Frequently Asked Questions

ACES uses natural language processing to group free text reasons for admissions into one of 80 syndromes. This is referred to as syndromic surveillance and these syndromes are not clinical diagnoses. More information about syndromic surveillance, ACES, and natural language processing can be found here.

Within the ACES Pandemic Tracker, four syndromes are monitored: pneumonia, chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF) and sepsis (see brief descriptions below). COPD and CHF are reported together.  These syndromes have been selected because they are known to be symptoms/complications of COVID-19 infection. It is expected that increases in admissions for these syndromes to higher than expected values could indicate the occurrence of community spread in a particular geography.

  • Pneumonia is an infection of one or both lungs that causes cough, fever, chills, and difficulty breathing. Signs and symptoms vary from mild to severe. Mild symptoms are similar to those of a cold.
  • COPD is a lung disease that causes a chronic cough, shortness of breath, and chest tightness. Emphysema and chronic bronchitis are the two most common conditions that contribute to COPD.
  • CHF occurs when the heart muscle doesn’t pump blood as well as it should. Fluid builds up around the heart and causes it to pump inefficiently.
  • Sepsis is a potentially life-threatening condition caused by the body’s overactive and toxic response to an infection. Sepsis occurs when the body’s response to an infection is out of balance, triggering changes that can damage multiple organ systems

The black line shows the 2020 7-day moving average for hospital admissions for the dates, LHIN(s), and COVID-19-related syndrome(s) selected. For comparison purposes, the teal line shows the 2018-19 historical moving average, along with lines for +1 and +2 standard deviations (yellow and maroon, respectively).

When the black line showing the 2020 data rises above the yellow and maroon lines showing the historical data +1 and +2 standard deviations, it is possible that this indicates that community spread of the novel coronavirus is occurring. One day of data above +1 or +2  standard deviations may be due to chance and is not evidence that community spread is occurring. There is strong evidence that community spread is occurring if the 2020 black line is consistently over the historical standard deviation lines.

Users can filter by LHIN and syndrome at the top of the graph, and by date range (at the bottom of the graph). Hovering over a date on the graph will provide the exact values.

Depending on filters selected in the graph, the map will display the corresponding current moving average by LHIN for the date selected below the graph and shown in the title. Any dates between February 1st, 2020 and the previous day can be shown. If a date after the previous day is selected, the map will show the most recent data available.

The map colour-codes LHINs based on whether the 2020 7-day or 30-day moving average is at or under 1 standard deviation above the historical moving average (teal) – Level 1; greater than 1 standard deviation, but less than or equal to 2 standard deviations above the historical moving average (yellow) – Level 2; or greater than 2 standard deviations above the historical moving average (maroon) – Level 3. Yellow flags possible high abnormal activity and maroon flags high abnormal activity.

Click and scroll with your mouse to zoom in and move around the map.

In syndromic surveillance, line graphs displaying real-time syndromic data over time offer a simple way to explore current trends and  compare to previous years (historical trends). Comparison to previous years can identify changes to the norm.

The calculation and plotting of moving averages (or rolling averages) allows for the smoothing of short-term fluctuations, so that longer-term trends are easier to recognize.

The 7-day moving average is the simplest approach to remove day-of-the-week variation. The first observation point in a 7-day moving average graph is the average of the first seven days. The second observation point is the average of day two to eight. This is continued so that each set of seven consecutive days is averaged.

The 30-day moving average, in contrast, removes day-of-the-month variation. The first observation point in a 30-day moving average graph is the average of the first 30 days. The second observation point is the average of day two to 31. This is continued so that each set of 30 consecutive days is averaged.

The standard deviation is a statistical measure that shows how much data vary from the average. If there is a lot of variability, the standard deviation will be high. If there is low variability, the standard deviation will be low. The standard deviation of a 7-day moving average, for example, is calculated by taking the difference of the data values on each of the seven days from the moving average for the 7-day period, squaring these differences and summing together, then dividing by 6 (the number of days minus 1) and taking the square root of the result. The standard deviation of the 30-day moving average is calculated in a similar way.

The historical moving average provides a comparison to the current trend. Even with the use of moving averages, there can be random variation in the data. Plotting the historical moving average and standard deviation (both +1 and +2 standard deviations) allows us to take this random variation into account in our comparisons and helps us decide if the current trend is higher than normal.

