Your figures are from a formula, the variables of which consist of guesses*. You cannot support it.
An actual survey by SAMHSA shows a decrease.
*such as a numerical factor for the relative truthtelling of heroin users vs. cocaine users.
"Cautious evaluation of this data is necessary because the NHSDA cannot accurately measure rare or stigmatized drug use, relying as it does on self-reporting and on people residing in households. In alternate research, the number of hardcore* users of heroin in 1998 was estimated to be 980,000,"
"Estimates of heroin use from the NHSDA are considered very conservative due to the probable underreporting and undercoverage of the population of heroin users."