Spine J 2008 (Jan); 8 (1): 8–20 ~ FULL TEXT
Simon Dagenais, DC, PhDb, Jaime Caro, MDd, Scott Haldeman, DC, MD, PhD
Division of Orthopaedic Surgery and Department of
Epidemiology and Community Medicine,
Faculty of Medicine,
University of Ottawa, Ottawa, ON, Canada.
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 (n==18) or private insurer (n==9) 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.
From the FULL TEXT Article:
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.  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.  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.
Cost of illness studies
The economic burden of a disease is the sum of all costs
associated with that condition which would not otherwise
be incurred if that disease did not exist. Given the many categories of costs that must be considered, it can be challenging to fully estimate the economic burden of an illness as data are often unavailable. The term "cost" in health economics refers to the value of the consequences of using
a particular good or service rather than its price. That value
corresponds to the best alternative use of those resources,
which is termed the opportunity cost. The difference between
cost and price can be demonstrated by comparing
what a clinician charges for his services versus the amount
actually reimbursed for those services; the latter may be
a better proxy of true cost than the former. Despite this example, it should be made clear that estimating the economic
burden of a disease is not simply a matter of tabulating the
amount reimbursed for all clinician services related to a particular diagnosis. The total cost of illnessdor economic
burdendhas three components: (1) direct (medical and
nonmedical) costs; (2) indirect costs; and (3) intangible
Direct costs refer to those that, at least in principle, involve
a monetary exchange. Direct medical costs are most
familiar to readers and commonly include costs incurred for
physician services, medical devices, medications, hospital
services, diagnostic testing, etc.  The term direct healthcare
costs may be more accurate because allied health,
complementary and alternative medicine (CAM), and other
nonphysician costs are included. Direct medical costs are
typically the easiest to estimate because records are kept
of such transactions by clinicians, third-party payers,
employers, or patients.
Direct nonmedical costs are those related to goods and
services consumed directly because of the illness but which
are not considered to be health care related. They include,
for example, transportation or other travel costs to attend
medical appointments, meals eaten outside the home when
receiving health care, renovations to make a house more accessible for those who may become physically disabled
secondary to a disease, and so on. These expenses are easy
to overlook when considering the economic burden of a disease
but can constitute an important source of related costs.
Indirect costs are those reflecting the economic value of
consequences for which there is no direct monetary transfer.
They commonly include costs related to employment
and household productivity. Employment costs include
both work absences resulting in foregone productivity
(termed "absenteeism") and decreased productivity for
those who continue to work despite being affected by their
condition (termed "presenteeism").  Indirect costs are
often more difficult to measure than direct costs. For example,
it is nearly impossible to determine whether presenteeism
is in fact occurring and which medical condition, if any,
is primarily responsible for its occurrence. Such productivity
losses can be estimated by interviewing workers and
asking them to evaluate their own reduced productivity as
it relates to their health status, or by interviewing their
supervisors and asking for their opinions about whether
a particular worker’s productivity has changed secondary
to a disease. Rarely can such presenteeism costs be measured
objectively or with certainty. Absenteeism costs are
generally easier to estimate because employer records and
workers’ compensation insurance systems will note
health-related work absences.
Human capital and friction cost approaches
There is debate among economists about the preferred
methods for estimating indirect costs, the majority of which
are often composed of lost productivity.  The most common
method is the human capital approach, which assumes
that the economic value of an employee’s productivity is
equal to the cost of their salary and benefits.  Lost productivity is therefore estimated by calculating earnings lost during work absences, regardless of the length of absence. This method would therefore calculate lost productivity for disability-related early retirement as the value of lost earnings from the date of disability until that worker would have reached normal retirement age (eg, 65 y). Employers, however, will eventually replace those employees that are absent for extended periods and regain the value of the original employee’s lost productivity.
