OBJECTIVE: There is increasing interest in using artificial intelligence (AI) in pathology to improve accuracy and efficiency. Studies of clinicians' perceptions of AI have found only moderate acceptability, suggesting further research is needed regarding integration into clinical practice. This study aimed to explore stakeholders' theories concerning how and in what contexts AI is likely to become integrated into pathology. MATERIALS AND METHODS: A literature review provided tentative theories that were revised through a realist interview study with 20 pathologists and 5 pathology trainees. Questions sought to elicit whether, and in what ways, the tentative theories fitted with interviewees' perceptions and experiences. Analysis focused on identifying the contextual factors that may support or constrain uptake of AI in pathology. RESULTS: Interviews highlighted the importance of trust in AI, with interviewees emphasizing evaluation and the opportunity for pathologists to become familiar with AI as means for establishing trust. Interviewees expressed a desire to be involved in design and implementation of AI tools, to ensure such tools address pressing needs, but needs vary by subspecialty. Workflow integration is desired but whether AI tools should work automatically will vary according to the task and the context. CONCLUSIONS: It must not be assumed that AI tools that provide benefit in one subspecialty will provide benefit in others. Pathologists should be involved in the decision to introduce AI, with opportunity to assess strengths and weaknesses. Further research is needed concerning the evidence required to satisfy pathologists regarding the benefits of AI.

Adapting dashboard design to different contexts of use is an open question in visualisation research. Dashboard designers often seek to strike a balance between dashboard adaptability and ease-of-use, and in hospitals challenges arise from the vast diversity of key metrics, data models and users involved at different organizational levels. In this design study, we present QualDash, a dashboard generation engine that allows for the dynamic configuration and deployment of visualisation dashboards for healthcare quality improvement (QI). We present a rigorous task analysis based on interviews with healthcare professionals, a co-design workshop and a series of one-on-one meetings with front line analysts. From these activities we define a metric card metaphor as a unit of visual analysis in healthcare QI, using this concept as a building block for generating highly adaptable dashboards, and leading to the design of a Metric Specification Structure (MSS). Each MSS is a JSON structure which enables dashboard authors to concisely configure unit-specific variants of a metric card, while offloading common patterns that are shared across cards to be preset by the engine. We reflect on deploying and iterating the design of OualDash in cardiology wards and pediatric intensive care units of five NHS hospitals. Finally, we report evaluation results that demonstrate the adaptability, ease-of-use and usefulness of QualDash in a real-world scenario.

This article describes the design and evaluation of two generations of an interface for navigating datasets of gigapixel images that pathologists use to diagnose cancer. The interface design is innovative because users panned with an overview:detail view scale difference that was up to 57 times larger than established guidelines, and 1 million pixel “thumbnail” overviews that leveraged the real estate of high-resolution workstation displays. The research involved experts performing real work (pathologists diagnosing cancer), using datasets that were up to 3,150 times larger than those used in previous studies that involved navigating images. The evaluation provides evidence about the effectiveness of the interfaces and characterizes how experts navigate gigapixel images when performing real work. Similar interfaces could be adopted in applications that use other types of high-resolution images (e.g., remote sensing or high-throughput microscopy).

