Abstract Aim Techniques that predict species potential distributions by combining observed occurrence records with environmental variables show much potential for application across a range of biogeographical analyses. Some of the most promising applications relate to species for which occurrence records are scarce, due to cryptic habits, locally restricted distributions or low sampling effort. However, the minimum sample sizes required to yield useful predictions remain difficult to determine. Here we developed and tested a novel jackknife validation approach to assess the ability to predict species occurrence when fewer than 25 occurrence records are available. Location Madagascar. Methods Models were developed and evaluated for 13 species of secretive leaf‐tailed geckos ( Uroplatus spp.) that are endemic to Madagascar, for which available sample sizes range from 4 to 23 occurrence localities (at 1 km 2 grid resolution). Predictions were based on 20 environmental data layers and were generated using two modelling approaches: a method based on the principle of maximum entropy (Maxent) and a genetic algorithm (GARP). Results We found high success rates and statistical significance in jackknife tests with sample sizes as low as five when the Maxent model was applied. Results for GARP at very low sample sizes (less than c. 10) were less good. When sample sizes were experimentally reduced for those species with the most records, variability among predictions using different combinations of localities demonstrated that models were greatly influenced by exactly which observations were included. Main conclusions We emphasize that models developed using this approach with small sample sizes should be interpreted as identifying regions that have similar environmental conditions to where the species is known to occur, and not as predicting actual limits to the range of a species. The jackknife validation approach proposed here enables assessment of the predictive ability of models built using very small sample sizes, although use of this test with larger sample sizes may lead to overoptimistic estimates of predictive power. Our analyses demonstrate that geographical predictions developed from small numbers of occurrence records may be of great value, for example in targeting field surveys to accelerate the discovery of unknown populations and species.
We present a 5.3‐Myr stack (the “LR04” stack) of benthic δ 18 O records from 57 globally distributed sites aligned by an automated graphic correlation algorithm. This is the first benthic δ 18 O stack composed of more than three records to extend beyond 850 ka, and we use its improved signal quality to identify 24 new marine isotope stages in the early Pliocene. We also present a new LR04 age model for the Pliocene‐Pleistocene derived from tuning the δ 18 O stack to a simple ice model based on 21 June insolation at 65°N. Stacked sedimentation rates provide additional age model constraints to prevent overtuning. Despite a conservative tuning strategy, the LR04 benthic stack exhibits significant coherency with insolation in the obliquity band throughout the entire 5.3 Myr and in the precession band for more than half of the record. The LR04 stack contains significantly more variance in benthic δ 18 O than previously published stacks of the late Pleistocene as the result of higher‐resolution records, a better alignment technique, and a greater percentage of records from the Atlantic. Finally, the relative phases of the stack's 41‐ and 23‐kyr components suggest that the precession component of δ 18 O from 2.7–1.6 Ma is primarily a deep‐water temperature signal and that the phase of δ 18 O precession response changed suddenly at 1.6 Ma.
In a follow-up study of incarcerated Connecticut youth, 69 subjects were interviewed during young adulthood. On follow-up, 26 gave histories of abuse discrepant with histories obtained from records and interviews conducted in adolescence. Eleven subjects agreed to an additional clarification interview, at which time they were apprised of the discrepancies. Of these, eight had adolescent records indicating that abuse had occurred but denied abuse during the adult follow-up interview. The remaining three had adolescent records indicating no abuse had ever occurred, but, on follow-up, reported having been abused. The additional clarification interviews revealed that all 11 subjects with discrepant histories had, in fact, been abused. Reasons for these discrepant data and strategies to enhance the investigator's ability to obtain accurate data regarding abuse are discussed. 69 personnes qui avaient été emprisonnées dans le Connecticut autrefois ont été soumises à un entretien alors qu'elles avaient atteint l'âge adulte. Dans ce suivi, 26 individus ont donné des récits de sévices subis qui n'étaient pas les mêmes que les récits extraits de leur dossier et entretiens enregistrés pendant l'adolescence. 