https://pharm-community.com/wp-content/uploads/wpforo/defaul...
предупреждающий предел (alert limit): Установленные критерии, заранее предупреждающие о возможном отклонении от нормальных условий, которые не обязательно ...

https://pharmadvisor.ru/term/action-limit/

--------------------------------------------------
Note added at 2 hrs (2025-04-14 10:37:59 GMT)
--------------------------------------------------

https://studfile.net/preview/6010925/page:2/
предел, требующий принятия мер (action limit): Установленные критерии, при превышении которых необходимо немедленное принятие дополнительных мер и корректирующих действий.

предупреждающий предел (alert limit): Установленные критерии, заранее предупреждающие о возможном отклонении от нормальных условий, которые не обязательно являются основанием для решительного корректирующего действия, но требуют дополнительного расследования.

--------------------------------------------------
Note added at 2 hrs (2025-04-14 10:42:18 GMT)
--------------------------------------------------

https://iris.who.int/bitstream/handle/10665/80681/WHO_IVB_11...
I.2.2
Quality control charts
In order to evaluate properly the data generated in a quality control programme,
values for each of the controls in each test run and test results should be listed on charts to permit a trend analysis. Such charts provide a visual representation of performance of the assay, enable comparisons to made, and clearly show any deviations, trends and shifts. A chart should be prepared for each control material. The most commonly used chart for internal quality control is the Shewhart control chart, also called the x chart, which is prepared in the following way.
The collection of data should be a continuous process, which should be done regularly
by capturing the raw data in the specified document (e.g. Excel, word data sheets or
a computer program). The test results should be obtained from the raw data for a
particular test/parameter for which trend analysis needs to be performed, for a specified
time period/interval and compiled data should be made available for analysis.
As a minimum, at least 20 values/data points of each quality parameter should be
collected and the mean, standard deviation (SD), mean and ± 2 and 3 SD values for this
data should be calculated. In case 20 data points/values are not available, the trends
may be established by using the maximum number of data points/values available.
These estimates may be revised on a regular basis or when more data is available and
new cumulative values should be used to revise the previously established trends,
until a minimum of 20 data points/values are completed. Subsequently, all the values
should fall within these established trends and the comments should be given,
based on these established trends.
The chart displays concentration on the y-axis (vertical) versus time on the x-axis
(horizontal). Horizontal lines are drawn for the mean and for the upper and lower
control limits, which are calculated from the standard deviation. The two upper
control limits are the mean plus three and two standard deviations (+3 SD and +2 SD)
and the lower control limits are the mean minus two and three standard deviations (-2
SD and -3 SD), respectively (see Figure I.2.1). The individual results from subsequent
assays are then plotted on the chart after every valid run.
The graph is an extension of a Gaussian normal distribution curve (bell curve) laid on
its side. Statistically, 95% of the control values are expected to fall within 2 SD of the
calculated mean value, and the data generated by the different runs should be randomly
distributed around the mean. Results falling outside the range of 2 SD (warning limit)
should be cause for alert, especially if the results are from two consecutive runs.
Any value falling outside mean ± 3 SD (action or rejection limit), but within the
specification, shall be considered as “out of trend” and should be notified immediately
to all the concerned staff.