RESEARCH ARTICLE A subgroup analysis to compare patients with acute low back pain classified as per treatment‐based classification G. Shankar Ganesh 1 | Pradeep Kumar Sahu 2 | Sakti Prasad Das 3 | Chittaranjan Mishra 4 | Sapna Dhiman 5 1 Department of Physiotherapy, Composite Regional Centre for Persons with Disabilities, Lucknow, India 2 Neurosciences Centre, All India Institute of Medical Sciences, New Delhi, India 3 Department of Physical Medicine and Rehabilitation, Swami Vivekanand National Institute of Rehabilitation Training and Research, Cuttack, India 4 Department of Physiotherapy, Swami Vivekanand National Institute of Rehabilitation Training and Research, Cuttack, India 5 Department of Physiotherapy, Delhi Pharmaceutical Sciences and Research University, New Delhi, India Correspondence G. Shankar Ganesh, Lecturer in Physiotherapy, Composite Regional Centre for Persons with Disabilities, Lucknow, Uttar Pradesh, India. Email: shankarpt@rediffmail.com Abstract Objectives: The evidence for the effectiveness of interventions targeting acute low back pain (LBP) is suboptimal. It is difficult to identify those patients who are more likely to develop chronic pain and disability after an acute episode of LBP. These shortcomings may be attributed to considering LBP as one homogenous condition. Methods: In this quasi‐experimental study, we examined and analysed a prospec- tive cohort of 267 patients with first‐onset LBP and classified them into one of the groups based on treatment‐based classification: direction‐specified exercises (Group 2), manipulation (Group 3), stabilization exercises (Group 4), traction (Group 5), and a physician care group (Group 1). Disability and pain were assessed at baseline, after treatment, and at 6 months using the Oswestry Disability Index and the Numerical Rating Scale, respectively. Comparisons were made between the groups, and we predicted measures of disability and pain intensity at 6 months with age, gender, fear avoidance behaviour, centralization phenomenon (CP), expectations about recovery, CP, group classification, baseline pain, and disability. Results: Analysis showed that all the heterogeneous groups of LBP improved their outcomes with the respective treatment provided. However, when the entire sample was considered as one homogenous group of LBP, the results showed improvement with time (p < 0.05) only and no difference was found between groups (p > 0.05). None of the studied factors, except baseline pain (R = 0.227, R 2 = 0.051, p < 0.05), were able to accurately predict the development of chronic pain in our study sample. Conclusion: Though our results showed no differences between the subgroups in the reduction of pain and disability, we conclude that classifying and treating patients with LBP into subgroups based on signs and symptoms produce better outcomes. Baseline pain alone may predict a small percentage of people who may develop chronic pain. KEYWORDS clinical decision‐making, low back pain, manipulation, prognosis, stabilization, traction 1 | INTRODUCTION The management of low back pain (LBP) varies widely at the initial point of consultation, from guidelines recommending a delay in the referral to physical therapy for several weeks (Koes et al., 2010; Mur- phy, van Teijlingen, & Gobbi, 2006) to management including relative rest, activity modification, nonsteroidal anti‐inflammatory drugs, patient education, imaging, and referrals including physical therapy Received: 21 May 2018 Revised: 17 July 2018 Accepted: 15 August 2018 DOI: 10.1002/pri.1747 Physiother Res Int. 2018;e1747. https://doi.org/10.1002/pri.1747 © 2018 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pri 1 of 11