What Fake Surgery Taught Us About Real Surgery

What is sham surgery in research?

Sham surgery is a randomised trial in which the control group receives a convincing imitation of an operation. They are prepared, anaesthetised, and taken to theatre. Skin is incised. The instruments, the sounds, and the elapsed time match the real procedure. What is withheld is the single step thought to be therapeutic — the trimming, the shaving, the removal.

A sham is a placebo with a scalpel: everything the patient experiences is held constant except the ingredient under test. Participants consent in advance to the possibility of receiving one, and are not told which they had until the trial ends. The people scoring the outcomes are blinded too, which matters more than it sounds — an assessor who knows who was operated on is a source of the very effect being measured.

The design is uncomfortable, and it is meant to be. Ethics committees permit it only under conditions that are themselves informative, and those conditions turn out to explain most of what sham evidence can and cannot say.

Why would anyone test an operation against a fake one?

Because a great many things make people better after surgery, and only one of them is the surgery. Symptoms that fluctuate tend to be at their worst when someone finally seeks help, and drift back toward their average afterwards. Expectation is powerful, and an operation is expectation's most elaborate form: the referral, the theatre, the recovery, the months of careful attention. Time does its own work.

All of that lands on a patient whether or not the therapeutic step happened. So "my patients improve" is a true observation and an empty argument. The comparison it needs is not the patient beforehand but a matched patient who got everything except the step. That is what a sham supplies and what nothing else does.

The first orthopaedic result to land this way concerned the knee. Among people aged 35 to 65 with a degenerative medial meniscus tear and no osteoarthritis, arthroscopic partial meniscectomy was no better than sham surgery for relieving symptoms 1Ref 1Sihvonen R, Paavola M, Malmivaara A, et al. (FIDELITY) (2013).Arthroscopic Partial Meniscectomy versus Sham Surgery for a Degenerative Meniscal Tear.In patients aged 35-65 with a degenerative medial meniscus tear and no osteoarthritis, arthroscopic partial meniscectomy was no better than sham surgery for symptom relief — used as the primary worked example of a placebo-controlled surgical trial and its result..

What did the sham trials actually find?

Two heavily performed orthopaedic operations were tested this way, and neither beat its imitation. The knee result stood on its own; the shoulder result arrived with an extra arm that made it harder to set aside. A later high-certainty synthesis reached the same place. These are not preliminary signals — they are the strongest evidence the field has produced about these procedures.

The shoulder trial's third arm is the part worth pausing on. By including a group that received nothing, it separated two questions that are usually welded together: does the operation beat the ritual, and does the ritual beat leaving it alone? The answers were no and no 2Ref 2Beard DJ, Rees JL, Cook JA, et al. (CSAW) (2018).Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial.Arthroscopic subacromial decompression provided no clinically important benefit over placebo (arthroscopy only) or over no treatment for subacromial shoulder pain — used as the second sham-controlled trial and, via its three-group design, to separate the operation from the ritual and from doing nothing.. The CSAW trial is why that procedure is now taught as a cautionary tale rather than as an indication.

A placebo control can reveal harm, not just absence of benefit

This is the half of the story that gets lost. A placebo comparison does not only ask whether a treatment beats nothing; it can show that a treatment does worse than nothing. For tennis elbow, corticosteroid injection produced worse outcomes at one year and higher recurrence than a placebo injection 4Ref 4Coombes BK, Bisset L, Brooks P, Khan A, Vicenzino B (2013).Effect of Corticosteroid Injection, Physiotherapy, or Both on Clinical Outcomes in Patients With Unilateral Lateral Epicondylalgia: A Randomized Controlled Trial.For lateral epicondylalgia, corticosteroid injection produced worse one-year outcomes and higher recurrence than placebo injection — used to show that a placebo control can detect harm, not merely the absence of benefit, and that an uncontrolled series could not have found it.. The early relief was real. It was also borrowed.

Without the placebo arm, that trial would have shown exactly what everyone already believed: people felt better within weeks. The arm is what revealed that the same people were worse a year on than those who had received nothing active. An uncontrolled series cannot find that. It is structurally incapable of finding it, because it has nothing to be worse than.

