Högskolan i Skövde

Introduction: Healthcare-associated infections (HAI) increase antibiotic use and contribute to antibiotic resistance, particularly in low- and middle-income countries (LMIC). Hygiene interventions are important and cost-effective strategies for preventing HAIs, reducing both infections, and the need for antibiotics.

Aim: To investigate how hygiene measures in healthcare settings in low- and middle-income countries (LMIC) can contribute to reduced antibiotic use and reduced incidence of healthcare-associated infections (HAI), and to analyze these measures using sense of coherence.

Method: A structured literature review of eleven scientific articles was thematically analyzed based on the three components of SOC comprehensibility, mangeability, and meaningfulness.

Results: Seven subthemes were identified and linked to three themes consisting of the three components of the "sense of coherence" (SOC): comprehensibility, manageability and meaningfulness. Comprehensibility was influenced by healthcare staff's understanding of hygiene routines and guidelines. Manageability was strengthened when resources and organizational support were available but was weakened by resource shortages and high workload. Meaningfulness was linked to the motivation of healthcare workers and the perception that hygiene measures contributed to improved health and reduced transmission of infections.

Conclusion: Hygiene measures in LMICs can reduce HAIs and antibiotic use. Long-term effects require education, resources, organizational support, and more local studies adapted to the conditions in LMICs. From a salutogenic perspective, hygiene measures can strengthen healthcare workers’ sense of coherence (SOC) and contribute to combating antimicrobial resistance.

The purpose of this study is to examine how employees with ethnic minority backgrounds in Swedish working life interpret situations in which inclusion or exclusion becomes relevant in everyday organizational settings, and how these interpretations are reflected in their actions and responses.

The study adopts a qualitative research approach and is based on twelve semi-structured interviews with employees who identify as belonging to an ethnic minority in Sweden. The empirical material was analyzed using thematic analysis following Braun and Clarke (2006). The analysis was guided by Shore et al.’s (2011) inclusion framework and Weick’s (1995) sensemaking perspective.

The findings indicate that experiences of inclusion and exclusion emerge through concrete organizational situations involving issues of belonging, recognition and participation. The study further shows that organizational conditions do not influence individuals directly, but gain meaning through the interpretations individuals make of them. Participants described different ways of responding to such situations, highlighting how participation in working life is shaped through the interaction between organizational conditions, individual interpretations and perceived opportunities for action.

The study contributes to research on inclusion in Swedish working life by deepening the understanding of how employees with ethnic minority backgrounds experience, interpret and respond to situations of inclusion and exclusion in organizational contexts.

The purpose of this study is to examine how middle managers perceive and manage internal crisis communication during the Covid-19 pandemic. The study is based on the assumption that middle managers play a central role in organizational communication by acting as a link between strategic leadership and operational activities. Previous research has primarily focused on external crisis communication or employees’ perspectives, leaving middle managers’ experiences relatively underexplored.

The study was conducted as a qualitative case study with an abductive approach. Data were collected through 15 semi-structured interviews with middle managers and individuals with communicative responsibilities within three industry-related organizations. The material was analyzed using thematic analysis supported by theories of organizational change, sensemaking, sensegiving, and internal crisis communication.

The findings show that internal crisis communication during the pandemic was less concerned with the mere transmission of information and more focused on creating understanding, coordination, and stability in a rapidly changing and uncertain situation. Prior to the pandemic, communication was characterized by physical presence and oral communication channels. During the pandemic, communication practices changed through increased digitalization, improvised solutions, and intensified information flows, making middle managers’ interpretive and translational role increasingly important. After the pandemic, a more hybrid and distributed communication structure emerged in which digital and physical forms of communication were combined.The study also demonstrates that the pandemic exposed pre-existing communicative vulnerabilities, such as person-dependent information flows and unequal access to information, while simultaneously contributing to organizational learning related to digitalization, coordination, and communication practices.

The study contributes by highlighting how middle managers’ communicative work changes across different crisis phases and how internal crisis communication is shaped through the interplay between organizational structures, local interpretations, and sensemaking processes.

Background: Canine parvovirus (CPV) is a leading cause of acute haemorrhagic gastroenteritis and mortality in European dogs. Secondary bacterial coinfection with Escherichia coli, Clostridium perfringens, and Salmonella spp. is widely recognised as a complicating factor, yet no systematic synthesis had previously quantified its independent impact on CPV outcomes in European canine populations.

Objectives: To estimate pooled in-hospital mortality of CPV-infected European dogs, quantify the effect of bacterial coinfection on mortality risk, and characterise geographic patterns of CPV disease severity across Europe.

Methods: A PRISMA 2020-compliant systematic review and meta-analysis was conducted, with the protocol registered on PROSPERO. Searches of PubMed/MEDLINE, Scopus, Web of Science, and DiVA identified 30 eligible European studies (2015–2026). Eight studies (n = 1,107 dogs) were included in the primary random-effects meta-analysis (DerSimonian-Laird method). Quality was assessed by the Newcastle-Ottawa Scale. Heterogeneity (I², Cochran's Q) and publication bias (Egger's test) were formally assessed.

Results: Pooled in-hospital mortality was 19.7% (95% CI: 13.2%–28.4%), representing 206 deaths among 1,107 dogs. Substantial heterogeneity was detected (I² = 87.1%, p < 0.0001); no significant publication bias was found (Egger's p = 0.877). Bacterial coinfection was associated with markedly increased mortality: OR = 4.09 (Alves et al., 2020; p = 0.024) and OR = 8.00 (Duijvestijn et al., 2016; 95% CI: 2.40–26.66; p = 0.001). A North-South geographic mortality gradient was identified, consistent with the predominance of the virulent CPV-2c variant in Southern Europe.

Conclusions: CPV mortality in Europe averages 16–20% under standard hospital care. Bacterial coinfection was strongly associated with markedly increased mortality risk, with odds ratios ranging from 4.09 to 8.00 across the available European studies. These findings support systematic bacteriological assessment in CPV-infected dogs and highlight the need for European randomised controlled trial evidence on targeted antibiotic protocols.

The purpose of this study was to examine how line managers in state-governed organizations experience their workload in relation to personnel responsibility and HR-related administration. The study is based on a qualitative approach using semi-structured interviews with eight line managers. The empirical material was analyzed thematically and interpreted using the “job demand-control-support-model”, the “three-legged stool model”, and “role theory".

The results show that workload is perceived as complex and variable rather than just uniformly high. Some line managers describe periods of high workload, particularly during times of increased work intensity or unclear demands, while others experience their work situation as more manageable over time. Workload is shaped through the interaction between organizational structures, role expectations, and individual conditions, where multiple areas of responsibility frequently need to be handled simultaneously.

The results further show that HR-related tasks have largely been integrated into line managers roles but without a corresponding reduction in other responsibilities, bringing about increased role complexity. The experience of workload is strongly influenced by job control, access to support, and clear understanding in the distribution of responsibilities. Limited support or unclear roles between line managers and HR functions can increase the perceived workload, while for example experience helps structure and manage work tasks without necessarily reducing the overall workload.

Overall, the study shows that line managers workload cannot be understood only in terms of the amount of work, but rather as the result of how demands, resources, and organizational conditions interact. To create a progressively sustainable work situation, measures at the organizational level are therefore required.

The aim of this study is to analyze, through a personnel economics process assessment based on the “Right from the Start” perspective, how HR support within the temporary onboarding process affects the operational time capacity of line and white-collar managers. As well as their conditions for maintaining operational quality within a manufacturing industrial group. The research questions of the study are as follows: What constitutes an onboarding process within a manufacturing industry, and which parts of this process are necessary for a basis that enables a personnel economic analysis? How do managers perceive the contribution of HR support to their operational time capacity, and what obstacles are identified in the collaboration during onboarding? What hidden opportunity costs can be made visible through a personnel economic calculation of managers' time consumption, and how does this affect managers' ability to maintain operational quality according to the principle of "right from the start"? To answer the purpose of the study, a mixed-methods approach was used, meaning that the empirical material was collected through semi-structured interviews with HR as well as line and white-collar managers, supplemented by the managers' time tracking using logbooks. The analysis was conducted in three steps: qualitative thematization of the interview material, quantification of time consumption, and a personnel economic calculation of management payroll and opportunity costs. The results indicate that onboarding is a strategically important support process; certain elements are more general and can be standardized or handled by HR, a mentor, or another support function, while other elements require the operational expertise of managers. The salary-based opportunity cost is calculated to be approximately SEK 700 per new employee. However, this calculation only captures a portion of the total cost; significant hidden costs also consist of displaced management work, production losses, mentorship and extended learning curves.

