Research Report Abstracts - Gesundheit

eS478
research funding supported from the Faculty of Associated
Medical Sciences, Chiang Mai University.
Ethics approval: This study was approved by the Human
Experimental Committee, Faculty of Associated Medical
Sciences, Chiang Mai University.
Research Report Platform Presentation
Number: RR-PL-2087 Wednesday 22 June 11:15
RAI: Elicium 1
IDENTIFYING GENERIC PREDICTORS OF
OUTCOME IN PATIENTS PRESENTING TO
PRIMARY CARE WITH MUSCULOSKELETAL
PAIN
Henschke N. 1,2 , Ostelo R. 2,3 , Terwee C. 2 , van der Windt D. 4
1The George Institute for International Health, Musculoskeletal
Division, Sydney, Australia, 2EMGO+ Institute
for Health and Care Research, VU University Medical Center,
Epidemiology & Biostatistics, Amsterdam, Netherlands,
3VU University, Department of Health Sciences, Amsterdam,
Netherlands, 4Keele University, Arthritis Research Campaign
National Primary Care Centre, Keele, United Kingdom
Purpose: The aim of the study was to identify which predictors
are most strongly associated with a poor outcome in
patients with musculoskeletal pain, regardless of the location
of pain. We tested the hypotheses that pain location does
not add predictive value to these generic prognostic models
and that such prognostic factors are equally important across
different pain locations.
Relevance: Musculoskeletal disorders are one of the most
common causes of disability, especially in older people. The
prevalence of these disorders and associated musculoskeletal
pain is expected to increase dramatically in coming decades
as the population ages. In order to provide optimal care to
patients with musculoskeletal pain, it is important that primary
care clinicians can identify patients who have a higher
risk of poor outcome.
Participants: Patients were eligible for participation in the
study if they met the following inclusion criteria: visited their
general practitioner with a new episode of pain in the neck,
shoulder, elbow, wrist, hand, arm, hip, knee, ankle, or foot;
were 18 years or older, and were capable of filling in Dutch
language questionnaires.
Methods: This study is based on a prospective observational
cohort of primary care patients with non-spinal musculoskeletal
pain in the Netherlands. Participants were assessed
at baseline and followed-up at 3 months and 12 months after
the initial consultation in primary care, with questionnaires
being mailed to all participants.
Analysis: Multiple logistic regression analyses were performed
to determine which factors, irrespective of location,
were associated with the risk of “poor” outcome. The analysis
was carried out in three steps: (1) derivation of predictive
models including generic factors only; (2) investigation of
added predictive value of pain location; and (3) investigation
of effect modification by pain location.
Results: The cohort included 1123 patients with non-spinal
musculoskeletal pain. For patients with acute musculoskeletal
pain, factors which predicted a poor outcome included:
lower level of education, having had the same complaint in
the past year, having multiple areas of musculoskeletal pain,
and having lower scores on the SF-36 vitality subscale. For
patients presenting with chronic musculoskeletal pain, factors
which predicted a poor outcome included: having had
the complaint before in the past year, low level of education,
being more bothered by the complaint in the past 3 months,
and higher scores on the SF-36 “physical” subscale. Variables
regarding the location of pain did not significantly add to the
models and few locations of pain acted as effect modifiers for
these predictors.
Conclusions: When estimating which patients with musculoskeletal
pain are at risk of a poor outcome, a range of generic
factors are likely to be useful, irrespective of the specific location
of pain. The similarities observed in the prognosis of
different musculoskeletal pain complaints and the identification
of consistent predictive factors supports a move towards
the development of a core set of features for the assessment
of all musculoskeletal pain conditions.
Implications: The evaluation of generic predictors can assist
clinicians to identify patients who are at risk of a poor
outcome and may lead to the development of improved management
strategies.
Keywords: Musculoskeletal pain; Prognosis
Funding acknowledgements: NH is a research fellow of the
National Health and Medical Research Council of Australia.
The original data was collected in a study supported by the
Dutch Arthritis Association.
Ethics approval: The Medical Ethics Committee of the VU
University Medical Centre approved the original study protocol.