Challenges in e-prescribing

JAMIA's article "Transmitting and processing electronic prescriptions:experiences of physician practices and pharmacies" by Grossman et al. talks about the challenges pharmacies and physician practices face regarding e-prescribing. From my experience in retail pharmacy, it is pretty accurate in highlighting the current state of the art.

IT standards and interoperability are a major focus, as well as the human factors engineering that also present unique challenges.

One area concerns e-prescription renewals. While two way electronic transmissions eliminate a lot of handoffs, the loop is not entirely closed. Offices will habitually send new e-prescriptions in response to renewal requests, leaving the renewal request sitting indefinately in the pharmacy system queue. If the pharmacist doesn't manually correlate that the new prescription is related to the renewal, physician offices may be continually pestered for that particular renewal. If the request isn't directly acted upon, the pharmacy may then send additional faxes or phone calls, leading to frustration for both parties. When asked why this occurs, the diffusion of responsibility is no surprise:

"Practice respondents were more likely to attribute the problem to pharmacy staff being inadequately trained to identify new e-prescriptions, with problems reportedly diminishing as pharmacy staff gained e-prescribing experience. Pharmacists, however, believed that incoming e-prescriptions were easily identified and were more likely to point to physician delays in transmitting the prescriptions or, less commonly, transmission to the wrong pharmacy."

For pharmacists working in the community, this statement should be familiar:

"Three prescription fields commonly required manual manipulation- medication name, quantity, and patient instructions"

Each field represents a challenge from an IT/informatics perspective as to why it is not 100% accurate in autopopulation.

Medication Name

There are issues when a physician is both too vague or too specific in ordering a drug. If they are too vague, ie.) 'metoprolol', the pharmacist will need to clarify 'metoprolol tartrate' vs 'metoprolol succinate.' If they are too specific, ie.) 'Auralgan', the pharmacist cannot dispense generic 'antipyrine/benzocaine' without calling the physician. Aong with endless permutations of amoxicillin doses/dosage forms from an infinite amount of manufacturers, you being to see the problem. NDCs are not adequate in solving this dilemma. RxNorm, as a medication coding standard, seeks to strengthen the electronic clinical signal between pharmacists and physicians without hampering the dispensing operations (eg., third party, formulary, stock, etc.). RxNorm may or may not be the panacea. As the article says, we want to do a better job "...conveying physicians’ clinical intent without requiring them to over-specifytheir choice."

Quantity

Standard pack sizes and dispense quantities are necessary in order to address issues like '1 inhaler' vs '17 gm' of albuterol HFA. Prescribers do not usually know how drugs are dispensed. When a provider writes for '1' of epoetin alpha, is that 1 syringe? 1ml? 1 box? 10,000 units? The monetary difference could be thousands of dollars.

Patient Instructions/Sig

Sigs are usually free text and subject to much variability. Even '1QD' could be presented a bunch of different ways. Prednisone taper directions were always an issue no matter what system I used. NCPDP is trying to develop a set of codified sigs, but ambiguities still linger.