An acceptance limit is a limit you set for one estimate, for example "Bias within ±2 mg/dL". Only limits you enter or adopt count. Each limit gets one outcome, and the outcomes combine into one study status. On Results, the Why panel under the status lists each limit with its observed value and outcome; select one to open it on Set up. To set limits, see Acceptance limits.
The five outcomes#
| On screen | In the report | Meaning |
|---|---|---|
| Met | ✓ Meets | The estimate is within the limit |
| Not met | ✗ Does not meet | The estimate is outside the limit |
| Undecided | ! Undecided | The limit applies but could not be decided |
| Not evaluated | ○ Not evaluated | The limit could not be assessed at all |
| Not applicable | ○ Not applicable | The limit is for another group or study type |
The study status reads Criteria met, Criteria not met, Undecided or Not evaluated; lists and the Reports page shorten these to Met, Not met, Undecided and Not evaluated. The outcome box on report page 1 uses the same words, or reads No applicable criteria or Descriptive – no criteria.
Telling Undecided and Not evaluated apart#
Neither is a pass. Either one makes the study Undecided unless another limit is not met. The reason sits under the outcome on screen and in a note in the report. Typical reasons:
| Outcome | Typical reason |
|---|---|
| Undecided | Only 1 independent specimen. At least 2 are needed to decide the criterion. A limit resting on fewer than 2 independent units is never met or failed; the reason gives the minimum number needed. What counts as a unit depends on the study: see How n is counted. |
| Undecided | Criterion unit must be … The limit's unit differs from the estimate's: a CV limit must be in %, a slope in ratio, a concentration in the study unit. |
| Undecided | The estimate could not be calculated. Too few results, no interval, or a failed fit. |
| Undecided | A specimen is missing a result or a limit. |
| Undecided | A method or instrument comparison has no limit for each specimen yet. |
| Undecided | This check needs absolute limits. A percentage limit on bias, relative bias, recovery, slope, intercept or a mean difference. These take absolute limits. Percentage limits work for specimen rules, bias at a medical decision level, and a mean or median against an assigned value. |
| Undecided | Percentage limits need a ratio scale. A percentage limit on an interval scale, which has no true zero. |
| Undecided | No limit value is entered, or a relative limit has no target value. |
| Not evaluated | No limit segment covers the target concentration. |
| Not applicable | Criterion applies to a different level. The limit is for another level, or the study type does not report that statistic. |
To resolve one, fix the setup on Set up or the data, save and choose Recalculate results.
How the study status is decided#
The status on the summary line, the project study list and report page 1 combines every limit in every group:
- Not applicable limits are set aside. A limit for a level that isn't in the data, such as
01when the data say1, counts as Not evaluated when another limit applies. - If nothing is left, the status is Not evaluated.
- If no results were included, the status is Undecided.
- If any limit is not met, the study does not meet its limits.
- Otherwise, any Undecided or Not evaluated limit makes the study Undecided.
- The study meets its limits when every remaining limit is met.
Method, instrument and lot comparison always include a Specimen agreement check. It stays Undecided, and the study Undecided, until you set a limit for each specimen.
Two checks on the five results in descriptive.csv (mean 6.0 mg/dL, SD 3.16 mg/dL, all at level 1):
| Limits set | Outcomes | Study status |
|---|---|---|
Mean 5 to 7 mg/dL; SD at most 3 mg/dL; a CV limit entered in mg/dL; a mean limit for level 2; a slope limit | Met; Not met; Undecided (wrong unit); Not applicable; Not applicable | Criteria not met |
Mean 5 to 7 mg/dL; a mean limit for level 01 | Met; Not applicable (no level 01) | Undecided |
In the second, the study notes ask you to correct the level. The study stays Undecided until you do and recalculate.
In the project's study list, a study without limits shows No acceptance criteria set, and a derived sigma study shows Not evaluated. See when no pass or fail is intended.
Decisions use unrounded estimates#
Every decision uses the full-precision estimate. If the estimate and limit look equal after rounding, the value is marked †. Display decimal places in Calculation settings on Set up shows more digits. The decision stays the same. Specimen rules compare each difference with its allowance in exact decimal arithmetic.
The report gives its rounding rule in the appendix section Decision rules and display precision.
Inclusive and exclusive limits#
| Written as | Meaning |
|---|---|
≤ 6 mg/dL, ≥ 5 mg/dL | Inclusive: a value exactly on the limit meets it |
< 6 mg/dL, > 5 mg/dL | Exclusive: a value exactly on the limit does not meet it |
With a mean of exactly 6 mg/dL, "Mean at most 6 mg/dL" was met and "Mean below 6 mg/dL" was not.
In 21-position carryover, carryover equal to the error limit is not met. See Carryover.
Criteria on a confidence bound#
- Agreement lower bounds. Qualitative agreement can compare overall, positive or negative agreement at the lower end of its Wilson CI.
- Interference effects. Met when the whole CI of the effect lies within the limits. A point estimate inside the limits with a wider or missing interval is Undecided.
The same data can meet a point-estimate criterion and fail a bound-based one. On the example agreement file, overall agreement was 91.9% with a 95% Wilson CI of 78.7–97.2%:
| Limit name | Limit | Value compared | Outcome |
|---|---|---|---|
| Overall agreement | ≥ 90% | 91.9% | Met |
| Overall agreement lower bound | ≥ 90% | 78.7% | Not met |
Use a bound-based criterion when you need confidence that performance reaches the limit.
Specimen-by-specimen criteria#
In method, instrument and reagent lot comparison, the specimen rule checks each specimen's difference against its allowance, then counts. The count decides. Means and confidence intervals are supporting statistics.
| Study | Specimens that must be within their limits | Minimum specimens | Also required |
|---|---|---|---|
| Method comparison | At least 95%: all of 5–19, 19 of 20, 24 of 25 | 5 | — |
| Instrument comparison | At least 95%, for each analyzer | 5 | — |
| Reagent lot comparison | Every specimen | 2 | Lot plan complete |
The observed value is the count, for example 19/20 (95%), and Specimen decisions lists every specimen. A specimen missing a result or a usable limit still counts, as one that could not be evaluated. The rule is met only when every specimen was evaluated. It is not met as soon as the failures alone exceed the number allowed. With 20 specimens, 1 failure plus 1 not evaluated is Undecided, because 19 could still pass. 2 failures plus 1 not evaluated is Not met. See Method comparison, Instrument comparison and Reagent lot comparison.
Decision levels outside the measured range#
In method comparison, bias at a medical decision level comes from the fitted line. A decision level outside the measured comparative range is extrapolated and stays Undecided.
On the example comparison file (comparative results 52.1 to 365.9 mg/dL), a limit at 600 mg/dL was Undecided, and the bias table marks that level as outside the studied range.
The intercept is the fitted value at 0, so an intercept limit is Undecided whenever 0 is outside the measured range. The fix is more specimens across the range.
Other reasons a limit is undecided#
The reasons in the table above cover the common cases. Analysis problems says what to do about each.
Sign-off#
Criteria met means the saved results meet the limits you entered.
Signing off is optional and never changes an outcome. The outcome box keeps the calculated outcome, so an undecided or descriptive result stays that way, and the reviewer's name and date fill the signature lines at the foot of page 1. See Signing off.