LEARN BUSINESS INTELLIGENCE

Building a Discharge Prescription Review Dashboard

A discharge prescription review dashboard should show each prescription once, with the medicine items and review information behind it. Keep prescription totals separate from item counts so readers can see what needs investigation.

Here’s how to combine the data, avoid duplicate rows and test the view against the prescribing system.

  
DEFINE THE WORKLIST

Decide Which Prescriptions Belong in the View

A pharmacy reader needs to know why a discharge prescription appears, which items need attention and where to investigate them. Start with that question before choosing a chart or building a count.

We’ll follow a fictional hospital pharmacy team reviewing discharge prescriptions. Its worklist includes current prescriptions with items confirmed as awaiting pharmacy review, or items whose review status is unavailable. It shows those reasons separately and counts each prescription once, even when both apply. The team uses the dashboard to find the work and returns to the prescribing system to investigate and record it.

An Electronic Prescribing and Medicines Administration system, often shortened to EPMA, records prescribing and medicine administration electronically. Its prescription statuses, item histories and review events depend on the system and local configuration. Map those fields with a pharmacy subject specialist and the reporting owner.

For this worklist, establish what identifies a prescription, how its current version is selected and which event records the required review for each item. Decide how cancelled prescriptions, discontinued items and replaced versions affect the view. Keep these rules together so the summary and detail describe the same work.

Separate prescription review from the wider discharge process. A review status describes the recorded pharmacy step; it does not by itself establish that the patient is ready to leave hospital or that every discharge task is complete.

Name the unit in the heading

“Prescriptions awaiting review” and “Medicine items awaiting review” answer different questions.

Keep the unit visible wherever readers see a total.

  
FOLLOW ONE PRESCRIPTION

Keep the Prescription and Its Items Distinct

Consider RX-101, a fictional discharge prescription on Ward A with three current medicine items. The statuses and flags below are illustrative reporting data, not a clinical ranking. The pharmacy team has defined which source events establish review completion.

Fictional prescription RX-101: three current medicine items
ItemCurrent review statusExample item flag
AReview recordedNo applicable flag—confirmed
BAwaiting reviewAnticoagulant
CAwaiting reviewControlled drug

The summary should show one prescription, three current items and two items confirmed as awaiting review, with no unavailable review statuses. It can also show that this prescription contains an anticoagulant item and a controlled-drug item. Those two flags describe RX-101; they do not make it two prescriptions.

Add a second fictional prescription, RX-102, on Ward A. It has two current items: D is confirmed as awaiting review, and E has an unavailable review status. The combined ward worklist should show two prescriptions, five current items, three items confirmed as awaiting review and one item with review status unavailable.

For RX-101, record the required review for item B in the test source. Once that change reaches the dashboard, RX-101 still appears once, with three current items and one item awaiting review. Its anticoagulant flag remains, because that describes the item rather than its review status.

If item C is then reviewed and no other rule requires attention, RX-101 leaves this example’s outstanding-review worklist. Its source record and review history still exist. Readers should be able to distinguish a prescription leaving the worklist from a prescription being deleted.

This gives you a small set of expected results to check before adding more wards or more flags.

Test two flags together

In the test data, give item B a second flag as well as its anticoagulant flag.

RX-101 should still count as one prescription, and item B should still count as one item.

  
PREPARE THE DATA

Combine Item Results Without Multiplying Rows

The source may store prescription headers, medicine items, review events and flags in separate tables. Joining those records can multiply rows: one item with two flags and several history entries may appear repeatedly.

Build the item result first. Use identifiers that distinguish the prescription, its relevant version and the current medicine item. Select the review information that applies to that item under the agreed rules. Counting joined history rows would answer a different question from counting items awaiting review.

Then produce one summary row per current prescription. Derive the item total from distinct current items. Keep separate counts for items confirmed as awaiting review and items with review status unavailable. For each prescription flag, determine whether any relevant current item has that flag, rather than adding another summary row for every match.

Keep the item-level detail available using the same prescription and version identifiers. A reader selecting RX-101 should see the three items that explain its summary, rather than a mixture of current and replaced records.

Recent changes need a separate definition. For example, a flag could mean that a relevant prescription change was recorded within the 48 hours before the extract time. Agree which changes qualify and state that time window clearly. The flag adds context; it should not silently exclude older prescriptions that still have outstanding review work.

