By Chris Salisbury, University of Bristol, on behalf of the 3D trial team.
.
Full report of the 3D study helps us interpret the findings
.
The 3D approach was designed to improve care for patients with multimorbidity. Based on a patient-centred care approach, it promotes continuity of care, offers a comprehensive holistic review which focuses on problems that most matter to the patient, and seeks to reduce inappropriate polypharmacy. The reviews involve a six-monthly multi-disciplinary review from a nurse, pharmacist and GP, leading to a health plan with specific goals agreed between the patient and GP. The 3D model incorporates most of the strategies recommended in international guidelines on multimorbidity.
We conducted a large cluster randomised controlled trial comparing the 3D approach and usual care, and the main trial results were published in the Lancet in July 2018 [1]. We found that the 3D intervention was effective at improving patient centred care, but did not result in improvements in patient’s quality of life, health outcomes or polypharmacy.
How should we make sense of these counter-intuitive results? Are the current international guidelines misconceived? Perhaps the 3D approach was the wrong intervention, perhaps it was not effectively implemented or not provided for the long enough to make a difference. Or maybe we chose the wrong outcome measures. Interestingly, the conclusion that the 3D approaches improves patient-centred care but not quality of life is consisent with most previous trials of interventions for multimorbidity.
The Lancet paper has generally been interpreted as reporting a negative trial. The full report of the 3D study has now been published [2], and provides a more rounded perspective on the findings. It includes a process evaluation based on interviews with patients and staff, along with direct observation in case-study practices to help us understand how the 3D approach was implemented and how it might be improved. The full report also includes an economic evaluation of cost-effectiveness.
Through the process evaluation we found that practices were strongly supportive of the principles underlying the 3D approach, but they found implementing it logistically difficult. Many patients in the trial did not receive the full ‘dose’ of the intervention. Only half of the patients received two 3D reviews over 15 months as intended, while three-quarters received at least one review. This incomplete implementation was related to the pressures that general practices in the UK currently face, which made introducing any kind of change very difficult. Trying to do so within the context of a trial made it even more difficult. Introducing a new way of working for a limited period for a sub-set of patients, within practices which had well-established systems for offering single-disease care designed to meet the requirements of the Quality and Outcomes Framework, meant that not everything worked as planned. For example, some practices offered 3D reviews as well as, rather than instead of, single-disease reviews. However, practices did identify ways in which the 3D model might be improved, for example by more selectively targeting patients with the most complex problems, more training for staff and tailoring the frequency of reviews according to patients’ needs.
The economic evaluation showed that the 3D intervention was associated with small improvements in quality-adjusted life years (QALYs) along with small increases in NHS costs. The cost per QALY was £18,499, just below the threshold of £20,000 commonly used to justify new interventions in the NHS. Therefore the economic case for introducing 3D is arguable, and could be justified given that it provided care in a way that patient’s preferred and which they felt met their needs.
In summary, the report describes the advantages and limitations of the 3D approach, and ways in which it might be improved. There doesn’t appear to be a simple magic bullet to improve care for multimorbidity and no model of care has yet been convincingly shown to be effective in randomised trials. Paradoxically, one key finding from the report is that the 3D approach would probably need to become normal practice and offered over several years before the benefits became apparent, but testing this hypothesis in an affordable randomised trial would almost certainly be impossible.
.
This project was funded by the National Institute for Health Research Health Services and Delivery Research Programme (project number 12/130/15). The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the HS&DR Programme, NIHR, NHS or the Department of Health.
.
One comment
Reply to Chris Salisbury
Chris’s summary giving a fuller understanding of the 3D study is really helpful and chimes in so many ways. These include the ‘dose’ of the intervention being only partly administered – something that was very clear in our recently published (observational) evaluation of the South London Integrated Care Pilot (https://bmjopen.bmj.com/content/9/3/e024220) where it took them more than twice as long to get some aspects of the intervention going as they’d planned. I’ve also always been very struck by the Massachusetts General’s evaluation of their risk profiling/integrated care intervention. When they tried it in the first district it increased costs for the first year, stabilised by 18 months then started to show benefits in terms of cost reduction by 24 months. So they then thought “We know how to do it now, so when we roll it out to other districts, we’ll be able to get the benefits earlier”. Not a bit of it – exactly the same happened when they rolled it out more widely. So Chris makes a really important point that we’re often unable to give these complex interventions which may require cultural change enough time to develop fully – either because they’re in the context of a trial or (in the UK) because policy-makers come up with another ‘bright idea’ in the meantime.
Professor Martin Roland
Emeritus Professor of Health Services Research University of Cambridge