Health care is growing more complex and difficult to manage due to factors such as a rapidly expanding body of knowledge, a shift towards more people living with chronic disease and multi-morbidity [1], challenges in coordinating multiple providers and actors, and, not least, the need to include the preferences and values of the individuals seeking health care [2]. For an increasing number of individuals, interacting with health care means addressing more than just one illness or disorder, undergoing more than one treatment, and collaborating with more than one care provider [1].

Addressing and adapting to complexity might be especially challenging in health care focusing on individuals with severe mental illnesses such as schizophrenia who are disproportionately affected by comorbid medical conditions [3]. With typical onset in early adulthood and a lifelong course, schizophrenia is among the top 10 disorders in terms of disability-adjusted life years lost [4]. Characteristic symptoms are hallucinations, delusions, and disturbances of thought. These features tend to make it challenging to establish and maintain relationships [5]. Treatment failure is common, increasing the risk of suicide [6], despite access to effective treatments [7]. Health and social services for persons with schizophrenia are marked by a high level of complexity. They involve a multimodal treatment approach with a range of treatments from a multiprofessional team, often requiring coordination with other providers (eg, primary care, care for other chronic disorders, social support, housing, and vocational rehabilitation) [5].

Systems can be described either as simple (straight forward and predictable, with few components), complicated (predictable but with more interacting components), or complex (unpredictable and dynamic, where the whole is more than its constituent parts) [8]. Traditional linear cause-and-effect-thinking is not sufficient when studying systems that evolve in ways that are hard or even impossible to predict. Knowledge of complex adaptive systems can aid in understanding and studying health services [9]. A complex adaptive system is a collection of interconnected agents with the freedom to act based on their own internalized rules, affecting each other. These rules, in human-based complex adaptive systems, could be instinct and implicit mental models. Agents adapt in various ways through interactions, which causes the system to change over time [2]. Capability among individual agents in a complex system, that is “the extent to which an individual can adapt to change, generate new knowledge, and continue to improve their performance” [10], can be supported by minimum specifications (simple rules to guide behavior) and feedback loops, letting individuals gradually upgrade their internalized rules through experience. Relevant feedback on performance is crucial to enable continuous learning and improvement at all levels of a health care system, from the level of individual patients to organizational management and policy levels, potentially enabled by the use of new technologies [11-15].

The use of new technologies, argue Pavel et al [16], is essential to achieving personalized, evidence-based, and economically viable health care. Meanwhile, experiences so far reveal significant challenges. Uptake of “disruptive” technologies in health care is slow [17,18], and fundamental quality, safety, and cost problems have not been resolved by digitalization [19]. Moreover, the failure rate of technology projects in health care is high; large and complex projects often tend to fail to deliver anticipated results [20-22]. There seems to be a gap between the development of technology and usefulness in practice within health care organizations that needs to be bridged if technologies are to support health care rather than further increase its complexity [19].

Greenhalgh et al [23] argued that adoption, scale-up, and spread of new technologies often fail due to complexity. They employed theories on complex adaptive systems and on the diffusion of innovations in health care to create a framework for using principles or rules to facilitate the development and application of technological innovations in complex contexts. The Nonadoption, Abandonment, and Challenges to Scale-Up, Spread, and Sustainability (NASSS) framework and complexity assessment tool (NASSS-CAT) were developed to help stakeholders identify and manage complexity in technology innovation projects in health care [8,23].

A university hospital in west Sweden started to develop a point-of-care dashboard to support patients and health care professionals in schizophrenia care. Successful prototyping and pilot testing led to the decision to scale-up the initiative to all the department's outpatient units. Challenges started to accrue when planning for larger-scale development and deployment of the dashboard and related tools. This case study evaluates the stakeholders’ use of the NASSS-CAT to inform the development and deployment of the dashboard and reports on the complexity profile of the project, stakeholders’ experiences when using a pilot version of the NASSS-CAT, and practical implications. This study also aims to inform a future multinational study with multiple cases having maximum intercase variation to field-test the NASSS-CAT [24].

