RESUMEN
BACKGROUND: Electronic health records (EHRs) contain patients' health information over time, including possible early indicators of disease. However, the increasing amount of data hinders clinicians from using them. There is accumulating evidence suggesting that machine learning (ML) and deep learning (DL) can assist clinicians in analyzing these large-scale EHRs, as algorithms thrive on high volumes of data. Although ML has become well developed, studies mainly focus on engineering but lack medical outcomes. OBJECTIVE: This study aims for a scoping review of the evidence on how the use of ML on longitudinal EHRs can support the early detection and prevention of disease. The medical insights and clinical benefits that have been generated were investigated by reviewing applications in a variety of diseases. METHODS: This study was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A literature search was performed in 2022 in collaboration with a medical information specialist in the following databases: PubMed, Embase, Web of Science Core Collection (Clarivate Analytics), and IEEE Xplore Digital Library and computer science bibliography. Studies were eligible when longitudinal EHRs were used that aimed for the early detection of disease via ML in a prevention context. Studies with a technical focus or using imaging or hospital admission data were beyond the scope of this review. Study screening and selection and data extraction were performed independently by 2 researchers. RESULTS: In total, 20 studies were included, mainly published between 2018 and 2022. They showed that a variety of diseases could be detected or predicted, particularly diabetes; kidney diseases; diseases of the circulatory system; and mental, behavioral, and neurodevelopmental disorders. Demographics, symptoms, procedures, laboratory test results, diagnoses, medications, and BMI were frequently used EHR data in basic recurrent neural network or long short-term memory techniques. By developing and comparing ML and DL models, medical insights such as a high diagnostic performance, an earlier detection, the most important predictors, and additional health indicators were obtained. A clinical benefit that has been evaluated positively was preliminary screening. If these models are applied in practice, patients might also benefit from personalized health care and prevention, with practical benefits such as workload reduction and policy insights. CONCLUSIONS: Longitudinal EHRs proved to be helpful for support in health care. Current ML models on EHRs can support the detection of diseases in terms of accuracy and offer preliminary screening benefits. Regarding the prevention of diseases, ML and specifically DL models can accurately predict or detect diseases earlier than current clinical diagnoses. Adding personally responsible factors allows targeted prevention interventions. While ML models based on textual EHRs are still in the developmental stage, they have high potential to support clinicians and the health care system and improve patient outcomes.
Asunto(s)
Aprendizaje Profundo , Diagnóstico Precoz , Registros Electrónicos de Salud , Aprendizaje Automático , Humanos , Estudios LongitudinalesRESUMEN
Teledentistry offers possibilities for improving efficiency and quality of care and supporting cost-effective healthcare systems. This umbrella review aims to synthesize existing systematic reviews on teledentistry and provide a summary of evidence of its clinical- and cost-effectiveness. A comprehensive search strategy involving various teledentistry-related terms, across seven databases, was conducted. Articles published until 24 April 2023 were considered. Two researchers independently reviewed titles, abstracts and full-text articles. The quality of the included reviews was critically appraised with the AMSTAR-2 checklist. Out of 749 studies identified, 10 were included in this umbrella review. Two reviews focusing on oral-health outcomes revealed that, despite positive findings, there is not yet enough evidence for the long-term clinical effectiveness of teledentistry. Ten reviews reported on economic evaluations or costs, indicating that teledentistry is cost-saving. However, these conclusions were based on assumptions due to insufficient evidence on cost-effectiveness. The main limitation of our umbrella review was the critically low quality of the included reviews according to AMSTAR-2 criteria, with many of these reviews basing their conclusions on low-quality studies. This highlights the need for high-quality experimental studies (e.g., RCTs, factorial designs, stepped-wedge designs, SMARTs and MRTs) to assess teledentistry's clinical- and cost-effectiveness.
