Clinical library · Clinical reasoning
Screening for Medical Referral
Physiotherapists seeing patients without a doctor first is well studied, and the results are consistently favourable. The caveat that matters most is buried in the methods: across 13 studies of direct access, only one actually assessed clinical safety.
In one line. Screening for medical referral is the process of deciding that a patient's presentation is not primarily a physiotherapy problem, or is not only one, and routing them accordingly.
It becomes a defining competence wherever patients reach a physiotherapist without a doctor having filtered them first. The evidence below is about whether physiotherapists do this well. It is largely reassuring, and it is thinner on the one question — harm — that reassurance would need to rest on.
Scope of practice is jurisdictional
The studies here come from health systems where direct access and, in some cases, imaging referral are legally defined. What a physiotherapist may do on first contact differs by country and by state, and this page makes no claim about what is permitted in any particular jurisdiction. Check your own regulator. The finding that clinicians are frequently unclear about their own scope [5] is a reason to check rather than assume.
What the evidence shows
| Question | Finding | Source and quality |
|---|---|---|
| Is direct access safe? | No adverse events were noted in any of the included studies [1] and no harms were reported [2] | But the second review states explicitly that only one study assessed the clinical safety of direct access. Absence of reported harm is not the same as measured safety [2] |
| Do referrals go to the right place? | Direct access showed high referral accuracy and a reduction in the rate of return visits | 28 articles. Risk of bias intermediate for the randomised trials; observational studies averaged 6 of 9 on the Newcastle-Ottawa Scale [1] |
| Are health outcomes worse? | No difference between models on health outcomes [1]; no difference in patient-reported outcomes in two randomised trials of physiotherapist-led orthopaedic triage [3] | Equivalence, not superiority, is the correct reading [1][3] |
| What changes? | The medical model had higher use of imaging, drugs and referral to another specialist [1]. Direct access involved fewer physiotherapy treatments, fewer physician visits, less imaging, fewer non-steroidal anti-inflammatories and less secondary care [2] | Two independent reviews reaching the same pattern, one of 28 and one of 13 studies [1][2] |
| Do physiotherapists and surgeons agree? | High agreement on diagnosis (9 cohort studies), on referral for investigation (5 cohort studies) and on treatment approach (8 cohort studies) | 2 randomised trials and 11 cohort studies, 1,357 participants; certainty of evidence low to moderate; methodological quality 14 to 23 of a possible 28 [3] |
| How good is the agreement, numerically? | In a paediatric orthopaedic triage clinic, mean raw agreement 82% (95% CI 73 to 87), with 12 of 15 individual codes at 90% or above (range 57 to 100). 87% of onward referrals were judged appropriate by the consultants; surgical conversion rate 23% | Retrospective review, two advanced practice physiotherapists and four consultants at a single site [4] |
| Does triage change the surgical list? | Physiotherapist-led triage had higher conversion-to-surgery rates — 55% to 91% versus 22% to 38% — with shorter waiting times and lower costs | One randomised trial and three cohort studies for the conversion figure. Three of the review's authors had authored one of the included trials, and state that the quality appraisal of those papers was performed by the other two [3] |
| Do physiotherapists know their own scope? | Only 42.0% of those practising where imaging referral was permitted were aware of the privilege. Awareness was 71.4% where it came via legislation and 25.2% where it came via the state board | US survey, 2020 to 2021. Those aware were far more likely to use it (44.5% versus 3.2%). The author declares payment for teaching musculoskeletal imaging courses [5] |
| Who practises direct access most? | Doctors of Physical Therapy, residency or fellowship-trained, and board-certified physiotherapists all reported greater frequency | Self-reported and cross-sectional, so this is an association with training, not evidence that training causes competence [5] |
Where it misleads
1. "No adverse events" is the weakest form of a safety claim
Both reviews report it, and both are honest about what it rests on: one states directly that only a single study assessed the clinical safety of direct access. [2] Adverse events in this setting are rare, delayed, and occur after the patient has left the study — a missed malignancy does not present as an adverse event in a physiotherapy cohort. The reviews report an absence of evidence of harm. That is worth having, and it is not the same as evidence of absence.
