Severity of chronic pain in an elderly
population in Sweden-impact on costs and
quality of life
Lars Bernfort, Björn Gerdle, Mikael Rahmqvist, Magnus Husberg and Lars-Åke Levin
Linköping University Post Print
N.B.: When citing this work, cite the original article.
Original Publication:
Lars Bernfort, Björn Gerdle, Mikael Rahmqvist, Magnus Husberg and Lars-Åke Levin,
Severity of chronic pain in an elderly population in Sweden-impact on costs and quality of life,
2015, Pain, (156), 3, 521-527.
http://dx.doi.org/10.1097/01.j.pain.0000460336.31600.01
Copyright: Elsevier / Lippincott, Williams & Wilkins
http://www.elsevier.com/
Postprint available at: Linköping University Electronic Press
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-116965
Research Paper
Severity of chronic pain in an elderly population in
Sweden—impact on costs and quality of life
Lars Bernfort
a,
*, Bj ¨orn Gerdle
b,c
, Mikael Rahmqvist
a
, Magnus Husberg
a
, Lars-
˚
Ake Levin
a
Abstract
Chronic pain is associated with large societal costs, but few studies have investigated the total costs of chronic pain with respect to
elderly subjects. The elderly usually require informal care, care performed by municipalities, and care for chronic diseases, all factors
that can result in extensive financial burdens on elderly patients, their families, and the social services provided by the state. This
study aims to quantify the societal cost of chronic pain in people of age 65 years and older and to assess the impact of chronic pain
on quality of life. This study collected data from 3 registers concerning health care, drugs, and municipal services and from 2 surveys.
A postal questionnaire was used to collect data from a stratified sample of the population 65 years and older in southeastern
Sweden. The questionnaire addressed pain intensity and quality of life variables (EQ-5D). A second postal questionnaire was used to
collect data from relatives of the elderly patients suffering from chronic pain. A total of 66.5% valid responses of the 10,000 subjects
was achieved; 76.9% were categorized as having no or mild chronic pain, 18.9% as having moderate chronic pain, and 4.2% as
having severe chronic pain. Consumed resources increased with the severity of chronic pain. Clear differences in EQ-5D were found
with respect to the severity of pain. This study found an association between resource use and severity of chronic pain in elderly
subjects: the more severe the chronic pain, the more extensive (and expensive) the use of resources.
Keywords: Age, Costs, Elderly, Intensity, Pain, Severity
1. Introduction
By 2050, 30% of the European popu lation w ill be aged 65 years
or older.
10
The European Commission has the goal to increase
the average number of healthy lif e years in the European Union,
a goal that requires addressing chronic pain in the elderly.
32
There is no consensus with respect to the relationships
between prevalence of chronic pain and age; some studies
note an increase in pain with age, some a decrease with age,
some a peak in middle age, and some found no age
dependence.
15,19,29,33
In addition, a systematic review and
a later study reported that only the prevalence of severe pain
conditions increased with age.
15,45
The literature suggests that the prevalence of chronic
musculoskeletal pain conditions in the elderly varies considerably.
In a short Swedish epidemiologic study not specifically focused
on the elderly population, the prevalence of chronic pain
(irrespective of pain intensity) was 56% among people aged 65
years and older,
24
but other studies report a higher prevalence of
pain in older adults (60%-86%).
34,35
A recent review concluded
that it was not possible to provide a general estimate because of
heterogeneity of the studies.
20
Chronic pa in is associa ted with large societal costs. The total
cost of moderate-to-severe chronic pain f or Swedish society in
2003 was calculated t o be 87.5 billion SEK, of w hich 7.5 billion
SEK con sisted of direct health care costs and 80 billion SEK
consisted of indirect costs related to production loss because
of illness-related sick leave.
44
In another Swedish study
concerning diagnoses possibly related to pain, the total costs
were 4 times higher.
26
In both studi es, the indirect costs
constituted the majority of costs. Some studies, including
studies that focus on health care use, indicate that very large
costs are associated with chronic pain and especially severe
chronic pain.
12,18,23,27,40
However, very li ttle is known about
health care costs in the elderly population. Obviou sly, the
elderly rarely require sick leave, so these costs are marginal.
However, costs related to informal care and care performed by
municipalities may be significant for this age group. Moreover,
increased age is associated with increased prevalence of
chronic diseases and co- and multi morbidities,
3,38
complica-
tions that require more complex health care and greater
personal and state expenses.
