No abstract

BackgroundIt is standard practice to review all patients following discharge at a follow-up clinic but demands on all health services outweigh resources and unnecessary review appointments may delay or deny access to patients with greater needs.AimsThis randomised trial aimed to establish whether a virtual outpatient clinic (VOPC) was an acceptable alternative to an actual outpatient clinic (OPC) attendance for a broad range of general surgical patients following a hospital admission.Patients and methodsAll patients admitted under one general surgical service over the study period were assessed. If eligible for inclusion the rationale, randomisation and follow-up methods were explained, consent was sought and patients randomised to receive either a VOPC or an OPC appointment.ResultsTwo-hundred and nine patients consented to study inclusion, of which 98/107 (91.6%) in the VOPC group and 83/102 (81.4%) in the OPC group were successfully contacted. Only 6 patients in the OPC group and 10 in the VOPC group reported ongoing issues. A further follow-up indicated 78 of 82 (95%) VOPC patients were very happy with their overall experience compared with 34/61 (56%) in the actual OPC group (p<0.001). A significant proportion of both cohorts—68/82 (83%) in VOPC group and 41/61 (67%) in OPC group (p = 0.029)—preferred a VOPC appointment as their future follow-up of choice.ConclusionsThe majority of patients discharged from a surgical service could be better followed up by a virtual clinic with a significant proportion of patients reporting a preference for and a greater satisfaction with such a service.

Viewing the trajectory of a patient as a dynamical system, a recurrent neural network was developed to learn the course of patient encounters in the Pediatric Intensive Care Unit (PICU) of a major tertiary care center. Data extracted from Electronic Medical Records (EMR) of about 12000 patients who were admitted to the PICU over a period of more than 10 years were leveraged. The RNN model ingests a sequence of measurements which include physiologic observations, laboratory results, administered drugs and interventions, and generates temporally dynamic predictions for in-ICU mortality at user-specified times. The RNN's ICU mortality predictions offer significant improvements over those from two clinically-used scores and static machine learning algorithms.

No abstract