Robot Pharmacists

Makarand writes "The next time you visit a pharmacy your prescription may be filled by a robot according to a TechTV article. Hospitals and drugstores are now increasingly relying on automated technology to count, bottle, and label prescription drugs in a faster and more accurate way. The technology uses a bar-code system similar to those used to read prices in grocery stores. Doctors enter prescription details directly into the pharmacy computer. The robot springs into action when an order is recieved. Riding on a conveyor belt, the robot picks up an empty vial, identifies the bar code of the chosen drug, and automatically fills the drug bottle."

Comments

This is bad

[dangermouse]

Human pharmacists have always served as information resources for their customers, and even as a check on poorly-chosen prescriptions from doctors.

I can understand automating away the cashier or the janitor, but automating away a job where human judgment is so crucial is a terrible idea.

Re:This is bad

[Daniel Dvorkin]

I've never understand the point of pharmacists. Really. To reply to dangermouse, shouldn't the doctor be an information resource? Also, the doctor should be picking the right prescriptions in the first place. I don't want to upset the pharmacists here, but it's just something I've never understood. I say, bring on the robots!!

To put it very simply: it's the doctor's job to decide what medications you should take, it's the pharmacist's job to tell you how to take it, and just as importantly, how not to take it.

This is an oversimplification, of course. A good physician needs to be able to give a patient advice on using medications. But doctors are busy; very, very often the doc needs to shoo the patient out of the exam room to get to the next patient, and the (often very confused) patient goes down to the pharmacy to get everything straightened out. This is particularly important for older people who may be a bit confused anyway, and on a whole shitload of different medications.

The problem of patients on multiple meds brings up another valuable function of pharmacists, which is catching drug interactions. Ideally, every time you went to the doctor -- any doctor, in any specialty, at any hospital or clinic, for every problem -- that doctor would a) know everything about your medical history (pharmaceutical and otherwise) and b) know every possible drug interaction and contraindication. The real world, of course, doesn't live up to this ideal. The fact is that people are more consistent in their pharmacist than in their doctor; old people with multiple conditions very often go to see several different specialists, but get all their meds from the same place. When a patient has a list of meds that fills both sides of a sheet of paper (and believe me, I've seen it) it's nice to have that extra sanity check.

Finally, good clinics and hospitals are increasingly finding it useful to have at least one or two PharmD's on staff to advise the physicians before the prescription is issued. "Hey, I've got this guy who's on Alizadol and Corvabarin, but his liver enzymes are running kind of high; what do you think of giving him Zelarin instead of the Alizadol?" (Drug names are made up, of course -- hey, it's New Year's, and my own biochemistry is a little altered right now; you want me to think of real ones right off the bat? ;) MD's know a great deal about the human body, but they very often find it impossible to keep up with the dizzying variety of chemicals designed to be put into the body.

When it's your body they're screwing with, it's good to have several sources of human judgement. Doctors can and do make mistakes, either because of a lack of specialized knowledge or (very often, especially in the case of interns) because of exahustion. Pharmacists very often keep those mistakes from killing people.

Re:This is bad

[CrystalFalcon]

The point of pharmacists should be obvious. You NEED to study for four years to understand the kind of handwriting they have to read to get you the right bottle.

Re:This is bad

[LostCluster]

I think a very big point that needs to be made about the kind of system we're discussing is that it's only designed to be a tool that helps pharmacists with the boring repetitive pill-counting part of their job.

The phamacist will still have the responsiblity for double-checking this machine's actions (hey, Viagra's not supposed be yellow...) and also still have the responsibility to consult tell the patient what they should be doing before/after using the medication and what to do if they forget about a dose.

My brother recently had his wisdom teeth removed, and when we took the perscription to the phamacist she was visibily worried that she was out of stock on the perscribed pain-killer, because she realized my brother was in pain at home for every second that was wasted. She went to her computer and was relieved to learn that she had it in the back room, but finding it in that back room and then counting was a process that still costs time. I'm sure that she would have loved to be able to type the number into the computer and have the bottle pop out onto her desk so we could be sent back to my brother faster.

Re:This is bad

[InsaneCreator]

To put it very simply: it's the doctor's job to decide what medications you should take, it's the pharmacist's job to tell you how to take it, and just as importantly, how not to take it.

A guy I know has been told once by the pharmacist to "take the pills 3 times a day with a bucket of water". Apparently the doctor prescribed him some horse tablets. Go figure...

Re:This is bad

[Jace of Fuse!]

I can understand automating away the cashier or the janitor, too, but I really don't agree with it. Even "lowly" jobs such as these still add a social element to the jobs. You see a person when you are checking out, or you know that there is an accountable person doing the cleaning. By having people do work, even if it's inferior work, the environment the work is being done in doesn't feel sterile or dead.

Not to mention the whole issue of unemployment. I hate to see people lose their jobs to machines. I hate to see it, because even before I lost my job it was threatened with automation, and in a way, it still is.

Eventually even yours may be as well no matter what it is you do.

Re:This is bad

[dangermouse]

Eventually even yours may be [threatened by automation] as well no matter what it is you do.

Er, probably not. I'm sort of in the business of automating things-- I'm a software developer.

My view on job automation is that if it can be done, it should be done; people should be doing work that can only be done by people, and machines should do work where they can. But I am strongly opposed to removing human judgment from systems that require it, which means I'm against things like the "robot pharmacist" (as opposed to the "robot bottle-filler-and-labeler") and mandatory sentencing legislation.

I realize there's an ethical question when you consider that some people are only as competent as machines, through no failing of their own. I don't know where I stand on that yet, but I do think that most people are more capable than machines and will always be so. Automating tasks both enables (through cost and time reduction) us to approach more complex tasks and simultaneously frees up the people necessary to undertake them. This is how we have always progressed, especially since the industrial revolution.

Re:This is bad

[Jace of Fuse!]

There is an ethical question when you consider some people are competent in many things, but skilled in certain ones. For example, sheet metal fabrication. A skilled machinist can be replaced (at least to some degree) with a person. That skilled work isn't easy, and I promise you that many of those workers do things you absolutely can not. You could be trained of course, and I'm sure eventually you too could become as skilled as they. A properly programmed machine could come along and replace all of them overnight.

