Economic Analysis of MRIs
Baker, Laurence and Shreibati, J. B.. 2011. The relationship between low back magnetic resonance imaging, surgery, and spending: Impact of physician self-referral status. Health Services Research, Vol. 46 no. 5, page(s) 1362-1381
Background
The main question here is what is the relationship between having MRI procedures in a physician's own office, and whether or not this will lead to changes in their decision making and in the amount of spending on lower back pain versus referring patients outside of their practice for MRI.
Intro
The advent of medical imaging has introduced valuable information to patient care but may be overused and results in subsequent procedures which are unneeded and costly with little to no tangible benefit. A similar conundrum is that people with more serious health conditions are more likely to receive treatments and imaging will, in turn, bias the analysis. According to (Deyo et al. 2009) lumbar MRI use increased 300 percent from 1994 to 2006 while imaging for the lower back has not been shown to improve clinical outcomes. Before going into surgery for back pain, patients often receive an MRI to find the abnormality that is causing this pain; this is not a perfect science as that abnormality may not be the cause of the pain and just a strange coincident abnormality.
Identification
The focus of this paper is how an outpatient experience will change when seeing an orthopedic surgeon or a primary care physician as these people deal the most with patients experiencing lower back pain. The only people considered where ones who had a diagnosis of lower back pain with no previous claims related to back pain(using [ICD-9] codes for classification). Specifically non specific low back pain as MRI and surgery for this type of back pain is not recommended.
Statistical Strategy
To analyse the effects of MRI imaging, the surgery rates and spending in patients before and after MRI use was compared. This will see whether or not there is a shift of MRI use and how this is associated with spending and health characteristics. The following regression was used for changes in MRI use.
The following regression was used to estimate the effects of MRI use on surgery utilization and spending.
Results
Patients of acquiring physicians, when accounting for differences in age, gender and race, had similar demographic characteristics, prior year health spending, and percent medicare enrollment as compared to before and after MRI imaging. This suggests that the patients in the traditional as compared to MRI imaging groups had a similar disease burden. After accounting for some correlation, it was found that there was no evidence of an effect of receiving MRI and having surgery for primary care doctors. This does not hold true for orthopedists where the probability of surgery increased 34 percentage points within 6 months of an MRI. As for changes in spending, the result for orthopedic patients for receiving an MRI was associated with an increase of USD $4,161 which accounts for the increased percentage of surgeries and MRI imaging. Estimates for primary care patients, and outpatient locations (Hospitals and Emergency) are non significant. Finally, Primary care physicians who did MRI images in house rather than referring them out where found to use MRI more for lower back pain which is associated with higher spending and more surgeries in orthopedic surgeons offices.
Conclusion
Having MRI imaging at a non primary care doctor increases cost, frequency of scanning, and leaves the patent with the same disease burden. There is no health benefit to the increased cost according to the data.
-----------------------------------------------
Baker, Laurence. 2010. Acquisition of MRI equipment by doctors drives up imaging use and spending. Health Affairs (Project Hope), Vol. 29 no. 12, page(s) 2252-2259
Background
In the early 2000's many neurologists and orthopedists got MRI machines. Part of the concern here is the changes in imaging use and whether or not overall spending has gone up. Under careful observations it would appear that this medical equipment is being used inappropriately. Another important concern is to what extent are physicians influenced by financial gains to use the expensive equipment that they have.
Intro
The change initially started in the mid 1990's with more and more physicians owning MRI machine. But even as recently as 2005, these people only accounted for 20% of the whole MRI market. The benefits of having an MRI machine is obvious as it is more convenient, the physicians can bill the insurance companies. Backed by several other articles, it is shown that the frequency of MRI's have increased over the years and as a result MRI reliant medical procedures. This also hold true for medicare.
Identification
Some of the data in this article is from medical providers, specifically from 1998-2005 medicare claims records. From this, the physicians who put claims for an in house MRI were recorded. Following this, each physician was searched for when a patient had an outpatient visit with the physician after 1 year, only applying for patients who have not seen that physician for one year. To study spending, total medicare spending was recorded for 90 days and 1 year following the index visit. To compare to a base rate, the data was compared to physicians with no records of use of an MRI machine.
