Electronic Prescribing – What is Needed to Move Forward?

Both in the USA and in the UK there are major efforts underway to increase the amount of prescribing done electronically. Additionally, as reported a little while ago in an article here, the Commonwealth is moving to ensure that by early next year there will be no regulatory or legislative barriers to introduction of E-prescribing (EP) in Australia.

The reason adoption is sought is that it is now very good evidence that EP has the capacity to substantially reduce the frequency of Adverse Drug Events (ADEs) and that if this is achieved considerable human suffering and cost can be avoided.

The objective of EP is to deliver to the dispensing pharmacist a prescription from the prescriber that is error free and which can then be dispensed in an error free fashion to the patient. Subsidiary objectives may include maximising the use of generic medication, where appropriate (to also reduce treatment costs), and assisting the pharmacist with accurate and timely claiming of the costs of medication from either the Pharmaceutical Benefits Scheme or the patient depending on the circumstances.

Analysis of the prescribing process identifies three distinct phases (plus financial settlement and inventory management which will not be discussed further as these are simple business processes which are already quite highly automated and well understood.)

Phase 1 is Prescription Creation.

Prescription creation can be done electronically in a number of ways. The simplest is to use a computer to capture the patient details, and then from a data base of available preparations select the appropriate medications and package size and print out a form ready for manual signing. This is now quite widely done and offers the benefit of providing the pharmacist a legible prescription from which to work to dispense. This approach is grossly sub-optimal in 2006 as there is no error prevention or decision support contained in this basic model and the risk of errors in re-entry of the medication information in the dispenser’s system is real and demonstrable, despite the legibility.

The much to be preferred option is that the clinical encounter which is to result in an EP should be documented in an Electronic Health Record for that patient that will contain a range of relevant information to assist with the quality of the prescribing. The system ideally will have a list of the patients active diagnoses, the patients current problem, current regular medication (and ideally complementary substances taken) and basic physiological and other information such as height, weight, age, renal function and so on as well as such things as known allergies etc.

With this information available, and with access to tools such as the electronic Therapeutic Guidelines, as the prescription is developed the clinician not only is alerted to possible interactions and other contra-indications but can easily confirm the appropriateness of the therapy proposed. The literature evidence that well designed systems with these capabilities save lives and money is now unequivocal (see article on this site from a month or so ago).

Such EHR based systems are really the basic client the prescribing clinicians should be using in 2006, especially as the requirements for such systems and what is needed for utility and interoperation is well understood.

Phase 2 is Prescription Transport.

In essence once the quality prescription is created there are two possibilities for . Firstly the prescription can be printed out and given to the patient to present for dispensing or, if legally acceptable, a secure token, containing the prescription, can be given to the patient to be used by the dispenser’s system.

The simplest way of linking the prescriber and dispensing computers I believe is to have the prescribing system write a two dimensional barcode onto the printed prescription. Such barcodes can hold the full written information on the prescription and can be scanned into the dispensing system is less than a second or so. The dispensing system then displays the medications to be dispensed for checking and one the pharmacist is satisfied the dispensing and label printing is completed. The prescription is then marked as having been dispensed and if repeats are authorised the necessary paper work is created. This process benefits the patient (accurate communication with pharmacist) and the pharmacist (avoids the need for data entry). Fraud is prevented by using appropriate encryption of the information stored on the barcode to ensure the clear-text contents of the prescription matches the barcode.

The second approach would be to electronically transmit the prescription directly from the prescriber to the dispenser. This is done in the USA to an increasing extent. In Australia community pharmacy is very concerned that direct transmission might give the prescriber too much commercial influence and so the Pharmacy Guild is totally opposed to any proposal of this sort – despite its obvious attraction in a technical sense.

It seems likely that this problem is best overcome by the creation of a national “store and forward” EP Network. The patient presents at the pharmacy of their choice and provides the paper script or a token and the full prescription is then pulled down from the network. (The 2 dimensional barcode seems simpler and cheaper I must say)

This would seem likely to involve a greater cost than simple point to point messaging of barcoding and it would seem reasonable that any additional costs be bourn by the pharmacists as a whole.

A network of this sort would also have major privacy and security implications, as it creates a large database of essentially every prescription in the country, and may need to be very carefully thought through. Secondary use of this data may however provide some useful and clinically relevant information.

Phase 3 is Prescription Dispensing.

Community pharmacy has been using pharmacy systems for dispensing for many years and once an electronic copy of the prescription is available within the pharmacists computer dispensing and labelling can proceed as usual. The advantage for the pharmacist is avoiding any data entry and any possibility of error in the prescription assessment process. Of course the wrong pack can still be picked off the shelf but this is known to be very rare indeed in the presence of a clearly legible prescription.

Before concluding it is important to realise this article has oversimplified a number of aspects of EP. The medication terminologies to be used in prescription communication are still to be standardised and are important as is some remaining work on other aspects of EP Standards, not to say the excellent work has already been done via the MediConnect Trials, flawed though they were in their final implementations.

Standards Australia have already published work covering the use of HL7 messaging to transfer prescription information the it is important this work is progressed and utilized.

Essentially there a major community benefits to flow from a considered and comprehensive implementation of EP in Australia, there are very few significant barriers remaining to its progress. The work needs to be proceed as soon as possible under the agreed standards frameworks.

What is needed now is the availability of effective, quality EP client systems which provide reliable, consistent and safe decision support and the definition of the technical approach(s) to be adopted the get the prescription from the clinician to the pharmacist.

David.