Table business rules

Missing references in WinNonlin
Intext Table Template tab
Table data display using decimal alignment
Table splitting
Display of different sample sizes in intext tables
Significance level display in statistical tables
Footnote about model time
RNT predose replacement
Display analyte and matrix information
Time deviations
Multiple analyte handling

Missing references in WinNonlin

In the Missing References in WinNonlin tab under Table business rules, administrators can specify how missing or non-existing data points are handled.

Turn on the Enabled checkbox to apply the corresponding rule.

Missing values

For existing WinNonlin missing values of '.', '-', 'missing,' or ‘blank’ in study data, values are replaced with the string listed in the Value Replacement field. A footnote is also added below the table.

Footnote defaults for specific table types
Table Type	Missing Values and Footnotes
Table Type	Body of Table	Statistical Block	Footnote
Time Concentration	NS	NA	NS=No Sample
Demographic	NR	NA	NR=Not Recorded
PK Parameter	NC	NA	NC=Not Calculated

As shown in the table above, missing values summary statistics use NA in two different ways, NR, and NC with corresponding footnotes. Footnotes associated with other business rules such as LOQ are discussed separately under each of the rules.

Note:The wildcard character for context-sensitive footnotes is * (asterisk).

Non-existent values

For non-existent time points, administrators can choose a replacement value and a footnote instead of leaving the cell blank. For example, if a study design indicates that concentration readings are taken every 15 minutes for three subjects, and only every hour for one subject. The single subject will have three missing values for each hour. In this type of situation, users can replace the empty data field with the text string listed in Value Replacement.

Missing descriptive statistic values

For missing descriptive statistic values of '.', '-', 'missing,' or ‘blank’, values are replaced with the string listed in the Value Replacement field, ‘NA’ by default.

Intext Table Template tab

The Admin Module allows users to configure the intext tables to display values using one of the templates below and sets which PK parameters, if any, are considered time dependent. A list of possible time-dependent PK parameters is displayed with the following selected as defaults: Tmax, Tlag, Tmin, Tlast, Tmax_Rate, and Mid_Pt_last.

Intext templates
Template	Template Subgroups
	Template 1A (not time dependent)	Template 1B (time-dependent)
Template 1	A ± B	A (B – C)
Template 2	A (B)	A (B, C)
Template 3	A B	A (B – C) A (B, C)
Template 4	A ± B (C)	A (B – C)
Template 5	A (B) [C]	A (B,C)
Template 6	A ± B (C) [D]	A (B – C)
Template 7	A (B) (C) [D]	A (B,C)
Template 8	A (B – C)	A (B, C)
Template 9	A (B, C)	A (B – C)

In the table above, A, B, C, and D in the Template 1A column and A, B, and C in the Template 1B column represent unique summary statistics, e.g., A ± B could represent Mean ± SD in Template 1A and A (B – C) could represent Median (Min – Max) in Template 1B. Select the summary statistic for each variable using the pull-down menus beneath the template equation. Use the Location in Output option to define where the template information is to be displayed.

Additional rules

Table data display using decimal alignment

This rule allows all tables to be created with decimal alignment. Admins can set this rule as a default, or allow the user to select it on a run-by-run basis. This selection is made to alter the horizontal alignment of numerical values in tables such that all values in a given column are aligned around their decimal points.

The exceptions to this rule are:

Values that have an associated footnote are not decimal aligned
Negative values are not decimal aligned

Table splitting

In some cases, all information within a specific table might not fit within the standard margins of a single page due to a large number of subjects, treatments, and others. In such cases, admins can allow tables to be split across multiple pages. If splitting is allowed, the user sets whether to split tables for each automation or comparison run. This selection affects all tables in a run and appends a split segment to table file names for that run. All tables can be split except for PK_Stats.

