What the STarT Back Risk Groups Actually Predict

Risk of what, over what timescale

The thing being forecast is persistent, disabling back pain: pain still getting in the way of ordinary life long after most episodes have quietened down 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. It is not the risk of anything dangerous, and the tool contains nothing that would detect a fracture, an infection or a tumour.

The timescale matters, because the word persistent does a lot of unexamined work. Back pain is conventionally described as acute when it lasts days to weeks, and chronic once it runs beyond twelve weeks 2Ref 2National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) (2023).Back Pain (Symptoms, Types & Causes).The conventional timescale that gives the word persistent its meaning: back pain described as acute when it lasts days to weeks, and as chronic once it continues beyond twelve weeks.. The risk groups point at that second territory — whether you are still dealing with this in three months, six months, a year.

The groups estimate a trajectory, not a severity. They answer how is this likely to go, not how bad is it and not what is wrong.

That is why the labels get misread. Handed a form with the word risk on it, most people assume risk of something bad happening to their spine; the instrument means something narrower. It has nothing to say about whether this is muscle or disc pain, and the start back screening tool was never built to look.

What a low-risk result means

Low risk is the group for a total of three or fewer out of nine, and the second score is never consulted to get there 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. It means the questionnaire found few of the features associated with an episode dragging on: little referred leg pain, little pain elsewhere, limited reported disability, limited distress about it.

What it is not is a guarantee. A forecast built from nine questions answered on one day has an error rate, and some people in the low-risk group do go on to have a long and difficult time. If your pain has not behaved the way the label implied, the label was wrong about you — a normal thing for a screening instrument to do.

It is also not a clearance certificate. A low-risk result says nothing about whether your pain has a serious cause — a completely separate question, answered by clinical assessment rather than by any score. The features that prompt concern about a serious cause are the low back red flags, and they appear nowhere on this form.

In practice a low-risk result means less care rather than none: an explanation, encouragement to keep moving, and a plan for what to do if things do not follow the expected course.

What a medium-risk result means

Medium risk is the group for a total of four or more out of nine, where three or fewer of those points came from the five questions covering bothersomeness, catastrophic thinking, fear of movement, anxiety and low mood 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. The score is elevated, and most of what raised it is the physical picture: pain travelling into the leg, pain elsewhere, difficulty doing things.

This is the group people find hardest to interpret, because medium sounds like a midpoint on a single dial and it is not. Medium and high are separated by where the points came from, not by how high the total went — someone in the medium group can have a higher total than someone in the high group.

What the label is saying is that this episode has enough going on to warrant more than reassurance, without the distress pattern that predicts the slowest recoveries. In the design of the tool, medium is the group routed toward structured physical rehabilitation.

Landing in the middle is not a statement that your case is unremarkable; it says the obstacle in front of you looks primarily physical.

What a high-risk result means

High risk requires two things at once: a total of four or more out of nine, and four or more of the five points available on the distress questions 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. Both conditions have to be met, which is why the high-risk group is defined by the character of the answers rather than by their quantity.

Those five ask how bothersome the pain has been, whether the person catastrophizes about it, whether they fear movement, and whether they are anxious or low in mood 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. A high-risk label means most of the elevated score came from that cluster.

This is a statement about a statistical pattern, not about your character, your honesty, or whether your pain is real. The pattern was identified because it predicts, not because anyone decided that distressed people deserve a label. Pain that is severe, frightening and unexplained is a perfectly rational thing to be distressed about, and the tool makes no claim about which came first.

What the label is meant to trigger is care that addresses both halves of the problem — the physical obstacles and the psychological ones — rather than physical treatment alone 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. The obstacles the questionnaire detected will still be there when the exercises finish.

Why the highest-risk group is not the worst-pain group

Because pain intensity is barely in the arithmetic. Of the nine items, four cover the physical picture and five the psychological one, and it is the second set that decides the top category 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. Someone in severe, well-localised pain with no fear about it and no low mood will not reach the high-risk group, however much it hurts.

The high-risk group is the group with the most obstacles to recovery, not the group with the most pain.

The reverse case is the one that matters: a person whose pain is moderate but whose fear of movement is total can land in the high-risk group. Nothing in how their back behaves would single them out, and a clinician sorting on pain alone would route them to the lightest pathway.

What each group is routed toward, and how good the evidence is

The point of sorting is to match the intensity of care to the likely obstacle: reassurance and advice at the low end, structured rehabilitation in the middle, combined physical and psychological treatment at the top 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument.. What sits inside those pathways is worth seeing honestly, because the effects are real and moderate rather than dramatic.

