Pain 73 (1997) 393–400

Pain epidemiology and health related quality of life in chronic non-malignant

pain patients referred to a Danish multidisciplinary pain center

Niels Becker a ,*, Annemarie Bondegaard Thomsen a, Alf Kornelius Olsen a,

Per Sjøgren a, Per Bech b, Jørgen Eriksen a
a
Multidisciplinary Pain Center, Danish National Hospital, DK-2100 Copenhagen Ø, Denmark
b
Frederiksborg General Hospital, DK-3400 Hillerød, Denmark.

Received 2 April 1997; revised version received 24 July 1997; accepted 6 August 1997

Abstract

This paper presents the results of a detailed study of the pain epidemiology and health related quality of life (HRQL) in 150 chronic non-
malignant pain patients consecutively referred to a Danish multidisciplinary pain center. Mean pain severity was 71.6 (SD = 18.5) on the
VAS scale. Forty-two percent reported poor quality of sleep. HRQL was evaluated with the Medical Outcome Study-Short Form (SF-36),
the Hospital Anxiety and Depression scale (HAD) and the Psychological General Well-Being Scale (PGWB). Compared with the normal
population (NP) both SF-36 scores and PGWB scores were significantly reduced (P , 0.001) indicating that physical, psychological and
social well-being were severely reduced. On the HAD scale 58% were found to have a depressive or anxiety disorder. Statistically
significant but modest correlations were found between pain severity and HRQL. Psychological and social well-being was closely
correlated. Sixty-three percent of the referred patients had neurogenic pain conditions. Of these, only 25% were treated with antidepressants
or anticonvulsants at referral. Seventy-three percent were treated with opioids at referral. Mean opioid consumption was 64 mg of morphine
per day (range 1–280 mg). Compared with the NP the chronic pain patients had used the health care system five times more often in the
years prior to referral (P , 0.001). The study confirms the severe multidimensional impact of chronic pain and demonstrates that HRQL of
chronic non-malignant pain patients is among the lowest observed for any medical condition.  1997 International Association for the
Study of Pain. Published by Elsevier Science B.V.

Keywords: Pain center; Pain epidemiology; Chronic pain; Health-related-quality-of-life; Neuropathic pain

1. Introduction measures described in the literature show that selection of

relevant parameters is difficult (Aaronson, 1990; Turk et al.,
Chronic pain is one of the most widespread and difficult 1993). The socio-economic impact of chronic non-malig-
problems the medical community has to face (Latham and nant pain is often assessed by measuring employment status,
Davis, 1994). In addition to the suffering of the patients, use of health care services or use of analgesics. These para-
enormous medical and social resources are spent on this meters may be relatively easy to quantify, but they are
patient category (Latham and Davis, 1994). This has led strongly influenced by cultural and socio-economic struc-
to increasing demands for effectiveness of therapy, and dur- tures in different communities (Turk et al., 1993).
ing the past decades multidisciplinary treatment principles Symptoms often accompanying chronic pain are depres-
for chronic non-malignant pain have received increasing sion, anxiety, physical dysfunction and social isolation
attention. For evaluation of treatment outcome relevant (Rudy et al., 1988). Methods used for assessment differ
assessment methods are necessary. The large variety of between studies, but the important aspects concerning the
patient’s physical, psychological and social well-being are
* Corresponding author. H:S Multidisciplinary Pain Center, Copenhagen
often defined as the patient’s health related quality of life
University Hospital, Tagensvej 18B, 7122, DK-2200 Copenhagen N, Den- (HRQL) (Schipper, 1990). It seems relevant to supplement
mark. Tel.: +45 35457136; fax: +45 35455349. pain epidemiological and socio-economic assessment meth-

