The Oxford Shoulder Score, Explained

What the Oxford Shoulder Score measures

Twelve questions about pain and function in your shoulder, answered by you and nobody else 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.. No one measures your range of motion, no one tests your strength, and no imaging goes into it. The form takes a few minutes and produces one number, which is the whole design: a measure of what the shoulder is costing you, recorded in your own answers rather than filtered through an examination.

Where the items came from matters more than it sounds. Dawson, Fitzpatrick and Carr built the questionnaire out of interviews with shoulder patients — the content came from people describing what their shoulders had taken from them, not from a committee deciding what ought to be asked. They then validated it prospectively in 111 patients assessed before shoulder surgery and again six months afterwards, running it alongside the SF-36, the Stanford Health Assessment Questionnaire, and the surgeon-rated Constant score 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs..

It held up. Internal consistency was high, test-retest reproducibility satisfactory, and it correlated in the expected direction with the Constant score — the evidence that it measures the thing it claims to. On responsiveness it did better than the instruments it was tested against: a standardised effect size comparing favourably with the SF-36 and the HAQ, and the best discrimination of the patients who said their shoulder was "much better" 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.. The questionnaire's single advantage over the general health measures is that it notices improvement they miss.

The shoulder this form was not designed for

Shoulder stabilisation was explicitly excluded from the population the Oxford Shoulder Score was developed and validated in 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.. If your shoulder's problem is that it comes out of joint — that it has dislocated, or subluxes, or that you have learned which movements to avoid because of what happens at the end of them — this is not the instrument for you. The Oxford group published a separate one for instability three years later.

Shoulder instability means the joint slips partly or fully out of its socket, as distinct from a shoulder that is painful or stiff but stays where it belongs. The distinction is not academic. After a dislocation the shoulder is prone to recurrent instability, a Bankart lesion of the labrum is a common associated injury, and treatment ranges from rehabilitation through to surgical stabilisation 2Ref 2American Academy of Orthopaedic Surgeons (OrthoInfo) (2024).Chronic Shoulder Instability and Dislocation.After a shoulder dislocation the joint is prone to recurrent instability; a Bankart (labral) lesion is a common associated injury; treatment ranges from rehabilitation to surgical stabilization. — a different problem, a different operation, and a different set of questions worth asking about the result.

A questionnaire used outside the population it was validated in still returns a number. That is the trap. Nothing on the form will tell you it is the wrong form, and a plausible score for the wrong construct is harder to catch than no score at all. If you are handed one in a clinic and instability is the reason you are there, it is worth asking whether the instability version was meant instead.

Does your score run to 48 or to 60?

Both conventions exist and they point opposite ways, which makes the range the most important thing on the page. The original 1996 scoring gives each of the twelve items a value from 1 to 5, sums them to a total between 12 and 60, and treats 12 as the best possible shoulder — the lowest score means the fewest symptoms 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.. The scoring in general use now runs from 0 to 48 with 48 as the best shoulder, and it was published later, by the same authors, who described it as running in the opposite direction from their original method.

Subtracting the 0-48 score from 60 converts it back. But the arithmetic is a footnote next to the practical point:

• A bare number is not a result. "22" is a poor shoulder on one form and a good one on the other.
• The two are easy to confuse because both produce ordinary-looking two-digit totals, and papers and clinic letters do not always say which they used.
• A series that switches convention mid-way reads as a dramatic deterioration that never happened.

If you are comparing a score from one clinic against a score from another, or a score from this year against a form you filled in years ago, the ranges have to match before the numbers mean anything at all.

Why nobody publishes a good Oxford Shoulder Score

Because the instrument does not have one. The founding paper establishes no severity cutoffs and no threshold for a good result 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs., and the later rescoring paper did not supply them either — its authors argued against categorising scores at all, on the grounds that any cut-off point is approximate and shifts with the population you apply it to. There is no line where a shoulder becomes officially bad.

The same gap runs through the change statistics. The 1996 paper reports no minimal clinically important difference and no minimal detectable change 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.; it demonstrated responsiveness through effect sizes, which tells you the score moves when patients improve without telling you how many points constitute an improvement in one person. More than a decade after publication the same authors were still describing work on MCID estimates as in progress.

The absence of a cutoff is not a hole in your care. It reflects an instrument being honest about a shoulder score being one input among several.

What this leaves you with is the comparison that was always the useful one: your own shoulder against your own shoulder, on the same form, at two points in time. That is the comparison the questionnaire was validated to make — before surgery and six months after 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs. — and it is the only one where the number is doing work that could not be done by a conversation.

The Oxford Shoulder Score and the constant-murley score

They measure overlapping things by opposite methods, which is exactly why the Oxford was validated against the Constant 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs. rather than replacing it. The Constant-Murley score is a composite: 100 points across four domains — pain, activities of daily living, range of motion, and strength — combining what the patient reports with what the examiner measures, and a higher score means better shoulder function 3Ref 3Constant CR, Murley AH (1987).A clinical method of functional assessment of the shoulder.The Constant-Murley score is a composite shoulder-function measure combining subjective and objective components, scored out of 100 points across pain, activities of daily living, range of motion and strength, with higher scores indicating better function; the method applies irrespective of diagnosis or radiologic abnormality, records individual parameters while yielding an overall functional assessment, is reproducible between different observers, and is sensitive to small changes in function; it reports no MCID, MDC or severity cutoffs..

That design has consequences either way.