Most of these said metal workers have spent many years getting where they are now, and practically have no other valuable work skills. This doesn't mean they are "only as competent as machines", it means they are highly competent in one area. They are skilled at work that can and will likely be eventually taken away from them. After that point, they'll be condemned to either spend many years learning a new trade that will too likely one day be replaced with machines or they will be working shit jobs alongside highschool students.

I'm not saying automation is totally bad, but sometimes the whole debate about whether or not we should automate certain things is driven by the bottom line of profit.

I learned a long time ago that when your motives are purely profit, nothing good can come of your actions.

Re:This is bad

[dangermouse]

Of course. Whenever you automate a task that formerly required special skills-- which is probably most of the time-- the skilled workers who perform that task are going to lose their jobs and quite possibly be relegated to working as "unskilled" laborers.

But those people losing their jobs is inescapable, really, if you wish to automate their tasks. What you do with those people is a matter of ethics that is quite aside from the decision to automate their tasks. If you replace a bunch of skilled workers with machines, you then make further choices regarding whether or not those people will be retrained, moved to other positions, given severance packages, etc.

Remember, though, that such a move will only affect the current generation of workers. That's a small price to pay for the forward progress that automation allows us. Where would we be if we had refused to adopt automatic looms, for instance, on the grounds that the weavers would suddenly be robbed of their economic niche?

Of course automation is driven by the bottom line of profit; profit is return on expenditure, or essentially benefit from work. Work always boils down, in the end, to human work. The more you automate, the more you're getting from the same amount of human work. This can only be bad when the automated task itself is compromised (ie, when the task requires human judgment), in which case the task cannot truly be said to have been automated; in such cases, the task has really been eliminated. But there will always be more work for people to do, because we create work.

To sum up my position, I'll provide a hypothetical scenario. A factory owner decides to automate a task performed by some of his highly skilled workers. I would argue that if such automation is possible, he should do so. I would also argue that he has a moral obligation to try to compensate those workers for making their skills irrelevant, and that such compensation should come in the form of an effort to restore their value as workers. But I consider the decisions to automate and to compensate to be distinct.

Re:This is bad

[Jace of Fuse!]

I can see your point, but you will have to agree with one thing. More often than not, the choice to automate results in lost jobs, without compensation.

I can see how it's easy to say the two decisions are distinct, but realistically I think we both know that the first one normally results in a second one that is unfavorable to the workers. Exceptions exist of course, but retaining and retaining one or two people happens more often than retaining and retraining many dozens.

I will agree that when automation is done to take workers out of hazardous work the debate becomes different. It's no longer just profit that is the driving force. It's the well-being of the individuals that helps make that choice, even if sometimes it's a second concern after profit. Companies will always do what they do for profit.

Happening on a smaller scale as well...

[mcgroarty]

This is also happening on a smaller scale.

Many hospitals use electronic equipment to fill prescriptions for patients inside of the hospital now, as well. Hundreds of deaths are attributed each year to wrong dosages or wrong medications given to patients, and robotic filling equipment virtually eliminates the problem.

I'd wager that it only takes avoiding a single lawsuit to pay for the equipment.

Quality of service?

[psi_diddy]

Of course robots may break down, too, especially since they're only as good as the humans who program them. As an added safety feature, a camera takes a digital picture of pills in each bottle so a pharmacy screener can compare the pills in the picture to the doctor's orders and make sure the robot made no mistakes.

Having seen one of these in action, the device is only as good as the guy filling the drug bins, who thanks to this marvelous technology will be a college intern who stayed up late last night studying for his exams. He will also be the guy checking the picture at the end. "Yup, those are pills."

Great first step!

[Chris_Stankowitz]

This is a great first step in avoiding complications, however the problems are caused by human error. They have not solved that problem if "Doctors enter prescription details directly into the pharmacy computer." Adding some sort of check&balance to this system would greatly reduce the change of prescription error related illnesses.

Re:Great first step!

[LostCluster]

All of the national drug store change have computer systems that record the perscriptions they've filled for a given person, and will kick and scream when an undesirable interaction will occur between the previous perscriptions and the new one. So long as you use the same drug store chain for all your medications, you should be well protected.

Re:Great first step!

[MacAndrew]

"Doctors enter prescription details directly into the pharmacy computer."

Judging from my doctor's handwriting, not to mention computer skills, this is a very bad sign. If a pharmacist serves a function, it's decoding the doctor's intentions from chicken-scratch heiroglyphics.

*

I should add that computers have been doublechecking dosages and interactions for quite a while, and are probably better at it than human pharmacists. I've seen the computer flag a couple of my prescriptions, and god knows my insurance company is very careful that I don't get my refills more than 5 days in advance. (It's about money not safety: I'm welcome to pay cash and apply for reimbursement later. No, I don't see how this saves them money.)

Re:And what happens when

[LostCluster]

The robot is less likely to make that mistake than humans.

Why fill Bottles?

[EABinGA]

I don't understand why in the US pharmacists need to fill bottles. It seems such a waste of time, and would seem to increase the cost of medication dramatically.

It would be much easier to have them prepackaged at the manufacturer, so the pharmacist simply reaches in the shelf and grabs the prepackaged box of whatever the doctor prescribed.

Re:Why fill Bottles?

[rehabdoll]

And thats how we are already doing it in the rest of the world.

Feel welcome to join us, USA.

Re:Why fill Bottles?

[mcgroarty]

It would be much easier to have them prepackaged at the manufacturer, so the pharmacist simply reaches in the shelf and grabs the prepackaged box of whatever the doctor prescribed.

I've wondered about this as well. An English coworker told me that in the UK, every medication he'd seen was sold this way. Given that most all medication is prescribed in one month intervals at one of just a few dosages, it seems to make perfect sense.

Being able to break apart the blister packs they're usually on also makes it much more convenient to pocket doses when going out and what not.

Re:Why fill Bottles?

[LostCluster]

Because doctors can perscribe any whole number of pills to any given individual. (In fact, they can perscribe half pills at times too.) The law requires that the pharmacist give exactly the number of pills perscribed, no more and no less.

