NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Santaguida PL, Gross A, Busse J, et al. Complementary and Alternative Medicine in Back Pain Utilization Report. Rockville (MD): Agency for Healthcare Research and Quality (US); 2009 Feb. (Evidence Reports/Technology Assessments, No. 177.)

Cover of Complementary and Alternative Medicine in Back Pain Utilization Report

Complementary and Alternative Medicine in Back Pain Utilization Report.

Show details

3Results

From an initial inclusion of 8,347 unique citations, 8,108 were excluded as they were not on topic, not English language publications, or not undertaken within the countries of interest (see Figure 2). The remaining 239 were identified as potentially being about complementary and alternative medicine (CAM) and about utilization, costs, patient or provider views, recommendations about optimal frequency and duration of visits, or clinical practice guidelines (CPG).

Figure 2. Flow of publications from initial search to final eligibility.

Figure

Figure 2. Flow of publications from initial search to final eligibility.

The full text publications for these 239 citations were retrieved and from these 95 were excluded as not addressing utilization, costs, or recommendations for use of CAM. Following full text screening, we further excluded non-U.S. publications for cost-related outcomes (n = 19), as comparability with American costs for CAM use would be limited. Our complete eligibility criteria were met by 103 publications for utilization, recommendations, and costs related to CAM use.

We partitioned CPG reports or publications related to recommended use of CAM (Figure 2). From these, eight CPG were excluded as not providing any specific recommendations on CAM use for back pain. Additionally we separated primary publications (n = 22) of provider or patient views about CAM utilization. Eighteen of these publications on provider or patient views were excluded as these did not provide information on “recommended” use of CAM therapies.

We grouped eligible publications according to their primary focus on utilization, recommendations for frequency of use, or costs associated with use of CAM; these publications were further divided into those undertaken within the U.S. and those from the other countries eligible for this systematic review. Figure 2 shows the overlap.

Question 1. What is the Relative Utilization for the Different CAM Therapies?

A total of 94 publications provided some information on utilization specifying either the prevalence of visits to CAM practitioners or more detailed service use based on specific therapies. These include publications based both within the U.S. and within other eligible countries including: Canada, United Kingdom, Europe, Australia, and New Zealand. Approximately one-third of the publications presented utilization of CAM but did not stratify these results specifically for persons with back pain. The remaining publications were partitioned to address: 1) relative utilization of CAM therapies evaluated concurrently, 2) trends over time, and 3) individual CAM therapy utilization for the most prevalent provider-based CAM therapies. There was significant overlap of publications among the research questions addressed in this review. The majority of publications (52 of 65) reported some aspect of use related to chiropractors, or chiropractic, or manipulation, or mobilization. Despite the overlap we summarize findings and quality assessment for each section and for the utilization publications as a whole.

Utilization of CAM, Data not Stratified by Back Pain or Therapy

Twenty-nine publications did not stratify CAM utilization data with respect to use by persons with back pain. Rather than exclude these publications, we chose to identify and detail some aspects of these studies; many of these publications are recognized as contributing significantly to our current understanding of CAM utilization. Our intention was to highlight that these studies were lacking in utilization results specific to persons with back pain, despite being the sources that originally identified back pain patients as a key population that utilize CAM services. Similarly, we wished to highlight that several of the non-stratified studies evaluated different subgroups of persons that experience significant back pain related to spinal cord injuries, cancer, or pregnancy. Nineteen publications6, 47, 71, 76, 91105 were based on utilization data from the U.S. and 10 publications106115 were from other eligible countries.

Utilization not Stratified by Back Pain Within the United States

Several American studies are seminal works based on large national samples6, 71, 76, 98, 99, 102, 105 that surveyed CAM utilization and established that back and neck pain were important reasons for seeking CAM treatments. However, the utilization data presented did not link the specific type of CAM therapy to persons with back pain; rather the prevalence of back pain within the samples was reported. Four publications101104 were based on the same study cohort, representing survey results from a national sample.

Three publications focused on persons with spinal cord injuries. Two of the publications92, 93 on spinal cord injury patients showed that back and neck pain was frequently a major source of their chronic pain (varying from 46 to 75 percent), but did not present stratified CAM utilization information. One study92 reported utilization of CAM combined with medical injections to manage pain.

One study91 evaluated the use of CAM for cancer and other chronic pain in American Veterans. Other publications identified CAM use within patients presenting to the emergency department,95 a specialized CAM clinic within a university hospital,100 or concurrent use of CAM in patients attending outpatient physiotherapy departments.96 Another study97 indicated the use of mind body therapies among patients with musculoskeletal pain. One study47 assessed the use of CAM by pregnant women; although rates of back pain are very high within this population, the publication did not specify the trimester or whether or not the whole sample was experiencing back pain at the time of the study.

Summary of U.S. publications on utilization not stratified by back pain. Nineteen publications were based on utilization data from the U.S. Although, these establish that back pain was an important symptom for which patients sought services, they do not present results on utilization of CAM therapies stratified for persons with back pain. Several studies evaluated subgroups of patients with non-musculoskeletal related back pain and would suggest that CAM therapies are commonly used by these populations.

Utilization not Stratified by Back Pain Within Other Countries

There were 10 publications from other eligible countries that did not present utilization data stratified for persons with back pain. One study106 evaluated neck and shoulder pain in newspaper workers, another the general population in Canada.115 Three publications108, 110, 111 evaluated physiotherapy outpatients and another107 patients at osteopathic clinics. One study109 evaluated the use of a chiropractic activator and another, general population with back pain in the United Kingdom.114 Another study evaluated persons with spinal cord injuries in Sweden.112 Finally, one study evaluated the general population in Spain.113

Summary of non-U.S. publications on utilization not stratified by back pain. Ten publications were based on utilization data outside the U.S. No clear pattern emerges from these studies; there are differences in populations, providers, and CAM therapies evaluated that contribute to this heterogeneity.

Relative use of CAM Therapies

We identified publications that presented the relative frequency of CAM utilization in order to rank the most to least frequently used therapies by persons with back pain. We identified four U.S.912 and four non-U.S. publications116119 that reported utilization data for a minimum of four different CAM therapies or practitioners. An additional study30 reporting more than four CAM therapies provided solely by a physical therapist, was not included in this grouping. Electrotherapies provided by conventional providers (such as physical therapists) were defined as non-CAM therapy; however, this publication provided utilization data for spinal manipulation and is detailed in the specific CAM therapy section below.

We did not select publications where subjects were recruited from the practices of CAM providers. The utilization patterns described within these studies would reflect practice patterns of concurrent CAM therapies and not that of relative utilization of CAM services. For example, utilization of differing CAM therapies reported from subjects currently receiving care by a chiropractor would reflect the manner in which this particular CAM practitioner combines various CAM modalities rather than the frequency with which persons with back pain would choose to use one CAM therapy over another. The manner in which CAM therapies are concurrently combined for treatment is described within the specific CAM therapy sections that follow.

Table 1 compares reports of the use of four or more different CAM therapies for back pain. Note that the proportion of use was not exclusive to a single category; most subjects reported using more than one CAM therapy for lifetime use. What was generally not specified within these publications was the order in which CAM therapies were selected for single or multiple episodes of care.

Table 1. Studies listing the use of 4 or greater different CAM therapies specific to back pain.

Table 1

Studies listing the use of 4 or greater different CAM therapies specific to back pain.

Relative Utilization of CAM Therapies Within the United States

Of the four U.S. publications for which relative utilization could be evaluated, all but one9 measured lifetime use of CAM (Table 1). All subjects within these publications had chronic or recurrent back pain. The sample sizes for those seeking care varied from 186 to 2,374 subjects. A single study9 reported the use of CAM therapies for a subgroup of the chronic back patients (69 percent seeking care within the last 6 months); the other publications reported on CAM use for the entire study sample with back pain. One publication12 reported on combined neck and back pain and the remaining ones on LBP. A single publication12 was based on a randomly selected population sample and another on North Carolina residents.9 The remaining publications were based on patients with chronic pain.10, 11 All publications evaluated utilization with respect to the proportion of patients who used any of several CAM therapies and all but one study10 reported on the percent accessing at least one CAM provider.

Table 1 shows the relative utilization rates across the four U.S. publications. Chiropractor/chiropractic was the CAM therapy that was most frequently used by patients with chronic LBP or combined neck and back pain.1012 In one study9 massage was the most frequently used CAM therapy; massage was also highly utilized in the other publications ranking second most frequent (Table 1). Although the use of ultrasonography, electrotherapy, and heat and cold modalities was high (18 to 68 percent), the provider of these therapies was not specified and treatment may have been provided by non-CAM practitioners. Other frequently used modalities were prayer and spirituality (27 to 42 percent) and glucosamine (26 to 46 percent); but these were reported in only two of the four publications. Both prayer and glucosamine were likely to be self-administered. It is probable that the empty cells within Table 1 can be accounted for by differences in study protocols with regards to which, of many potential CAM therapies, were probed.

Three publications911 specified that there was overlap in use of the various CAM therapies, but type of therapy overlap was not detailed. One study12 indicated that more than one CAM practitioner may have provided the therapy in the estimates of utilization. Similarly, overlap with conventional providers was not stratified for specific CAM therapies; a single study specified overlap with chiropractors alone but not other CAM practitioners.9

Summary of relative utilization of CAM therapies in persons with back pain in the U.S. There are a limited number of publications (n = 4) based on population-based samples that show the relative use of CAM therapies in persons with back pain. Two of these publications recruited subjects from large general population-based samples, and two recruited samples with chronic LBP. In general, these publications show that chiropractic/spinal manipulation is the most frequently utilized provider based CAM therapy followed by massage and acupuncture. Prayer and herbals and food supplements were also frequently used (self-practiced) CAM therapies although they were only evaluated in half the publications. Differences in study protocols for probing the type of CAM therapies used likely account for missing utilization data from the range of CAM services available in Table 1. Rates of utilization may also vary because of differences in populations and in location of back pain. The methodological quality of these four studies is generally good with minimal biases and valid results.

