Steady Hands

07:42. The patient is already asleep when the console is powered up. She is fifty-eight, she has a tumour in the lower left of her colon, and she was told nine days ago that the operation would be performed by a robot. The consultant corrected her, gently and for the third time, during the pre-operative discussion.

"I will be performing the operation," he said. "The machine is the instrument. It does not decide anything and it does not move unless I move."

This is not modesty. It is an accurate description, and the persistent public misunderstanding of it is one of the small frustrations of the speciality.

08:05. The abdomen is insufflated with carbon dioxide, raising the abdominal wall away from the viscera to create a working cavity — a dome perhaps fifteen centimetres high. Four ports are placed through incisions of eight millimetres. A camera enters through one. Through the others go three instruments: a grasper, a vessel sealer, and a pair of scissors with an integrated diathermy connection.

The robotic arms are then wheeled in and docked to the ports. Docking is the step most likely to cause difficulty, and it is entirely mechanical: the arms must be positioned so that their working ranges do not collide either with each other or with the patient's hip, and a poor configuration will announce itself forty minutes later as an instrument that cannot reach where it needs to go. The registrar has done this eleven times. She takes six minutes. The consultant, watching from two metres away, says nothing at all until the last arm is locked, and then says, "Good."

08:31. The consultant sits down at the console, which is positioned across the room and faces away from the table.

This is the arrangement that non-specialists find most difficult to accept. The operating surgeon is not at the patient's side. He is seated, his forearms supported, his head in a binocular viewer, his feet on a set of pedals. The assistant and the scrub nurse remain at the table; if the patient requires an immediate open procedure, the instruments are withdrawn, the arms undocked and the abdomen opened conventionally, and the team rehearses this sequence regularly for exactly that reason.

What the console offers, in exchange, is threefold.

The view is stereoscopic and magnified approximately tenfold. Structures that are difficult to distinguish through a standard laparoscopic camera — the plane between the mesentery and the retroperitoneum, the ureter running beneath it — are here separated with a clarity that shortens the operation and reduces the principal risk of this procedure, which is injury to a structure the surgeon did not see.

The instruments are wristed. A conventional laparoscopic instrument is a rigid rod through a fixed point, pivoting like an oar in a rowlock, which means that to move the tip left the surgeon must move the handle right and that the tip cannot articulate at all. The robotic instrument has an additional joint a centimetre from its end, giving it a range of motion comparable to a human wrist inside a cavity the size of a grapefruit. Suturing, in particular, becomes a substantially different task.

And the motion is filtered. Physiological tremor — present in every human hand, of small amplitude and higher frequency than voluntary movement — is removed algorithmically. Motion is additionally scaled: a five-centimetre movement of the surgeon's hand produces a one-centimetre movement of the instrument tip. The surgeon operates at a precision his own hands do not possess.

09:14. The mobilisation of the sigmoid colon proceeds along the correct embryological plane, which is bloodless if found and troublesome if not. The consultant narrates for the registrar in short phrases.

"Watch the tissue, not the instrument. There. That's the plane. It wants to open — I'm not cutting, I'm letting it separate."

There is a pause. Something has caught.

"That's tethered. I can see it's tethered because the tissue on the left isn't following. I can't feel that it's tethered."

This is the system's principal deficiency, and it is not a minor one. Force feedback — haptics — is largely absent. The surgeon receives no sensation of tension, resistance or texture. The judgement of how hard a suture may be pulled before it tears, which an open surgeon makes through the fingers, must here be made through the eyes, by watching tissue deform. Experienced operators develop this substitution to a remarkable degree and are frequently unaware that they have. Trainees break sutures. The learning curve for the technique is measured in dozens of cases, and the curve for haptic substitution is the steepest part of it.

10:48. The specimen is removed through a small extension of one port site. The anastomosis — the rejoining of the two ends of bowel — is performed with a circular stapler and then inspected, tested for leaks under saline, and inspected again.

11:20. Undocked. Closed. Total operative time three hours eighteen minutes, which is roughly twenty minutes longer than the same consultant's average for the equivalent laparoscopic procedure. That difference is consistent with the published literature, and it is one half of the central controversy about this technology.

The other half is cost. A system represents a capital outlay in the low millions with a substantial annual service contract, and each case consumes instruments that are licensed for a fixed number of uses. Against this, randomised comparisons across several procedures have repeatedly found outcomes — complication rates, oncological adequacy, length of stay — that are equivalent to good laparoscopic surgery rather than superior to it. Where robotic assistance shows clearer advantage is in operating deep within the pelvis, where the working space is narrow and rigid instruments are at their worst, and in reducing conversions to open surgery in technically difficult cases.

There is a further benefit that appears in no outcome table. Laparoscopic surgery is performed standing, with the arms elevated and the neck extended towards a monitor, often for many hours, and the resulting rates of musculoskeletal injury among surgeons are high enough to shorten careers. The console surgeon is seated and supported.

11:35. The consultant stands, stretches once, and goes to write the operation note. The registrar begins the undocking checklist for the second case.

"The robot doesn't get tired," he says, on the way out. "That isn't the interesting part. The interesting part is that neither do I, now, by four in the afternoon."