Abstract
BACKGROUND CONTEXT: The economic burden of low back pain (LBP) is very large and
appears to be growing. It is not possible to impact this burden without understanding the strengths
and weaknesses of the research on which these costs are calculated.
PURPOSE: To conduct a systematic review of LBP cost of illness studies in the United States and
internationally.
STUDY DESIGN/SETTING: Systematic review of the literature.
METHODS: Medline was searched to uncover studies about the direct or indirect costs of LBP
published in English from 1997 to 2007. Data extracted for each eligible study included study design, population, definition of LBP, methods of estimating costs, year of data, and estimates of direct, indirect, or total costs. Results were synthesized descriptively.
RESULTS: The search yielded 147 studies, of which 21 were deemed relevant; 4 other studies and 2
additional abstracts were found by searching reference lists, bringing the total to 27 relevant studies. The
studies reported on data from Australia, Belgium, Japan, Korea, the Netherlands, Sweden, the UK, and
the United States. Nine studies estimated direct costs only, nine indirect costs only, and nine both direct
and indirect costs, from a societal (n518) or private insurer (n59) perspective. Methodology used to derive both direct and indirect cost estimates differed markedly among the studies. Among studies providing a breakdown on direct costs, the largest proportion of direct medical costs for LBP was spent on
physical therapy (17%) and inpatient services (17%), followed by pharmacy (13%) and primary care
(13%). Among studies providing estimates of total costs, indirect costs resulting from lost work productivity represented a majority of overall costs associated with LBP. Three studies reported that estimates
with the friction period approach were 56% lower than with the human capital approach.
CONCLUSIONS: Several studies have attempted to estimate the direct, indirect, or total costs associated with LBP in various countries using heterogeneous methodology. Estimates of the economic costs in different countries vary greatly depending on study methodology but by any
standards must be considered a substantial burden on society. This review did not identify any studies estimating the total costs of LBP in the United States from a societal perspective. Such studies
may be helpful in determining appropriate allocation of health-care resources devoted to this
condition. 2008 Elsevier Inc. All rights reserved.
Keywords:
Introduction
The focus of this special focus issue of The Spine Journal is on the management of chronic low back pain
(CLBP). This topic was chosen partly because of its imposing socioeconomic burden, which appears to be increasing
rapidly despite technological advances in diagnosis and
the introduction of numerous interventions in recent years.
The general pessimism surrounding the prognosis of CLBP
is such that few clinicians, researchers, or third-party
payers would dispute any proposed cost estimate, no matter
how large it may appear. Such views are understandable
when faced with a prevalent, disabling, clinically challenging, and seemingly expensive condition such as CLBP.
However, it is essential to understand the precise magnitude
of the economic burden of CLBP before examining potential cost-saving solutions or comparing the cost effectiveness of competing interventions. This can be achieved by
reviewing cost of illness studies.
Cost of illness studies summarizing the economic burden of a particular disease must be considered by all
stakeholders, including patients, clinicians, and third-party
payers when deciding on the allocation of scarce healthcare resources [1]. It should be noted that cost of illness
studies serves a different purpose than health economic
evaluations (eg, cost-benefit analysis, cost-effectiveness
analysis, cost-utility analysis), which are focused on evaluating the costs of interventions rather than estimating the
cost of a particular disease [1]. The purpose of this study
was to conduct a systematic review of CLBP cost of illness
estimates in the United States and internationally. To help
readers understand some of the basic principles of health
economics pertinent to the studies summarized in this review, a brief overview of important concepts related to cost
of illness studies is presented below.
10
Methods
A search of Medline was conducted on July 1, 2007, for
studies pertaining to the costs of low back pain (LBP) using
the following strategy:
1.
2.
3.
4.
5.
6.
7.
8.
*Back Pain/Economics
low back pain.mp. or exp Low Back Pain/
health care costs.mp. or exp Health Care Costs/
cost of illness.mp. or exp Cost of Illness/
health expenditures.mp. or exp health expenditures/
exp Health Resources/Economics, Utilization
exp Sick Leave/Economics
1 or (2 and (or/3-7))
Cost perspectives
The cost of an illness may be viewed from various perspectives and depends on who bears the costs. Costs could
For studies reporting detailed direct costs, portions attributable to each of the following (or similar) categories
were extracted:
1. Chiropractic (and osteopathy for studies outside the
United States)
2. CAM (eg, acupuncture, homeopathy, massage, and
naturopathy)
3. Emergency department (ED)
4. Imaging
5. Inpatient
6. Mental health
7. Other
8. Outpatient
9. Pharmacy
10. Physical therapy (PT)
11. Primary care
12. Specialists
13. Surgery.
For studies reporting detailed indirect costs, portions attributable to each of the following (or similar) categories
were extracted:
1.