Having them in pre-packaged units works for some medications, but for the vast majority you'd still find the phamacist having to rip the blister packs in half in order to get the exact number of doses the doctor requested.

Re:Why fill Bottles?

[EABinGA]

Being able to break apart the blister packs they're usually on also makes it much more convenient to pocket doses when going out and what not.

Actually you would be in violation of federal law if you were to carry around medication like that, because it is not in its labelled container. Many people that carry their daily dose of pills around in pill organizers aren't even aware of that. Usually its no big deal, but it can get you busted big time.

Re:Why fill Bottles?

[nhavar]

It helps to cut down on waste, reduces storage costs/space and allows more individually taylored dosages to a particular patients needs.

Re:Why fill Bottles?

[rjamestaylor]

• It would be much easier to have them prepackaged at the manufacturer, so the pharmacist simply reaches in the shelf and grabs the prepackaged box of whatever the doctor prescribed.

Yes and no. First, note that my company, Dispensing Solutions, Inc., is a repackaging company specializing in point-of-care dispensing (meaning the physician or other prescribing healthcare provider hands the patient a bottle of the medication s/he prescribes instead of a prescription). I am not a pharmacist nor a healthcare professional; I'm the CTO and developer of the Internet-based point-of-care dispensing application (basically a real-time inventory control system with a procedurally controlled dispensing component), which is used in physician practices and community health clinics.

The problem with the mfg making pre-packed bottles of drugs for instant dispensing is that doctors (more properly called "providers" since not only MD's can dispense drugs legally) don't prescribe the same number of pills and of the same strength for the same diagnosis. For example, Amoxicillin may come in 250 MG caplets or 500 MG caplets. A provider may prescribe 30 caps of Amoxicillin 250 MG,with one cap to be taken 4 times a day. Another may prescribe 28 caps. If the Rx says "28 caps" you can't dispense 30 caps just because that's the only bottle size you have on hand. Each provider has his/her own way, so mfg cannot pre-determine the dosing without practicing medication themselves.

This then is where repackaging companies such as the one I work for come in. We work directly with the providers to determine the top 20% of prescriptions and work to standardize their prescribing habits, according to their direction. This way we can provide high-quality prepacked drugs for point-of-care dispensing.

What's the benefit? One is quality control. With ever-increasing workloads and shrinking numbers of pharmacists the number of Rx fills per day is increasing, as is the number of errors and related injuries and deaths. In the same line, we are regulated by the FDA and licensed as a drug manufacturer. A pharmacist is regulated in each state by the Board of Pharmacy. The rules on cross-contamination are much more severe on us than on your local pharmacist. For example, penicillin is processed in the same area as all other drugs in a pharmacy, but we are forced to use a negative-air flow clean room for filling penicillin-type products and are not allowed to have any cross-contamination. It is unfortunately not uncommon for people allergic to penicillin to suffer anaphylactic shock from a cross-contaminated non-penicillin drug filled at a pharmacy.

Another benefit is financial to the provider. Providers, by law, cannot receive money from prescriptions written to be filled by a pharmacist. But, as long as the service is provided to the provider's own patients in their own practice/clinic, they can make money dispensing drugs they would otherwise send out to be filled. Some providers make an extra one or two hundred thousand dollars a year just dispensing their most commonly prescribed drugs.

There is a benefit to the insurer, or, more accurately, the pharmaceutical benefit manager (PBM). Each plan has a preferred formulary (say chosing Allegra over Clarinex for allergies). By having the preferred drug available in house the PBM could "give away" the preferred drug and save money over paying for the higher drug even with a higher co-pay.

Another therapeutical benefit is in therapy compliance. Usually, a provider has no way of knowing if a prescription written has been filled. However, if the provider hands the patient the drug directly it has been shown that there is a higher probability of that patient completing the therapy prescribed.

Lastly ('cause this is too long), there is a benefit for pharmacists, too. Precisely because the best use of a pharmacist's time is counselling not filling, we work with pharmacists to pre-pack common movers in their pharmacies. Thus the pharmacist needs only grab the right bottle matching the drug, strength, and count for the prescription. No counting (but they still collect the "fill fee"). This way the pharmacists can fill more prescriptions--safely.

The filling machines are neat. We have several in our facility, including one that has the capacity to fill 200,000 bottles a day. While these machines are slick, they're expense is hard to swallow, especially when companies like DSI are able to fill the need. Neat article. Good to see it on Slashdot.

Re:Why fill Bottles?

[King_TJ]

Actually, I think some of this depends on the situation. I did some temp. work recently for ExpressScripts (mail order pharmacy), and worked on one of their drug programs that issues some medications for free to the poor/needy.

In this particular program, we certainly would send the doctors prescriptions that came in pre-packaged bottles. (Typically, 100 count bottles, but sometimes 50 count.) We only filled 3 month (90 day) prescriptions at a time, but if a doctor specifically asked for 90 pills (1 a day for a 3 month supply), we'd send off the bottle of 100.

THX-1138 here we come!

[corebreech]

The real prequel to Star Wars

Re:And what happens when

[JaredOfEuropa]

Well, the lawyers get richer, of course!

For what it's worth they reckon that the robot is more accurate than a human. It might well be true: people preparing medicine doses in hospitals or drug stores aren't 100% accurate anyway, and mistakes do happen. Sometimes with deadly consequences.

Slick Technology

[edlong]

This company, http://www.innovat.com (skip flash intro), innovation associates, has some cool technology that does this. For example, if the doctor mistakenly chooses pill A, instead of pill B, the machine will not dispense. Also, you can't fill the Tylenol Aspirin tray with anything but that. It uses some fancy recognition software; it can tell the difference between a skittle and and M&M, plus it won't dispense if pills are deformed (chiped etc.) cool stuff.

Re:Slick Technology

[JaredOfEuropa]

"it can tell the difference between a skittle and and M&M"

Thank god for that!

Re:Slick Technology

[ross.w]

Hopefully this technology can overcome the situation my wife found herself in, after being prescribed a series of drugs that weren't compatible. This despite the doctor having a complete record. We don't go to that doctor anymore...

