Skip to content
English
  • There are no suggestions because the search field is empty.

Inpatient Occupancy Forecaster > Historical Analysis

Review how accurately your hospital’s inpatient occupancy forecasts have performed over time.

Location in SystemView: SystemView > Explore > Beds > Inpatient Occupancy Forecaster > Historical Analysis

In this article:


What is it?

The Historical Analysis component reviews the accuracy of past inpatient occupancy forecasts from the Live Inpatient Occupancy Forecast tool.
It compares predicted versus actual occupancy to show how reliable the forecasts have been across different wards, divisions, and timeframes.

Screenshot 2025-10-01 095949

Why it matters

Learn from the past to forecast smarter.

This component helps users validate the reliability of their inpatient occupancy forecasts.
It highlights where forecasting is most accurate and where improvement opportunities exist, ensuring hospitals can make the best possible use of their predictive insights.

Key benefits:

  • Understand the accuracy of historical occupancy forecasts.
  • Identify divisions or wards with the most variation between forecasted and actual occupancy.
  • Improve confidence in forecast reliability.
  • Support continuous model refinement and data quality improvement.

How to use it

Filter to focus your view

Use filters to narrow your forecast to a specific hospital, division, or ward:

  • Facility/Hospital – Select the facility you want to focus on.
  • Division – Choose one or more divisions.
  • Ward – Choose one or more wards.
  • Type – Filter by bed type (e.g. Standard Beds, Special Purpose Beds).
  • Forecast Days – Select the forecast days you'd like to assess.

💡 Tip: Start with the hospital or division level to view overall forecast accuracy, then drill down to specific wards for more detail.

Explore key performance metrics

Each tile presents a different view of historical forecast accuracy:

Tile Name What It Shows
Occupancy Forecast Metrics Overview of forecast accuracy across the selected timeframe.
Average Forecast % Distance from Actual Values by Weekday Compares forecast accuracy by day of the week.
Average Forecast % Distance from Actual Values by Forecast Days Compares how forecast accuracy changes with time into the future.
Average Forecast % Distance from Actual Values by Division Displays forecast accuracy for each division.
Average Forecast % Distance from Actual Values by Ward Displays forecast accuracy for each ward.
Forecast vs Actual Comparisons Visualises forecasted versus actual occupancy 1–4 weeks prior, showing how predictions aligned with real outcomes over time

ℹ️ Note: Depending on your SystemView environment and local scheduling practices, you may only see comparisons between actual occupancy and forecasts made one week prior. Sites that schedule activity further in advance may have access to longer historical forecast ranges.


How it works

The component measures how closely previous forecasts matched actual hospital occupancy.
It shows where forecasts have been most accurate and where there’s been more variation, such as by ward, division, or day of the week.
By reviewing these differences, hospitals can better understand the reliability of their forecasts and gain confidence when planning for future demand by:

  • Identifying where the model predicts most accurately (e.g., certain wards or weekdays).
  • Understanding the typical margin of forecast error over time.
  • Using these insights to strengthen confidence in the Live Inpatient Occupancy Forecaster.

How it helps you

  • Build trust in forecasting data: Validate how accurate your previous occupancy predictions have been.
  • Spot areas for improvement: Identify where forecast accuracy could be refined by division or ward.
  • Enhance planning confidence: Make future occupancy planning decisions informed by measurable accuracy data.
  • Support continuous improvement: Feed insights into future model updates and scheduling practices.

Best practices

How often should I use it?

What to Do How Often Who Should Do It Why It Helps
Review historical forecast accuracy across the hospital Monthly Bed Managers, Patient Flow Coordinators Monitor how well past forecasts have aligned with actual occupancy to improve forecasting reliability.
Analyse ward and division-level trends Monthly NUMs / Ward Leaders, Division Managers Identify areas where occupancy or forecast accuracy varies and plan operational improvements.
Compare historical data with current forecasts Monthly Hospital Operations Managers, Patient Flow Leads Validate model performance and strengthen confidence in ongoing forecasting accuracy.
Present findings in leadership or planning meetings Quarterly Executives, Hospital Operations Managers Support data-driven planning, service reviews, and continuous improvement discussions.

Pair with these components

Tips for success

  • Review both division and ward-level data to understand where forecast accuracy varies most.
  • Use weekday trends to identify patterns in forecasting differences (e.g., weekends vs weekdays).
  • Pair with the Live Inpatient Occupancy Forecaster to compare current predictions with previous performance.

ℹ️ FYI: Forecast accuracy naturally varies depending on the level of scheduling data available at each site. Hospitals with limited forward scheduling may see greater variance in long-range forecasts.


❓FAQs / Troubleshooting

Q: What does “Average Forecast % Distance from Actual Values” mean?
A: It shows how close the predicted occupancy was to the actual occupancy. Lower percentages indicate more accurate forecasts.

Q: Why do some wards show higher error percentages?
A: Forecast accuracy can vary based on admission patterns, discharge practices, and how far in advance activity is scheduled.

Q: Why do I only see comparisons between actual occupancy and forecasts made one week prior?
A: The length of historical forecast comparisons depends on your site’s scheduling practices. If activity is only scheduled around a week in advance, SystemView can only compare actual occupancy against the forecasts available for that period. Sites with longer scheduling windows may have access to extended historical forecast data.