Background: This study examines how generative artificial intelligence (GenAI) affects accounting work in accounting firms. The background is based on the fact that digitalisation and AI have transformed the industry by automating routine tasks. At the same time, they have created new opportunities for analysis, information processing, and decision support. This development has led to accounting work increasingly being characterised by control, review, and follow-up of AI-generated outputs. This in turn places higher demands on technical understanding, analytical skills, and professional judgement. Despite increased automation, the need for human control remains to ensure the quality and reliability of financial information.

Purpose: The purpose of the study is to investigate how generative AI changes accounting work in accounting firms and how employees experience these changes in their daily work. The study is based on a socio-technical perspective, where the interaction between technology and humans is central to understanding how work processes and roles are changing.

Theoretical framework: The theoretical framework shows that previous research describes how GenAI both increases efficiency in work tasks and changes the organisation of work. The technology affects not only efficiency but also responsibility allocation and decision-making processes. The boundary between human and technical work thus becomes less distinct. This means that accounting professionals increasingly need to interpret and quality-assure AI-generated outputs. At the same time, it is emphasised that the impact of technology depends on organisational factors such as competence, routines, and trust in the systems.

Method: This study uses a qualitative method. The empirical data was collected through semi-structured interviews with six professionals within the accounting industry. The material was analysed using thematic analysis and complemented with peer-reviewed scientific literature. The study is case-based which means that the results are therefore not statistically generalisable, but aim to provide a deeper understanding of the phenomenon. Ethical principles have been considered through anonymisation, informed consent and voluntary participation.

Conclusion: The conclusion shows that generative AI increases efficiency in accounting work by automating routine tasks and thereby reducing manual work. This leads to a shift in roles from data entry towards more control, analysis, and quality assurance. At the same time, the demands for technical understanding, critical thinking, and the ability to work with AI systems increase. The study shows that this development should be understood as a socio-technical change where technology and organisation interact. Limitations include a small sample size and a qualitative method, meaning that the results cannot be generalised to the entire industry.

Background: Threats and violence against community-supporting professions, such as teachers and social workers, are an increasing problem. The consequences are work environments consisting of unsafety and fear of recurring components of unpredictability in the encounter with third parties.

Purpose: The study examines which working methods middle managers can apply to tribute to the perceived security for employees in jobs where they encounter human third parties.

Method: The study was conducted through a qualitative research approach with semi-structured interviews as the research method. The collected empirical data and the theoretical frame of reference are analyzed through an abductive method. Antonovsky's (2005) 'sense of context' (SoC) is used as an analyzemodel for how middle managers can create safety for employees, who are at risk of being exposed to threats, violence or other unpredictability from third parties, through the dimensions; comprehensibility, manageability and meaningfulness.

Conclusion: The sense of safety is not created through one single action but through a relationship-oriented work-method where the middle manager's behavior is crucial. This study shows that through three central methods, the middle manager can create safety for employees in jobs where they encounter third parties that can be unpredictable. These work-methods are; to be present, to be relationship-oriented and to maintain a clear structure in the workplace.

Introduction: Digital tools are increasingly used in psychiatric care to support documentation, information management and clinical decision support. In compulsory psychiatric care, where care situations are often complex and patients may be particularly vulnerable, it is important to understand how these tools affect both patient and staff safety.

Aim: The aim of this study was to analyze how digital tools in compulsory psychiatric care influence patient and staff safety and to identify opportunities and risks associated with digital tools in this context. The study also aimed to examine which digital tools are used in compulsory psychiatric care to support patient and staff safety, as well as their impact on documentation, information flow and decision support.

Method: The study was conducted as a structured literature review with a qualitative approach. A total of 20 scientific original articles were included, identified through systematic searches in the databases PubMed, CINAHL, Scopus and PsycINFO. The articles were analyzed using thematic analysis according to Braun and Clarke (2006).

Results: Four themes were identified: (1) Digital tools in compulsory psychiatric care, (2) Impact on documentation, information flow and decision support, (3) Opportunities with digital tools, and (4) Risks and challenges with digital tools. The results show that digital tools can improve risk identification, documentation structure and clinical decision support, and contribute to safer care processes. At the same time, risks related to usability, implementation and information management were identified, which may lead to increased workload and erroneous decisions.

Conclusion: Digital tools can both strengthen and challenge safety in compulsory psychiatric care. The effects are context- dependent and influenced by how the systems are implemented and used in practice. A holistic perspective is required to ensure safe and effective care.

This systematic review examines structural and functional neural differences between violent offenders and non-violent offenders, with additional comparison to non-offending controls. Using magnetic resonance imaging and functional magnetic resonance imaging evidence published between 2000-2026, 13 studies were synthesized for the results to investigate brain regions associated with emotion regulation, impulse control, and executive functioning. Findings indicate that violent offenders exhibit consistent alterations in fronto-limbic networks, including reduced connectivity between amygdala and medial prefrontal cortex, alongside heightened limbic reactivity. Structurally, they display both decreases and increases in grey matter across prefrontal, temporal, and subcortical regions, highlighting heterogeneous neural profiles. These findings are frequently linked to psychopathy, antisocial and impulsive traits, and may be influenced by psychiatric comorbidities. In contrast, non-violent offenders tend to exhibit more stable prefrontal regulation and adaptive connectivity during emotional provocation. Despite converging evidence for impaired top-down control in violent offenders, findings also suggest that regulatory functioning is context-dependent rather than uniformly deficient. Developmental and genetic factors, including early network disruptions and serotonergic mechanisms, further contribute to variability in aggressive behaviour. Overall, the results support a multidimensional neurocognitive framework in which violent behaviour arises from dysregulated integration of emotional, cognitive, and social processes. However, these neural characteristics should be interpreted as probabilistic risk factors rather than deterministic causes. This highlights the importance of integrating neuroscientific, psychological, and environmental perspectives in forensic assessment, while emphasizing the need for larger, more diverse, and longitudinal studies to improve understanding, prevention, and rehabilitation strategies.

Stress-related non-communicable diseases are a significant global health problem. This study aimed to investigate the impact of virtual reality (VR) environments on physiological and psychological stress recovery, particularly in comparing nature and urban environments, while also considering the impact of presence. Using a counterbalanced within-subject design, 20 participants were exposed to five-minute virtual nature and urban environments. Objective stress was measured with the Root Square of Successive Differences (RMSSD) of heart rate variability, while subjective stress was measured using the Visual Analogue Scale (VAS) and the Stress-Energy Questionnaire (SEQ). Presence was measured using the Igroup Presence Questionnaire (IPQ). Results showed no significant differences in RMSSD across conditions. However, the VAS scores were significantly higher in the urban condition compared to baseline and nature. Additionally, the SEQ showed significantly lower energy levels in the urban condition compared to nature. No significant results were found in experienced presence between the environments, which suggests that the results were not influenced by varying levels of immersion. In conclusion, while virtual nature did not significantly reduce physiological stress, the urban environment had a negative impact (i. e., it increased perceived stress and decreased energy). These findings suggest that an urban environment may provoke a stress response or ‘urban discord’ rather than nature providing immediate recovery in this specific experiment.

Introduction: Municipalities diverge from laws and recommendations that exist in municipal accounting, resulting in non-uniform annual reports. In addition, the information disclosed is insufficient, which means that the stakeholders’ primary source of information, the annual report, is not satisfactory. Previous research finds that municipalities compliance and harmonization with laws and standards is deficient from both national and international perspectives. 

Purpose: The purpose of this study is to examine compliance and thus the harmonization of a specific recommendation issued by the Council for Municipal Accounting among individual municipalities. 

Method: A qualitative document study, along with a complementary interview study, has been conducted to examine how compliance with the recommendation R14 Drift- och investeringsredovisning appeared in the municipalities of Mariestad, Falköping, Ulricehamn and Skara. The empirical data were analyzed using thematic analysis. 

Empirical Findings: Based on the document study, it is possible to illustrate both compliance with and deviations from the normative sentences of the recommendation. Furthermore, the interviews reveal that the recommendation is not fully integrated into the budgeting process or the preparation of the annual report but are instead treated as a reference document. All municipalities describe participating in networks and education where the recommendation can be discussed. Analysis and 

Discussion: The accounting practices observed are not always in line with the accounting regulation, which hinders compliance and thus harmonization. This study confirms the results of previous studies indicating that compliance and harmonization is not complete. 

Conclusion: Compliance with and thus the harmonization of recommendation R14 is not complete, as all municipalities deviate from the recommendation.  