• Physical therapy. The clinical practice guideline for acute and chronic low back pain recommends exercise, manual therapy and manipulation, and patient education, each graded by the strength of its evidence 3Ref 3George SZ, Fritz JM, Silfies SP, et al. (2021).Interventions for the Management of Acute and Chronic Low Back Pain: Revision 2021 (Clinical Practice Guidelines).That the physical-therapy clinical practice guideline for acute and chronic low back pain recommends exercise, manual therapy and manipulation, and patient education, each graded by the strength of the supporting evidence — the content of the rehabilitation pathways the medium- and high-risk groups are routed toward..
• Spinal manipulation. A systematic review and meta-analysis found effects on pain and function similar to other recommended therapies for chronic low back pain, with generally minor and transient adverse events 4Ref 4Rubinstein SM, de Zoete A, van Middelkoop M, et al. (2019).Benefits and harms of spinal manipulative therapy for the treatment of chronic low back pain: systematic review and meta-analysis of randomised controlled trials.That spinal manipulative therapy produces effects on pain and function similar to other recommended therapies for chronic low back pain, with generally minor and transient adverse events..
• Anti-inflammatory medication. A Cochrane review found non-steroidal anti-inflammatory drugs slightly better than placebo for short-term pain and disability in chronic low back pain, with an effect that was small and may not be clinically important 5Ref 5Enthoven WTM, Roelofs PDDM, Deyo RA, van Tulder MW, Koes BW (2016).Non-steroidal anti-inflammatory drugs for chronic low back pain.That non-steroidal anti-inflammatory drugs are slightly more effective than placebo for short-term pain and disability in chronic low back pain, but that the effect is small and may not be clinically important..

Spinal manipulation, in meta-analysis, performed about as well as the other recommended therapies for chronic low back pain — not better 4Ref 4Rubinstein SM, de Zoete A, van Middelkoop M, et al. (2019).Benefits and harms of spinal manipulative therapy for the treatment of chronic low back pain: systematic review and meta-analysis of randomised controlled trials.That spinal manipulative therapy produces effects on pain and function similar to other recommended therapies for chronic low back pain, with generally minor and transient adverse events..

Read together, these say something the label alone does not: no pathway contains a decisive treatment. Stratification buys a better-aimed intervention, not a stronger one.

What the label does not authorise

A risk group is not an indication for imaging. Scanning low back pain in the first six weeks does not improve outcomes and adds cost, and imaging is reserved for cases with red flags such as a progressive neurological deficit or a suspected serious underlying condition 6Ref 6American Academy of Family Physicians (Choosing Wisely) (2021).Don't do imaging for low back pain within the first six weeks, unless red flags are present.That imaging for low back pain within the first six weeks does not improve outcomes and increases cost, and should be reserved for cases with red flags such as a progressive neurologic deficit or a suspected serious underlying condition — establishing that a prognostic risk label is not an indication for a scan.. A high-risk result is not one of those triggers; the tool is not looking for serious pathology at all.

The instinct runs the other way: being told you are high risk for anything makes a scan feel overdue. But the mri for back pain question is decided on the history, the examination and the neurological findings, and a prognostic label contributes nothing to it.

Nor does the label authorise anyone to treat your account of your pain as inflated. If a clinician's response to a high-risk result feels dismissive, that is a misuse of the tool rather than its intended output, and it is reasonable to say so.

And it does not settle what your low back pain is. The tool was built for ordinary nonspecific low back pain — no specific structural culprit identified — and assumes that question was handled by someone who examined you.

Can your group change?

Yes, and that is among the more useful things to understand about it. Several of the nine items describe how you feel about the pain right now, and those move — with sleep, with a stressful week, with a frightening remark about your spine, with an explanation that finally makes sense. Two administrations a month apart can land in different groups.

That mobility describes what the tool measures rather than a flaw in it. But it does limit what re-scoring can tell you, and the developers are explicit: the tool reports no minimal clinically important difference and no minimal detectable change, because it is a screener rather than an outcome measure 1Ref 1Hill JC, Dunn KM, Lewis M, et al. (2008).A primary care back pain screening tool: identifying patient subgroups for initial treatment.The definition of the three risk subgroups and what they predict: a nine-item tool scored 0-9 with a five-item psychosocial subscale scored 0-5; low risk at a total of 0-3, medium risk at a total of 4 or more with a subscale of 3 or less, high risk at a total of 4 or more with a subscale of 4 or more; higher scores indicating greater risk of persistent disabling low back pain; the tool being prognostic rather than diagnostic or a severity measure; the item content including referred leg pain, comorbid pain, two disability items, bothersomeness, catastrophizing, fear, anxiety and depression; the intended matching of care intensity to subgroup; and the explicit absence of any MCID or MDC for this instrument..

A falling STarT Back score is not a validated measure of progress, and was never meant to be used as one.

Tracking recovery is done with something else — a disability or function questionnaire built for that job. If a clinic re-administers this one and your group has moved, the honest reading is that the sorting decision may deserve revisiting, not that a number fell by a meaningful amount.