Human resource planning in home healthcare is gaining importance day by day since companies in developed and developing countries face serious nurse and caregiver shortages. In the problem considered in this paper, the decision of patient assignment must be made immediately when the patient request arrives. Once patients have been accepted, they are serviced at the same days, times and by same nurse during their episode of care. The objective is to maximise the number of patient visits for a set of nurses during the planning horizon. We propose a new heuristic based on generating several scenarios which include current schedules of nurses, the new request under consideration, as well as randomly generated future requests to solve three decision problems: first, do we accept the patient? If so, which nurse services the patient? Finally, which days and times are weekly visits of the patient assigned to? We compare our approach with a greedy heuristic from the literature by considering some real-life aspects such as clustered service areas and skill requirements, and empirically demonstrate that it achieves significantly higher average daily visits and shorter travel times compared to the greedy method.