Relative Utilization of CAM Therapies Outside the United States

Three publications116118 evaluating CAM use in Canadians and one119 in Australians provided information on at least four different therapies (Table1). These publications evaluated a combination of neck and back pain,116, 117 unspecified back pain,118 or LBP.119

Two companion publications116, 117 reported utilization with respect to the proportion of back pain patients using specific CAM therapies rather than the proportion relative to all persons with back pain. Three publications116118 found chiropractic to be the most frequently used modality/practitioner, with massage in second place (Table 1). In another study119 massage was the most frequently used CAM therapy. Acupuncture was the third most frequently used modality in two studies.116118 The use of prayer and spirituality was reported in three of the publications but varied significantly. The use of herbals and food supplements was reported in a single study.116, 117 These resources were excluded from the operational definition of CAM within another study.118

Two related publications116, 117 did not specify complementary or combined CAM therapy use. Two publications118, 119 evaluated patients with back pain alone and presented the degree of overlap with other CAM therapies. One of these publications118 showed a degree of overlap with conventional providers that varied from 6 percent with any physician to 1 percent with specialists; similarly, overlap with other conventional providers varied from 2 percent with physiotherapists to 0.3 percent with psychologists.

Summary of relative utilization of CAM therapies in persons with back pain. There are four publications based on population-based samples that show the relative use of CAM therapies in persons with back pain outside the U.S. These publications reported predominately on combined or unspecified back pain of variable duration. All publications undertaken outside the U.S. were based on larger, randomly selected, population-based samples. Three of these publications show that chiropractic/spinal manipulation is the most frequently utilized provider-based CAM therapy followed by massage and acupuncture. Prayer and herbals and food supplements were also frequently utilized CAM therapies although the rates of use were highly variable. Differences in study protocols with regards to probing the type of CAM therapies used likely account for missing utilization data from the range of CAM services available in Table 1. Rates of utilization may also vary because of the differing populations, as duration and location of back pain were combined or not specified in these studies. In general, these cross-sectional studies are rated as high quality with few biases and valid results.

Trends in CAM Utilization

We identified 10 publications within the U.S. 5, 19 and other countries112, 117, 120125 containing data on the trends across time in the utilization of CAM interventions for back and neck pain. These papers were published between 1996 and 2007 and examined trends between 1962 and 2006. They had various designs and included patients with neck and/or back pain. Specifics regarding these publications are located in summary Tables 2 and 3.

Table 2. Summary table of studies evaluating trends in utilization of CAM therapies over time for persons with low back pain.

Table 2

Summary table of studies evaluating trends in utilization of CAM therapies over time for persons with low back pain.

Table 3. Summary table of studies evaluating trends in utilization of CAM therapies over time for persons with combined neck and back pain.

Table 3

Summary table of studies evaluating trends in utilization of CAM therapies over time for persons with combined neck and back pain.

Trends in Utilization Within the United States

For those with LBP only, study19 showed that in 1987, 41 percent used chiropractic medicine and in 1997, 31 percent. Another study found that in 1990, 36 percent of individuals with unspecified back pain used a CAM therapy while 20 percent saw a CAM practitioner; in 1997, 48 percent used CAM for back pain while 30 percent saw a CAM practitioner.5 This same study found that in 1997, 57 percent of individuals with neck pain used a CAM therapy and 37 percent saw a CAM practitioner.5

Summary of trends in CAM utilization within the U.S. The limited literature on trends over time would suggest that utilization of chiropractic/spinal manipulation for LBP in the U.S. has decreased, while the use of CAM generally has increased, although data on specific CAM therapies was not reported. More research is required to highlight trends in the utilization of CAM therapies for back pain in the U.S.

Trends in Utilization Outside the United States

One study of LBP patients in the United Kingdom found that osteopathy was used by 49 percent of patients over an 18 month period.122 Note that osteopathy practices outside of the U.S. may differ in the types of CAM therapies administered. Another study found that over a 4-year period acupuncture was used for LBP by 9 percent of patients from teaching hospitals and 32 percent of patients from general hospitals.125 One study from the Netherlands found that between 1989 and 1992, physical therapists (PT) administering massage and passive mobilizations were used by 12 percent of those suffering from LBP. This increased to 15 percent in 2002 to 2003.121 A Canadian study found that between 1996 and 2002 glucosamine was used by 5 percent of men and 10 percent of women with back pain.120 A study from Norway found that in 1992, 59 percent of non-referred and 62 percent of referred chiropractic patients had LBP.124

For those suffering from neck and back pain, in Canada, “in the past 12 months” the following utilization rates of CAM therapies were reported: chiropractic care 61 percent; massage 39 percent; acupuncture 30 percent; energy healing 16 percent; yoga 14 percent; relaxation techniques 10 percent; and prayer/spiritual practice 5 percent.117 In Denmark, between 1962 and 1999 patients visiting chiropractors increased by over 7 percent for the primary complaint of LBP alone, decreased by almost 9 percent for LBP with sciatica; increased by over 3 percent for neck pain, and remained essentially the same for neck and arm pain.123

Summary of trends in CAM utilization outside the U.S. The limited literature on trends in countries outside the U.S. would suggest that utilization rates differ between countries and between specific CAM interventions. More research is required to determine the extent of these differences. No data were reported on the usual costs for these therapies per treatment or for the prescribed course of treatments.

Specific CAM Therapy Utilization: Acupuncture

Eighteen publications provided data on utilization of acupuncture by back pain patients. Of these, eight were specific to the U.S.;1012, 15, 16, 18, 44, 45 two of these publications were based on the same sample.15, 44 Ten publications provided utilization outcomes from other eligible countries;51, 116119, 125129 one publication117 was a companion to a previous study116 as it incorporated the findings of the initial survey and utilization from a resampling of new subjects 6 years later. The variation in the samples recruited for these studies is noteworthy and includes subjects from large population-based surveys or administrative databases and subjects from small clinical practices, or populations limited only to patients with back pain. As expected, the estimates of utilization varied substantially based on the type of denominator used to estimate prevalence.

Acupuncture Utilization Within the United States

There were eight U.S. publications that provided data on acupuncture utilization in persons with back pain. Three of these were of cross-sectional design using interviews.12, 18, 45 Three publications from two study cohorts used a single group prospective cohort study design based on interviews15, 44 and self-administered questionnaires.11 One study was a retrospective design using data from a health claims database16 and another was a randomized controlled trial10 that used a self-administered questionnaire to evaluate use of CAM.

Two publications12, 45 reported CAM utilization based on samples from the general population. Two publications recruited subjects from health claims databases in Washington state.16, 18 Two publications15, 44 randomly sampled licensed acupuncturists and patients in their practices from seven states, including Washington and Massachusetts. One study recruited subjects from large urban specialty clinics for orthopedic or neurosurgery consultation11 and another recruited a small sample of back pain subjects who had access to the internet.10

Utilization was primarily defined as a visit to an acupuncturist or having received or “ever tried” acupuncture based on self-report with or without information from health records. Three publications queried use “within the last 12 months”,10, 12, 45 three11, 18, 45 reported “lifetime” use, and two15, 44 reviewed visits to practitioner over a 1-year interval.

Only one of the publications15 provided any detail as to the type of acupuncture administered. This same study was the sole publication to provide information about the mean duration of the visits (60 minutes) but it did not specify the frequency of visits over the course of treatment. The training or type of acupuncturist was specified as “licensed acupuncturist”15, 18, 44 in three publications; the remaining eligible publications provided no further details about the practitioner.

Acupuncture utilization based on anatomical region. Tables 4A to 7A show utilization of acupuncture in the U.S. specific to neck pain (n = 2), unspecified back pain (n = 4), LBP (n = 3) and combined back and neck pain (n = 1). Neck pain utilization was reported in two publications44, 45 from general population samples and the use of acupuncture varied from 7 to 14 percent of persons with neck pain of unspecified severity and duration. Four publications10, 15, 44, 45 reported utilization for unspecified back pain (Table 5A) and found widely varying rates for the use of acupuncture, from 2 percent (recurrent but not disabling back pain from small sample) to 34 percent (severity and location not specified).

Table 4A. Utilization of acupuncture in persons with neck pain.

Table 4A

Utilization of acupuncture in persons with neck pain.

Table 7A. Utilization of acupuncture in persons with combined neck and back pain.

Table 7A

Utilization of acupuncture in persons with combined neck and back pain.

Table 5A. Utilization of acupuncture in persons with unspecified back pain.

Table 5A

Utilization of acupuncture in persons with unspecified back pain.

Table 5B. Utilization of acupuncture in persons with unspecified back pain.

Table 5B

Utilization of acupuncture in persons with unspecified back pain.

Similarly, the three publications11, 16, 18 evaluating the use of acupuncture for LBP patients (Table 6A) showed rates ranging from 3 to 11 percent of the total sample of subjects; two of these publications11, 18 indicate the back pain was chronic in nature. One study12 (Table 7A) reported acupuncture use for patients with combined neck and back pain and showed a utilization rate of 0.9 percent (95 percent CI, 0–1.9), and an estimated total of 1.6 million visits (mean of 2.6 visits); this study also reported no difference in the relative frequency of CAM use by location of the back pain. However, a single study45 that segregated rates for persons with neck and back pain found that acupuncture was used more frequently for back pain than for neck pain (34 versus 14 percent).

Table 6A. Utilization of acupuncture in persons with low back pain.

Table 6A

Utilization of acupuncture in persons with low back pain.

Table 6B. Utilization of acupuncture in persons with low back pain.

Table 6B

Utilization of acupuncture in persons with low back pain.

It is important to note whether the denominator used to estimate utilization was based on the entire sample enrolled or just a specific subset of patients with back pain. Two studies12, 45 were based on general population samples and, not surprisingly, showed markedly lower rates of acupuncture utilization for both neck and unspecified back (1 to 4 percent) and combined neck and back pain (1 percent) than studies that recruited smaller samples of persons all of whom had back pain.

Acupuncture use as complementary care or combined with other CAM. The majority of publications did not collect or report sufficient information to determine concurrent use of conventional with CAM therapies for back pain. Two publications provided information on use with conventional care but were not specific to those respondents with back pain44 or specific to those receiving acupuncture12 (Table A2 and A4). Although not stratified for persons with back pain, up to 53 percent of acupuncture patients saw a medical or osteopathic physician, but only 10 percent of persons with back pain reported speaking to their doctor about using acupuncture.44 Another study12 suggested that CAM alone had been used by 29 percent of persons with back and neck pain; however, this was not specific to those receiving acupuncture.