Greaaat...

[mcgroarty]

The technology uses a bar-code system similar to those used to read prices in grocery stores.

Great.

*boop* Milk, 2% - $2.69...

*boop* pretzel sticks - $1.39...

*boop* ... ??? *boop* ... ??? *boop* *boop* *boop* ...??? "Just a moment, sir..."

*skreeeeeeee* "PRICE CHECK ON 200MG TYLAMANEX ANTIDIURETIC INCONTINENCE AID(-id-id-id)!!!??? PRICE CHECK ON 200MG TYLAMANEX ANTIDIURETIC INCONTINENCE AID(-id-id-id-id)!!!???"

Re:And what happens when

[gmack]

I once had a pharmasist give me siniquan(anti phsychotic) instead of singular(asthma drug).

Didn't even get an appology for a week of hell and almost losing my job because I couldn't be motivated to do anything.

The way I see it a robot would be one less source of screwups provided the doctors move to an electronic perscription system.

Re:And what happens when

[JaredOfEuropa]

Over here, the doctors already use an electronic prescription system most of the time. A doctor will ask you which pharmacy you go to, enter the details in his computer, and by the time you get to the pharmacy, the prescription will be ready and waiting for you. Having a robot doing the actual filling would eliminate another source of human error.

Robots and mistakes

[goombah99]

While it's probably true that robots will make fewer casual mistakes than humans, the human mistakes will be at least random, and probably infrequently of a fatal type. That is to say, on any given day there are probably tens of thousands of anti-biottic prescriptions filled. If some chain store uploaded rev 2.2.4 robo-phamacist with a bug that replaced erythromycin with warfarin we could have a massive computer sized problem. On the other hand the chance of a human doing somthing like that more than once in their career is probably very likely but only a singleton event at each occurence.

likewise while a phamicist might give out the wrong dosage perscription occasionally, if any bulk bottle were miss barcoded the computer would give out the wrong prescription every time. Moreover there's a chance the human pharmacist would actually write on the bottle the wrong dosage thereby giving the patient a chance to catch the error. the computer would just repeat the same error on the bar code onto the bottle.

Now obviously the drug manufacturer's themselves face the problem of getting the right pills in the right bottles all the time and I would guess rarely make mistakes. That is all automated and it works, because they have the resources to do it right. But taking the automation down to the local level is scarey to me. In addition to the possibilty of mistakes there is also the opportunity for maliscious or crimminal attempts. E.g. someone wants to steal some oxy codon, breaks in and accidentally re-orders the bottles. A human phamacist would at least detect the problem.

finally what happens at the major hospital when the power fails, or the computer crashes or the mechanism jams in an earthquake? will they have enough staff available to be called in to do it all by hand. Frankly I doubt it, as similar medical autmation disasters have happened when record keeping systems and test result automation have broken down.

I'm not trying to be a luddite here. But not only is the severity of the issue much larger than say a broken robot on a GM car line but the financial pressure to prematurely implement such a system is very high in the medical community.

Re:Robots and mistakes

[Jennifer E. Elaan]

Barcode reading errors are extremely rare with a proper lase reader. The only way the robot would mismatch barcodes is if the bottles were labelled incorrectly, and that would be a human error. They still put cameras to catch such glitches though, so the pharmacist should still notice.

As for the question of resources, if these machines cost a half million dollars each, I'm willing to bet that they are nearly as good as what the drug manufacturers use.

I would expect, although I'm not sure, that a half-million dollar piece of medical equipment would be jam packed with batteries in case of a power failure. In the unlikely occurance of a jam, or a natural disaster taking down most of the electronics (this sounds rather like a straw-man argument to me), there has been a relatively good track record with extra staff. The recent network packet storm in a major hospital comes to mind, everyone was busy but service was not particularly degraded, even with the network completely out of commission.

Besides this, for a regular pharmacy, delays of a day or two to get a prescription filled are not unusual in some cases. Only in emergency cases is this a mission-critical system, and emergency drugs are usually available in places other than the main dispensory.

Re:Robots and mistakes

[Twirlip of the Mists]

As for the question of resources, if these machines cost a half million dollars each, I'm willing to bet that they are nearly as good as what the drug manufacturers use.

Don't forget, though, that drug manufacturers still have human beings sitting there watching every pill go by. The machines do most of the labor, but they have yet to replace human eyeballs for quality control.

Re:Robots and mistakes

[Jennifer E. Elaan]

These devices have cameras so that the pharmacist can do the same thing. This is designed more for large pharmacies that employ many pharamcists than the smaller ones. There is still a need for at least one human pharmacist.

Re:And what happens when

[chunkwhite86]

How so?

I'm not trying to sound anti-technology here, but can you name any computerized system with 100% uptime and 0% error rate?

While the DEC/Compaq/HP Himalaya with the non-stop kernel comes damn close, it's certainly priced out of reach for your typical drug store.

I think the solution would be to augment a human pharmacist with computerized/robotic equiptment to do his job more efficiently - but I think replacing the pharmacist all together is a mistake.

Just my 2 cents.

Re:And what happens when

[nneul]

Similar thing here - Walmart pharmacy gave my infant daughter "ear allergen drops" instead of an albuterol. Both were in labelled boxes on the shelf near each other.

Not only did they get the wrong box, but they put the prescription label for albuterol OVER TOP of the labelling on the box that said ear allergen drops.

Re:And what happens when

[Jennifer E. Elaan]

Read the article.

Towards the end it says they use a digital camera to take pictures of the filled bottles so the pharmacist can check that it's the right kind of medication.

And humans make plenty of mistakes. No human is going to get 9 9's, but designed to medical tolerance, these machines might. The phone company manages, after all.

Robotic System Errors

[msobkow]

I spent over two years working at NorTel's Bramalea site, which had robotic slide-lines for manufacturing surface-mount component boards. The equipment used tape hoppers of parts, which had individual components mounted on tape similar to a belt-fed machine gun. Once programmed, the robots themselves were flawless, doing exactly what they had been instructed.

The errors that typically cropped up after an assembly program was put in production were caused by good old fashioned human error: loading the wrong parts tape in a hopper (e.g. resistors with the wrong ohm value.)

