Tag: EWMA

PRRS Exponentially Weighted Moving Average by State

This week,we are sharing a report by Dr. Juan Sanhueza from the MSHMP team regarding EWMAs by state.

Key points

  • Across states, different EWMA patterns continue to be observed.
  • The expected high PRRS incidence during fall/winter was not as marked both in duration and magnitude in some states during the 2018-2019 season.

Reminder: What is the EWMA?

The Exponential Weighted Moving Average (EMWA) is a statistical method that averages data over time, continually decreasing the weight of data as it moves further back in time.  An EWMA chart is particularly good at monitoring processes that drift over time and is used to detect small shifts in a trend.

In our project, EWMA is used to follow the evolution of the % of farms at risk that broke with PRRSV every week. EWMA incorporates all the weekly percentages recorded since the beginning of the project and gives less and less weight to the results as they are more removed in time. Therefore, the % of farms at risk that broke with PRRSV last week will have much more influence on the EMWA than the % of farms at risk that broke with PRRSV during the same week last year.

The charts depict: 1)the number of new cases (green dots – secondary Y axis) during a specific week and 2)the percentage of farms that broke during that week of the total in the MSHMP project in a smoothed way (blue line/Y axis). The red horizontal line indicates the threshold (upper confidence limit  – UCL). This UCL is calculated based on the average of cases during the lowest PRRS months in the year, June, July and August and is recalculated every two years.
When  there are more cases than expected, the blue line crosses the threshold (red line) indicating there is an epidemic.

The formula used in the EWMA chart is the following:

EMVA formula


where E is the smoothed % of infected herds, lambda the constant smoothing the curve, I the % of infected herds during that week and Et-1 is the smoothed % of  infected herds during the previous week.

EWMAs by state

Minnesota: As expected, this season EWMA crossed the epidemic threshold by the end of October/beginning of November 2018. However, the magnitude and duration of the epidemic was lower and shorter than in the 2017-2018 season. Incidence dropped during December 2018-January 2019 but lingered above the epidemic threshold until Mid-February, 2019.

Iowa: The 2018/2019 PRRS season started slightly earlier than in Minnesota and it was higher in magnitude than in any other of the assessed states. Although it reached a similar peak than in the 2017-2018 PRRS season in Iowa, its duration appeared to have been shorter since the EWMA went below the epidemic threshold during Mid-February 2019, which was about three months earlier than in the previous season.

North Carolina: As the 2017-2018 PRRS season, the epidemic begun on Nov of 2018. However, the 2018-2019 PRRS season was about three months shorter than the 2017-2018 PRRS season. The EWMA crossed the epidemic threshold momentarily in September/October 2018 but dropped below it until Mid-November when it crossed the epidemic threshold again to remain above it for about four months.

Oklahoma: Had a drastically different PRRS pattern than the 2017Ͳ2018 MSHMP season. The 2017-2018 PRRS season continued well into the summer, and only stayed below the epidemic threshold for about two months. PRRS incidence during the 2018-2019 season has been drastically lower than the one during the 2017-2018 season, staying at around 0.5%, and moving below the epidemic threshold several times.

Nebraska: The 2018-2019 fall/winter were characterized by a pattern of fewer, and more intermittent cases than the 2017-2018 fall/winter. No obvious PRRS season can be observed since PRRS outbreaks occurred sporadically throughout the year.

Illinois: Had a long PRRS season in 2017-2018 with the EWMA remaining mostly above the epidemic threshold for almost a year. In comparison, the 2018-2019 PRRS season started almost three months later in the year and its incidence has been lower than in 2017-2018, with the EWMA sporadically crossing above the epidemic threshold during the 2018Ͳ2019 fall and winter.

Science Page: Update on EWMA all versus EWMA original 13

This is our Friday rubric: every week a new Science Page from the Bob Morrison’s Swine Health Monitoring Project. The previous editions of the science page are available on our website.

This week,  we are sharing an update on the EWMA comparison by the MSHMP team.

Key points:

  • Even though small differences between both EWMAs exist, the EWMA of the original 13 participating systems is still a good indicator of the overall PRRS EWMA.
  • Questions from participants are always welcome and help us to provide answers and insights to all of you.

REMINDER: WHAT IS THE EWMA?

The Exponential Weighted Moving Average (EMWA) is a statistical method that averages data over time, continually decreasing the weight of data as it moves further back in time.  An EWMA chart is particularly good at monitoring processes that drift over time and is used to detect small shifts in a trend.

In our project, EWMA is used to follow the evolution of the % of farms at risk that broke with PRRSV every week. EWMA incorporates all the weekly percentages recorded since the beginning of the project and gives less and less weight to the results as they are more removed in time. Therefore, the % of farms at risk that broke with PRRSV last week will have much more influence on the EMWA than the % of farms at risk that broke with PRRSV during the same week last year.

EWMA vs EWMA 13

Results from this year’s comparison

EWMA 13 is still a good representation of the overall EWMA. The reason that the EWMA 13 is still representative may be because they cover a wide area of the States and they still represent a high percentage of the final EWMA. A minor difference occurred in 2017’s summer as some farms of the 13 experienced outbreaks. However, as we have discussed in previous science pages each state or region seems to have a different EWMA pattern.