Most studies did not report concurrent or combined use of different CAM therapies in a detailed manner. Overall, when conventional or combined therapy details were reported, the results were not stratified with respect to using acupuncture or specifically for those with back pain (Tables 4B to 7B). There was one exception15 and this publication indicated that acupuncture was used predominately with other CAM therapies such as east Asian massage, cupping, heat, and herbs.

Table 4B. Utilization of acupuncture in persons with neck pain.

Table 4B

Utilization of acupuncture in persons with neck pain.

Table 7B. Utilization of acupuncture in persons with combined neck and back pain.

Table 7B

Utilization of acupuncture in persons with combined neck and back pain.

Quality assessment of U.S. acupuncture publications. Only one10 of the eight U.S. publications on the utilization of acupuncture for back pain used a study design with a true comparator group; as such the focus of our quality assessment is on the potential for selection and reporting biases. All but a single publication16 are based on self-reported utilization and therefore subject to recall bias; heterogeneity with respect to survey versus interview and the types of CAM services queried are also important sources of reporting bias.

Two12, 45 of three cross-sectional publications were of high quality, reporting CAM utilization based on random sampling of the general population with adequate response rates (74 and 63 percent respectively). The third study18 recruited subjects from a health database and then subsequently interviewed them. This study had a low response rate (36 percent) and had the potential for selection bias as the eligible sample had to have back pain (44 percent were ineligible due to no longer having back pain). One study, based on a large health insurance claims database, would also be considered high quality as it used standardized coding for back pain classification, provider visits, and treatments.16

There is greater potential for selection bias in studies using samples obtained from provider clinics. Two publications15, 44 randomly sampled licensed acupuncturists using a recruitment strategy based on the National Ambulatory Medical Care Survey. This methodology collects service visit characteristics close to the time of each patient encounter thereby minimizing longterm provider recall, errors from chart extraction, and retrospective administrative data analysis. The two remaining studies recruited chronic back pain subjects who had been referred for orthopedic or neurosurgery consultation or from urban specialty clinics11 and a small sample of back pain subjects who had access to the internet.10 As noted previously, the rates of use for patients already seeking treatment differ from those in the general population; additionally, those already seeking CAM therapy services may possess inherent differences in defining characteristics that are not easily determined in the absence of comparison groups.

Across all eight U.S. publications, the methodological quality is rated as fair, indicating that they are susceptible to some biases but these, being somewhat endemic to self-report, are not sufficient to negate the results.

Summary of utilization for back pain for U.S. acupuncture publications. Eight publications evaluated the use of acupuncture in persons with back pain and all but one used self-report methods.16 Two of these studies12, 45 were based on randomly selected general population samples and, as expected, these showed markedly lower rates of utilization for both neck and unspecified back (1 to 4 percent) and combined neck and back pain (1 percent). Similar rates of utilization were observed (1 percent) in a study evaluating a large administrative database. All other studies based primarily on subjects from clinical practices or with back pain showed higher rates of utilization (7 to 36 percent). Rates of use also varied by the timeframe for recall (12 months to lifetime prevalence), the interval of analysis for administrative databases, and recruitment of subjects from clinical practices.

The majority of publications evaluated LBP or unspecified pain, with utilization for neck or combined neck and back problems being less commonly assessed. No clear pattern emerges for acupuncture utilization rates as a function of region of the back; in part this is due to the limited number of studies, but it may also be related to the lower prevalence of neck pain relative to LBP.

Only one of the publications15 provided any detail as to the specific type of acupuncture provided. A single publication44 presented information on concurrent conventional care with acupuncture. When reported, conventional care overlap was generally not stratified for acupuncture. Similarly, a single publication15 sufficiently detailed combining of other alternative therapies with acupuncture. Overall these studies were rated as having fair methodological quality, indicating that they are susceptible to some biases. These threats to validity are somewhat endemic to the use of self-report of utilization but are not sufficient to negate the results.

Acupuncture Utilization Outside the U.S

Of the 10 publications on utilization outside of the U.S., three116118 were based in Canada, three125, 128, 129 in the United Kingdom, three119, 126, 127 in Australia, and one51 in Germany.

Six of the 10 publications were of cross-sectional design with utilization outcomes derived equally from interviews116118 or self-administered questionnaires (Tables 4A to 7A).119, 127, 128 Three publications used medical chart data and a cohort study design with non-concurrent (historical) control,129 a single group prospective study design,126 and a before after study design.125 Another study used a single group prospective cohort design51 based on all subjects enrolled in a randomized controlled trial (RCT) and using interview data.

Half the studies recruited participants from the general population, including a national sample within Canada,116118 one within Australia,119 and one of Australian women.127 The remaining studies recruited subjects from specialized outpatient or pain clinics,125, 126, 128 and general medical practices.51, 129

Utilization was primarily defined as a visit with an acupuncturist or having received or “ever tried” acupuncture. In addition to a visit to a CAM practitioner, one study118 included “discussed CAM with, a non-mainstream practitioner” and “looked for a support group when faced with a health problem” as utilization. Half the publications51, 116118, 127 queried use within the last 12 months, and two publications119, 125 reported use within 6 months or less. One study128 did not report a timeframe and another126 reviewed visits over a 6-year interval.

None of the publications provided details as to the type of acupuncture provided. Only two publications indicated that the provider was either a medical acupuncturist129 or was likely a conventional practitioner.125

Acupuncture utilization based on anatomical region. None of the non-U.S. publications provided utilization specific to the neck and the use of acupuncture. In the three publications that did not specify the location of the back pain (Table 5A), rates of acupuncture use varied from 3 percent127 to 29 percent of those with back pain.126 Table 6A shows the five publications that evaluated acupuncture use in persons with LBP, and rates varied from 2 percent119 to 19 percent.129 Two related publications116, 117 (Table A4) that combined neck and back pain reported that 28 to 30 percent of back pain patients used acupuncture. No clear trend emerges for the utilization of acupuncture as a function of back region; neck pain specifically was not evaluated in these eligible studies and the other back regions showed a similar range of rates of use. In general, as one would expect, rates of use of acupuncture were lower in population-based samples than those publications with samples from specialty clinics which tended to be smaller in number. It is this factor, rather than the back region that accounts for the majority of observed variability; the timeframe interval and methods used to collect rates of use are also important factors in explaining variability.

Acupuncture use as complementary care or combined with other CAM therapies. Seven publications did report some information regarding use of conventional care with CAM. Three of these116118 did not stratify the proportions specific to persons reporting back pain and one did not stratify by practitioner type119 (Table 4A and 7A). Three publications51, 125, 128 reported use of acupuncture with conventional care and found rates of overlap varying from 90 percent128 to 63 percent.51 Three publications indicated that patients used medications and injection therapies128 (16 to 75 percent), and mixed therapies (medications, nerve blocks, electrotherapy, and transcutaneous electrical nerve stimulation).51, 125

The majority of publications did not provide detail regarding concurrent or additional use with other CAM therapies. Two publications51, 119 indicated use of acupuncture with other CAM therapies including chiropractic (46 percent) and massage (37 percent).51

Quality assessment for non-U.S. acupuncture publications. Seven51, 116119, 126, 127 of 10 non-U.S. publications used self-reported utilization data and were subject to some degree of recall bias. The remaining three publications125, 126, 129 used medical health record information. Six of the 10 publications were of cross-sectional design with utilization outcomes derived equally from interviews or self-administered questionnaires; of these, five publications116119, 127 were derived from random population samples, with one127 being limited to older Australian women. However, the response rate was low (19 to 26 percent)116, 117 or not reported118, 127 in four of these publications. Three publications using non-comparative study designs recruited patients from general practices129 or specialized pain clinics.125, 126 For these 10 non-U.S. studies the overall methodological quality is fair, suggesting some susceptibility to bias, but not sufficient to negate all the results.

Summary of utilization for back pain for non-U.S. acupuncture publications. Ten publications undertaken in Canada,116118 the United Kingdom,125, 128, 129 Australia,119, 126, 127 and Germany51 evaluated the use of acupuncture in persons with back pain. Three of these125, 126, 129 retrieved utilization data from health records rather than using self-report methods; the quality of utilization data retrieved was limited by lack of standardized extraction methods. Six of the publications51, 116119, 127 using interview or survey methods were cross-sectional studies. Five of these selected random population samples and found utilization rates varying from 2 percent119 to 30 percent.116, 117 Differences in timeframes queried may account for this variation.

Of the 10 non-U.S. publications, five51, 116118, 127 probed use within the last 12 months, and two119, 125 reported use within 6 months or less. One publication128 did not report a timeframe and another126 reviewed visits over a 6-year interval.

The majority of publications evaluated LBP (n = 5) or unspecified pain (n = 3) and none evaluated or reported specifically on neck pain. No clear trend for utilization rates as a function of region of the back was observed in the non-U.S. studies. None of the publications provided details as to the type of acupuncture provided and only two publications specified the type of practitioner.125, 129 Although seven publications did report some information regarding use of conventional care with acupuncture, four of these did not either stratify results by persons with back pain116118 or specify the CAM practitioner;119 the remaining three publications51, 125, 128 showed overlap varying from 63 to 90 percent. The majority of publications did not provide sufficient detail regarding concurrent or additional use with other CAM therapies. Two publications51, 119 indicated use of acupuncture with other CAM therapies including chiropractic (46 percent) and massage (37 percent).51 For these 10 non-U.S. studies the overall methodological quality is fair, suggesting some susceptibility to bias, but not sufficient to negate all the results

Specific CAM Therapy Utilization: Massage

Nineteen publications provided information on the utilization of massage in persons with back pain. Of these, 12 were specific to the U.S.9, 11, 12, 15, 16, 18, 20, 22, 23, 30, 33, 44 Seven of these publications are based on three patient study cohorts from Oregon State medical and chiropractic practices,20, 22, 23 the practices of licensed CAM practitioners in Connecticut and Washington,15, 44 and North Carolina.9, 30 Eight publications provided utilization outcomes from eligible non-U.S. countries;51, 116119, 121, 130, 131 one publication117 was a companion to a previous study116 as it incorporated the findings of the initial survey and utilization from a resampling of new subjects 6 years later.

