AbstractRisk reduction processes in healthcare remain at the core of 21stcentury health care provision, though the continuing scale of the problem gives little room for complacency. While other areas of complex technological activity such as air transportation can demonstrate improvements in safety performance, comparable progress eludes modern healthcare. A review of risk reduction techniques within healthcare identifies that there exists a lack of tools involving simulation of risk. It has been necessary in the context of the research to establish many wholly original information structures representing healthcare activity and associated risk related interactions.
This Thesis describes a new risk simulation environment for the Critical Care Unit of University Hospital, Coventry which is a 1200 bed modern acute hospital which fully opened in 2006. Available sets of patient admission/discharge information and records of patient treatment records used for cost charging together with extensive direct observation of clinical activity are used to create simulated patient episodes within the Critical Care environment. Specific patient interventions are sub divided into a series of up to 7 sub tasks which are associated with sub competencies and a linked adverse effect. Such sub competencies can be coded to reflect three levels of task complexity. Separate codes can be allocated to identify sub competencies which are supervised and sub competencies for which additional competency can be requested from other team members.
A fuzzy logic framework has been adopted to combine empirically derived mathematical functions which for a specific sub task, translate values of individual effectiveness, distraction, competency mismatch of individual/team together with the level of supervision to a specific risk value for each adverse effect. This fuzzy logic framework, referenced as the ‘risk engine’ has specific responses for levels of sub task complexity and can be modified by indicators relating to sub task supervision and competency sharing. In addition, each sub task/competency is associated with an adverse effect whose probability of occurrence can be reduced through identified safe working practices which are referenced as ‘preventive measures’. Individual effectiveness is identified as being influenced by cirdadian rhythm, physical effort, emotional/stress effort, intellectual effort, sleep deficit and long term factors. Organisational factors influencing individual effectiveness are identified as patient admission and shift handover.
The risk simulation process is implemented within a 10 bed Critical Care Unit which utilises a specifically designed nurse rostering process for 12 hour shift periods. Sub grades of nurse skills (1 to 15) are used to structure skill mix within each rostered group and which are based on representative nurse grades (band 5, 6 and 7).Available competencies of nursing staff for a specific sub task are allocated on the basis of sub grade value and the parameter of individual competency mismatch is derived from values of required competency and available competency for each sub task. The team competency mismatch for a specific sub task linked to a specific individual is derived from the maximum available competency within the active nursing team. Nursing staff are allocated to patients on the basis of clinical need at the start of each shift.
A novel feature of the model identifies modes of interaction between nursing individuals on a ‘bed to bed’ basis as relating to parameters of distraction, supervision and competency sharing and which are related to the physical layout of the active clinical area. A fuzzy logic sub system for determining values of such interaction coefficients and which uses the same design methodology as the ‘risk engine’ is described.
The risk simulation model is operated for a sequence of 9 months of simulated clinical activity and the outcome expressed in a number of ways including the relative occurrence of types of adverse effects based on occurrences per patient day stay. Comparison is made with the level of occurrence of locally reported clinical adverse events within the Critical Care Unit at University Hospital Coventry using the coding system of types of adverse effects of the simulation system. The lack of agreement between the two sets of data is attributed to mismatch between the basic information content of the two data sets, under reporting of the local clinical adverse incident reporting system (as confirmed with comparison with results of relevant clinical studies) and the need of further refinement in the complex process of simulation of clinical interventions. Modes of reporting of simulated risk activity are also described in the context of a normalised patient day where the resulting risk profile is related to the patterns of clinical activity simulated within the model. This replicates some of the expected characteristics of simulated data such as circadian factors and of the morning shift changeover but may indicate the need for further refinement in the process of simulation of clinical interventions.
|Date of Award||2009|