Effect of a centralised transfer service on characteristics of inter-hospital neonatal transfers.

OBJECTIVE: To determine the effect of a centralised neonatal transfer service on numbers of neonatal transfers and the time taken for teams to reach the baby. DESIGN: Prospective census of neonatal inter-hospital transfers between May and July 2004. Comparison with a previous census undertaken before introduction of the service. Analysis of requests for antenatal in-utero transfer to the regional emergency bed service. SETTING: Geographically defined area in London and southeast England. PATIENTS: Babies transferred to or from a neonatal unit. INTERVENTIONS: Introduction of a centralised neonatal transfer service. MAIN OUTCOME MEASURES: Numbers of transfers, time taken for teams to arrive to the baby (response time). RESULTS: During the census there were 835 transfers with an increase of 34% from the previous census (n = 619). Most of the increase was in urgent transfers for neonatal intensive care. There was a mean of 4.4 urgent transfers a day, with 3.9 elective and 0.8 short-term transfers. Over the same period in-utero transfers decreased. Response times improved from a median of 2 h in 2001 to 1.45 h in 2004 (p<0.05). The 90th centile fell from 6 h to 4.9 h. CONCLUSION: Following the introduction of a centralised neonatal transfer service, response times improved significantly. An increase in the numbers of transfers for medical intensive care was associated with a reduced number of in-utero transfers. To balance the improved safety and accessibility of neonatal transfer, similar developments may be needed to facilitate in-utero transfer.

Identification of compartmental models for perturbed cell populations using state-space parameter estimation techniques.

Multiple compartment models for describing synchronous cell kinetics in which cell populations are characterized by DNA content are reviewed. These models are useful for understanding and predicting a cell population's dynamic response to perturbations induced by drugs or radiation. A practical approach is proposed for determining the parameters of these models from empirical cell-cycle data. Specifically, a state-space parameter estimation algorithm based on the maximum likelihood method--developed and coded primarily for engineering applications and commercially available for personal computers and minicomputers--can be applied to DNA specific cell-cycle measurements from synchronized cell populations to produce model parameter estimates. This is demonstrated using published data from a cell-cycle experiment. The results show that the procedure works well, and that with careful experimental planning even better results should be possible. Since the compartmental model is often used to represent biological systems, this approach is widely applicable.