Reliability and validity of manual palpation for the assessment of patients with low back pain: a systematic and critical review

Key points

1. 01Inter-rater reliability for joint and bony structure palpation (lumbar facets, spinous processes, sacroiliac joints) is poor, with most kappa values well below 0.6
2. 02Soft tissue palpation reliability is inconsistent: sciatic nerve palpation (k=0.80) and multifidus contraction (k=0.75-0.81) are reasonably reliable, but lumbar paraspinal muscle palpation is not (k=0.34)
3. 03Posterior-to-anterior palpation to detect spinal stiffness performed no better than chance (sensitivity 38%, specificity 45%)
4. 04Gluteal trigger point palpation showed good accuracy for distinguishing radicular from non-radicular low back pain (sensitivity 74.1%, specificity 91.4%)
5. 05Only 14 of 2023 screened articles had low risk of bias, highlighting the generally poor quality of evidence in this field

How it was conducted

Design

Systematic review and critical appraisal with best-evidence synthesis

Databases searched

MEDLINE, CINAHL, PubMed, Cochrane Central Register of Controlled Trials, SPORTDiscus (2000-2019)

Articles screened

2023 articles reviewed; 14 with low risk of bias included in evidence synthesis

Quality appraisal tools

Modified QAREL (reliability studies) and QUADAS-2 (validity studies)

Validity classification

Sackett and Haynes Phase I-IV framework applied to validity studies

Primary outcomes

Inter-rater and intra-rater reliability (kappa); validity (sensitivity, specificity, likelihood ratios)

What they found

• Inter-rater reliability for eliciting pain from lumbar facet joints: 0.38 <= k <= 0.73; lumbar spinous processes: 0.21 <= k <= 0.57; sacroiliac joints: 0.14 <= k <= 0.59
• Inter-rater reliability for static joint segmental mobility: lumbar facet joints -0.17 <= k <= 0.17; lumbar spinous processes -0.02 <= k <= 0.26; sacroiliac joints -0.11 <= k <= -0.10
• Prone instability test inter-rater reliability ranged from k=0.10 to k=0.87 depending on study and test phase
• Sciatic nerve palpation inter-rater reliability: k=0.80 (95% CI 0.39-0.94); multifidus lift test: k=0.75-0.81; gluteal tender points: k=0.51-0.68; lumbar paraspinal: k=0.34
• Posterior-to-anterior spinal stiffness palpation (L1-L5): sensitivity 38% (95% CI 21-59%), specificity 45% (95% CI 28-62%), +LR 0.69 (95% CI 0.37-1.31), -LR 1.38 (95% CI 0.82-2.33)
• Sciatic nerve palpation for mechanosensitivity (vs SLR and slump tests): sensitivity 85% (95% CI 75-95%), specificity 60% (95% CI 46-74%), +LR 2.25, -LR 0.25
• Gluteal trigger point palpation for radiculopathy (phase III): sensitivity 74.1% (95% CI 67.7-80.3%), specificity 91.4% (95% CI 86.8-96.0%), +LR 8.62, -LR 0.28, ROC 0.827 (95% CI 0.781-0.874)
• Motion palpation of sacroiliac joints: inter-rater reliability 0.14 <= k <= 0.75; sacral base position test kappa: flexion k=0.37, extension k=0.05

Limitations

• Only 14 of 2023 screened studies had low risk of bias, making the evidence base very thin and potentially unrepresentative
• Motion palpation studies had particularly small sample sizes (validity n=50, reliability n=49), limiting precision
• Search was restricted to English and French publications from 2000 onwards, potentially missing relevant work
• Most validity studies are only Phase I or II, meaning they cannot confirm a test is clinically valid on their own

Why it matters

For patients

Patients should be aware that most hands-on spinal feeling tests used to locate their pain source have limited agreement between practitioners and little proven accuracy, so a positive or negative finding may not be reliable.

For clinicians

Clinicians should exercise caution when using joint palpation to guide diagnosis or treatment decisions in low back pain; gluteal muscle trigger point palpation is the best-supported test for distinguishing radicular pain, but most other palpation tests lack sufficient validity evidence.

For readers

This review highlights a significant gap: manual palpation is widely used in clinical practice for low back pain assessment, yet robust evidence for its reliability and validity is almost entirely absent, underscoring the need for high-quality Phase III and IV diagnostic accuracy studies.