City Research Online

Abstract

There has been a significant increase in the demand for temporary skilled workers in the health sector. They provide volume flexibility, but are generally more expensive than their permanent counterparts. A balance must therefore be struck between staffing cost and service quality by recruiting the right mix of temporary and permanent healthcare workers. Focusing on periods of highly uncertain demand, in this thesis, we propose optimization models aiming to inform permanent and temporary recruitment decision making for settings in which all patients must be served. We pursue this under two different scenarios, a mid-term planning horizon and a long-term planning horizon.

The first part of the thesis [1] is devoted to recruitment decision making in a mid-term planning horizon. The main trade-off in this case is between recruitment lead times and staffing costs of temporary and permanent workers. More specifically, permanent skilled workers are cheaper for the healthcare provider than equivalent temporary workers, but have a substantially longer recruitment lead time. Longer recruitment lead time of permanent workers implies that providers face a higher level of demand uncertainty when making permanent recruitment decisions and a higher likelihood of not being able to fill the created positions. Considering a single-interval planning horizon, we propose a two-stage stochastic optimization framework to capture this fundamental trade-off. The first stage of our framework identifies the number of permanent positions to advertise, and the second stage determines the number of temporary workers to recruit. Our framework accounts for the uncertainty in the number of permanent vacancies that will be filled, stochasticity of the service delivery process, and imperfect demand information at the time of advertising for permanent positions. Under a general setting of the problem, we characterize the optimal first- and second-stage decisions analytically, propose fast numerical methods for finding their values, and prove some insensitivity and monotonicity properties for the optimal decisions and their corresponding costs. The benefit/loss of delaying the advertisement for permanent positions to obtain a more accurate demand information, at the expense of a higher risk of not filling the advertised positions, is also investigated. A case study based on data from a geriatric ward illustrates the application of our framework to an inpatient department, and further managerial insights are developed using a combination of analytical and numerical results.

The second part of the thesis is dedicated to recruitment decision making in a long-term planning horizon. In addition to the different staffing costs and recruitment lead times of temporary and permanent workers captured in the first part, we consider the difference in their placement durations. This is because permanent workers have substantially longer contracts which may cover periods of low demand, hence in the long run, they are likely to be more expensive to the provider than temporary workers. We capture this by a multi-interval optimization framework which involves a two-stage decision making, similar to the two-stage decision making of the first part, repeated in each interval. The time-varying nature of demand over different intervals is also incorporated into this framework. Using a Markov decision process formulation, we prove that the optimal recruitment policy for permanent healthcare workers in this context has a hire-up-to structure. Numerical experiments then investigate the sensitivity of the hire-up-to value to different system parameters. The potential benefits of using the long-term (multi-interval) recruitment model as compared to the mid-term (single-interval) recruitment model is also evaluated numerically.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management H Social Sciences > HJ Public Finance R Medicine > RA Public aspects of medicine
Departments:	Bayes Business School > Bayes Business School Doctoral Theses Bayes Business School > Management Doctoral Theses

Export

Downloads