2.
3.
4.
5.
Early retirement
Household
Inactivity
Presenteeism
Sick leave.
11
Results
The search strategy yielded 147 studies, of which 12
(8%) were relevant, 114 (78%) were irrelevant because they
did not meet eligibility criteria, and 21 (14%) were of uncertain relevance based on information contained in the
search records (eg, title, abstract). When full-text articles
were retrieved for the latter group, an additional eight studies were deemed relevant; five additional studies were also
located by searching references of the studies obtained via
Medline. In addition, two relevant conference proceedings
were located, which were only available as abstracts.
The methodology for the 27 relevant studies included in
this review is summarized in Table 1.
Studies from Australia [6], Belgium [7], the Isle of Jersey [8], Japan [9], Korea [10], the Netherlands [2,11], Sweden [4,5,12,13], the United Kingdom [14], and the United
States [1,3,1527] were obtained. All non-US studies
(n512) [2,414] and 5 of the US studies [1,15,2022] examined costs from a societal perspective; the other 10 US
studies used an insurers perspective to estimate costs. Most
[1,3,4,69,11,1316,1927] took a top-down approach and
allocated portions of total costs from national public or private insurer databases according to related international
classification of diseases (ICD)-9 or ICD-10 diagnostic codes. Five top-down studies [4,6,7,13,14] also relied on data
provided in prior surveys or utilization studies to help allocate costs for specific interventions to LBP. Because only
three [2,12,21] had a temporal component (eg, O6 mo) to
their case definition of LBP, it was not possible to limit this
review to studies examining specifically CLBP.
The prevalence of LBP as defined in each study ranged
from 5% to 65%, with a mean of 18.7% and a standard
deviation (SD) of 4.6%. The mean time lag between the
year of data examined to derive cost estimates and the year
of study publication was 4.6 years (SD 0.5), with a range of
1 to 11 years. Four of the US top-down studies [1,17,26,27]
used the Medical Expenditures Panel Survey (MEPS) and
one [21] used its predecessor, the National Medical Expenditures Survey. There were three prospective cohort studies
[2,5,10] that examined costs incurred by a defined group of
patients over time based on health utilization or disability
benefit records. Only one study [2] used a bottom-up approach with self-reported patient diaries documenting the
use and cost of health services related to LBP to supplement information contained in utilization databases. One
cross-sectional study [12] interviewed physicians to inquire
about the perceived use of health services in some of their
patients with LBP, rather than interviewing patients directly. Two studies [17,18] compared the use of health services by those with LBP to a control group of people
without LBP.
National LBP cost of illness estimates uncovered in this
review are summarized in Table 2.
Eight studies from five countriesdAustralia, Belgium,
Japan, Sweden, and the UKdestimated the total national
12
Table 1
Methodology of included cost of LBP studies
Country
Year of
data
Cost types/cost
perspective
[6]
Australia
2001
[7]
Belgium
1999
[8]
1994
[9]
Isle of
Jersey
Japan
1994
[10]
Korea
1997
[11]
Netherlands
1991
[2]
Netherlands
2002
[13]
Sweden
1994
[5]
Sweden
[4]
Sweden
19941995
(2 y)
2001
[12]
Sweden
2002
[14]
UK
1998
[26]
United
States
1996
Direct only
National
[27]
United
States
1996
Indirect only
National
[21]
United
States
United
States
United
States
United
States