This study examines the challenges that small and medium-sized enterprises in Sweden face when integrating artificial intelligence into their existing accounting systems. The study is based on a case study with semi-structured interviews from two companies.The findings show that AI integration is largely controlled by system suppliers, which limits the companies’ ability to influence the solutions. AI is mainly used in standardized work tasks, where it can increase efficiency but still requires human control. As a result, accounting work shifts from manual tasks to more review, assessment and control. The study also shows that the use of AI varies depending on customers’ working methods and digital maturity. AI can make work more efficient, but it can also make tasks more monotonous and affect how new employees learn the profession. Time, competence, training and employee involvement are important factors for how AI can be implemented and used in practice. Data security and regulations are also central issues, as accounting work involves sensitive financial information. In conclusion, the study shows that AI is not only a technical change, but also affects the organization, professional roles and the skills required from employees. The study contributes knowledge about how companies can prepare for AI integration, for example by planning training, creating routines for controlling AI, protecting data and understanding how dependence on system suppliers can affect the business.

Introduction: Tenant owner associations stand before a regulatory challenge of changing accounting frameworks from the rules-based K2 to the principles based K3. The motivation for the change is, according to BFN, to give a more representative picture of the association's financial position and to make comparisons easier between them and for buyers. But the responsibility for implementation of these changes that the more complex K3 brings falls to the boards of the associations. The change of accounting frameworks creates new requirements of comprehension and their perceptions on the accounting changes' effects on the management, as these boards often consist of laymen.

Purpose: This essay aims to map how the boards in tenant owned associations view the transition to K3 and how this influences the economic management. The purpose is to increase the understanding of the challenges and opportunities that appear when accounting requirements change.

Method: This essay has used a qualitative method. It has used a strategic selection of respondents where the head chairperson in housing cooperatives have been prioritized and included two accountants with specialization in housing cooperatives have also been included for data triangulation. Semi structured interviews were used to gather data which has been analyzed with the help of a thematic analysis on the empirical data and with the help of our analysis model in order to visualize the sensemaking process.

Results: The results show, through thematic analysis, that the transition has brought some changes in the management process. The board members' own competencies have influenced their view in the change and the need for external help is found to be large, and that the use and comparability of the accounting seems to have been improved.

Analysis and discussion: The analysis shows that the boards work with the accounting change, based on the sensemaking attributes, and that these show up in the actual management of the housing cooperatives. The discussion centers around how K3 gives better information to external stakeholders, but that the boards seem to have difficulties in seeing the benefit in their management.

Conclusions: The perceptions about the accounting framework change to K3 and its influence on the management of the tenant-owner associations are scattered. For some associations it has led to changes in their practical management and to introduce their accounting as a comparison tool between associations. But it has also been seen as just a technical change by others in the regulations that bring no practical benefit to the management.

Bakgrund: Personalomsättning är en återkommande utmaning för många organisationer och kan medföra konsekvenser såsom kompetensförlust, minskad kontinuitet och ökade rekryteringskostnader. Tidigare forskning har visat att arbetsutformning kan påverka både arbetsengagemang och medarbetares intention att stanna kvar i organisationen. Forskning har dock främst fokuserat på statistiska samband mellan olika arbetsrelaterade faktorer, medan mindre uppmärksamhet riktats mot hur medarbetare själva upplever dessa samband i sin arbetsvardag.

Syftet: Syftet med studien är att beskriva och analysera hur medarbetares upplevelser av arbetsutformning påverkar intentionen att stanna kvar i organisationen.

Teoretisk referensram: Studien utgår från den arbetskaraktäristiska modellen (Hackman & Oldham, 1976), där autonomi, variation och återkoppling används för att analysera och förstå arbetsutformning. Arbetsengagemang används som en mellanliggande variabel för att analysera hur upplevelser av arbetsutformning relateras till medarbetares intention att stanna kvar i organisationen.

Metod: Studien genomfördes som en kvalitativ fallstudie vid ett IKEA - varuhus. Det empiriska materialet samlades in genom semistrukturerade intervjuer med åtta medarbetare från olika avdelningar. Materialet analyserades med stöd av studien analysmodell, där autonomi, variation och återkoppling relateras till arbetsengagemang och intentionen att stanna kvar.

Resultat och analys: Resultatet visar att autonomi, variation och återkoppling upplevs ha betydelse för medarbetarnas arbetsengagemang. Möjligheten att påverka det egna arbetet, möta varierande arbetsuppgifter och få återkoppling genom kunder, kollegor och chefer bidrar till att arbetet upplevs som stimulerande, utvecklande och meningsfullt. Analysen visar att dessa aspekter av arbetsutformning bidrar till arbetsengagemang, vilket i sin tur relateras till medarbetarnas intention att stanna kvar i organisationen. Resultatet visar även att autonomi, variation och återkoppling inte upplevs som isolerade faktorer utan som sammanlänkade delar av arbetsvardagen.

Slutsats: Studien bidrar med en kvalitativ förståelse av hur arbetsutformning upplevs av medarbetare och hur autonomi, variation och återkoppling tillsammans kan bidra till arbetsengagemang och stärkt intention att stanna kvar i organisationen. Resultatet fördjupar därmed förståelsen av sambandet mellan arbetsutformning, arbetsengagemang och retention ur ett medarbetarperspektiv.

This study aims to highlight how leaders' communicative practices in hybrid work contribute to creating direction, relationships, and cohesion within organizations. The study is based on a communicative perspective on leadership, where leadership is understood as something that is constituted through interaction between leaders and employees. To analyze these relationships, trust is used as an analytical concept, where competence, benevolence, integrity, and legitimacy constitute analytical dimensions.

The theoretical framework is based on the perspective of the Communicative Constitution of Organizations (CCO) and Relational Leadership Theory (RLT). The study is based on a qualitative method, and the data collection was conducted through twelve semi-structured interviews with leaders working in organizations where hybrid work is practiced.

The results show that communication constitutes a central prerequisite for leadership in hybrid work. Leadership is shaped through recurring communicative actions such as meetings, check-ins, and feedback, which contribute to creating structure, coordination, and direction in work. The results also show that hybrid work places greater demands on leaders' ability to actively create and maintain relationships with employees. Trust emerges as a fundamental prerequisite for leadership in hybrid work and develops through communicative processes in which competence, benevolence, integrity, and legitimacy are continuously reinforced in the interaction between leaders and employees.

The conclusion shows that communication does not merely function as a tool for leadership but constitutes the very foundation through which leadership is created and maintained in hybrid work.

Background: Remote work can contribute to flexibility, work engagement and organizational benefits such as higher performance and reduced employee turnover. At the same time, the working form changes the conditions for managers' work environment responsibility through reduced physical presence, increased demands on communication and trust, and difficulties in identifying and managing psychosocial risks. Previous research has primarily focused on employees' experiences of remote work and its work environment consequences, and therefore this study examines how managers adapt their leadership to achieve work environment responsibility when working remotely.

Purpose: The purpose of the study is to create an understanding of how managers in the public sector adapt their leadership to manage work environment responsibilities during remote work, with a particular focus on promoting a good psychosocial work environment for employees. The study is based on a managerial perspective with the following question: How do managers adapt their leadership based on employees' psychosocial work environment during remote work?

Design/methodology/approach: The main method of the study was a qualitative method where data was collected through semi-structured individual interviews, with ten managers who exercise work environment responsibility for employees working remotely in the public sector.

Findings: The most important result shows that managers adapt their leadership to remote work through communication, relationship-building work, task orientation and clear boundaries between work and private life. Digital tools enable follow-up and collaboration, but at the same time bring challenges related to social interaction and the opportunity to pay attention to employee well-being. However, physical meetings appear as an important complement to remote work to strengthen relationships and enable earlier detection of ill health.

Cybercrime presents increasing challenges to digital societies, yet population-level research on offenders remains limited, particularly in Sweden. This longitudinal study used national register data (2000–2022) to examine demographics of Swedish cybercrime offenders. Cybercrimes included online fraud, crimes against children, online threats, and other digital offenses. The proportion of offenders increased steadily over time, with men twice as likely as women to be involved. Most offenders were aged 20–39 and had low income and low educational attainment. Prior offline crime was common, suggesting overlapping criminal trajectories. Divorced and never-married individuals had higher odds of committing cybercrime. Socioeconomic profiles varied by offense type; child exploitation offenders were typically older with fewer offline crimes. Findings highlight the need for targeted interventions and further data to understand and prevent cyber-offending effectively.

Hybrid work has become an increasingly common way of organizing work in knowledge-intensive organizations. Previous research has primarily focused on the challenges associated with hybrid work, particularly regarding leadership, communication and coordination. Against this background, this study aims to examine how everyday leadership practices are enacted in hybrid work within knowledge-intensive organizations and how these practices enable and constrain the creation of shared direction in hybrid work. 