Massage Utilization Within the U.S

Three of the 12 eligible U.S. publications used cross-sectional study designs based on interviews.9, 12, 18 Two studies were based on retrospective health claims database review16 or chart review.33 Another study9 was designed as a single group prospective study but presented only baseline data for the subgroup of chronic pain patients and as such the results are cross-sectional. The remaining seven publications established CAM utilization using a single group prospective design based on telephone interviews,15, 30, 44 self-administered questionnaires,11 and health record audit with administrative billing records.20, 22, 23

The majority of populations were from specialized clinical practices or from the practices of licensed CAM providers. Three publications reported CAM utilization based on subjects from either a national sample,12 or regional samples from North Carolina.9, 30 Two publications recruited subjects from health claims databases in Washington state,16, 18 and three from family medical and chiropractic practices in Oregon state.20, 22, 23 Similarly, two publications15, 44 randomly sampled licensed massage therapists and patients within their practices from the states of Washington and Connecticut. One study33 recruited patient charts from chiropractic practices within California and another study recruited subjects from large urban specialty clinics for orthopedic or neurosurgery consultation.11

Utilization was primarily defined as a visit with a CAM practitioner, having received or “ever tried” massage, or use from a health claims database or health record from a CAM practitioner. Persons with back pain were asked about use of CAM within the last 12 months,9, 10, 12, 20, 22, 23 lifetime use,11, 18 or visits to practitioner over a 1-year interval.15, 16, 33, 44

Two publications16, 20 using data from large health claims databases showed low rates of massage utilization for LBP, varying from 4 to 5 percent of the total sample. One study12 reported massage use for patients with combined neck and back pain and showed a utilization rate of 14.1 (95 percent CI, 10.8–17.4), an estimated 32.8 million visits (mean of 5.4 visits). In contrast, publications that evaluated use of massage within smaller samples, showed rates of use varying from 16 to 52 percent.

Only one publication15 provided any detail as to the type of massage administered to back pain patients. This publication also reported the mean duration of the visits as 60 minutes. The training or type of massage therapist was specified as a “licensed massage therapist”,15, 18, 44 provided within a chiropractor practice and thus likely provide by a chiropractor,20, 22, 23, 33 or provided within a physical therapy practice thus possibly from a physical therapist.30 The remaining four publications9, 11, 12, 16 provided no details regarding the massage practitioner.

Massage utilization based on anatomical region. Tables 8A to 11B indicate the number of publications specific to regions of the back and these include: one study evaluating massage utilization for neck pain,44 two for unspecified back pain,15, 44 nine for LBP,9, 11, 16, 18, 20, 22, 23, 30, 33 and one study12 that combined back and neck pain. A single study44 evaluating patients in licensed massage therapist practices found rates of use of massage for neck pain to vary between 13 and 20 percent for the states of Connecticut and Washington (combined 17 percent). This same study found slightly higher rates of massage use for unspecified back pain at 20 percent for both states.

Table 8A. Utilization of massage in persons with neck pain.

Table 8A

Utilization of massage in persons with neck pain.

Table 8B. Utilization of massage in persons with neck pain.

Table 8B

Utilization of massage in persons with neck pain.

Table 9A. Utilization of massage in persons with unspecified back pain.

Table 9A

Utilization of massage in persons with unspecified back pain.

Table 9B. Utilization of massage in persons with unspecified back pain.

Table 9B

Utilization of massage in persons with unspecified back pain.

Table 10A. Utilization of massage in persons with low back pain.

Table 10A

Utilization of massage in persons with low back pain.

Table 10B. Utilization of massage in persons with low back pain.

Table 10B

Utilization of massage in persons with low back pain.

Table 11A. Utilization of massage in persons with combined neck and back pain.

Table 11A

Utilization of massage in persons with combined neck and back pain.

Table 11B. Utilization of massage in persons with combined neck and back pain.

Table 11B

Utilization of massage in persons with combined neck and back pain.

The majority of publications9, 11, 16, 18, 20, 22, 23, 30 reported on massage use in persons with LBP and all of these publications evaluated chronic LBP, except two20, 30 which had combined acute and chronic groups. Two publications16, 20 using data from large health claims databases showed low rates of utilization, varying from 4 to 5 percent of the total sample. In contrast, publications that evaluated use of massage within smaller samples with chronic back pain, showed rates of use varying from 24 to 52 percent. One study12 reported massage use for patients with combined neck and back pain also reported no difference in the relative frequency of massage use by location of the back pain.

Massage use as complementary care or combined with other CAM therapies. The majority of publications did not collect or report sufficient information to determine use of complementary or other CAM therapies specific to patients receiving massage. There was one exception15 and this study indicated that massage was used predominately with movement re-education and less frequently with energetic work, and reflexology.

Two publications reported use with other CAM therapies but were not specific to those respondents with back pain44 or specific to those receiving massage.12 Although not stratified by back pain, up to 29 percent of subjects receiving massage saw a medical or osteopathic physician.44 Another study12 suggested that CAM alone had been used by 25 percent of persons with back and neck pain; however, this was not specific to those receiving massage.

Quality assessment for U.S. massage publications. Three of the 12 eligible U.S. publications used cross-sectional study designs based on interviews.9, 12, 18 One publication12 was based on a randomly sampled population cohort with adequate response rate while another18 obtained a low response rate (36 percent) creating the potential for selection bias. Another study recruited subjects from North Carolina, and reported baseline data for subjects with chronic back pain9 or acute back pain.30 Four publications used retrospective administrative database information and employed methods to standardize coding for back pain classification, provider visits, and treatments; one was based on a large health plan database16 and the others from chiropractic practices across a state.20, 22, 23 There is greater potential for selection bias in studies stemming from samples obtained from provider clinics. Two related publications15, 44 randomly sampled licensed massage therapists and patients within their practices but was essentially a single group prospective design. The remaining studies employed single group prospective designs from specialized settings.11, 33 There is significant overlap with the studies included for acupuncture. These studies evaluating the use of massage within the U.S. are prone to some biases but these do not negate the validity of the results; the overall quality is rated as fair for these studies.

Summary of use for back pain in U.S. massage publications. Twelve U.S. publications evaluated the use of massage in persons with back pain. The majority of populations were from specialized clinical practices of back pain patients or CAM providers. Two publications16, 20 using data from large health claims databases showed low rates of utilization, varying from 4 to 5 percent of the total sample. One study with a population based sample12 reported massage use for patients with combined neck and back pain as 14.1 percent (95 percent CI, 10.8–17.4). Publications that evaluated use of massage in smaller samples, showed rates of use varying from 17 to 52 percent. The timeframe of reported use was 1 year, with the exception of two studies evaluating lifetime use.11, 18 Only one of the publications,15 provided any detail as to the type of massage administered to back pain patients.

From 11 publications, the majority reported on low back problems (n = 9) and mostly for chronic pain; only one study44 reported on utilization of massage in neck pain. A single publication12 reported no difference in the relative frequency of massage use by location of the back pain.

Most publications did not collect or report sufficient information to determine use of complementary or other CAM therapies by patients receiving massage. There was one exception15 and this study indicated that massage was used predominately with movement re-education and less frequently with energetic work, and reflexology. Although several studies indicated that massage was used in combination with other CAM therapies, none indicated which other therapies were combined with massage.

Massage Utilization Outside the United States

From the eight publications undertaken outside of the U.S., four were from Canada,116118, 131 two from the Netherlands,121, 130 one from Germany,51 and one from Australia.119 All the publications undertaken within Canada were cross-sectional design from randomly selected samples, predominately at a national level; all but one study131 was undertaken by telephone interview. The Australian study119 was also a population-based cross-sectional study using self-administered questionnaire. The two Dutch publications were both based on a single group prospective questionnaire completed by subjects selected from general medical practices130 and from a retrospective analysis of an administrative database for physical therapy.121 Similarly, the study from Germany51 interviewed a single group prospective cohort from general practices.

Utilization of massage therapy was primarily defined as a visit with a CAM practitioner, or use as indicated in a health claims database. All subjects were asked about use of CAM within the last 12 months with the exception of three publications, which evaluated use within the last 6 months,119 past 4 weeks,131 or lifetime use.130

Five publications116119, 131 were cross-sectional in design recruiting randomly selected national or provincial samples. In these studies utilization rates varied from 7 percent131 within the past 4 weeks to 42 percent lifetime use.116 Generally, studies of single group prospective design showed higher rates of utilization from 56 percent from an administrative database121 to 30 percent.51 The exception to this was a single study that evaluated massage use in persons with neck pain reporting a 7 percent rate of use at baseline.130

None of the publications provided detail as to the specific type of massage received by persons with back pain. The providers of massage were physical therapists in one publication,121 massage therapist in another,119 and were not specified in the remaining publications.

Massage utilization based on anatomical region. A single non-U.S. study130 found that utilization specific to the neck and the use of massage was 7 percent. A single study118 evaluating unspecified back pain reported a rate of 55.5 percent (95 percent CI< 54.1–57.0) for massage use. Similarly, the three publications51, 119, 121 evaluating LBP reported high rates of use that varied from 15 to 56 percent. Three publications116, 117, 131 evaluated rates of massage use reported for neck and back pain combined; in one study131 7 percent of respondents saw massage therapist. In the other two publications 39 to 42 percent received massage.116, 117

Massage use as complementary care or combined with other CAM therapies. Of the five publications providing some information regarding conventional care with CAM, three116118 had not stratified the proportions specific to persons reporting back pain and one did not stratify by practitioner type.119 A single study131 indicated that 4 percent of combined neck and back patients saw a medical doctor and a massage therapist; similarly, 3 percent saw a medical doctor, a chiropractor, a physiotherapist and a massage therapist.131

The majority of publications did not provide detail regarding concurrent or additional use of other CAM therapies with massage, specific to persons with back pain and stratified by therapy. In one study51 37 percent of patients receiving massage, also received acupuncture concurrently. Similarly, one study131 reported concurrent use of massage and chiropractic in 5 percent of the sample. Another study130 of persons with neck pain indicated that up to 12 percent of patients used other CAM therapies, including Reiki, energy healing, and acupuncture.