Thus, productivity losses are truly only incurred from the worker’s absence for health reasons until a replacement worker is hired, trained, and has reached the productivity of the disabled employee. The length of time required for this transition to occur has been termed the friction period and varies for different industries and economic conditions based on factors that may influence the time required to replace an employee (eg, unemployment level, job skill and education requirements, and location).  Evaluating the cost of lost productivity with
this assumption has been termed the friction cost approach.  Alternatively, the cost of hiring replacement workers for employees on disability leave may also be used to estimate lost productivity.  A combination of methods may
be most appropriate to estimate indirect costs by, for example,
using the human capital approach for temporary productivity
losses from absences that are shorter than the
friction period, and using the friction cost approach for
longer-term productivity losses.
In addition to lost work productivity, individuals with an
illness may also incur productivity losses at home if they
are unable to complete routine household tasks (eg, cleaning,
cooking, and maintenance). This is true whether they
must rely on paid outsiders to complete such tasks on their
behalf, or whether unpaid household members must do so.
These indirect costs can be estimated using the earnings of
a hired household worker or by using an estimate of the national
or regional value of leisure time, which is typically
less than the mean hourly wage.
The third type of cost that may be considered when estimating
the total cost of illness for a particular disease is
termed intangible costs. These costs reflect the value of decreased enjoyment of life because of illness. However,
these costs are rarely included when estimating the economic
burden of an illness because of general societal discomfort
with placing a monetary value on these aspects of a disease.
The cost of an illness may be viewed from various perspectives
and depends on who bears the costs. Costs could be estimated from a patient perspective (eg, out-of-pocket costs), employer perspective (eg, cost of worker’s compensation insurance premiums and lost productivity), insurance company perspective (eg, cost of claims paid), government perspective (eg, cost of public health services), or societal perspective, which would include all related costs. The latter is the most comprehensive and broadest perspective and avoids underestimating costs when only considered from a narrower perspective.
Sources of cost data
The perspective chosen will also impact the types and
sources of data used in the cost of illness study. Employer,
insurer, or government perspective may derive acceptable
cost estimates by apportioning claims in large utilization databases to specific diagnoses, which has been termed a top-down approach.  Alternatively, the patient perspective may extrapolate costs from interviews or diaries about health-care utilization and costs, which has been termed a bottom-up approach.  The societal perspective
could combine both approaches to capture more
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. *Back Pain/Economics
2. low back pain.mp. or exp Low Back Pain/
3. health care costs.mp. or exp Health Care Costs/
4. cost of illness.mp. or exp Cost of Illness/
5. health expenditures.mp. or exp health expenditures/
6. exp Health Resources/Economics, Utilization
7. exp Sick Leave/Economics
8. 1 or (2 and (or/3-7))
The search was limited to studies published in English
from 1997 to 2007 to uncover relatively recent cost of illness
studies. Search results were screened for relevance
according to the eligibility criteria outlined below.
1. Back pain or LBP
2. Monetary estimate of direct or indirect costs
3. Societal or insurer perspective.
1. Neck pain only
2. Economic evaluations of interventions (eg, cost-effectiveness
3. Details concerning only some specific direct costs
(eg, prescription costs)
4. Indirect costs expressed as lost work days without
5. Elderly patients only (eg,O65 y)
6. Review articles without new data.
If information provided in search results was insufficient
to determine study eligibility, full-text articles were retrieved for further screening. The following data were extracted for all included studies:
1. Case definition of LBP
2. Cost perspective
3. Cost types
5. Estimated direct costs
6. Estimated indirect costs
7. Estimated total costs
8. Prevalence of LBP according to case definition
9. Study design and data sources
10. Study population
11. Year of data.
For studies reporting detailed direct costs, portions attributable to each of the following (or similar) categories
1. Chiropractic (and osteopathy for studies outside the
2. CAM (eg, acupuncture, homeopathy, massage, and
3. Emergency department (ED)
6. Mental health
10. Physical therapy (PT)
11. Primary care
For studies reporting detailed indirect costs, portions attributable to each of the following (or similar) categories
1. Early retirement
5. Sick leave.
Because of considerable methodological heterogeneity, results were synthesized descriptively.
