J Manipulative Physiol Ther 2005 (Oct); 28 (8): 564–569 ~ FULL TEXT
Nelson CF, Metz RD, LaBrot T
Health Services Research,
American Specialty Health,
San Diego, CA 92101, USA.
OBJECTIVE: The aim of this study was to measure the effects of a managed chiropractic benefit on the rates of specific diagnostic and therapeutic procedures for the treatment of back pain and neck pain.
DESIGN: This study is a retrospective analysis of claims data from a managed-care health plan over a 4–year period. The use rates of advanced imaging, surgery, inpatient care, and plain-film radiographs were compared between employer groups with and without a chiropractic benefit.
RESULTS: For patients with low back pain, the use rates of all 4 studied procedures were lower in the group with chiropractic coverage. On a per-episode basis, the rates in the group with coverage were reduced by the following: surgery (–32.1%); computed tomography (CT)/magnetic resonance imaging (MRI) (–37.2%); plain-film radiography (–23.1%); and inpatient care (–40.1%). On a per-patient basis, the rates were reduced by the following: surgery (–13.7%); CT/MRI (–20.3%); plain-film radiography (–2.2%); and inpatient care (–24.8%). For patients with neck pain, the use rates were reduced per episode in the group with chiropractic coverage as follows: surgery (–49.4%); CT/MRI (–45.6%); plain-film radiography (–36.0%); and inpatient care (–49.5%). Per patient, the rates were surgery (–31.1%); CT/MRI (–25.7%); plain-film radiography (–12.5%); and inpatient care (31.1%). All group differences were statistically significant.
CONCLUSIONS: For the treatment of low back and neck pain,
the inclusion of a chiropractic benefit resulted in a reduction in the rates of surgery, advanced imaging, inpatient care, and plain-film radiographs. This effect was greater on a per-episode basis than on a per-patient basis.
From the Full-Text Article:
In all study categories, there were statistically significant reductions in the rates of surgery, advanced imaging, inpatient care, and plain-film radiography. Previously published data from this study showed that chiropractic care was used almost entirely as a substitution for medical care for back pain and neck pain complaints. [27, 28] The clinical treatment of back and neck pain complaints tends to be highly variable. [29–33] Specifically, the use rates of the procedures here under investigation tend to vary significantly in different care delivery systems, and these rates are not driven solely, or even principally, by clinical variables. Given the amount of discretion that is exercised in the use of these procedures, it is not surprising to find that when the option of accessing chiropractic care is provided, it would produce the reductions seen in this study.
The difference in rates is greater as measured on a per-episode basis than on a per-patient basis. Also, as demonstrated in a previous publication, chiropractic care (which is only present in cohort A1) tends to generate more episodes of care per patient than medical care and thus artificially further reduces the rate per episode. The per-patient comparison is the more valid and meaningful because it captures the actual probability of an individual patient's likelihood of receiving any of the specified interventions.
It is also notable that the rates of plain-film radiographs are lower in the group with chiropractic coverage. Unlike the other 3 procedure categories, chiropractors themselves provide plain-film radiographs. Previous studies have shown that chiropractors may use plain-film radiographs at a higher rate than medical physicians, and thus, the substitution phenomenon might produce an increase rather than a decrease in these rates.  The reductions seen here are attributable to specific treatment policies designed to reduce the rates of use of radiographs in chiropractic care and may not reflect the rate of use in an unmanaged system.
Given the significant differences in the rates of procedures found in this study, it is important to consider which rates represent a more optimal level of care for back pain and neck pain. These differences exist within the same geographic area and between reasonably similar patient groups. It is a plausible assumption that if the same standard of care were to be applied to these two groups, very similar use rates would result. However, it is not possible to answer this question directly because there are simply no clinical outcomes measured in this study.
There are also no standard benchmarks for use rates against which these rates can be compared. However, there is a large body of scientific literature and guidelines on the treatment of back pain that provides some insight into this question. This body of evidence coalesces around 3 recurrent themes:
1) most back pain episodes are categorized as “mechanical” or “uncomplicated” and do not require aggressive interventions,
(2) existing guidelines specify that less aggressive and invasive diagnostic and therapeutic treatment of back pain will lead to better clinical outcomes, and
(3) there is poor compliance with these guidelines and departures from the guidelines are most likely in the direction of overuse of invasive procedures.
Given these observations, it is probable that the rates of use of these procedures are greater than optimal in both groups and any change in the direction of decreased use may result in positive effects on health-care costs, outcomes, and patient safety. The literature on neck pain is less comprehensive than that on low back pain, but it also points in the same direction. [34–37] The conclusion that the presence of a chiropractic benefit results in more appropriate use diagnostic and therapeutic procedures is suggested but not proven by this study.
The cohort with chiropractic coverage was slightly younger with slightly fewer comorbidities than the cohort without chiropractic coverage. These differences may have contributed to the reductions in procedure rates. However, the magnitudes of these demographic and clinical differences are quite small compared with the magnitudes of the reductions, and it is unlikely that they account for more than a small percentage of the changes.
The results of this study cannot be generalized beyond the specific health care systems involved in the study. It is well understood that the use rates of specific diagnostic and therapeutic procedures are highly dependent upon variables such as local practice habits, economic incentives, and other health plan characteristics. A different set of these variables would undoubtedly produce a different result. However, the direction of these results (a reduction in rates among those with chiropractic coverage) might be more robust relative to these variables. As long as chiropractic care is being used as a substitute for medical care and as long as chiropractors do not directly provide these procedures (with the exception of plain-film radiographs), it is likely that a reduction in advanced imaging, surgery, and inpatient care would be seen. Finally, there are no clinical outcomes measured in this investigation. No inferences can be made on the relative clinical benefits of the different procedure rates except as suggested by the existing literature on this subject.
Among employer groups with chiropractic coverage compared with those without such coverage, there is a significant reduction in the use of high-cost and invasive procedures for the treatment of low back pain and neck pain. The presumed mechanism of this effect is the substitution of chiropractic care for medical care for the treatment of back and neck pain. The resultant chiropractic care is far less likely to lead to the use of these invasive procedures. This reduction is more pronounced when measured on a per-episode basis than on a per-patient basis.