0304-3959/97/$17.00  1997 International Association for the Study of Pain. Published by Elsevier Science B.V.
PII S0304-3959 (97 )0 0126-7
394 N. Becker et al. / Pain 73 (1997) 393–400

ods with HRQL measures for characterization of chronic Regarding the general practitioner questionnaire, pain
non-malignant pain patients. Sufficiently validated ques- was considered the main reason for referral. The GP was
tionnaires designed to measure HRQL of patients with a asked to rate the importance of secondary referral reasons
diversity of diseases allow comparison with a background on a scale where 0 represented no importance and 5 was
population and other patient groups. Apart from demon- equal to very important. Any additional reasons for referral
strating the multidimensional nature in which chronic pain were also recorded, and the referring GPs were asked on a 5-
affects the patients, studies have shown a close relation point scale to rank how they rated the impact of physical,
between pain and depression (Romano and Turner, 1985). psychological and social factors on the pain condition.
However, the relationship between pain and different The patient questionnaire dealt with three main topics:
HRQL dimensions is still a subject of research and debate pain, sleep, and HRQL. Measurements of pain intensity
(Carsten et al., 1995; Turk et al., 1995). were performed using a visual analog scale (VAS) and a
The aim of this study was to characterize chronic non- 5-point Likert scale (0 = no pain, 1 = mild, 2 = moderate,
malignant pain patients referred to a Danish multidisciplin- 4 = severe and 5 = unbearable pain). Quality of sleep dur-
ary pain center using not only pain epidemiology and socio- ing the week prior to the first consultation in the pain center
demographic data but also HRQL. Furthermore, this study was evaluated as good, fair or poor. If the sleep was inter-
investigated the relationship between chronic pain and rupted by pain, the patients were asked to rate the interrup-
HRQL. tion as mild, moderate or severe.
HRQL instruments were selected to cover three dimen-
sions: physical function (disability), psychological well-
2. Methods being, and social well-being (Bech, 1993). They included
the Hospital Anxiety and Depression Scale (HAD) (Zig-
During the study period May 1994 to August 1995 all mond and Snaith, 1983), the Psychological General Well-
patients suffering from chronic non-malignant pain condi- Being Scale (PGWB) (Dupuy, 1984), and the MOS 36-Item
tions referred to the Pain Center and living in the commu- Short Form Health Survey (SF-36) (Stewart et al., 1988).
nity of Copenhagen were consecutively included in the HAD identifies milder cases of depression and anxiety in
study if they fulfilled the inclusion criteria: all relevant medically ill patients. It was developed and validated on
medical or surgical investigations and treatments should non-psychiatric medical patients. Items relating to both
be completed prior to referral, age above 18 years, no mood disorder and physical illness have been eliminated.
major mental disorders present and no illegal use of opioids. HAD consists of a depression and a anxiety subscale. Scores
Ethically the investigation was in accordance with the on each subscale range from 0 to 21. Scores above 8 indicate
Helsinki Declaration and approved consent was given by that a depressive or anxiety disorder is likely to be present
the Ethical Committee of Copenhagen. (Zigmond and Snaith, 1983).
Immediately before the first visit to the pain center the PGWB is a 22-item inventory originally designed to mea-
patient and the referring general practitioner (GP) received a sure subjective psychological well-being in population
questionnaire asking for information concerning the based studies. It is extensively validated and has been pro-
patient’s pain, physical and psychological conditions. At ven to posses good psychometric properties in several clin-
the pain center the pain specialists made a medical record, ical studies within indications such as hypertension (Omvik
a pain analysis including classification according to the et al., 1993) and gastrointestinal symptoms (Dimenäs et al.,
IASP coding system, and filled in a questionnaire concern- 1993). PGWB is comprised of six subscales providing eva-
ing sociodemographic data. The pain center evaluations luations of anxiety, depression, vitality, positive well-being,
were performed by three pain specialists. self-control and general health. Each subscale has three to
Sociodemographic data concerning the normal popula- five items. The subscales range from 0 to 15 or 20 or 25. The
tion (NP) of the city of Copenhagen (n = 462 000) were overall PGWB index score ranges from 0 to 110. The norm
supplied by the Copenhagen Bureau of Statistics. To com- for the American normal population is 82 ± 15, the higher
pare civil status and employment rate of the NP and the the better (Dupuy, 1984). The norm value for the Danish
patient population (PP) a standardization procedure was version has been confirmed on well-controlled diabetes
used utilizing the ‘indirect principle’ (Foldspang et al., patients (Naylor, 1996). Validation has been performed by
1981). comparing it with the Beck Depression Inventory and the
The mean number of in-hospital days in the PP during a Hopkins SCL-90. Correlation coefficients about 0.7
5-year period prior to referral was compared with the mean (Dupuy, 1984) have been demonstrated. SF-36 is a general
number of in-hospital days for the NP of Copenhagen. Data health questionnaire evaluating the physical, social and
were achieved from the Danish National Board of Health. mental aspects of HRQL. SF-36 includes eight subscales:
The questionnaires were administered as a part of a major Physical Functioning, Role Functioning-Physical, Bodily
ongoing evaluation and treatment outcome study and con- Pain, Social Functioning, Mental Health, Role Function-
tained several other questions. Details of the assessments ing-Emotional, Vitality and General Health Perceptions.
considered relevant for this investigation are given below. The range for each subscale is 0–100. SF-36 has been vali-
 N. Becker et al. / Pain 73 (1997) 393–400 395

