OuluREPO University of Oulu repository
Not a member yet
54977 research outputs found
Sort by
TDoA Extraction Methods for Location Estimation of Wireless Capsule Endoscopy in Gastrointestinal Examination
No full textAbstract
Wireless Capsule Endoscopy location estimation plays a significant role in investigating and diagnosing gastrointestinal diseases, as well as in futuristic applications, such as sample collection for biopsy, location-oriented drug administration, and surgery. Although received signal strength and time-of-arrival are the usual parameters involved in location estimations, this paper deals with the time-difference of arrival (TDoA) parameters. The transmitted signal is an ultra-wideband (UWB) pulse sent by a wireless capsule antenna from 3 different positions in the abdominal cavity. The signal is received by a set of 9 on-body receivers, including a reference receiver. The propagation media is a human voxel model provided by the CST Studio Suite. Four TDoA extracting methods are considered for the received signals at the on-body receivers. The in-voxel TDoA is compared with the actual TDoA, and the histograms are presented for three different positions of the wireless capsule endoscope (WCE). The TDoA error is observed to be lower for the nearby receivers and is found to increase for the faraway receivers. From the statistical mean and standard deviation of the TDoA error, the first cross-correlation peak yields the lowest TDoA error, which is reconfirmed by the mean square of TDoA error in all three positions of the WCE.Abstract
Wireless Capsule Endoscopy location estimation plays a significant role in investigating and diagnosing gastrointestinal diseases, as well as in futuristic applications, such as sample collection for biopsy, location-oriented drug administration, and surgery. Although received signal strength and time-of-arrival are the usual parameters involved in location estimations, this paper deals with the time-difference of arrival (TDoA) parameters. The transmitted signal is an ultra-wideband (UWB) pulse sent by a wireless capsule antenna from 3 different positions in the abdominal cavity. The signal is received by a set of 9 on-body receivers, including a reference receiver. The propagation media is a human voxel model provided by the CST Studio Suite. Four TDoA extracting methods are considered for the received signals at the on-body receivers. The in-voxel TDoA is compared with the actual TDoA, and the histograms are presented for three different positions of the wireless capsule endoscope (WCE). The TDoA error is observed to be lower for the nearby receivers and is found to increase for the faraway receivers. From the statistical mean and standard deviation of the TDoA error, the first cross-correlation peak yields the lowest TDoA error, which is reconfirmed by the mean square of TDoA error in all three positions of the WCE
Negotiating Ethics: How Gen Z Tourists Balance Moral Intentions and Practical Limits?
No full textAbstract
Research on Gen Z often presents them as ethically conscious consumers, yet little is known about how moral intentions translate into practice under real constraints. Existing studies seldom address how economic limits, emotional strain, and social expectations shape the contradictions of ethical and political consumption among this cohort. This study examines these issues through semi-structured interviews with a sample of Gen Z tourists in a developed country context, framed by lifestyle politics and generational cohort theory. Findings show that while participants voiced strong concern for sustainability, human rights, and responsible travel, their actions were often restricted by cost, time, fatigue, and limited access to ethical options. Ethical consumption emerged as negotiated and inconsistent rather than coherent or habitual. By linking individual choices to structural and emotional conditions, the study advances understanding of lifestyle politics and argues that encouraging ethical tourism among youth requires systemic support beyond individual moral commitment.Abstract
Research on Gen Z often presents them as ethically conscious consumers, yet little is known about how moral intentions translate into practice under real constraints. Existing studies seldom address how economic limits, emotional strain, and social expectations shape the contradictions of ethical and political consumption among this cohort. This study examines these issues through semi-structured interviews with a sample of Gen Z tourists in a developed country context, framed by lifestyle politics and generational cohort theory. Findings show that while participants voiced strong concern for sustainability, human rights, and responsible travel, their actions were often restricted by cost, time, fatigue, and limited access to ethical options. Ethical consumption emerged as negotiated and inconsistent rather than coherent or habitual. By linking individual choices to structural and emotional conditions, the study advances understanding of lifestyle politics and argues that encouraging ethical tourism among youth requires systemic support beyond individual moral commitment
Radiolaitteita koskevan direktiivin 3 artiklan 3 kohdan päivitys : vaikutukset ja täytäntöönpano älylaitteiden osalta
No full textArticle 3(3) (d-f) of the Radio Equipment Directive (RED) introduces new requirements for smart devices relating to cybersecurity, immunity protection, and misuse. The directive has already been published, but its implementation and application practices continue to pose challenges for manufacturers. This thesis applied the Design Science Research (DSR) method presented by Peffers et al. (2007) to build a roadmap. The roadmap clearly illustrates all the steps that organizations need to take to implement the requirements of RED 3(3) (d-f) in their smart devices. The aim is to provide clear theoretical guidance to various stakeholders.