Constant and Murley's method records the individual parameters while still producing an overall functional assessment, is reproducible between different observers, and is sensitive to small changes in function 3Ref 3Constant CR, Murley AH (1987).A clinical method of functional assessment of the shoulder.The Constant-Murley score is a composite shoulder-function measure combining subjective and objective components, scored out of 100 points across pain, activities of daily living, range of motion and strength, with higher scores indicating better function; the method applies irrespective of diagnosis or radiologic abnormality, records individual parameters while yielding an overall functional assessment, is reproducible between different observers, and is sensitive to small changes in function; it reports no MCID, MDC or severity cutoffs.. What it cannot be is filled in at your kitchen table, and a measured strength figure is not available for the shoulder you had in 2019. The Oxford's twelve questions can be answered anywhere, which is why serial tracking tends to fall to instruments like it.

Neither carries an MCID or a normative band from its founding paper 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.3Ref 3Constant CR, Murley AH (1987).A clinical method of functional assessment of the shoulder.The Constant-Murley score is a composite shoulder-function measure combining subjective and objective components, scored out of 100 points across pain, activities of daily living, range of motion and strength, with higher scores indicating better function; the method applies irrespective of diagnosis or radiologic abnormality, records individual parameters while yielding an overall functional assessment, is reproducible between different observers, and is sensitive to small changes in function; it reports no MCID, MDC or severity cutoffs.. Both leave interpretation to the person reading it next to everything else they know about you.

The other shoulder questionnaires you may meet

Shoulder pain arrives in clinic attached to a lot of different problems, and the questionnaires have specialised accordingly. Which form you are handed says something about what the person handing it to you is trying to find out, and it is a fair question to ask them.

• The DASH measures symptoms and physical function across upper-extremity musculoskeletal disorders generally 4Ref 4Hudak PL, Amadio PC, Bombardier C (Upper Extremity Collaborative Group) (1996).Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder, and hand).The DASH is a validated self-reported measure of symptoms and physical function across upper-extremity musculoskeletal disorders. — arm, shoulder and hand together, rather than the shoulder alone. Useful when the problem is not confined to one joint, or when the elbow and wrist are part of the story.
• The oxford hip score is the elder sibling of this whole family: twelve items covering pain and function in one composite, developed and validated prospectively in 220 patients before total hip replacement and at six months 5Ref 5Dawson J, Fitzpatrick R, Carr A, Murray D. (1996).Questionnaire on the perceptions of patients about total hip replacement.The Oxford Hip Score is a 12-item patient-completed questionnaire covering hip pain and function in a single composite score, developed and validated prospectively in 220 patients before total hip replacement and at six-month follow-up; its original scoring sums twelve 5-point items to 12-60 with lower = better.. The shoulder score followed the same architecture the same year, and the knee version two years after that. Same twelve-item shape, same original 12-60 scoring, same later rescoring 1Ref 1Dawson J, Fitzpatrick R, Carr A (1996).Questionnaire on the perceptions of patients about shoulder surgery.The Oxford Shoulder Score is a 12-item patient-completed questionnaire developed from patient interviews and validated prospectively in 111 patients before shoulder surgery and at six months, alongside the SF-36, the Stanford HAQ and the surgeon-rated Constant score; its intended population is shoulder surgery excluding stabilisation; high internal consistency, satisfactory test-retest reproducibility, construct validity via correlation with the Constant score, and responsiveness with a standardised effect size comparing favourably with the SF-36 and HAQ plus best discrimination of patients reporting their shoulder was much better; the original scoring runs 12-60 with 12 = best; the paper reports no MCID, no MDC and no severity cutoffs.5Ref 5Dawson J, Fitzpatrick R, Carr A, Murray D. (1996).Questionnaire on the perceptions of patients about total hip replacement.The Oxford Hip Score is a 12-item patient-completed questionnaire covering hip pain and function in a single composite score, developed and validated prospectively in 220 patients before total hip replacement and at six-month follow-up; its original scoring sums twelve 5-point items to 12-60 with lower = better..

What none of them does is diagnose. If shoulder pain has brought you here without an explanation attached, that is a different question — one that starts with an examination. Impingement and rotator cuff tendinitis, one of the more common reasons a shoulder hurts, are managed first without surgery: rest, anti-inflammatory medication, physical therapy, and sometimes an injection 6Ref 6American Academy of Orthopaedic Surgeons (OrthoInfo) (2024).Shoulder Impingement / Rotator Cuff Tendinitis.Shoulder impingement and rotator cuff tendinitis are commonly managed nonsurgically with rest, NSAIDs, physical therapy and injections.. A questionnaire is how the result of that gets measured. It is not how the decision gets made.

Filling one in before your appointment

The most useful copy of the Oxford Shoulder Score is usually the first one, filled in before anything is done — and the most commonly missed one, because at that point there is nothing to compare it against and it feels like paperwork. Six months later it is the only record of where you started, and without it the after-score floats free.

Three habits make the series readable:

• Answer it about the week you actually had, not the good day you are having in the waiting room. A form completed on your best morning of the month makes the next one look like a decline.
• Note the range alongside the number every time, and keep to one version.
• Bring the old form, not the memory of the old form. Two sheets side by side show a shape that no recollection of "about the same, maybe a bit better" can.

It is also worth knowing what the score is not measuring. It does not ask what you do for work, whether you sleep on that side, or whether you have stopped reaching for things without noticing you stopped. The number is a summary of twelve answers, not a summary of your shoulder — the things it leaves out are precisely what the appointment is for.