Sure, but that is true for ever job then. An unknown and hidden confounding factor explaining job choice and the problems of the job can always exist.

The general principle may apply to any job, but you wanted to study the population of sex workers in particular, and doing so requires collecting and analyzing data, in regard to sex workers, properly and sufficiently, toward a conclusion.

I only suggested that your conclusion may not be robust if sex work is disproportionately represented by populations that carry broader vulnerabilities to some of the difficulties that you inferred were directly consequences of sex work.