Although a relationship between pain and low quality of life
(HRQoL) has mainly been reported in middle-aged cohorts, little is
known about this relationship in the elderly.
32
This is the first report
of an epidemiologic study aimed at investigating the elderly
population with respect to the prevalence, intensity, variability,
and frequency of pain and how such pain characteristics correlate
with sociodemographic factors, comorbidities, activities, and
health. To this end, this study quantifies societal costs for chronic
Sponsorships or competing interests that may be relevant to content are disclosed
at the end of this article.
a
Division of Health Care Analysis, Department of Medical and Health Sciences,
Link ¨oping University, Link ¨oping, Sweden, Departments of
b
Pain and Rehabilitation
Centre and,
c
Department of Medical and Health Sciences, Link ¨oping University,
Link ¨oping, Sweden
*Corresponding author. Address: Department of Medical and Health Sciences,
Link ¨oping University, SE-581 85 Link ¨oping, Sweden. Tel.: 146 1034987; fax: 146
1034995. E-mail address: lars.bernfort@liu.semailto (L. Bernfort).
PAIN 156 (2015) 521–527
© 2015 International Association for the Study of Pain. This is an open-access article
distributed under the terms of the Creative Commons Attribution-NonCommercial-
NoDerivatives 3.0 License, where it is permissible to download and share the work
provided it is properly cited. The work cannot be changed in any way or used
commercially
http://dx.doi.org/10.1097/01.j.pain.0000460336.31600.01
March 2015
·
Volume 156
·
Number 3 www.painjournalonline.com 521
Copyright
Ó
2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
pain care for people 65 years and older. In addition, this study
investigates the relationship between chronic pain and HRQoL.
2. Methods
2.1. Design and subjects
This study collected data from 3 registers and 2 surveys. The first
survey was a cross-sectional postal questionnaire used to collect
data from a stratified sample of elderly—65 to 69 years, 70 to 74
years, 75 to 79 years, 80 to 84 years, and 85 years and
older—residing in the 2 largest cities of a county (
¨
Osterg ¨otland) in
southeastern Sweden. These subjects were selected from
a sampling frame based on the Total Population Register. A
questionnaire was mailed in October 2012 and was followed by
up to 2 postal reminders (at 2-week intervals). The collection of
questionnaires closed in January 2013. Data were collected by
Statistics Sweden (SCB). The sample population consisted of
49,320 people. From this sample, 2000 subjects were randomly
selected from each age stratum, so the final study included
10,000 subjects. The first survey asked the subject whether they
needed help from relatives because of their chronic pain. If they
answered “yes,” then a second postal survey was sent to the
relative(s) mainly responsible for their care. Two postal reminders
(at 2-week intervals) were also mailed for the second survey.
The study was approved by the Regional Ethics Research
Committee in
¨
Osterg ¨otland, Sweden (Dnr: 2012/154-31).
2.2. Main questionnaire
The questionnaire sent to the elderly subjects included several
variables and instruments. All variables and instruments will be
described in detail in forthcoming studies. This study used the
pain intensity and quality of life variables described below.
2.3. Pain intensity
The respondents assessed their pain intensity by marking an 11-
grade (0-10) numeric rating scale; zero indicated no pain at all and
10 indicated worst possible pain.
21
For analytical purposes, the
population was subdivided into 3 groups with respect to having
chronic pain or not and with respect to level of pain during the last
week: no or mild pain (0-4), moderate pain (5-7), or severe pain
(8-10).
6,28
An alternative approach to using numbers with 2 verbal
anchor points is to label different sections of the scale with verbal
descriptions such as “mild,” “moderate,” and “severe” chronic
pain.
14,36
The reason for this approach would be that such labels
may more adequately represent the respondents’ perceptions.
However, in this study, we choose to follow the recommendations
of IMMPACT and used numeric rating scale with 2 anchor points.
17
2.4. Quality of life
In the main questionnaire, the instrument EQ-5D for measuring
health-related quality of life was included. The EQ-5D instrument
is a descriptive system of health-related quality of life states
consisting of 5 dimensions—mobility, self-care, usual activities,
pain/discomfort, and anxiety/depression. For each dimension,
the subject records the level of severity—no problems/some or
moderate problems/extreme problems.
