Background Identification of homogenous subgroups of patients with low back pain (LBP) and classification-based treatment have been recommended by some researchers and primary care clinicians. However, evidence regarding the effectiveness of this approach is not conclusive; one reason for this controversy appears to be the heterogeneity of trials in this context..Methods The aim of this study was to determine the effectiveness of classification-specific physical therapy in patients with LBP. The included trials were investigated in more homogeneous categories with respect to their classification scheme. Electronic databases including Medline, Cochrane, Ovid, Scopus, and PEDro were searched systematically for English-language randomized controlled trials (RCTs), published from 1980 to October 3, 2015. We included studies on LBP cases, which aimed to compare classification-specific physical therapies with non-specific treatments lacking patient classification. PEDro scoring was used to check the quality of the included trials, and the GRADE approach was used to evaluate the overall quality of evidence. Data on participants’ characteristics, sample size, and inclusion/exclusion criteria were extracted to obtain an overview of the included RCTs..Results A total of 12 RCTs were identified and categorized into four classification schemes. Some evidence supporting classification-specific treatment was found in each of the schemes. However, the reported evidence was conflicting predominantly due to differences in the study design. Also, GRADE quality assessment indicated the low quality of evidence for both approaches..Conclusions Categorization of trials based on their classification scheme to investigate the efficacy of classification-based physical therapy could reduce the heterogeneity of trials and allow researchers to understand the contradictory results in this context..