There is substantial variability in the clinical course of people presenting with acute low back pain (LBP) – pain that lasts for up to 3 months. The majority of patients with acute LBP recover from the episode within a few weeks or months;[1] however, substantial variability exists between patients: some will recover within a few days, others will recover more slowly and others will not recover at all.[2-4]
Recent guidelines[5,6] recommend the use of tools to help predict probable outcomes after an episode of LBP to guide the management in primary care. These tools – also known as “clinical prediction models” – use individual’s characteristics to estimate probabilities of outcomes.[7,8] The use of these tools can help clinicians in making informed decisions about the amount and type of treatment to provide, and potentially reduce costs.
There are existing clinical prediction models that estimate the prognosis of patients with LBP, but have important limitations. Most focus on predicting persistent pain or non-recovery at 12-months. While this information is useful, focusing on predicting an outcome at 1-year may not be highly relevant to clinicians making decisions about short and intermediate term management of people with acute LBP. Also, most clinical prediction models lack acceptable discrimination (the ability of the model to differentiate patients who have the outcome and those who do not have the outcome) and/or accuracy (whether the observed frequencies agree with the predicted probabilities).
We recently published two studies about a new clinical prediction model for patients with acute LBP. The first study[9] developed the clinical prediction model which aimed to predict the probability of recovery from pain at 1-week, 1-month and 3-months post 1-week clinical review in patients who still have LBP when reviewed 1-week after initially seeking care. The prediction model was developed to be used at the “baseline” of 1 week after the patient initially sought care as it allowed the pain intensity change over the first week to be included as a predictor. Pain intensity change over the first week has been previously found to be predictive of outcome in patients with LBP.[10,11] Another reason to use this clinical prediction model at this “baseline” is that it aligns with clinical guidelines for acute LBP management, which recommend minimal intervention with a review in the first 1 to 2 weeks to check if additional treatment is necessary.[5,12]
The study found that the developed clinical prediction model, using five variables that can be easily collected as part of routine care, was able to predict the likelihood of pain recovery in patients with acute LBP at 1 week, 1 and 3 months after “baseline” 1-week review. The model had good discrimination (according to the of C-statistic=0.76, 95% CI=0.69, 0.81) and calibration.
Then, we conducted a second study[13] to validate the prediction model in a different sample. Three variables of the development dataset (from the previous study) needed minor re-categorization to enable testing in the new data set. The discrimination of the prediction model with re-categorized variables in the development dataset was good (C‐statistic=0.76, 95% CI=0.70, 0.82). The discrimination of the model using the validation dataset resulted in a C‐statistic of 0.71. The calibration for the validation sample was acceptable at 1-month. However, at 1-week the predicted proportions of patients who would recover overestimated the observed recovery proportions, and at 3-months the predicted proportions tended to underestimate the observed recovery proportions.
The findings of our studies indicate that this tool can provide important information to patients and clinicians and may help in shared decision making. For example, a patient with a favourable prognosis and high likelihood of recovery by 1-week and 1-month, may be reassured and decide to continue simple baseline care rather than receive additional intervention. Alternatively, a patient with low probability of recovery by 3-months may be more likely to decide to receive additional intervention. However, the information provided by our studies does not indicate that the use of the clinical prediction model improves health outcomes and costs associated with acute LBP. Definitive evidence for this can only be gained by testing the tool in a randomised controlled trial.
About Tati Mota
Tati has recently completed her PhD on the topic of predicting recovery for acute LBP, and risk of recurrences of LBP. Her PhD was completed at Macquarie University, under supervision of Prof Mark Hancock, with a scholarship from Brazil’s government. Her research interests include prediction of recovery of acute low back pain, understanding the risk of recurrences of LBP, effectiveness of interventions for LBP, and investigating the use of evidence-based practice by clinicians. Tati will continue her work as a post-doctoral researcher at Universidade Cidade de Sao Paulo (UNICID), Brazil.
References
[1] da CMCL, Maher CG, Hancock MJ, McAuley JH, Herbert RD, Costa LO. The prognosis of acute and persistent low-back pain: a meta-analysis. CMAJ. 2012;184(11):E613-624.
[2] Downie AS, Hancock MJ, Rzewuska M, Williams CM, Lin CW, Maher CG. Trajectories of acute low back pain: a latent class growth analysis. Pain. 2016;157(1):225-234.
[3] Henschke N, Maher CG, Refshauge KM, et al. Prognosis in patients with recent onset low back pain in Australian primary care: inception cohort study. BMJ. 2008;337:a171.
[4] Itz CJ, Geurts JW, van Kleef M, Nelemans P. Clinical course of non-specific low back pain: a systematic review of prospective cohort studies set in primary care. Eur J Pain. 2013;17(1):5-15.
[5] Maher C, Underwood M, Buchbinder R. Non-specific low back pain. Lancet. 2017;389(10070):736-747.
[6] In: Low Back Pain and Sciatica in Over 16s: Assessment and Management. London2016.
[7] Laupacis A, Sekar N, Stiell IG. Clinical prediction rules. A review and suggested modifications of methodological standards. JAMA. 1997;277(6):488-494.
[8] Reilly BM, Evans AT. Translating clinical research into clinical practice: impact of using prediction rules to make decisions. Ann Intern Med. 2006;144(3):201-209.
[9] da Silva T, Macaskill P, Mills K, et al. Predicting recovery in patients with acute low back pain: A Clinical Prediction Model. Eur J Pain. 2017;21(4):716-726.
[10] Dunn KM, Jordan K, Croft PR. Characterizing the course of low back pain: a latent class analysis. Am J Epidemiol. 2006;163(8):754-761.
[11] Heymans MW, van Buuren S, Knol DL, Anema JR, van Mechelen W, de Vet HC. The prognosis of chronic low back pain is determined by changes in pain and disability in the initial period. Spine J. 2010;10(10):847-856.
[12] Koes BW, van Tulder M, Lin CW, Macedo LG, McAuley J, Maher C. An updated overview of clinical guidelines for the management of non-specific low back pain in primary care. Eur Spine J. 2010;19(12):2075-2094.
[13] da Silva T, Macaskill P, Kongsted A, Mills K, Maher CG, Hancock MJ. Predicting pain recovery in patients with acute low back pain: Updating and validation of a clinical prediction model. Eur J Pain. 2019;23(2):341-353.