Knee orthopaedic referral criteria.
When does a patient with knee pain need an orthopaedic opinion?
A patient with knee pain needs an orthopaedic opinion in four cases. The first is a problem an operation can fix, in a patient who wants it fixed. The second is a knee that is mechanically obstructed (blocked). The third is a structural injury that is repairable and time-critical. The fourth is when non-surgical management has genuinely been completed, and the patient's function remains substantially impaired. For everything else, the question is what to do first, not whether to refer.
National guidance is more specific than most local pathways suggest. The National Institute for Health and Care Excellence (NICE) recommends considering referral for joint replacement when two things are true. Joint symptoms are substantially impacting quality of life. And non-surgical management has been ineffective or unsuitable. NICE also says explicitly that the referral decision should be made on clinical assessment, not on a numerical severity score.
The harder question is the one guidance does not answer, and it is the reason this page exists. What counts as non-surgical management having been completed? And what does not, by itself, justify a referral at all?
Written for the clinician seeing the patient first
This page is written for general practitioners (family doctors), musculoskeletal (MSK) first contact practitioners, MSK physiotherapists and advanced practitioners. They are the clinicians who actually make knee referral decisions, and who rarely get to see how those decisions turn out.
It is not written for patients, and it is not a substitute for local pathways or for clinical judgement. Where your service has agreed criteria, those govern.
The questions that come before the referral question
Serious pathology (disease or injury) in a knee is uncommon. It does not go through an elective (planned) referral or an advice service. It goes through the local urgent or emergency pathway, on the day. The categories worth holding in mind at first contact are these.
Same day
A hot, swollen, acutely painful knee with restricted movement, fever or systemic upset (feeling generally unwell). Any new pain or effusion (fluid swelling) in a knee that has a prosthesis (joint replacement) in it. Significant trauma (injury) with deformity, or a knee dislocation even if it has already reduced (gone back into place). A genuinely locked knee that cannot be fully straightened goes to A&E (the emergency department), for the on-call orthopaedic team.
Urgent, not routine
Inability to lift the straight leg off the bed. A gap you can feel in the extensor mechanism (the muscles and tendons that straighten the knee). A calf or knee presentation that could be a deep vein thrombosis (a blood clot) rather than a joint problem.
Different pathway entirely
Unexplained night or rest pain, unexplained weight loss or a history of cancer. An inflammatory pattern: early morning stiffness over thirty minutes, other joints involved, psoriasis (a skin condition) or inflammatory bowel disease. And pain from the hip or lumbar spine (lower back) that is felt in the knee (referred pain).
How to recognise serious musculoskeletal pathology is set out nationally by the Arthritis and Musculoskeletal Alliance (ARMA). Its guidance on urgent and emergency musculoskeletal conditions requiring onward referral is the document to work to, rather than this page. Guidance from NHS England (National Health Service England) is explicit that serious pathology should not be routed through specialist advice services at all.
What does not, by itself, justify a referral
Every day, four things shape knee referral decisions. National guidance says none of them should decide a referral on its own. Naming them is not a criticism of referrers. It is a criticism of a system in which nobody tells you what happened to the last one.
The X-ray
Radiographic severity (how bad the X-ray looks) is not the referral trigger. A patient coping well with a poor-looking X-ray does not need an operation because of the picture. Delay in that patient costs very little. NICE recommends deciding on clinical assessment rather than on numerical severity scoring. The same logic applies to grading an X-ray.
A tear on a magnetic resonance imaging (MRI) report
A degenerate (wear-related) tear of the meniscus, the knee's cartilage cushion, is a common finding. It is not a reason for surgery. The British Association for Surgery of the Knee has a national guideline on this. It says the large majority of meniscal tears are treated non-operatively first. It also says arthroscopic meniscectomy (keyhole surgery to trim the meniscus) is not a first-line treatment. The trigger is mechanical obstruction, a repairable traumatic tear, or a root tear in a knee without established arthritis. It is not the word tear in a report.
Age, weight, smoking or other health conditions
NICE is clear on this. People with osteoarthritis should not be excluded from referral for joint replacement because of their age, sex or smoking. Nor should other health conditions (comorbidities), or being overweight or obese, exclude them. These may affect the outcome of surgery. They are not obstacles to referral. Using them as a filter blocks access to care, and it is not a clinical reason.
A score
NICE recommends using clinical assessment when deciding to refer someone for joint replacement, rather than systems that numerically score severity of disease. Oxford Knee Scores and similar instruments are useful for measuring change. They were not recommended as the basis for deciding who is referred.