Statistical Strategy
The analysis of the data focused on the before and after changes of MRI use of the physicians who had begun billing for MRI in house. This was compared to changes over time in traditional MRI uses. OLS was used to make a regression for change in MRI use. The model controlled for characteristics including age, sex, race, prior year healthcare spending, medicare status, the presence of comorbidities, and some other clinical classifications. The overall time trends were controlled by the year and month.
Results
One of the initial results is that the patients characteristics did not change as compared to before and after the MRI machine was installed, as they shouldn't. One of the exhibits show that the rate of MRI use increased from 1999 to 2005 for both neurologists and orthopedists. When testing for MRI use after starting to bill for it, it was shown that for orthopedists it increased 38%. This increase was mostly evident 30 days after the index visit. Even 18 months after acquiring MRI machines, the physicians used the machine more than their colleagues. The effect of spending was also there, there was a 2 percent increase in overall spending per episode. The two components that resulted in increased overall spending was increased spending on MRI and increased spending on physician procedures. After 90 days the outpatient spending for neurologists increased by 6% after being billing for MRI.
Conclusion
By installing an MRI machine, the use of it increased by 38% and at the same time the average cost goes up on average for all patients by approximiatley 2%.
-----------------------------------------------
Hillman Bruce J., Catherine A. Joseph, Michael R. Mabry, Jonathan H. Sunshine, Stephen D. Kennedy and Monica Noether. 1990. "Frequency and Costs of Diagnostic Imaging In Office Practice: A Comparison of Self-Referring and Radiologist Referring Physicians." New England Journal of Medicine 323:1604-1608.
Intro
There are conflicts of interest when it comes to the increased rate and cost of MRI imaging and there has been some government legislation enacted when it comes to the ownership of MRI imaging by the physician. Differing from previous studies which had methodological faults, small populations, and a lack of control this study includes a large database of private insurance claims and controlled for the difference in patients conditions and the physician specialities.
Identification
Getting reliable data was done by buying access to a database of the insurance information of a half million employees of several large american companies and their dependents. This was chosen due to the uniformity and completeness of the data, the fact that there was no copayments, and the age groups were not heavily weighted to an older age group. The specific events which were categorized and analyzed were acute upper respiratory symptoms, pregnancy, lower back pain, and difficult urinating for men with specifically questioning whether radiography, ultrasonography, cystography, or urography was performed.
Using the (ICD-9-CM) classification codes, the previously mentioned events were picked out. Using the index date, the patient care was only considered for a specific date date: four weeks for respiratory, 9 months for pregnancy, 6 weeks for lower back pain and 6 weeks for difficulty urinating. The specific dates needed to begin after January 1, 1986, and end before June 1, 1988. The following physician categories were considered self referring physicians,and radiologist referring physicians only for the 1 physician episodes.
Statistical Strategy
By applying maximum likelihood methods to the information we derived from our data we can adjust the observed frequencies to account for the physicians who did no imaging. This is based on the assumption that the data was homogeneous which is not guaranteed. To account for the heterogeneity in the physicians imaging practices, estimated biased up and down was developed to show the results aren't affected qualititatively by the adjustment of the physicians who did no imaging. For the analysis of the 1 physician and 2 physician episodes the differences were assessed between the self referring and other physicians for the proportion of images, and average charges. The t-statistic was done to find the significance of the differences between the two different types of radiologists and the null hypothesis was rejected at a p-level of .05.
Results
For the one physician episodes, the mean imaging charges was significantly higher if the self referring physicians was observed as compared to the radiologists for all events except for difficulty urinating. Depending on the clinical presentation, the self referring physicians resulted in imaging occuring 4-4.5 times as frequently, with average charges per episode 4.4-7.5 times higher than the radiologist. For the two physician episodes, the findings follow the one physicians episodes as they are 1.7-3.7 times as likely to perform imaging. Depending on the specific specialty and the clinical presentation, imaging occured 2.4-11 times as likely and was 3-17 times as expensive.
Conclusion
This data set is much more detailed and accounts for various ailments and different imaging methods. The data support the conclusion that by referring the patient to an in house scanning, the cost will be increased and the frequency is also increased. There is no mention of the effect of the increased rate of imaging and the effects of this on health.