1,8
Each combination
(denoted EQ-5D profile) can be transformed into an index value
corresponding to a quality-adjusted life-year weight using the UK
population–based value set.
16
The EQ-5D also contains a visual
analog scale (VAS) health scale (denoted EQ-5D-VAS) that
ranges between 0 and 100, where zero is the worst imaginable
health state and 100 is the best imaginable health state. The
EQ-5D-VAS is an alternative direct method for eliciting an index
value. This study presents results for both the EQ-5D-index
(including the 5 items) and EQ-5D-VAS.
2.5. The relatives’ questionnaire
In the main questionnaire, the respondents were asked to
volunteer contact information for relatives who provided them
informal care and support. If the respondents volunteered contact
information, a second questionnaire was mailed to the relatives
who they identified as their main caregivers. This second
questionnaire asked these relatives about what kind of support
they provided and how many hours they spent caring for their
elderly relative. This support included domestic chores, shopping,
getting around outside the home, taking medication, and handling
finances. These caregivers were also asked to estimate the amount
of work time and leisure time they devoted to their charges.
2.6. Informal care costs
Informal care costs were calculated according to the human
capital approach.
49
The cost of a workday spent caring for the
relative was calculated as average gross wages (Sweden, 2012)
plus payroll taxes (230 EUR corresponding to 29 EUR per hour).
1
Leisure time was valued as half the average net wages (7.78 EUR
per hour).
2.7. The health care registers used for retrieving health
care costs
Registers were used to retrieve information on consumption of
health care and health care costs. Data on inpatient (hospital days
and treatments) and outpatient (visits to various categories of
personnel) care were retrieved from the Health Care Register in
¨
Osterg ¨otland (HCR
¨
O). The HCR
¨
O is an administrative database that
records date, name, unique personal identification number, and
diagnosis (main diagnosis and up to 9 secondary diagnoses)
according to the International Classification of Diseases, 10th
version (ICD-10), of every health care visit or hospital stay within
the County of
¨
Osterg ¨otland. This database has previously been
described in greater detail.
50
The health care consumption related to
the persons participating in this study was retrieved from the HCR
¨
O
by matching the personal ID numbers to the database. To determine
costs, we linked health care consumption data from the HRC
¨
Owith
data in a regional cost database provided by
¨
Osterg ¨otland County
Council. The regional cost per patient database contains costs for all
health care services within
¨
Osterg ¨otland County and, combined with
the CDW
¨
O, offers individual-based cost data that may be
summarized for certain populations or clinics.
The National Board of Health and Welfare (Socialstyrelsen)
maintains a register on drug prescription. The register contains
information on prescribed medications that have been retrieved by
patients. The register can be merged with other registers to
connect prescribed and retrieved drugs to specific diagnoses or
individuals. For each prescription, the drug register contains
information on the patient, prescribed drug, relevant diagnosis,
cost, profession, and workplace. Information on all prescribed and
then retrieved drugs related to the participants in our study were
gathered by matching the personal ID numbers to the register.
Information on costs incurred by the municipalities for home
help services and special housing was retrieved from a register
provided by the Link ¨oping municipality. The register contains all
522 L. Bernfort et al.
·
156 (2015) 521–527 PAIN
®
Copyright
Ó
2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
consumption of the different home help services with hours and
prices attached to each service. In addition, the register contains
the number of days in special housing. Total costs can be
calculated by the use of the standard cost per day. The municipal
registers also contain personal ID numbers, which were used to
match the data to persons participating in our study. All costs are
expressed as costs per year (2012 prices). Cost data were
gathered over 1 year, from November 2011 to October 2012.
2.8. Statistics
All data files and analyses of data were handled with SPSS version
IBM Statistics 20 (IBM Corp., Armonk, NY). The outcome in costs
for the 3 chronic pain groups was tested using analysis of variance,
and the result was adjusted for sex and age. The age component,
in particular, should be adjusted when testing the different costs
across the groups. Analysis of variance was used for the statistical
analysis of costs for health care, drug acquisition, and municipal
home services. The P value was #0.05.
3. Results
3.1. Participation rate
Of the 10,000 subjects selected from the sample population,
3261 did not participate or did not return the questionnaire; 58 of
these subjects were not reachable. The net response rate was
67.3%; however, some respondents failed to answer a number of
vital questions, leaving a total of 6611 (66.5%) valid responses.