A control group is not a formality. It is the only instrument that can tell a treatment that works from one that borrows against the future.

What sham trials cannot test, and the operations that will never have one

No one will ever run a sham trial of fixing a broken hip, washing out an infected joint, or decompressing a spinal cord that is losing power by the week. Those trials do not exist and should not: a sham requires genuine uncertainty about whether the operation helps, and for these there is none. The condition is dangerous, waiting causes the damage, and the control arm would itself be a harm.

That selection is the most useful thing about the sham literature and the least quoted. The operations that have been sham-tested are, by construction, the ones where doubt was already reasonable — elective, offered for pain, in people who could safely wait. Sham evidence is concentrated exactly there, and it does not generalise outward.

Where the question was never open:

• A displaced or unstable fracture, or one failing to unite.
• Infection in or around a joint or the spine.
• Cauda equina syndrome — saddle numbness with new bladder or bowel dysfunction. An emergency, not a decision.
• Progressive neurological deficit — weakness measurably worse than last month, not pain that comes and goes.
• A joint mechanically locked or unstable — a knee that will not straighten, a shoulder that leaves its socket.
• A joint destroyed by end-stage arthritis, where the operation replaces the failed part rather than reshaping a painful one.

And surgery sometimes wins its comparisons outright. Hip arthroscopy for femoroacetabular impingement syndrome produced modestly better patient-reported hip function at twelve months than personalised physiotherapist-led conservative care, at substantially higher cost 5Ref 5Griffin DR, Dickenson EJ, Wall PDH, et al. (UK FASHIoN) (2018).Hip arthroscopy versus best conservative care for the treatment of femoroacetabular impingement syndrome (UK FASHIoN): a multicentre randomised controlled trial.Hip arthroscopy for femoroacetabular impingement syndrome produced modestly better patient-reported hip function at 12 months than personalised physiotherapist-led conservative care, at substantially higher cost — used as the counterweight showing an operation outperforming its comparison.. The sham literature does not overturn that. It was never about it.

You don't always need a sham to isolate the active ingredient

Where a sham is impossible or unnecessary, trials isolate the ingredient another way: they compare the operation with and without the part in question, or against the serious alternative. Adding instrumented fusion to decompression for lumbar spinal stenosis did not improve outcomes at two or five years compared with decompression alone, while adding cost and operative burden 6Ref 6Försth P, Ólafsson G, Carlsson T, et al. (2016).A Randomized, Controlled Trial of Fusion Surgery for Lumbar Spinal Stenosis.Adding instrumented fusion to decompression for lumbar spinal stenosis did not improve clinical outcomes at two or five years versus decompression alone, but increased cost and operative burden — used as the example of a component-isolation trial design that tests an ingredient without requiring a sham..

That is a component test. It leaves the shared parts in both arms and varies one thing — the same logic a sham applies to the whole procedure. It answers a narrower question and requires no deception, which is why most surgical evidence looks like this rather than like a placebo-controlled operation.

The third shape compares an operation against the real non-surgical alternative: acl surgery vs rehab is that question, and so is the achilles rupture decision. None of these are shams. All of them ask the same thing — what does the operation add that the alternative does not?

How to read the next surgery study you are handed

The first question is what the operation was compared with, because that single fact bounds everything the study can claim. Against nothing, it can only tell you that people improved. Against usual care, it can tell you whether it beat the alternative. Against a sham, and only then, it can tell you the therapeutic step is what did the work.

• What was the control? Sham, usual care, a waiting list, or no comparison at all.
• Were the outcome assessors blinded? Unblinded scoring of a subjective outcome measures belief as much as biology.
• How much crossover? When many controls cross over to surgery, the comparison blurs and the result drifts toward no difference.
• What was the primary outcome, and what counted as a meaningful change on it?
• How long was the follow-up? Treatments that separate at three months have often converged by a year.

Reading surgical trials this way is a skill rather than a specialty, and it survives contact with the consulting room. It is also worth remembering what none of these trials measure: what a recovery costs, when you could return to work after surgery, and who carries the months in between. The vertebroplasty evidence is its own long chapter of the same argument.