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 , Belgium , the Isle of Jersey , Japan , Korea , the Netherlands [2, 11], Sweden [4, 5, 12, 13], the United Kingdom , and the United
States [1, 3, 15–27] were obtained. All non-US studies (n=12) [2, 4–14] and 5 of the US studies [1,15,20–22] examined costs from a societal perspective; the other 10 US studies used an insurer’s perspective to estimate costs. Most [1, 3, 4, 6–9, 11, 13–16, 19–27] 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  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  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  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
National LBP cost of illness estimates uncovered in this
review are summarized in Table 2.
Eight studies from five countries – Australia, Belgium,
Japan, Sweden, and the UK – estimated the total national
economic burden of LBP. [2, 4–7, 9, 13, 14] An additional four [1, 10, 17, 26] reported only national estimates of direct costs, whereas seven [8, 11, 15, 19, 21, 22, 27] reported only national indirect costs. Most results are reported in the study’s local currency, though one study from the Netherlands  reported its estimates in US dollars. Country populations according to the year of the study are provided for comparisons. Three of the eight studies reporting combined direct and indirect costs of LBP originated in Sweden. [4, 5, 13] The two Swedish studies reported in euros (V) used different designs (prospective cohort  vs. top-down database approach [4, 12]), different definitions of LBP (low back or neck problems with minimum 28-d work incapacity  vs. not defined ), and reported different results (V3.35 billion over 2 y  vs. V1.86 billion ). In the eight studies that provided estimates of both direct and indirect costs, mean direct costs accounted for only 22% of total costs, indicating that indirect costs are a much larger contributor to the cost of LBP. There were no national estimates in the United States that combined direct and indirect costs associated with LBP.
Allocation of direct medical costs for LBP
The allocation of direct medical costs for LBP is summarized
in Table 3.
Fourteen studies reported on the allocation of direct
medical costs to various categories of health services [2, 4–7, 12–14, 17, 18, 20, 23, 24, 26] used for LBP. The categories differed slightly among studies and were combined where appropriate to summarize results. For example, costs for the categories general practitioner, primary care, and nonspecialist office visits were combined under the rubric of ‘‘primary care,’’ and costs for the categories X-rays and computed tomography/magnetic resonance imaging were combined into ‘‘imaging.’’ In one study reported in
an abstract , only three categories were mentioned (office based, inpatient, outpatient); the remainder of costs was classified as ‘‘other.’’ The number of categories to which medical costs were allocated ranged from 4 to 10 with a mean of 6.7. The most commonly reported category of direct medical costs was pharmacy (n=13), followed by inpatient (n=511), PT (n=510), primary care (n=510), imaging
(n=58), outpatient (n=58), specialists (n=57), other (n=57),
chiropractic (n=56), mental health (n=55), ED (n=54), surgery
(n=54), and CAM (n=52). Only one study  specified that its direct medical cost estimate did not include out-ofpocket expenses, over-the-counter medication, CAM, or mental health. It was unclear in many instances how or why only specific categories of direct medical costs were included in estimates.
The largest mean cost components for care of LBP were
PT (17%) and inpatient services (17%), followed by pharmacy
(13%) and primary care (13%), outpatient services
(8%), diagnostic imaging (7%), specialists (7%), surgery
(5%), chiropractic (including osteopathy) (5%), other
services (5%), CAM (2%), ED (1%), and mental health
(1%). Mean outpatient physician services (primary care,
specialists) in these studies accounted for only 20% of
costs, whereas nonphysician services (PT, chiropractic,
CAM, mental health, other) accounted for 30%. Hospital
costs (inpatient, outpatient, surgery, ED) were 30% of overall
direct medical costs, whereas pharmacy costs were more
than twice those of diagnostic imaging. It should be noted
that these means assumed that there were no costs in categories
for which no estimates were reported; these estimates
were therefore influenced by the number of studies
that included a particular category of direct medical costs.
Allocation of indirect costs for LBP
Eighteen studies [2–8, 11–15, 19, 21, 22, 24, 25, 27] estimated indirect costs associated with LBP, and five categories of indirect costs were considered. The most common was sick leave (n=518), followed by early retirement (n=56), lost household productivity (n=53), presenteeism (n=53), and inactivity (n=51). Methodology differed considerably, though the majority of studies used methods consistent with the human capital approach by estimating indirect costs according to lost wages as determined by disability payments. Only one study  used solely a friction period approach, where work productivity losses were valued using a 22-week friction period and a national average wage of V9-18 per hour depending on age and gender.