2. The patients who self-refer are not the patients who worry you
Direct access patients were younger, more educated, presented with less severe clinical conditions and more acute spine-related complaints. [2] Safety demonstrated in that population does not transfer automatically to an older, more comorbid, less health-literate one. This is selection, not bias in the studies — but it constrains what the result means.
3. Agreement with a consultant is not accuracy
The agreement studies compare the physiotherapist's diagnosis to the orthopaedic surgeon's diagnosis. [3][4] The surgeon is the reference standard, not the truth. Where both would miss the same non-musculoskeletal cause, agreement is perfect and both are wrong. High agreement establishes that the physiotherapist reaches the same conclusion as the specialist, which is what triage needs — it does not establish that either conclusion is right.
4. A higher surgery conversion rate is an efficiency result, not a health result
Conversion rose from 22% to 38% under standard care to 55% to 91% under physiotherapist-led triage. [3] That means fewer people saw a surgeon unnecessarily. It does not mean patients ended up better: patient-reported outcomes were no different in the randomised trials. [3] The same review's own conclusion is that health outcomes are equivalent.
5. Screening competence is uneven and the profession does not know its own boundaries
Fewer than half of physiotherapists practising in jurisdictions permitting imaging referral knew they could do it, and the figure fell to a quarter where the authority came from a board rather than legislation. [5] Screening for referral requires knowing both what to look for and what you are permitted to do about it. The second half is apparently not reliable.
6. Single-site agreement figures are not profession-wide competence
The paediatric study involved two advanced practice physiotherapists and four consultants at one hospital, retrospectively. [4] Its mean agreement of 82% is genuine and its range across individual diagnostic codes was 57% to 100% — so agreement was poor for some diagnoses even in a well-performing clinic. Reporting the mean without the range overstates the consistency.
What the evidence supports — and what it does not
Supported
- Equivalent health outcomes under direct access and physiotherapist-led triage compared with the medical model. [1][3]
- Less imaging, less medication and fewer onward specialist referrals. [1][2]
- Lower cost and shorter waiting times. [1][2][3]
- High diagnostic and triage agreement with orthopaedic specialists, including in paediatrics. [3][4]
- Higher patient satisfaction and perceived quality of care. [1][2][3]
Not supported
- "Direct access is proven safe." Only one study assessed clinical safety directly. [2]
- Generalising to older or more complex populations. Self-referring patients were younger and less severe. [2]
- Treating agreement with a surgeon as diagnostic accuracy. [3][4]
- Claiming better clinical outcomes. The finding is equivalence. [1][3]
- Assuming clinicians know their own scope of practice. [5]
- Any statement about what is permitted in a given country on the basis of this literature.
How certain is this?
Evidence grade: Low.
The grade is set by the weakest link, which is safety — the claim most often made from this literature. One study across 13 assessed clinical safety directly. [2] Everything else about harm is an absence of reports rather than a measurement, and the outcome in question is rare and delayed enough that the studies were unlikely to capture it.
The effectiveness and agreement evidence is better. The triage review is GRADE-assessed at low to moderate certainty across 2 randomised trials and 11 cohort studies with 1,357 participants, and reports its own limitations plainly, including that three of its authors wrote one of the included trials. [3] The two direct access reviews are independent, reach the same pattern, and are honest about intermediate risk of bias and Newcastle-Ottawa scores averaging 6 to 6.4 of 9. [1][2]
The agreement studies are small and single-site, [4] and the scope-awareness survey is cross-sectional and self-reported with a declared commercial interest. [5]
What would change the grade: prospective safety studies with follow-up long enough and linkage complete enough to detect delayed diagnosis, in unselected populations rather than self-referring ones.
Common questions
Is it safe for patients to see a physiotherapist first?
