: This research will address the following questions and hypotheses: Question 1.a What is the risk stratification rule that optimizes identification of patients with high fracture risk? Question 1.b How do the sensitivity and specificity of the new rule qualitatively compare to BMD testing alone? Hypothesis 2.b Information constructs important to clinicians will affect the likelihood of prescribing by at least 25% Hypothesis 3.a The decision tool will result in a 25 percentage-point increase in prescribing among high risk patients This project focuses on knowledge engineering for the creation of decision support in the area of osteoporosis;broad-based implementation is beyond the scope of the work. The overall goal is to create a robust method for designing decision support that, in large part, incorporates the needs of clinicians in order to minimize cognitive burden. It is established that poorly designed decision support tools greatly tax clinician working memory.27 These researchers showed that the common approach to design of CPOE (focused on functions or tasks) does not necessarily reflect physician mental models for conducting clinical tasks and physician preferences for structuring tasks and navigating CPOE systems. A decision support tool that was designed using a heuristic approach (a commonly used mnemonic for considering order types) also posed an unnecessary cognitive burden. This suggests that task analysis is an important but frequently overlooked step in the careful design of decision support. Barriers to adequate identification and treatment of chronic disease include process barriers and information barriers. Several process barriers that are specific to computerized decision support also include unclear role responsibilities for clinicians and nurses related to clinical reminders, concerns that interacting with the computer while with the patient will negatively impact the provider-patient interaction, and using paper forms during the clinic encounter rather than directly entering notes and orders into the computer.28 A great deal of work has been done to assess process barriers to decision support implementation, and investigators have proposed the use of several 'facilitators'to overcome them, such as strategically locating computer workstations in exam rooms and integrating reminders into clinic workflow.28, 29 While process barriers are an important part of implementation research, addressing them is not the focus of the proposed research. Our goal is to focus on information barriers. The proposed work involves the adaptation of assessment rules from clinical practice guidelines to the targeted population and available data in order to create and validate a measure of assessment that is specific for that population. This is followed by the incorporation of clinician information needs-assessment. The ultimate deliverable is a decision rule that should increase the decision-making capacity of clinicians without appreciably increasing cognitive load. A model for phases in testing of decision support technology includes a) testing the knowledge base, b) testing the knowledge base and execution engine in concert, and c) testing the clinical decision support system architecture 1.30 The scope of the proposed work lies entirely within the realm of creating and testing the knowledge base.
This project focuses on knowledge engineering for the creation of decision support in the area of osteoporosis;broad-based implementation is beyond the scope of the work. The overall goal is to create a robust method for designing decision support that, in large part, incorporates the needs of clinicians in order to minimize cognitive burden. It is established that poorly designed decision support tools greatly tax clinician working memory.27 These researchers showed that the common approach to design of CPOE (focused on functions or tasks) does not necessarily reflect physician mental models for conducting clinical tasks and physician preferences for structuring tasks and navigating CPOE systems. A decision support tool that was designed using a heuristic approach (a commonly used mnemonic for considering order types) also posed an unnecessary cognitive burden. This suggests that task analysis is an important but frequently overlooked step in the careful design of decision support. Barriers to adequate identification and treatment of chronic disease include process barriers and information barriers. Several process barriers that are specific to computerized decision support also include unclear role responsibilities for clinicians and nurses related to clinical reminders, concerns that interacting with the computer while with the patient will negatively impact the provider-patient interaction, and using paper forms during the clinic encounter rather than directly entering notes and orders into the computer.28 A great deal of work has been done to assess process barriers to decision support implementation, and investigators have proposed the use of several 'facilitators' to overcome them, such as strategically locating computer workstations in exam rooms and integrating reminders into clinic workflow.28, 29 While process barriers are an important part of implementation research, addressing them is not the focus of the proposed research. Our goal is to focus on information barriers. The proposed work involves the adaptation of assessment rules from clinical practice guidelines to the targeted population and available data in order to create and validate a measure of assessment that is specific for that population. This is followed by the incorporation of clinician information needs-assessment. The ultimate deliverable is a decision rule that should increase the decision-making capacity of clinicians without appreciably increasing cognitive load. A model for phases in testing of decision support technology includes a) testing the knowledge base, b) testing the knowledge base and execution engine in concert, and c) testing the clinical decision support system architecture 1.30 The scope of the proposed work lies entirely within the realm of creating and testing the knowledge base.