Quality assessment for non-U.S. massage publications. Five publications116119, 131 were cross-sectional in design recruiting randomly selected national or provincial samples; although derived from large samples, the response rates were relatively low varying from 19 to 69 percent. The survey methodology employed within these studies was of high quality, using pre-tested questionnaires. Three studies employed single group prospective design with subjects recruited from general medical practices51, 130 or from an administrative database for physical therapy;121 one of these130 recruited a small sample relative to the other studies. The overall quality of these studies would be rated as fair.

Summary of use for back pain for non-U.S. massage publications. Of the eight publications undertaken outside of the U.S., four were from Canada,116118, 131 two from the Netherlands,121, 130 one from Germany,51 and one from Australia.119 Five publications116119, 131 were cross-sectional in design recruiting randomly selected national or provincial samples. Within these studies utilization rates varied from 7 percent131 for seeing a massage therapist within the past 4 weeks to 42 percent for lifetime use.116 Generally, studies of single group prospective design showed higher rates from 56 percent from an administrative database121 to 30 percent.51 The exception to this was a single study that evaluated massage use in persons with neck pain which reported a 7 percent rate of use at baseline.130

All but one publication130 queried the use of massage within 12 months or less (as recent as 4 weeks). None of the publications provided detail as to the specific type of massage received by persons with back pain, or the type of provider.

The sole study evaluating neck pain130 reported a rate of 7 percent utilization. Rates of use for LBP or combined neck and back pain were generally higher, varying from 15 (6-month use) to 42 percent (lifetime use). Based on this single study for neck pain, it is difficult to establish whether or not the rate of massage use varies based on anatomical location.

While five publications provided some information regarding conventional care with CAM, four of these did not present results stratified by back pain or practitioner. A single study131 indicated that 4 percent of combined neck and back patients receiving massage also saw an M.D. concurrently; similarly, 3 percent saw an M.D., a chiropractor, and a physiotherapist in addition to a massage therapist.131 Only three studies reported on concurrent use of massage with other CAM therapies such as acupuncture51 (37 percent), chiropractic131 (5 percent), or mixed therapies130 (12 percent receiving either Reiki, energy healing, and acupuncture).

Specific CAM Therapy Utilization: Chiropractic/Spinal Manipulation

We retrieved the full text of 52 articles that potentially provided information on utilization of chiropractic/spinal manipulative therapy for back pain. One publication83 is not summarized below, as this study included comparison data from three other publications already included. Of the 51 remaining publications, 31 were undertaken solely within the U.S., 15 in other eligible countries, and five included both U.S. and Canadian data. Although, one of the publications25 with U.S. and Canadian data did not stratify results by country, we classified this as U.S. as the majority of data was not Canadian.

Chiropractic/Spinal manipulation utilization within the U.S. Thirty-six publications on chiropractic and spinal manipulation, were undertaken in the U.S. The study design varied and included 15 single group prospective cohort study publications,11, 15, 17, 20, 2225, 27, 30, 34, 38, 41, 43, 44 10 retrospective reviews of patient records or claims data,16, 21, 28, 32, 33, 36, 37, 39, 40, 42 and 10 cross-sectional publications.9, 1214, 18, 19, 26, 29, 31, 35 One study was a randomized controlled trial.10

The majority (n = 24) focused on LBP.9, 11, 1624, 27, 28, 3033, 35, 3742 Three publications presented information on neck pain,25, 40, 44 one the thoracic spine,40 six on combined spinal sites1214, 26, 36, 39 and seven unspecified back region.10, 15, 25, 29, 34, 38, 44 Sixteen of 24 publications on LBP and the one study of neck pain provided details on the duration of patients' complaints. Twelve of the publications reported on chronic complaints,9, 11, 18, 21, 22, 24, 27, 28, 31, 37, 41, 42 three on acute pain,19, 20, 30 and two on a range of duration.20, 33

Chiropractors were the providers of spinal manipulative therapy in 34 of the publications,929, 3134, 3644 and physiotherapists in the remaining two.30, 35

Details on the type of spinal manipulation was provided in three publications;15, 24, 33 one clarified that 75 percent of manual therapy provided was “manual, high-velocity, low-amplitude manipulation”,24 and one defined the manual therapy provided as “specific, short-lever dynamic thrusts (or spinal adjustments) or non-specific, long-lever manipulation”.33 The study by Sherman et al.,15 noted that the predominant manipulation provided was of the diversified type.

Eligible U.S. trials on LBP populations found that rates of attending for chiropractic/spinal manipulative therapy ranged from 10 to 47 percent.9, 11, 18, 19, 28, 30, 32, 35, 41 Between 79 and 94 percent of treatments provided by chiropractors included spinal manipulative therapy.9, 11, 22, 23, 28, 33, 37 In two studies in which physiotherapists were the treatment providers, 21 to 34 percent of patients with LBP were treated with spinal manipulative therapy either in isolation or with other modalities.30, 35 The number of visits for chiropractic care or other providers of spinal manipulative therapy ranged from a mean of 4.3 to 15.79, 19, 2224, 28, 33 with the exception of one study of disability claimants who attended for an average of 41 visits,37 and a median of 4 to 716, 28 (Tables 12A to 16B). In one study in which the mean and median number of visits was provided, the mean was approximately double the median.28 In one study in which two similar cohorts of adults with LBP were surveyed 10 years apart, the rates of chiropractic utilization declined by 10 percent between 1987 (40 percent) and 1997 (31 percent).19

Table 12A. Utilization of chiropractic/manipulation in persons with neck pain.

Table 12A

Utilization of chiropractic/manipulation in persons with neck pain.

Table 12B. Utilization of chiropractic/manipulation in persons with neck pain.

Table 12B

Utilization of chiropractic/manipulation in persons with neck pain.

Table 13A. Utilization of chiropractic/manipulation in persons with unspecified back pain.

Table 13A

Utilization of chiropractic/manipulation in persons with unspecified back pain.

Table 13B. Utilization of chiropractic/manipulation in persons with unspecified back pain.

Table 13B

Utilization of chiropractic/manipulation in persons with unspecified back pain.

Table 14A. Utilization of chiropractic/manipulation in persons with low back pain.

Table 14A

Utilization of chiropractic/manipulation in persons with low back pain.

Table 14B. Utilization of chiropractic/manipulation in persons with low back pain.

Table 14B

Utilization of chiropractic/manipulation in persons with low back pain.

Table 15A. Utilization of chiropractic/manipulation in persons with combined neck and back pain.

Table 15A

Utilization of chiropractic/manipulation in persons with combined neck and back pain.

Table 15B. Utilization of chiropractic/manipulation in persons with combined neck and back pain.

Table 15B

Utilization of chiropractic/manipulation in persons with combined neck and back pain.

Table 16A. Utilization of chiropractic/manipulation in persons with thoracic back pain.

Table 16A

Utilization of chiropractic/manipulation in persons with thoracic back pain.

Table 16B. Utilization of chiropractic/manipulation in persons with thoracic back pain.

Table 16B

Utilization of chiropractic/manipulation in persons with thoracic back pain.

Chiropractic/spinal manipulation utilization based on anatomical region. A single study40 provided rates per 1,000 members episodes of care for all three spinal regions and showed a gradient with the thoracic spine being the region for which there was lowest use (10 to 20 percent), the neck, the second most common region (31 to 45 percent), and the low back, the most frequent region receiving chiropractic care (58 to 64 percent). A second study44 reported rates of use for two spinal regions and showed lower rates for the neck (23 to 25 percent) relative to unspecified back (41 to 44 percent). A third study25 showed a similar gradient for neck pain (17 percent for all subjects and 9 percent for those aged over 55 years) and unspecified back pain (38 percent for all subjects and 33 percent for those aged over 55 years).

Chiropractic/spinal manipulation use as complementary care or combined with other CAM therapies. Only two of the 36 eligible U.S. trials provided detailed information on patterns of care seeking; the majority of publications did not present complementary use stratified by persons with back pain. Use of chiropractic care as an alternative to conventional care ranged from 4 to 33 percent and use of chiropractic as complementary therapy ranged from 13 to 16 percent.16, 17 One study16 explored the care seeking patterns of 104,358 Washington State residents who made a claim for LBP and found that 45 percent sought conventional care only (M.D. or PT), 33 percent sought chiropractic care only, 13 percent sought other CAM providers only (aside from chiropractors), and 12 percent pursued both conventional and CAM providers. Another study34, 43 found that one-third of patients with unspecified back pain chose to attend a chiropractor, that chiropractors were the primary provider (defined as the provider who delivered the majority of care) for 40 percent of episodes of back pain, and that 92 percent of chiropractors' patients chose their services again for future episodes of back pain (Table 13A).

Although many authors described treatments provided to patients with back pain, only one study35 provided details on how spinal manipulative therapy is combined with other modalities (Table 14A). Of 2,328 American patients discharged from physiotherapy outpatient services 34 percent received spinal manipulative therapy either in isolation or with other therapy, 10 percent received spinal manipulative therapy with exercises and physiotherapy modalities, and 8 percent received spinal manipulative therapy and exercises.

Quality assessment for U.S. chiropractic/spinal manipulation publications. From 36 publications, 10 cross-sectional studies are based predominately on random samples,9, 12, 14, 26, 31 systematic samples,13, 18, 29 national probability samples,19 or not reported.35 Response rates varied from 36 to 95 percent; the methods of the surveys and interviews were generally of high quality.

There are 15 single-group prospective cohort study publications;11, 15, 17, 20, 2225, 27, 30, 34, 38, 41, 43, 44 two of these15, 44 employed a random sampling of chiropractors. The majority of these papers recruited samples from chiropractic or other professional clinical practices, and as such represent selective groups of patients. From 10 retrospective publications, four were based on patient health records,21, 28, 33, 42 two were from large administrative health databases,16, 40 and four from workers' compensation databases;32, 36, 37, 39 the latter using standardized coding for back injury diagnosis and utilization. Overall, the quality of these studies would be evaluated as fair, noting that biases are present but not sufficient to negate the validity of results.

