Realizing E-Prescribing’s Potential to Reduce Outpatient Psychiatric Medication Errors

Despite the shortcomings of e-prescribing in the ambulatory setting, a growing literature indicates that e-prescribing may reduce the overall rate of medication errors. However, the current structure of outpatient psychiatric practice poses unique challenges to widespread implementation. Most outpatient psychiatry is delivered in non–clinic-affiliated, private practice settings, where an e-prescribing system may be perceived as too expensive, cumbersome, and complex to warrant its use. Furthermore, in a 2012 survey (9) of U.S. outpatient psychiatrists, 26% of respondents reported either not using or not feeling comfortable using electronic devices for clinical tasks, including e-prescribing. Therefore, any effort to increase the use of e-prescribing systems needs to account for up-front and maintenance costs as well as the technological literacy of the prescriber.

In contrast to the sleek, streamlined design of many popular consumer social apps (Facebook, Instagram, etc.), e-prescribing interfaces are often cluttered, text heavy, and redundant. Many systems include complex field-based entry formats, which require careful attention to detail to prevent erroneous entry or omission of important information. Possible solutions include replacing prompts with clear graphical user interfaces, integrating required drop-down menus (for route of administration and units of strength), and autocompleting a medication’s administered amount (calculated by the frequency and number of days prescribed). Revision of automated error alerts, such as pop-up boxes to guide and confirm potentially harmful medication interactions or abnormally high medication strengths, may also prevent these types of errors. However, these automated alerts need to be balanced against “pop-up fatigue” in which prescribers may habituate to and ignore frequent computer-generated warnings.

Likewise, e-prescribing systems should integrate design features to simplify medication reconciliation (matching what the patient is currently taking against the medical record), such as including the computerized medication list in the “plan” part of the note template. Maintaining accurate, computerized medication lists is a prerequisite for facilitating cross-talk between e-prescribing and retail pharmacy systems; a rigorous, collaborative initiative may be required to improve the quality of outpatient medication reconciliation.

A detailed analysis of system- and user-specific medication errors related to the use of e-prescribing has been presented elsewhere (7). Here, the study of human factors—that is, understanding interactions between humans and technology—can improve e-prescribing systems. For instance, a human-factors design approach would systematically assess prescriber and pharmacist current practices, needs, and priorities; e-prescribing system capacities; and the areas of mismatch between the two.

Patients often seek care from multiple clinicians who may use e-prescribing systems that fail to communicate with each other. This fragmented system of care increases the risk of medication duplication and coadministration of various medications. In this context, Pandolfe et al. (10) proposed to augment health information exchanges (HIEs) to include a “patient-adjudicated” medication list. This list can be managed through a centralized, digital medication database that maintains an accurate list of the patient’s medications. Within this digital ecosystem, prescribers, hospitals, pharmacies, and the patients themselves could view and modify the list on an ongoing basis. This HIE functionality, if implemented, would improve transparency between various e-prescribing systems for a given patient and reduce harmful drug-drug interactions. Patients who have limited facility with or access to technologies, at the very least, should be encouraged to select one pharmacy instead of having customer profiles with several retail chains. If possible, patients with specialty care needs, such as comorbid psychiatric and medical conditions, should work closely with community pharmacists specialized in these fields (for example, pharmacists with HIV-focused practices for patients with comorbid HIV and a psychiatric disorder).

Although prescribers can confirm that an e-script has been transmitted to and even received by an external pharmacy, they currently cannot view the patient’s list of medications or the number of refills remaining. Kaiser Permanente and the Veterans Health Administration system are two integrated health systems that integrate prescriber and pharmacy records; however, these systems are challenging to replicate among independent and often competing entities. Solutions include enabling prescribers to view the most current list of medications for a patient in their retail pharmacy’s system (and vice versa), allowing for real-time chat between prescribers and pharmacists (reducing the need for time-consuming phone calls), and allowing for e-discontinuation (thus ensuring that medications are stopped).

Likewise, most e-prescribing systems do not allow for e-prescription of controlled substances (EPCS), such as psychostimulants (schedule II) and benzodiazepines and hypnotics (schedule IV), which are commonly used in outpatient psychiatric practice. Although the Drug Enforcement Administration issued a ruling in 2010 that allows for EPCS, stringent security mandates have limited implementation even among those who use e-prescribing tools.

Implementation of these changes in outpatient psychiatric settings will require targeted incentive and penalty programs, like those initiated by the Centers for Medicare and Medicaid, and tied to insurance payments. Effective advocacy can push the government to enact changes that will pressure the behavioral health care industry to adopt bidirectional e-prescriptions and enhanced pharmacy-provider communication. In addition, all stakeholders—patients, payers, prescribers, and pharmacists—will need to advocate for these changes. Once changes are enacted, outcomes can be measured before and after implementation by comparing the number of medication orders, number of communication errors, medication adverse effects, physician and pharmacist call volume, and administrative burden.

Education for providers around technological literacy and e-prescribing implementation can further bolster demand and advocacy. Such educational efforts should start during residency training. However, the Accreditation Council for Graduate Medical Education has not established general psychiatry milestones that explicitly address clinical informatics (including e-prescription) or bidirectional pharmacy-to-prescriber communication skills. Inculcating such skills may prevent safety mishaps when transmitting e-prescriptions.

Notwithstanding the potential benefits of e-prescribing in outpatient psychiatric settings, several limitations persist. First and foremost, evidence supporting the benefits of e-prescribing specifically in outpatient psychiatric settings is limited; outpatient studies have so far been restricted to nonpsychiatric settings. Therefore, we advise gradual and careful implementation of e-prescribing systems in outpatient psychiatric settings on a case-by-case basis. Second, increased interoperability among e-prescribing systems may reveal too much private information to patients’ psychiatric and nonpsychiatric providers. Although these improved lines of communication may reduce medication errors (duplicate medications, harmful drug-drug interactions), some patients may perceive an all-inclusive e-prescription system as an invasion of privacy for exposing their psychiatric care to their nonpsychiatric medical providers. A potential solution consists of adding a layer of security in the record in regard to psychotropic medications, akin to “break-the-glass” digital firewalls embedded in many EHRs. Third, it is unclear whether expansion of state-run prescription drug monitoring programs (PDMPs), which currently track only controlled substances, would confer similar benefits to those of an interoperable e-prescription system. However, reliance on a PDMP would require cross-referencing an external database, leaving room for yet another source of communication error and increasing pharmacist and prescriber burden. Fourth, up-front costs of e-prescribing implementation prevent many practices, especially small groups and solo practitioners, from using these systems. State and federal mandates to use e-prescribing systems should be accompanied by financial incentives or assistance to overcome this barrier to entry.

Well-designed, controlled studies comparing e-prescribing with traditional prescription practices in outpatient psychiatric settings will provide useful information about the benefits and limits of these emerging technologies. However, preliminary evidence from inpatient and nonpsychiatric outpatient settings already demonstrates that e-prescribing is associated with decreased medication error rates. By gradually replacing paper- and phone-based medication prescriptions with more robust, better designed e-prescribing systems for psychiatry, we may address medication safety issues, medication adverse events, pharmacy-prescriber communication problems, and administrative burdens.

This work was supported in part by grant R25-MH060482 from the National Institute of Mental Health to Dr. Hirschtritt.