Each study answers one question. The New study dialog lists the studies by topic and sets up the matching study type. If you already have a results file, Import a file can suggest the study instead; see importing a file.
Starting a study#
- Open a project and select New study. In an empty project, select Or choose a study first. In an empty workspace, the Projects page opens on Start a study instead: select a study, then Start this study, and the dialog opens with that study chosen.
- Fill in Study title (optional), Analyte and Unit. Opened from a project, the dialog fills in the analyte and unit most of its studies use.
- Under Study type, select a study, or type part of its name in Find a study type…. Its question, Suggested collection and what You get appear below the list, with a Run this study link to its guide. Some studies ask a few more details there (below); these can also wait until Set up.
- Select:
- Create study to open the study at Set up, where you check the collection plan and record your acceptance limits before collecting, or
- Create and import data if you already have results. A sigma study has only Create study. In an empty workspace, or when you pick New project in the dialog, a project named after the analyte, for example Glucose evaluation, is made for the study.
Which study answers your question#
| Your question | Choose | Chapter |
|---|---|---|
| How much do results vary within runs, between runs and between days? | Multi-day precision | Multi-day precision |
| How closely do repeated results of one material agree within a single run? | Repeatability | Repeatability |
| Do two methods or instruments give the same results on the same patient samples? | Method comparison | Method comparison |
| Do two or more instruments agree when your file has one result per row? | Instrument comparison | Instrument comparison |
| Does a new reagent lot give the same patient results as the current lot? | Reagent lot comparison | Lot comparison |
| How far are results from the assigned value of a reference or control material? | Bias check | Bias against an assigned value |
| Is the response a straight line across the measuring range? | Linearity | Linearity |
| Does a diluted high sample recover its expected value after correction? | Dilution verification | Dilution verification |
| What is the lowest amount the method can detect or quantify? | Detection limits | Detection limits |
| Does hemolysis or another added substance change results? | Interference | Interference |
| Does a high sample raise the result of the low sample run after it? | Carryover | Carryover |
| How long do specimens, reagents, calibrators or controls stay stable in storage? | Stability | Stability |
| Do healthy-subject results fit an existing reference interval or limit? | Verify a reference interval | Reference interval verification |
| What reference interval or limit do your healthy-subject results support? | Establish a reference interval | Reference interval establishment |
| Do positive and negative results agree with a comparative method? | Qualitative agreement | Qualitative agreement |
| How reliably does a qualitative test call samples near its cutoff? | Near-cutoff precision | Near-cutoff precision |
| What sigma do your known bias and imprecision give against the allowable total error? | Sigma metric | Sigma metric |
| What are the mean, SD and CV of a set of results? | Descriptive summary | Descriptive summary |
Once a study exists, both reference-interval choices show the study type Reference interval, and both qualitative choices show Qualitative agreement or near-cutoff precision.
What each form asks#
Every study asks for Analyte and Unit and an optional Study title. Qualitative agreement needs only Analyte. A blank title becomes, for example, Glucose · repeatability. Some studies ask more:
| Study | Extra questions |
|---|---|
| Method comparison | Comparative method or instrument, Test method or instrument |
| Instrument comparison | Reference analyzer |
| Bias check | Assigned value |
| Dilution verification | Lowest reportable value before dilution, Highest reportable value before dilution, Expected value of the undiluted sample |
| Reagent lot comparison | Previous lot name, New lot name, Required matched pairs. Set each specimen's allowable error and the concentration groups on Set up. |
| Verify a reference interval | Reference limits (two-sided or one-sided), Claimed lower reference limit, Claimed upper reference limit |
| Establish a reference interval | Reference limits (two-sided or one-sided) |
| Detection limits | Which limit (limit of blank, detection or quantitation) |
| Interference | Interferent, Interferent concentration unit |
| Stability | Time unit, What is stored, Same material measured at every time?, Largest acceptable change from baseline (%) (optional) |
Type numbers in these fields with a decimal point (1.5). The study's Blank template (CSV) on Set up and Data has the columns it needs; your own columns can have any names and order.
After you create the study#
Create study opens the study at Set up, with its collection plan open. Create and import data opens the study at Data with Import a file open. A carryover study also offers Start a blank 21-position run on Data, and a sigma study takes values instead of a data table.
If the calculation needs a setting the dialog didn't ask for, such as a lot study's allowable error or the quantitation goal for LoQ, the study header shows Finish set up and Set up names what's missing. You can change everything later on Set up.
The study design is fixed once data exist#
You can change Study type only until the study has rows or a saved calculation. After that, create a new study and import the same file again. To remove the old study, choose More, then Move study to Trash; see deleting a study.
Start again with the same setup#
- Open the study you want to reuse. Choose More → New study from this setup. Any unsaved changes are saved first.
- Enter a Study title, choose the Project and check the Study date. Select Create study.
- Check the new study's Set up stage before collecting or importing results. The study details, calculation settings, assigned values and acceptance limits are copied. Enter the new lot names and check any settings that depend on this experiment.
The new study starts with empty data and its own version history. Its source study, saved calculations and reports are unchanged. Reagent and calibrator lots, the previous and new lots of a lot comparison, and links to earlier calculations are cleared; select the precision evidence again if the new comparison needs a linked Deming lambda. A sigma study starts with an empty entries table.