Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Classifying chronic lower back pain groups using a time series model of lifting

Slaboda, Jill Christina (2007) Classifying chronic lower back pain groups using a time series model of lifting. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Primary Text

Download (2MB) | Preview


A classification procedure was developed that uses hidden Markov models (HMMs) to identify sub-groups within a chronic lower back pain (CLBP) patient population based on their time series of lifting patterns during a repetitive lifting task. Based on clinical observations of a repetitive lifting task, our approach assumed that the patient population was composed of two groups: one group that performed lifts more similar to controls than to other patients and another group that lifted differently from control subjects. Two HMMs were designed to describe the repetitive lifting data, one derived from the control subject data and one derived from the CLBP subject data. The HMMs were designed based on the results of a data reduction procedure that reduced and combined the multidimensional lifting parameters into discrete lifting patterns using factor analysis and cluster analysis. Simulation studies were performed to demonstrate that the HMMs could reliably identify subjects from one group that were intentionally mislabeled as the other group. When the HMMs were applied to clinical data, 35 of the 81 CLBP subjects were classified to the control HMM and 46 were classified to the CLBP HMM. For the control group, 46 of 53 control subjects were classified to the control HMM and only seven were classified to the CLBP HMM. The CLBP groups were found to use different lifting patterns during the task. The CLBP subjects that were classified to the CLBP HMM were found to use a lifting pattern that involves slow, controlled movements. Self-reported measures of the two groups of CLBP subjects were compared and self-reported pain intensity, pain severity and perceived self-efficacy found to be statistically different. The CLBP subjects that were classified to the CLBP HMM reported higher pain intensity and pain severity, and lower self-efficacy suggesting that the CLBP population is heterogeneous and that the HMM classification procedure can successfully identify two meaningfully different sub-groups of CLBP patients.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Slaboda, Jill Christinajcsst46@pitt.eduJCSST46
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairBoston, J.Robertboston@engr.pitt.eduBBN
Committee CoChairRudy,
Committee MemberEl-Jaroudi, Amroamro@ee.pitt.eduAMRO
Committee MemberMiller, Markmcmllr@engr.pitt.eduMCMLLR
Committee MemberRedfern,
Committee MemberCooper, Roryrcooper@pitt.eduRCOOPER
Date: 12 June 2007
Date Type: Completion
Defense Date: 29 March 2007
Approval Date: 12 June 2007
Submission Date: 16 March 2007
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Institution: University of Pittsburgh
Schools and Programs: Swanson School of Engineering > Bioengineering
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: hidden Markov models; chronic lower back pain; data reduction
Other ID:, etd-03162007-163608
Date Deposited: 10 Nov 2011 19:32
Last Modified: 15 Nov 2016 13:37


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item