dated extensively on general populations and different 3. Results

diseases demonstrating high reliability and good con-
struct validity (McHorney et al., 1994). Calculation of Two hundred and ten patients (mean age 58 years, range
scores were performed according to Ware et al. (1993). 23–89) with chronic non-malignant pain conditions were
The Danish SF-36 norm values were used to estimate the included in the study. Sixty patients (29%) who were unable
SF-36 subscores for the NP of Copenhagen. Calculations or unwilling to fill in the questionnaires were excluded,
were made according to the ‘indirect principle’ (Foldspang leaving 150 patients to take part in the study. Patients
et al., 1981). excluded from the study did not differ from the included
The pain specialist questionnaire included sociodemo- patients with respect to age, gender, reasons for referral or
graphic data and information concerning the use of medi- pain duration and pathophysiology. Sociodemographic data
cine 1 week prior to the first consultation. Drugs, daily on the PP and the NP are presented in Table 1.
doses, administration routes and temporal aspects of drug
intake were registered. All opioid doses are given as milli- 3.1. Pain epidemiology
grams of oral morphine. For opioids other than morphine the
equipotency table published by Clausen et al. (1995) was Data on reasons for referral given by the GPs were
used for conversion. obtained in 113 patients (75.3%) (Fig. 1). The importance
of the physical, psychological and social factors on the pain
2.1. Statistical methods condition rated by the GPs is shown in Table 2. Physical
factors were rated as the most important reasons in 64% of
Poisson confidence limits were used to test for sociode- the patients. Psychological factors were rated among the
mographic differences between PP and NP. PGWB and SF- most important reasons in 58% of the patients.
36 scores were compared with the norm values using t-tests.
Associations among the pain and HRQL variables were 3.2. Pain pathophysiology
assessed using the Spearman correlation and linear regres-
sion functions of the SAS program for statistical analysis. Data on pain pathophysiology classified by the pain spe-
Spearman correlations were used for assessing intrapersonal cialists were obtained in all 150 patients. Many patients had
variation in response to identical items. both primary and secondary pain conditions. The distribu-

Table 1
Age, gender and sociodemographic data for the patient population (PP) and the normal population (>18 years) of Copenhagen (NP)

PP PP-expa NP

Number Percent Percent Percent

Age distribution (years)

18–24 2 1 13
25–34 10 7 23
35–44 21 14 13
45–54 35 23 11
55–64 23 15 8
65–74 28 19 8
75– 31 21 10

Total 150 100

Gender
Female 98 65* 53*
Civil status 149
Married 70 47** 39 27
Unmarried 34 23** 24 53
Divorced/widowed 45 30** 37 20
Occupational status (,67 years) 91
Working 35 39* 64
Unemployed 7 7** 7 8
Disability pension 41 45* 10 10
Other/unknown 8 9 (ns) 18
Mean age (years) 58* 45*
a
PP-exp: Expected frequencies as they had been, if the age and gender stratified frequencies were as in the NP.
*Difference between groups PP and NP (P , 0.05).
**No difference between PP and PP-exp.
396 N. Becker et al. / Pain 73 (1997) 393–400

mity pain as the primary pain location, 27 patients (20%)

had low back pain, 19 patients (14%) head or facial pain, 16
patients (12%) abdominal pain, nine patients (7%) thoracic
pain and five patients (4%) rectal pain. Fourteen patients
had more than three pain locations.