United
States
1987
Indirect only
Societal
Indirect only
Societal
Direct only
Societal
Direct only
Societal
Indirect only
Societal
National
National
National
National
[15]
[1]
[17]
[22]
[19]
United
States
1995
1996
1998
2002
2004
Prevalence
(%)
Study population
National
NR
NR
Nonspecific LBP O6 mo
21
National
ICD-10 81
NR
NR
NR
36
Indirect only
Societal
National
National
National
National
National
National
National
National
National
National
National
65
NR
16
NR
NR
22
18
NR
NR
Back pain
15
Reference
NR
21
[16]
[18]
2005
2002
2001
2001
[23]
13
NR
7
Sixty-six related ICD-9 codes
United
States
United
States
United
States
United
States
United
States
[25]
[3]
19971999
Indirect only
Insurer
Direct only
Insurer
Direct only
Insurer
Direct only
Insurer
Direct only
Insurer
NR
NR
United
States
United
States
[24]
1992
14
Table 2
National estimates of total, direct, or indirect costs for low back pain
Total costs
Reference Country
Year
Population
Currency National
[6]
[7]
[9]
[8]
[10]
[11]
[2]
[13]
[5]
[4]
[14]
[15]
[26]
[27]
[21]
[1]
[17]
[22]
[19]
2001
1999
1994
1994
1997
1991
2002
1994
19941995
2001
1998
1995
1996
1996
1996
1996
1998
2002
2004
19,357,954
10,182,034
124,712,000
82,000
45,948,811
15,022,393
16,067,754
8,730,290
8,778,461
8,909,128
58,970,119
260,713,585
263,814,032
263,814,032
263,814,032
263,814,032
270,311,756
280,562,489
293,027,571
AUD
V
Yen
Won
$
V
SEK
V
V
$
$
$
$
$
$
$
$
Australia
Belgium
Japan
Jersey
Korea
Netherlands
Netherlands
Sweden
Sweden
Sweden
UK
United States
United States
United States
United States
United States
United States
United States
United States
Direct costs
Per
capita National
Indirect costs
%
Per
capita National
9,174,931,649 474
1,025,840,000 11
53
1,179,605,000 116
187,005,000 16
18
6,022,403,378
48
2,713,454,390 45
22
d
d
d d
d
d 349,742,900,000 d 7612
d
d
d d
6,418,744,458 399
4,236,371,342 66 264
25,089,000,000 2874
832,000,000 3
95
3,346,300,485 381
234,241,034 7
27
1,860,000,000 209
297,600,000 16
33
12,332,000,000 209
1,632,000,000 13
28
d
d
d d
d
d 14,701,417,650 d
56
d
d
d d
d
d
d d
d
d 12,200,000,000 d
46
d
d 90,600,790,000 d 335
d
d
d d
d
d
d d
8,149,091,649
992,600,000
3,308,948,988
1,287,204
d
4,613,000,000
2,182,373,116
24,257,000,000
3,112,059,451
1,562,400,000
10,700,000,000
13,925,940,000
d
18,533,583,620
28,170,000,000
d
d
19,800,000,000
7,400,000,000
Per
capita
89 421
84
97
55
27
d
16
d
d
d 307
34 136
97 2778
93 355
84 175
87 181
d
53
d
d
d
70
d 107
d
d
d
d
d
71
d
25
based on daily absenteeism and disability costs for the actual duration of work absences, but only $1.5 billion when
assuming a friction period of 3 months and labor elasticity
of 80% (ie, only 80% of productivity is lost during worker
absences because other employees can partially compensate for that loss). The study by Maniadakis and Gray
[14] estimated work productivity losses in the UK of 9.1
billion with the human capital approach based on 116 million lost work days and average earnings of 78, but only
3.4 billion based on a friction period of 90 days, much
shorter than the mean work absence duration of 232 days.
The study by Walker et al. [6] estimated work productivity
losses in Australia at AUD$8.1 billion with the human capital approach based on 62,441,052 lost work days and mean
daily earnings of AUD$130, but only AUD$5.1 billion assuming a friction period of 2.4 months for males and 1.4
months for female. The friction period method yielded estimates that were 56% (SD 9%) lower, on average, than the
human capital approach.
The allocation of indirect costs of LBP is summarized in
Table 4.
