What is stratified care and what clinical conditions have there been trials of stratified care for?

Stratified care is a model of healthcare where patients are subgrouped according to specific characteristics and then each subgroup is matched to treatment.

Subgroups can be based on causal mechanisms, prognostic characteristics, treatment responsiveness or a combination.

To date, most stratified care trials in musculoskeletal pain have focused on low back pain and use prognostic subgrouping.

Did these trials show that stratified care is more effective than usual care?

A series of studies from the UK that used the STarT Back approach to prognostic stratified care showed that a simple self-reporting tool could identify low back pain patients’ risk of persistent disabling pain and categorise patients as being at low, medium or high risk of a poor outcome.

They also showed that matched treatments for each patient subgroup could be agreed on by clinicians.

The STarT Back randomised trial demonstrated that when subgrouping and matched treatments are combined and delivered with high fidelity, clinical, work and cost outcomes are superior (published in The Lancet in 2011).

It makes sense logically that stratified care should be more effective. Do you think we have given it a fair test with these trials?

The STarT Back program has inspired many similar trials around the world, including in Denmark, the USA, the Netherlands and Australia.

These trials are challenging and one key challenge has been intervention fidelity.

For example, in several trials few patients have been subgrouped systematically and/or patients’ clinical care has not changed to the matched treatment intended.

Professor Nadine Foster.

It is then tricky to decide whether inadequate intervention fidelity explains the lack of differences between stratified care and the comparison in these trials.

What do we need to do better?

More careful staging of research is needed. While this can be challenging due to funding constraints, it seems that several large pragmatic randomised trials have been conducted without first doing key initial work on intervention fidelity, feasibility or pilot testing.

An increase in the use of internal pilot phases within large multicentre randomised controlled trials could help, along with conducting external feasibility and pilot trials where intervention fidelity could be specified as a key criterion for progression to a larger trial.

Have the existing trials of stratified care investigated the fidelity of their stratified care intervention (subgrouping and matching treatment)?

Some have, yes.

Examples include the USA-based MATCH trial, which reported that only about half of patients were subgrouped as intended, and the TARGET trial, in which only 36 per cent of participants in the stratified care arm were referred for matched treatment.

The Dutch OCTOPuS trial in knee osteoarthritis reported frequent protocol violations in the stratified care arm.

Do you think the problem could be that usual physiotherapy care is already quite tailored to individual characteristics?

Most physiotherapists would argue that the care they provide for patients is already highly individualised.

The challenge is that the same patient, consulting a range of different clinicians, may receive different tailored treatments.

Stratified care is about matching subgroups of patients to treatments and it sits in the middle of a spectrum between a ‘one-size-fits-all’ approach and a personalised medicine approach.

We have not made much progress in researching truly individualised treatments for low back pain.

This may improve as we continue to identify and harness multiple biomarkers or prognostic variables in very large datasets and to use machine learning and other tools to identify more nuanced subgroups.

So what should future trials of stratified care do?

Learn from the trials completed to date and spend more time in the early stages of specifying how the components of stratified care will be put into operation and delivered with high fidelity, followed by feasibility testing in the new setting or context.

Specify progression criteria that include intervention fidelity prior to conducting the (expensive) large multicentre randomised controlled trials.

>> Professor Nadine Foster is a physiotherapist and a NHMRC investigator whose research focuses on musculoskeletal pain, including low back pain, and on developing, testing and implementing treatments and models of care. She is Director of the Surgical, Treatment and Rehabilitation Service Education and Research Alliance between the University of Queensland and Metro North Health and Academic Director of the University of Queensland’s Clinical Trials Centre.