Umeå University's logo

Circulatory dysfunction and shock in critical illness has a high mortality. Hypovolemia can contribute to inadequate cardiac output in critical illness. One clinical challenge is that hypovolemia can be difficult to identify. Many hypotensive critically ill patients are hypovolemic, thus need fluid resuscitation, though not all hypotensive patients are hypovolemic. Fluid management is thought to influence patient outcome in critical illness. There is a very common clinical diagnostic dilemma: does the critically ill patient in question need additional fluid administration or have enough or too much already been given. Cardiac 2-dimensional strain offers new parameters for assessment of chamber mechanical function and possibly of left ventricular filling. These novel measurements could be used in the assessment of hypovolemia and cardiac function.

Objectives

The aim of this thesis is assessment of load dependence of left heart chamber wall strain and the interrelation between left atrial and left ventricular strain, with a particular focus on left atrial contraction strain, in healthy individuals and in patients with cardiac illness (cardiac amyloidosis).

Methods

Papers 1 and 2 were conducted in a cohort of healthy volunteers to evaluate the effects of varying loading conditions and sympathetic activation on left atrial contraction strain. Changes in preload were induced through controlled alterations of airway pressure using a CPAP manoeuvre to decrease preload, a passive leg raise to increase preload, and by a Valsalva manoeuvre to decrease preload with a simultaneous increase in sympathetic tone. Paper 3 comprised a retrospective analysis of a study-cohort with cardiac amyloidosis, in which a passive leg raise was used to assess load-dependent changes with a particular focus on left atrial contraction strain. Paper 4 examined the relationship between left atrial and ventricular strain in relation to preload. Atrial and ventricular strain were measured within the same cardiac cycle, and linear regression analysis was used to assess a left atrial/left ventricular strain curve, describing their interdependence under varying loading conditions.

Results

Across the four papers included in this thesis, acute alterations in preload did not result in measurable changes in left atrial contraction strain. This finding was consistent across different preload-modifying interventions, including passive leg raise, CPAP, the Valsalva manoeuvre; and this was observed in both healthy ivindividuals and patients with cardiac amyloidosis. Furthermore, the relationship between left atrial and left ventricular strain during the atrial contraction phase was not affected by changes in preload.

Conclusion

In this thesis, left atrial contraction strain demonstrated stability in response to acute preload alterations in both healthy individuals of different ages and patients with cardiac amyloidosis. These findings support the concept that left atrial contraction strain is largely preload independent within clinically relevant loading ranges and therefore may serve as a robust marker of intrinsic left atrial contractile function.

Bottom-up building stock modelling (BBSM) approach is widely used to assess the energy performance of urban building stocks by modelling individual buildings in detail and aggregating them to larger spatial scales. It plays an important role in supporting urban energy transition planning and policymaking. However, existing BBSM studies are constrained by several limitations, including incomplete building performance datasets, reliance on archetype-based averages in retrofit prioritization, simplified representations of occupant behaviour and building properties, and limited integration of modelling into public engagement tools. These limitations obscure inherent heterogeneity that determines which buildings benefit most from specific measures, leading to one-size-fits-all retrofit strategies and biased estimation of energy-saving potentials and policy effectiveness of energy transition initiatives.

This thesis advances heterogeneity-aware BBSM through an integrated and cumulative methodological pipeline that fuses incomplete building-performance datasets, enables localized retrofit prioritisation, and supports evaluation and communication of demand-side behavioural impacts. First, a data-fusion framework combines multiple incomplete datasets using probabilistic record linkage and inverse modelling, filling data gaps by transferring information across sources rather than relying on archetype-level averages. This improves stock representation by capturing building-to-building variation within the same urban context. Second, the thesis integrates data fusion with ensemble machine learning and explainable AI (SHAP) to identify impactful envelope retrofit measures in a local, data-driven manner. Across 81 building-stock clusters in three Swedish municipalities, the results demonstrate substantial variation in the most influential thermal components across municipalities and climate zones, underscoring the need for local-specific retrofit prioritisation. Third, the thesis incorporates occupant-behaviour diversity alongside building heterogeneity via an enhanced DOB-HUBS framework based on representative clustering, automated physics-based simulations, and surrogate machine learning. Empirical results of the Umeå building stock demonstrate that oversimplified behavioural and homogeneous assumptions can bias energy outcomes by up to 15% (standard deviation 3.06%). The framework is further used to assess Sweden’s forthcoming 2027 capacity-based electricity tariff, indicating that behavioural adaptations could reduce peak electricity demand by 6–17%, with heterogeneous impacts across clusters.