11 de ces personnes ont accepté d'avoir un entretien de clarification supplémentaire au cours duquel on a évalué les discordances avec leurs anciens dossiers. Sur ces 11, 8 avaient des dossiers d'adolescent indiquant que des sévices avaient été subis; au cours de l'entretien de suivi à l'âge adulte, ils ont rétracté ces affirmations. Les 3 autres avaient indiqué qu'ils n'avaient pas subi de sévices comme adolescent mais au contraire à l'âge adulte ils ont dit qu'ils avaient en fait été maltraités. L'entretien de clarification additionnel a pu démontrer que tous ces 11 sujets sans exception avaient en fait bien été l'objet de sévices. Les raisons pour ces données discordantes et les stratégies à utiliser lorsqu'on interroge ce genre de personne font l'objet d'une discussion puisque l'on désire obtenir des données fiables. En una investigación complementaria dejóvenes encarcelados en Connecticut, 69 sujetos fueron entrevistados cuando eran adultos jóvenes. En las entrevistas, 26 dieron historias de abuso discrepantes con las historias obtenidas de documentos y entrevistas obtenidos durante su adolescencia. Once sujetos estuvieron de acuerdo con participar en una entrevista adicional de clarificación durante la cual se les informó de las discrepancias. Los documentos obtenidos durante la adolescencia de ocho de ellos indicaron que había ocumdo abuso, pero este fué negado durante la entrevista complementaria cuando adultos. Los documentos acerca de los otros tres indicaron que no había ocurrido abuso durante la adolescencia, pero ellos afirmaron en la entrevista complementaria cuando adultos que sí había ocurrido. Entrevistas clarificadoras subsecuentes revelaron que todos los once con historias discrepantes habían, en realidad, sido abusados. Se comentan las causas de las discrepancias así como estrategias para acrecentar la capacidad del investigador para obtener datos exactos con respecto al abuso.
BACKGROUND: Despite a consensus that the use of health information technology should lead to more efficient, safer, and higher-quality care, there are no reliable estimates of the prevalence of adoption of electronic health records in U.S. hospitals. METHODS: We surveyed all acute care hospitals that are members of the American Hospital Association for the presence of specific electronic-record functionalities. Using a definition of electronic health records based on expert consensus, we determined the proportion of hospitals that had such systems in their clinical areas. We also examined the relationship of adoption of electronic health records to specific hospital characteristics and factors that were reported to be barriers to or facilitators of adoption. RESULTS: On the basis of responses from 63.1% of hospitals surveyed, only 1.5% of U.S. hospitals have a comprehensive electronic-records system (i.e., present in all clinical units), and an additional 7.6% have a basic system (i.e., present in at least one clinical unit). Computerized provider-order entry for medications has been implemented in only 17% of hospitals. Larger hospitals, those located in urban areas, and teaching hospitals were more likely to have electronic-records systems. Respondents cited capital requirements and high maintenance costs as the primary barriers to implementation, although hospitals with electronic-records systems were less likely to cite these barriers than hospitals without such systems. CONCLUSIONS: The very low levels of adoption of electronic health records in U.S. hospitals suggest that policymakers face substantial obstacles to the achievement of health care performance goals that depend on health information technology. A policy strategy focused on financial support, interoperability, and training of technical support staff may be necessary to spur adoption of electronic-records systems in U.S. hospitals.
Spatial thinning of species occurrence records can help address problems associated with spatial sampling biases. Ideally, thinning removes the fewest records necessary to substantially reduce the effects of sampling bias, while simultaneously retaining the greatest amount of useful information. Spatial thinning can be done manually; however, this is prohibitively time consuming for large datasets. Using a randomization approach, the ‘thin’ function in the spThin R package returns a dataset with the maximum number of records for a given thinning distance, when run for sufficient iterations. We here provide a worked example for the Caribbean spiny pocket mouse, where the results obtained match those of manual thinning.
BACKGROUND: Most US medical records lack socioeconomic data, hindering studies of social gradients in health and ascertainment of whether study samples are representative of the general population. This study assessed the validity of a census-based approach in addressing these problems. METHODS: Socioeconomic data from 1980 census tracts and block groups were matched to the 1985 membership records of a large prepaid health plan (n = 1.9 million), with the link provided by each individual's residential address. Among a subset of 14,420 Black and White members, comparisons were made of the association of individual, census tract, and census block-group socioeconomic measures with hypertension, height, smoking, and reproductive history. RESULTS: Census-level and individual-level socioeconomic measures were similarly associated with the selected health outcomes. Census data permitted assessing response bias due to missing individual-level socioeconomic data and also contextual effects involving the interaction of individual- and neighborhood-level socioeconomic traits. On the basis of block-group characteristics, health plan members generally were representative of the total population; persons in impoverished neighborhoods, however, were underrepresented. CONCLUSIONS: This census-based methodology offers a valid and useful approach to overcoming the absence of socioeconomic data in most US medical records.
