The extent of acceptance of the short MR imaging examination by clinicians may also have introduced a bias. We noticed that clinicians were sometimes hesitant to use the results of the short MR imaging examination in their decisions. This acceptance curve of the new technology may have biased the outcome of the study in favor of the reference strategy. Nevertheless, we did find that if a short MR imaging examination had been performed in the acute setting, patients were less likely to undergo further examinations during follow-up.
In this study, a short MR imaging protocol was used to make the examination time acceptable for routine use in the emergency department, as well as to reduce costs. A shortening of the duration of the MR imaging examination comes at the cost of some loss in image quality, but in a pilot study (32), we found that the image quality was acceptable to demonstrate most clinically important lesions. It is likely that, by using this short MR imaging protocol, some injuries will be missed that would have been detected if a standard protocol had been used. It is therefore important to realize that the goal of the short MR imaging examination was not to replace standard MR imaging but to diminish diagnostic uncertainty at the initial evaluation of the patient, thereby resulting in a decrease in total costs and in a potential gain in effectiveness. If the lower quality of the short protocol MR images results in a wrong diagnosis, this would likely lead to an increase in costs and a decrease in effects, including a decrease in quality of life. If the overall effect of an initial diagnostic strategy, including routine short MR imaging examination, results in an overall decrease in costs without loss of effectiveness, or even an increase in effectiveness, the new strategy is worth considering.
Although we could not demonstrate with statistical significance that the application of a short MR imaging examination in all patients with knee injuries is cost saving from a societal perspective, our results suggest that this might well be the case because the overall costs were considerably higher in patients with knee injuries if MR imaging was not performed. For patients with wrist injuries, the strategy, including the short MR imaging examination, increased the total costs without a change in effectiveness. For patients with ankle injuries, the application of MR imaging did not change the total costs and had little influence on the effects.
In conclusion, compared with radiography alone, MR imaging in patients with acute wrist or ankle injury is neither cost saving nor more effective in expediting the work-up or improving quality of life, but, in patients with acute knee injury, MR imaging shortens the diagnostic work-up, reduces the number of diagnostic procedures during follow-up, improves quality of life in the first 6 weeks, and, although not statistically significant, may reduce costs associated with lost productivity.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: SF-36 = Short Form 36
See Materials and Methods for pertinent disclosures.
Author contributions: Guarantor of integrity of entire study, M.G.M.H.; study concepts and design, all authors; literature research, J.J.N.; clinical studies, J.J.N., E.H.G.O.; data acquisition and analysis/interpretation, all authors; statistical analysis, J.J.N.; manuscript preparation and definition of intellectual content, all authors; manuscript editing, J.J.N.; manuscript revision/review and final version approval, all authors
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. Groot Koerkamp, J. J. Nikken, E. H. Oei, T. Stijnen, A. Z. Ginai, and M. G. M. Hunink Value of Information Analysis Used to Determine the Necessity of Additional Research: MR Imaging in Acute Knee Trauma as an Example Radiology, February 1, 2008; 246(2): 420 - 425. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
