MP-SIR Parameters Tab

Use the Parameters Tab in an MP-SIR session to specify your desired settings for the dates and length of your study, the confidence intervals used in your analysis, and the cut-points for the time-dependent variables in the analysis.

Study Time

Choose a date to use for each of these parameters from their drop-down lists.

A subject begins accruing person time at risk, and SEER*Stat begins counting their events, at the latest of these three dates:

• the Exposure Date plus the Latency Exclusion Period
• the Start Date
• the study's Cutoff Start and End Dates

If you are using the SEER databases, the Exposure Date and Start Date are always the person's Date of diagnosis recode, which is the date of diagnosis for their index record (as defined on the Selection tab).

• The Exposure Date is the date on which a person qualifies to be in the cohort. It is the first day on which the person is considered to be at risk for having a second event.
• The Latency Exclusion Period is the number of months after the Exposure Date during which person time at risk and potential events are ignored. Two months is the default. Adjust the number by typing in the field or using the up and down arrow buttons next to it.
• The Start Date is a person-specific date before which person time at risk and potential events are ignored. It is the subject's earliest possible date of entry into the study, subject to the Latency Exclusion Period and Cutoff Start Date. Note that, if you are using the SEER databases, the Start Date is irrelevant since it will always be the same as the Exposure Date.
• The Cutoff Start Date is the date on which your study begins. Events and person time at risk are not counted before this date.
• The Cutoff  End Date is the date on which your study ends. Every subject exits the study on the Cutoff End Date, if they have not exited before then.

Cutoff dates default to the beginning and ending dates in the database; you cannot choose cutoff dates that are outside of the database's range.

Cut-points for Time-dependent Variables

In this section of the Parameters tab, set the cut-points for any time-dependent calculated variables you are using in your analysis.

It is not necessary to set cut-points for variables that you are not using in your analysis. If you set a cut-point for a variable, but do not use that variable on the Table tab, SEER*Stat will issue a warning when you execute the session. However, it is necessary to set cut-points for variables you are using in your table, and SEER*Stat will not allow you to execute the session if you have not.

Type the cut-points you want to use in the field. SEER*Stat remembers any sets of cut-points you have used recently, as well as some default sets, and you can access them by clicking the arrow at the right of the field. Cut-points should be expressed as a list of numbers separated by commas.

Latency cut-points mark intervals of time since the Exposure date. They can occur at intervals of months or years; add an "m" or "y" after the number to indicate which you mean. Example: "4m,8m,1y,3y". The first latency interval starts at the end of the Latency Exclusion Period defined above. There is an implicit total latency category.

Latency is useful in determining how long a subject can survive after exposure without developing a cancer.

Attained Age cut-points mark the intervals for individuals' ages as they are tracked through the study. For example, to use ten-year age groups, write "10,20,30,40,50,60,70,80". Note that the highest age group in the SEER incidence data is "85+".

Attained age is useful in determining whether the relative risk of developing a cancer differs depending on the age of the subject at the time of the subsequent event.

Attained Calendar Years cut-points delimit groups of calendar years, which can also be tracked for individuals throughout the study. Use four-digit years. Example: "1975,1980,1985,1990,1995". Note that the earliest year for which SEER data was reported is 1973.

Attained calendar year is useful in determining whether cancer incidence in a cohort reflects known patterns of disease in relation to SEER reporting through the years.