The evidence points that way and is thinner than it looks. Across 28 studies no adverse events were noted, [1] and across 13 studies no harms were reported — but that second review states only one study assessed clinical safety. [2] Patients who self-refer are also younger and less severely affected. [2] The honest statement is that no harm has been detected in the populations studied, which is not the same as safety having been established.
Do physiotherapists reach the same diagnosis as specialists?
Usually. High agreement with orthopaedic surgeons on diagnosis, referral for investigation and treatment approach across nine, five and eight cohort studies respectively. [3] In a paediatric clinic, mean raw agreement was 82% with 12 of 15 diagnostic codes at 90% or above — but the range across codes ran from 57% to 100%, so agreement was weak for some diagnoses. [4]
Does direct access give patients better results?
No — it gives equivalent results more efficiently. Health outcomes showed no difference between models, [1] and patient-reported outcomes were no different in two randomised trials. [3] What changed was less imaging, fewer drugs, fewer onward referrals, shorter waits, lower cost and higher satisfaction. [1][2][3] That is a system argument, and it is a good one; it is not a claim of superior recovery.
What should trigger a referral back to a doctor?
Features raising the probability of serious pathology, a presentation that does not fit a musculoskeletal pattern, or failure to respond as expected. The specific checklist items are weaker than they are usually taught — see red flags for which few carry real diagnostic weight and why guidelines disagree with each other.
Can a physiotherapist order imaging?
It depends entirely on jurisdiction, and this page does not assert what is permitted anywhere in particular. Note that in one US survey only 42.0% of physiotherapists practising where the privilege existed were aware of it — 71.4% where it came from legislation, 25.2% where it came from a board. [5] Verify with your own regulator rather than with colleagues. See also imaging in low back pain for when it is worth requesting at all.
Does more training make a difference?
The survey found Doctors of Physical Therapy, residency or fellowship-trained and board-certified physiotherapists all reported practising direct access more frequently, and direct access frequency was positively associated with imaging skill performance and imaging referral practice. [5] This is cross-sectional and self-reported: it shows who does it, not that the training produced the competence.
References
- Gallotti M, Campagnola B, Cocchieri A, et al. Effectiveness and Consequences of Direct Access in Physiotherapy: A Systematic Review. Journal of Clinical Medicine. 2023 Sep 7;12(18):. doi:10.3390/jcm12185832 PMID 37762773 Systematic review
- Piscitelli D, Furmanek MP, Meroni R, et al. Direct access in physical therapy: a systematic review. Clinica Terapeutica. 2018 Sep-Oct;169(5):e249–e260. doi:10.7417/CT.2018.2087 PMID 30393813 Systematic review
- Samsson KS, Grimmer K, Larsson MEH, et al. Effects on health and process outcomes of physiotherapist-led orthopaedic triage for patients with musculoskeletal disorders: a systematic review of comparative studies. BMC Musculoskeletal Disorders. 2020 Oct 10;21(1):673. doi:10.1186/s12891-020-03673-9 PMID 33038935 Systematic review (GRADE)
- Ó Mír M, O'Sullivan C, Lennon O, et al. An evaluation of diagnostic agreement rates between advanced practice physiotherapists and paediatric orthopaedic consultants for children with musculoskeletal complaints. Musculoskeletal Care. 2018 Dec;16(4):433–439. doi:10.1002/msc.1357 PMID 30109917 Retrospective agreement study
- Mabry LM, Keil A, Young BA, et al. Physical therapist awareness of diagnostic imaging referral jurisdictional scope of practice: an observational study. Journal of Manual & Manipulative Therapy. 2024 Aug;32(4):435–445. doi:10.1080/10669817.2023.2296260 PMID 38130076 Cross-sectional survey
About this resource
Using this in clinic
Every figure here is traceable to its source.
The favourable headline on direct access rests on a safety literature of one study, and this page says so rather than repeating the headline. Where a value could not be verified against the paper it came from, it is not on this page, and the omission is stated rather than filled with a number from a secondary source.