Incremental costs
A study analyzed data from the 1998 MEPS, a nationally
representative sample of the US population, to estimate the
cost of medical care provided to those with back pain (ICD9 codes 720724, 805806, 839, 846, 847) [17]. To overcome perceived weaknesses in prior estimates that included
only costs directly related to back pain, this study used multivariate regression to compare the overall costs of medical
care in those with back pain to a similar population without
back pain. Those with back pain had total medical care
15
Table 3
Allocation of direct medical costs for low back pain
CAM
Reference (%)
Chiropractica
(%)
ED Imaging
(%) (%)
Inpatient
(%)
Mental
health
(%)
Other
(%)
Outpatient
(%)
Pharmacy
(%)
PT Primary Surgery
(%) care (%) (%)
Specialists
(%)
Total
(%)
[2]
[12]
[4]
[5]
[13]
[17]
[14]
[18]
[26]
[20]
[7]
[23]
[6]
[24]
Mean
d
16
d
2
d
d
15
d
d
d
d
10
20
3
5
d
d
d
d
d
3
2
4
d
5
d
d
d
d
1
15
30
11
d
28
31
13
14
32
35
d
d
16
7
17
11
d
d
d
d
d
d
1
d
2
d
d
2
1
1
d
6
d
d
d
12
13
1
15
d
d
d
2
20
5
d
d
27
d
27
13
17
11
14
3
4
d
d
d
8
19
6
7
8
16
15
7
20
d
19
19
29
9
2
13
17
17
55
19
d
d
15
d
d
3
61
22
14
20
17
22
d
d
25
7
d
4
24
d
9
d
d
5
d
7
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
9
d
d
d
d
d
d
d
d
d
d
d
16
d
2
d
4
d
7
d
d
5
3
d
2
d
39
6
25
7
8
21
d
15
22
26
9
9
39
22
d
d
11
d
13
d
d
d
24
d
d
d
13
d
d
16
d
d
22
5
CAM5complementary and alternative medicine (includes acupuncture, massage, homeopathy, and similar services); ED5emergency department
(includes ambulance); PT5physical therapy (includes occupational therapy, rehabilitation).
a
Chiropractic includes osteopathy.
Table 4
Allocation of indirect costs for low back pain
Reference
Method
[2]
[12]
[4]
[3]
[15]
[5]
[25]
[11]
Friction
Human capital
Not specified
Not specified
Not specified
Human capital
Not specified
Human capital
Friction
Human capital
Human capital
Friction
Human capital
Not specified
Not specified
Not specified
Not specified
Human capital
Friction
Not specified
Not specified
[13]
[14]
[27]
[19]
[21]
[22]
[7]
[6]
[8]
[24]
Household (%)
Inactivity (%)
31
12
27
Presenteeism (%)
18
31
33
0
62
15
31
85
95
70
Total (%)
42
54
48
100
100
69
100
67
100
38
85
69
100
15
5
30
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
16
coding and subdivided into specific (sciatica, IVD disorders), nonspecific (backache, lumbago), and injuries. Data
were obtained from a health survey and public databases
for health utilization and disability benefits. Length of disability was longer in specific LBP (45.1 d) than nonspecific
back pain (35.8 d) and lowest in back injuries (23.1 d). Less
than 1% of new claims had not returned to work after 1
year. Disability benefit data were skewed, with 35% of
cases returning to work after 1 day and 3% of those disabled for more than 6 months accounting for 33% of benefit
costs.
A study examined the total costs of occupational LBP
(ICD-9 721724, 805, 806, 839.20, 846, 847) in four US
states (Oregon, Illinois, Pennsylvania, and Florida) [24].
Data were obtained from the detailed claim information database that includes claims in which there is compensable
lost work time because of a work-related injury. A random
sample of 520 closed claims was selected from the four
states for the years 1998 to 1992 to examine costs over
time. The distribution of work disability duration was as
follows: !30 days (50%), 30 to 90 days (25%), 91 to 180
days (12%), and O180 days (13%). Costs increased substantially with work disability duration, from $1,146 for !30
days, to $6,286 for 31 to 90 days, $16,284 for 91 to 180
days, and $32,555 for O180 days.
Prescription medication use for LBP
A study in Sweden asked 302 patients who presented to 16
physicians at 14 randomly selected outpatient clinics to complete questionnaires on health-care resource use in the past 6
months [12]. Direct costs were calculated based on unit costs
from hospital, pharmacy, and national health sources. Annual
direct costs per patient for pharmaceuticals were V183 and
patients took an average of two medications for CLBP. The
most common class of medication was analgesics (59%), followed by nonsteroidal anti-inflammatories (NSAIDs) (51%),
muscle relaxants/anxiolytics (11%), proton pump inhibitors
(8%), antidepressants (8%), cyclo-oxygenase 2 (COX-2) inhibitors (5%), antacids (2%), prostaglandins (1%), H2 antagonists (1%), and others (1%).
A study was conducted to examine health-care utilization
in patients with mechanical LBP as defined by one of 66 related ICD-9 codes [20]. Data were obtained from utilization
records of Kaiser Permanente Colorado, a health maintenance organization with 410,000 enrollees in the Denver
area. Costs were measured from the insurers perspective using actual costs in 1999, adjusted to 2005 with the consumer
price index. Pharmacy records indicated that 31% of patients
had a claim for NSAIDs and 29% for opioids.
