• Chiropractic and Medical Care Costs of Low Back Care: Results ...

  • FileName: AJMC2002sepStano802_809.pdf [read-online]
    • Abstract: Table 1 shows the ChiroCode and Medicare RVUs. for the leading chiropractic procedure ... ChiroCode RVUs are generally the same as or very. similar to Medicare ...

Download the ebook

Chiropractic and Medical Care
Costs of Low Back Care: Results From a
Practice-Based Observational Study
Miron Stano, PhD; Mitchell Haas, DC; Bruce Goldberg, MD;
Paul M. Traub, MBA; and Joanne Nyiendo, PhD
Objective: To compare the 1-year costs for patients treated indirect costs associated with absenteeism and lost
for acute and chronic ambulatory low back pain by medical productivity are even greater than the direct treat-
physicians and chiropractors. ment costs.2
Study Design: Prospective, practice-based observational Nonmedical providers, particularly chiropractors,
study undertaken in 13 general medical practices and 51 chi-
account for a significant share of back care provid-
ropractic community-based clinics.
ed.3-7 It is thus important to determine what chiro-
Patients and Methods: Of 2872 study patients, 2263 had
complete 1-year records of services. Service data, collected practic patients are getting for their spending and
from billing records, chart audits, and provider questionnaires, whether chiropractors are relatively cost-efficient
were assigned relative value units that were converted into providers of care. These issues are becoming more
1995 dollar costs. Prescription drug costs for medical patients relevant as chiropractic and other forms of alterna-
were included. Patient data on health status, pain and disabil- tive medicine are being increasingly integrated into
ity, and socioeconomic characteristics were obtained from self- managed care.8-11
administered questionnaires. Using literature reviews12 as well as original stud-
Results: The direct office costs of treating both chiropractic ies,13-15 Shekelle and colleagues provide a concise
and medical patients over a 1-year period were relatively summary of the evidence on chiropractic cost effec-
small. Forty-three percent of chiropractic patients and 57% of
tiveness.16 They acknowledge that most studies have
medical patients incurred costs of less than $100. However,
found that chiropractic care is relatively cost efficient
the mean costs associated with chiropractic patients ($214)
were significantly higher than those for medical patients compared with medical care. Nevertheless, they also
($123), especially when compared with medical patients who argue that chiropractic’s favorable evaluation “has not
were not referred for further treatment or evaluation ($103). been convincingly established” because “most studies
Chiropractic patients had somewhat lower baseline levels of have failed to compare equivalent patients, measure
pain and disability than nonreferred medical patients, but the clinically useful outcomes, and include both direct
2 groups were relatively similar on most patient characteris- and indirect costs in the comparisons.”16(p6) Concerns
tics. There also were no statistically significant differences in about the methodologic rigor of existing chiropractic
the improvements in pain and disability between these 2 studies have been echoed elsewhere.17,18 Some inves-
groups of patients.
Conclusion: The results of this study indicate that patients
treated in chiropractic clinics incur higher costs over a 1-year
period, but have about the same degree of relief as nonreferred From the Department of Economics, School of Business
Administration, Oakland University, Rochester, Mich (MS); the
patients treated in medical clinics.
Division of Research, Center for Outcomes Studies, Western States
(Am J Manag Care 2002;8:802-809)
Chiropractic College, Portland, Ore (MH, JN); the Department of
Family Medicine and Department of Public Health and Preventive
Medicine, Oregon Health & Science University, Portland, Ore (BG);
and the Corporate Economics Department, DaimlerChrysler
Corporation, Auburn Hills, Mich (PMT).
This study was supported by the Health Resources and Services
he economic costs of back pain are substan- Administration, Department of Health and Human Services (grant
T tial. A 1994 US Department of Health and
Human Services report placed this nation’s
annual healthcare bill for back problems in the $20 to
no. R18AH10002) and by a challenge grant from the Foundation for
Chiropractic Education and Research (grant no. 940502).
Address correspondence to: Miron Stano, PhD, School of Busi-
ness Administration, Oakland University, Rochester, MI 48309.
