Evaluation of a treatment-based classification algorithm for low back pain

Tasha R Stanton1, Julie M Fritz2, Mark J Hancock3, Jane Latimer4, Chris G. Maher4, Benedict M Wand5, Eric Parent6
1 Sansom Institute for Health Research, University of South Australia, Adelaide; Neuroscience Research Australia, Sydney
2 Department of Physical Therapy, College of Health, University of Utah, United States
3 Universityof Sydney, Faculty of Health Sciences, Discipline of Physiotherapy, Australia
4 Musculoskeletal Division, The George Institute for Global Health; The University of Sydney, Australia
5 School of Physiotherapy, University of Notre Dame Australia
6 Department of Physical Therapy, Faculty of Rehabilitation Medicine, University of Alberta, Canada

Background Methods
Many individual studies have investigated subgroups of low back pain (LBP) 250 patients with acute/subacute LBP were recruited from USA and
that respond best to certain interventions. A classification algorithm was Australia
created to combine these individual study sub-grouping criteria into a clinical Each patient underwent a standardized assessment and was classified
decision-making guide that is mutually exclusive and comprehensive. In order into a treatment group using the classification algorithm
to do this, certain changes to the individual study criteria had to be made. Each patient was also classified using the individual study subgrouping
This study aims to evaluate these changes. criteria
31 patients were assessed twice to determine the reliability of the
Individual Study Sub-grouping Criteria algorithm.

Main Results
Prevalence rates: Individual Study Sub-grouping Criteria

The classification algorithm prioritises

manipulation over specific exercise.

25% of patients do not meet any of the

treatment subgroups based on the individual
study criteria suggesting when using the
algorithm with these patients, we would need
to use the bottom table. However, 34% of
patients have unclear classifications when
using the algorithm = 9% of patients that
Classification Algorithm
should have clear classification but dont.
NO Doespatienthavesymptomsdistaltothebuttock
YES
andsignsofnerverootcompression?
25% of patients meet more than one
Doesthepatient:
1.Centralizewith2ormoremovementsinthesame Traction
Doesthepatient:
1.Peripheralize withextensionmovement treatment subgroup based on the individual
direction(ie,flexionorextension) Classification OR
OR
2.Centralizewithmovementinonedirectionand
SpecificExercise
Classification
2.Positivecrossedstraightlegraisetest
study criteria. This confirms the need for
peripheralize withanoppositemovement
Doesthepatient: hierarchical ordering of the algorithm so that a
1.Centralizewith2ormoremovementsinthe
Doesthepatient:
1.havedurationofsymptoms<16days
AND
Manipulation
SpecificExercise
Classification
samedirection(ie,flexionorextension)
OR
patient is only assigned to one subgroup. It is
Classification 2.Centralizewithmovementinonedirectionand
2.Nosymptomsdistaltotheknee
peripheralize withanoppositemovement Prevalence rates: The Classification Algorithm unclear what treatment a patient should get
Doesthepatienthave3ormore:
1.AverageSLRROM>91
Stabilization
first if they meet more than one subgroup.
2.Positiveproneinstabilitytest
Classification
3.Positiveaberrantmovements
4.Age<40years Having an unclear classification when using
WhichSubGroupdoesthepatientfitbest? WhichSubGroupdoesthepatientfitbest? the algorithm appears to adversely affect the
reliability of the classification decision.
MANIPULATION STABILIZATION SPECIFICEXERCISE TRACTION
Factorsfavoring Factorsagainst Factorsfavoring Factorsagainst Factorsfavoring Factorsagainst Factorsfavoring Factorsagainst Conclusions
Hypomobility Symptomsbelow Hypermobility Discrepancyin Directional Lowbackpain Peripheralization Lowbackpain
withspringtesting
LowFABQscores
theknee
Increasing
withspringtesting
Increasing
SLRROM(>100)
LowFABQscores
preferencefor
extensionor
only(nodistalsx)
StatusQuowith
ofsymptomswith
noabilityto
only(nodistalsx)
Nosignsofnerve
This provides important clinical data to guide potential revisions
(FABQW<19)
Hipinternal
episodefrequency
Peripheralization
episodefrequency
3ormoreprior
(FABQPA<9) flexion
Centralization
allmovements centralizewith
movement
rootcompression
to the algorithm. That 25% of patients meet more than one
rotationROM>350 withmotion episodes withmotion
testing
Nopainwith
testing
Peripheralize in subgroup has implications for sequencing of treatments in the
springtesting directionopposite
centralization
Reliability of the Classification Decision using the Algorithm algorithm. The finding that 25% of patients do not meet any
Top boxes = clear classification; Bottom table = unclear classification
subgroups will help guide revisions to the bottom table of the
Specific Aims Inter-rater reliability was moderate algorithm (which guide unclear classifications).
To determine the prevalence of patients meeting each treatment (kappa = 0.52; 95%CI 0.27 to 0.77)
subgroup/none of the subgroups/more than one subgroup using the Future directions of research:
individual study criteria and to compare these rates to the classification Reliability of clear classification was good Further research is needed to explore the treatment options for
algorithm treatment subgroup prevalence rates. (kappa = 0.69; 95% CI 0.42 to 0.96) patients that meet more than one subgroup which treatment
should they get? Future research should also explore the
To determine the reliability of the classification decision when using the Reliability of unclear classifications was poor potential addition of other treatments to the algorithm (as
classification algorithm. (kappa = 0.23; 95% CI -0.21 to 0.66) evidence permits).
 Doespatienthavesymptomsdistaltothebuttockand
NO YES
signsofnerverootcompression?

Doesthepatient: Doesthepatient:
1.Centralizewith2ormoremovementsinthesame Traction 1.Peripheralizewithextensionmovement
direction(ie,flexionorextension) Classification OR
OR 2.Positivecrossedstraightlegraisetest
SpecificExercise
2.Centralizewithmovementinonedirectionand
Classification
peripheralizewithanoppositemovement

Doesthepatient:
1.Centralizewith2ormoremovementsinthe
Doesthepatient: samedirection(ie,flexionorextension)
1.havedurationofsymptoms<16days SpecificExercise
Manipulation OR
AND Classification 2.Centralizewithmovementinonedirectionand
2.Nosymptomsdistaltotheknee Classification
peripheralizewithanoppositemovement

Doesthepatienthave3ormore:
1.AverageSLRROM>91
2.Positiveproneinstabilitytest Stabilization
3.Positiveaberrantmovements Classification
4.Age<40years

WhichSubGroupdoesthepatientfitbest? WhichSubGroupdoesthepatientfitbest?

MANIPULATION STABILIZATION SPECIFICEXERCISE TRACTION

Factorsfavoring Factorsagainst Factorsfavoring Factorsagainst Factorsfavoring Factorsagainst Factorsfavoring Factorsagainst
Hypomobilitywith Symptomsbelowthe Hypermobilitywith DiscrepancyinSLR Directional Lowbackpainonly Peripheralizationof Lowbackpainonly
0
springtesting knee springtesting ROM(>10 ) preferencefor (nodistalsx) symptomswithno (nodistalsx)
LowFABQscores Increasingepisode Increasingepisode LowFABQscores extensionorflexion StatusQuowithall abilitytocentralize Nosignsofnerve
(FABQW<19) frequency frequency (FABQPA<9) Centralizationwith movements withmovement rootcompression
Hipinternalrotation Peripheralization 3ormoreprior motiontesting
0
ROM>35 withmotiontesting episodes Peripheralizationin
Nopainwithspring directionopposite
testing centralization