A study was conducted to estimate the direct medical
costsdfocusing on analgesic medication usedfor mechanical LBP as defined by one of 66 ICD-9 codes [23]. Data
were obtained from utilization records of 255,958 commercial members enrolled in the University of Pittsburgh Medical Center Health System in 2001. Costs were examined
from the health plans perspective. Among the 9,417 patients (56%) with a pharmacy claim for analgesics related
to their LBP, the most frequent analgesic was opioids alone
(33%), followed by NSAIDs alone (27%), opioids and
NSAIDs (26%), opioid and other analgesics (9%), or
COX-2 inhibitors alone (3%). Most opioid use was short
term, with a duration of 1 to 30 days (71%), though a substantial portion had a duration of 31 to 90 days (14%).
Smaller proportions of patients used opioids for 90 to 179
days (6%), and more than 180 days (9%). Most costs for
analgesics were for opioids (61%), followed by COX-2 inhibitors (23%), NSAIDs (13%), and other (2%). Mechanical LBP represented 48% of total health system costs for
opioids, 24% of costs for NSAIDs, and 28% of costs for
COX-2 inhibitors.
Nonconventional health-care use
A study examined the costs of health care for individuals
with back pain (related ICD-9 codes) [16]. Utilization data
were obtained from two large insurers in the state of Washington for the year 2002. Of the 497,597 eligible insured,
104,358 (22%) made 652,593 visits related to LBP that represented 15% of all outpatient visits. Most visits were made
to chiropractors (49%), followed by medical doctors (20%),
massage (13%), PT (11%), other providers (11%), acupuncture (2%), naturopathy (1%), and nurse practitioner/
physicians assistant (1%). The largest group of insured
(45%) sought care from only conventional providers and
had mean costs of $506, whereas those who sought care
from only CAM providers (43%) had mean costs of
$342, and a third group of insured who saw both conventional and CAM providers (12%) had mean costs of
$1,079. The average cost allowed per visit was $128 for
conventional providers and $50 for CAM providers.
Health-care utilization assumptions for LBP
Several studies using top-down approaches made assumptions about health-care utilization for LBP to allocate
costs to that illness. These assumptions are summarized below to provide insight into study methodology and additional information about the proportion of specific health
resources consumed by patients with LBP. A study in Sweden [4] assumed that 2.5% of total outpatient visits were the
result of LBP, 42% of all public PT visits and 60% of private PT visits were the result of LBP, and 25.0% of pharmaceutical costs for musculoskeletal diseases were for LBP. A
study was undertaken to estimate the total costs of back
pain (ICD-9 codes 702724, 846847) in the UK in 1998
[14]. Data were obtained from various sources, including
prior surveys on the epidemiology of back pain, prior studies
on the use of care for back pain, and health utilization claims
from the national and private health-care databases. It was
estimated in this study that of those with back pain, 36%
use topical creams and sprays, 19% use over-the-counter
17
Discussion
Numerous LBP cost of illness studies were identified in
this review and provided valuable information for readers
18
19
Conclusion
This review identified several studies that have previously attempted to estimate the direct, indirect, or total
costs associated with LBP, both in the United States and internationally. Study methodology differed considerably,
making direct cost comparisons across studies and between
countries difficult. The largest components of direct medical
costs were PT, inpatient services, pharmacy, and primary
care. From studies conducted outside the United States, it
appears that direct medical costs represent only a small portion of the total costs of LBP, suggesting that interventions
that are able to reduce LOD may present the opportunity
for cost savings from a societal perspective. Further studies
are required in the United States to estimate the total costs
of LBP and inform decision making for this complex and
challenging condition.
References
[1] Druss BG, Marcus SC, Olfson M, Pincus HA. The most expensive
medical conditions in America. Health Aff (Millwood) 2002;21:
10511.
[2] Boonen A, van den HR, van TA, et al. Large differences in cost of illness
and wellbeing between patients with fibromyalgia, chronic low back
pain, or ankylosing spondylitis. Ann Rheum Dis 2005;64:396402.
[3] Goetzel RZ, Hawkins K, Ozminkowski RJ, Wang S. The health and
productivity cost burden of the top 10 physical and mental health
20
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]