Asunto(s)
Análisis Costo-Beneficio , Salud Bucal , Telemedicina , Humanos , Telemedicina/economía , Telemedicina/métodos , Salud Bucal/economía , Odontología/métodosRESUMEN
ABSTRACT: Since the first Oxford Survey of Childhood Cancer's results were published, people have become more aware of the risks associated with prenatal exposure from diagnostic x rays. As a result, it has since been the subject of many studies. In this review, the results of recent epidemiological studies are summarized. The current international guidelines for diagnostic x-ray examinations were compared to the review. All epidemiological studies starting from 2007 and all relevant international guidelines were included. Apart from one study that involved rhabdomyosarcoma, no statistically significant associations were found between prenatal exposure to x rays and the development of cancer during 2007-2020. Most of the studies were constrained in their design due to too small a cohort or number of cases, minimal x-ray exposure, and/or data obtained from the exposed mothers instead of medical reports. In one of the studies, computed tomography exposure was also included, and this requires more and longer follow-up in successive studies. Most international guidelines are comparable, provide risk coefficients that are quite conservative, and discourage abdominal examinations of pregnant women.
Asunto(s)
Neoplasias Inducidas por Radiación , Niño , Estudios de Cohortes , Femenino , Humanos , Neoplasias Inducidas por Radiación/epidemiología , Neoplasias Inducidas por Radiación/etiología , Embarazo , Radiografía , Tomografía Computarizada por Rayos X/efectos adversos , Rayos XRESUMEN
The development of eHealth is dramatically changing the way healthcare is provided and organized. eHealth applications are used not only by healthcare professionals but also by patients specifically to self-manage their health condition. The development of eHealth applications requires a new methodological approach, departing from the more conventional methods dedicated to designing health information systems. There is a gap between theories to design persuasive eHealth applications and practices. In the Netherlands, eHealth innovation emerges from three areas. In research, the development of eHealth application often remains in a pilot phase. Healthcare organizations are also keen to innovate but do not always have the know-how. We further witness technology push from business and industry, undermining the co-creation process of the innovation. We consequently advocate an integrated, systematic and practical but scientifically based methodology to design effective persuasive eHealth applications. This approach is being successfully embedded in our educational health informatics program.
Asunto(s)
Informática Médica/educación , Informática Médica/instrumentación , Comunicación Persuasiva , Automanejo/psicología , Humanos , Informática Médica/métodos , Modelos Teóricos , Países Bajos , Automanejo/métodos , Telemedicina/instrumentación , Telemedicina/métodos , Telemedicina/normasRESUMEN
A survey was conducted among 20 Dutch hospitals about radiation protection for interventional fluoroscopy. This was a follow-up of a previous study in 2007 that led to several recommendations for radiation protection for interventional fluoroscopy. The results indicate that most recommendations have been followed. However, radiation-induced complications from interventional procedures are still often not recorded in the appropriate register. Furthermore, even though professionals with appropriate training in radiation protection are usually involved in interventional procedures, this often is not the case when these procedures are carried out outside the radiology department. Although this involvement is not required by Dutch law, it is recommended to have radiation protection professionals present more often at interventional procedures. Further improvements in radiation protection for interventional fluoroscopy may come from a comparison of dose-reducing practices among hospitals, the introduction of diagnostic reference levels for interventional procedures, and a more thorough form of screening and follow-up of patients.