3.4. Quality of sleep

Data on quality of sleep were obtained from 149 patients

(99.3%). Sixty-two patients (42%) reported poor, 69 patient
(46%) fair, and 18 (12%) good quality of sleep. Sleep was
interrupted by pain in 125 patients (83.9%). During the
week prior to the first consultation 11% rated the distur-
bance as mild, 47% as moderate and 42% as severe.

3.5. Use of analgesics

Data on use of analgesics and tranquilizers was obtained

from 150 patients. At referral 109 patients (73%) were trea-
ted with opioids. Mean opioid consumption for the patients
Fig. 1. Importance of secondary referral reasons rated on a 0–5 scale:
treated with opioids was equivalent to 64 mg of morphine
0 = no importance, 5 = very important (n = 150).
per day (range 1–280 mg). Twenty-four of the 109 patients
tion of primary pain pathophysiologies is shown in Fig. 2. (22%) were in a stable treatment using only long-acting
Sixty-nine patients (46%) suffered primarily from neuro- opioids. Short-acting opioids were used by 71 patients
pathic pain, and additionally 27 patients (17%) had neuro- (65%) mostly on demand basis (68% of the patients).
pathic pain secondary to other pain conditions. Twenty- Twenty-five percent of the patients with primary or second-
seven patients (18%) suffered from somatic pain secondary ary neuropathic pain conditions were treated with anticon-
to other pain conditions. Twenty-three patients (15%) had vulsants and/or antidepressants versus 17% of the patients
psychogenic pain as a secondary pain condition. with other pain conditions (no statistically significant dif-
ference).
3.3. Pain intensity, duration and location Acetylic acid, paracetamol or NSAIDs were used by 90
patients (60%). Fifty-four patients (36%) received tranqui-
Data on pain levels measured on the VAS scale were lizers (benzodiazepines). Only 11 patients (7%) did not use
obtained from 143 patients. Mean pain level was 71.6 analgesics or tranquilizers.
(SD = 18.5). On the 5-point Likert scale 109 of 150 patients
(73%) reported severe or unbearable pain. Forty-one 3.6. Use of health care services
patients (27%) reported moderate pain. None of the patients
reported none or mild pain. The VAS and verbal pain rating During a 5-year period prior to referral the mean number
scores were highly correlated (P = 0.73). Mean duration of of in-hospital days was 11.38 days per year (range 0–181)
pain was 8.3 years.
Data concerning primary locations of pain was obtained
in 135 patients (90%). Forty-five patients (33%) had extre-
Table 2
Significance of physical, psychological and social factors

Significance of factors rated

by the general practitioner

Not Very
important important

0 1 2 3 4 5

Physical factors 3 10 14 24 38 24
Psychological 7 11 18 28 31 18
factors
Social factors 21 21 24 26 15 6

n = 113. Fig. 2. Primary pain pathophysiology (n = 150).

 N. Becker et al. / Pain 73 (1997) 393–400 397

questions in the SF-36 Mental Health subscale and PGWB

questionnaire were relatively low. Spearman correlations
ranged from 0.52 to 0.66 indicating acceptable reliability
of the questionnaires.

3.8. Relationships between health related quality of life and

pain.

Relationships between pain (VAS and Likert scale) and

the scales measuring HRQL were relatively weak. Fig. 5
displays the patterns of correlations among pain (VAS)
and subscales representing the various HRQL dimensions.
Psychological well-being (PGWB-total) and social function
(SF-36-Social Function) were relatively highly correlated
(r = 0.64). Pain correlated closest with SF-36-Physical
Function (r = 0.39). In a linear regression analysis, varia-
tion in pain score predicted only 2.87% of variation in SF-
36-Social Function and 4.8% of variation in PGWB-total,
while 12.7% of the variation in SF-36-Physical Function
could be explained. Physical function accounted for 8.6%
of the observed variation in PGWB-total.