Three studies [6, 11, 14] used both methods and in all instances, the cost estimates derived were much lower with the friction cost approach than the human capital approach. The study by Hutubessy et al.  estimated work productivity
losses at $4.6 billion with the human capital approach
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  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.  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
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 (ICD-
9 codes 720–724, 805–806, 839, 846, 847).  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 costs that were US$1,015 greater (US$3,498 vs. US$2,178) than those without back pain. Incremental medical care costs attributable to back pain totaled US$26 billion and included higher costs for office-based visits (US$11.1 billion),
outpatient services (US$4.7 billion), inpatient care
(US$4.5 billion), prescription drugs (US$3.9 billion), and
ED visits (US$1.1 billion). Cost distribution was skewed,
with patients in the highest 10% of costs accounting for
99% of inpatient care costs, 90% of ED visits, 87% of outpatient
services, 53% of prescription drugs, and 52% of
Another study examined direct inpatient, outpatient, and
pharmacy costs for chronic back pain (ICD-9 codes 722,
724, 738, 739, 756, 846, 847) among members of a health
maintenance organization.  Data were obtained from utilization records of the Lovelace Health Plan with 240,000 members in New Mexico. Costs were examined from the health plan’s perspective using charges appearing on claims in 2001. There were 3,864 patients with at least two outpatient or one inpatient claim with chronic back
pain as the primary or secondary diagnosis during the 12-
month study period. Utilization records for that group were
compared with 11,557 controls without chronic back pain
or arthritis to estimate incremental use attributable to
chronic back pain. Patients with chronic back pain were
three times more likely (odds ratio 3.08; 95% confidence
interval 2.70 to 3.53) to be admitted to hospital than controls.
Total incremental costs for the 3,864 patients with
chronic back pain were estimated at US$14.5 million.
A study was conducted to examine lost productivity related
to exacerbations of back pain in US workers 
within the broader Caremark American Productivity Audit,
a telephone survey of lost productivity because of pain conditions in a nationally representative random sample of
adult US workers previously reported above.  The 2-week prevalence of back pain in the survey was 15%, of whom 17% reported lost productivity. There were 320 participants aged 40 to 65 years with self-reported clinically meaningful LBP in the previous 2 weeks, and 91 ageand gender-matched controls without back pain or arthritis.
Lost productivity costs for any health reason in those with
back pain were estimated at $23.5 billion, and were $16.9
billion higher than controls without back pain or arthritis.
Length versus cost of claims for LBP
A study examined workers’ compensation claims for
LBP for the years 1988, 1990, 1992, 1994, and 1996 from
an insurer representing 10% of the private US workers’
market.  The database included both direct costs for medical claims and indirect costs for wage replacement while on disability leave. Claims for LBP were compared with non-LBP claims over 8 years to note general trends. Although the proportion of LBP claims did not change
from 1988 (15%) to 1996 (15%), the proportion of costs
for LBP decreased from 38% in 1988 to 23% in 1996.
The length of disability (LOD) claims for 1996 were
heavily skewed, with an average of 61 days and a median
of 0 days. Costs were also skewed, with 60% of LBP claims
in 1996 costing less than $500 and 12% costing more than
$5,000. There was a clear association between LOD and
cost because 86% of all claims with LOD less than 1 month
in 1996 accounted for only 11% of all costs, and 5% of
claims with LOD for longer than 1 year accounted for
65% of total costs.
A study was conducted to examine the association between
duration of treatment and costs of occupational
LBP in Korea using 9,277 claims from a worker’s compensation
database.  Duration and costs of claims were skewed, with the 51% of claims less than 6 months accounting for only 10% of costs and the 6% of claims longer than 2 years accounting for 29% of costs. After 50 months, 418 (4.5%) claims were still open.
A study was undertaken to estimate the costs of disability
pay resulting from back pain due on the isle of Jersey in
1994, a Crown possession of the UK, which operates under
its own system of health care and disability benefits.  Back pain was defined according to proprietary diagnostic 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
A study examined the total costs of occupational LBP
(ICD-9 721–724, 805, 806, 839.20, 846, 847) in four US
states (Oregon, Illinois, Pennsylvania, and Florida).  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%), andO180 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 >180 days.