The study is grounded in Leadership as Practice (LAP), where leadership is understood as something that emerges through everyday organizational practices and interactions rather than through individual leadership styles or managerial traits. To examine how leadership practices influence the organization of work, responsibility, autonomy and participation are used as analytical dimensions. These dimensions are operationalized through the concepts of positioning, acting spacing and co-orientation. 

The study applies a qualitative research approach based on nine semi-structured interviews with managers in knowledge-intensive organizations where hybrid work has been established over time. The empirical material was analyzed through thematic analysis using an abductive approach. 

The findings show that leadership in hybrid work is largely practiced through organizational and communicative practices such as meeting structures, digital coordination, follow-up processes, clear responsibility distribution and continuous dialogue. The study also demonstrates that hybrid work is not solely dependent on flexibility and autonomy, but on how work is organized through shared practices, coordination and clear structures. Furthermore, the findings indicate that autonomy and control do not exist as opposites, but rather coexist through different leadership practices and organizational frameworks. 

The conclusion suggests that hybrid work cannot only be understood as a flexible work arrangement, but rather as an organizational transformation that changes the conditions for leadership, coordination and followership. The study highlights how everyday leadership practices enable and constrain responsibility, autonomy and participation and how these dimensions together create conditions for shared direction in hybrid work. 

The indie game industry is growing quickly, nevertheless creating professional visual art remains a challenge for small teams and solo developers. This study investigates whether players perceive a difference in stylistic consistency between AI-generated and human-created 2D pixel art assets when experienced within a playable game environment. Two versions of a simple 2D platformer prototype were made in Unity, one with assets generated by PixelLab and one with a human-created asset pack from Unity’s asset store. Eleven participants played both versions without being told which is which, they then completed a structured survey. Findings indicate a preference towards the human-created version in terms of stylistic consistency, with nine out of eleven participants rating it as more consistent. Player feedback pointed to blurriness, animation issues, and general lack of polish in the AI version. Which suggests that AI-generated pixel art, without manual refinement, does not yet meet players' visual expectations from indie games.

Background: Digitalization has transformed the role of the Chief Financial Officer (CFO) from traditional accounting into a more strategic leadership position. Generative AI now offers new opportunities to support decision-making by rapidly processing and structuring large volumes of unstructured data. At the same time, the technology introduces new and complex challenges related to limited transparency, information security, and the risk that users rely uncritically on AI-generated results. Understanding how CFOs actually manage this balance between efficiency and control in practice remains limited.

Purpose: The purpose of this study is to describe the role that generative AI plays in CFOs’ decision-making, and to contribute to a nuanced understanding of how the technology is used in practice when working with financial decision-support material.

Method: The study was conducted using a qualitative research approach and a deductive perspective. Empirical data was collected through semi-structured interviews with six CFOs working in different industries. The material was then analyzed thematically based on Herbert Simon’s decision-making model and the theory of bounded rationality.

Conclusion: The results show that generative AI primarily serves as support in the early phases of the decision-making process—the intelligence and design phases—where the technology is used to structure information and generate decision alternatives. At the same time, AI does not eliminate bounded rationality, but may instead create new cognitive risks such as confirmation bias, anchoring bias, and uncritical trust. The final decision in the choice phase, however, still requires human judgment. Consequently, the CFO’s role shifts from administrative data collection toward a more strategic, supervisory, and critically evaluative function.

Background: Artificial Intelligence (AI) is being embedded into the hospitality industry by using tools like AI-powered chatbots, recommendation systems and personalized marketing, Hotels implement these tools to foster operational efficiency and improve personalized service to its guests to build customer trust. While AI has its benefits in terms of optimizing service quality, concerns regarding ethics, privacy and transparency have remained. Most existing studies on this topic were quantitative analysis focused on managerial aspects of AI adoption, leaving a gap for a qualitative research for better insight into customer’ perception of AI in hotels.

Purpose: This study investigates the perception and experiences of hotel customers with AI-powered tools and its influence on customer trust in the hotel. We apply the SERVQUAL theory which seeks to find the gap between guests’ expectation and actual perception of experience or service received. The Commitment-Trust theory of relationship which stipulates the relevance of commitment and trust in building long-term customer relationship is used to complement SERVQUAL. With these two theories, the study aims to examine whether quality of Service rendered by AI has any impact on their trust.

Method: Using a purposive and convenience sampling method, interviews were conducted with 20 participants who have resided in hotels in Sweden. After transcription and anonymization of collected data, an inductive thematic analysis method was used to generate code. Related codes were then grouped to form distinctive themes. Created themes were backed by participant quotes.

Conclusion: After in-depth investigations, the findings suggest that although customers recognize the relevance of AI for increased operational efficiency, they valued human interaction, especially in a service-oriented establishment like a hotel over AI-powered tools. Evidence indicates that the quality of service rendered by AI tools in hotels was not perceived as a key factor influencing customer trust in the hotel, but rather other factors such as service quality received from human employees, the price and overall physical condition of the hotel had a bigger impact on customer trust. Findings suggest that relational trust in the hospitality industry continues to be intimately dependent on empathy and accountability only humans can offer, and the grade of physical service offered in a hotel.

Introduction: Overweight and obesity are rapidly growing global health challenges that are largely shaped by socioeconomic disparities and not solely by individual lifestyle choices.

Aim: The aim of this study was to investigate how socioeconomic factors influence the prevalence of childhood obesity in Sweden.

Methods: A structured literature review was conducted based on 20 original scientific articles published between 2021 and 2026. Data analysis was performed using thematic analysis according to Braun and Clarke (2006).

Results: Based on the thematic analysis, two main themes were identified: Socioeconomic inequalities and Lifestyle habits. The results showed that the parents’ educational level, economic status and living environment are strongly associated with higher BMI in children. Furthermore, it emerged that factors such as parent’s activity patterns, diet-related behaviors, and psychosocial stress increase the risk of overweight and obesity among children in Sweden.

Conclusion: Children’s health is influenced by the complex interplay between living conditions and underlying social structures. To create the conditions for and promote health equity, interventions are needed to address these inequalities, which arise early in children’s lives.

Diabetes mellitus is characterized by dysregulated glucose homeostasis involving the pancreatic hormones insulin and glucagon. While these hormones are traditionally regarded as physiological antagonists, previous studies have suggested that direct molecular interactions between them may also occur under certain conditions. The aim of this study was to investigate whether insulin–glucagon complex formation could be detected in surface-based fluorescence assays and whether such interactions influence molecular stability or antibody epitope accessibility. Experiments were performed using LigandTracer technology together with antibody-based sandwich assays, disturbance assays and serial dilution experiments. Multiple assay configurations involving anti-glucagon and anti-insulin antibodies were evaluated using fluorescently labelled ligands and nonspecific control proteins. Several sandwich assays demonstrated altered fluorescence following insulin addition, suggesting possible modulation of glucagon epitope accessibility. In addition, FITC-insulin serial dilution experiments showed reproducible concentration associated signal progression that differed from nonspecific control conditions. Control experiments further indicated that the observed signal depended on the presence of glucagon. Due to the complexity of the experimental system and technical limitations associated with surface-based fluorescence measurements, the findings should be interpreted cautiously. Nevertheless, the collective results are consistent with possible insulin associated modulation of glucagon and support further investigation of insulin–glucagon interactions using complementary biophysical approaches.

The study examines managers’ conditions for a sustainable working life with a focus on self-leadership. The background to the study is the high workload experienced by managers, which often leads to stress and emotional exhaustion. Self-leadership is often viewed as a tool for managing the work situation, but there is a lack of knowledge about how it is influenced by the social context. The purpose of the study is to contribute to an understanding of how the social context affects managers’ ability to engage in self-leadership in order to promote a sustainable working life. The study has a qualitative design and is based on nine semi-structured interviews with unit managers within municipal healthcare services. The empirical material has been analyzed through thematic analysis. In addition, the study’s theoretical framework consists of a combination of the Demand-Control-Support Model (DCS) and Self-Determination Theory (SDT). By using these theories, the interaction between the organizational context, that is, the external factors, and the individual’s basic needs for autonomy, competence, and relatedness, that is, the internal factors, has been analyzed. The results show that managers’ self-leadership is not an isolated individual ability, but is largely shaped, enabled, and constrained by the surrounding environment. The study concludes that responsibility for a sustainable working life cannot be placed solely on the individual manager’s self-leadership. For self-leadership to function in a developmental way, an organizational environment is required that balances demands with sufficient decision latitude and social support. The study thus contributes to shifting the focus from individual performance to the organization’s responsibility for creating sustainable conditions.