Asunto(s)
Fluoroscopía/efectos adversos , Hospitales/estadística & datos numéricos , Enfermedades Profesionales/prevención & control , Traumatismos por Radiación/prevención & control , Protección Radiológica/métodos , Radiografía Intervencional/efectos adversos , Humanos , Países Bajos/epidemiología , Enfermedades Profesionales/epidemiología , Enfermedades Profesionales/etiología , Dosis de Radiación , Exposición a la Radiación/análisis , Traumatismos por Radiación/epidemiología , Traumatismos por Radiación/etiología , Protección Radiológica/normasRESUMEN
Atherosclerosis is the development of lipid-laden plaques in arteries and is nowadays considered as an inflammatory disease. It has been shown that high doses of ionizing radiation, as used in radiotherapy, can increase the risk of development or progression of atherosclerosis. To elucidate the effects of radiation on atherosclerosis, we propose a mathematical model to describe radiation-promoted plaque development. This model distinguishes itself from other models by combining plaque initiation and plaque growth, and by incorporating information from biological experiments. It is based on two consecutive processes: a probabilistic dose-dependent plaque initiation process, followed by deterministic plaque growth. As a proof of principle, experimental plaque size data from carotid arteries from irradiated ApoE[Formula: see text] mice was used to illustrate how this model can provide insight into the underlying biological processes. This analysis supports the promoting role for radiation in plaque initiation, but the model can easily be extended to include dose-related effects on plaque growth if available experimental data would point in that direction. Moreover, the model could assist in designing future biological experiments on this research topic. Additional biological data such as plaque size data from chronically-irradiated mice or experimental data sets with a larger variety in biological parameters can help to further unravel the influence of radiation on plaque development. To the authors' knowledge, this is the first biophysical model that combines probabilistic and mechanistic modeling which uses experimental data to investigate the influence of radiation on plaque development.
Asunto(s)
Apolipoproteínas E/deficiencia , Fenómenos Biofísicos , Modelos Biológicos , Placa Aterosclerótica/metabolismo , Traumatismos por Radiación/metabolismo , Animales , Progresión de la Enfermedad , Femenino , Ratones , Placa Aterosclerótica/patología , Traumatismos por Radiación/patologíaRESUMEN
A survey about radiation protection in pediatric radiology was conducted among 22 general and seven children's hospitals in the Netherlands. Questions concerned, for example, child protocols used for CT, fluoroscopy and x-ray imaging, number of images and scans made, radiation doses and measures taken to reduce these, special tools used for children, and quality assurance issues. The answers received from 27 hospitals indicate that radiation protection practices differ considerably between general and children's hospitals but also between the respective general and children's hospitals. It is recommended that hospitals consult each other to come up with more uniform best practices. Few hospitals were able to supply doses that can be compared to the national Diagnostic Reference Levels (DRLs). The ones that could be compared exceeded the DRLs in one in five cases, which is more than was expected beforehand.
Asunto(s)
Hospitales/estadística & datos numéricos , Pediatría/estadística & datos numéricos , Pautas de la Práctica en Medicina/estadística & datos numéricos , Exposición a la Radiación/estadística & datos numéricos , Protección Radiológica/estadística & datos numéricos , Radiografía/estadística & datos numéricos , Niño , Preescolar , Femenino , Encuestas de Atención de la Salud , Humanos , Lactante , Recién Nacido , Masculino , Países Bajos/epidemiología , Radiología/estadística & datos numéricosRESUMEN
Epidemiological miner cohort data used to estimate lung cancer risks related to occupational radon exposure often lack cohort-wide information on exposure to tobacco smoke, a potential confounder and important effect modifier. We have developed a method to project data on smoking habits from a case-control study onto an entire cohort by means of a Monte Carlo resampling technique. As a proof of principle, this method is tested on a subcohort of 35,084 former uranium miners employed at the WISMUT company (Germany), with 461 lung cancer deaths in the follow-up period 1955-1998. After applying the proposed imputation technique, a biologically-based carcinogenesis model is employed to analyze the cohort's lung cancer mortality data. A sensitivity analysis based on a set of 200 independent projections with subsequent model analyses yields narrow distributions of the free model parameters, indicating that parameter values are relatively stable and independent of individual projections. This technique thus offers a possibility to account for unknown smoking habits, enabling us to unravel risks related to radon, to smoking, and to the combination of both.