$50 billion range.1 Other findings suggest that the E-mail: [email protected]
Costs of Low Back Care
tigators have found that chiropractic care costs even Washington State, all other medical and chiropractic
more than treatment by primary care physicians19,20 clinics are located in Oregon. The vast majority of par-
while providing marginal patient benefits.21 ticipating chiropractors were in solo practice; only 8
In the face of unclear and sometimes contradicto- practices consisted of 2 chiropractors. In contrast,
ry results, more work is needed to guide patients, the medical doctors generally worked in group prac-
providers, and third-party payers. The research com- tices. Four medical clinics were Oregon Health &
munity faces a formidable challenge to developing a Science University academic practices.
more complete picture of the economic impact of The study enrolled 2872 patients (1950 chiro-
chiropractors and other practitioners. The costs of practic and 922 medical) with acute and chronic
back care are difficult to measure due to the vari- ambulatory low back pain of mechanical origin
ability in the needs for treatment and the long-term between December 8, 1994, and June 30, 1996. The
duration of treatment for chronic patients. In addi- mean numbers of patients treated by the medical
tion, it is difficult to distinguish between care for the and chiropractic clinics, respectively, were 52.9 (SD
back and care for unrelated conditions or other = 93.9, median = 11.5, interquartile range = 7.0-
conditions that are aggravated by the patient’s back 61.0) and 33.1 (SD = 37.6, median = 17, interquar-
problems. For example, chiropractors who routine- tile range = 6.3-49.8). Information was collected on
ly perform full-spine manipulation may include patient demographics, insurance type (though not
manipulation of the lumbar spine in the treatment whether it was managed care), health status, psy-
of neck pain. chosocial characteristics, complaint characteristics,
Long-term, randomized clinical trials are widely and physicians’ practice activities. Patient data were
recognized as the gold standard for determining effi- obtained using self-administered questionnaires at
cacy. But, as the back pain research indicates, such the initial visit and at 5 follow-up periods (1 month,
trials do not necessarily ensure high quality.18 They 3 months, 6 months, 9 months, and 1 year). Data on
also are expensive, and in the case of chiropractic physician practice activities were obtained by ques-
treatment, best suited to comparisons of methods of tionnaire at each patient visit for treatment of low
manipulation similar to chiropractic methods.22 As a back pain and by chart audit at the end of the study.
result of these limitations, investigators have adopt- Two patient outcomes measures were adopted:
ed other approaches, such as administrative and sur-
vey databases, that can provide useful information 1. Severity of present pain, as assessed by a 100-mm
about the efficacy and efficiency of alternative treat- visual analog scale (VAS) score with the descrip-
ments for back problems.23 tive anchors “no pain” (0) and “excruciating
pain” (100).
This report is derived from an ongoing longitudi-
nal, prospective, practice-based observational study 2. Functional disability, as measured with the
undertaken in general medical practices and chiro- Revised Oswestry Disability Questionnaire
(OSW), a 10-item instrument designed to mea-
practic community-based clinics. A prospective,
sure the effects of low back pain on daily activi-
observational study of clinical activities and associat-
ties such as personal care, lifting, walking, sitting,
ed patient outcomes offers a pragmatic approach to sleeping, and social life. For each question,
assessment of therapeutic modalities by defining and patients choose 1 of 6 descriptive statements
quantifying the clinical problems seen by the physi- indicating the degree of dysfunction. The OSW
cian in practice and the nature of the interaction score ranges from 0 to 100, with higher values
between the physician’s approach and the patient’s reflecting greater disability.
response to treatment. An observational study also
can be a useful complement to randomized clinical These 2 outcome measures, widely used in back
trials because the latter may not be generalizeable pain research,28 are analyzed in other reports for this
when therapist, setting, and patients are atypical.24 study.26,27 Here, we focus on the costs of care that
were provided in the participating clinicians’ offices
over a 1-year period and on cost comparisons
METHODS between chiropractic and medical patients.