Summary of utilization in U.S. chiropractic/spinal manipulation publications. Thirty-six publications944 on chiropractic and spinal manipulation, were undertaken in the U.S. The majority (n = 24) reported on LBP and the fewest on neck (n = 3) and thoracic spine (n = 1). Our review suggests that chiropractic care/spinal manipulation is commonly sought by patients with back pain. From 10 cross-sectional studies9, 1214, 18, 19, 26, 29, 31, 35 based predominately on random or systematic samples, rates of utilization varied from 16 percent14 to 45 percent.18 Eligible U.S. trials on LBP populations found that rates of attending for chiropractic/spinal manipulative therapy range from 10 to 47 percent.9, 11, 18, 19, 28, 30, 32, 35, 41 The number of visits for chiropractic care or other providers of spinal manipulative therapy ranged from a mean of 4.3 to 15.79, 19, 2224, 28, 33 Between 79 percent and 94 percent of treatment by chiropractors included spinal manipulative therapy.9, 11, 22, 23, 28, 33, 37 Two studies provided some evidence for a gradient of use based on anatomical regions; LBP was treated most often,25, 40, 44 followed by neck, and then thoracic pain.40

Chiropractors were the providers of spinal manipulative therapy in most publications (n = 34), and physiotherapists in two publications.30, 35 Details on the type of spinal manipulation provided was provided in three only publications.15, 24, 33

The available information suggests that roughly equal proportions of LBP patients use chiropractic as complementary to conventional care and as an alternate. Current literature provides little insight into the manner in which chiropractic/spinal manipulation is combined with other therapies for back pain. Despite the common use of chiropractic/spinal manipulation by patients with back pain the current literature provides limited data on utilization.

Chiropractic/spinal manipulation utilization in non-U.S. publications. Nineteen publications14, 21, 28, 42, 116119, 122, 124, 126, 128, 129, 131136 on the utilization of chiropractic and spinal manipulation, were undertaken within other eligible countries. The majority of these publications (n = 11) focused on LBP.21, 28, 42, 119, 122, 124, 128, 129, 132134 Three publications presented information on neck pain,124, 134, 135 one on the thoracic spine,134 three on combined spinal sites,116, 117, 131 and three on unspecified back regions.118, 126, 136

Study designs varied and included two single-group prospective cohort study publications,122, 126 five retrospective reviews of patient records or claims data,21, 28, 42, 132, 135 one study that used both patient records and single-group prospective cohort data,129 and 11 cross-sectional publications.14, 116119, 124, 128, 131, 133, 134, 136 Four publications did not provide information on duration of pain.14, 118, 122, 136 Ten studies evaluated chronic pain (greater than 3 months) in the back,126 combined back and neck,116, 117, 131 and the low back.21, 28, 42, 119, 128, 133 Three studies evaluated acute pain in acute whiplash (less than 30 days),135 in the LB for pain greater than 24 hours,132 and less than 6 weeks.129 Two studies evaluated mixed duration populations.124, 134

Chiropractors were the providers of spinal manipulative therapy in 14 of the publications,14, 21, 28, 42, 116119, 124, 126, 131, 133135 physiotherapists in one study,132 osteopathic doctors in one study,122 chiropractors and osteopaths in two publications,128, 136 and osteopathic doctors, chiropractors and physiotherapists in one study.129 Details on the type of spinal manipulation was provided in one study;122 this study noted provision of osteopathic manipulation defined as “passive articulation of the lumbar spine and high-velocity thrust techniques”.122

The study that reported on neck pain (specifically, acute whiplash) found that 18 percent of respondents sought chiropractic care.135 Eligible trials on LBP populations found that rates of attending for chiropractic/spinal manipulative therapy ranged from 11 to 40 percent.119, 128, 133 One publication reported that 81 percent of treatment by chiropractors included spinal manipulative therapy.28 In the single study in which physiotherapists were the treatment providers, 3 percent of patients with LBP were treated with spinal manipulative therapy.132 A single study reported the average number of visits for chiropractic care was 10.5 and the median was six.28 A national survey found that overall use of chiropractic services among Canadians increased from 36 percent in 1997 to 40 percent in 2006.116, 117

Chiropractic/spinal manipulation use based on anatomical region. A single study134 evaluated back pain in three regions and showed a gradient from the least utilization for the thorax (7 percent), to the neck (9 percent) and the low back (40 percent). One study124 showed lower rates of use for neck pain (24 to 25 percent) than for unspecified back pain (71 to 74 percent).

Chiropractic/spinal manipulation used as complementary care or combined with other CAM therapies. Of the 19 eligible non-U.S. publications, the five that provided information on patterns of care-seeking all showed that chiropractic care is used both as a complementary therapy and an alternative to conventional care, often in similar proportions.119, 128, 131, 135, 136 The proportion of patients with back pain that attended a chiropractor in isolation ranged from 6 to 29 percent, and use of chiropractic as a complementary therapy ranged from 2 to 36 percent.

Quality assessment for chiropratic/spinal manipulation. Of 19 publications, 11 were cross-sectional publications.14, 116119, 124, 128, 131, 133, 134, 136 Two studies did not report response rates118, 134 while the remaining studies reported rates from 19 to 100 percent. All but two studies124, 134 recruited random samples from the general population. In addition, these two studies did not report the method of questionnaire development. For the remaining eight publications, study design varied and included two single-group prospective cohort study publications,122, 126 one study that used both patient records and single-group prospective cohort data,129 and five retrospective reviews of patient records or claims data.21, 28, 42, 132, 135 All of the retrospective database studies obtained data from practitioner sources, and one study135 also used insurance claim source from persons following a motor vehicle accident. These studies recruited from select back pain patient groups. Overall, the quality of these studies was rated as fair.

Summary of utilization in non-U.S. chiropractic/spinal publications. Nineteen publications14, 21, 28, 42, 116119, 122, 124, 126, 128, 129, 131136 on the utilization of chiropractic and spinal manipulation, were undertaken in other eligible countries. The majority of these publications (n = 11) focused on LBP.21, 28, 42, 119, 122, 124, 128, 129, 132134

Chiropractors were the providers of spinal manipulative therapy in most publications (14 of 19), with other providers included physiotherapists or osteopathic doctors. Only one study provided details on the type of spinal manipulation used.122

One study reported use for acute whiplash neck pain at 18 percent.135 Eligible trials on LBP populations found that rates of attending for chiropractic/spinal manipulative therapy ranged from 11 to 40 percent.119, 128, 133 A single trial reported that 81 percent of treatment by chiropractors included spinal manipulative therapy.28 A national survey found that overall use of chiropractic services among Canadians increased from 36 percent in 1997 to 40 percent in 2006.116, 117 Two studies showed some evidence of lower rates of use for neck/thoracic pain relative to the LBP.124, 134 Only 5 publications provided information on use as complementary or combined care. The proportion of patients with back pain that attended a chiropractor in isolation ranged from 6 to 29 percent, and use of chiropractic as a complementary therapy ranged from 2 to 36 percent.

Specific CAM Therapy Utilization: Naturopathic Medicine and Related Interventions

We identified eight publications examining the utilization of naturopathic medicine and other related CAM therapies. Five of these10, 12, 16, 44, 137 were undertaken in the U.S. and three118120 in other eligible countries. These papers were published between 2002 and 2007 and included 184,907 participants aged 15 to over 90.

Naturopathic medicine utilization within the U.S. There were three publications from the U.S that clearly examined naturopathic medicine utilization. Of these, two were single group prospective study designs44, 137 where data was collected through a self-administered questionnaire,44 or a telephone interview.137 One study was a retrospective review16 of an administrative database. Two publications sampled practitioners44, 137 and one study extracted data from a health insurance plan database.16 All publications defined the utilization of CAM (of which naturopathic medicine was a subset) as ‘a visit to a practitioner’.

One study44 found that approximately 5 percent of patients with LBP saw a naturopathic doctor (ND) in Connecticut and Washington states in 1998 and 1999. A reanalysis of this study undertaken for back and neck pain,137 found that licensed NDs provided a wide variety of interventions for patients with these conditions. However, the frequency of use of the differing ND treatment modalities was not reported specifically for patients with back or neck pain. In the reanalysis, it was reported that 4 to 7 percent of visits to NDs were primarily due to symptoms diagnosed as back problems (Table 17A). Similarly, 2 to 3 percent of patients who visited NDs presented with neck problems as the primary complaint or symptom. A third study,16 unrelated to the previous two, indicated that 2 percent of patients visited an ND at least once in the previous year (2002) for back pain (mean visits = 2). Only 1 percent of people with back pain visited only an ND.16

Table 17A. Utilization of naturopathic medicine in persons with unspecified back pain.

Table 17A

Utilization of naturopathic medicine in persons with unspecified back pain.

Table 17B. Utilization of naturopathic medicine in persons with unspecified back pain.

Table 17B

Utilization of naturopathic medicine in persons with unspecified back pain.

Naturopathic medicine utilization outside of the U.S. One study with a cross-sectional design was conducted on the general Australian population in 2001119 and found that for those with LBP, approximately 3 percent had visited an ND in the past 6 months. No specifics regarding naturopathic treatments were reported in the study. It should be noted that the educational and licensing requirements for NDs in North America and Australia are drastically different, the former mimicking M.D. training in primary care very closely.

No data were reported on the usual costs for these therapies per treatment or for the prescribed course of treatments suggested. These data are insufficient to make any conclusions regarding the utilization of NDs for those with back pain. More research is needed.

Dietary Supplements and Homeopathy Utilization Within the United States

This section focuses on those publications that assessed the utilization of CAM related to the use of dietary supplements and homeopathy for back pain. These interventions are often suggested and delivered by NDs but not all of these papers specifically mentioned that that was the case.

One study was cross-sectional12 and used a telephone interview and one study was an RCT10 that collected data with a self-administered questionnaire. Both publications sampled the general population in the U.S. and defined utilization of dietary supplements and homeopathy as ‘ever having tried’ (Tables 18A and 18B).

Table 18A. Utilization of naturopathic medicine in persons with low back pain.

Table 18A

Utilization of naturopathic medicine in persons with low back pain.

Table 18B. Utilization of naturopathic medicine in persons with low back pain.

Table 18B

Utilization of naturopathic medicine in persons with low back pain.

