Validation of general pain scores from multidomain assessment tools in stroke

Myzoon Ali*, Holly Tibble, Marian C. Brady, Terence J. Quinn, Katharina S. Sunnerhagen, Narayanaswamy Venketasubramanian, Ashfaq Shuaib, Anand Pandyan, Gillian Mead

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

13 Downloads (Pure)


PURPOSE: We describe how well general pain reported in multidomain assessment tools correlated with pain-specific assessment tools; associations between general pain, activities of daily living and independence after stroke.

MATERIALS AND METHODS: Analyses of individual participant data (IPD) from the Virtual International Stroke Trials Archive (VISTA) described correlation coefficients examining (i) direct comparisons of assessments from pain-specific and multidomain assessment tools that included pain, (ii) indirect comparisons of pain assessments with the Barthel Index (BI) and modified Rankin Scale (mRS), and (iii) whether pain identification could be enhanced by accounting for reported usual activities, self-care, mobility and anxiety/depression; factors associated with pain.

RESULTS: European Quality of Life 3- and 5-Level (EQ-5D-3L and EQ-5D-5L), RAND 36 Item Health Survey 1.0 (SF-36) or the 0-10 Numeric Pain Rating Scale (NPRS) were available from 10/94 studies (IPD = 10,002). The 0-10 NPRS was the only available pain-specific assessment tool and was a reference for comparison with other tools. Pearson correlation coefficients between the 0-10 NPRS and (A) the EQ-5D-3L and (B) EQ5D-5 L were r = 0.572 ( n = 436) and r = 0.305 ( n = 1,134), respectively. mRS was better aligned with pain by EQ-5D-3L ( n = 8,966; r = 0.340) than by SF-36 ( n = 623; r = 0.318). BI aligned better with pain by SF-36 ( n = 623; r = -0.320). Creating a composite score using the EQ-5D 3 L and 5 L comprising pain, mobility, usual-activities, self-care and anxiety/depression did not improve correlation with the 0-10 NPRS.

DISCUSSION: The EQ-5D-3L pain domain aligned better than the EQ-5D-5L with the 0-10 NPRS and may inform general pain description where resources and assessment burden hinder use of additional, pain-specific assessments.

Original languageEnglish
Article number1328832
JournalFrontiers in Neurology
Publication statusPublished - 25 Jan 2024


  • stroke
  • pain
  • assessment optimisation
  • EQ-5D-3L
  • EQ-5D-5L
  • NPRS stroke
  • EQ-5D
  • assessment

ASJC Scopus subject areas

  • Clinical Neurology
  • Neurology


Dive into the research topics of 'Validation of general pain scores from multidomain assessment tools in stroke'. Together they form a unique fingerprint.

Cite this