3.2. Characteristics of the responders
The valid study population (6611 subjects) consisted of 54% women
and 46% men. Average age was 76.2 years (range, 65-102 years).
The intensity of chronic pain broke down as follows: 76.9% were
categorized as having no or mild chronic pain, 18.9% as having
moderate chronic pain, and 4.2% as having severe chronic pain.
3.3. Comparisons between responders and nonresponders
Results were calibrated with respect to imbalances in respon-
siveness associated with certain variables. Calibrations were
made for results to be representative of the population studied.
Table 1 shows variation in responsiveness for the most important
background variables.
3.4. Costs
Results on resource uses are presented in Table 2. The total
average annual cost to society in terms of health care, drugs,
municipal services, and informal care for the average person 65
years or older was 5090 EUR. It is clear that the more severe the
chronic pain, the more resources were consumed (Table 2). The
additional costs of chronic pain that was more severe than mild,
that is, for moderate and severe chronic pain, were on average
570 EUR SEK per person 65 years and older. The total annual
costs were more than doubled in the subgroup with severe
chronic pain compared with the group with no or mild chronic
pain (ie, 10,200 EUR vs 4510 EUR). Moderate chronic pain also
showed a higher cost than no or mild chronic pain (ie, 6250 EUR
vs 4510 EUR). Health care costs generally increase with both age
and severity of chronic pain, as presented in Figure 1 .
With the exception of municipal costs, the statistical analyses
showed highly significant differences between the groups for the
investigated cost items (Table 2). However, when municipal
costs were divided according to home help services and special
housing, the tests indicate differences concerning home help
services but not special housing.
In the first questionnaire, 249 respondents stated that they
received help from relatives. In the relatives’ questionnaire, 128
relatives completed the questionnaire: 38 (30%) were relatives of
persons with no or mild chronic pain, 55 (43%) were relatives of
persons with moderate chronic pain, and 35 (27%) were relatives
of persons with severe chronic pain. The costs of the subgroup of
help receivers were distributed over the entire study population of
6611 persons. The results clearly show that increased severity of
chronic pain is associated with increased informal care costs.
When interpreting results from surveys, caution should be
taken regarding the representativeness of the respondents. In this
study, particular caution should be taken in interpreting results
considering informal care costs as the number of elderly persons
reporting to receive help from relatives was relatively small. In
addition, the response rate among relatives identified as helpers
was relatively low. These facts might indicate that informal care
costs were underestimated.
3.5. Quality of life
The health-related quality of life of persons with different grades of
chronic pain measured by EQ-5D is presented in Table 3.
Differences in health-related quality of life were very clear between
the 3 pain categories, both for the EQ-5D-index and the EQ-VAS,
and for each of the 5 dimensions of the EQ-5D-health state
profile. The least evident difference concerns the dimension of
self-care (Table 3). The differences detected clearly show the
association between presence and degree of chronic pain and
the health-related quality of life experienced.
EQ-5D-index values for persons with different levels of chronic
pain in different age groups are presented in Figure 2. Generally,
Table 1
Proportion responding with respect to different
characteristics.
Characteristics Response rate, %
Gender
Men 73.1
Women 66.0
Age
65-69 74.9
70-74 75.4
75-79 70.7
80-84 64.0
85 and above 51.1
Marital status
Married 75.5
Other 61.5
Education
Compulsory school 59.2
Secondary school 72.2
University 82.5
Income (SEK per y)
0-150,000 57.0
150,001-220,000 69.1
220,001 and above 80.8
Birth country
Sweden 71.1
Rest of Europe 54.8
Rest of the world 33.2
March 2015
·
Volume 156
·
Number 3 www.painjournalonline.com 523
Copyright
Ó
2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
the expected relationship of lower HRQoL with increasing age
can be seen. The most remarkable result is the very low HRQoL
(0.2) experienced by those with severe chronic pain in the oldest
age group.
4. Discussion
This study found that both costs and the aspects of HRQoL
investigated were clearly associated with the existence and
severity of chronic pain. This relationship was evident for all kinds
of resource use. Hence, health care costs and HRQoL aspects
differed significantly between the 3 groups—no or mild,
moderate, and severe chronic pain.
As briefly reviewed by Lalonde et al.,
31
economic burden
increases with pain severity and/or pain disability. This observation
was confirmed in this study’s population, elderly subjects. Our
results indicate that the cost to society in terms of health care,
drugs, municipal services, and informal care for the average person
65 years or older with or without chronic pain is 5090 EUR per year.