Building on these modelling advances, the thesis extends toward public engagement by developing data-driven benchmarking and interactive visual analytics platform that translate bottom-up modelling outputs into user-facing insights, including peer comparison and ‘what-if’ exploration of retrofit and behavioural scenarios. Collectively, the thesis contributes methods and empirical evidence for more credible, locally tailored, and publicly actionable building-stock analytics. This supports in designing targeted retrofit strategies, effective behavioural measures, and informed public participation in urban energy transition planning.

Vascular surgery may involve manipulation and occlusion of arteries supplying blood to the brain. Obstructions in these feeding arteries can disturb cerebral blood flow, risking hypoperfusion and ischemic brain injury. To maintain balanced cerebral perfusion, the cerebral vasculature contains alternative pathways that can perfuse territories affected by an obstruction, known as the cerebral collateral circulation. However, collateral capacity varies among individuals, meaning that surgical risk is patient-specific. Currently, no reliable method exists for predicting an individual patient’s cerebral collateral capacity during vascular surgery. Computational modelling of hemodynamics offers a powerful tool for assessing patient-specific blood flow and perfusion pressure. This type of modelling provides a predictive dimension to the analysis and could potentially be used to assess intraoperative cerebral collateral circulation to inform preoperative planning and decision-making in vascular surgery.

This work aimed to develop and evaluate a computational method combining computational fluid dynamics (CFD) with segmented arterial trees from computed tomography angiography (CTA) and arterial flow rates from magnetic resonance imaging (MRI), to predict patient-specific cerebral hemodynamics during vascular surgery.

As cerebrovascular resistance is a key parameter in computational predictions of cerebral hemodynamics, we first studied 48 patients with stroke or transient ischemic attack, with and without symptomatic carotid stenosis (mean age 71 years). The aim was to assess patient-specific resistances and evaluate whether they are affected by feeding artery obstruction due to carotid stenosis. Territorial cerebrovascular resistances were estimated based on the simulated cerebral perfusion pressure in the major cerebral arteries and the MRI-measured arterial flow rates. Total cerebrovascular resistance was also compared with that of 56 controls (mean age 74 years). No differences were observed between patients and controls, nor between hemispheres in patients with symptomatic carotid stenosis (Paper I).

To test the predictive capabilities of our model, we next studied a prospective cohort of 20 patients (mean age 67 years) scheduled for elective aortic arch surgery with selective antegrade cerebral perfusion (SACP). Surgery with SACP presents a highly relevant scenario in which the cerebral collateral capacity is crucial for maintaining brain perfusion. SACP can be applied unilaterally, relying on the collateral circulation, or bilaterally, perfusing both hemispheres directly but requiring additional arterial manipulation. There is no consensus on a preferred technique, although neurological injury might be minimized if SACP use was tailored to the patient’s collateral capacity. Our method simulated perfusion pressure in the arterial tree using measured surgical parameters, for each patient obtained during both techniques, as well as downstream vascular resistances. First, a feasibility study (N=5) showed that predictions were possible but were sensitive to arterial size (Paper II). The full prediction analysis (N=19) showed no difference between measured and predicted left-side perfusion pressure for bilateral SACP, while it was slightly underestimated for unilateral SACP. Measured and predicted pressures were highly correlated for both techniques (Paper III). To make the modelling workflow more clinically feasible, these same patients also underwent time-of-flight magnetic resonance angiography (TOF-MRA) (Paper IV). Using the same segmentation method as for CTA, a signal intensity threshold was identified that produced arterial geometries similar to the CTA-based segmentations (N=19). Subsequent simulations showed no difference between modalities in simulated baseline perfusion pressure or pressure laterality during unilateral brain inflow. However, simulations were for both modalities sensitive to segmentation of the collateral arteries.

In summary, we presented a method to determine cerebrovascular resistances and found that carotid stenosis had minimal effect on global or local cerebrovascular resistance in stroke and TIA patients. Using surgical input data, we then proposed a method for computing cerebral perfusion pressure during unilateral and bilateral SACP in aortic arch surgery. The method showed acceptable accuracy in predicting left-side cerebral perfusion pressure, with predicted values highly correlated with measured values. Finally, TOF-MRA offers a promising alternative to CTA, reducing imaging requirements and improving clinical feasibility. These studies demonstrate the potential of hemodynamic modelling to predict cerebral collateral capacity and support individualized clinical decision-making in vascular surgery. 