During a run, AutoPilot Toolkit determines whether splitting is warranted for each table and determines the splits needed to conform to the appropriate page margins using the following rules:

Table headers are carried over to each file in a split table.
Data columns that are underneath a set of merged cells can be separated across split table files with the exception of the Intext tables. Columns that are underneath a set of merged cells within an Intext table are considered a group. These groups cannot be split; the entire group is moved to the next split table file when needed.
Splitting across rows within the summary statistics section is not allowed.
Identical footnotes are displayed at the bottom of each split table, independent of whether the associated value is contained in that specific split table. For example, a table containing the footnote “BQL - Below quantification limit” is split into four smaller tables. Each of these split tables has the identical footnote “BQL - Below quantification limit” even if the BQL value is not present in each of the smaller split tables.
The split table file names indicate the position of each file in the whole table following the scheme mapped out below. Two numbers are appended to each the end of each file name. The first number identifies the split table’s position by row; the second, by column, as follows:

1_1 1_2

2_1 2_2

Examples of file names with table splitting
Original (un-split) Table File Name	Table Splitting Selected in User Module	Total Number of Files After Splitting and Description	Final (split) File Names
Concentration_Trt1	No	1; no splitting allowed	Concentration_Trt1
Concentration_Trt1	Yes	1; no splitting needed	Concentration_Trt1_1_1
Concentration_Trt1	Yes	2; original table too wide	Concentration_Trt1_1_1 Concentration_Trt1_1_2
Concentration_Trt1	Yes	2; original table too long	Concentration_Trt1_1_1 Concentration_Trt1_2_1
Concentration_Trt1	Yes	4; original table too long and too wide	Concentration_Trt1_1_1 Concentration_Trt1_1_2 Concentration_Trt1_2_1 Concentration_Trt1_2_2

Display of different sample sizes in intext tables

For intext tables, summary statistics on PK parameters might have different sample sizes within the same treatment. The representative sample size for each treatment, presented in a row at the top of the intext table, is based on the number of subjects with Cmax values for each treatment. PK parameter summaries that are based on a different sample size are footnoted. Administrators can set the following options:

Whether or not to enable this rule. The default is Enabled.
If this rule is used, then administrators can choose to flag footnotes using letters or numbers. The default is letters.

Significance level display in statistical tables

P-values below a certain threshold can be overwritten by a configurable text string. Administrators can configure the following:

Whether to implement this rule. The default is Enabled.
A global threshold value below which p-values are overwritten. The default is to replace values less than 0.05.
The text value with which to replace the p-values. The default is “< 0.05”.

Footnote about model time

Values that are based on Relative Nominal Time rather than the default, Relative Actual Time, can be automatically flagged and footnoted. All tables except demographics are displayed with a footnote indicating that Nominal Times were used instead of Actual times. Administrator configurations include the following:

Turn on the Enabled checkbox in the Footnote about Model Time area to set this rule. The default is enabled.
In the NCA model time footnote field, users can enter a new footnote. The default footnote is “Nominal Times used as time variable for NCA”.

RNT predose replacement

RNT values that are less than zero can be replaced with a text value indicating predose values. This option is enabled by default to replace all RNT values less than zero with the text value Predose. Administrators can choose to replace values that are less than or equal to zero, and can edit the replacement text value. To use negative RNT values, the Enabled checkbox must be cleared.

Display analyte and matrix information

Matrix and analyte information can be included in the column headers of output tables by turning on the Matrix and/or Analyte checkbox below the desired table types. Administrators can use the Column Mapping tab to map the matrix abbreviation for concentration columns in the study data to a full name, which is displayed in the final output. The concentration column headers follow the nomenclature of [Matrix]_(AnalyteID)_CONC, where the [Matrix] defines the sample matrix. The [Matrix] information can be abbreviated in the column header; with the aid of the mapping, the entire text can be displayed in the output. For tables, the display is in the form of an additional header.

Administrators can make this functionality available as user options by turning on the Available to user checkbox.

Time deviations

Administrators can configure settings to calculate and flag deviations from planned time as follows:

Choose a formula for the calculation of the Time Deviations parameter
Set a threshold for flagging values
The abbreviation (flag) and footnote to use are set in the corresponding fields

The default is to use the formula: ((Relative Actual Time – Relative Nominal Time)/Relative Nominal Time)*100, with a threshold of 15. Time deviations are then flagged with an asterisk and the footnote “Actual sampling time exceeds +/-#VALUE#% of nominal sampling time threshold”.

Multiple analyte handling

Administrators can select whether or not to produce separate tables for each analyte. They can also select whether or not to group table columns by treatment and then by analyte.