I cannot see a robotic prescription-filler avoiding this problem. If someone fills the Atenolol (high blood pressure medication) hopper with Viagra, the robot is going to happily count out the correct number of the wrong pills, label them as Atenolol, and leave it up to humans to notice the error.

Cashiers/assistants (sometimes part-time high school or university students) usually hand the packages to customers, not the pharmacist. Even if the bottles aren't pre-bagged, the assistants are very unlikely to notice the pills are the wrong color, size, or shape -- they don't know what the pills are supposed to look like.

While I can see the benefit of a manually fed pill-counter device, all I can forsee from full automation are inevitable mis-filled prescriptions resulting in injury or death of the patients. This is one of the best examples of over-automation I've seen to date.

Some jobs just shouldn't be fully automated, even if we have the technology to do so.

Re:And what happens when

[Ashran]

> I'm not trying to sound anti-technology here, but can you name any computerized system with 100% uptime and 0% error rate?
Now show me a human with 100% uptime and 0% error rate.

"robot"

[mattdm]

Hmmm. Seems like grandiose use of terminology. The headline makes me picture C3PO standing there handing out medication, but in actuality, it sounds only slightly more advanced than one of those coffee vending machines at the bus station. Sure, it's a machine, but what exactly makes this a robot?

Around for Years - Pharmacist STILL required

[chicagothad]

I work for a large (Fortune 500) prescription benefit management company. We use automation in all of our mail order facilities, pumping out 60000 prescriptions A DAY. We have had this technology in place for many years and is the only way mail order pharmacies are cheap and effective. The processing has several quality controls, but at the end of the line you still have a human pharmacist checking the drug. This is required by state laws...And yes, even with automation and 7 checks by machinery including an automated image comparison the pharmacist STILL catches errors.

Re:Around for Years - Pharmacist STILL required

[stoops]

"And yes, even with automation and 7 checks by machinery including an automated image comparison the pharmacist STILL catches errors."

that's pretty surprising. if you assume these checks are relatively unreliable - say they're 98% accurate.

at 7 checks, you have an overall accuracy of 99.999999999872%. in other words, with these 7 checks, you're likely to miss an error once in 781,250,000,000 checks. if you assume that 1 in 10 prescriptions have en error in them, you'll fill about 7,8 trillion prescriptions before making an error. so if you're pumping out 60,000 prescriptions a day, you'll have an error once every 35,673 years.

in other words, with 7 automated checks, you shouldn't really have any errors. what can we conclude? the automation sucks a lot.

Re:Around for Years - Pharmacist STILL required

[FredGray]

in other words, with 7 automated checks, you shouldn't really have any errors. what can we conclude? the automation sucks a lot.

Alternative conclusion: the 7 checks aren't uncorrelated, as your calculation assumes. Each check could have a 99.9 percent efficiency, but that same 0.1 percent could get past all 7. If they're all basically asking variations on the theme of "is this a little yellow pill?" then this scenario isn't unlikely.

Shrug.

[Jennifer E. Elaan]

Was I the only one left thinking *shrug* so what?

Honestly, this is not the most impressive form of automation I've heard of, it was bound to happen and sounds pretty obvious to me. Warehouse management tasks (and that includes microwarehouses like in pharmacies) can be done quickly and accurately by machines, in a properly designed system.

This is news?

Good and Bad

[iiioxx]

One the one hand, the robot will probably count faster than old Mr. Mulligan at my local drug store, so I'll get my prescription in 5 minutes instead of 30 minutes.

On the other hand, the robot will probably count more accurately than old Mr. Mulligan, so I won't get the extra 5 or 10 pills that always seem to end up in my prescription bottle.

Mr. Mulligan counting: "15... 20... 25... 30... 25... 30... 35... I think they're making these bottles smaller. I remember when one of these would hold 50 pills... Damn Koreans! I fought a war so we could have decent sized pill bottles..."

I just hope they wait until he croaks before giving the robot his job. He's like 90 or something, it shouldn't be long.

Re:Read The Label

[BabyDave]

I happened to get the other guy's pills.

At least you got a guy's prescription.

[/me fondles his breasts]

Y'know, maybe this isn't so bad after all ...

Re:OK, I'll finally ask...

[phoebus1553]

Then you obviously don't know what a pharmacist really does. The 8 or so that I know don't count pills, that's what pharmacy techs do. The real pharmacist looks at the prescription and makes sure the doctor wasn't on crack when he wrote it.

A doctor has to know how the body works and know all the warning signs for about half a million diseases, and the tests for the signs. Are they supposed to know all about 3/4 of a million drugs, their side effects, their dangerous interactions with the other drugs the patient is taking, and the proper dosage?

I don't know how many times I've talked to the pharms' that I know that said some doctor prescribed something that was way too much/little for the patient's body or would've given them side effects that were worse than the origional problem.

Hell, if the town my Grandma's hospital was in had a decent pharm. maybe she'd still be alive, rather than the drugs (to treat the side effects of the drugs (to treat the side effects of the drugs)) causing her kidneys to rot out of her body.

From the trenches�

[(H)elix1]

This seems like a great idea, until I read the line - "Doctors type prescriptions into a pharmacy computer system and the orders send the robot into action." This implies a physician (or health care provider if you prefer a more derogatory term the insurance world uses) will actually use technology.... Not just a computer, but a handheld, tablet PC, something with a browser... the chances are slim to none with most of the baby boomers out there.

My first real world experience with CORBA was connecting a Cobol program on a mainframe to a Java application. The Java app would then shoot drug interaction and formulary data to be used by the Physician's winCE handheld. As a former biochemist that spent a serious amount of time working with MD/PhD's in a hospital, the app seemed like a killer idea - response time was fast, the data was secure, and the information was personalized. Then they tried to get them to use the thing. I understand 'you can lead a horse to water' on a whole new level.

Anyhow, there is the human factor as well. It is hard to beat a well designed smart system, but I like the idea of requiring two humans - one to order the drug, another to confirm. The physician spends a surprisingly little amount of time in the real sciences compared to a pharmacist. Different skill set (filtered on what I consider real science). As a side note, pharmacists were pretty good with technology - since they were using Power Builder thick clients and later web interfaces to deal with co-pays, adjudication, interaction warnings, etc.