The roadmap was evaluated using the Framework for Evaluation in Design Science Research (FEDS) presented by Venable et al. (2016). The evaluation was based entirely on expert feedback. Based on the results, the roadmap was theoretically consistent, useful in practice, and feasible for manufacturers of different sizes, despite that there are many identifiable risks associated with regulatory changes, technological developments, and applicability.
The theoretical contribution of the study demonstrates how the FEDS framework can be applied to the assessment of cybersecurity-related regulatory requirements. The practical contribution shows that the roadmap can provide manufacturers and authorities with a structured path for the implementation of the updated RED requirements
The reliability of artificial intelligence in forecasting investment returns and model risk in time series validation
No full textTämä pro gradu -tutkielma tarkastelee, miten tekoäly- ja erityisesti koneoppimismalleja voidaan arvioida ja hyödyntää luotettavasti epästationaarisessa aikasarjaympäristössä ja millä ehdoilla niiden tuottama tilastollinen etu muuttuu käytännössä merkitykselliseksi lisäarvoksi. Tapausalueena ovat sijoittaminen ja rahoituksen analyysi, joissa tekoälyn käyttö on yleistynyt esimerkiksi signaalinmuodostuksessa, salkunrakennuksessa, riskienhallinnassa ja päätöstuessa. Tutkimuksen tavoitteena on jäsentää teknisiä, taloudellisia ja eettisiä vaikutuksia sekä arvioida vaihtosuhteita ennustetarkkuuden, selitettävyyden, kapasiteetin ja käytännön toteutettavuuden välillä.
Tutkielma perustuu kirjallisuuskatsaukseen ja sitä täydentävään kvantitatiiviseen empiiriseen tutkimukseen. Kirjallisuuskatsaus kokoaa yhteen ajankohtaisia vertaisarvioituja tutkimuksia, joissa tekoälyä sovelletaan signaalinmuodostukseen, salkunrakennukseen, riskimittareiden arviointiin ja roboneuvontaan sekä tarkastelee näiden ratkaisujen luotettavuus-, läpinäkyvyys- ja eettisiä ulottuvuuksia. Empiirisessä osassa kirjallisuudesta johdettu arviointikehikko testataan vertaamalla kahta itse rakennettua ennustemallia, koneoppimismallia ja perinteistä sääntöpohjaista mallia, jotka ennustavat S&P 500-indeksin seuraavan kolmen pörssipäivän avaus-, ylin-, alin- ja päätöskurssin yhden vuoden päivätason historiadatan perusteella.
Kirjallisuuskatsauksen tulokset osoittavat, että tekoäly voi parantaa sijoituspäätösten teknistä laatua erityisesti korkeaulotteisissa ja ei-lineaarisissa ympäristöissä, joissa perinteiset mallit kärsivät spesifikaatiovirheistä. Koneoppimisen ja luonnollisen kielen käsittelyn avulla voidaan tunnistaa markkinasignaaleja, riskejä ja rakenteita, jotka jäisivät perinteisiltä menetelmiltä piiloon, mutta tilastollinen etu realisoituu taloudellisesti ja eettisesti mielekkääksi lisäarvoksi vain, jos arviointi ja validointi toteutetaan kurinalaisesti. Ajallisesti oikea testaus, datavuotojen estäminen, transaktiokustannusten ja kapasiteetin huomioiminen, malliriskin hallinta sekä selitettävyys ja datan alkuperän dokumentointi ovat keskeisiä edellytyksiä. Empiirinen tutkimus tukee tätä kokonaiskuvaa. Koneoppimismalli suoriutui lyhyen aikavälin indeksiennusteissa keskimäärin paremmin kuin perinteinen sääntöpohjainen malli, mutta erot jäivät maltillisiksi, ja myös yksinkertainen, läpinäkyvä sääntömalli ylsi lähelle monimutkaisemman mallin suorituskykyä. Tämän vuoksi tekoälyn lisäarvoa on arvioitava määrällisten tulosten ohella suhteessa selitettävyys- ja operatiivisiin vaatimuksiin. Tutkielma korostaa, että tekoälyn menestyksekäs hyödyntäminen sijoittamisessa edellyttää teknisen ja taloudellisen osaamisen yhdistämistä, datan laadun ja validoinnin systemaattista varmistamista sekä eettisesti kestäviä käytäntöjä, ja että tekoäly on tarkoituksenmukaisinta nähdä päätöksenteon tukena eikä ihmisen harkintakyvyn korvaajana
5G-testiautomaation tehostaminen NOVA-ympäristössä : visuaalinen UE-sekvenssien luontimenetelmä
No full textArtificial intelligence–driven smartphone imaging for early detection of oral cancer: How far have we come? A comprehensive literature review
No full textAbstract
Objective:
This review aimed to evaluate the evidence from previous studies on smartphone-based imaging and artificial intelligence (AI)–assisted mobile health (mHealth) technologies for the early detection of oral potentially malignant disorders (OPMDs) and oral cancer, focusing on their diagnostic performance and feasibility in low-resource settings.
