Common Notes, Warnings and Errors with POP_RATE macro. 1. The bygroup paramter is only used to divide the numerator data into separate groups. Rates are then provided for each of these groups for the whole population. To generate rates for groups within both the numerator and population data create a new variable for the AREA parameter prior to calling the macro. 2. If a standard popuation has been defined then indirect rates should not be used. They are not adusted to the standard population, but the total of the numerator and denominator populations. The program will continue to run after this point. WARNING: A Standard Population dataset has been defined; that is different than the population corresponding ; the numerator. Do not use the Indirect rates. ; The Calculation for indirect rates does not use ; the provided standard population.; 3. A standard Rates data set can be defined separately. The indirect rates are then standardized using these rates. The program will continue to run after this point. NOTE: A Standard Rates dataset has been defined ; The indirect rates will be standardized using ; the rates in this data set ; 4. If you want a comparison region defined then both the compdef and compval parameters must be passed to the macro. The program will continue to run more errors will be returned. The output in this case should not be used. WARNING: Comparison AREA defined without comparison value ; WARNING: Comparison value defined without comparison definition ; 5. After summarizing and merging the numerator and population data the program checks how many strata in the population file have no corresponding events in the numberator files. If you get this warning the ASRATES data should be printed to determine how many strata have zero events. Generally this does not affect the rates calculations, but it can cause significant variation in the standard errors. Direct rates are more unstable with sparse data (missing a lot of fields). If you can not re-adjust your data you may want to consider using indirect rates. Check if the values for AREA/CONF1/CON2 match in both the numerator and denominator data sets. WARNING: 12 strata had 0 events in the ; numerator dataset. This situation can lead to ; incorrectly calculated standard error values. ; Check the ASRATES dataset to see how many strata; are missing. As a general rule of thumb there; should be no more than 2 missing strata for any; specific region or area. ; The rates macro has re-set the number of events ; to 1 for calculating st. errors. ; 6. Population adjusted warning. This warning is often generated in when running rates on small populations, or very large utilization data sets (e.g. medical claims). This occurs when there are more individuals in the popuation data. The records where this occurs are printed to the log. This error often occurs when an individual identify is not defined. This can cause significant changes in the rates. Check if the values for AREA/CONF1/CON2 match in both the numerator and denominator data sets. WARNING: 12 POP values where changed. More individuals ; occured in the event data than in the population. ; The population value was increased to match the ; event file ; NOTE: You have not indicated an individual identifier ; This may be the cause of the above mentioned ; problem. Please if your are using binary variables; in a file that may contain multiple records per; individual use an individual identifier ; 7. The overall rates (direct, crude, indirect) should be the same for the overall rates when no standard population (or rate) has been defined. Check to ensure that all values of your AREA value exist in both data sets using PROC FREQ with no formats. WARNING: There may be a problem with the rates calculations. ; The crude and direct rate for the overall &area (&taval) ; are not the same. This may occur if there are missing ; values of &area on the numerator data set. ; 8. If an incorrect SAS expression has been used to defene the comparison region the following error may be generated. The comparison region should read where EXPRESSION ; Where EXPRESSION is provided in the macro call as compdef= The 'if' and 'then' are provided by the macro. There are some restrictions on the use of an '=' sign in a macro call. An 'EQ' should be used instead. WARNING: There was an error during the execution of the macro.; The output data sets will be empty or incomplete ; There was and ERROR creating the summarized numerator; dataset with sub-area, and total area summaries.; 9. The parameter passsed for the overall by group does not match the type of the bygroup variable. WARNING: Thee was an error during the execution of the macro.; The output data sets will be empty or incomplete ; The group or bygroup variable on the numerator dataset is ; numeric but an alphabetic (A-Z) value has been requested ; for a group total value. ; The overall bygroup value should be greater then any of the values of the bygroup variable This is a common error and is not currently checked in this program. 10. The parameter passed for the overall area does not match the type of the area varable. WARNING: There was an error during the execution of the macro.; The output data sets will be empty or incomplete ; The area variable on the numerator dataset is ; numeric but an alphabetic (A-Z) value has been requested ; for an area total or subarea values. ; The overall area value should be greater then any of the values of the area varaible. This is a common error and is not currently checked in this program. 11. Statistics option incorrectly used. If the statistics option is incorrectly used there may be no error registered, but standard errors and variance will be incorrect. This problem is often related to item 6 above. When using the statistics option (stats=stats). - Individual level numerator data must be used. - If there are multiple records/individual an id= option must be set or every record is treated as an individual. Summarized data may be use the numerator (stats=sum). - Events can only happen once (e.g. mortality). 12. The macro writes out some basic information when the program has finished. This information should be checked against what you know about the input data. All of the notes have been put here - those marked with an * require a parameter to be set. Populis Rates Macro finished. Input Data: Numerator data = EVENT Denominator data = POP Standard Pop data = STPOP * Standard Rate data = * Confounding vars = AGEGRP SEX * Output Data: ; Summary data = SUMMDATA Strata Specific Rates = ASRATES T^2 values = T2 Notes: Number of rate variables= 1 REGION values= 10 Overall REGION value= Z Comparison &area= Y * BYGRP values= 3 Overall BYGRP value= Z Time period event data represents= 2 * Confidence Interval = 95 * T^2 values have been log transformed * Stats option off *