BackgroundThere are various methodological approaches to identifying clinically important subgroups and one method is to identify clusters of characteristics that differentiate people in cross-sectional and/or longitudinal data using Cluster Analysis (CA) or Latent Class Analysis (LCA). There is a scarcity of head-to-head comparisons that can inform the choice of which clustering method might be suitable for particular clinical datasets and research questions. Therefore, the aim of this study was to perform a head-to-head comparison of three commonly available methods (SPSS TwoStep CA, Latent Gold LCA and SNOB LCA).MethodsThe performance of these three methods was compared: (i) quantitatively using the number of subgroups detected, the classification probability of individuals into subgroups, the reproducibility of results, and (ii) qualitatively using subjective judgments about each program’s ease of use and interpretability of the presentation of results.We analysed five real datasets of varying complexity in a secondary analysis of data from other research projects. Three datasets contained only MRI findings (n = 2,060 to 20,810 vertebral disc levels), one dataset contained only pain intensity data collected for 52 weeks by text (SMS) messaging (n = 1,121 people), and the last dataset contained a range of clinical variables measured in low back pain patients (n = 543 people). Four artificial datasets (n = 1,000 each) containing subgroups of varying complexity were also analysed testing the ability of these clustering methods to detect subgroups and correctly classify individuals when subgroup membership was known.ResultsThe results from the real clinical datasets indicated that the number of subgroups detected varied, the certainty of classifying individuals into those subgroups varied, the findings had perfect reproducibility, some programs were easier to use and the interpretability of the presentation of their findings also varied. The results from the artificial datasets indicated that all three clustering methods showed a near-perfect ability to detect known subgroups and correctly classify individuals into those subgroups.ConclusionsOur subjective judgement was that Latent Gold offered the best balance of sensitivity to subgroups, ease of use and presentation of results with these datasets but we recognise that different clustering methods may suit other types of data and clinical research questions.

Purpose Modic changes (MCs) have been suggested to be a diagnostic subgroup of low back pain (LBP). However, the clinical implications of MCs remain unclear. For this reason, the aims of this study were to investigate how MCs developed over a 14-month period and if changes in the size and/or the pathological type of MCs were associated with changes in clinical symptoms in a cohort of patients with persistent LBP and MCs. Methods Information on LBP intensity and detailed information from MRI on the presence, type and size of MCs was collected at baseline and follow-up. Changes in type (Type I, II, III and mixed types) and size of MCs were quantified at both time points according to a standardised evaluation protocol. The associations between change in type, change in size and change in LBP intensity were calculated using odds ratios (ORs). Results Approximately 40 % of the MCs followed the expected developmental path from Type I (here Type I or I/II) to Type II (here Type II or II/III) or Type I to Type I/II. In general, the bigger the size of the MC at baseline, the more likely it was that it remained unchanged in size after 14 months. Patients who had MC Type I at both baseline and 14-month follow-up were less likely to experience an improvement in their LBP intensity as compared to patients who did not have Type I changes at both time points (OR 7.2, CI 1.3-37). There was no association between change in size of MCs Type I and change in LBP intensity. Conclusions The presence of MCs Type I at both baseline and follow-up is associated with a poor outcome in patients with persistent LBP and MCs.

To estimate the prevalence of degenerative lumbar spinal stenosis (LSS) in adults, identified by clinical symptoms and/or radiological criteria. Method Systematic review of the literature. Pooled prevalence estimates by care setting and clinical or radiological diagnostic criteria were calculated and plotted. [PROSPERO ID: CRD42018109640] Results In total, 41 papers reporting on 55 study samples were included. The overall risk-of-bias was considered high in two-thirds of the papers. The mean prevalence, based on a clinical diagnosis of LSS in the general population was 11% (95% CI: 4-18%), 25% (95% CI: 19-32%) in patients from primary care, 29% (95% CI: 22-36%) in patients from secondary care and 39% (95% CI: 39-39%) in patients from mixed primary and secondary care. Evaluating the presence of LSS based on radiological diagnosis, the pooled prevalence was 11% (95% CI: 5-18%) in the asymptomatic population, 38% (95% CI:-10-85%) in the general population, 15% (95% CI: 13-18%) in patients from primary care, 32% (95% CI: 22-41%) in patients from secondary care and 21% (95% CI: 16-26%) in a mixed population from primary and secondary care. Conclusions The mean prevalence estimates based on clinical diagnoses vary between 11% and 39% and the estimates based on radiological diagnoses similarly vary between 11% and 38%. The results are based on studies with high risk-of-bias and the pooled prevalence estimates should therefore be interpreted with caution. With an growing elderly population there is a need for future low risk-of-bias research clarifying clinical and radiological diagnostic criteria of lumbar spinal stenosis.