Methods:
A structured literature review was conducted, including 23 peer-reviewed articles published between 2015 and 2024. Eligible studies investigated smartphone-based imaging, tele-dentistry platforms, dual-modality autofluorescence systems, mobile cytology tools, mHealth community programs, and AI-driven models. Data extraction was based on study design, sample size, assessment tool, reference standard, and diagnostic outcomes.
Results:
The 23 studies demonstrated that smartphone-based imaging and mHealth platforms are not only feasible but also effective for OPMDs and oral cancer screening in both community and clinical settings. These platforms reported sensitivity ranged from 70% to 99%, specificity from 64% to 100%, and accuracy from 81% to 97%, depending on the device used, the screening context, and operator expertise. Moreover, AI-driven models such as DenseNet, HRNet, MobileNet, and CNN-based ensembles achieved diagnostic accuracies of 84–95%, in some cases even approaching specialist-level performance. Similarly, dual-modality autofluorescence and white-light imaging systems enhanced classification accuracy, reaching 79–87%. Importantly, community-based mHealth programs involving frontline health workers showed high diagnostic agreement with specialists (κ up to 0.92), thereby enabling large-scale, cost-effective screening. Moreover, MeMoSA® applications demonstrated strong concordance with conventional oral examination (sensitivity 92–94%, specificity up to 95.5%), thus supporting their integration into structured referral pathways.Abstract
Objective:
This review aimed to evaluate the evidence from previous studies on smartphone-based imaging and artificial intelligence (AI)–assisted mobile health (mHealth) technologies for the early detection of oral potentially malignant disorders (OPMDs) and oral cancer, focusing on their diagnostic performance and feasibility in low-resource settings.
Methods:
A structured literature review was conducted, including 23 peer-reviewed articles published between 2015 and 2024. Eligible studies investigated smartphone-based imaging, tele-dentistry platforms, dual-modality autofluorescence systems, mobile cytology tools, mHealth community programs, and AI-driven models. Data extraction was based on study design, sample size, assessment tool, reference standard, and diagnostic outcomes.
