A randomised controlled equivalence trial to determine the effectiveness and cost-utility of manual chest physiotherapy techniques in the management of exacerbations of chronic obstructive pulmonary disease (MATREX)

To estimate the effect, if any, of manual chest physiotherapy (MCP) administered to patients hospitalised with chronic obstructive pulmonary disease (COPD) exacerbation on both disease-specific and generic health-related quality of life. To compare the health service costs for those receiving and not receiving MCP. A pragmatic, randomised controlled trial powered for equivalence. It was not possible to blind participants, clinicians or research staff to study arm allocation during the intervention. Four UK hospitals in Norwich, Great Yarmouth, King's Lynn and Liverpool. 526 participants aged 34-91 years were recruited between November 2005 and April 2008; of these, 372 provided evaluable data for the primary outcome. All persons hospitalised with COPD exacerbation and evidence of sputum production on examination were eligible for the trial providing there were no contraindications to performing MCP. Participants were allocated to either MCP or no MCP on an intention-to-treat (ITT) basis. However, active cycle of breathing techniques (ACBT) was used in both arms. Participants allocated to the intervention were guided to perform ACBT while the physiotherapist delivered MCP. Participants allocated to the control arm received instruction on ACBT only. The primary outcome was COPD-specific quality of life, measured using the St George's Respiratory Questionnaire (SGRQ) at 6 months post randomisation. The European Quality of Life-5 Dimensions (EQ-5D) questionnaire was used to calculate the quality-adjusted life-year (QALY) gain associated with MCP compared with no MCP. Secondary physiological outcome measures were also used. Of the 526 participants, 261 were allocated to MCP and 264 to control, with 186 participants evaluable in each arm. ITT analyses indicated no significant difference at 6 months post randomisation in total SGRQ score [adjusted effect size (no MCP - MCP) 0.03 (95% confidence interval, CI -0.14 to 0.19)], SGRQ symptom score [adjusted effect size 0.04 (95% CI -0.15 to 0.23)], SGRQ activity score [adjusted effect size -0.02 (95% CI -0.20 to 0.16)] or SGRQ impact score [adjusted effect size 0.02 (95% CI -0.15 to 0.18)]. The imputed ITT and per-protocol results were similar. No significant differences were observed in any of the outcome measures or subgroup analyses. Compared with no MCP, employing MCP was associated with a slight loss in quality of life (0.001 QALY loss) but lower health service costs (cost saving of 410.79 pounds). Based on these