Predicting the long-term mechanical degradation of automotive thermoplastics from accelerated ageing data is a challenging problem in industrial materials informatics, especially when random validation does not account for the hierarchical structure of material grades. This thesis investigates how well machine learning can predict elongation‑at‑break retention in polycarbonate/acrylonitrile‑butadiene‑styrene (PC/ABS) and polypropylene (PP) under thermal ageing, and how feature design and validation strategy influence performance. A dataset of 331 specimens from 32 mixed grades provided by Volvo Cars was analysed using six machine-learning algorithms (Lasso, support vector regression, random forest, XGBoost, Gaussian process regression, and a multilayer perceptron) across three scenarios: group‑aware cross-validation, condition‑based forecasting, and temperature‑transfer prediction. The features included polymer class, filler content, ageing descriptors, Arrhenius‑inspired thermal dose, mechanical retention measurements, and FTIR‑derived degradation indices.

XGBoost achieved the highest predictive performance across all the validation scenarios. The results showed an R² value of 0.35 for group-aware validation by material grade, whereas the performance was higher for condition-based forecasting (R²=0.69) and temperature‑transfer (R²=0.73). These results demonstrate that the estimated predictive performance is strongly influenced by validation design and increases when material-specific information remains partially represented across the train-test split. Feature selection analysis was done to assess the contribution of different features groups. The results showed that mechanical retention features contribute substantially to predictive performance while FTIR-derived degradation indices did not provide a consistent additional predictive signal both with and without mechanical-retention features. The SHAP analysis identified filler content, mechanical retention features and physics-inspired Arrhenius thermal dose as most influential predictors. Class-specific analysis showed that the predictive performance differs largely between PP and PC/ABS. The larger PP subset supports stronger forecasting and temperature-transfer performance, while the smaller PC/ABS subset shows weaker and less stable performance, mainly in condition-based forecasting.

Overall, the work provides a rigorous evaluation framework for machine-learning based prediction in small, heterogeneous and complex polymer‑ageing datasets. The finding shows how validation design, feature representation, polymer class, and data sparsity influence the reliability of ML‑based degradation prediction.

This thesis explores how inherited and persistent mechanics can function as worldbuilding structures within a narrative artefact, without relying on explicit exposition or traditional progression systems. The study is based on Soul Art, an original practice-based narrative and system design artefact centered on reincarnation, inherited abilities, soul-bound weapons, and irreversible consequence. Using research-through-design and reflective practice, the thesis examines how mechanics can communicate world logic, identity, memory, and sacrifice through systemic continuity. The artefact was evaluated through qualitative reflective analysis and a small-scale pilot test with three participants. The findings suggest that participants were able to infer key aspects of the world’s structure, death system, and thematic meaning from the mechanics alone. The study argues that persistent and inherited mechanics can operate not only as gameplay systems, but as narrative and worldbuilding structures within speculative game design. 

In recent years, anime-style games have developed rapidly, but in-depth research on how players perceive narrative in this genre remains lacking. This study explores how players perceive narrative in anime-style games. The research employed a mixed approach, combining questionnaires (N=105), open-ended questions (N=54), and semi-structured interviews (N=9). The results show that narrative is considered the most important element, even surpassing gameplay and art. Players view narrative as the “soul” of the genre, driving emotional investment, retention, and payment. However, players actively skip low-quality or repetitive narrative, suggesting a selective engagement pattern. Anime art style is inseparable from the narrative experience. The study concludes that narrative is a core dimension in evaluating anime-style games, but its influence depends on the quality and presentation of the narrative.

Environments in interactive media are increasingly designed to respond to participant behaviour, yet the relationship between the logic of that responsiveness and the narrative meanings participants construct remains poorly understood. This study investigates how predefined and adaptive forms of environmental responsiveness influence narrative interpretation in immersive interactive environments. A purpose-built virtual environment was developed in Unreal Engine 5, implemented in two versions identical in appearance but differing in their responsiveness logic. Five participants explored one version each in a between-subjects design. Behavioural data, quantitative ratings, and qualitative reflective responses were collected and analysed. Predefined responsiveness produced rule-based, object-anchored narrative interpretation with higher perceived predictability. Adaptive responsiveness produced atmospheric, history-oriented narrative interpretation with higher perceived meaningfulness and intentionality despite lower predictability. The findings suggest that predictability and intentionality are separable dimensions of environmental experience, and that opacity in a responsive system can generate deeper narrative engagement than transparency.

Background: Electronic health records (EHRs) play a vital role in healthcare organizations to manage patients’ health information. Implementing or transitioning to EHRs is often a large-scale implementation with many challenges. Organizational and individual readiness is crucial for successful adoption and implementation of an EHR system. The objective of this study was to explore clinicians’ and administrative staff’s experiences of a large-scale EHR implementation (Millennium) that was terminated shortly after its initial deployment.

Methods: A qualitative design involving individual online interviews with 73 clinicians and staff investigated end users’ experiences after a large-scale implementation of EHRs in Sweden. A semi-structured interview guide was used during the interviews and included broad questions on individual and collective readiness to change before, during and after the implementation. Interviews were recorded and transcribed. The data were analysed using reflexive thematic analysis.

Results: The findings show that experiences from a failed implementation highlight inadequate communication with the management level, inadequate preparedness in understanding the change and education on how to use the system. The results also show that the end users felt deceived, disappointed and overwhelmed with stress. End users highlighted the need to be involved in the implementation process and understand the whole picture. Lack of dialogue and feedback between co-workers and top management negatively affected organizational and individual readiness. Incomplete readiness and system contributed to anxiety, feelings of abandonment and being in an abstract reality.

Conclusions: The findings provide an empirically grounded account of a failed large-scale implementation, reinforcing and extending key insights from implementation science. The study underscores that implementation is inherently a socio-organizational process, where meaningful end-user involvement and continuous, bidirectional communication between frontline staff and management are not peripheral but constitutive conditions for successful change. Anchored in Organizational Readiness for Change, the findings illustrate how limited readiness, combined with misalignment between system design and local context, can drive non-adoption and rapid abandonment. The system’s positioning as a comprehensive, modern solution raised expectations that were not met in practice, resulting in a sense of emptiness and disappointment among the participants, echoing the tale of the emperor’s new clothes.

Background: Sterile water injections have been shown to be safe and efficacious in the management of childbirth pain, particularly labour back pain. The only commonly experienced side effect of the technique is the significant injection pain. Cryotherapy, in the form of vapocoolant sprays or cold packs has been shown in previous studies to reduce injection pain generally. However, it is not known if cryotherapy would mitigate the pain of sterile water injections.

Objective: To assess the efficacy of vapocoolant spray or chemical cold pack in reducing sterile water injection pain for labour back pain.

Design: Three-arm multicentre open label superiority randomised controlled trial. Setting: Three maternity hospitals in Brisbane, Australia.

Methods: Consenting participants were allocated 1:1:1 using an independently produced randomisation schedule to either the application of a vapocoolant spray, chemical cold pack or standard care (no cryotherapy). The primary outcome was the difference between groups in self-reported visual analogue pain scale at time of injection. Secondary outcomes included skin temperature at the time of injection and visual analogue pain scores for back pain up to 120 min following injection.

Results: Between February 2024 and June 2025 a total of 133 women randomised to either vapocoolant spray (n = 45), chemical cold pack (n = 45) or control group (n = 44). Consent forms were not completed for four participants, and four withdrew. There was no difference in visual analogue pain scores at injection between groups (control: mean 84.73 mm (SD 22.75); vapocoolant: mean 81.23 mm (SD = 24.94); chemical ice pack: mean 86.89 mm (SD = 18.21)) (p = 0.51). Skin temperature in both the vapocoolant (median 29.05 IQR 25.9, 31.8) and chemical cold pack (median 27.23, IQR 23.8, 30.5) was lower than the control group (median 32.75, IQR 31.8, 34.2) (p = <0.001). All three arms showed lower post injection pain scores across all time points.

Conclusion: Neither the application of a vapocoolant spray or chemical cold pack prior to sterile water injections resulted in a statistically significant reduction in associated pain. The use of cryotherapy to the mitigate pain of sterile water injections cannot be recommended. Trial registration: The trial is registered at the Australian and New Zealand Clinical Trials Registry (Trial ID: ACTRN12623000585628). Registration date: 29/05/2023. First recruitment 21st February 2024.

