Many existing manufacturing systems still rely heavily on human workers as the backbone of their production processes. Such systems are commonly termed labor-intensive. Developing Digital Twins for labor-intensive manufacturing lines is a complex and challenging task as human involvement adds another level of uncertainty. While Digital Twins offer numerous benefits, such as improved efficiency, reduced downtime, and enhanced decision-making, they also come with unique challenges when they need to be developed for labor-intensive manufacturing systems. In this paper, we discuss the main challenges and their implications that arise from existing research. Considering these challenges, we propose a framework for developing data-driven Digital Twins of labor-intensive manufacturing systems as an initial step towards addressing these challenges. We illustrate the challenges associated with Digital Twins of labor-intensive manufacturing systems through a practical case study derived from our collaboration with two companies. In the case study, we make necessary preparations for developing Digital Twins for decision support in job scheduling in a hybrid machine-worker environment while considering the well-being of workers.