Counting pills is not hard. Knowing that pill x with pill y scores you a role on the potion miscibility chart is priceless. A real easy fix to "has to read doctor's handwriting" problem? Get them to type the bloody prescription. Wait that would require them to use some funky software ranging from a typewriter to a computer thing...
nah... too hard.

Had this been a toolkit for a pharmacist, it would really take off.

Re:From the trenches�

[(H)elix1]

Granted, it was a sweeping generalization. I've had the pleasure to work with many MD's who showed me the Zen of statistics and coding tools to deal with drudgery. They are defiantly the expectation to the rule, however. The bitterness comes with knowing a desktop was given to many physicians and never even turned on.

I suspect this will improve over time too.

robots

[east coast]

They missed the boat on this one. I work for a mail order pharmacedical company that has been using robotics to do the same thing for many years (atleast the 7 that i've been there).
The most common misconception is that it replaces a pharamcist but by law (atleast here in the Pittsburgh PA area) the pharmacist still looks over the pills in the vial. But a pharmacists time is at a premium and machines are much more capable of doing the job of counting with less errors and faster than a human phar tech.
Overall our "defect" rate is on par with your mom and pop store but our RXs per hour count is much higher.

Re:robots

[GigsVT]

Yeah, I agree with you. I hang out with an older guy on IRC that owned several pharmacies in his town in PA. He got robots in several years ago. All it does is count pills from hoppers, there is still a lot of human checks involved. All it lets him do is now have so many workers counting pills by hand all day.

Something other people have pointed out, pharmacists are usually pretty tech-savvy. This guy had unix machines in his pharmacies to communicate with the different branches across town 20 years ago.

He's sold out to CVS within the last year and retired now... oh well, score one for big corporations.

Viagra trivia

[Raetsel]

• "...refills grandpa's heart medication with viagra..."

Funny you should choose that phrase...

Sildenafil citrate (Viagra) was orininally researched as a heart medication. Pfizer only found out about the, uh, uplifting side effects when their male test patients were extremely reluctant to return test formulations of the drug.

I think this one truly fits the definition of a "happy accident!"

Ok but

[TerryAtWork]

The robot is going to fill out the prescriptions totally perfectly until it fills one out totally wrong.

And sooner or later this is going to happen. UNLESS every pill has a teeny weeny little bar code on it - and I can see this happening - sooner or later there's going to be a Its-A-Wonderful-life level tragedy.

pill recognition

[chunkwhite86]

I think a necessary addition to this robotic scheme would be a small camera that could verify that the pills are indeed the correct pills.

Facial recognition systems do exist - it shouldn't be terrible difficult to have a small camera and computer that can make sure that the pills being dispensed look like the correct pills. This idea would of course just be another failsafe in the system - it wouldn't make the system infallible.

If your drug is supposed to be a little green square pill, the camera could just verify that the pills being dispensed match this description so that your fate doesnt rest in a bar code and a college-intern assistant.

Robot dispensers probably less error prone

[popocatapetl]

A pharmacist once made up a label for amoxicillin for my son (an infant then) with three times the regular dose - essentially listing the total daily input as a single dose. I only noticed it when I saw the medicine level in the bottle dropping faster than it should have for a 10 day course. Fortunately, amoxicillin is not toxic in high doses, but other drugs could have killed him.

Let's face it - neither humans not machines are perfect, but machines are less prone to error for routine tasks like prescribing medication.

My friend owns one of these robots

[jht]

I have a good friend who owns a large independent pharmacy in New England (I won't be more specific, because I'm not sure if he'd want it mentioned here - though his store and system have been profiled in both trade and general newspapers). He loves it - it handles his top 200 medications, and in his own paraphrased words:

"It lets the pharmacists (of which he is one) spend more time with the patients, and less time counting pills."

He is able to keep a couple fewer pharmacy techs on hand then he used to need for his volume, but it gives the pharmacists a nice assist. He worries less about mistakes - the sanity checks these machines have are a lot more reliable than a human's would be. He's told me he sleeps better at night, knowing the likelihood of a potentially fatal mistake is far lower because of the robot. He sees the role of a pharmacist as being to advise and dispense - with a strong emphasis on the first job.

Here's a true anecdotal story supporting robot usage: I take a Priloec every day. One time I went to my local CVS to get a refill - when I got home and opened the bottle I found Prozac instead (the prescription label was correct - they just filled it with the wrong drug). They took care of the problem immediately, but imagine if I wasn't bright enough to realize that those pills weren't mine (I joke that I'd still have had stomach trouble, but I'd have been happy about it). Fortunately, taking a Prozac per day wouldn't have killed me, but what if it had been something that could have?

That's where I can see robots helping the typical pharmacist. It'll prevent those sort of mistakes from happening, and ensure that the drug ordered is the drug given.

The point of this shouldn't be to. . .

[kfg]

replace the pharmacist, for all the reasons you've stated and more. The point is to automate one of the routine jobs of the pharmacist where mechanical means is less prone to error and removes an act of pure labor from the job.

The pharmacist should, as a matter of course, double check on the work of the robot, because even robots can make mistakes.

This isn't like replacing the pharmacist. It's like giving a ditch digger a backhoe to replace his shovel, or automating a daily incremental system backup so the admin can spend his time and attention somewhere more profitable.

KFG

I work for one of these companies...

[SplendidIsolatn]

McKesson Automation does this for hospitals. I should know, I work for them.

Hospitals like using robots in their pharmacy for a variety of reasons. Safety is reason #1. Barcodes are used, errors are almost non-existant. When there are errors, they are usually due to a human error at some point down the line. Second, it eliminates human need to do that task--freeing time and money.

In addition to dispensing drugs at the pharmacy level, we also make a product that sits at the nursing station level and does fundamentally the same thing dispensing the proper drug. We also have a nurse hand-held barcoding system to ensure the proper medication gets to the proper patient via barcoding -- it checks the nurses barcode, the patients barcode, and the drug barcode.

Check out the link above and go to 'products and services' to see what I'm talking about.

