topic Syntax for exponents in New SAS User

Syntax for exponents

_maldini_ — Thu, 07 Jul 2022 22:19:03 GMT

What is the correct syntax for the following calculation (see image)?

Thanks!

Re: Syntax for exponents

PaigeMiller — Thu, 07 Jul 2022 22:26:01 GMT

Use the EXP() function.

Re: Syntax for exponents

Reeza — Thu, 07 Jul 2022 22:31:05 GMT

1-0.9144**(exp(60.69-61.18))

EXP or CONSTANT('e') function for e

** for exponential

Re: Syntax for exponents

_maldini_ — Fri, 08 Jul 2022 18:36:31 GMT

Whoops. I'm getting an "Invalid argument to function" error. What am I doing wrong here?

Re: Syntax for exponents

Reeza — Fri, 08 Jul 2022 19:17:59 GMT

The number being generated is too large for a computer to represent. E to those exponents in a regular calculator errors out so something is wrong somewhere in your formulas. Guessing units?

Re: Syntax for exponents

FreelanceReinh — Sat, 09 Jul 2022 10:23:38 GMT

Hello @_maldini_,

I suspect that there are two errors in your calculation:

Looking at the source of your formula, the 2013 ACC/AHA Guideline on the Assessment of Cardiovascular Risk, it appears that the "mean" value occurring in the argument of the EXP function is a race- and sex-specific constant based on older studies. This is confirmed by the "Calculation Details and Variables" under the online calculator found in https://www.merckmanuals.com/medical-calculators/ACCAHA2013-ja.htm. Only the "individual sum" term is calculated from individual measurements. The formulas for ind_sum_blk_f and mean_blk_f in your code, however, show that you calculate both the sum and the mean from the same set of values, which is definitely wrong. According to either of the two sources linked above the relevant mean value (here: for black females) is mean_blk_f=86.61.
I would expect realistic "individual sum" values to be about the same order of magnitude as the constant "mean" values. Note that all continuous measurements (such as cholesterol) are logarithmized in the formula. So, extreme outliers leading to individual sums in the thousands (cf. your 11187, etc.) should be completely impossible and indicate that one or more values used must be "way off," like not logarithmized or having the wrong unit, as Reeza has already suggested.