Breast cancer is the most commonly diagnosed cancer worldwide, yet its proteomic profile shows substantial overlap with other cancer types, making differential diagnosis challenging. This study aimed to identify a panel of plasma protein biomarkers capable of distinguishing breast cancer from eleven other cancer types using a publicly available pan-cancer proteomic dataset (Álvez et al., 2023), comprising 1,477 patients and approximately 1,463 proteins measured by the Proximity Extension Assay (PEA). 

Three supervised machine learning (ML) models were applied: Elastic Net Regression (ENR), Random Forest (RF), and Support Vector Machine (SVM), to identify the most discriminative proteins for breast cancer classification. Variable importance scores from all three models were combined to generate a consensus-ranked panel of 70 proteins, with five proteins (FLT3, CD244, ENPP2, JCHAIN, and AHSP) consistently identified across all models. Each model was trained and evaluated using four protein subsets (top 10, top 30, top 70, and all proteins) and assessed using accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Model robustness was further evaluated using repeated 10-fold cross-validation (20 repetitions). 

Classification performance improved with increasing protein panel size across all models. When all proteins were included, the ENR model achieved the best performance, with an accuracy of 0.72 and an AUC of 0.78 on the independent test dataset. A reduced 30-protein panel retained competitive discriminative ability (AUC = 0.71–0.73), supporting its potential for future targeted assay development. In repeated cross-validation, ENR demonstrated the highest and most stable performance (mean accuracy = 74.2%, mean Kappa = 0.48), outperforming RF. Functional enrichment analysis revealed that the top candidate proteins were enriched in cell adhesion, immune regulation, and IL-17 signaling pathways, consistent with known mechanisms of tumor invasion and immune evasion in breast cancer. However, the identified proteins should be considered candidate biomarkers, and their diagnostic utility requires validation in independent external cohorts and prospective clinical studies. 

These findings demonstrate that plasma proteomics combined with ML can distinguish breast cancer from other cancer types in a pan-cancer setting and identify promising candidate proteins for further experimental validation and assessment in future clinical studies. 

In line with the rapid digitalization and technological development, it is becoming increasingly important to ensure that digital products and services are accessible and usable by everyone. Despite significant progress in research, standards, and legislation related to digital accessibility, challenges regarding practical implementation still remain, while websites continue to exhibit extensive accessibility issues. The emergence of artificial intelligence (AI) has therefore attracted interest as a potential solution to streamline and automate accessibility work by identifying accessibility issues and suggesting improvements. The integration of these AI tools may impact User Experience Design (UXD), as it constitutes a central area in the development of accessible digital products and services. At the same time, AI’s focus on efficiency and automation can contrast with UXD’s focus on user-centeredness. Previous research exists on AI and digital accessibility as well as AI and UXD, while the combination of these fields remains underexplored. Therefore, this study investigates how UX practitioners perceive the use of AI-based tools in accessibility work within the user-centered UXD process. To address the study’s purpose and problem definition, a literature review was conducted, followed by a survey and interview study to capture UX practitioners’ perspectives, perceptions, and experiences. The data were subsequently analyzed using thematic analysis and descriptive statistics. The results show that UX practitioners have predominantly positive attitudes toward AI tools for accessibility work, as the technology is viewed as a complementary support that can contribute to efficiency by identifying accessibility issues and suggesting improvements. At the same time, significant risks related to AI’s reliability emerged, such as hallucinations, bias, misinterpretations of accessibility guidelines, and a lack of understanding of context and human needs. Therefore, a clear sense of caution regarding AI use was also expressed, with UX practitioners emphasizing that AI cannot replace human expertise, user-centered methods or the professional responsibility required to critically review and ensure accessible and user-centered solutions. Furthermore, the study provides three recommendations for the future development and implementation of AI tools to support accessibility efforts within the UXD process.

Digital accessibility has become increasingly important as more essential societal functions have shifted from the physical to the digital world. Ensuring equal access to digital content and services is essential for creating an inclusive society, particularly for individuals with cognitive disabilities whose needs are often insufficiently addressed. To regulate digital accessibility, Swedish laws such as the Act on Accessibility to Digital Public Services (2018:1937) and the Accessibility Requirements for Certain Products and Services Act (2023:254) refer to compliance with the Web Content Accessibility Guidelines (WCAG). Compliance is commonly evaluated using automated accessibility testing tools. Although these tools are convenient and widely accessible for both organizations and individuals, their accuracy is limited and many accessibility issues remain undetected.

This study aims to address the limitations of automated accessibility tools by exploring the use of heuristics as a manual complement to improve the accuracy of accessibility evaluations, with a particular focus on cognitive accessibility. The study was conducted through a literature review, evaluations of which accessibility issues automated tools could and could not identify, and the development of a complementary heuristic set validated by experts in UX design and digital accessibility.

The study addresses the research question: “How can automated accessibility evaluation tools and heuristic methods be combined to ensure digital accessibility and compliance with WCAG on websites?” with the sub-questions: “Which cognitive accessibility problems are overlooked by automated accessibility evaluation tools when evaluating website accessibility?” and “How can heuristics be designed to identify the cognitive accessibility problems that automated tools fail to detect?”

The evaluations showed that automated tools frequently failed to detect accessibility issues related to text alternatives for video content, navigation, and deficiencies in error handling and form instructions. The results also demonstrated that a substantial number of WCAG success criteria were either not evaluated at all or incorrectly assessed by the tools. The outcome of the study is a set of heuristics intended to complement automated tools when evaluating websites’ compliance with WCAG success criteria. The set of heuristics was validated by experts in UX design and digital accessibility, who described it as relevant and useful while also contributing suggestions for further development. The experts particularly emphasized that several aspects of cognitive accessibility require human interpretation and are therefore difficult to identify through automated testing alone.

In recent years, virtual reality (VR) has increasingly expanded, as it enables a safe learning environment without the risks that may occur in a real context. Research within user experience (UX) in a VR context is based on creating interfaces that continuously guide the user. The CAMIL framework also highlights cognitive and affective factors that together form a foundation for motivation and engagement, two important cornerstones of learning. At the same time, research within cognitive psychology and the testing effect shows that memory retrieval is crucial in learning of information, and that designers should be cautious when implementing unlimited support.

This study investigates the balance between digital support and guidance in order to optimize learning. The purpose is to understand how the degree of guidance affects participants' learning performance, as well as how it affects their user experience and self-efficacy. The method consists of an experiment in which two test groups perform a motor task under either alternating guidance and test sessions (retrieval group) or continuous guidance (guidance group). The participants performance is measured through completion time, while user experience and self-efficacy are examined through qualitative interviews and observations. The analysis of quantitative data was conducted through descriptive statistics and statistical comparisons. This was combined with a thematic analysis of the interviews as well as identification of patterns and themes in the observations.

The analysis of the quantitative data showed that the guidance group demonstrated more stable performances during the sessions, while the retrieval group performed slightly better on the follow-up test. However, no statistically significant difference was found between the test groups. Both groups generally had a positive user experience, felt present, and perceived the design of the guidance as effective. The same applied to the participants self-efficacy, which increased throughout the sessions. All participants experienced the task as relatively simple and requiring low cognitive load, which may have been reflected in the results. Further research would be needed to establish direct causal relationships without the influence of limitations such as sample size.

En avgörande faktor för användarupplevelsen i e-handel är laddningstid. Då bilder ofta står för en stor del av sidans data, undersöker studien hur ett serverbaserat middleware med bildoptimering påverkar laddningstiden hos en e-handelsplattform för brädspel. En webbplats för brädspel utvecklades i PHP och JavaScript, där en AI-genererad bild laddades upp för varje spel. Först laddades spelen med bilderna upp i PNG-format utan filkonvertering för att sedan laddas upp, och via ett middleware, automatiskt konverteras till WebP-format. Ett automatiserat mätscript utvecklades för att fånga upp laddningstider över upp till 1000 iterationer för varje mätserie. Resultatet av mätningarna visar att middleware innehållande bildoptimering minskar delvis medelladdningstid, men även variation i laddningstiden. Studien visar att effekten av bildoptimering blir högre när bildfilerna är större, och att middleware är en relativt enkel lösning för att förbättra webbprestanda. Förslag på framtida arbete är att undersöka samarbete med andra optimeringslösningar, och analysera regelbundna spikar i mätningarna.

