Related Experiment Videos
Low back pain recollection versus concurrent accounts: outcomes analysis
Edgar G Dawson1, Linda E A Kanim, Parveen Sra
1UCLA Comprehensive Spine Center and Department of Orthopaedic Surgery, UCLA School of Medicine, Los Angeles, California 90404, USA. lkanim@espineinstitute.com
Spine
|April 30, 2002
Summary
Patient recall of low back pain and function over 10 years showed fair agreement with initial reports. Accuracy varied by question type, with location and frequency being more reliable than severity.
Area of Science:
- Orthopedics
- Pain Management
- Clinical Epidemiology
Background:
- Assessing patient outcomes via recall is subject to biases like age, gender, and current health.
- The impact of retrospective data on spinal patient outcome analyses remains under-explored.
Purpose of the Study:
- Compare patient-recalled data with prospectively collected data for low back pain.
- Identify reliable, symptom-specific questions for long-term outcome assessment in low back pain patients.
Main Methods:
- Utilized data from the National Low Back Pain prospective follow-up study (1986-1991).
- Compared 1998 mail follow-up responses with recalled data from a 1996 telephone interview and initial 1986-1991 questionnaire responses.
- Evaluated agreement between recalled and initial responses using Cohen's kappa.
Main Results:
- 144 patients completed the 9.4-year follow-up.
- Overall agreement (kappa=0.37) between initial and recalled data was "fair".
- Kappa values ranged from 0.12-0.58 for pain location, 0.28-0.48 for radicular symptoms, and 0.11-0.30 for pain severity.
Conclusions:
- Agreement between recalled and initial low back pain reports was "fair" to "moderate".
- Questions on pain frequency, location, and affecting activities demonstrated higher accuracy.
- Recalled data showed a bias towards underestimating pain severity; careful question selection is crucial for accurate outcome measures.