Cooperative perception has emerged as a key paradigm for enhancing environmental understanding in multi-agent systems by fusing sensory information from multiple agents to achieve more comprehensive and accurate perception than single-agent approaches.Despite its demonstrated benefits, existing cooperative perception methods face critical limitations in practical deployments, primarily due to model heterogeneity, latency, and limited communication bandwidth.

This Ph.D. thesis addresses the gap between the theoretical promise of cooperative perception and its practical deployment by systematically investigating how to design cooperative perception systems that are robust, efficient, and scalable under realistic constraints. The main objective of this research is to develop unified frameworks that enable effective multi-agent perception.

To this end, the thesis proposes a series of novel methods targeting these challenges.First, as a foundational study, InputMix is proposed to balance the contributions of heterogeneous sensors in joint training scenarios. Second, an intermediate model-agnostic cooperative perception framework is introduced to enable modular training and seamless collaboration among agents with heterogeneous models. Third, the Latency-Robust Cooperative Perception (LRCP) framework is developed to mitigate the adverse effects of temporal misalignment among agents. Fourth, a lightweight, codebook-free feature compression framework is designed to reduce communication overhead while preserving perceptual performance. Finally, these components are integrated into a unified framework.

Extensive experiments on public benchmark datasets demonstrate that the proposed methods achieve perception performance comparable to the ideal scenario under latency constraints, while enabling effective collaboration among heterogeneous agents and substantially reducing communication bandwidth.

The main contributions of this thesis lie in establishing practical cooperative perception frameworks that collectively address multiple fundamental challenges in multi-agent perception. The findings of this research have broader implications for large-scale autonomous systems, including connected autonomous vehicles and distributed robotic platforms, where reliable cooperative perception under communication and system heterogeneity constraints is essential.

This thesis investigates the diagnostic accuracy of non-echo-planar diffusion-weighted magnetic resonance imaging (non-EPI DW MRI) for cholesteatoma detection in routine clinical practice. It also evaluates the audiological and patient-reported outcomes of active transcutaneous bone-conduction devices (active transcutaneous BCDs) in otologic care.The thesis focuses on cholesteatoma diagnostics and hearing rehabilitation in individuals with conductive hearing loss, mixed hearing loss, and single-sided deafness.

The first paper evaluated the diagnostic accuracy of non-EPI DW MRI for the detection of cholesteatoma in routine clinical practice. Overall, diagnostic accuracy was moderate in routine clinical practice but improved when examinations were interpreted by experienced radiologists and when standardised EAONO/JOS diagnostic criteria were applied.

The subsequent papers assessed audiological and patient-reported outcome measures following implantation of active transcutaneous BCDs. The second paper evaluated the short-term outcomes of an active transcutaneous BCD and included an assessment of postoperative complications. The results demonstrated substantial short-term improvements in audiological performance and in quality of life among individuals with conductive and mixed hearing loss, while the complications rate was low. Participants with single-sided deafness generally experienced less benefit from the active transcutaneous BCD. However, the observed outcomes were in line with those reported for alternative rehabilitation options in this population.

The third paper examined the long-term efficacy and safety of active transcutaneous BCDs. Hearing improvements and patient-reported outcomes were maintained over time, with a low incidence of adverseevents.

The fourth paper expanded the evidence base by comparing the audiological performance of two commercially available active transcutaneous BCDs in participants with conductive and mixed hearing loss. The results showed that both systems produced substantial improvements in audiological performance. However, one system achieved significantly lower aided sound-field thresholds at several frequencies, indicating more efficient sound transmission. Both devices led to significant improvements in speech recognition in noise, with comparable gains across groups.

Overall, the findings indicate that non-EPI DW MRI has limitations in routine clinical practice, but diagnostic accuracy improves when examinations are interpreted by experienced radiologists using standardised assessment criteria. Active transcutaneous BCDs represent a safe and effective rehabilitation option for individuals with conductiveand mixed hearing loss and single-sided deafness.

In the Swedish constitutional system, lower-court judges play a central role in the application of law, yet little is known about how they understand this role in relation to the role of the national legislator. This thesis explores how Swedish lower-court judges understand their role in legal application, with particular attention to its relationship to the role of the national legislator in the constitutional system. The study focuses on judges’ own conceptions of what it means to apply the law, rather than on their actual practices or the outcomes of judicial decisions.