Frontier Large Language Models (LLMs) are rapidly integrating into software development lifecycles (SDLCs); however, this integration comes an inherent risk when deploying the LLM generated code into production environments. Because LLMs stochastically generate code without any inherent security reasoning, the introduction of critically vulnerable code becomes a statistical likelihood. This thesis serves to fill an academic gap by providing quantified, raw empirical data regarding standardized measurements for the vulnerability rate of LLM generated code and their baseline of security reasoning. This thesis evaluated four frontier contemporary LLMs: Gemini 3.1 Pro, Claude Opus 4.7, Llama 3.3 70B, and Codestral v25.08. This thesis developed and deployed an automatic code audit framework utilizing an orchestration script to standardize prompt dispatch and token ingestion. This pipeline collected, isolated, and split functional Python files originating from the raw LLM-generated code blocks. The subsequent audit utilized the CodeQL Command Line Interface (CLI) for static application security testing (SAST) which generates a Static Analysis Results Interchange Formatted (SARIF) report detailing the vulnerabilities within the Python files. The metrics were parsed into two primary domains of evaluation: a Common Vulnerability Scoring System (CVSS) calculated security severity score and the prevalence rate of Common Weakness Enumerations (CWEs) found in the code, mapped in alignment with their categorical status within the contemporary Open Worldwide Application Security Project (OWASP) 2025 Top 10 framework. Through the final analysis, it was found that a grand mean of 39.67% of the analyzed files possessed at least one security-based defect, correlating with a grand mean CVSS severity score of 7.04. Through inferential statistical tests, it was ultimately determined that there is no statistically significant difference between the targeted models’ security reasoning or vulnerability rate.

This project evaluates and compares the performance of an underactuated anthropomorphicrobotic hand with conventional two-finger grippers in pick-and-place and object discrimination tasks using only proprioceptive sensory data. The project aims to simplify sensory implementations for robotic grasping and object discrimination, since these tasks frequently require external sensors, such as force or visual sensors.The experimental setup consisted of a Universal Robots UR5 on which the anthropomorphic hand, the qb SoftHand Research, and the OnRobot RG6 grippers were mounted. Also, a Universal Robot UR10e with a Robotiq 2F-85 was needed due to hardware failures. The pick-and-place and object discrimination comparisons were done under controlled conditions, using two different sets of objects for each part. The qualitative metrics used for the pick-and-place task were repeatability, success rate and slip ratio. For the object discrimination task, a classification model based only on internal sensory feedback was implemented, using motor current and the actual position. The supervised machine learning algorithms Support Vector Machine (SVM) and RandomForest (RF) were trained but only validated for the Robotiq 2F-85 gripper due to the anthropomorphic hand’s failure during testing.The obtained results and their subsequent analysis show the behavioural differences between the grippers in pick-and-place tasks. Furthermore, they demonstrate that object discrimination using only internal data is feasible, under controlled conditions and with a reduced object set, highlighting limitations related to overfitting or the dataset size.

This study highlights the discrepancy between the expansion of the Swedish Armed Forces and the challenge of retention. Understaffing causes heavy workload for the current officers, who work at the limit of their capacity. Paradoxically the upscaling of the organization results in loss of competency rather than expansion. The study illustrates the relationship between job satisfaction and wages together with the intention of officers to stay in the Swedish Armed Forces.Semistructured interviews revealed experiences regarding structural inconsistencies and a complex wage system where competency and experience plays a minor part. The symbolic significance of organizational actions proved to be central for perceived fairness. The empirical data could be related to current research and the underlying principles of Equity theory. In the end, the study concluded that job satisfaction and wage are complementary factors for retention. Balance must exist in relation to the opportunity cost i.e. the loss of value when one course of action is chosen over another.

Background: The availability of digital technologies such as virtual patient simulation has nurtured innovation in higher healthcare education. However, the quality of validated questionnaires related to virtual patient simulation in higher healthcare education remains limited. In particular, there is a need for validated questionnaires for virtual patient simulations focusing on intimate partner violence. Thus, the aim was to cognitively validate questions in the Questionnaire Virtual Patient (QVP) among undergraduate nursing students.

Method: A think-aloud design was employed. Eleven undergraduate nursing students from a single school of nursing in western Sweden participated in an individual digital cognitive interview. Collected data were analyzed using a qualitative conventional analysis. The QVP consists of eleven background questions and additional eighteen questions, structured into two themes: (1) User friendliness virtual patient and (2) Virtual patient as an integrated part of educational module.

Result: The QVP was perceived as clear, well-constructed, and easy to understand. At the same time, some areas for refinement were emphasized. Thus, original questions and revised versions of the questions in QVP are presented in the alignment with the structure of the QVP.

Conclusion: The cognitive validation of the QVP highlights the need for methodological rigor and thoughtful planning in survey design. Ongoing refinement of questionnaires can contribute to the development of evidencebased learning didactics in an increasingly digital healthcare educational context.

In today’s technological landscape, robots are becoming an increasingly common feature of everyday life, making seamless and engaging human-robot interaction essential. A simple and effective way to test this communication is through straightforward games.This thesis aims to study, using Tic-tac-toe—a game that provides a clear structure of rules and turns—how to develop this interaction in the most efficient way. To this end, a system has been created that combines various necessary elements, ranging from the creation of the physical board to the use of a graphical interface to communicate the status of the game and the robot to the user, via perception systems, robot motion control and AI decision-making. To verify that the system met its objectives, two types of tests were carried out: technical tests to validate the functionalities, and user tests to ensure a smooth and engaging interaction. Finally, the results obtained, the system’s limitations and the sustainabilityimpact are presented.

Background: The ongoing technological development, known as Industry 4.0, involves increased integration of digital technology within the manufacturing industry. While this transformation is replacing certain routine tasks, it is simultaneously creating a need for new, more advanced technical expertise. Previous research has often emphasized technical aspects, while the social dimension and HR’s specific role in managing the interaction between the social and technical systems have received less attention. HR therefore finds itself in a complex position where it needs to balance operational and strategic aspects in a technology-dominated work environment.

Purpose: The purpose of this study is to analyze which HR roles appears in a socio-technical context shaped by the Fourth Industrial Revolution, based on Ulrich’s model (1997), which addresses four HR roles.

Method: The study uses a qualitative methodological approach and is based on a deductive framework in which socio-technical systems theory and Dave Ulrich’s (1997) four HR roles serve as analytical frameworks. Data collection was conducted through semi-structured interviews with respondents in the manufacturing industry. The material was then processed through a thematic analysis to identify patterns in how HR describes and performs its role in relation to technological development.

Conclusion: The study shows that HR roles manifest themselves with varying degrees of prominence and are shaped by the technological context as well as the organization’s specific circumstances. In summary, the study shows that the dominance of technology makes it difficult to achieve full joint optimization, as the social system often has to adapt to the premises of the technological system. The role of HR appears to be clearly shaped by the context.

The relationship between attention and consciousness is a central question in cognitive neuroscience, with growing evidence that these processes can be dissociated. The present study conducted a systematic review and meta-analysis of electroencephalography (EEG) studies on unconscious visual selective attention, defined as attentional modulation occurring without conscious awareness of either the attentional cue or the target. A search across Scopus, Web of Science, and APA PsycInfo identified 25 EEG studies, of which 21 contributed quantifiable effect sizes. A three-level multilevel random-effects model was used to synthesize 55 effect sizes (Hedges’ g) from 20 studies in the main analysis and 1 additional study (Baier & Ansorge, 2020) in the sensitivity analysis only. The overall pooled effect was g = 0.60, 95% CI [0.50, 0.70], indicating a medium-sized neural signature of unconscious attentional selection. Subgroup analyses revealed broadly consistent effects across endogenous (g = 0.67) and exogenous (g = 0.55) attention, as well as across unconscious cue (g = 0.60) and unconscious target (g = 0.62) paradigms, with no significant moderation by attention type or unconscious element. Sensitivity analyses assigning g = 0 to non-significant results with unreported statistics confirmed these findings (g = 0.51). These results provide converging neural evidence that both top-down and bottom-up attentional mechanisms can operate without conscious awareness, supporting theoretical accounts that propose a partial dissociation between attention and consciousness, particularly the Recurrent Processing Theory. A funnel plot and formal tests suggested some potential publication bias, though sensitivity analyses supported the robustness of the pooled estimate.

The EU Directive NIS2 (Network and Information Security Directive 2), transposed into Swedish law through the Cybersecurity Act, imposes strengthened requirements on supply chain security for organisations in critical sectors. For energy companies dependent on external IT suppliers, this means that security efforts cannot be limited to the organisation itself but must also encompass systems and services provided by external actors. 

This study examines how a Swedish energy company manages security in its IT supply chain in relation to the requirements of the NIS2 Directive. The study is guided by two sub-questions: how the organisation operationalises requirements for risk management, incident handling, incident reporting and continuity, and how supplier dependencies affect the ability to comply with the continuity requirement. The study was conducted as a qualitative case study with triangulation between fourteen internal governance documents and semi-structured interviews with five respondents. The analysis is based on pattern matching against a theoretical framework comprising eleven propositions grounded in the NIS2 Directive and prior research on vendor lock-in, exit-readiness and digital sovereignty. 