Fig. 3. PGWB subscores, mean (SD) (n = 150). X, American norm values;

ANX, anxiety; DEP, depression; GEN, general health; POS, positive well-
4. Discussion
being, SEL, self-control; VIT, vitality. *P , 0.001.

for the patient population. The mean number of in-hospital In this study we have investigated a population of
days per year for the general population of Copenhagen in a chronic, non-malignant pain patients referred to a multidis-
corresponding age group is 2.4 days. The difference is sta- ciplinary pain center situated in the community of Copenha-
tistically significant (P , 0.001). gen. Our patients do not represent the general population of

3.7. Health related quality of life

PGWB and HAD questionnaires were obtained from all

150 patients. On the HAD depression subscale, 40% had
scores above 8 (range 0–20) indicating depressive disorder.
On the HAD anxiety subscale 50% had scores above 8
(range 0–21) indicating anxiety disorder. The presence of
anxiety and depression in patients were highly correlated
(0.65, P , 0.0001). Forty-two percent of the patients had
scores indicating neither depressive nor anxiety disorders.
The subscores for psychological well-being measured by
PGWB are shown in Fig. 3. Mean PGWB total score was
47.9 (SD = 20). This is much lower than the norm value of
the American population which is 82 (SD = 15) (P ,
0.001).
Data obtained from the SF-36 questionnaires are shown
in Fig. 4. The eight SF-36 subscores for the pain patients
were all much lower than the norm values for a Danish
population of similar age and sex distribution (P , 0.001).
The SF-36 subscores concerning psychological well-
being were highly correlated with the corresponding
PGWB and HAD subscales. Correlation coefficients for
the depression, anxiety, mental health and vitality subscales
Fig. 4. SF-36 subscores, mean (SD) (n = 150). X, Danish norm values; BP,
of the three questionnaires ranged from 0.69 to 0.79 body pain; GH, general health; MH, mental health; PF, physical function-
(P , 0.0001) indicating good concurrent validity. ing; RE, role emotional; RP, role physical; SFA, social functioning; VIT,
Intrapersonal variations in responses to almost identical vitality. *P , 0.001.
398 N. Becker et al. / Pain 73 (1997) 393–400