Results:
The 23 studies demonstrated that smartphone-based imaging and mHealth platforms are not only feasible but also effective for OPMDs and oral cancer screening in both community and clinical settings. These platforms reported sensitivity ranged from 70% to 99%, specificity from 64% to 100%, and accuracy from 81% to 97%, depending on the device used, the screening context, and operator expertise. Moreover, AI-driven models such as DenseNet, HRNet, MobileNet, and CNN-based ensembles achieved diagnostic accuracies of 84–95%, in some cases even approaching specialist-level performance. Similarly, dual-modality autofluorescence and white-light imaging systems enhanced classification accuracy, reaching 79–87%. Importantly, community-based mHealth programs involving frontline health workers showed high diagnostic agreement with specialists (κ up to 0.92), thereby enabling large-scale, cost-effective screening. Moreover, MeMoSA® applications demonstrated strong concordance with conventional oral examination (sensitivity 92–94%, specificity up to 95.5%), thus supporting their integration into structured referral pathways
Modeling complexity in wildlife populations : a Bayesian hierarchical framework
No full textAbstract
Bayesian statistics provides a flexible framework for understanding the complexity of wildlife populations by modeling ecological processes. In this work, I developed and applied Bayesian hierarchical models to key questions in population dynamics. First, I connected classical open population models with genetic information to model population dynamics close to extinction, aiming to quantify parameters such as population size and survival probabilities. Second, I integrated different data sources to model species distribution across large spatial and temporal scales, combining opportunistic observations with structured surveys to produce spatial distribution maps and abundance estimates. Third, I implemented methodologies to infer habitat preferences and associations, linking environmental variables to demographic rates and the distribution of individuals. By taking advantage of the Bayesian hierarchical structure, it is possible to, firstly, diagnose challenges in the modeling of small populations, including weak identifiability of covariate effects, secondly, provide a scalable approach for estimating abundance and distribution over broad spatio-temporal extents and thirdly, quantify the effect of environmental variables in the distribution of wild individuals. The resulting analyses were translated into conservation actions, aiming to support the decision-making process and improve wildlife management. Original papers Rondon, D., Mäntyniemi, S., Aspi, J., Kvist, L., & Sillanpää, M. J. (2024). A Bayesian multi-state model with data augmentation for estimating population size and effect of inbreeding on survival. Ecological Modelling, 490, 110662. https://doi.org/10.1016/j.ecolmodel.2024.110662 https://doi.org/10.1016/j.ecolmodel.2024.110662 Self-archived version Rondon, D., Ollila, T., Karabanina, E., Aspi, J., Kvist, L., & Sillanpää, M. J. (2025). Multimodal data help in identifying spatio-temporal patterns and habitat associations of Aquila chrysaetos (Golden Eagle) in Finland. Ornithological Applications, duaf047. https://doi.org/10.1093/ornithapp/duaf047 https://doi.org/10.1093/ornithapp/duaf047 Self-archived version Rondon, D., Ollila, T., Aspi, J., Kvist, L., & Sillanpää, M. J. (2025). Seasonal space use and habitat associations of Golden Eagles in Finland: Insights from a point process modelling framework. Manuscript in preparation. Tiivistelmä
Bayesilainen tilastotiede tarjoaa joustavan viitekehyksen luonnonvaraisten eläinpopulaatioiden monimutkaisten ekologisten prosessien ymmärtämiseen mallinnuksien avulla. Tässä työssä kehitin ja sovelsin bayesilaisia hierarkkisia malleja populaatiodynamiikan keskeisiin kysymyksiin. Ensiksi yhdistin klassisen avoimen populaation mallin geneettiseen informaatioon mallintaakseni sukupuuton partaalla olevan populaation dynamiikkaa, tavoitteena kvantifioida ekologisian parametreja kuten populaatiokokoa ja eloonjäämistodennäköisyyksiä. Toiseksi integroin erilaisia tietolähteitä mallintaakseni lajien levinneisyyttä laajoilla ajallis-spatiaalisilla asteikoilla, yhdistäen opportunistiset havainnot strukturoiduista kartoituksista saatuihin aineistoihin tuottaakseni spatiaalisen levinneisyyden karttoja ja runsausarvioita. Kolmanneksi hyödynsin menetelmiä selvittääkseni elinympäristöpreferenssejä ja -assosiaatioita, yhdistäen ympäristömuuttujat populaation koon muutoksiin ja yksilöiden sijaintiin. Bayesilaisen hierarkkisen rakenteen avulla on mahdollista 1) diagnosoida pienten populaatioiden mallintamisen haasteita, mukaan lukien kovariaattivaikutusten heikko identifioitavuus, 2) tarjota skaalautuva lähestymistapa runsauden ja levinneisyyden estimointiin laajoilla spatio-temporaalisilla alueilla ja 3) kvantifioida ympäristömuuttujien vaikutusta luonnonvaraisten