10.4225/03/58ae2c7e6e84b Stehlik, Paulina Paulina Stehlik A needs-based approach to patient-relevant information delivery Monash University 2017 ethesis-20141212-101611 Medication management Information requirements Clinical decision support system 1959.1/1133700 Open access Complex patients Decision support system 2014 monash:148856 thesis(doctorate) Inappropraite prescribing 2017-02-23 00:27:39 Thesis https://bridges.monash.edu/articles/thesis/A_needs-based_approach_to_patient-relevant_information_delivery/4683754 The plethora of medical studies that are published annually leaves many health professionals (HPs) overwhelmed with new health information. To assist HPs, information from research is often summarised and presented in a practice ready fashion through disease specific guidelines, drug monographs, and up-date newsletters. Although these have been effective in improving the quality of medical care, criticisms of these resources are focused on their inability to deal with complex patients such as those with comorbidities and the aged. Resources tailored specifically to the aged are either simplistic and do not take into account all issues surrounding complex patient management, drug therapy individualisation and end of life care or are impractical to use at the point of care. The hypothesis of this research was that HPs’ information needs were not currently being met and would require the identification, development and testing of an information delivery framework that would address their needs. The overall aim of this research was to identify HP information needs when delivering healthcare to complex patients, such as the aged, and was conducted in three phases. The first phase of this research used fifteen personal interviews with key HPs (geriatricians, GPs and accredited pharmacist) from the Melbourne metropolitan area to explore whether they felt currently available information resources meet their needs during complex patient healthcare delivery. Specifically it explored what HPs consider when making a clinical decision during healthcare delivery, what resources they use to assist in decision making and what features they desire in an information resource. Findings suggested that currently available resources did not meet HPs information needs. Issues identified included exclusion of complex patients from the literature, lack of relevance to the Australian environment and inability to access the literature due to lack of time to search for appropriate information and costs associated with subscriptions. Other identified hurdles included HPs inability to interpret and contextualise the literature and incomplete patient information that would otherwise influence their management choices. Features that were thought to make an information resource useful included clear formatting, simplicity, use of peer-reviewed evidence-based recommendations, local relevance and ready access via an easy to use electronic interface. A new framework for delivering disease state management information to HPs that took into account all the complexities of care but did not jeopardise clinical autonomy was designed and developed in the second phase of this research. Past successes and failures of other electronic systems were identified to inform the basis of system design. These fundamental elements included adherence to usability guidelines when developing user interface, provision of recommendations based on up-to-date reputable information sources accompanied by rationale and additional information links or references, identification of all important interactions clearly identifiable by level of importance, good integration into workflow and HP involvement throughout the design and development process. In order to take into account all of the complexities of appropriate care, the clinical decision making process was modelled by reviewing Australian and international literature regarding quality prescribing, medication review, disease state management and drug monographs. Features common across all sources included making a correct diagnosis, making a decision whether to treat a patient with non-pharmacological or pharmacological treatment, documenting care, and reviewing progress. Common considerations that influence treatment choice can be divided into patient characteristics and drug attributes. The clinical decision making model developed was used to design the system framework. Issues with updating the system knowledge and subsequent retesting were overcome by developing a business rule engine rather than coding decision making algorithms. Clinical autonomy was maintained by allowing HPs to choose the guideline for the system to use as the basis of decision support, as well as display of recommendations and associated potential issues based on patient data rather than providing absolute recommendations. The usefulness of the way in which information is provided by the system was evaluated in phase three using ‘think out aloud’, questionnaire and focus group techniques. Five pharmacy academics were asked to think-out aloud while using the system to make decisions regarding the management of newly diagnosed osteoarthritis in a complex patient in order to give insight into the usability, aesthetics and usefulness of information presentation. Although there were some minor issues identified, overall participants found the way in which information was displayed useful. A validated modified computer system usability questionnaire (mCSUQ) was used to gain feedback from pharmacists and medical practitioners regarding system usefulness after a demonstration of the programme. The 17 question mCSUQ used a 7-point scale (1 = strongly disagree, 7 = strongly agree) and allows for open-ended feedback to measure user satisfaction with the system overall, its usefulness, the information and interface quality. When validated it demonstrated good internal consistency and congruent validity. Fifty-two HPs completed the mCSUQ and were invited to participate in focus group feedback. The system scored well on all domains: overall usability (mean 5.05, SD 1.07), system usefulness (mean 5.06, SD 1.11), interface quality (mean 4.84, SD 1.25) and highest on information quality (mean 5.09, SD 1.24). Two HPs were successfully recruited for focus group feedback. Results were used to confirm findings from open-ended question answers. Participants felt that the system interface was simple, clear and could be easy to learn but required colour to help identify important issues. Positive aspects to the system included linking information to patient data and the comprehensive nature of the information given. Additional information desired included statistical information on treatment efficacy and provision of medication review and deprescribing guidance. Major determinants to using the system included ability to integrate with electronic health records (EHR) and use of high quality information within the knowledge base. Barriers identified to CDSS implementation in general included lack of access to accurate and complete patient information and inconsistent medical terminology use among HPs. This research defines the types of information required during clinical decision-making as well information delivery preferences of HPs. It has resulted in the development of a new framework for delivering the information HPs need when requiring guidance during complex patient healthcare delivery. The framework developed is universal as the clinical decision making model is based on local and international literature. Only alterations to knowledge are required, rather than changes to code, to reflect local practice and drug availability. Future work could expand system knowledge, develop EHR integration capability, implement identified desired features and further evaluate the usability of the system and whether it is able to improve quality of prescribing.