Introduction

The following is a description of the constraints applied when calculating age and/or setting up age groups using the Manitoba Centre for Health Policy (MCHP) databases.

When programming and doing health data analysis, it is common to either organize data into particular age groups or to limit research to certain ages. (e.g. organize the data into the age groups of 0-20, 21-50 and 51+; keeping only those people aged 60 and over). Depending on the type of database being used, there will be certain ways of limiting the data by age or organizing the data into age groups.

Notes:

1. Three types of ages can be defined:



• Age as of some constant date (eg. 92/12/31)
• Age at month end of claim month
• Age for a particular day assuming birth date includes day


2. The age variable can be suffixed by a 4, 6 or 8 (i.e. AGE4, AGE6 or AGE8), depending on whether the date variable used is 4 characters (cy), 6 characters (ccyymm) or 8 characters (ccyymmdd) long. [Where cc=century, yy=year and mm=month].

3. The type of age used is study-dependent. Often, age at hospital admission is used, except for statistical rates, where age at December 31 would be used.

A. AGE AS OF SOME CONSTANT DATE

This means picking a specific constant date, e.g. 92/12/31, and subtracting the birth year from this value.


See SAS CODE below.

B. AGE AT MONTH END OF THE CLAIM MONTH

Hospitals and physicians submit claims to Manitoba Health for services provided in order to be reimbursed (for more information read the file MCHP Database). To calculate the age at the month end of the claim month, the date admitted to the hospital for services is often used. However, the date that is chosen to be used is study dependent. The admit date is often used because a claim for that service will likely be submitted within the month of service.



See SAS CODE below.

C. EXACT AGE FOR A PARTICULAR DATE

If a particular date is preferred, for example, age at date of discharge, then birth date is subtracted from the chosen date. In this example the date of admission is used.



See SAS CODE below.

D. CENTURY ADJUSTMENT

Some of the birth dates in the data contain badly coded century values. An estimate of the year must be made. The following provides a stable estimate of the century. For this example the claim year is found from admit date.



See SAS CODE below.

E. CALCULATING AGE IN YEARS AND DAYS (from SAS.com)

1. Version 8 base SAS software includes a new function called YRDIF . See SAS Code below for an example.
   
   
2. Another function available in version 8 and later, DATDIF is very useful for calculating the number of years and days between any two dates. See SAS Code below for an example.