Prescription medication use for LBP
Astudy 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.  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 (n=SAIDs) (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.  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 insurer’s 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.  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 plan’s 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
Nonconventional health-care use
A study examined the costs of health care for individuals
with back pain (related ICD-9 codes).  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/physician’s 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  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 702–724, 846–847) in the UK in 1998.  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 medication, 10% use outpatient clinics, 10% have X-rays,
9% use PT for an average of seven visits, 5% visit osteopaths
for an average of five visits, 3% visit private consultants,
2% visit chiropractors for an average of five visits,
2% use the ED, 1% visit occupational therapists, 1% visit
acupuncturists, and 2% visit other specialists. A study
was conducted to estimate the direct medical costs for
LBP in Belgium in 1999.  Data were gathered from a variety of sources including national health utilization databases and prior health utilization studies. Assumptions made regarding the use of health services for LBP were that 29% of all sick days are the result of LBP, 56% of PT is for LBP, 55% of transcutaneous electrical nerve stimulation units are for LBP, 70% of epidural steroid injections are for LBP, and 60% of intrathecal pumps are for LBP.
Costliest health conditions
A study was undertaken to estimate the 15 most costly
conditions in America using results from the 1996 MEPS.  Back problems (n=o definition provided) were ranked sixth in national costs at $12.2 billion, behind ischemic heart disease, motor vehicle accidents, acute respiratory infections, arthropathies, and hypertension. Using a summary disability score calculated from mean ranks in number of bed days, missed work days, and rates of impairment in activities of daily living and instrumental activities of daily living, back problems were ranked third in associated disability, behind arthropathies and mood disorders. This study suggested that resources devoted to back problems
(#6) are proportionally smaller than its disability toll (#3).
A study was conducted to determine the 20 costliest health
conditions for employers in the United States using databases
from six large corporations with 374,799 employees in 43
states.  Costs were allocated to specific conditions based on proprietary algorithms related to ICD-9 codes and temporal patterns of utilization. Mechanical LBP (n=o definition provided) was the fourth most expensive health condition behind angina pectoris, essential hypertension, and diabetes mellitus, with an estimated prevalence of 52.7 per 1,000 employees and costs of $90.24 per employee.
A study examined the indirect costs associated with
common pain conditionsdincluding back paindin the
US workforce.  Data were obtained from the Caremark American Productivity Audit, a telephone survey of lost productivity because of pain conditions in a nationally representative random sample of adult US workers. Using the human capital approach, indirect costs for back pain were estimated at $19.8 billion and were higher than those of headache ($19.6 billion), arthritis ($10.3 billion), and other pain conditions ($11.6 billion) considered in the survey.
Numerous LBP cost of illness studies were identified in
this review and provided valuable information for readers
to understand the magnitude of its economic burden from
different perspectives in various countries. Despite differences
in study methods producing a wide range of cost of
illness estimates, it was apparent that LBP represents an
important economic burden wherever it was studied. One
of the most notable findings in this review is that the methodology used to derive LBP cost of illness estimates varied
considerably among the 27 studies examined. More importantly,
methodological differences that may not always be
readily apparent to LBP stakeholders appeared to exert
a substantial influence on the resulting cost of illness estimates. Some of these methodological differences that
should be considered when appraising cost of illness studies
related to LBP include the cost perspective adopted, delays
in publication, case definitions of LBP, sources of data for
estimating direct medical costs, direct medical cost categories
considered, assumptions about health utilization to
apportion costs, total versus marginal medical costs, and
approaches used to estimate indirect costs from lost productivity. Each is briefly discussed below.