The research is based on interviews with currently serving lower-court judges, capturing how they view and reflect on the application of law and use of different legal sources. Using thematic analysis, three ideal-type roles are constructed to illustrate variations in these conceptions: The Civil Servant, The Guardian, and The Creator. The Civil Servant emphasises strict application of legislation and loyalty to the legislator. The Guardian highlights responsibility for maintaining the integrity of the legal system, including adherence to principles and to the hierarchy of norms. The Creator expresses a conception of the judge as also contributing to the development of law through judicial law-making. 

The constructed roles are analysed in relation to institutional and constitutional conditions and prior discussions about the role of the Swedish judge, as well as compared to each other. The analysis shows how the different role conceptions can be understood as expressions of different ways of relating to the institutional and constitutional conditions shaping the Swedish constitutional system. Furthermore, the comparison between the roles demonstrates that the different role conceptions can be seen as positioned within a tension between tradition, change, and the practical realities of everyday judicial work, ranging from a view of the judge as primarily serving the legislator to one of acting as a guardian of the legal system or, to some extent, its co-creator.

Overall, by constructing, analysing, and comparing these role conceptions, the dissertation contributes to a nuanced view of the diversity of the self-understandings of Swedish lower-court judges regarding their role in legal application. 

Background:

Lung function and its trajectory throughout the life-course - starting from lung growth and maturation, followed by the attainment of peak capacity, and subsequent decline - are influenced by different interrelated factors, including smoking, overweight, obesity, traffic exposure in childhood, and asthma, with the latter being one of the most common chronic diseases in the world. Both lung function impairment and asthma carry a significant public health burden due to increased risk of morbidity, mortality, and lower quality of life. Although characterized by chronic airway inflammation and variable airflow limitation, asthma is a heterogeneous disease encompassing diverse clinical presentations and symptoms, including wheezing, difficulty breathing, and coughing. Understanding how these factors impact lung function is critical for identifying high-risk individuals in order to improve asthma outcomes and survival.

Aims:

The overall aim of this thesis is to investigate how specific respiratory symptoms and changes in body mass index (BMI) during adulthood, as well as traffic exposure in childhood, are associated with lung function and asthma. A further aim is to evaluate the association between different spirometry patterns and mortality among adults with asthma.

Methods:

This thesis is based on longitudinal data from the Obstructive Lung Disease in Northern Sweden (OLIN) research program, which used two separate population-based cohorts: the OLIN adult asthma cohort, examined at study entry (1986-2001) and at one follow-up (2012-2014), and the OLIN pediatric cohort II, recruited in 2006 at age 8 and followed up in 2016-2017 at age 19. Four prospective studies were conducted:

1. Respiratory symptoms: An analysis of 977 adults with asthma, who were followed for an average of 18 years to determine the association between the number and type of respiratory symptoms at study entry (recurrent wheeze, dyspnea, longstanding cough and productive cough) and lung function at follow-up in terms of the forced expiratory volume in one second (FEV₁) and also the annual decline in FEV₁, adjusting for other potentially associated factors by linear regression.

2. Body mass index (BMI): A study of 945 adults with asthma that investigated the impact of annual BMI change on the annual decline in FEV1, forced vital capacity (FVC) and FEV₁/FVC separately in those with normal weight (BMI 18.5–24.9) and those with overweight/obesity (BMI ≥25) at study entry. Regression models were used to adjust for sex, age, smoking, inhaled corticosteroids (ICS) use and occupational exposure to gases, dust or fumes.

3. Spirometry patterns: A mortality study of 2028 adults with asthma, where spirometry patterns at study entry were classified as normal, preserved ratio impaired spirometry (PRISm) and obstruction. Mortality data until the 31^st of December 2020 (n=720 deceased) were linked from a national register and grouped by cause: respiratory, cardiovascular, cancer and other. Cox and Fine-Gray regression models were used to estimate hazard ratios (HRs) for spirometry patterns in relation to all-cause and cause-specific mortality, adjusted for age, sex, ICS use, education level, smoking habits and BMI.

4. Traffic exposure: A 10-year follow-up of a pediatric general population cohort (n=1056), that examined the impact of childhood exposure to vehicle traffic, in an area with low traffic-flows, on lung function in young adulthood in terms of FEV₁, FVC and FEV₁/FVC. Different exposure thresholds were defined based on residence proximity (within a 200 m radius of the home address) to a road with the highest daily count of heavy vehicles (≥250 and ≥500) or any type of vehicle (≥4000 and ≥8000). The association between vehicle traffic exposure and lung function was analyzed by linear regression adjusting for potential confounders.