yksilöiden levinneisyyteen. Tuloksia voidaan hyödyntää luonnonpopulaatioiden tilaa tukevissa suojelutoimissa ja luonnonsuojeluun liittyvissä päätöksentekoprosesseissa. Osajulkaisut Rondon, D., Mäntyniemi, S., Aspi, J., Kvist, L., & Sillanpää, M. J. (2024). A Bayesian multi-state model with data augmentation for estimating population size and effect of inbreeding on survival. Ecological Modelling, 490, 110662. https://doi.org/10.1016/j.ecolmodel.2024.110662 https://doi.org/10.1016/j.ecolmodel.2024.110662 Rinnakkaistallennettu versio Rondon, D., Ollila, T., Karabanina, E., Aspi, J., Kvist, L., & Sillanpää, M. J. (2025). Multimodal data help in identifying spatio-temporal patterns and habitat associations of Aquila chrysaetos (Golden Eagle) in Finland. Ornithological Applications, duaf047. https://doi.org/10.1093/ornithapp/duaf047 https://doi.org/10.1093/ornithapp/duaf047 Rinnakkaistallennettu versio Rondon, D., Ollila, T., Aspi, J., Kvist, L., & Sillanpää, M. J. (2025). Seasonal space use and habitat associations of Golden Eagles in Finland: Insights from a point process modelling framework. Manuscript in preparation. Academic dissertation to be presented with the assent of the Doctoral Programme Committee of Technology and Natural Sciences of the University of Oulu for public defence in the OP-Pohjola auditorium (L6), Linnanmaa, on 6 March 2026, at 12 noonAbstract
Bayesian statistics provides a flexible framework for understanding the complexity of wildlife populations by modeling ecological processes. In this work, I developed and applied Bayesian hierarchical models to key questions in population dynamics. First, I connected classical open population models with genetic information to model population dynamics close to extinction, aiming to quantify parameters such as population size and survival probabilities. Second, I integrated different data sources to model species distribution across large spatial and temporal scales, combining opportunistic observations with structured surveys to produce spatial distribution maps and abundance estimates. Third, I implemented methodologies to infer habitat preferences and associations, linking environmental variables to demographic rates and the distribution of individuals. By taking advantage of the Bayesian hierarchical structure, it is possible to, firstly, diagnose challenges in the modeling of small populations, including weak identifiability of covariate effects, secondly, provide a scalable approach for estimating abundance and distribution over broad spatio-temporal extents and thirdly, quantify the effect of environmental variables in the distribution of wild individuals. The resulting analyses were translated into conservation actions, aiming to support the decision-making process and improve wildlife management.Tiivistelmä
Bayesilainen tilastotiede tarjoaa joustavan viitekehyksen luonnonvaraisten eläinpopulaatioiden monimutkaisten ekologisten prosessien ymmärtämiseen mallinnuksien avulla. Tässä työssä kehitin ja sovelsin bayesilaisia hierarkkisia malleja populaatiodynamiikan keskeisiin kysymyksiin. Ensiksi yhdistin klassisen avoimen populaation mallin geneettiseen informaatioon mallintaakseni sukupuuton partaalla olevan populaation dynamiikkaa, tavoitteena kvantifioida ekologisian parametreja kuten populaatiokokoa ja eloonjäämistodennäköisyyksiä. Toiseksi integroin erilaisia tietolähteitä mallintaakseni lajien levinneisyyttä laajoilla ajallis-spatiaalisilla asteikoilla, yhdistäen opportunistiset havainnot strukturoiduista kartoituksista saatuihin aineistoihin tuottaakseni spatiaalisen levinneisyyden karttoja ja runsausarvioita. Kolmanneksi hyödynsin menetelmiä selvittääkseni elinympäristöpreferenssejä ja -assosiaatioita, yhdistäen ympäristömuuttujat populaation koon muutoksiin ja yksilöiden sijaintiin. Bayesilaisen hierarkkisen rakenteen avulla on mahdollista 1) diagnosoida pienten populaatioiden mallintamisen haasteita, mukaan lukien kovariaattivaikutusten heikko identifioitavuus, 2) tarjota skaalautuva lähestymistapa runsauden ja levinneisyyden estimointiin laajoilla spatio-temporaalisilla alueilla ja 3) kvantifioida ympäristömuuttujien vaikutusta luonnonvaraisten yksilöiden levinneisyyteen. Tuloksia voidaan hyödyntää luonnonpopulaatioiden tilaa tukevissa suojelutoimissa ja luonnonsuojeluun liittyvissä päätöksentekoprosesseissa
Barrier Self-Efficacy Moderates the Relationship Between Self-Compassion and Self-Reported Leisure-Time Physical Activity
No full textAbstract
