Time, tenacity and trials: improving the quality of research

Medical research studies are increasing in number but evaluating efficacy, effectiveness, and public health impact of new interventions remains difficult because of inconsistent reporting in scientific journals. The CONSORT guidelines for reporting randomised trials were recommended over 15 years ago, ^{Reference Schulz1} and were adopted by many journals across disciplinary boundaries and specialties. ^{Reference Newcombe2-Reference Bennett5} The guidelines have clear implications for trial design and execution. ^{Reference Leitao and Goldsmith6} Further guidelines have emerged for the reporting of systematic reviews and observational studies. ^{Reference Brand7} These guidelines aim to improve transparency in research to inform decisions about clinical effectiveness and treatment recommendations. Have these guidelines had sufficient impact on the reporting and the accessibility of research to the wider public, as well as clinicians and researchers? It appears that the answer to this is no, at least not in psychiatry ^{Reference Han, Kwak, Marks, Pae, Wu and Bhatia8,Reference Motraghi, Seim, Meyer and Morissette9} nor for pharmacological ^{Reference Dasi, Navarro-Garcia, Jimenez-Heredia, Magraner, Vina and Pallardo10} and nonpharmacological trials in general, ^{Reference Nagendran, Harding, Teo, Camm, Maruthappu and McCulloch11} for trials of treatments for cancer ^{Reference Peron, Maillet, Gan, Chen and You12} nor for some trials in anaesthetics. ^{Reference Farr13} Psychiatric studies are especially challenging as the trials can include pharmacological and non-pharmacological interventions, devices, surgery, psychological treatments, complex multi-component interventions in hospital or community settings, public health and policy interventions in single or cluster randomised designs, and more. Are these multiple study designs a reason why the guidelines are not more widely used in mental health research? The CONSORT group have been busy generating evidence for effective ways of improving uptake of the guidelines, not least creating guidelines for parallel group designs, ^{Reference Moher, Hopewell, Schulz, Montori, Gotzsche and Devereaux14} social and psychological interventions, ^{Reference Grant, Mayo-Wilson, Melendez-Torres and Montgomery15} addictions research, ^{Reference Grant, Mayo-Wilson, Hopewell, Macdonald, Moher and Montgomery16} trials in child and adolescent mental health services, ^{Reference Gardner, Mayo-Wilson, Montgomery, Hopewell, Macdonald and Moher17} cluster randomised trials of those with cognitive impairment, ^{Reference DiazOrdaz, Slowther, Potter and Eldridge18} and pragmatic trials in general. ^{Reference Zwarenstein, Treweek, Gagnier, Altman, Tunis and Haynes19}

The number of trials in psychiatric and mental health research is increasing, as seen in the trends in the BJPsych; for example, Chien & Thompson (pp.) in this month’s Journal show that mindfulness-based group psychoeducation for patients with schizophrenia leads to improvement in symptoms, function, insight and readmissions profile at 2-year follow-up. Other designs are still needed in psychiatric and mental health research; for example, in this issue there are studies that seek to identify biomarkers to improve diagnosis (see Howes & Kapur, pp.; Pearlman et al, pp.; and Li et al, pp.), and to understand aetiology and developmental pathways to mental disorders in order to identify new interventions (Stringaris et al, pp.; Hung et al, pp.; Gumley et al, pp.; Muralidharan et al, pp.). Two important issues are raised by observational studies that show a higher mortality risk associated with antipsychotic use in people with cognitive impairment (see Gerhard et al, pp. and the linked editorial by Ballard et al, pp.) and in behavioural management in people with intellectual disabilities (see editorial by Glover et al, pp.). Prescribing without any scientific rationale is still common, so better evidence is needed about mechanisms by which psychiatric illness emerges and can be prevented and treated.

The skills, tenacity, temperament and time required to progress research along the pathway towards new and effective interventions always seem more than patients, researchers, clinicians, commissioners and policy makers would like. As a refreshing counterpoint to this proliferation of best evidence, Patterson et al (pp.) argue that including service users in research can lead to realisation of benefits. Indeed, the way research is assessed for impact is rapidly evolving, and patient involvement in both the design and execution of clinical research may well be the way we can be assured of more meaningful progress ^{Reference Greenhalgh20} while ensuring that the evidence-based agenda remains relevant to everyday experiences of patients and clinicians. ^{Reference Greenhalgh21}