Moving beyond a limited follow-up in cost-effectiveness analyses of behavioral interventions