Open Access Highly Accessed Open Badges Research article

Reliability of the Multidimensional Pain Inventory and stability of the MPI classification system in chronic back pain

Martin L Verra123*, Felix Angst2, J Bart Staal4, Roberto Brioschi2, Susanne Lehmann2, André Aeschlimann2 and Rob A de Bie3

Author Affiliations

1 Department of Physiotherapy, Inselspital, Bern University Hospital, 3010, Bern, Switzerland

2 Rehabilitation clinic “RehaClinic”, 5330, Bad Zurzach, Switzerland

3 Department of Epidemiology and Caphri Research School, Maastricht University, Maastricht, The Netherlands

4 Scientific Institute for Quality of Healthcare, Radboud University Medical Centre, Nijmegen, The Netherlands

For all author emails, please log on.

BMC Musculoskeletal Disorders 2012, 13:155  doi:10.1186/1471-2474-13-155

Published: 24 August 2012



This cross validation study examined the reliability of the Multidimensional Pain Inventory (MPI) and the stability of the Multidimensional Pain Inventory Classification System of the empirically derived subgroup classification obtained by cluster analysis in chronic musculoskeletal pain. Reliability of the German Multidimensional Pain Inventory was only examined once in the past in a small sample. Previous international studies mainly involving fibromyalgia patients showed that retest resulted in 33–38% of patients being assigned to a different Multidimensional Pain Inventory subgroup classification.


Participants were 204 persons with chronic musculoskeletal pain (82% chronic non-specific back pain). Subgroup classification was conducted by cluster analysis at 4 weeks before entry (=test) and at entry into the pain management program (=retest) using Multidimensional Pain Inventory scale scores. No therapeutic interventions in this period were conducted. Reliability was quantified by intraclass correlation coefficients (ICC) and stability by kappa coefficients (κ).


Reliability of the Multidimensional Pain Inventory scales was least with ICC = 0.57 for the scale life control and further ranged from ICC = 0.72 (negative mood) to 0.87 (solicitous responses) in the other scales. At retest, 82% of the patients in the Multidimensional Pain Inventory cluster interpersonally distressed (κ = 0.69), 80% of the adaptive copers (κ = 0.58), and 75% of the dysfunctional patients (κ = 0.70) did not change classification. In total, 22% of the patients changed Multidimensional Pain Inventory cluster group, mainly into the adaptive copers subgroup.


Test-retest reliability of the German Multidimensional Pain Inventory was moderate to good and comparable to other language versions. Multidimensional Pain Inventory subgroup classification is substantially stable in chronic back pain patients when compared to other diagnostic groups and other examiner-based subgroup Classification Systems. The MPI Classification System can be recommended for reliable and stable specification of subgroups in observational and interventional studies in patients with chronic musculoskeletal pain.

Reliability; Back pain; Cluster; Subgroup; MPI; Classification