The RCT10 compared an email discussion group behavioral intervention for LBP with usual care over a one year period. At baseline, 46 percent of the intervention group reported having ‘ever tried’ glucosamine compared to 26 percent in the control group. Thirteen percent of the intervention group and 8 percent of the control group reported trying glucosamine during the last 6 months of the study.

Wolsko12 in a national survey of back and neck pain from 1997 to 1998, found that those suffering from any back and or neck pain used the following CAM therapies over the preceding 12 months: 3 percent used homeopathy; 2 percent used vitamins, 1 percent used herbal medicines, and 1 percent used a special diet.

Dietary Supplements and Homeopathy Utilization Outside the United States

Two Canadian publications sampling the general population, used cross-sectional designs to collect information on CAM use via telephone interview118 and face-to-face interview (Tables 19A and 19B).120 One study found that of the patients with chronic back pain (n = 3259), 39 percent had used a CAM therapy (massage, acupuncture, homeopathy, relaxation, reflexotherapy, or spiritual therapy) in the last year, 18 percent had used homeopathy and 6 percent had used herbal medicine.118 Another Canadian study120 looking at glucosamine use over a 5 year period (1997 to 2002) found that 5 percent of men and 10 percent of women had used glucosamine at some point for back pain. No data were reported on the usual costs for these therapies per treatment or the prescribed course of treatments suggested.

Table 19A. Utilization of naturopathic medicine in persons with combined neck and back pain.

Table 19A

Utilization of naturopathic medicine in persons with combined neck and back pain.

Table 19B. Utilization of naturopathic medicine in persons with combined neck and back pain.

Table 19B

Utilization of naturopathic medicine in persons with combined neck and back pain.

Other CAM Therapies Based on Self-Care

Three publications focused on the use of single CAM therapies that are generally self-administered, including yoga,138 prayer,139 and mind body therapies.140 All three of these publications were subgroup analyses from a larger population-based survey undertaken in 1998.5 All evaluated both lifetime use and use within the last 12 months. Utilization reported in these studies was limited to the proportion of the sample who used these therapies and had neck or back pain.

Forty-two percent (standard deviation (SD) 4.5 percent) of those participating in yoga had back or neck pain; the concurrent or lifetime use of other CAM therapies for persons using yoga was not stratified by back pain. Prayer was used by 18 percent of persons with back or neck pain and 59 percent of these found it to be very helpful; 40 percent of these patients also saw a physician for their condition.139 Approximately 18 percent (standard error (SE) 2.6) of persons with back or neck pain also used mind body therapy within the last 12 months and 40 percent (SE 8.2) found this to be very helpful for their condition; mind body therapy can include relaxation techniques such as meditation, stress response, guided imagery, and biofeedback.140 The specific types of mind-body therapies used by persons with back or neck pain were not identified.

Assessment of Quality of Utilization Publications

Given the degree of overlap of studies among the differing CAM therapies, we thought it useful to summarize quality for all the studies reporting utilization of any CAM therapy. As noted previously, single group cohort study designs are prone to a variety of biases and do not have specified criteria to assess methodological quality.

Quality of Utilization Publications Undertaken Within the United States

A total of 46 publications on CAM utilization were from the U.S. Of these, 10 were based on retrospective patient records or claims data16, 21, 28, 32, 33, 36, 37, 39, 40, 42 and 15 were cross-sectional publications.5, 9, 1214, 18, 19, 26, 29, 31, 35, 45, 138140 A single publication was based on an RCT.10 The remaining publications were single group cohort publications, for which quality was not assessed.

Of the publications based on retrospective health database or chart records, three were from large workers' compensation claims databases36, 37, 39 where the primary interest was in chiropractic visits. Three publications16, 32, 40 were from large health insurance claims or managed care network databases. All of these publications have standardized coding for classifying back pain (predominately ICD 9 codes), provider visits, and enrollee characteristics and treatments; given the pre-specified coding systems, linkages between elements within the database were likely limited in errors. Four publications used health record extraction from chiropractic charts;21, 28, 33, 42 and standardized data collection forms were used.

All the U.S. cross-sectional publications used self-report methods of ascertaining utilization of CAM. Table 20 details these publications with respect to sampling strategy, response rate, mode of administration of the survey, questionnaire design (which we operationalized as pre-testing of questions) and the time frame for recall. Several publications were related to a single national survey,5, 12, 138140 and two to residents of North Carolina.9, 31 Response rates varied from 95 to 60 percent and all but one5 would be considered acceptable; an acceptable response rate is from 65 to 75 percent.84 Recall periods were 6 or 12 months, with the exception of two papers29, 35 that collected current episode or the past 2 weeks. The majority of publications used telephone or in person interviews suggesting less potential for bias. Limited information about questionnaire design was provided.

Table 20. Quality characteristics of cross-sectional studies using self report utilization data from U.S.

Table 20

Quality characteristics of cross-sectional studies using self report utilization data from U.S.

Quality of Utilization Publications Undertaken in Other Countries

A total of 22 publications were undertaken within eligible countries outside the U.S. with an additional four overlapping with U.S. data.14, 21, 28, 42 Of these, eight21, 28, 42, 121, 125, 129, 132, 135 were based on retrospective patient record or claims data and 14 were cross-sectional publications.14, 116120, 123, 124, 127, 128, 131, 133, 134, 136 The remaining four publications were single group cohort publications, for which quality was not assessed.51, 122, 126, 130

Among the publications from retrospective health database or chart records, one study135 used claims from a provincial insurance database for persons with whiplash. Others used, patient health records from physiotherapy practices,121, 132 a pain clinic125 and chiropractic practices.21, 28, 42, 129 It was not clear in all of these publications if standardized forms were used to collect utilization information.

The 14 cross-sectional publications all used self-report methods of ascertaining utilization of CAM. Table 21 shows these publications and their quality characteristics. Four of these publications116118, 133 were based in Canada and were undertaken at a national level using random sampling; similarly, two in Australia were nationally-based random sample surveys.119, 127 Three publications recruited patients from practitioner clinics.124, 128, 134

Table 21. Quality characteristics of cross-sectional studies using self report utilization data from non-American Studies.

Table 21

Quality characteristics of cross-sectional studies using self report utilization data from non-American Studies.

Response rates varied from 19 to 100 percent when reported. Recall periods included lifetime use,120 12 months,116118, 127, 133 3 to 6 months,119, 131, 136 4 weeks,131 and current episode.124, 128, 134 Three publications127, 128, 133 did not report the mode of data collection for utilization data. The remaining publications all used mailed questionnaires or telephone interviews with the exception of one study that used face-to-face interviews.120 Limited information about questionnaire design was provided.

Question 2. What is the Utilization Recommended by Different Types of Healthcare Providers?

To address this question we searched primarily for relevant U.S. based Clinical Practice Guidelines (CPG), and also publications that specified provider views on the recommended use of CAM therapies for patients with back pain.

U.S. Clinical Practice Guidelines

We reviewed 19 U.S. CPG on neck or back pain. Of these, three were excluded as their target population was adults with spinal pain related to neurological conditions.141143 Four were excluded as they did not provide any recommendations for the use of CAM therapies for back pain.144147 A single publication148 compared CPG from 11 different countries; this study was excluded as the U.S. guideline was developed in 1994.

Eleven guidelines related to the management of back pain included the use of CAM and the characteristics of these CPGs are detailed in Appendix C Table 1. Of these guidelines, three were published by professional organizations (acupuncture46 and chiropractic organizations149, 150), and three from payer organizations.151153 Only one CPG46 made specific utilization recommendations for acute, subacute, chronic and recurrent flare-up patient conditions. The remaining 10 made general recommendations for using specific CAM therapies either by treatment type (n = 10), practitioner type (n = 2), or back region (n = 9) as follows:

By treatment type including. 1) Manipulation/Mobilization,149152, 154158 2) Heat or Cold,154, 155, 157 3) Relaxation,155, 156 4) Acupuncture/electro-acupuncture,46, 153, 156 5) Massage,151, 156 6) Manual Therapy,152, 158 and 7) Yoga.156

By practitioner. 1) Chiropractor,151, 152 2) Massage Therapist,151, 152 and 3) Manual Therapist.152

By spinal region. 1) Neck (none), 2) Neck/thorax,152, 154 3) Thorax,151 4) Low back,149, 151, 155157 5) Mixed back regions,46, 150, 158 and 6) Back region not specified.153

For the single CPG46 with utilization recommendations, the recommended frequency for electro-acupuncture was for two to three times weekly for a duration of 4 weeks for acute and subacute back pain and 6 to 8 weeks for chronic back pain and recurrent/flare-up, Appendix C Table 1. Although, these recommendations were not specific to either neck or back problems, the CPG was developed to apply to all these conditions. Although very similar in its recommendations, this guideline distinguishes between utilization with respect to initial course of treatment and a continuing course of treatment. The CPG recommends re-evaluation after 12 weeks of treatment irrespective of the duration of the condition. Two independent raters scored this CPG using the AGREE scoring instrument.(see Appendix C Table 2) For this guideline, stakeholder involvement scored 6 out of 12, applicability 8 out of 12, and rigor and development 25 out of 32. Based on the AGREE89 score, this CPG would be acceptable with greater clarification and specifically a stronger link to the evidence for recommendations for utilization.

Summary of recommended use of CAM therapies. A single CPG, for electro-acupuncture, provided recommendations for the frequency of treatment that was stratified by duration, and recurrence. The guideline was developed from a provider organization and was acceptable based on the AGREE quality score.

Provider Views About Recommended Use of CAM

Of the 22 publications that focused primarily on patient or provider views on CAM, 16 were excluded because they had no direct or indirect information on recommendations for CAM specific to back pain and in the context of optimal frequency and duration of utilization. Two publications47, 48 were U.S. and four4952 were from other eligible countries.

Of the two U.S. publications providing direct recommendations, one48 did not provide information stratified for back pain. The other publication47 recommended the type of CAM treatment for LBP in pregnancy and included heat/cold, yoga, and massage; no information on the utilization was provided. This recommendation was based on the views of American health care providers (physicians, nurses, midwives, prenatal educators).