are among the most frequent users of the Danish health care
system.
One of the main objectives of this investigation was to
characterize the HRQL of chronic, non-malignant pain
patients referred to a Danish multidisciplinary pain center.
Our patients suffered from a wide range of pain conditions.
Therefore, HRQL questionnaires designed for use on the
general population or patient populations with a diversity
Fig. 5. Correlations between pain and subscales representing the various of diseases were chosen. SF-36 especially is widely used
HRQL dimensions. PGWB, psychological well-being; SF-36 PF, physical
and validated (McHorney et al., 1994). Although SF-36 is a
function; SF-36 SF, social function; VAS, pain.
measure of general health status rather than HRQL, the
chronic non-malignant pain patients in Denmark. The struc- eight subscales cover physical, psychological and social
ture of the Danish health care system, having only few well-being. Physical well-being is assessed by the subscales
multidisciplinary pain treatment units with a limited treat- rating physical function (disability) and role limitations due
ment capacity, has undoubtedly an important influence on to physical and emotional problems. In the assessment of
the referral and visitation policy. psychological well-being, the SF-36 Mental Health scale
Surprisingly more than 60% of our patients suffered from was supplemented with the HAD and the PGWB question-
neuropathic pain conditions. This is far above the preva- naires in order to measure the psychological well-being of
lences reported by other pain clinics where neuropathic the patients in more detail.
pain conditions were found in 30–40% of the patients The present study confirmed the multidimensional reduc-
(Bowsher, 1991; Ekter-Andersen et al., 1993). Diagnosis tion in HRQL reported by other studies on chronic non-
of neuropathic pain conditions in our study were based on malignant pain patients (Rudy et al., 1988; Flor et al.,
the criteria given by Hansson (1994). Traditionally, the 1992). However, the severity of impairment in physical,
highly extended rheumatological services in Denmark social and psychological well-being was remarkable.
take care of most of the chronic musculoskeletal pain con- HRQL measured by SF-36 was equally low or lower than
ditions. No doubt this is the main explanation for the high scores obtained in patients having severe cardiopulmonary
percentage of patients with neuropathic pain conditions diseases or major depression (Stewart et al., 1989; Wells et
referred to our pain center. Another factor may be limited al., 1989; Ware et al., 1994). SF-36 scores in our study were
knowledge of neuropathic pain conditions. According to a similar to those found by Garratt et al. (1993) in chronic low
recent questionnaire survey, most Danish physicians lacked back pain patients referred to a variety of outpatient clinics
knowledge of even basic pharmacological treatment princi- in Scotland. Both Garratt et al. (1993) and Lyons et al.
ples of neuropathic pain conditions in cancer patients (1994) have reported less severely reduced SF-36 scores
(Sjøgren et al., 1996). In accordance with these findings for chronic back pain patients who had not been referred
this study showed that the use of antidepressants and/or to pain centers. No doubt this illustrates the importance of
anticonvulsants in patients with neuropathic pain conditions considering chronicity and selection criteria when results
was sparse and did not differ statistically significantly from are interpreted. Compared to the results reported by Nilges
that of patients with other pain conditions, indicating these et al. (1996) the scores of our patients seem to be matched
drugs primarily were used for other purposes than pain only by the scores of the most chronified subpopulation of
treatment. chronic, non-malignant pain patients referred to a number of
Denmark has the highest legal opioid consumption per German multidisciplinary pain centers.
inhabitant in the world (Clausen et al., 1995). This is pri- Concerning psychological well-being the PGWB ques-
marily due to a very liberal legislation on prescription of tionnaire confirmed the severe reduction demonstrated by
opioids. Still, it is remarkable that about 75% of the patients the SF-36 Mental Health and SF-36 Vitality scores. Further-
were using opioids at referral. Most of these patients used more, the PGWB subscores in our study showed that the
short-acting opioids on an on-demand basis. The objective patients’ positive well-being and self-control were equally
for pain clinicians in Denmark is not when to initiate an low. To our knowledge PGWB has not been used for rating
opioid treatment in chronic non-malignant pain patients, chronic pain patients. Studies on patients with hypertension
but to convert and stabilize uncontrolled, on-demand use (Omvik et al., 1993) and gastrointestinal symptoms (Dime-
of short-acting opioids into a stable, controlled and regular näs et al., 1993) reported much better PGWB scores than
administration of long-acting opioids, or when possible, to obtained in our study.
reduce or eliminate irrelevant opioid use. The very high incidence of anxiety and depression as
Our patients had spent an average of 11.4 days/year in recorded by the HAD is consistent with other studies.
hospital. In a study on a Danish population of 30 426 per- Romano and Turner (1985) report that approximately 50%
sons aged 17–49 years the number of in-hospital days of chronic pain patients display significant levels of depres-
among the 1.0% most frequently hospitalized was 9 days/ sion.
year (Fink, 1989). This indicates that chronic pain patients Only a moderate direct association between pain intensity
 N. Becker et al. / Pain 73 (1997) 393–400 399