Services provided in the medical and chiropractic
Described in detail elsewhere,25-27 the project clinics were collected from billing records, chart
involves 111 medical physicians in 13 general medical audits, and provider questionnaires. The services
clinics and 60 chiropractic physicians in 51 chiro- were assigned Current Procedural Terminology (CPT)
practic clinics. Except for 1 medical clinic located in codes that were converted to relative value units
(RVUs) using 1995 Medicare RVUs for medical
physicians29 and 1995 RVUs from The ChiroCode
Book for chiropractors.30 ChiroCode RVU values Table 1 shows the most frequent chiropractic
correspond closely to Medicare RVUs but provide a and medical services provided to all 1920 chiro-
more complete list than Medicare. If a procedure practic and 952 medical patients and the total num-
code did not have an RVU, an RVU value was inter- ber of services provided. Of the 23 procedure codes
polated based on the national charge for that proce- used by chiropractic physicians more than 100
dure code compared with the national charge31,32 for times, regional manipulation is by far the dominant,
the most common procedure code for each group: accounting for 31% of the total chiropractic services
an office visit (CPT code 99213) for medical physi- provided. (Eighty-four percent of chronic chiro-
cians and regional manipulation (CPT code 97260) practic patients received manipulation at some
for chiropractors. We used the same methods to point in time.26) Chiropractors also use a wider set
interpolate RVUs for procedure codes with no of therapies than medical physicians. The top 5
Medicare or ChiroCode RVUs and no national procedure codes accounted for only 64% of all chi-
charge data. Based on relative charges from the ropractic services. By contrast, an office outpa-
billing information we collected, RVUs were inter- tient visit for an established patient is easily the
polated from the mean billing charge of that proce- leading medical procedure code; and the top 5 med-
dure code compared with the mean billing charges ical procedure codes, consisting mainly of office
for CPT codes 99213 for medical physicians and visit codes, account for 85% of all medical services.
97260 for chiropractors. Chiropractors also used far more services per
Costs were cumulated for each medical patient patient (14.4 vs 2.7 for medical patients) over the
based on the RVUs and the national Medicare con- course of care.
version factors for 1995. Prescription drug costs for Table 1 shows the ChiroCode and Medicare RVUs
medical patients, based on 1995 Red Book prices,33 for the leading chiropractic procedure codes and the
were calculated separately. As the RVUs for chiro- Medicare RVUs for the medical procedure codes.
practors are often the same as or very similar to ChiroCode RVUs are generally the same as or very
Medicare RVUs, a different conversion factor was similar to Medicare RVUs, but the ChiroCode list is
developed for chiropractors to reflect their lower more complete. The cost assigned to each procedure
fees. For example, CPT code 97260 (regional code shown in Table 1 was determined by using the
manipulation) is assigned 0.41 RVUs for both med- methods described in the Methods section.
ical physicians29 and chiropractic physicians.30 Table 2 shows the mean and median costs per
However, the 1995 median national fee for CPT patient for those with complete 1-year cost records.
code 97260 provided by chiropractic physicians32 The medical patients are further divided into 2
was $19 compared with $27 for medical physi- groups: those who were referred for evaluation or
cians31—a 70% ratio. Similarly, the respective treatment to a surgeon or physical therapist and
national fees for CPT code 99212 (office/outpatient those who were not referred. Of the 128 referred
visit for an established patient) were $26 and $36— patients, 33 were referred to surgeons only, 80 to
a 72% ratio. Based on the average fee ratios for pro- physical therapists only, and 15 to both. There were
cedure codes that were the most common to both too few chiropractic patients with complete cost
medical physicians and chiropractic physicians, records who were referred to a surgeon (n = 3) or to
the Medicare conversion factor was multiplied by a physical therapist (n = 12) to warrant separate
0.71 to cumulate the costs of the RVUs provided to analysis.
each chiropractic patient. Overall, the mean costs were relatively low.
Of the 2872 patients enrolled in the study, 2263 Nevertheless, the mean for the chiropractic group
(1524 chiropractic and 739 medical) had complete ($214) was nearly double that of the total medical
cost records, as determined from billing and chart group ($123), although it was slightly less than that
abstraction, with no missing values for any of the of the medical referred group ($217). These cost
data collection points over the 1-year period. Of data, however, do not capture the costs of any refer-
those patients with complete 1-year cost records, ral treatment, including possible surgical and post-
1360 (916 chiropractic and 444 medical) patients surgical care, as well as the costs of advanced
had both their baseline and 1-year VAS scores, and imaging. The cost values also do not include the
1372 (925 chiropractic and 447 medical) had both costs of care that may have been independently
their baseline and 1-year OSW scores. sought by either chiropractic or medical patients.
