Inspired EHRs: Designing for Clinicians


E-Prescribing and
Computerized Physician Order Entry (CPOE)

E-Prescribing offers an opportunity to improve user accuracy and efficiency, a satisfying experience.

E-Prescribing (eRx) can be one of the most satisfying tasks of the physician because it saves the duplication of effort involved in hand-writing prescriptions, updating the list of medications, and including the changes in the office notes. E-Prescribing can also be frustrating when the EHR does not provide adequate data entry support. E-Prescribing offers the opportunity to reduce the medication errors that can result from pharmacists misreading prescriptions, dispensing an incorrect dose, or even prescribing the wrong drug because its name was similar to the name of the drug the physician actually intended.

The back-end process of e-prescribing sends discrete electronic data to a central hub, which then distributes the prescription message to the target pharmacy electronically (or via fax, if the target pharmacy lacks e-prescribing capabilities). The pharmacy can also send messages for renewal request to the prescribing physician. A new feature, not yet widely adopted, allows prescribers to send a message electronically to a pharmacy to cancel a previously prescribed medication or prescription.

Clinical Scenario — New prescription for Newly Diagnosed Diabetes

Mr. Martin is a 60-year-old construction supervisor. Three months ago, he was diagnosed with diabetes by Dr. Barnes, his family physician. Despite some healthy lifestyle changes, his weight is unchanged. His fingerstick blood sugars are improving, but are still too high at around 200. His goal is 80-140.

Dr. Barnes wants Mr. Martin to take a new medication named metformin to control his blood sugar. Together they look at the EHR screen and see that metformin is on Mr. Martin’s insurance formulary, and has the lowest-tier co-pay. Both are pleased. Mr. Martin wants to start with just a 30 day prescription from his local pharmacy in case he has any side-effects. The new prescription is sent electronically to the local pharmacy.

5.2 Computerized Physician Order Entry (CPOE)

A CPOE (also sometimes referred to as Computerized Provider Order Entry) is an electronic entry of patient care orders that electronically transmits itself to the departments (lab, radiology, etc.) or outside organizations that will fulfill it. CPOE orders can be distributed more quickly than their predecessors. They eliminate errors based on hand-writing, and can prevent duplicate orders by checking new orders against existing orders.

Clinical Scenario — Placing Future Lab Orders

Mr. Martin has achieved good control of his diabetes, blood pressure, and lipids. For the past year, he's been on stable doses of his medications, and his lab results have been stable as well. He can now settle into a more predictable routine, and won't need to visit the office or undergo lab tests as frequently. In about six months, Mr. Martin will need to come in for a hemoglobin A1c lab test, and then a visit. Another six months after that, he'll need to come in for further tests (another hemoglobin A1c a fasting lipid profile and a urine microalbumin test, both for his diabetes), and then another office visit.

5.2.1 Display Pre-Existing Orders to Prevent Duplication

A patient's EHR often contains unfulfilled orders. Patients forget tests, or can't find the time to get them done. If physicians can't see patients' existing future orders or recent lab results, they may accidentally order tests that have already been done or ordered, or very similar tests. This would duplicate their colleagues' work and spend resources wastefully.

When users place new orders, they can simultaneously be able to see the work that's already been done, without navigating away from their own unfinished orders and losing their work.

Figure 5.5 Before: Interruptive Dialog Box — Doesn’t allow the physician to see existing prescription orders
Dialog doesn’t allow the physician to see existing orders
Figure 5.6 After: Non-interruptive Dialog Box — Allows users to see existing orders using a separate panel, or by making the dialog box non-interruptive.
See Existing Orders in a separate panel, or a non-interruptive dialog box

5.2.2 Make It Easy to Find the Right Orders

Naming orders can be a challenge, because tests and procedures can have several commonly-used names. Different organizations may use different names for the same test. The physician ordering procedures might not be familiar with precise names listed in the EHR order catalog. The EHR might formally call a chest x-ray "XR chest". A physician, however, might look for it under:

  • chest x-ray (or variant spellings like “xray”)
  • chest XR
  • XR chest
  • X-ray chest
  • Chest x-ray 2 views
  • Chest x-ray PA and Lateral
  • CXR (fastest way to hand-write the order)

These are all correct ways to name a chest x-ray. Thus, the interface could allow physicians to find tests and procedures listed under their various commonly-used designations.