and psychological well-being was observed. This is in Wiklund, I., Quality of life in patients with upper gastrointestinal
accordance with the cognitive-behavioral mediation model symptoms, Scand. J. Gastroenterol., 28 (1993) 681–687.
Dupuy, H.J., The Psychological General Well-Being (PGWB) Index. In:
(Kerns and Turk, 1984) in which depression rather than N.K. Wenger, M.E. Mattson, C.F. Furberg and J.A. Elinson (Eds.),
being directly associated with pain is mediated through per- Assessment of Quality of Life in Clinical Trials of Cardiovascular
ceived ‘life control’ and ‘interference’ (e.g., disability). This Therapies, LeJacq, New York, 1984, pp. 170–183.
would explain why SF-36 Physical Function, a disability Ekter-Andersen, J., Janzon, L. and Sjølund, B., Chronic pain and the socio-
measure, in our study seemed to predict psychological demographic environment: results from the pain clinic at Malmø Gen-
eral Hospital in Sweden, Clin. J. Pain, 9 (1993) 183–188.
well-being better than pain intensity. However, more ade- Fink, P., Admissions of persons aged 17–49 years to non-psychiatric
quate measures of ‘life control’ or ‘interference’ would have departments, Ugeskr. Laeger, 151 (1989) 307–310.
to be used in order to test the role of cognitive-behavioral Flor, H., Fydric, T. and Turk, D.S., Efficacy of multidisciplinary pain
mediation in this patient sample. The apparent paradox of a treatment centers: a meta-analytic review, Pain, 49 (1992) 221–230.
relatively weak correlation between pain intensity and Foldspang, A., Juul, S., Olsen, J. and Sabroe, S., Analytisk epidemiologi.
In: A. Foldspang, S. Juul, J. Olsen and S. Sabroe (Eds.), Epidemiologi.
HRQL in spite of severe and uniform reduction in HRQL Sygdom og befolkning, Munksgaard, Copenhagen, 1981, pp. 114–
found among the chronic pain patients makes it relevant to 116.
question the scaling properties of the questionnaires used in Garratt, A.M., Ruta, D.A., Abdala, M.I., Buckingham, J.K. and Russel,
this study. However, the questionnaires used in this study I.T., The SF-36 health survey questionnaire: an outcome measure sui-
were all validated, and scores were highly correlated within table for routine use within the NHS?, Br. Med. J., 306 (1993) 1440–
1444.
constructs. Furthermore, it seems reasonable to expect only Hansson, P., Neurogenic pain: diagnosis and treatment, Pain Clin.
a moderate correlation between pain and HRQL, because Updates, II (3) (1994) 1–2.
several factors other than pain intensity influence HRQL. Kerns, R.D. and Turk, D.C., Depression and chronic pain: the mediating
In conclusion, the present study on chronic, non-malig- role of the spouse, J. Marr. Fam., 46 (1984) 845–852.
nant pain patients referred to a Danish multidisciplinary Latham, J. and Davis, B.D., The socio-economic impact of chronic pain,
Disabil. Rehabil., 16 (1994) 39–44.
pain center showed: (i) the majority of the patients had Lyons, R.A., Lo, S.V. and Littlepage, B.N.C., Comparative health status of
severe pain and a very low HRQL; (ii) the prevalence of patients with 11 common illnesses in Wales, J. Epidemiol. Commmun.
undiagnosed neurogenic pain conditions was high; (iii) the Health, 48 (1994) 388–390.
majority received substantial doses of short-acting opioids; McHorney, C.A., Ware, J.E., Lu, J.F. and Sherbourne, C.D., The MOS 36-
and (iv) chronic pain patients are among the most frequent item Short-Form Health Survey (SF-36): III. Tests of data quality, scal-
ing assumptions, and reliability across diverse patient groups, Med.
users of the health care system. The study confirms the Care, 32 (1994) 40–66.
severe multidimensional impact of chronic pain and demon- Naylor, A.S., Epilepsy and Psychiatric Disorder – A Comorbidity Study,