This case study [25,26], informed by the principles of action research [27] and action evaluation [28], involved stakeholders’ use of the NASSS-CAT to inform the development and deployment of the point-of-care dashboard for patients and health care professionals in schizophrenia care. Another study will specifically evaluate patients’ experiences while using the dashboard at the point of care.

The developers of NASSS noted that it was crucial to understand the sociotechnical interaction between individuals, organizations, technology, and policy to explain why a new technology is adopted and sustained (or not) in health and social care [23,29]. The NASSS framework features 7 key domains, identified through systematic hermeneutic literature review and refined through empirical case studies of technology implementation [8]. It is intended to be used to guide and evaluate the success of technology deployment in sociotechnical systems. By addressing questions in the tool’s domains, properties of the technology and adopting system are placed along a continuum ranging from simple to complicated to complex (Figure 1) [8]. Knowledge of a technology project’s domain-specific complexity can aid stakeholders to respond adaptively, lessen complexity, and strengthen their capability to handle complexity [30]. Principles or “simple rules” can act as recommendations to guide further development (Textbox 1) [30].

The NASSS-CAT has two primary components. The initial component, based on the NASSS framework, supports the development of a rich narrative that surfaces key areas of uncertainty and interdependence in the project. The second component, based on an adapted version of the complexity assessment tool by Maylor et al [31], consists of a series of questions to support emergent project planning and evaluation and, in particular, to prompt project teams to consider how they might either reduce or manage complexity across the different NASSS domains.

The Department for Schizophrenia Spectrum Disorders at Sahlgrenska University Hospital in Gothenburg, Sweden (the Department) delivers specialized care for people with psychotic disorders in the metropolitan Gothenburg area (population of approximately 600,000 people). It serves 2600 patients, with schizophrenia as the most common diagnosis, at 8 outpatient units. About 20% of these patients need acute inpatient care at one of the Department’s 5 wards each year. To support patient engagement at the point of care, the Department developed a digital dashboard to visualize key indicators of each patient’s health and care status. The dashboard is one of several connected applications and displays to visualize data fed by several systems developed for several years and piloted at 2 outpatient units with some 400 patients (Figure 2). It includes team tools for care planning and management and tools to support coproduction of health and care among patients, their family members, and psychiatry staff. These tools include a unit-level overview of quality indicators identifying patients at risk, triage and planning tools including support for patient coproduction, a dashboard to be jointly reviewed at the point of care by patients and case managers/psychiatrists to support evaluation and planning, outcomes questionnaires, and patients’ care plans.

Below is a vignette from the fictitious patient Ana’s perspective [32] of a follow-up visit with her case manager at the outpatient clinic. It illustrates how the imagined dashboard-enhanced service model would differ from the traditional service model (pre-dashboard, Textbox 2).

A workshop using NASSS-CAT was set up for the project stakeholders. The focus of the workshop was to increase understanding of the influence of complexity in digitalization projects, determine the complexity profile of the dashboard project, explore the usefulness of the NASSS-CAT, and reflect on ways to manage complexity. The 11 participants were line managers, department directors, organization developers, and programmers representing different professions: psychiatrist, psychologist, occupational therapist, and IT developer. Before the workshop, participants received written and oral information about the study and the voluntary nature of their participation. All invited participants gave their informed consent and agreed to join. The day included both small breakout group sessions and large group discussions. Data collection was framed according to the NASSS-CAT and included observations, field notes, notes from the participants, and audio recordings of discussions during the workshop that were subsequently transcribed. The use of NASSS-CAT yielded a complexity profile and the second part of NASSS-CAT, with questions to prompt consideration of how to handle complexity, was particularly used to identify preliminary practical implications. The complexity profile and practical implications were later presented at two feedback sessions to department directors, managers, developers, and assistants involved in planning the future deployment of the dashboard project. Their discussion of the analysis served to validate the findings as a form of member checking [33] and to deepen understanding of practical implications in relation to the project. Data related to the use and usefulness of the NASSS-CAT from the workshop were analyzed by the authors (AG, ACA) using an inductive thematic approach inspired by Braun and Clarke [34]. The resulting themes and selected illustrative quotes are reported in the Results section in the subsection Experiences With Using NASSS-CAT. The piloting of the NASSS-CAT tools in health care settings within the United Kingdom had been approved by the UK Health Research Authority, Health Research Wales, and Health Research Scotland (IRAS no. 258679; REC no. 19/LO/0550). No formal ethical review was required for piloting the tools in Sweden.

