Idaho’s coronavirus caseload is climbing — so how does the state see a downward trend?
Idaho’s daily coronavirus caseload has increased by more than 50% through Stages 2 and 3 of Gov. Brad Little’s reopening plan — yet the Idaho Department of Health and Welfare likely will report a “downward trend” in new cases Thursday as Little makes his decision on Stage 4.
It might seem nonsensical, but the finding is rooted in the department’s decision to change how it determines the trend for new cases.
IDHW told the Idaho Statesman on May 1 that it would use a 14-day moving average of new cases reported to determine Idaho’s trend, which produces “a more stable and reliable trend compared with looking at day-to-day fluctuations.”
Instead, the department is using those day-to-day numbers in the report it produces for Little to determine whether Idaho has met the criteria to advance through the stages of his reopening.
At the end of Stage 2, Idaho’s 14-day moving average had increased from 24.4 cases per day through Stage 1 to 29 per day — but the department identified a “downward trend.”
Through Tuesday, that average had climbed to 37.2. Yet, given the department’s methodology, it’s likely another “downward trend” will be reported this week when Little announces his decision on Stage 4. His press conference is scheduled for 10 a.m. Thursday and will be aired and streamed by Idaho Public Television.
In both cases, the number of cases during the 14-day evaluation window peaked in the ninth or 10th day, which created a downward trend when comparing just the daily numbers within those two weeks.
And in both cases, the trend line for the moving average was pointing up.
“The moving average is no longer being used because it was too complicated for many people trying to keep track, based on feedback we received,” IDHW spokesperson Niki Forbing-Orr wrote in an email. “We really wanted it to be something the public can understand, so public health staff are using case counts for that metric.”
The moving-average approach wouldn’t have prevented Idaho from moving to Stage 3, because the state’s criteria allows for moving forward even with increasing case counts as long as the positive testing percentage is below 5%.
The same is true this week. As long as the positive-test percentage remains below 5% — that data isn’t publicly available for the latest evaluation period — and the state meets the rest of the criteria, Stage 4 can begin regardless of case counts. The testing percentage was 3.37% during Stage 2, and it has been below 5% for the past eight reported weeks.
Here’s how the two methods produce different results:
(Note: For this story, we used IDHW’s daily numbers of confirmed and probable cases posted on the state’s coronavirus dashboard rather than the numbers we report each day based on health districts.)
May 13-26
IDHW considered only the number of cases per day within the 14-day evaluation period leading up to Stage 3. The case numbers spiked on May 22, but two of the lowest days of reported cases were the last two days of the period. IDHW applies a trend line — produced by a mathematical formula, or a spreadsheet app — that shows a slight downward trend.
It’s notable that these numbers have been changed from when IDHW produced its report (in epidemiology, data frequently is adjusted based on new information). The trend is still downward, but not by as much as IDHW showed in the gating criteria report on May 27. The big difference: That report showed more than 100 cases reported on May 22, but dozens of those cases apparently have been distributed to other days.
The 14-day average, meanwhile, smooths out the big changes from day to day. It’s also more connected to what happened in the prior weeks, which is why the average increased on the 13th day even though the case number was relatively low. Each new day replaces what happened 14 days prior in the formula, and 19 cases was an increase. The average dropped the last day but was significantly higher the last three days of the cycle than it was the first three days, and the trend line pointed up.
“The metric is calculated using the discrete 14-day period with a trend line fitted — to provide a look back at the previous 14 days,” Forbing-Orr wrote Wednesday. “For movement into each stage, the metrics are focused on what has happened during the stage we are currently in and each stage is a 14-day period rather than a comparison to previous stages.”
| Date | New cases | 14-day average |
| 5/13 | 31 | 24.3 |
| 5/14 | 27 | 24.0 |
| 5/15 | 38 | 25.3 |
| 5/16 | 30 | 25.6 |
| 5/17 | 17 | 25.5 |
| 5/18 | 19 | 24.9 |
| 5/19 | 21 | 24.9 |
| 5/20 | 30 | 24.9 |
| 5/21 | 28 | 25.4 |
| 5/22 | 61 | 27.9 |
| 5/23 | 31 | 28.3 |
| 5/24 | 39 | 29.9 |
| 5/25 | 19 | 30.3 |
| 5/26 | 15 | 29.0 |
May 27-June 9
IDHW likely will release its report for this date range Thursday, but we’ve done our best to predict the outcome based on the available data. There’s a collection of high case counts in the middle of the period with a drop-off the last few days, which creates a downward trend.
However, the 14-day average was higher Tuesday than it has been at any time during the governor’s reopening plan. That’s a clear upward trend.
| Date | New cases | 14-day average |
| 5/27 | 32 | 29.1 |
| 5/28 | 38 | 29.9 |
| 5/29 | 34 | 29.6 |
| 5/30 | 36 | 30 |
| 5/31 | 57 | 32.9 |
| 6/1 | 10 | 32.2 |
| 6/2 | 27 | 32.6 |
| 6/3 | 57 | 34.6 |
| 6/4 | 64 | 37.1 |
| 6/5 | 57 | 36.9 |
| 6/6 | 28 | 36.6 |
| 6/7 | 28 | 35.9 |
| 6/8 | 22 | 36.1 |
| 6/9 | 31 | 37.2 |
This story was originally published June 10, 2020 at 4:02 PM.