The municipalities studied together have approximately 50,000
inhabitants 65 years and older, which would translate into a total
cost of 255 million EUR, assuming that the 2 municipalities are
representative for Sweden (approximately 1.8 million inhabitants 65
years and older). This would imply that costs attributable to people
65 years and older add up to 9.2 billion EUR for Sweden. These
costs can be seen as the cost of an aging population.
The additional cost associated with moderate and severe
chronic pain was on average 570 EUR per person 65 years and
older. This means that for Link ¨oping and Norrk ¨oping together the
additional annual costs of the presence of chronic pain that is
worse than mild is 28.5 million EUR. This cost difference means
that for Sweden (10 million citizens) the additional annual costs for
chronic pain worse than mild in people 65 years and older
amounts to more than 1 billion EUR.
In a Swedish study, Gustavsson et al.
26
estimated the direct
and indirect costs of patients with a diagnosis related to chronic
pain. The mean total cost per patient each year was estimated at
6400 EUR (SEK 56,000); this study found the mean total cost per
year of persons 65 years or older to be 5090 EUR. The estimation
of diagnosis related to chronic pain is not intended to estimate the
costs of pain, as many diagnoses are not associated with pain,
a confounder that could lead to an overestimation. Gustavsson
et al.
26
discussed possible reasons for this discrepancy and
pointed out the difficulty of strictly dividing the costs into pain-
related and nonpain–related costs. From an international
perspective, many cost estimations are hampered by difficulties
identifying patients with chronic pain.
7
Thus, a strength of our
epidemiologic population–based study is that the pain classifi-
cation is based on the respondents’ own evaluations.
Outpatient care costs are most reasonably underestimated
because of the fact that data from private caregivers were not
included in the data capture. However, private care is generally only
a small part of the total health care in the 2 municipalities
investigated. A fact that distinguishes this study from other similar
ones is that we have been able to include costs incurred by the
municipalities, making this study more complete compared with its
predecessors. However, municipal costs could only be retrieved
from one (Link ¨oping) of the 2 included municipalities. Concerning
the municipal costs, the group with the most severe chronic pain
stands out as a large consumer of resources. Differences in
municipal costs between the group with no or mild chronic pain and
the group with moderate chronic pain were quite small, possibly
because of the elderly population consuming resources irrespective
of the existence of chronic pain. Chronic pain may become
a problem only for the group with the most severe chronic pain.
A possible weakness in the study lies in the data quality on
informal care. In the relatives’ questionnaire, some internal loss of
data was experienced. A related issue is how to value the
relatives’ loss of leisure time because of informal care efforts.
Equating loss of production with loss of time from paid labor
would imply that the value of leisure time is zero. We do not agree
Table 2
Mean annual health care and municipal costs (Euros, rounded to the nearest 10) per person across subgroups with different
conditions of chronic pain.
Average cost (CI 95%) Without pain or with mild
chronic pain (n 5 5077)
Moderate chronic pain
(n 5 1245)
Severe chronic pain
(n 5 289)
P
Inpatient care 1220 (1030-1410) 1710 (1330-2090) 2110 (1320-2900) 0.012
Outpatient and primary care
Doctor’s visits 1230 (1160-1300) 1760 (1610-1900) 2000 (1700-2300) ,0.001
Total outpatient care 1710 (1610-1800) 2500 (2300-2700) 3100 (2690-3500) ,0.001
Drugs 440 (410-480) 640 (580-710) 900 (760-1040) ,0.001
Total health care 3370 (3120-3610) 4860 (4360-5360) 6110 (5080-7130) ,0.001
Home help services 280 (180-370) 340 (170-520) 1180 (800-1560) ,0.001
Special housing 680 (530-810) 530 (240-820) 1200 (610-1790) 0.136
Total municipal costs 960 80 2380 NA
Informal care 180 520 1710 NA
Total costs 4510 6250 10,200 NA
NA 5 not applicable.
Figure 1. Health care costs divided into different age groups categorized with
respect to the level of chro nic pain, and it shows the variation in health care
costs because of age and divided into groups of severity of chronic pain.
524 L. Bernfort et al.
·
156 (2015) 521–527 PAIN
®
Copyright
Ó
2015 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