What non-surgical management having been completed actually means
Referral criteria almost universally require that non-surgical management has been tried. Almost none of them say what that means. In practice, the phrase can mean twelve weeks of supervised strengthening. It can also mean a photocopied exercise sheet handed over eighteen months ago. Those are not the same answer, and a surgeon reading the referral cannot tell them apart.
Three things make the difference clear to whoever reads the referral next. How many weeks of exercise, and whether it was supervised or done on their own. What pain relief was tried, and what it did. And if there was an injection, not just that one was given, but how long the response lasted. That is the detail that tells patients apart, and it is almost never recorded.
Therapeutic exercise is core first-line treatment for osteoarthritis under NICE, alongside weight management where relevant. Nobody has failed non-surgical management until they have actually had it.
What to request before you refer
For suspected osteoarthritis in a patient over forty-five, the useful test is a weight-bearing (standing) X-ray, not an MRI. An MRI in that patient will find degenerate change (wear and tear). Finding it changes neither the diagnosis nor the management.
Where the images are held matters as much as whether they were done. Image transfer between organisations is named repeatedly as a reason a decision cannot be made at the first appointment. So say where the images are held, and attach the report.
Would this patient consider surgery if it were offered?
Ask it, at first contact, every time. First, say plainly that most knee problems do not need an operation. Say too that the answer changes nothing about what is offered today.
It is never a reason to refuse or delay a referral, and it must never be used as one. It changes the order of the plan, not the patient's access. A patient who says no in March may say yes in October. So the answer is recorded with its date, and the question is asked again. NHS England reports that in 2019 to 2020, almost a third of patients were discharged after their first orthopaedic appointment. A great many referrals are already requests for advice. Very few say so.
What a knee referral needs to contain
Apart from knee replacement, there is no national standard for what a knee referral must contain. In July 2026, Getting It Right First Time (GIRFT) set out what a knee replacement referral should show, in its knee replacement pathway. Local pathways ask referrers to show that core treatment was offered, and stop there. That is why referrals are rejected for missing information that was never specified in the first place.
This page is written by Dr Arj Imbuldeniya, a consultant orthopaedic knee, hip and trauma surgeon. In his view, a knee referral becomes decidable when it carries five things. First, the safety screen, answered rather than left blank. Second, the presentation, including how any injury happened and whether there is true locking, catching or giving way. Third, what the patient cannot do now and what they want back, in their own words. Fourth, what has been tried and for how long, including how long any injection helped. Fifth, the imaging, with the date and where the images are held. Add one line saying what you actually want. That might be advice, a diagnostic opinion, a surgical assessment or a second opinion for the patient.
None of that requires a new examination. All of it is already known to the person writing the referral.
Limits, stated plainly
This page covers the adult knee at first contact, for elective care. It does not cover trauma, children, the painful or infected joint replacement, or inflammatory arthritis. Nor does it cover spine or hip problems that show up as knee pain. Each of these follows a different route.
It is general information for health professionals about referral pathways. It is not clinical advice, it is not a protocol, and where your local pathway differs, your local pathway governs.
Where this sits in the pathway
The wider policy context is on MSK referral optimisation. It explains why a referral decision can be wrong in two directions. It also explains why the number of referrals is a poor measure of their quality. The category of software that supports this decision is explained on what MSK clinical decision support is. Where these decisions are made in the NHS is covered on MSK triage explained and first contact practitioners and MSK roles.
Where this came from
Osteoarthritis in over 16s: diagnosis and management, guideline NG226, National Institute for Health and Care Excellence, 2022. Referral for joint replacement, and the recommendation to use clinical assessment rather than numerical severity scoring. Read it here.
Abram SGF, Beard DJ, Price AJ, et al. Arthroscopic meniscal surgery: a national society treatment guideline and consensus statement. Bone and Joint Journal, 2019. British Association for Surgery of the Knee. Read it here.
National consensus on the definition, investigation and classification of meniscal lesions of the knee, 2018. Read it here.
Urgent and emergency musculoskeletal conditions requiring onward referral, Arthritis and Musculoskeletal Alliance, version 2, 2020. Read it here.
Musculoskeletal orthopaedic approach to referral optimisation, NHS England, publication reference PR2103, 2023. Read it here.
Knee replacement pathway, Getting It Right First Time (GIRFT), July 2026. Read it here.
Surgeon-governed decision support for first contact MSK clinicians.
The demo uses draft rules awaiting final sign-off, and example patients only.