Comparisons of cost of illness estimates between countries
were severely limited by differences observed in cost
perspectives adopted in different studies. Although the majority
of economic burden studies originated from the US,
no studies attempted to accurately estimate the national total
cost of illness of LBP from a societal perspective. None
of the studies conducted outside the United States examined
costs solely from a private insurer’s perspective,
whereas several of the studies from the United States did
so. Although costs of illness estimates from a narrower perspective are no doubt informative for insurers who sponsor
such studies, they are of limited value when attempting to
understand the economic burden of a disease from a societal
perspective. Of particular concern is that any decreases observed over time because of cost containment efforts by insurers could simply indicate that costs are simply being
shifted to other payers. The importance of private health insurers in the United States should not preclude examining
costs from a societal perspective, as evidenced by a study
in the UK that included direct medical costs from both public
and private health insurers. Differences in perspectives
adopted and health-care systems made it difficult to interpret
the wide variations in the direct or indirect costs of
LBP among the countries studied.
Delays between data collection and study publication
should be examined closely to ensure that readers do not
draw conclusions on outdated data. For example, two studies
published in the same year could be reporting estimates
based on data from as little as 2 years ago or much as 11
years ago. The year of publication is therefore not sufficient
to identify how current the cost estimate might be. Although
some authors may attempt to compensate for these
delays by adjusting older cost data for inflation with a consumer price index, such methods are likely prone to error
and will fail to capture other changes that may have
occurred in the ensuing years.
The case definition of LBP differed considerably among
the studies. Some of the studies specified that algorithms
were used to estimate the probability of medical service utilization being related to LBP based on 66 ICD-9 codes,
whereas others did not provide any operational definition
of LBP. Comparisons across studies must therefore pay attention
to whether the cost estimate was for all spinal pain
(eg, neck and back problems) or only subsets of LBP (eg,
trauma-related occupational LBP). Because much of the
difficulty surrounding the management of CLBP is related
to nonspecific pain of mechanical origin, estimates focusing
on that category of disease appear warranted. It was
also not possible to examine the temporal component of
LBP because few studies included this aspect in their case
definition. Evidence that the costs of LBP are heavily
skewed by the most expensive claims with the longest duration
suggests that most costs for LBP are for CLBP.
The sources of data on which estimates of direct medical
costs were based varied among the studies reviewed, most
of which used public or private health insurer claims databases
(top-down approach). Although a few bottom-up studies were conducted in Europe using patient diaries or interviews, there were no such studies identified in the United States to capture out of pocket medical or other expenses that contribute to the overall cost of LBP and may be missed by top-down approaches.
Such studies could help validate some of the assumptions
made regarding the use of specific interventions by
patients with LBP. For example, although it has been reported
in the United States that 31% of patients with LBP who seek care will consult a chiropractor , a study in the UK based its LBP cost estimates on the assumption that only 2% of those seeking care for LBP would use chiropractic. It would be important to understand the reasons for such discrepancies in utilization estimates before making assumptions when estimating direct medical costs attributable to LBP. Bottom-up studies could also provide data to support prior utilization surveys, which have reported that back pain was the leading reason for using interventions such as CAM, on which a total of $27 billion was spent out of pocket in 1997 ; only two studies included costs related to CAM in their estimates.
The number of categories of direct medical costs that
were considered by a study was likely to influence the magnitude
and precision of their estimates. Whereas many studies
reported costs related to PT, primary, outpatient, and
inpatient care, few reported costs for surgery, mental health,
ED, allied health, chiropractic, or CAM. Given that studies
which included these latter categories did in fact reported
associated costs, studies limited to the fewer categories
are likely to have underestimated the true direct medical
costs of LBP. Because care of LBP is often decentralized,
more comprehensive direct medical cost estimates should
include multiple categories of health providers and interventions, whether reimbursed by national or public insurers
or paid for by patients out of pocket.
Another challenge that was apparent in estimates based
on top-down approaches from utilization databases was determining whether costs for those with a history of LBP
were in fact receiving specific medical services for that
condition. Patients who develop LBP are likely to continue
receiving care for other medical conditions (eg, hypertension,
diabetes, headache). It would therefore be incorrect to attribute all medical costs for those with LBP solely to
that condition. A few studies attempted to address this issue
by comparing health utilization to age- and gender-matched
control groups without LBP. Estimates for incremental –
rather than total – costs attributed to LBP were substantially
lower making this an important consideration for future LBP cost of illness studies. This issue may also be addressed by collecting data directly from patients about the primary reason for seeking a particular medical service.
