Layer 1

‘Dispensing robots can fail. Pharmacy should be cautious with them’

"Humans may rely on robots before they're competent enough"

As automated dispensing advances, pharmacists should be aware of the risks of giving responsibility to robots, says Malcolm Brown

Do you understand how dispensing works? “Of course,” you retort. You may not witness every step of the process, such as each action of your accuracy checking technician or dispensing robot, but you are comfortable with it.

You trust them, notably your technician who is trained and bound by financial, legal and moral responsibilities. But what if your team member was a machine with artificial intelligence?

Until recently, computers worked by brute number crunching. Now, clever algorithms and deep learning enable them to solve problems that were once insoluble within the lifetime of a human.

New technology is enticing. It tells us what we want to know, without us fully understanding how it works. We undertake some tasks less ourselves, as machines perform them, so our skills in those areas go unused and atrophy. The equipment’s output is so increasingly useful that we cease to care that we are not sure how it functions.

Soon, a completely automated dispenser may be available to buy. It will check the doses, the interactions, the patient identity and give advice. It will need no lunch break. Importantly, it will make fewer errors than humans do. The pharmacy may over-hype the robot to increase sales.

Humans may rely on robots before they're competent enough. Machines, like people, go wrong. Is a human mechanic available to repair it? Is buying a new machine cheaper? Power cuts and other disasters may produce consequences for robots on workplaces, some imaginable and some not presently imaginable.

It may be too late to avoid automatic dispensers. Perhaps my caution is old-fashioned. However, aircraft pilots have been known to reject certain ultra-sophisticated autopilots. The philosopher Daniel Dennett argues that humans should design robust safety precautions into machines, just as a sword’s guard separates its grip from the blade.

We should use dispensing robots with our eyes wide open, with full training on their limitations and risks, and after signing a legal agreement accepting full liability for them.

Malcolm Brown is a sociologist and retired community, hospital and industrial pharmacist

5 Comments

Emmanuel Chisadza, Community pharmacist

It's all very well when everything works. I recently worked in a pharmacy that uses off-site dispensing. They have very little stock which I suppose is great from a stock management perspective. Problems arose when they lost broadband connection for a week. This meant no EPS, no off-site dispensing and no stock orders via the dispensary software. Starting from a low stock holding base, it was a nightmare. Head office did nothing to help, not even extra support staff.

Benie I, Locum pharmacist

Not so sure the government are listening. They're doing away with human beings as we speak or paying derisory wages but the end goal is the same.

Chris Green, Hospital pharmacist

We've had an automated dispensing system for 11 or 12 years and we're just about to buy our second.  A wonderful bit of kit. Yes, it's a machine so the stuff that comes with that happens sometime, but overall, it's one of the best things we've ever done. It's (almost) 2020 people!!' 

We’re all doomed, Locum pharmacist

 

Fantastic did you personally fund the £80k+ investment?

On a serious note I worked in a hybrid Dr Dispensary. The registered pharmacy technicians thought it quite appropriate that I checked their Dr dispensing output  despite my protestations.

When I presented back to an individual a signed for dispensed item delivered by the robot because it was 100mg instead of 50mg the response was 'no way'. I replied you've signed for it but not taking the accountability? Well the robot got it wrong ! Errr no! Yes robots do make mistakes....

Dave Downham, Manager

Hospital environment?

Job of the week

Pharmacist Manager
Yarm
Competitive