It's a Wonderful Machine

[dexter riley]

SCENE: Gower's Pharmacy, Bedford Falls

George Bailey: Mr. Gower?
Mr. Gower: Zzzz...ehh, whuzza meh damn kid...
George: Uh, you put the wrong pills in the vial...
Mr. Gower: Uh? Ehh, aska drugbot, mumble mumble.
George: M-mister Drugbot?
DrugBot3000: SLEEP MODE
George: You, you put the wrong pills in the bottle...
DrugBot3000: INVALID COMMAND, PLEASE REPEAT YOUR REQUEST
George: I said, you put the wrong pills in the bottle.
DrugBot3000: INVALID COMMAND, PLEASE REPEAT YOUR REQUEST
George: The pills! You put bad medicine in the bottle!
DrugBot3000: INVALID COMMAND, PLEASE REPEAT YOUR REQUEST
George: Why, you're nothing but a big fraud!
DrugBot3000: ENTERING SLEEP MODE
Gower: Zzzzz....
George: ...aww, screw this, I'm going to Martini's.

Please note that parent poster didn't say. . . .

[kfg]

the robot *wouldn't* make mistakes. He said the robot was less *likely* to.

This is really a pretty undeniable fact. Robots have become ubiquitous in manufacturing largly just because of this. Robots don't come in Monday morning hung over after a fight with their wife. Robots aren't thinking about the upcoming weekend while performing critical tasks. Robots don't even drink beer so they don't have to hide the can of the one they snuck onto the job in the new Chevy's door panel.

Do robots make mistakes? Of course they do. That's why quality control is always best handled by a person. Robots are better at labor, people are better at judgment.

A wise employer puts his people where they will do best, even if his "people" are machines.

You're also making the error of equating large, general purpose computing systems with small, specialized and embeded systems. These systems are *far* more robust than a general purpose computer. They're made that way. They can operate for decades without a single, *computer* related failure ( most failures of robotic systems come from old fashioned mechanical wear of the physical componants).

How much do such systems cost, can a pharmacy afford one? Well, if you look at the low end of such embeded robotic systems you'll find that every pharmacy already has one. It's called a thermostat.

In this case were're only talking about something on the order of sophistication somewhat lower than a desktop CNC mill. Those go for under $10K now.

KFG

Handwriting?

[nautical9]

Does this mean hand-writing-recognition has reached a huge milestone to actually understand what the doctor writes?

Re:And what happens when

[Twirlip of the Mists]

Having a robot doing the actual filling would eliminate another source of human error.

Maybe, but it would also remove a source of assurance. Doctors are human, too; sometimes they prescribe in error. Pharmacists are there to double-check prescriptions and dosages.

Electronic prescription systems, incidentally, increase rates of error rather than decreasing them. It's a lot easier to type a "2" instead of a "3" than to write a prescription incorrectly in longhand.

This is good

[simetra]

The goofballs at my pharmacy have yet to master COUNTING. The primary, base job of a pharmacist, or pharmacy intern, whoever it is, is to count pills. However, they somehow manage to screw this up. Although, once I got 100 pills for the price of 30 due to their incompetence, which saved me like $60! Heh, of course I didn't complain that time.

Responsibility

[kfg]

"If some chain store uploaded rev 2.2.4 robo-phamacist with a bug that replaced erythromycin with warfarin we could have a massive computer sized problem."

Which would be a *human* error of fairly massive proprotion involving multiple people over a chain of events.

You do realize you put your life in the hands of such systems every day already, don't you? Who do you think built your car? Who do you think controls the traffic lights? People? Get real. Robots and computers do it, and they do it better.

Ok, lets take it a bit closer to the exact subject at hand. *How do you think these drugs are made in the first place*? A couple of old crones stiring chemicals in vats with old 2x4's?

No, computer controled robotic systems make the drugs because it's proven *safer,* even on the grand scale.

If you want to worry, worry about the pharmacies giving you the right bottle with wrong drug in it because the computer at the manufaturing plant screwed up. Man, that would be a more widespeard disaster than just one chain having a bug. It would be world wide.

Hasn't happened. Relax. As long as people watch the machines the machines make you safer.

KFG

Re:And what happens when

[frank_adrian314159]

Maybe, but it would also remove a source of assurance. Doctors are human, too; sometimes they prescribe in error. Pharmacists are there to double-check prescriptions and dosages.

I assume that there will still be a human in the loop to check the prescription. With luck, that human will be able to do more checking and less counting.

Electronic prescription systems, incidentally, increase rates of error rather than decreasing them. It's a lot easier to type a "2" instead of a "3" than to write a prescription incorrectly in longhand.

And that's why well-designed systems of this type have checks built in to make sure that too much/little is not prescribed, drugs prescribed match the diagnosis, drugs in the proper form are prescribed, etc.

Uh oh!

[JoeCommodore]

I hope they get 'robot insurance,' I recall an ad where it was proven robots attack people to get their perscription drugs - which is what they run on....

Cool..

[grub]

.. so now I can watch my pharmacist perform Tai Chi whilst filling my prescriptions.

bar codes, eh?

[Hubert_Shrump]

some work from a sharpie and I'm rolling in crack for life.

wait...

I'm rolling in legal prescription narcotics for life.

"Bob! Bob! HEY Bob!"...

[tlambert]

"Bob! Bob! HEY Bob! So which hopper does the ``Estrogen'' go in?"
-- Anonymous 18 year old grocery clerk

-- Terry

European pharmacy system better than US

[baz00f]

I lived in Europe for a few years (about 10 years ago) and prescription drugs were primarily obtained in packaged bubble packs (20, 30, etc. pills), much like you get many OTC drugs in the US. The advantage, I believe, is better quality control at the delivery end: not having to rely on an overworked pharmacist to dump and count the correct pills. Without a PDR (Physicians Desk Reference) you just can't verify that the pills you get are the ones that should have been prescribed. The US way just seems very backward and labor intensive.

Plus when you get a factory sealed box of pills, you get the package insert and all the information about the drug compound and side effects etc. In the US you always have to ask for that - it should be mandatory!