Keep a prescription in this example’s worklist while any current item has an unavailable review status, even if no item is confirmed as awaiting review. Label the reason “Review status unavailable” and show that item count separately. A prescription meeting both reasons still contributes only once to the overall prescription count.

For category flags, distinguish a confirmed absence of applicable flags from unavailable information. Use “No applicable flag—confirmed” only when the source supports that result; otherwise show “Unavailable.” Resolve uncertain mappings with the source owner instead of treating missing data as a negative result.

A recent change is not a review

A change timestamp tells you that something changed. A review event tells you what was reviewed.

Compare their scope and timing under the local workflow instead of using one as a substitute for the other.

  
BUILD THE READER VIEW

Connect the Ward Summary to the Prescription Detail

Give the reader a ward view with separate prescription and item totals, followed by a prescription worklist. For each prescription, show its identifier, ward, separate counts for confirmed awaiting-review items and unavailable review statuses, and relevant category flags. The supporting detail should explain which items produce each count.

Your prescribing system may already provide a suitable discharge-review list. Use it if readers can find the right prescriptions and understand the results. A separate dashboard becomes useful when the required reporting fields need to be brought together or several readers need a consistent shared view.

When a shared dashboard portal helps

IntelliFront BI is self-hosted business intelligence software from ChristianSteven Software. It provides a browser-based place to build and share dashboards on infrastructure your organization controls, keeping reporting data within your environment.

For this example, its grid visuals can present the prepared prescription summary and supporting items. A Combo Box visual can filter dashboard items by ward. The dashboard filter setup guide explains that configuration.

Use the prepared prescription and item datasets to design the summary-to-detail journey. Start with both sample prescriptions in their initial states.

Selecting RX-101 should show only its three current items: two confirmed as awaiting review and none with an unavailable review status. Returning to the ward summary should restore the combined totals for RX-101 and RX-102.

Pharmacy review continues to be recorded in the prescribing system.

Show a reason for inclusion beside each prescription and use text alongside any color. A category flag helps a reader investigate the record; the pharmacy team determines its clinical significance and priority through its agreed process.

Include an information-age label with a clear meaning, such as the time the source extract was produced. Readers then have context if the prescribing record has changed since the dashboard collected its information.

A ward filter is a viewing choice

Use a ward selector to focus the display, and configure the reader’s authorized access separately.

Test content permissions and underlying data restrictions using representative reader accounts.

  
TEST THE COMPLETE JOURNEY

Check Reviews, Changes and Replaced Prescriptions

Before operational use, compare the view with the prescribing system alongside a pharmacy subject specialist. Use fictional or approved test records for changes and follow your organization’s rules for patient information.

Begin with both sample prescriptions in their initial states. Ward A should show two prescriptions, five current items, three items confirmed as awaiting review and one item with review status unavailable. Select RX-101 and confirm that its detail contains only A, B and C, with two items awaiting review and no unavailable statuses.

Record item B’s review and check that RX-101 now has one item awaiting review. Separately, record item D’s review on RX-102: that prescription should remain visible because item E’s review status is unavailable, with zero confirmed awaiting-review items and one unavailable status. The two-flags test should leave the prescription and item totals unchanged.

Next, make a prescription change that requires review under the agreed workflow. Check whether the source updates an item, creates a new item version or replaces the prescription. The dashboard needs to select the relevant current record and its applicable review information, so an old review does not incorrectly clear new work.

Include a cancelled prescription, a discontinued item and missing review information in the test set. Write down the expected treatment with the pharmacy team, then compare the view with those expectations. This tests the inclusion rules rather than assuming every source status has the same meaning.

Let a representative reader move between the ward summary and each prescription’s detail, identifying both confirmed review work and unavailable statuses. Check that changing the ward selection updates the intended totals and supporting rows together. If something disagrees, compare the underlying identifiers and version rules before changing the visual.

Keep the definitions, reporting owner and expected sample results with the dashboard documentation. Revisit them when the prescribing workflow or source fields change.

Start with one ward and a prescription whose summary you can explain item by item. If you need a shared portal for that view, use the same example to see how pharmacy teams can bring work queues and reporting together.

Contributors:

Christian Ofori-Boateng

Christian Ofori-Boateng

CEO, ChristianSteven Software