  • Normal variation is considered to occur when the current moving average is at or under 1 standard deviation above the historical average (Level 1).
  • A signal for possible high abnormal activity occurs when the current moving average is greater than +1 standard deviation, but less than or equal to +2 standard deviations above the historical average (Level 2).
  • High abnormal activity is signaled when the current moving average is greater than +2 standard deviations above the historical level (Level 3)

Hospitals share data with ACES in real time as patients are registered. In order to accurately compare current data to historic data, only full days of data are displayed in the ACES Pandemic Tracker. Each day at midnight, all admissions from midnight the day before to 11:59pm are added (i.e., at 12:00am on March 22nd, all records from 12:00am – 11:59pm on March 21st are added). At 12:30am the dashboard is updated to display the new data.

This historical data consists of ACES admissions from 2018 and 2019.  For each current day, the average of admissions for that day in 2018 and 2019 are used for comparison. For example, if there were 5 admissions for pneumonia on January 4th, 2018 for the Central West LHIN and 9 admissions on January 4th, 2018, the historical average for January 4th would be 7. As there is no historical data in 2018 and 2019 for February 29th, the historical average for February 28th is used.

Sometimes there are one or two days of data outages for a hospital. On the graph these will show as sudden dips in the historical moving average and standard deviations. These are to be expected and there is no reason to be concerned when the current data is above the historical data in these situations.

ACES captures hospital data from across Ontario. There are currently 10 hospitals in Ontario that do not share data with ACES:

  • Almonte General Hospital (Champlain)
  • Deep River and District Hospital (Champlain)
  • Groves Memorial Community Hospital (Waterloo Wellington)
  • Haldimand War Memorial Hospital (Hamilton Niagara Haldimand Brant)
  • Louise Marshall Hospital (Hamilton Niagara Haldimand Brant)
  • Norfolk General Hospital (Waterloo Wellington)
  • Palmerston and District Hospital (Waterloo Wellington)
  • Francis Memorial Hospital (Champlain)
  • Stevenson Memorial Hospital (Central)
  • West Haldimand General Hospital (Hamilton Niagara Haldimand Brant)

The ACES Pandemic Tracker uses two years of historical data (2018 and 2019) to compare with current (2020) data. In order to make accurate and valid comparisons, complete data for these two years is needed. This means that any hospital that does not have complete admissions data for 2018 and 2019 is excluded from the Pandemic Tracker. Hospitals excluded for this reason are:

  • Alexandra Marine and General Hospital (South West)
  • Brantford General Hospital (Hamilton Niagara Haldimand Brant)
  • Children’s Hospital of Eastern Ontario (Champlain)
  • Clinton Public Hospital (South West)
  • Georgian Bat General Hospital (North Simcoe Muskoka)
  • Grand River Hospital (Waterloo Wellington)
  • Grey Bruce Health Services – Lion’s Head, Markdale, Meaford, Owen Sound, Southampton and Wiarton sites. (South West)
  • Hanover and District Hospital (South West)
  • Headwaters Health Care Centre (Central West)
  • Mackenzie Health – Richmond Hill (Central)
  • Muskoka Algonquin Healthcare – Huntsville District Memorial Hospital and South Muskoka Memorial Hospital (North Simcoe Muskoka)
  • The Ottawa Hospital – General and Civic sites (Champlain)
  • Perth and Smiths Falls District Hospital – Great War Memorial and Smith Falls sites (South East)
  • Orillia Soldiers Memorial Hospital (North Simcoe Muskoka)
  • Sault Area Hospital (North East)
  • The Scarborough Hospital – General and Birchmount Campuses (Central East)
  • Seaforth Community Hospital (South West)
  • South Grey Bruce Health Centre – Chesley, Durham, Kincardine and Walkerton sites (South West)
  • Southlake Region Health Centre (Central)
  • Stratford General Hospital (South West)
  • Mary’s General Hospital (Waterloo Wellington)
  • Mary’s Memorial Hospital (South West)
  • Joseph’s Health Centre Toronto (Toronto Central)
  • Joseph’s Healthcare Hamilton (Hamilton Niagara Haldimand Brant)
  • Sunnybrook Hospital (Toronto Central)
  • West Parry Sound Health Centre (North East)
  • Winchester District Memorial Hospital (Champlain)

The tooltip on the map within the Pandemic Tracker shows the number of hospitals included for the LHIN region in the dashboard out of the total number of hospitals within the region.

The ACES Pandemic Tracker focuses on hospital admissions data. We believe that this data source will be the best signal to monitor for indications of covid-19 community spread and severity.

We do not include emergency department visits for a number of reasons. Visits to emergency departments have been steadily decreasing since mid-March as people heed social distancing warnings. People with covid-related symptoms are now being directed to local assessment centers which do not share data with ACES. Therefore, data from emergency department visits may lead to false assumptions regarding covid-10 presentations.

Admissions related to pneumonia, congestive heart failure, chronic obstructive pulmonary disease, and sepsis are being monitored as these are the types of presentations that are expected to increase when widespread community activity is occurring.