Interesting fact: In Switzerland, pharmacists are licensed to identify toxic and edible mushrooms and verify them for you. So after your deep woods mushroom collecting, you stop by the pharmacy with your bag of shrooms and get rid of the bad ones before dinner. Don't remember what they say about psychoactive fungi.

You're misreading the article

[srmalloy]

Human pharmacists have always served as information resources for their customers, and even as a check on poorly-chosen prescriptions from doctors.

I can understand automating away the cashier or the janitor, but automating away a job where human judgment is so crucial is a terrible idea.

The problem is that pharmacies and patients don't use the same terms. When your doctor writes you a prescription, you take it to a pharmacy to get it filled. When the pharmacist takes your prescription script, they fill it and then dispense it to you.

When a pharmacy fills a prescription, what they are doing is to take the ordered quantity of the medication out of their stock and package it; when they dispense a prescription, they present it to the patient along with information about how and when to take it. Filling machines as described in the article don't take the place of a pharmacist -- they're taking the place of the pharmacy techs back inside the pharmacy who type up prescription labels, count out pills, and put them in prescription bottles. This means that the people working in the pharmacy don't have to take as much time preparing the prescription, and can spend more time with the patient.

Fill robots don't replace the pharmacist; what they do is eliminate the place where many medication mistakes occur -- selecting and measuring out the drug that the patient will receive. Drug manufacturers deliberately make pills and capsules with different shapes, colors, sizes and markings in order to help both the patients and the pharmacy staff tell them apart, but a pharmacy tech can still make a mistake and pick the wrong canister off the shelf when pulling a medication, and not notice that they've got the wrong little white pill, particularly when they've got lots of prescriptions to fill. A fill robot doesn't make that mistake; as long as the correct drug is in the correct hopper (and it's easier to make sure you've got the right drug when you're only touching the supply to put another 10,000 pills in the hopper, rather than for each of the 60-pill prescriptions that would be filled from that supply), the fill robot will always pick the correct drug.

And the situation is not as generally clear-cut as the article portrays. Where I work -- a major military hospital -- I am the manager for the pharmacy module of the medical information system at the hospital. The pharmacy has a large fill robot that processes refills; patients can either bring in a prescription for refill or use the phone- or web-based refill system to order their refills. Prescriptions entered for refill are processed by the main medical information system (checking to make sure that there are still refills available on that Rx, that the patient isn't trying to refill the Rx too soon, etc.), and then are sent to the fill robot, which fills and labels the refills, which are then distributed to the satellite pharmacies that the patients have selected to pick up their refills. The prescriptions don't need counselling or instructions, because the patient got those when they got the prescription initially.

A pharmacist's point of view

[Dominum]

I work in a retail setting and have actually seen and used one of these robots. This particular model was by Baker, and is a $125,000 piece of equipment. Most facilities implement one of these robots because of extremely high volumes, to assist in the mundane filling tasks associated with pharmacy. I am not aware of any Federal or State Regulation which allows robots to fill your prescription without a pharmacist on duty, so do not assume that automation of pharmacy means lack of supervision. The Baker unit I described above is sent information from a pharmacy computer, not from a doctor. At this point, I consider direct doctor-to-robot input sketchy at best, unless a prototype model. (Imagine your crusty old country physician firing up his Compaq iPaq and transmitting a prescription on his wireless network to the local robot pharmacy...yeah get the picture?) I know of towns in the rural southwest where the pharmacist literally rides into town once a week to fills rxs that day and that day only... But anyway, this particular robot holds about 180 drugs, holds the vials and caps, and actually fills, prints and labels your prescription. This model can fill about 100 rxs per hour and if it breaks, the ancillary staff has to pick up the slack. I like the analogy previously posted that describes the robot as giving a ditch digger a backhoe instead of a shovel. However, I am very concerned about how technology has affected pharmacy over the last 100 years. WE actually used to compound medication, not a pharmaceutical company in the MAJORITY of cases. Does this mean I want to go back to hand-rolling suppositories? (we used to do that too) No. All I ask is that the benevolent readers of /. be concerned that there is a licensed pharmacist on duty no matter WHO or WHAT fills your prescription. If not, then start to panic :-) Dom

Re:And what happens when

[0x0d0a]

the robot accidently refills grandpa's heart medication with viagra?

"Flood my IRC bot, will you? I'll hax0r your grandpa's pharmacist! We'll see who laughs last!"

Finally...

[BrodieBruce]

I can stop hearing my pharmacy school friends brag about making $70-80K a year in the US by counting pills on a daily basis.

Yeah...I know it's not a nice thing to say. But fsck that, and fsck the karma too.

Who checks the doctor?

[bluGill]

All well and good, but who checks the doctor. One advantage of a good pharmacist is they double check the doctors work. Pills often come in different dossages, it doesn't take much for the doctor to mix up and perscribe 5 500mg pills, when 1 500mg or 5 100mg is needed. The results of that can be deadly, so I want someone checking up. Rrobots can be programed to flag that situation, but if there are times when 5 500mg pills are needed then flagging it everytime will be an annoyance, while a pharmacist might know enough about the patient (without beind a doctor) to flag it as wrong only for her situation.

If done properly will be much safer.

[TheLink]

As long as the _people_ responsible for filling up the machine do their jobs properly, it is likely to be much less error prone for the same or lower cost.

In fact in addition to reading the label on the bottle and then using that to decide what drug to put in it, the robot could also print and paste a confirmation label on the reverse side. The person at the counter could then make sure both labels agree on the contents. So if the robot goes out of sync (same barcode but different drug/dose) or misreads the barcode there is a safeguard.

If they use the more expensive 2D barcodes they can do away with the extra print step because the 2D barcode has more space for error detection metadata.

This way it is VERY likely that what's printed on the label is what's inside. Whether it's the right drug is still between the pharmacist and the doctor tho.

But patients can then countercheck the prescription on the Internet e.g. www.rxlist.com

Since you don't have to fill the machine as often, more care could be given to the filling up- might even have more than one person to check and countercheck (just like filling up ATMs with cash). The trouble is some bosses are often tempted to skip counterchecks to save money. Or get some cheap person to fill the machine up, ignoring the great importance of that job.