Self-compassion, a compassionate attitude toward oneself, has been associated with positive physical activity (PA) outcomes. This cross-sectional study on the 33- to 35-year follow-up data of the Northern Finland Birth Cohort 1986 examined the associations between self-compassion and self-reported leisure-time PA (LTPA), accelerometer-measured PA, and self-reported sitting time and assessed the moderating effect of PA barrier self-efficacy on these associations. Self-compassion was associated with higher LTPA (N = 2,080), also after adjusting for key PA-related covariates, such as gender, perceived health, and level of education but not after adjusting for barrier self-efficacy. Self-compassion was also associated with higher accelerometer-measured PA (N = 811), and lower sitting time (N = 1,072), but not after adjusting for covariates. Barrier self-efficacy moderated the associations between self-compassion and self-reported activity behaviors, but not accelerometer-measured PA. Higher self-compassion was associated with higher LTPA and lower sitting time in individuals with high barrier self-efficacy. Conversely, higher self-compassion signified lower LTPA in individuals with low barrier self-efficacy. Accounting for barrier self-efficacy is thus necessary for the beneficial role of self-compassion in PA to be realized.Abstract
Self-compassion, a compassionate attitude toward oneself, has been associated with positive physical activity (PA) outcomes. This cross-sectional study on the 33- to 35-year follow-up data of the Northern Finland Birth Cohort 1986 examined the associations between self-compassion and self-reported leisure-time PA (LTPA), accelerometer-measured PA, and self-reported sitting time and assessed the moderating effect of PA barrier self-efficacy on these associations. Self-compassion was associated with higher LTPA (N = 2,080), also after adjusting for key PA-related covariates, such as gender, perceived health, and level of education but not after adjusting for barrier self-efficacy. Self-compassion was also associated with higher accelerometer-measured PA (N = 811), and lower sitting time (N = 1,072), but not after adjusting for covariates. Barrier self-efficacy moderated the associations between self-compassion and self-reported activity behaviors, but not accelerometer-measured PA. Higher self-compassion was associated with higher LTPA and lower sitting time in individuals with high barrier self-efficacy. Conversely, higher self-compassion signified lower LTPA in individuals with low barrier self-efficacy. Accounting for barrier self-efficacy is thus necessary for the beneficial role of self-compassion in PA to be realized
Parental Ability to Identify Severe Illnesses in Their Children
No full textAbstract
Importance:
Early parental recognition of severe illness in children and adolescents is crucial for timely management and improved outcomes in pediatric emergency care.
Objective:
To assess how accurately parents can identify severe illness in their children using a questionnaire completed shortly after arrival at the emergency department (ED).
Design, Setting, and Participants:
This diagnostic study was conducted in a tertiary pediatric ED in northern Finland. Data were collected in 2019 to 2021, and this analysis was conducted in May 2024 to May 2025. Children and adolescents whose parents completed the questionnaire before physician assessment were included.
Exposures:
A structured, 36-item parental questionnaire assessing symptoms and the child or adolescent’s overall condition.
Main Outcomes and Measures:
Severe illness was defined as 1 or more of the following: admission to the pediatric intensive care unit, hospital treatment of more than 24 hours, need for intravenous or nasogastric fluids, need for intravenous antibiotics for more than 24 hours, oxygen saturation less than 93% or the need for inhaled medications, anaphylactic shock, intoxication requiring hospital admission, or surgical intervention. Sensitivity and specificity were calculated for each question. To identify parental triage questions with the strongest diagnostic value, a machine learning analysis was conducted.
Results:
Among 2375 included children and adolescents (mean [SD] age, 5.4 [4.6] years; 1140 female [48.0%]), 567 individuals (23.9%) met criteria for severe illness. Moderate to high parental worry showed the highest sensitivity (91.0% [95% CI, 88.3%-93.2%]) but the lowest specificity (17.5% [95% CI, 15.8%-19.4%]). Other specific pediatric questions demonstrated modest diagnostic accuracy with limited additional value. The machine learning model (area under the receiver operating characteristic curve, 0.71; 95% CI, 0.65-0.77) identified parental worry (feature importance score, 0.047), parent assessments of child or adolescent’s general condition (feature importance score, 0.046), and need for treatment (feature importance score, 0.141) as the strongest predictors of hospital admission.
Conclusions and Relevance:
In this study, parental worry identified most cases of severe illness but had low specificity. These findings suggest that while parental concern may serve as an initial screening indicator, it should be complemented by clinical evaluation and objective measures to avoid unnecessary escalation of care.Abstract
Importance:
Early parental recognition of severe illness in children and adolescents is crucial for timely management and improved outcomes in pediatric emergency care.