Costs of Low Back Care
Table 1. Most Frequent Chiropractic and Medical Procedure Codes*
CPT Code CPT Description Frequency ChiroCode Medicare Cost ($)
97260 Regional manipulation 8712 0.41 0.41 10.08
97010 Hot or cold packs therapy 2744 0.45 0.34 11.06
99212 Office/outpatient visit, EST 2522 0.68 0.68 17.57
97014 Electric stimulation therapy 2205 0.42 0.40 10.32
99211 Office/outpatient visit, EST 1735 0.38 0.38 9.82
97124 Massage therapy 1449 0.41 0.47 10.08
97035 Ultrasound 1245 0.41 0.33 10.08
99070 Special supplies pepper patch 893 — — 6.57
97261 Supplemental manipulations 577 0.24 0.24 5.90
97122 Manual traction therapy 547 0.40 0.57 9.83
97118 Manual electric stimulation 541 — — 4.71
97128 Ultrasound therapy 528 — — 4.02
99213 Office/outpatient visit, EST 410 0.96 0.96 24.80
2000 Manipulation of spine 406 0.75 0.75 18.43
97250 Myofascial release 332 0.84 0.84 20.64
99203 Office/outpatient visit, new 298 1.72 1.72 44.43
99202 Office/outpatient visit, new 246 1.25 1.25 32.29
97032 Electric stimulation, manual 236 0.51 — 12.53
99201 Office/outpatient visit, new 231 0.79 0.79 20.41
72100 X-ray exam of lower spine 225 1.01 1.01 24.82
97110 Therapeutic exercises 30 min 186 0.52 0.60 12.78
97122 Traction, manual 151 0.40 0.57 9.83
99212 Office visit, EST, focused 149 0.68 0.68 17.57
Other procedure codes 1493
Total: all procedure codes 28 061
99213 Office/outpatient visit, EST 1404 — 0.96 34.93
99212 Office/outpatient visit, EST 319 — 0.68 24.74
99214 Office/outpatient visit, EST 195 — 1.48 53.85
72100 X-ray exam of lower spine 128 — 1.01 34.96
99202 Office/outpatient visit, new 62 — 1.25 45.48
81000 Urinalysis with microscopy 53 — — 9.35
99203 Office/outpatient visit, new 46 — 1.72 62.58
99201 Office/outpatient visit, new 25 — 0.79 28.74
Other procedure codes 235
Total: all procedure codes 2467
ChiroCode indicates the ChiroCode Book30; CPT, Current Procedural Terminology; EST, established; RVU, relative value unit.
*The frequencies are based on data for 1950 chiropractic and 922 medical patients.
Two other features are apparent. First, the cost of had costs of more than $1000 (maximum = $3111).
prescription drugs is an important component of In comparison, the majority of the medical patients
medical costs, accounting for nearly 30% of the total. (57%) incurred costs of less than $100 and fewer
Second, there are large discrepancies between mean than 2% incurred costs of more than $500 (maxi-
and median costs. These discrepancies arise mum = $1698).
because the distributions of total costs, especially We also assessed the potential role of patient demo-
for chiropractors, are highly skewed. Forty-three graphics and health indicators in costs. Table 3 shows
percent of the chiropractic patients with complete mean values for selected patient characteristics and
1-year costs incurred costs of less than $100, but the baseline pain (VAS) and disability (OSW) scores
nearly 10% had costs exceeding $500 and 2 percent for chiropractic, medical (nonreferred), and medical
(referred) patients.
