Rethinking the serological threshold for onchocerciasis elimination