Objective:
To assess how accurately parents can identify severe illness in their children using a questionnaire completed shortly after arrival at the emergency department (ED).
Design, Setting, and Participants:
This diagnostic study was conducted in a tertiary pediatric ED in northern Finland. Data were collected in 2019 to 2021, and this analysis was conducted in May 2024 to May 2025. Children and adolescents whose parents completed the questionnaire before physician assessment were included.
Exposures:
A structured, 36-item parental questionnaire assessing symptoms and the child or adolescent’s overall condition.
Main Outcomes and Measures:
Severe illness was defined as 1 or more of the following: admission to the pediatric intensive care unit, hospital treatment of more than 24 hours, need for intravenous or nasogastric fluids, need for intravenous antibiotics for more than 24 hours, oxygen saturation less than 93% or the need for inhaled medications, anaphylactic shock, intoxication requiring hospital admission, or surgical intervention. Sensitivity and specificity were calculated for each question. To identify parental triage questions with the strongest diagnostic value, a machine learning analysis was conducted.
Results:
Among 2375 included children and adolescents (mean [SD] age, 5.4 [4.6] years; 1140 female [48.0%]), 567 individuals (23.9%) met criteria for severe illness. Moderate to high parental worry showed the highest sensitivity (91.0% [95% CI, 88.3%-93.2%]) but the lowest specificity (17.5% [95% CI, 15.8%-19.4%]). Other specific pediatric questions demonstrated modest diagnostic accuracy with limited additional value. The machine learning model (area under the receiver operating characteristic curve, 0.71; 95% CI, 0.65-0.77) identified parental worry (feature importance score, 0.047), parent assessments of child or adolescent’s general condition (feature importance score, 0.046), and need for treatment (feature importance score, 0.141) as the strongest predictors of hospital admission.
Conclusions and Relevance:
In this study, parental worry identified most cases of severe illness but had low specificity. These findings suggest that while parental concern may serve as an initial screening indicator, it should be complemented by clinical evaluation and objective measures to avoid unnecessary escalation of care
Open-source, low-cost 3D-printable testbed for in-body optical wireless communications research
No full textAbstract
This hardware paper introduces an experimental testbed for in-body optical wireless communication (OWC) studies. The conventional version often relies on bulky optical benches and costly supporting equipment, which are often cost-prohibitive for many research institutions. The proposed testbed featured a small footprint, lightweight, a vertically aligned optical path (with fixed optical component placement), and ambient light shielding. It can be printed using commercial 3D printing, reducing costs compared to conventional optical benches. The 3D-printable testbed consists of a box-like chassis that securely positions a near-infrared (NIR) LED TX at the top and a photodetector RX at the bottom, with a tissue sample (e.g., ex-vivo porcine tissue or a tissue-mimicking phantom) held firmly in between. All design files, including CAD and STL formats, along with detailed assembly instructions, are made openly available. The inherent design structure enables faster alignment, and the shields can effectively protect against exposure to indoor ambient light (e.g., typical laboratory lighting), thereby improving experimental reliability. The modular nature of the testbed allows for easy customization to accommodate sensors of different wavelengths and different tissue models. The proposed testbed offers practical benefits and an accessible solution for researchers conducting in-body OWC studies, especially when access to high-end optical equipment is limited.Abstract
This hardware paper introduces an experimental testbed for in-body optical wireless communication (OWC) studies. The conventional version often relies on bulky optical benches and costly supporting equipment, which are often cost-prohibitive for many research institutions. The proposed testbed featured a small footprint, lightweight, a vertically aligned optical path (with fixed optical component placement), and ambient light shielding. It can be printed using commercial 3D printing, reducing costs compared to conventional optical benches. The 3D-printable testbed consists of a box-like chassis that securely positions a near-infrared (NIR) LED TX at the top and a photodetector RX at the bottom, with a tissue sample (e.g., ex-vivo porcine tissue or a tissue-mimicking phantom) held firmly in between. All design files, including CAD and STL formats, along with detailed assembly instructions, are made openly available. The inherent design structure enables faster alignment, and the shields can effectively protect against exposure to indoor ambient light (e.g., typical laboratory lighting), thereby improving experimental reliability. The modular nature of the testbed allows for easy customization to accommodate sensors of different wavelengths and different tissue models. The proposed testbed offers practical benefits and an accessible solution for researchers conducting in-body OWC studies, especially when access to high-end optical equipment is limited