Last year comparison of the EMWA

Science Page: PRRS EWMA analysis for years 2009 – 2018

This is our Friday rubric: every week a new Science Page from the Bob Morrison’s Swine Health Monitoring Project. The previous editions of the science page are available on our website.

This week, we are sharing a report by the MSHMP team regarding the EMWA analysis for years 2009-2018.

Key points:

  • The  EWMA chart is a smoothed chart of the percentage of farms that are breaking.
  • Newly added farms to MSHMP increase the denominator therefore diluting the estimate which affects the EWMA chart giving the impression that PRRS season has changed.

Reminder: What is the EWMA?

The Exponential Weighted Moving Average (EMWA) is a statistical method that averages data over time, continually decreasing the weight of data as it moves further back in time.  An EWMA chart is particularly good at monitoring processes that drift over time and is used to detect small shifts in a trend.

In our project, EWMA is used to follow the evolution of the % of farms at risk that broke with PRRSV every week. EWMA incorporates all the weekly percentages recorded since the beginning of the project and gives less and less weight to the results as they are more removed in time. Therefore, the % of farms at risk that broke with PRRSV last week will have much more influence on the EMWA than the % of farms at risk that broke with PRRSV during the same week last year.

prrs_emwa_analysis

MSHMP report chart 4 depicts: 1)the number of new cases (green dots – secondary Y axis) during a specific week and 2)the percentage of farms that broke during that week of the total in the MSHMP project in a smoothed way (blue line/Y axis). The red horizontal line indicates the threshold (upper confidence limit  – UCL). This UCL is calculated based on the average of cases during the lowest PRRS months in the year, June, July and August and is recalculated every two years.
When  there are more cases than expected, the blue line crosses the threshold (red line) indicating there is an epidemic.

The formula used in the EWMA chart is the following:

EMVA formula
where E is the smoothed % of infected herds, lambda the constant smoothing the curve, I the % of infected herds during that week and Et-1 is the smoothed % of  infected herds during the previous week.

If different smoothing factors are applied to the MSHMP data this would generate different trends and then we would place the threshold based on the sensitivity
that we consider that signals an epidemic.

EMVA graph with different parameters

Has the incidence of PRRS changed?

One possible reason the EWMA % of cases decreasing might be that the number of farms that are breaking expressed as a percentage is less. This can be due to the fact that the total number of farms sharing PRRS status has been increasing and these new farms might have a lower underlying incidence.

 

Science Page: EWMA by State

This is our Friday rubric: every week a new Science Page from the Bob Morrison’s Swine Health Monitoring Project. The previous editions of the science page are available on our website.

This week we are sharing a report by our MSHMP team regarding the evolution of the EWMA by state. The states of Illinois, Iowa, Minnesota, Nebraska, North Carolina, and Oklahoma are presented in today’s analysis.

Reminder: What is the EWMA?

The Exponential Weighted Moving Average (EMWA) is a statistical method that averages data over time, continually decreasing the weight of data as it moves further back in time.  An EWMA chart is particularly good at monitoring processes that drift over time and is used to detect small shifts in a trend.

In our project, EWMA is used to follow the evolution of the % of farms at risk that broke with PRRSV every week. EWMA incorporates all the weekly percentages recorded since the beginning of the project and gives less and less weight to the results as they are more removed in time. Therefore, the % of farms at risk that broke with PRRSV last week will have much more influence on the EMWA than the % of farms at risk that broke with PRRSV during the same week last year.

Key Points of the report:

  • Different states continue to have different EWMA patterns.
  • Even though winter is the high risk season, biosecurity measures should be enhanced during the whole year.
  • The state of Minnesota suffered two unusual peaks, one in spring and another one during the summer.

EWMA patterns by state
EWMA patterns by state

Science page: Comparing EWMAs

This is our Friday rubric: every week a new Science Page from the Bob Morrison’s Swine Health Monitoring Project. The previous editions of the science page are available on our website.

But first, we would like to congratulate our 4th-year student Hunter Baldry for receiving the ZinPro scholarship in recognition of her accomplishments as a food-animal student at the University of Minnesota. Keep up the good work, Hunter!

This week, the Science Page answers one of your questions: Is the trend of PRRSV outbreaks recorded for the original 13 participants still related to the PRRSV outbreaks evolution monitored for all the MSHMP participants?

Key points from this week edition:

  • The EWMA of the original 13 participating systems is still a good representation of the overall EWMA.
  • Questions from participants are always welcome!

Reminder: What is the EWMA?

The Exponential Weighted Moving Average (EMWA) is a statistical method that averages data over time, continually decreasing the weight of data as it moves further back in time.  An EWMA chart is particularly good at monitoring processes that drift over time and is used to detect small shifts in a trend.

In our project, EWMA is used to follow the evolution of the % of farms at risk that broke with PRRSV every week. EWMA incorporates all the weekly percentages recorded since the beginning of the project and gives less and less weight to the results as they are more removed in time. Therefore, the % of farms at risk that broke with PRRSV last week will have much more influence on the EMWA than the % of farms at risk that broke with PRRSV during the same week last year.

Take a look at the original 13 and overall EMWAs.