Asunto(s)
Neoplasias Pulmonares/epidemiología , Minería , Neoplasias Inducidas por Radiación/epidemiología , Exposición Profesional/efectos adversos , Radón/efectos adversos , Fumar/efectos adversos , Carcinogénesis , Estudios de Casos y Controles , Estudios de Cohortes , Alemania , Humanos , Método de Montecarlo , Enfermedades Profesionales/epidemiología , Factores de TiempoRESUMEN
BACKGROUND: Physical inactivity is a growing public health concern. Use of mobile applications (apps) may be a powerful tool to encourage physical activity and a healthy lifestyle. For instance, apps may be used in the preparation of a running event. However, there is little evidence for the relationship between app use and change in physical activity and health in recreational runners. The aim of this study was to determine the relationship between the use of apps and changes in physical activity, health and lifestyle behaviour, and self-image of short and long distance runners. METHODS: A cross sectional study was designed. A random selection of 15,000 runners (of 54,000 participants) of a 16 and 6.4 km recreational run (Dam tot Damloop) in the Netherlands was invited to participate in an online survey two days after the run. Anthropometrics, app use, activity level, preparation for running event, running physical activity (RPA), health and lifestyle, and self-image were addressed. A chi-squared test was conducted to analyse differences between app users and non-app users in baseline characteristics as well as in RPA, healthy lifestyle and perceived health. In addition, a multivariate logistic regression analysis was performed to determine if app use could predict RPA, perceived health and lifestyle, and self-image. RESULTS: Of the 15,000 invited runners, 28% responded. For both distances, app use was positively related to RPA and feeling healthier (p < 0.05). Also, app use was positively related to feeling better about themselves, feeling like an athlete, motivating others to participate in running, and losing weight (p < 0.01). Furthermore, for 16 km runners app use was positively related to eating healthier, feeling more energetic and reporting a higher chance to maintain sport behaviour (p < 0.05). CONCLUSIONS: These results suggest that use of mobile apps has a beneficial role in the preparation of a running event, as it promotes health and physical activity. Further research is now needed to determine a causal relationship between app use and physical and health related behaviour.
Asunto(s)
Estilo de Vida , Aplicaciones Móviles/estadística & datos numéricos , Carrera/estadística & datos numéricos , Autoimagen , Adulto , Estudios Transversales , Ejercicio Físico , Femenino , Conductas Relacionadas con la Salud , Humanos , Masculino , Persona de Mediana Edad , Motivación , Países BajosRESUMEN
Diagnostic reference levels (DRLs) for medical x-ray procedures are being implemented currently in the Netherlands. By order of the Dutch Healthcare Inspectorate, a survey has been conducted among 20 Dutch hospitals to investigate the level of implementation of the Dutch DRLs in current radiological practice. It turns out that hospitals are either well underway in implementing the DRLs or have already done so. However, the DRLs have usually not yet been incorporated in the QA system of the department nor in the treatment protocols. It was shown that the amount of radiation used, as far as it was indicated by the hospitals, usually remains below the DRLs. A procedure for comparing dose levels to the DRLs has been prescribed but is not always followed in practice. This is especially difficult in the case of children, as most general hospitals receive few children.
Asunto(s)
Diagnóstico por Imagen/normas , Implementación de Plan de Salud , Hospitales/normas , Protección Radiológica/métodos , Protección Radiológica/normas , Radiometría , Niño , Recolección de Datos , Humanos , Países Bajos , Dosis de Radiación , Estándares de Referencia , Rayos XRESUMEN
Exposures to ionizing radiation from medical examinations are on the rise. An important cause for this has been the advent and ever-increasing use of computed tomography (CT) scans for diagnostic purposes. It is often implied that population aging contributes significantly to this rise. Here, the trends in population statistics are compared to the trend in the number of CT scans in the Netherlands for the period 2002-2010. It is concluded that population growth and population aging cannot explain the observed rise in CT examinations. In fact, these factors contribute only 17% to this rise, indicating that there must be other factors that are far more important.