Four non-U.S. publications provided some recommendations for the treatment of back pain; three of these were from the United Kingdom49, 50, 52 and one from Germany.51 Two of these publications51, 52 made recommendations primarily with regard to the type of CAM appropriate for managing back pain. CAM practitioners in the United Kingdom recommended the Bowen technique, chiropractic, magnet therapy, massage, reflexology and yoga for LBP and the Bowen technique, chiropractic and massage for neck pain.52 The German publication51 cautioned that receiving acupuncture for acute, recurrent or chronic LBP seemed to be a function of availability and did not offset the use of other health care resources. In fact, there were both increased consultation rates and use of other health care services after adjusting for key patient characteristics.

Two related publications undertaken in the United Kingdom based on a cross-sectional questionnaire49 and a mixed methods design that included a qualitative component,50 captured attitudes toward LBP and included opinions about recommended utilization. These publications helped to explain the practice patterns observed in three provider groups (chiropractors, osteopaths, and physiotherapists). Based on provider survey views, these publications found that osteopaths and physiotherapists endorsed limiting the number of treatment sessions more than chiropractors did. Additionally, those working in private practice did not endorse limiting the number of treatment sessions as frequently as those working in a national health service setting. Those in private practice also endorsed a biomedical treatment oriented attitude advising their patient to be more vigilant and restrict activity versus a reactivation treatment orientation (items that concern return to work, daily activity and increasing mobility). These two publications would suggest that there may be moderating factors that influence the recommended use of CAM. These include availability of the CAM service, attitude of the practitioner (reactivation versus biomedical orientation) and a national health service versus private practice setting.

Summary of provider views on utilization. Six studies reported views on recommended use of CAM therapies. The two studies published in the U.S. did not stratify recommendations specifically for patients with back pain. Four publications from studies outside the U.S. provided some recommendations on use of CAM therapies specific to persons with back pain. None of the publications encompassing provider or patient views addressed recommendations with regards to the frequency or duration of treatment. These limited publications would suggest that the frequency of utilization may be influenced by the availability of the CAM service, attitudes of differing providers, the practice settings, and the type of CAM therapy.

Question 3. What are the Usual Costs for These Therapies per Treatment and for the Prescribed Course of Treatment?

Our review identified 18 publications that were associated with costs or economic evaluations related to back pain and the use of CAM within the U.S. One study5 provided information on costs and coverage for CAM therapies but did not present results stratified for back pain. Another study35 evaluated the use of manual therapy applied by physical therapists but did not present costs in relation to the specific therapies applied. Of the remaining publications, two were based on the same patient population cohort53, 54 and two on the same health claims database.32, 59 Another study57 incorporated data from three different publications and undertook an economic analysis of this combined data. Study characteristics and results are detailed in Appendix C Tables 3 and 4.

A variety of study designs were employed to collect cost related data including: retrospective analyses from administrative databases (n = 7),16, 32, 36, 39, 5961 single group prospective design (n = 5),20, 34, 53, 54, 57 RCTs (n = 3),55, 56, 58 and a case-control design.37

The majority of publications (n = 13) focused on costs associated with CAM services in persons with LBP.16, 20, 32, 37, 39, 5360 From these LBP studies, four5356 included acute pain varying from greater than 7 days 56 to less than 10 weeks.53, 54 Four studies16, 32, 39, 58 did not specify the duration. Three studies evaluated chronic back pain described as recurrent,59 chronic60 or disabling.37 Two studies combined chronic and acute back pain.20, 57 For the three studies that did not evaluate costs of CAM services in persons with LBP, two evaluated back pain location unspecified34, 36 and one combined neck and back pain;61 none of these three studies identified the duration of the pain.

The source of payment for the CAM therapies used in these studies included private health insurance (n = 5),16, 32, 34, 59, 61 workers' compensation insurance (n = 4),36, 37, 39, 60 and mixed sources including out of pocket (n = 4).20, 53, 54, 57 Three studies,55, 56, 58 that were randomized trials, had the study or the participants insurance pay for treatments.

Characteristics of Cost Evaluations for CAM Services

Fifteen of the 18 studies were not true economic analyses, but rather cost identification studies. A single study57 undertook a complete cost-effectiveness analysis (CEA) for patients with LBP. The perspective of this analysis was from that of the payer. This CEA had some limitations, as data for complete cost estimates were only available for 38 percent of chronic and 50 percent of acute patients. In addition, this study did not assess actual use of services for referred patients, rather they imputed some costs. This approach was based on previously used methods for charges based on per claimant data and adjusted for proportion of provider charges that were actually reimbursed. It was not clear if this study57 included the costs of adverse events related to visiting a chiropractor or M.D.. Although three studies53, 55, 56 did provide health outcome change scores and costs, they did not estimate cost-effectiveness ratios; one of these studies56 attempted to evaluate adverse events.

Comparators to CAM provider costs. Although selection of a comparator is not necessary in cost identification studies, all but two39, 61 undertook comparison of back treatment costs for CAM and non-CAM providers (Table 22). The overwhelming majority of studies (n = 13) evaluated chiropractors as the exclusive CAM provider.20, 32, 34, 36, 37, 39, 5355, 5759, 61 One study56 evaluated choice of a single alternative provider that may have included a chiropractor, acupuncturist, or massage therapist; another study16 included any of four CAM providers (chiropractors, massage therapists, acupuncturists, or NDs) and another60 may have included acupuncturists or massage therapists within the “other” category. Similarly, CAM provider costs were compared to a number of conventional practitioners including: M.D.s (n = 11 publications),16, 20, 32, 34, 36, 37, 53, 54, 56, 59, 60 orthopedic surgeons or internists,53, 54 physical therapists,55, 60 combined M.D.s and physical therapists,58 and “mixers” of both CAM and conventional providers.16

Table 22. Summary of studies evaluating costs in persons with low back pain.

Table 22

Summary of studies evaluating costs in persons with low back pain.

Direct costs for CAM services. The cost perspectives evaluated were predominately from that of the payer, and as such direct costs were collected. All studies included costs of visits to practitioners, but varied with respect to: a) inclusion of imaging or diagnostic tests, b) visits to specialists, c) back surgery costs and d) the use of medications. For studies with the workers' compensation insurance as the payer, some included costs for lost days36, 60 and others excluded these.37, 39 Studies from other payers such as private insurance could have attributed lost days as indirect costs, but none evaluated these. The studies using health administration databases were prone to including costs associated with other illnesses, due to the manner of establishing an episode of care.32

The years for which costs were estimated varied significantly; these included cost estimates from a single study in 1982,34 two from 1985,36, 37 five from 1995,20, 5355, 57 one from 2002,16 and one from 2003.56 Other studies used a range of years for cost estimates including: a) a two year interval from 1988-199032, 59, b) a four year interval from 1988-199260, c) another from 1997-2001,61 and d) from 1998 to 2002.60 One study did not specify any year but estimated costs over an 18 month interval.58 One study specified costs for each of the years between 1999 to 2002.39 In general, half of the cost estimates were based on data over 12 to 25 years and only three studies reported costs within the last 5 years (Table 23).

Table 23. Summary of studies evaluating costs in persons with unspecified back pain or combined neck and back pain.

Table 23

Summary of studies evaluating costs in persons with unspecified back pain or combined neck and back pain.

A variety of cost measures were used including: a) total costs of care per episode per patient,20, 34, 36, 39, 53, 56, 60 b) mean costs for index provider,54 c) costs per unit visit,16, 55 mean costs over an interval of time,55, 58, 61 percentage of total costs,37 first episode versus multiple episodes costs,32 and first and subsequent episode provider costs.59 A single study estimated the incremental cost effectiveness ratio for health outcomes of pain and disability reduction.57 Appendix C Table 3 details the dollar values for each of these cost outcomes. Although costs do vary with differing providers, interpretation of these cost differences is problematic in the absence of links to health benefits or harms. There are also problems with the, often implicit, assumption that the effects of the providers and their treatments are equivalent.

Indirect costs for CAM services. The cost perspective for all studies was that of the payer, and as such direct costs were collected and indirect costs were either considered to be of less importance or were intentionally excluded.55 No studies evaluated indirect costs with respect to time off work to attend therapy, or the costs associated with any adverse reactions from CAM therapy. The two studies in which the payer was workers' compensation insurance36, 60 included time off due to back pain as direct costs to the payer. Sick days taken for back pain in studies that estimated costs from other payers would not have captured these indirect costs to the patient or their employers.

Quality of studies evaluating costs. (Appendix C Table 4) A single study57 undertook a formal CEA and methodological quality was evaluated using the Quality of Health Economic Studies (QHES).90 This study scored low on criteria related to selection of perspective, specification of economic analyses, justification for these, and disclosing the source of funding for the study. The QHES could not be applied to the remaining studies as they were cost-identification evaluations and not true economic analyses. The limitations in the accuracy of cost estimates for publications using administrative databases16, 32, 36, 39, 5961 have been detailed previously. However, we noted that even those studies that employed RCT design, relied upon administrative health data for their cost estimates; as such these high level designs were also prone to the same potential confounding that results from using this source of cost data.56

Impact of insurance coverage on costs for CAM services. Three publications20, 53, 57 were not based on samples entirely covered by insurance from private health organizations or workers' compensation. One study53 indicated the proportion of persons with coverage for urban versus rural chiropractors and primary care providers, and orthopedic surgeons. Another study57 reported the proportion of those covered by differing insurance carriers (or lack of coverage) as a function of provider type (chiropractor versus M.D.) and the duration of the LBP (acute versus chronic). The findings of this study would suggest that the duration does not affect the relative proportions of coverage for either chiropractor or M.D.. However, there was a greater proportion of patients who had no coverage seeking chiropractor relative to M.D. services (42 to 47 percent versus 6 to 8 percent). Similarly, a smaller proportion of patients with private or Medicaid insurance had coverage for chiropractic relative to M.D. services for back pain. The third study20 reporting the impact of coverage suggested that up to 42 percent of persons with LBP pay out of pocket for chiropractic services. In contrast, those who seek medical services that include referral to a surgeon or physical therapist pay only 3 percent out-of-pocket; those seeking medical services and who are not referred to either a surgeon or physical therapist have slightly higher (7 percent) out of pocket costs.

Views

  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...