Table 2. Mean and Median Costs Per Patient Chiropractic patients
reported less pain and
Mean ± SD (Median) Cost in 1995 Dollars
disability at baseline
Type of Treatment CPT Code Prescription Total than the nonreferred
medical patients, even
Chiropractic (n =1524) 214 ± 284 (124) NA 214 ± 284 (124)
though a somewhat
Total medical (n = 739) 89 ± 80* (70) 34 ± 71 (17) 123 ± 128* (89)
higher proportion had
Nonreferred medical (n = 611) 78 ± 65* (60) 25 ± 44 (17) 103 ± 83* (78)
Referred medical (n = 128) 140 ± 116 (105) 77 ± 135 (45) 217 ± 228 (159)
a history of back
pain. Chiropractic and
CPT indicates Current Procedural Terminology; NA, not applicable.
nonreferred medical
*P < .01 compared with chiropractic costs. patients were generally
similar on most other
characteristics, includ-
Table 3. Patient Characteristics* ing some not shown in
Table 3 (eg, education,
Mean ± SD
occupation). The most
Characteristic Chiropractic Medical (Nonreferred) Medical (Referred) striking difference is
method of payment:
Baseline VAS score 52.0 ± 24.2† 56.1 ± 24.3 59.3 ± 22.1 nearly half of chiro-
Baseline OSW score 41.3 ± 17.4† 47.5 ± 17.5 51.4 ± 17.2‡ practic patients paid
Stage (%) out-of-pocket com-
Chronic 27.1 26.1 42.9† pared with only 7% of
History (%)
With history of back pain 89.7† 84.4 81.9
nonreferred medical
Location (%) ‡ patients (and just 3%
Pain in back only 49.0 44.7 34.6 of those who were
Pain travels into thigh 28.7 32.9 33.9 referred).
Pain travels below the knee 22.3 22.4 31.5 The referred pa-
Smoker (%)
Currently a smoker 23.5 24.8 22.9
tients appear to form a
Depression (%) distinct group with
2 or more weeks in past year 34.4 38.7 29.3‡ more serious back
2 years or more 23.3 24.1 16.3 problems. This group
Much of time in past year 16.7 19.5 14.8 had the highest base-
Socioeconomic line pain and disability
Age (y) 41.4 ± 12.8† 39.3 ± 12.3 39.2 ± 11.8 scores. Significantly
Sex (%)
higher proportions of
Male 50.1 52.0 62.5‡
Race (%) referred patients had
White, non-Hispanic 92.2 92.5 94.2 chronic conditions and
Income (%) † pain traveling below
Less than $12 000 7.1 12.2 5.9 the knee. The high pro-
$12 000-$35 999 37.2 33.2 36.5
portions that were
$36 000-$59 999 30.4 33.7 30.5
More than $60,000 25.4 20.9 27.1 male and covered
Health insurance (%) through workers’ com-
Patient has health insurance 83.8† 89.6 91.2 pensation also stand
Pay (%) † †
Out of pocket 42.1 6.8 2.9
Table 4 shows the
Insurance and other 50.7 86.4 81.4
Workers’ Compensation 7.2 6.8 15.7 change in pain (cal-
culated as the numer-
OSW indicates Revised Oswestry Disability Questionnaire; VAS, visual analog scale.
ical difference between
*Sample sizes vary due to missing observations. For chiropractic: minimum = 1212 (health insurance); end-of-year VAS and
maximum = 1524 (sex); For medical (nonreferred): minimum = 480 (health insurance); maximum = baseline VAS) and
611(age, sex); For medical (referred): minimum = 102 (health insurance, pay); maximum = 128 (age, sex). the change in disabil-

P < .01 compared with medical (nonreferred).

P < .05 compared with medical (nonreferred). ity (calculated as the
Costs of Low Back Care
numerical difference
between end-of-year Table 4. Health Outcomes and Costs
OSW and baseline
Mean ± SD (Median)
OSW). Chiropractic
and nonreferred Outcome Chiropractic* Medical (Nonreferred)† Medical (Referred)‡
medical patients
showed about the Change in VAS score 37.2 ± 28.5 (38.0) 38.7 ± 30.1 (38.0) 27.6 ± 31.8§ (26.5)
same average im- Change in OSW score 26.3 ± 21.0 (24.0) 27.2 ± 26.7 (26.0) 25.0 ± 21.8 (21.0)
provement in VAS, Change in VAS score per dollar §
0.46 ± 0.81 (0.22) 0.75 ± 0.76 (0.57) 0.38 ± 0.61§ (0.20)
whereas the improve-
Change in OSW score per dollar 0.32 ± 0.62§ (0.16) 0.53 ± 0.55 (0.37) 0.35 ± 0.45§ (0.17)
ment for the referred
group was substan-
tially and signifi- OSW indicates Revised Oswestry Disability Questionnaire; VAS, visual analog scale.
cantly lower. All 3 *The number of patients in each outcome group was 916 (change in VAS score and change in VAS score
groups had about the per dollar) and 925 (change in OSW score and change in OSW score per dollar).