In a recent JAMIA article, Otte-Trojel et al. 1 present a realist review of patient portals. We commend the authors for using this approach to synthesizing evidence, which is a divergence from traditional systematic review methodology. We believe realist approaches have much to offer the medical informatics community, providing a means to not only determine if health IT interventions provide benefit in terms of outcomes, but to understand why and in what contexts such benefits may occur. However, we feel it is important to address some concerns we have regarding the way in which the authors used realist methods in their review. Our intention is to encourage the authors to expand on this work and to clarify for readers some of the key concepts of realist reviews and how they differ from traditional systematic reviews. In this, we respond to the call of realist evaluators for collective scrutiny of each other’s work to drive the method forward.2 Realist reviews identify theories of how an intervention works, for whom, and in what circumstances, and then test and refine those theories through consideration of primary studies.3 For realists, interventions themselves do not produce outcomes. Rather, interventions offer resources; outcomes depend on how recipients respond to those resources, which will vary according to the context. Realist theories, referred to as Context Mechanism Outcome configurations, explain how different contexts trigger particular mechanisms (the reasoning and responses of recipients) which, in turn, give rise to a particular pattern of outcomes. An important initial stage in a realist review is “theory elicitation,” where reviewers explore the literature with the explicit purpose of identifying theories.4 Otte-Trojel et al. 1 undertook an exploratory review to “identify ways in which patient portals may contribute to health service delivery and patient outcomes.” In reporting the results of this initial review, the authors describe what could be considered a mixture of resources that patient portals might offer (patient access to information and services, patient decision-support) and possible outcomes (coordination of care around the patient; interpersonal continuity of care; health services efficiency; and service convenience to patients and caregivers). However, nothing has been reported about how patients might respond to those resources or how their responses might vary according to the context. Looking at the reference list, it seems the authors drew primarily on journal articles. We suggest that a broader search might have assisted in identifying theories; while journal articles can provide some insight, stakeholders’ theories about how patient portals work are likely to be found in editorials, websites of healthcare providers and patient portal vendors, medical informatics mailing lists, and patient information websites. In a realist review, it is only once the theories have been identified that identification of primary studies takes place. Searching should be purposive and iterative, driven not by the intervention but by the theories.4 For example, if one of the theories suggests that giving patients access to their health record will increase their understanding of their condition and thereby enable them to take a more active role in their care, a relevant search would not only look for primary studies on patient portals but also other interventions that seek to engage patients in their care by increasing their knowledge of their condition. Rather than taking this approach, the search strategy employed by Otte-Trojel et al. 1 is closer to that of a traditional systematic review, with search terms that describe the intervention. Similarly, the choice of outcomes to focus on should be driven where possible by the theories, rather than being based on an existing review as Otte-Trojel et al. 1 have done. In the results section of the paper, the authors describe four mechanisms. We would suggest that the authors’ descriptions of mechanisms focus on resources that the intervention provides, rather than the response of recipients. For example, the mechanism “interpersonal continuity of care” describes how patient portals allow patients to communicate asynchronously with a preferred provider but does not explain what would motivate a patient to do so. In describing context, the authors refer only to organisational context, stating that context at the service unit level and patient-provider level was rarely described in detail. We appreciate that studies do vary in the extent to which context is described. However, an important aspect of context is at the individual level in terms of nature and severity of the patient’s condition. While not identified as a context by the authors, they implicitly acknowledge this as a context when discussing outcomes, pointing to the emphasis in the studies on chronic disease patients and the modest outcomes for patients whose condition is already well controlled. Finally, we feel it is important to acknowledge that different study designs make different contributions to theory testing. From our reading of the paper, Otte-Trojel et al. 1 appear to have treated all studies as potentially providing evidence on contexts, mechanisms, and outcomes. Randomized controlled trials (RCTs) provide information on outcome patterns and, by examining differences in, for example, intervention delivery or patient population, some pointers to likely contextual differences might also be identified. However, RCTs seldom provide information about mechanisms as RCTs are concerned with identifying regularity between a particular intervention and a particular outcome, not with understanding how the intervention changed the reasoning and behavior of recipients. To understand how recipients respond to an intervention, it is necessary to look at qualitative studies, which explore these responses in detail. Realist reviewers would not typically look to qualitative studies for evidence on outcome patterns because such studies rarely explore outcomes and, where they do, small numbers and lack of standardised measurement make it difficult to draw reliable conclusions. We feel Otte-Trojel et al’s 1 findings would have produced more sharply defined Context Mechanism Outcome configurations if they had engaged in a process of knitting together different forms of evidence from different study types as we describe above. Contexts, mechanisms, and outcomes do not just fall out of the primary studies so the realist reviewer has to shuttle between theory and data, integrating the data in imaginative rather than mechanistic ways.5 None.

Concern has been reported about the safe use of medical abbreviations in documents such as handover sheets and medical notes, especially when information is being communicated between staff of different specialties (BBC 2008 BBC. 2008. Available from: http://news.bbc.co.uk/1/hi/health/7171453.stm [Accessed 23 September 2010] [Google Scholar], Sheppard et al. 2008 Sheppard, J. E. 2008. Ambiguous abbreviations: an audit of abbreviations in paediatric note keeping. Archives of Disease in Childhood, 93: 204–206. [Crossref], [PubMed] , [Google Scholar]). This article describes a study to investigate whether the use of symbols in handover documentation that is shared within and between multidisciplinary teams (MDTs) has similar safety implications. We asked 19 healthcare professionals from a range of specialties to identify 45 different combinations of 38 individual symbols. The symbols and combinations of symbols were extracted from 102 handover sheets taken from 6 different healthcare contexts in 4 London hospitals. Three symbols proposed in Microsoft's Common User Interface guidelines for alert symbols were also included. Results reveal that while some symbols are well understood, many others are either ambiguous or unknown. These results have implications for the safe use of symbols in medical documents, including paper and electronic handover documents and Electronic Patient Records (EPRs), especially where teams comprise individuals from different professional backgrounds, i.e. MDTs. We offer initial suggestions for standardisation and further research.

Handovers are a specific kind of multi-disciplinary team meeting. Shift handovers and transfers are both regular features of hospital work but there is currently great variation in how such handovers are conducted, presenting a challenge for those seeking to develop technology to support handover. This article presents the ConStratO model of handover, which captures aspects of the context that influence how the handover is conducted, a range of different handover strategies relating to different aspects of the handover and possible outcomes of handover. The model is based on detailed data collection in a range of clinical settings. We present the model as a tool for developing and evaluating technology support for handover.

New wireless, networked, and sensor-based technologies are increasingly pervasive within health care. We are seeing a growth not only in the availability of hardware and software solutions for heal...

Drawing on data from a multi-site case study of a range of clinical settings, this paper explores the form of nursing handover sheets and the processes through which they are created and updated. We argue that these documents function as both public and private documents, having relevance for the whole ward while also acting as a personal workspace. Such dual functionality needs to be supported by any technology that seeks to provide for the work of handover, if the handover sheet is to continue to act as a space for work, rather than just a repository of information.

This paper considers how adjustable alarms support collaborative monitoring work within the intensive care unit. Drawing on examples from an observational study, it hopes to stimulate new ways of thinking about the role that alarms play in supporting awareness of not only changes in the environment but also awareness of colleagues' actions. Adjustable alarms allow nurses to fit the alarm limits to both the patient state and the nurse's level of experience. The setting of alarm limits is an accountable activity, being visible to and observed by colleagues.