The results show that the organisation is in a transitional phase where new structures are actively being built but do not reach existing supplier relationships. Four patterns emerge: a temporal dimension where one business-critical system is governed by a contract from 2016 lacking security requirements that remains unchanged until 2028, a person-dependency where the most systematic supplier follow-up was tied to a single individual who has since changed roles, a contrast between the two systems where one underwent extensive requirements specification while the other system owner lacks insight into basic security matters, and a contrast between formal implementation and practical implementation. 

The study contributes to the empirical knowledge base on NIS2 implementation by demonstrating that compliance varies within an organisation depending on individual suppliers, and that person-dependency constitutes an underestimated mechanism limiting the sustainability of security efforts. 

This paper conducts a multimodal analysis of the environmental storytelling in Disco Elysium (2019). The analysis is done through the frame of the proposed term playspace which unifies the concepts of gameworld and virtual interface under one singular analytical frame. The analysis finds that Disco Elysium often engages multiple modes when communicating information about its world—primarily aural and textual in dialogue and visual and textual in gamespace. The analysis only finds a few examples of singular mode use—most notably the aural mode for non-diegetic music and as substitution for more graphically demanding sequences. The analysis highlights significant examples where several modes work together, what type of information they communicate and how they fit into the frame of playspace.

Flera organisationer har börjat införa någon typ av AI i deras verksamheter och inom den offentliga sektorn har AI-assistenter blivit allt mer populärt. Denna studie undersöker hur AI-assistenter skapar värde inom offentlig sektor, både genom tidsbesparingar och kvalitét, samt vilka faktorer som påverkar värdet av AI-assistenter inom en organisation, såsom tekniska, kulturella och organisatoriska aspekter.

AI-assistenterna som studien berör är Microsoft Copilot, ChatGPT och SVEA, med störst fokus på hur SVEA kan skapa värde inom offentlig verksamhet. Studien har tillämpat en kvalitativ fallstudie med en induktiv ansats. Den genomfördes i samarbete med Skaraborgs Kommunalförbund och bygger på semi-strukturerade intervjuer för att fånga användarnas upplevelser och känslor kring AI-assistenter. För att analysera det insamlade materialet har en tematisk analys genomförts, i syfte att hitta återkommande mönster och teman.

Crisis preparedness and response require coordination among diverse actors. Traditional crisis management literature emphasizes the ad-hoc and temporary nature of crisis collaborations, but thereby overlooks the extent to which crisis response builds on pre-existing structures and relationships. This paper applies the concept of organizationality—divided into structural and entitative organizationality—to analyze multi-actor crisis management in Sweden. Drawing on three qualitative real event case studies—a ferry grounding, a flooding incident and a forest fire—we demonstrate how organizationality can be used to understand crisis collaboration. The three cases illustrate strong entitative organizationality, enabling coordinated crisis response despite relatively weak but still present structural frameworks rooted in pre-existing institutional arrangements. Organizationality showcases how crisis management relies on latent structural conditions, but also needs trust and actorhood. In contexts like Sweden, where crisis management lacks hierarchical command structures, entitative organizationality plays a critical role in ensuring coordination.

Den pågående digitaliseringen förändrar den verklighet som dagens verksamheter har att förhålla sig till. Ny teknik innebär nya verktyg, men frågan är i vilken utsträckning organisationer faktiskt har förmågan att ta till sig dem. Eftersom det råder vetenskaplig konsensus både inom forskningen relaterat till rekrytering och Business Intelligence att datadrivna beslut överträffar mänsklig intuition, borde rekryteringsfunktioner rimligen vara bland de första att anamma dessa nya verktyg. Trots detta existerar det en tydlig diskrepans mellan HR-funktioners visioner och verkligheten. Syftet med denna studie är att ge insikter som kan underlätta verksamheters omställning till ett mer datadrivet arbetssätt vid rekrytering, och därigenom bidra till att lösa de utmaningar organisationer ställs inför vid implementering av digitala beslutsstöd. Detta uppnås genom att undersöka i vilken utsträckning rekryteringsfunktioner har tillgång till, samt tillämpar beslutsstöd vid rekrytering, och om det existerar motstånd mot användningen av beslutsstöd.

Studien har genomförts med en kvalitativ forskningsstrategi. Datainsamlingen består av fem semistrukturerade intervjuer med HR-specialister och rekryteringsansvariga inom både offentlig och privat sektor. Den insamlade empirin har analyserats genom en induktiv, tematisk analysmetod och strukturerats samt analyserats mot Breaughs m.fl. (2008) modell över rekryteringsprocessen samt Sharda m.fl. (2015) kategorisering av analysnivåer.

Studiens resultat visar på en övergripande nivå att samtliga undersökta verksamheter använder beslutsstöd, men att tillämpningen begränsas till en grundläggande deskriptiv nivå. Studien resulterar mer specifikt i flertalet intressanta fynd. För det första är det inte direkt möjligt att bedöma om beslutsstöd förbättrar rekryteringsprocessen, eftersom majoriteten av verksamheterna saknar en strukturerad utvärdering av sina rekryteringsresultat. För det andra identifieras ett omvänt organisatoriskt motstånd i relation till befintlig forskning (Davenport & Harris, 2017). HR-funktionerna uppvisar en tydlig vilja att arbeta mer datadrivet, motståndet existerar i stället hos linjecheferna. Linjecheferna förlitar sig i stället mer på intuition samt bedriver parallellt manuellt arbete utanför rekryteringssystemen, vilket försämrar datakvaliteten och låser fast organisationerna i en lägre analytisk mognadsgrad. Slutligen identifieras aktiva bortval av beslutsstöd baserat på strategiska val, där en av respondenterna medvetet väljer att prioritera personliga relationer och det mänskliga hantverket framför nya datadrivna verktyg i rekryteringsprocessen.

Despite continuous advancements in technical cybersecurity defence, the human user remains the most frequently exploited vulnerability. The persistence of the knowledge-behavior gap however demonstrates that providing users with technical awareness is rarely enough to ensure secure behavior. The aim of this exploratory study is to investigate the underlying psychological drivers that influence how individuals perceive and respond to digital threats, moving beyond traditional assumption that insecure behavior is simply a result of a lack of knowledge. To explore this, a mixed methods research design was used, combining a quantitative scenario-based survey (N = 40) with qualitative semi-structured interviews (N = 5). The study evaluated participants' responses across eight distinct digital scenarios, running from obvious external threats like phishing to routine habits like password reuse.

The findings indicate that while users are highly capable of identifying and avoiding severe external threats in calm environments, they actively choose to bypass security protocols during routine tasks. This insecure behavior is primarily driven by a desire to avoid digital friction, where effort minimization overrides logical risk analysis. Participants displayed a strong illusion of control in familiar environments, blindly transferring their security responsibilities to large institutions. The study confirms that under stress, emotional reactions consistently override analytical knowledge. These findings suggest that effective cybersecurity cannot rely on technical training alone; it must account for human psychology, respect human limits and stop blaming users for seeking convenience in high-friction digital environments.

Quantum computing has emerged as a potential long-term challenge to the cryptographic algorithms that currently secure digital financial systems. As financial institutions rely extensively on cryptography to protect communications, transactions, authentication processes, and sensitive information, the prospect of quantum computing has raised concerns about the future security of existing cryptographic infrastructures. This has created a need to better understand how organisations perceive and approach the transition toward post-quantum cryptography (PQC). 

This study explores how financial organizations in the United Arab Emirates perceive and approach the transition toward PQC. Guided by the Technology–Organization–Environment (TOE) framework,the study examines how technological, organizational, and environmental factors influence organizational perspectives and approaches toward future cryptographic transition. A qualitative research design was adopted using semi-structured interviews with nine participants in cybersecurity, technology, governance, and compliance roles across banks, financial service providers, payment organizations, and a regulatory authority within the UAE financial sector. 

The findings show that awareness of quantum-related risks is increasing across the sector; however, most organisations remain in the early stages of assessment and monitoring. Organisational approaches were shaped by technological, organisational, and environmental factors, including cryptographic visibility, governance, resource constraints, regulatory expectations, international standards, and ecosystem coordination. 

The study concludes that financial organisations in the UAE recognise PQC as an important future cybersecurity challenge but currently approach the transition through awareness-building, assessment, and monitoring rather than active implementation.Organizational responses were shaped by technological, organisational, and environmental factors, indicating that PQC transition extends beyond a purely technical challenge.