Complexity mapping of the 7 NASSS-CAT domains showed significant complexity in 6 of the 7 domains. The domain of intended adopters (ie, health care professionals and patients at the point of care) was perceived to be the least complex.

The thematic analysis of data for the stakeholders’ experiences of the use and usefulness of NASSS-CAT complexity mapping yielded 3 themes: new insights, threshold to start using NASSS-CAT, and inclusion of relevant stakeholders. These are presented with illustrative quotes from workshop participants. Quotes were translated from Swedish.

By using the second part of the NASSS-CAT in particular, with questions to prompt consideration of how to handle complexity, the simple rules for managing technology projects in complex systems (Textbox 1), and the feedback from directors and managers on the complexity analysis through the feedback and validation sessions, the following practical implications were identified, serving as project-specific recommendations:

The dashboard initiative’s complexity profile – with considerable complexity demonstrated in 6 of 7 domains – indicates that the initiative is unlikely to proceed successfully under current circumstances, reflecting the observation of Maylor et al [31]: “the greater the complexity posed by a project, the lower the chance that any successful outcome, let alone an innovative one, will be achieved.” Nevertheless, using complexity as a lens when assessing the initiative was perceived as meaningful since it revealed not only challenges but also strengths. Developing the dashboard locally resulted in a high level of engagement and readiness to participate among both health care professionals and patients in focus groups, user testing, and pilots. The collaboration and codesign of technologies and work routines in the local development project might explain why participating stakeholders in the workshop perceived the domain of “intended adopters” to be the least complex when considering further deployment of the dashboard. Perhaps the greatest remaining complexity in the dashboard project concerns the challenges of connecting top-down and bottom-up initiatives and processes, whether related to the development and spread of the dashboard or the related technology, goal alignment, or governance.

There are several limitations to this study. It is a small, single case study of an application of the NASSS-CAT to see if it could be useful in a local development and deployment project. It is restricted to use in the local project and does not test the usability and usefulness of the tool at higher levels in the organization, such as macro-level leadership or involvement from the regional IT department, even though the project has strong interdependencies to those parts of the organization. Patient involvement in the use of the complexity assessment tool could have helped gather further useful information, but the involvement of patients has been restricted to the development of the digital dashboard and not the use of NASSS-CAT. Assessing the complexity of a technology project can aid in understanding and planning but may not be enough on its own to identify what needs to be addressed to succeed with scale-up and spread of innovations [36]. Further research is needed to identify how complexity assessment, using the NASSS-CAT, at various levels of organizations from individuals to the top management can support the development and deployment of new technologies in complex health care contexts.

Complexity assessment of the dashboard project using the NASSS-CAT helped highlight important areas and challenges identified through rigorous research as important in the development and deployment projects of new technologies in health care settings. Experiences from the workshop and validation sessions showed that domains that otherwise might have been neglected received more attention and were brought forward to subsequent planning of the project. The assessment identified strengths of the dashboard initiative to further build upon, while also exposing wider organizational complexity that can challenge the process and spread of the initiative. The assessment helped stakeholders generate specific ideas for how to reduce complexity and strengthen the ability to manage any remaining complexity. This pilot testing of the NASSS-CAT in a real-life setting suggested that the NASSS-CAT can provide participants with a greater understanding of complexities in digitalization projects in general.