strates that HRQL of chronic non-malignant pain patients is FADL Publishers, Copenhagen, 1996, p. 108.
among the lowest observed for any medical condition. Nilges, P., Gerbershagen, H.U., Dietz, J. and Nasri, A.F., Pain stages are
necessary classification criteria for psychological research and treatment
evaluation. In: Abstracts, 8th World Congress on Pain, IASP Press,
Seattle, 1996, pp. 193–194.
Acknowledgements Omvik, P., Thaulow, E., Herland, O.B., Eide, I., Midha, R. and Turner,
R.R., Double blind, parallel, comparative study on quality of life during
The authors would like to thank Jacob Bjørner for supply- treatment with amlodipine or enalapril in mild or moderate hypertensive
ing data on Danish SF-36 norm scores. Thanks are also patients: a multicentre study, J. Hypertension, 11 (1993) 103–113.
Romano, J.M. and Turner, J.A., Chronic pain and depression: does
extended to Prof. Allan Krasnik at the Department of Social the evidence support a relationship?, Psychol. Bull., 97 (1985) 18–
Medicine, Copenhagen University for advice concerning 34.
interpretation of sociodemographic data. Rudy, T.E., Kerns, R.D. and Turk, D.C., Chronic pain and depression:
toward a cognitive-behavioral model, Pain, 35 (1988) 129–140.
Schipper, H., Guidelines and caveats for quality of life measurement in
clinical practice and research, Oncol. Huntingt., 4 (1990) 51–57.
References Sjøgren, P., Banning, A.M., Jensen, N.H., Klee, M. and Vainio, A., Man-
agement of cancer pain in Denmark: a nationwide questionnaire survey,
Aaronson, N.K., Quality of life research in cancer clinical trials: a need for Pain, 64 (1996) 519–522.
common rules and language, Oncol. Huntingt., 4 (1990) 59–66. Stewart, A.L., Greefield, S., Hays, R.D., Wells, K., Rogers, W.H., Berry,
Bech, P., Health related quality of life rating scales. In:P. Bech (Ed.), S.D., McGlynn, E.A. and Ware, J.E., The MOS short form general
Rating Scales for Psychopathology, Health Status and Quality of Life, health survey, Med. Care, 26 (1988) 724–735.
Springer-Verlag, Berlin, 1993, pp. 395–424. Stewart, A.L., Greefield, S., Hays, R.D., Wells, K., Rogers, W.H., Berry,
Bowsher, D., Neurogenic pain syndromes and their management, Br. Med. S.D., McGlynn, E.A. and Ware, J.E., Functional status and well-being of
Bull., 47 (1991) 644–666. patients with chronic conditions, J. Am. Med. Assoc., 262 (1989) 907–
Carsten, R.J., Parmelee, P.A., Kleban, M.H., Lawton, M.P. and Katz, I.R., 913.
The relationship among anxiety, depression, and pain in a geriatric Turk, D.C., Rudy, T.E. and Sorkin, B.A., Neglected topics in chronic pain
institutionalized sample, Pain, 61 (1995) 271–276. treatment outcome studies: determination of success, Pain, 53 (1993) 3–
Clausen, T.G., Eriksen, J. and Borgbjerg, F.M., Legal opioid consumption 16.
in Denmark 1981–1993, Eur. J. Clin. Pharmacol., 48 (1995) 321– Turk, D.C., Okifuji, A. and Scharff, L., Chronic pain and depression: role
325. of perceived impact and perceived control in different age cohorts, Pain,
Dimenäs, E., Gliese, H., Hallerbäck, B., Hernqvist, H., Svedlund, J. and 61 (1995) 93–101.
400 N. Becker et al. / Pain 73 (1997) 393–400

Ware, J.E., Snow, K.K., Kosinski, M. and Gandek, B., SF-36 Health Wells, K., Stewart, A., Hays, R.D., Burham, M.A., Rogers, W., Daniels,
Survey Manual and Interpretation Guide, New England Medical Center, M., Berry, S., Greenfield, S. and Ware, J., The functioning and well-
The Health Institute, Boston, MA, 1993. being of depressed patients, J. Am. Med. Assoc., 262 (1989) 914–
Ware, J.E., Gandek, B. and the IQOLA Project Group, The SF-36 Health 919.
Survey: development and use in metal health research and the IQOLA Zigmond, A.S. and Snaith, R.P., The Hospital Anxiety and Depression
Project, Int. J. Ment. Health, 23 (1994) 49–73. Scale, Acta Psychiatr. Scand., 67 (1983) 361–370.