Asunto(s)
Envejecimiento , Tomografía Computarizada por Rayos X/estadística & datos numéricos , Adolescente , Adulto , Anciano , Anciano de 80 o más Años , Niño , Preescolar , Humanos , Lactante , Recién Nacido , Persona de Mediana Edad , Países BajosRESUMEN
Breast cancer incidence in a tuberculosis fluoroscopy cohort has been modelled with a two-stage carcinogenesis model. The relatively simple model, in which hormonal influences only affect the number of sensitive target cells, fits the data very well. Under the assumption that individual hormonal differences average out, and with a relative biological effectiveness for mammographic X rays of 1, the model yields â¼10 fatal breast cancer cases induced by the entire Dutch screening programme over a period of 25 y. This is much lower than derived from standard ICRP risk estimates and should be compared with the number of lives saved, which is estimated at â¼350 y(-1). As the extent of screening is currently being reconsidered in The Netherlands and elsewhere, this is an important result.
Asunto(s)
Neoplasias de la Mama/epidemiología , Fluoroscopía/estadística & datos numéricos , Neoplasias Inducidas por Radiación/epidemiología , Modelos de Riesgos Proporcionales , Tuberculosis/diagnóstico por imagen , Tuberculosis/epidemiología , Neoplasias de la Mama/diagnóstico por imagen , Estudios de Cohortes , Comorbilidad , Femenino , Humanos , Persona de Mediana Edad , Países Bajos/epidemiología , Prevalencia , Medición de Riesgo , Factores de RiesgoRESUMEN
A mechanistic two-stage carcinogenesis model has been applied to model lung-cancer mortality in the largest uranium-miner cohort available. Models with and without smoking action both fit the data well. As smoking information is largely missing from the cohort data, a method has been devised to project this information from a case-control study onto the cohort. Model calculations using 256 projections show that the method works well. Preliminary results show that if an explicit smoking action is absent in the model, this is compensated by the values of the baseline parameters. This indicates that in earlier studies performed without smoking information, the results obtained for the radiation parameters are still valid. More importantly, the inclusion of smoking-related parameters shows that these mainly influence the later stages of lung-cancer development.
Asunto(s)
Neoplasias Pulmonares/epidemiología , Minería/estadística & datos numéricos , Neoplasias Inducidas por Radiación/epidemiología , Enfermedades Profesionales/epidemiología , Exposición Profesional/estadística & datos numéricos , Modelos de Riesgos Proporcionales , Radón/análisis , Adolescente , Adulto , Anciano , Niño , Alemania Oriental/epidemiología , Humanos , Masculino , Persona de Mediana Edad , Prevalencia , Medición de Riesgo , Factores de Riesgo , Fumar , Adulto JovenRESUMEN
A biologically based two-stage carcinogenesis model is applied to epidemiological data for lung cancer mortality in a large uranium miner cohort of the WISMUT company (Germany). To date, this is the largest uranium miner cohort analyzed by a mechanistic model, comprising 35,084 workers among whom 461 died from lung cancer in the follow-up period 1955-1998. It comprises only workers who were first employed between 1955 and 1989 and contains information on annual exposures to radon progeny. We fitted the model's free parameters, including the average growth time of one malignant cell into a lethal tumor. This lag time has an extraordinary value of 13 to 14 years, larger than that previously used or found in miner studies. Even though cohort-wide information on smoking habits is limited and the calendar-year dependence of tobacco smoke exposure was only implicitly accounted for by a birth cohort effect, we find good agreement between the modeled (expected) and empirical (observed) lung cancer mortality. Model calculations of excess relative lung cancer death risk agree well with those from the descriptive, BEIR VI-type exposure-age-concentration model for WISMUT miners. The large variety of exposure profiles in the cohort leads to a well-determined mechanistic model that in principle allows for an extrapolation from occupational to indoor radon exposure.