5.2.3 Preconfigure Orders with as Much Detail as Possible

Tests physicians order in ER or urgent care scenarios are often high priority and need done STAT. Routine tests physicians order in primary care settings will almost always be lower-priority. They'll need completed today, in the near future, or at some specified future time. An EHR that could be easily configured to a specific care setting, one that established default "priority" settings for the orders it processed, would be a powerful support tool.

Figure 5.7 Before: The physician needs to expend a lot of effort to fill in the many missing details. There are many opportunities for error.
More effort is required to fill in missing details
Figure 5.8 After: The EHR pre-completes key fields. Less frequently needed details are displayed less prominently.
Pre-complete key fields and Display less frequently needed details less prominently

5.2.4 Assign the Correct Diagnosis for an Order or Prescription

The EHR can make the problem and diagnosis lists readily available to physicians entering orders. It could also allow users to add new diagnoses on the fly, without having to exit the ordering tool to add them.

The EHR can provide clinical decision support by suggesting probable diagnoses based on patients' list entries and lab results.Some orders are almost exclusively associated with a single diagnosis, and in these cases the system could assign this diagnosis to these orders by default. For instance an order for a A hemoglobin A1c lab test will almost always be associated with a diagnosis of diabetes or hyperglycemia.

5.2.5 EHRs Can Adapt to Users, Not Users to EHRs

An adaptable EHR can let physicians “add to the shopping cart” (Figure 5.9), then “continue shopping” without checking out right away. This allows the physicians and patients to make preliminary decisions, and act on them immediately, but also allows them to make adjustments as the visit unfolds.

Figure 5.9 New Order Workflow — EHRs must offer users the option of building sets or collections of orders to facilitate their workflow.
Build sets or collections of orders to facilitate workflow

5.2.6 Building Groups of Related Orders Function like Checklists

EHRs must offer users the option of building sets or collections of orders to facilitate their workflow. This offers users personalized clinical decision support and frees them from having to depend on their memories or external reference materials to complete involved tasks. Well Child Visits, for example, follow predictable patterns. They involve immunizations at regular, predetermined intervals, specific counseling (anticipatory guidance), and set follow-up visits.

A four-month-old's Well Child Visit would include:

  • scheduling the next visit, at six months
  • administering several specific vaccines
    • hemophilus B
    • rotavirus
    • pneumococcal 13-valent
    • combination of diphtheria / hepatitis B / pertussis - acellular / polio / tetanus.

5.3 Summary

EHRs can:

  1. Make it easy for users to select new orders by offering a predictive search function that suggests appropriate results.
  2. Fill in the blanks with probable default settings (such as 30 or 90 day prescription supplies) where possible.
  3. Remember patients’ pharmacy preferences, and allow users to remove pharmacies patients no longer prefer from the list.
  4. Use preattentive attributes, such as color and typographic emphasis to enable physicians to find the results that are more likely to be relevant quickly. Use color iconography to graphically display medication renewal due-dates.
  5. Let physicians review e-prescriptions before sending so that they can correct any errors.
  6. Allow users to sort and filter medication lists to speed up the renewal process.
  7. Allow users to modify existing orders without forcing them to start over from scratch.
  8. Allow users to personalize the interface and build detail-rich collections of related orders that function like checklists.
  9. Allow users to select specific or flexible dates for new lab orders.

This book was last updated 10 Nov 2014.

The designs in this book were created by our team and reviewed by a national panel of clinical and human factors experts, but have not been empirically tested against existing designs.

Additional Resources

From the National Center for Cognitive Informatics & Decision Making in Healthcare

EHR Safety Enhanced Design Briefs:
Preventing Electronic Medication Order Errors (E-Prescribing)

  1. Choudhry NK, Fischer MA, Avorn J, et al. The Implications of Therapeutic Complexity on Adherence to Cardiovascular Medications. Arch Intern Med.2011;171(9):814-822. doi:10.1001/archinternmed.2010.495.
  2. Profile photo in interfaces by David Amsler