There were also notable differences in the assumptions
made to estimate the indirect costs of LBP attributable to
lost productivity, which was by far the largest component
of costs in studies reporting both direct and indirect costs.
The three studies that compared estimates of the human
capital and friction period approaches reported much lower
cost estimates with the latter. This difference should be explored in future studies because it is unclear whether assumptions made in this approach are valid with LBP, where injured workers may have repeated work absences of varying durations or be temporarily or permanently assigned to less physical job duties. Both scenarios may complicate or lengthen the process of hiring a replacement worker and therefore potentially impact the length of the friction period and cost of lost productivity.
Regardless of the differences in methods used to estimate
costs related to LBP, studies reviewed suggest that
the greatest cost savings from a societal perspective may
be obtained from interventions that promote early return
to work and minimize lost productivity. However, a potential
difficulty that may occur when attempting to implement
such strategies without a clearer understanding of all cost
components is that cost savings from reducing lost productivity
would be primarily observed from an employer perspective,
whereas costs related to interventions that facilitate an early return to work (eg, exercise, fear avoidance training) would be primarily borne from a health insurer perspective. Cooperation among injured workers, clinicians, employers, and third-party payers would be required to minimize such potential conflicts when evaluating competing interventions that may be more favorable from a narrower cost perspective but detrimental from a societal
There were no studies identified in this review that provided
a total LBP cost of illness estimate for the United
States. Those wishing to estimate the economic burden
for LBP in the United States may combine results of direct
($12.2–$90.6 billion) and indirect ($7.4–$28.2 billion)
costs to arrive at a combined total that ranges from $19.6
to $118.8 billion. However, estimates derived this way
are not materially different than those proposed in 1991
(eg, $50–100 billion) and are therefore likely inaccurate.  Alternatively, the total costs could be estimated by applying the median proportion of direct (14.5%) versus indirect (85.5%) costs obtained from eight international studies to those obtained in the United States. Assuming that US direct cost estimates of $12.2 to $90.6 billion represent only 14.5% of total costs, the total US costs attributable to LBP could be estimated at $84.1 to $624.8 billion, which is substantially higher than previous estimates. However, the assumption that the median ratio of direct:indirect costs obtained in other countries can be applied to the United
States has not been validated. Given the magnitude of the
problem, a more precise estimate of total costs of LBP in
the United States from a single study with robust methodology
appears necessary and worthy of additional research.
Only then can comparisons between countries be explored
to determine the effects (if any) of social, cultural, economic,
and health-care system differences on the relative
cost of illness of LBP.
The findings of this study are limited by having examined
only one biomedical database (Medline) for studies
published in English in the last 10 years. It is possible that
a broader search with multiple databases and no language or
date restrictions may have uncovered additional LBP cost
of illness studies with different results than those included
in this review.
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
Druss BG, Marcus SC, Olfson M, Pincus HA.
The most expensive medical conditions in America.
Health Aff (Millwood) 2002;21: 105–11.
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;64:396–402.
Goetzel RZ, Hawkins K, Ozminkowski RJ, Wang S.
The health and productivity cost burden of the ‘‘top 10’’ physical and mental health conditions
affecting six large U.S. employers in 1999.
J Occup Environ Med 2003;45:5–14.
Ekman M, Johnell O, Lidgren L.
The economic cost of low back pain in Sweden in 2001.
Acta Orthop 2005;76:275–84.
Hansson EK, Hansson TH.
The costs for persons sick-listed more than one month because of low back or neck problems.
A two-year prospective study of Swedish patients. [see comment].
Eur Spine J 2005;14:337–45.
Walker BF, Muller R, Grant WD.
Low back pain in Australian adults. Health provider utilization and care seeking.
J Manipulative Physiol Ther 2004;27:327–35.
Van ZJ, van KM.
Low back pain: from algorithm to cost-effectiveness?
Pain Pract 2005;5:179–89.
Watson PJ, Main CJ, Waddell G, Gales TF, Purcell-Jones G.