Asunto(s)
Neoplasias Pulmonares/etiología , Neoplasias Inducidas por Radiación/etiología , Enfermedades Profesionales/etiología , Radón/efectos adversos , Envejecimiento , Estudios de Cohortes , Humanos , Neoplasias Pulmonares/mortalidad , Minería , Modelos Biológicos , Mutación , Neoplasias Inducidas por Radiación/mortalidad , Enfermedades Profesionales/mortalidad , Riesgo , Fumar/efectos adversos , Factores de Tiempo , UranioRESUMEN
From studies of the atomic bomb survivors, it is well known that ionizing radiation causes several forms of leukemia. However, since the specific mechanism behind this process remains largely unknown, it is difficult to extrapolate carcinogenic effects at acute high-dose exposures to risk estimates for the chronic low-dose exposures that are important for radiation protection purposes. Recently, it has become clear that the induction of acute myeloid leukemia (AML) in CBA/H mice takes place through two key steps, both involving the Sfpi1 gene. A similar mechanism may play a role in human radiation-induced AML. In the present paper, a two-mutation carcinogenesis model is applied to model AML in several data sets of X-ray- and neutron-exposed CBA/H mice. The models obtained provide good fits to the data. A comparison between the predictions for neutron-induced and X-ray-induced AML yields an RBE for neutrons of approximately 3. The model used is considered to be a first step toward a model for human radiation-induced AML, which could be used to estimate risks of exposure to low doses.
Asunto(s)
Modelos Animales de Enfermedad , Leucemia Mieloide Aguda/genética , Modelos Biológicos , Mutación/efectos de la radiación , Neoplasias Inducidas por Radiación/genética , Animales , Relación Dosis-Respuesta en la Radiación , Leucemia Mieloide Aguda/etiología , Funciones de Verosimilitud , Masculino , Ratones , Neutrones/efectos adversos , Efectividad Biológica Relativa , Procesos EstocásticosRESUMEN
Mammography screening is an accepted procedure for early detection of breast tumors among asymptomatic women. Since this procedure involves the use of X rays, it is itself potentially carcinogenic. Although there is general consensus about the benefit of screening for older women, screening practices differ between countries. In this paper radiation risks for these different practices are estimated using a new approach. We model breast cancer induction by ionizing radiation in a cohort of patients exposed to frequent X-ray examinations. The biologically based, mechanistic model provides a better foundation for the extrapolation of risks to different mammography screening practices than empirical models do. The model predicts that the excess relative risk (ERR) doubles when screening starts at age 40 instead of 50 and that a continuation of screening at ages 75 and higher carries little extra risk. The number of induced fatal breast cancers is estimated to be considerably lower than derived from epidemiological studies and from internationally accepted radiation protection risks. The present findings, if used in a risk-benefit analysis for mammography screening, would be more favorable to screening than estimates currently recommended for radiation protection. This has implications for the screening ages that are currently being reconsidered in several countries.
Asunto(s)
Neoplasias de la Mama/diagnóstico por imagen , Mamografía/efectos adversos , Adulto , Anciano , Femenino , Humanos , Persona de Mediana Edad , Neoplasias Inducidas por Radiación , RiesgoRESUMEN
Two-mutation model fits to bone cancer mortality data from mice, rats and beagle dogs injected with (239)Pu or (226)Ra show that (1) it is possible to fit the radiation-related parameters for animals from different strains of the same species together; (2) for every species the same significant parameters are found in the models for (239)Pu and in the models for (226)Ra, and the only difference is in the value of the linear mutation coefficient; and (3) the toxicity ratio, when defined as the ratio of the linear mutation coefficients for (239)Pu over (226)Ra, has a relatively uniform value of approximately 8 for the species considered. This relatively constant ratio enables the development of a (239)Pu model for humans that is based on the radium dial painters and the toxicity ratio for beagles. The model predictions agree well with published risk estimates based on other data and derived using alternative approaches. This has two important implications: (1) The two-mutation model appears to be a useful tool in translating from animal models to humans in a meaningful way; and (2) once a two-mutation model for humans has been derived, radiation risks can be calculated that depend on doses, dose rates and ages at exposure. Such a model therefore supplements published risk estimates that often lack such dependences.