Results:

Respiratory symptoms associated with lung function - specifically, recurrent wheeze and a high number of concurrent symptoms - were strongly associated with lower lung function, particularly in women. Regarding weight, BMI change was significantly associated with lung function decline. This association was more pronounced in those who were already overweight or obese at study entry compared to those with a normal BMI. The mortality study showed that, in addition to smoking and obesity, both PRISm and obstruction patterns increased all-cause and respiratory mortality risks. Environmentally, even in areas with relatively low traffic flow, childhood exposure to ≥250 heavy vehicles/day and ≥8000 vehicles/day were associated with slight but significantly lower lung function in young adulthood, suggesting that there may be no "safe" threshold for traffic exposure in relation to lung function development.

Conclusions:

Early-life traffic exposure in childhood was found to be associated with slightly lower lung function in young adulthood. Among adults with asthma, recurrent wheeze was the respiratory symptom that was most strongly associated with low lung function. BMI increase was also associated with lung function decline, especially among those who were already overweight or obese. Regarding spirometry patterns, both PRISm and obstruction were associated with increased risk of all-cause and respiratory mortality. These findings highlight that symptom control and weight management among adults with asthma, as well as public health policies considering even low-level traffic exposure, are essential for preserving lung function and reducing mortality.

Background: Chronic pain affects multiple aspects of life, including employment, functioning, interpersonal relationships, and overall quality of life. Approximately one-fifth of the European population experiences chronic pain. Yet, research and public policy have devoted limited attention to this condition, despite its substantial societal costs, including reduced productivity and high healthcare utilisation. The Interdisciplinary Pain Rehabilitation Programme (IPRP) is an evidence-based treatment provided in specialist care. However, it remains largely underutilised in primary care, where the majority of chronic pain patients are managed. Implementing the IPRP requires coordinated professional efforts and substantial initial resources, which can hinder its adoption. Current health-economic evaluations are limited, short-term, and inconclusive, casting doubt on the programme’s long-term effectiveness. 

Aims: The overall aim of this thesis was to study the health economic implications of IPRPs in primary care from both a societal and healthcare provider perspective. Study I aimed to evaluate patient-reported outcomes and healthcare utilisation one year before and after a case manager-led IPRP. Study II aimed to analyse the cost-effectiveness of IPRP compared with care as usual. Study III aimed to analyse healthcare utilisation and costs one year before and after IPRP. Study IV aimed to examine whether participating in IPRP in primary or specialist care is associated with background variables, pain characteristics, quality of life, anxiety, and depression.

Methods: Study I compared patient-reported outcomes and healthcare utilisation one year before and after assessment using non-parametric analyses, the Wilcoxon Signed Rank and Mann-Whitney U tests. Study II applied a cost-utility analysis to evaluate the cost-effectiveness of IPRPs compared with usual care in primary care. In Study III, healthcare utilisation and costs during the 1 year before and after IPRP were analysed by linking regional registry data to participants. Paired t-tests were used for comparative parametric analyses (Study III). The distribution of resources was compared one year before and one year after IPRP (Studies I and III). Study IV used logistic regression to identify factors associated with participation in IPRP in primary or specialist care.

Results: In Study I, reduced healthcare utilisation after IPRP was associated with increased activity levels, improved health-related quality of life, and fewer general practitioner visits. Increased healthcare utilisation was associated with higher pain intensity and a lack of psychological support at baseline and greater use of specialist services. The cost-utility analysis carried out in Study II indicated that IPRP in primary care is cost-effective, particularly in the long term. In Study III, healthcare utilisation decreased by 16% and costs by 12% the year after IPRP. This was mainly due to fewer consultations with general practitioners and physiotherapists. Study VI showed that women, individuals with university education, and those with frequent general practitioner visits were more likely to participate in IPRP in specialist care. Persistent pain and multiple pain sites also increased the likelihood of specialist referral. In contrast, obesity, high pain intensity, higher pain catastrophising, and better general health were associated with participation in IPRP in primary care. 

Conclusion: IPRP enhances health-related quality of life and reduces sickness absence to an extent that supports its cost-effectiveness compared with usual care, especially in the long run. Reduced healthcare utilisation, especially visits to general practitioners and physiotherapists, generated cost savings and freed resources in the primary care centre. Early biopsychosocial intervention, including psychological support, may improve well-being and limit unnecessary healthcare use. Reorganising primary care resources could strengthen chronic pain management and support the broader implementation of IPRP. Socioeconomic factors appear to influence referral pathways, resulting in unequal access to healthcare and inefficient use of healthcare resources. Straightforward guidelines are needed to ensure that patients with lower rehabilitation needs receive treatment in primary care, while those with greater needs access specialist rehabilitation.