Medically certified work loss, recurrence and costs of wage compensation for back pain:
a follow-up study of the working population of Jersey.
Br J Rheumatol 1998;37:82–6.
Shinohara S, Okada M, Keira T, Ohwada M, Niitsuya M.
Prognosis of accidental low back pain at work.
Tohoku J Exp Med 1998 Dec;186:291–302.
Kim HS, Choi JW, Chang SH, Lee KS, Oh JY.
Treatment duration and cost of work-related low back pain in Korea.
J Korean Med Sci 2005;20:127–31.
Hutubessy RC, van Tulder MW, Vondeling H, Bouter LM.
Indirect costs of back pain in the Netherlands: a comparison of the human capital method with the friction cost method.
Pain 1999;80(1–2): 201–7.
Ekman M, Jonhagen S, Hunsche E, Jonsson L.
Burden of illness of chronic low back pain in Sweden: a cross-sectional, retrospective study in primary care setting.
Jonsson D, Husberg M.
Socioeconomic costs of rheumatic diseases. Implications for technology assessment.
Int J Technol Assess Health Care 2000;16:1193–200.
Maniadakis N, Gray A.
The economic burden of back pain in the UK.
Guo HR, Tanaka S, Halperin WE, Cameron LL.
Back pain prevalence in US industry and estimates of lost workdays.
Am J Public Health 1999;89:1029–35.
Lind BK, Lafferty WE, Tyree PT, Sherman KJ, Deyo RA, Cherkin DC.
The role of alternative medical providers for the outpatient treatment of insured patients with back pain.
Luo X, Pietrobon R, Sun SX, et al.
Estimates and Patterns of Direct Health Care Expenditures Among Individuals
with Back Pain in the United States
Spine (Phila Pa 1976) 2004 (Jan 1); 29 (1): 79–86
Mapel DW, Shainline M, Paez K, Gunter M.
Hospital, pharmacy, and outpatient costs for osteoarthritis and chronic back pain.
J Rheumatol 2004;31:573–83.
Ricci JA, Stewart WF, Chee E, Leotta C, Foley K, Hochberg MC.
Back pain exacerbations and lost productive time costs in United States workers.
Ritzwoller DP, Crounse L, Shetterly S, Rublee D.
The association of comorbidities, utilization and costs for patients identified with low back pain.
BMC Musculoskelet Disord 2006;7:72.
Rizzo JA, Abbott TA III, Berger ML.
The labor productivity effects of chronic backache in the United States.
Med Care 1998;36:1471–88.
Stewart WF, Ricci JA, Chee E, Morganstein D, Lipton R.
Lost Productive Time and Cost Due to Common Pain Conditions in the US Workforce
JAMA 2003 (Nov 12); 290 (18): 2443–2454
Vogt MT, Kwoh CK, Cope DK, Osial TA, Culyba M, Starz TW.
Analgesic usage for low back pain: impact on health care costs and service use.
Williams DA, Feuerstein M, Durbin D, Pezzullo J.
Health care and indemnity costs across the natural history of disability in occupational low back pain.
Hashemi L, Webster BS, Clancy EA.
Trends in disability duration and cost of workers’ compensation low back pain claims 1988–1996.
J Occup Environ Med 1998;40:1110–9.
Mychaskiw MA, Thomas J.
Direct costs of back pain in the United States: a national estimate.
Value Health 2002;5:508–9. Ref Type: Abstract.
Mychaskiw MA, Thomas J.
Indirect costs due to back pain in the United States.
Value Health 2002;5:136. Ref Type: Abstract.
Deyo R, Tsui-Wu Y.
Descriptive epidemiology of low back pain and its related medical care in the United States.
Eisenberg DM, Davis RB, Ettner SL, Appel S, Wilkey S, Van Rompay M, Kessler RC.
Trends in Alternative Medicine Use in the United States, 1990 to 1997:
Results of a Follow-up National Survey
JAMA 1998 (Nov 11); 280 (18): 1569–1575
Frymoyer JW, Cats-Baril WL.
An overview of the incidences and costs of low back pain.
Orthop Clin North Am 1991;22:263–71.
Return to LOW BACK PAIN