Prognostication in Low Back Pain:A Biopsychosocial Index

Abstract Introduction. The biopsychosocial model has developed over the last forty years since Melzac and Wall first stated the need to broaden the biomedical view of pain mechanisms. The biopsychosocial concept is now central in contemporary understanding of human health and functioning in general, and pain in particular. An extensive biopsychosocial literature on low back pain has emerged since the paradigm shift of nearly thirty years ago, and has been reflected in evidence-based clinical practice guidelines on musculoskeletal pain. The estimation of prognosis from an episode of low back pain is particularly important to clinicians, patients, employers and third party payers. The instruments currently available however, quantify prognostic factors from a biomedical perspective, rather than the contemporary biopsychosocial model. The successful development of the biopsychosocial instrument, reported in this thesis, fills this gap in measurement, and affirms the hypothesis advanced that it is feasible to develop a valid and reliable multidimensional index to estimate prognosis from a range of biopsychosocial variables, in low back pain. The thesis describes; • A review of pain mediation/modulation • A review of the literature on the biopsychosocial model to identify a range of variables for investigation • Consideration of different outcome measures as the gold standard criterion measure • The development of the BPIP, a novel biopsychosocial instrument designed for eventual clinical use, as a means of considering non specific low back pain from a biopsychosocial perspective, and informing prognosis in patients who do not have diagnosable levels of psychopathology Methods. The Biopsychosocial Index of Prognosis (BPIP) was developed from a content map derived from the International Classification of Functioning, Health and Disease (ICF) for use in non specific low back pain of variable duration. A prototype BPIP instrument was drafted, and a peer review process resulted in item reduction. The prototype BPIP instrument was then subjected to a twenty four hour test-retest assessment of stability, prior to assessment of the prototype BPIP’s validity and internal consistency. Questionnaire packs containing the prototype BPIP instrument, a range of other measurement instruments, an informed consent form and an instruction sheet were then issued. The data were captured at baseline, six weeks, and again at twelve weeks. Recruitment of respondents was originally intended to capture a homogeneous cohort drawn from general (medical) practice. Recruitment difficulties however resulted in two distinct cohorts, a larger Australian cohort (n = 91) and a smaller New Zealand cohort (n = 27). The larger Australian group was utilised as an elucidation cohort, and the smaller New Zealand cohort was treated as a small prospective validation of the BPIP. Item reduction was undertaken and resulted in an ordinal scale for correlation with the Roland and Morris Disability Index, the criterion measure. The resulting twenty four item BPIP scale was tested for internal consistency in both the Australian and New Zealand cohorts. The baseline data from the Australian elucidation cohort were then assessed for concurrent, predictive and construct validity against the twelve week data from The Roland and Morris Disability Index (RDQ), the selected criterion measure. The New Zealand cohort was treated as a small prospective validation of the BPIP scale, with the baseline data from the BPIP correlated with the change score from the baseline to twelve weeks, of the RDQ. Analysis was undertaken using the SPSS statistical package. Results. Correlation analysis of the baseline BPIP Australian data with the twelve week RDQ Australian data resulted in reducing the prototype scale to questions which correlated at or above 0.3, a total of twenty four questions were retained. Reliability coefficients for internal consistency of the twenty four item BPIP scale were: The Australian cohort, Cronbach’s Alpha = 0.8736. The New Zealand cohort, Cronbach’s Alpha = 0.8628. A further review of the correlation analysis of the baseline BPIP Australian data with twelve week RDQ Australian data for items that correlated at or above 0.4 resulted in further item reduction of the BPIP to twelve questions. Following this further item reduction, reliability coefficients for internal consistency were: The Australian cohort, Cronbach’s Alpha = 0.875. The New Zealand cohort, Cronbach’s Alpha = 0.776. Regression analysis of the Australian cohort based on the twelve item scale demonstrated that 61.7% of the variance in the RDQ score at twelve weeks was accounted for by the BPIP scale, with p = 0.0005. Regression analysis of the change score of the RDQ with the twelve item BPIP in the New Zealand cohort demonstrated that 78.2% of the variance in RDQ scores was accounted for by the BPIP scale, with p = 0.006. Forty five point five percent of the Australian cohort improved by more than 30%, (the proposed minimal clinically important difference of the RDQ), whilst 76.9% of the New Zealand cohort improved more than thirty percent. In a post hoc analysis of the Orebro data, the longer 25 item questionnaire accounted for 97.8% of the variance in the New Zealand cohort. Conclusion. The initial hypothesis that it would be feasible to develop a valid and reliable multidimensional instrument from a range of biopsychosocial variables into a valid instrument for estimating the prognosis of an episode of low back pain is supported by the results. The utility of a biopsychosocial instrument for routine clinical use lies in the potential to predict prognosis. As low back pain is typically a recurrent problem, information on both prognosis for recovery from episodes, and the likelihood of recurrences would be helpful to both clinician and patient. For the patient, a well communicated prognosis helps assure the patient about their future, reduces uncertainty about their pain, and establishes treatment goals within the domain of informed consent. From the clinician’s perspective, a prognostic approach shifts the focus from the pain history to future outcomes and provides a context for considering how risks of future pain and dysfunction may be reduced. The BPIP scale is a biopsychosocial, prognostic instrument, which accounts for a high degree of the variance in the RDQ scores in both cohorts. Pain amelioration and functional improvement are the two key aspects of prognosis which the BPIP has been demonstrated to be both valid and reliable in predicting at the clinically important three month time point from baseline assessment. The BPIP has been demonstrated in these cohorts, to provide a reliable estimate of prognosis from a biopsychosocial perspective. The reliability of the shortened BPIP scale remained acceptable, and allowed for the scale to be contained on a single A4 page, potentially increasing the clinical utility of the instrument. When BPIP score fails to change over time, psychosocial screening and intervention may be indicated. Future work will include further validation in other subgroups and clinical environments, identification of cut points for BPIP scores, and the performance of comparative studies of the relative value of different purported prognostic indices. The BPIP is the first instrument developed to specifically to estimate prognosis from an episode of low back pain, in primary care, within the contemporary biopsychosocial model.