Statistical Medication Prescription Review (SMPR) systems are used by people in several different roles:

  • Pharmacist Prescription Reviewers
  • Pharmacy Administrators and Leaders
  • Knowledge Managers for SMPR Systems
  • Technical SMPR System Administrators

It is crucial to take a role-based approach to the SMPR system user experience.

To begin that work, next consider an early “view” for an SMPR system meant for pharmacist prescription reviewers.


When doing prescription review, especially for the medication prescriptions that do NOT receive an “OK” from the SMPR system, a reviewing pharmacist could be presented with a view like this:

View of reviewing a new medication prescription for the drug amlodipine besylate.

In the view above, outputs from all of the major technical capabilities of SMPR systems are represented in the panel of 10 dashboard widgets below the details of the amlodipine besylate prescription. In essence, this view for the pharmacist prescription reviewer is a “dashboard of smaller dashboards.”

Looking at those 10 small panels starting in the upper left there are scales with medication prescription scores risk (low), complexity (very low), and typicality (high). To the right of those three scales are two panels comparing the dosage and frequency to a sample of 480 past amlodipine besylate prescriptions for similar patients. To the right of those is the Past Exposures panel. This panel indicates visually that the patient has been using amlodipine besylate at home for many years and has also been exposed to this drug during two prior hospital visits.

Next, along the bottom starting in the lower left is the Profile Fit readout indicating a good fit between this patient’s profile and this new amlodipine besylate prescription. There is a very high likelihood (93%) that this medication prescription would have been created for this patient. In the Rule Checks panel to the right, a whole battery of rule check results are summarized. Rule checks surfaced no issues with allergies (A), drug-drug interactions (DD), drug-food interactions (DF), drug-disease complications (Dx), pharmacogenomics (PGx), or pharmacy policies (P). Below the Rule Checks a computer-generated dose and frequency are given. Further to the right, some details about the origin and prescribing process are provided. The reviewing pharmacist can see this order was pre-built in the EHR and not typed anew by the prescriber. The order was not part of an order set. Also, the prescriber did not add any special comments to the order. Below these details of the prescribing process the reviewing pharmacist gets categorical information about the prescriber’s experience prescribing amlodipine besylate tablets. Finally, in the lower right corner the results of using a predictive model show the expected effects of this medication prescription on the patient’s systolic blood pressure (SBP). This patient’s SBP value is expected to decrease by about 4 mmHg after 3 days of exposure.

Almost none of the information shown in this SMPR system view is provided to reviewing pharmacists today. Most, if not all, of this information is needed to compare the outcomes of clinical versus statistical review of medication prescriptions in scientific studies.


Pharmacy shift leaders, managers, and directors need access to a variety of controls for setting the parameters that govern the operations of their SMPR systems. The main point of this design concept is to highlight the likely need for SMPR systems to operate with varying thresholds and modifiable capabilities throughout the course of a seven-day week. For example, SMPR systems may need to operate in a different way overnight on third shift or on the weekends than they do for weekday operations.

View of SMPR system settings spanning three timeframes over a week for one pharmacy

In the view above, rather than patient information, what the pharmacy administrator sees are three global settings for an SMPR system above and three longer lists of other settings associated with different parts of the week below.

The three global settings are straightforward. The SMPR system can be turned on or off, automated review enabled by the system can also be turned on or off, and pharmacy administrators can decide whether or not automated review requires every field in a medication prescription to be complete.

The twelve time-based and shift-based settings are more detailed. In the diagram, starting on the left there is a panel of settings that govern how the SMPR system functions Monday through Friday from 0730 to 2330. Not only can the pharmacy administrator operating the system turn specific SMPR system capabilities on or off they can also set thresholds for the capabilities that are turned on. These thresholds determine how automated review using the SMPR system will play out. For example, in the left-most panel the Risk scoring function is turned on and it is set to pass new medication prescriptions with risk scores less than or equal to 41 on a scale of 100 so long as all other checks are also passed. Likewise, for the five alert capabilities that are operable during this timeframe (allergy, drug-drug, diagnosis, genetics, and policy alerts) a new medication prescription will only pass if there are zero alerts in all five of these active alert categories.

Finally, looking carefully at the differences in the three time-based settings panels one can see why it may be very important to allow pharmacy administrators to have this granular control over an SMPR system. Notice how drug-food alerts may only be relevant on the weekends when perhaps expert clinical nutritionists are unavailable. Also notice how thresholds change so that, for example, on third shifts when new medication prescriptions are much fewer the SMPR system is configured only to automate the review of the very lowest risk and basic ones.

Allowing pharmacy administrators to have exquisite control over SMPR systems using familiar user experience feature designs from tablets and smartphones is the priority here.


SMPR systems are hybrid systems that use large production rule bases combined with a large number of statistical algorithmic or machine-learning models. Expert knowledge managers control and maintain these computable knowledge assets ensuring all rules are up-to-date and all models are properly calibrated.


SMPR systems are complex technical systems. Expert system admins keep SMPR systems secure, integrate them with other systems, and make sure they are always operating properly.