Modernizing Medication Management: Data-driven Approach to Pyxis MedStation

Delve into the significance of Pyxis MedStation in healthcare, highlighting its challenges and the data-driven solutions offered by Factspan. Discover how analytics improves medication management, saving costs and enhancing patient care in the process

If you’ve been around the healthcare industry, it’s likely that you would have come across Pyxis systems and the well-known Pyxis MedStation.

The need for precise and efficient medication management has always been paramount and the genesis of systems like Pyxis MedStation arose from this imperative, a response to the growing complexities of patient care and the increasing demand for accuracy.

BD defines the Pyxis MedStation as a leading automated dispensing system designed for decentralized medication management. It incorporates innovative communication and workflow optimization tools that assist pharmacy and nursing staff in their mission to ensure safe and high-quality patient care.

Pyxis MedStation can be likened to an ATM machine where the Nurses are our replacements and it works on a sign in technique, and it dispenses medicines specific to the patient requested for by the nurse.

Below is a representation of what Pyxis MedStation looks like.

                          Source: BD

Traditional to Innovation

To understand why Pyxis MedStation matters, let’s begin by understanding the complex medication distribution process through this graphical chart.

As is evident from the typical “Cart fill” process, This widely-used system presents challenges like first-dose delays, missing doses, billing inaccuracies, and time-consuming unused dose management.

The MedStation attempts to bridge all the above-mentioned challenges. Here’s a quick look at it:

Patient Care:

– Streamlines medication distribution, allowing more time for patient-focused care
– Ensures nurses have timely access to medications, boosting productivity and job satisfaction
– Increases safety by providing controlled medication access and real-time usage data
– Reduces medication errors through warnings and expiration date management


– Enhances safety for caregivers
– Provides real-time usage information
– Improves logistics and reduces time spent on medication management
– Offers medication error prevention with alerts


– Boosts efficiency and productivity.
– Improves the work environment and collaboration between departments
– Aids in compliance with regulatory requirements and hospital policies


– Helps manage costs related to patient hospital stays
– Facilitates data capture for decision-making and diagnosis-based reimbursement
– Helps manage costs related to patient hospital stays
– Facilitates data capture for decision-making and diagnosis-based reimbursement

Internal Analytics: A lot to be desired

It’s clear that the Pyxis Med station is valuable for medication management, but there is a lot of work to be done when it comes to internal analytics.

Let’s talk in specifics and take a couple of use cases that demonstrate its limited analytics capabilities.

  • Inaccurate Demand Prediction:
    The Pyxis inventory comprises over 2,000 different medicines, each with unique demand patterns. The internal analytics of Pyxis MedStation relied on moving averages and standard deviations, which often failed to accurately predict demand. This led to challenges like overstocking and frequent refills.
  • Inappropriate Periodic Automatic Replenishment (PAR) Levels:
    The PAR levels, representing minimum and maximum inventory thresholds, provided by the BD Insight tool were found to be inaccurate. This resulted in medication wastage and frequent refills, ultimately leading to financial losses due to expired medicines.
A chronological approach to Factspan addressing these challenges:
  • Selective Focus/Data Preparation:
    To manage the vast number of medicines effectively, we adopted a selective approach, focusing on the Top 50 medicines based on unit cost and consumption. This selection encompassed nearly 72% of the total medicine volume. We also introduced a dual level of granularity by considering both the medicine and the Pyxis machine station, creating a series ID combining both aspects.
  • AI/ML Implementation:
    We introduced machine learning (ML) techniques to enhance demand forecasting accuracy. Our ML model utilized the dispense volume for each medicine at a station as the target variable. These forecasts were then utilized to determine the PAR levels, using standard deviations and means specific to each medicine.
  • Experimentation:
    We rigorously tested various modeling techniques, including statistical time series like ARIMA, SARIMA and ensemble models, along with FB-Prophet and some  auto ML tools like Pycaret and Datarobot leveraging. External features such as holidays and derived features like  lags, trends, and rolling windows were incorporated to enhance the model’s capabilities.
  • Enhanced Accuracy:
    The final ML model, particularly for the Top 50 consumed medicines, achieved a significantly improved accuracy rate, jumping from approximately 55% (the previous method’s accuracy) to nearly 85%.
  • Presenting to the Pharmacy team:
    The results were presented via a Tableau dashboard, designed to be updated on a weekly or monthly basis according to the team’s needs. The model provides daily predictions, which can be aggregated to weekly or monthly levels as required.
  • Model Maintenance:
    To ensure the model remains effective and up to date, the latest data was retrained every 4 weeks and predictions are made for next 28 days and aggregated to next 4 weeks.
  • Business Success Criteria:
    The expected outcomes include annual savings of $3-$5 million and a reduction in the workload for the Pharmacy team. This translates to improved inventory management, with fewer expired medicines.

The Pyxis MedStation plays a crucial role as the backbone of medication management, ensuring the safe delivery of medications to patients. Its impact on patient care is undeniable, with pharmacists, nurses, and administrators depending on it for efficient and safe medication distribution.

However, there’s room for growth, particularly in harnessing the power of data analytics. The journey to fully leverage its potential is ongoing, with exciting developments on the horizon.

In the meantime, Factspan’s efforts are poised to make a significant impact on healthcare companies. Our focus is on enhancing the capabilities of systems like the Pyxis MedStation, translating data into actionable insights. As we move forward, we anticipate positive changes that will ultimately lead to better patient care and more efficient healthcare practices.

Most Popular
A Detailed Guide to Create a Snowflake Python Worksheet

How Can We Create a Snowflake Python Wor...

Streamlining ETL Pipeline with Snowflake...

Snowflake Cover image | Factspan

Stream & Merge for Incremental Load...

AI based Medical Imaging Tools 2024 | Factspan

Top AI Medical Imaging tools for Healthc...

Let’s Connect

    Work Email*

    Phone Number (Optional)


    Scroll to Top