same mean out- The number of patients in each outcome group was 366 (change in VAS score and change in VAS score
per dollar) and 371 (change in OSW score and change in OSW score per dollar).
comes according to ‡
The number of patients in each outcome group was 78 (change in VAS score and change in VAS score
the OSW scores: per dollar) and 76 (change in OSW score and change in OSW score per dollar).
none of the small §
P < .01 compared with medical (nonreferred).
differences were sta-
tistically significant.
The changes in the VAS and OSW scores per practic patient were significantly higher than 1-year
dollar also are shown in Table 4. The most relevant costs per medical patient, especially when chiroprac-
comparisons were between the chiropractic and tic patients are compared with medical patients not
nonreferred medical groups. In accordance with referred for further care. The former 2 groups of
our results showing that chiropractic patients patients appeared to be relatively homogeneous,
incurred higher costs and had about the same out- whereas referred medical patients appeared to have
comes as the nonreferred medical patients, the more severe problems. Patient improvement, as mea-
improvements in both VAS and OSW scores per sured by VAS and OSW scores, also was very similar
dollar were significantly lower for the chiropractic for chiropractic and nonreferred medical patients.
group. The mean values were about 60% of the The results found here are consistent with those
nonreferred medical ratios, and the ratios of the reported by Carey et al,20 who also conducted an
medians were even lower. observational study. The authors found that the total
Comparisons of the chiropractic patients with a direct outpatient costs among patients with acute
group consisting of all medical patients (not shown low back pain were highest for those treated by
in Table 4) had little effect on the results because orthopedic surgeons and chiropractors and lowest
the changes in outcomes per dollar for all medical for those treated by primary care providers. Patient
patient groups were just slightly smaller than those outcomes were similar among the 3 groups.
for the nonreferred group. The differences in the However, the costs in our study were considerably
mean changes in VAS or OSW scores per dollar lower than those reported by Carey et al. We used a
between chiropractic and all medical patients Medicare payments standard rather than actual
remain striking (0.46 vs 0.69 for the change in VAS charges, which typically are higher, sometimes much
per dollar, P < .01; 0.32 vs 0.50 for the change in higher, than Medicare reimbursements. Other limita-
OSW per dollar, P < .01). tions of our work that contribute to the lower costs
compared with those in the Carey study also caution
against strong conclusions that are favorable to med-
DISCUSSION ical treatment. Our cost data do not include costs for
imaging or referral services rendered (or indepen-
This project adopted an observational, practice- dently sought by patients) outside the sample
based approach to examine costs and outcomes of providers’ clinics. This explains why the chiropractic
patients treated for acute and chronic ambulatory low patients had no prescription costs.
back pain. Using standardized RVU costing methods, Costs for patients who might have undergone
we found that 1-year direct office costs per chiro- surgery also were not considered. These costs can
dwarf the cost of services in physician clinics or Our 60% response rates are below this threshold
offices. For example, hospital and physician charges value, but they seem adequate to avoid serious bias
per claim processed by a major insurer for surgical in light of the minimal differences between respon-
back hospitalizations averaged $13 990 in 199334 dents and nonrespondents.
(about $18 300 in year 2000 dollars after adjustment Finally, we recognize that more sophisticated
by the medical care services component of the modeling approaches need to be applied, especially
Consumer Price Index). Nonsurgical hospitalizations on longer-term outcomes and costs. In preliminary
averaged $7120 per admission in 1993 or about work, we found that very little variation in 1-year
$9300 in 2000 dollars. The rate of surgery for low costs could be explained (R2 values on the order of
back pain increased substantially in the 1980s,35 and 0.05) for either medical physicians or chiropractic

Use: 0.3661