To better define levels of accomplishment for publishing journal articles in strategic management, a bibliometric study was performed on the publication records of 96 doctorates in the field whose first post-degree job was in academics. By examining 20 journals that are outlets for research in strategic management, publication records were developed for each individual for the first 5–10 years following receipt of the doctoral degree. Two factors influenced the publication records of these new faculty. Having publications prior to receiving the doctorate and getting a first job at an institution with a graduate program in management were associated with more frequent publishing after an academic career began. As expected, the number of papers published was related to the likelihood of receiving tenure. However, despite the fact that they had produced more papers during the first 5 years than male faculty members and had higher citation rates, female faculty members were less likely to receive tenure. The findings are discussed in terms of institutional policy for hiring and evaluating new faculty.
A variety of geophysical records are examined to determine the dependence upon the lag s of a quantity called ‘rescaled range,’ denoted by R ( t , s )/ S ( t , s ). If there had been no appreciable dependence between two values of the record at very distant points in time, the ratio R / S would have been proportional to s 0.5 . But, in fact, as first noted by Edwin Hurst, the R / S ratio of hydrological and other geophysical records is proportional to s H with H ≠ 0.5. Hurst's original claims must be tightened and hedged, and his estimates of H must be discarded, but his general idea will be shown to be correct. We have shown elsewhere that this behavior of R / S means that the strength of long‐range statistical dependence in geophysical records is considerable.
We show that easily accessible digital records of behavior, Facebook Likes, can be used to automatically and accurately predict a range of highly sensitive personal attributes including: sexual orientation, ethnicity, religious and political views, personality traits, intelligence, happiness, use of addictive substances, parental separation, age, and gender. The analysis presented is based on a dataset of over 58,000 volunteers who provided their Facebook Likes, detailed demographic profiles, and the results of several psychometric tests. The proposed model uses dimensionality reduction for preprocessing the Likes data, which are then entered into logistic/linear regression to predict individual psychodemographic profiles from Likes. The model correctly discriminates between homosexual and heterosexual men in 88% of cases, African Americans and Caucasian Americans in 95% of cases, and between Democrat and Republican in 85% of cases. For the personality trait "Openness," prediction accuracy is close to the test-retest accuracy of a standard personality test. We give examples of associations between attributes and Likes and discuss implications for online personalization and privacy.
Predictive modeling with electronic health record (EHR) data is anticipated to drive personalized medicine and improve healthcare quality. Constructing predictive statistical models typically requires extraction of curated predictor variables from normalized EHR data, a labor-intensive process that discards the vast majority of information in each patient's record. We propose a representation of patients' entire raw EHR records based on the Fast Healthcare Interoperability Resources (FHIR) format. We demonstrate that deep learning methods using this representation are capable of accurately predicting multiple medical events from multiple centers without site-specific data harmonization. We validated our approach using de-identified EHR data from two US academic medical centers with 216,221 adult patients hospitalized for at least 24 h. In the sequential format we propose, this volume of EHR data unrolled into a total of 46,864,534,945 data points, including clinical notes. Deep learning models achieved high accuracy for tasks such as predicting: in-hospital mortality (area under the receiver operator curve [AUROC] across sites 0.93-0.94), 30-day unplanned readmission (AUROC 0.75-0.76), prolonged length of stay (AUROC 0.85-0.86), and all of a patient's final discharge diagnoses (frequency-weighted AUROC 0.90). These models outperformed traditional, clinically-used predictive models in all cases. We believe that this approach can be used to create accurate and scalable predictions for a variety of clinical scenarios. In a case study of a particular prediction, we demonstrate that neural networks can be used to identify relevant information from the patient's chart.
The widespread use of electronic health records (EHRs) in the United States is inevitable. EHRs will improve caregivers' decisions and patients' outcomes. Once patients experience the benefits of this technology, they will demand nothing less from their providers. Hundreds of thousands of physicians have already seen these benefits in their clinical practice.But inevitability does not mean easy transition. We have years of professional agreement and bipartisan consensus regarding the potential value of EHRs. Yet we have not moved significantly to extend the availability of EHRs from a few large institutions to the smaller clinics and practices where most Americans . . .
Secondary use of electronic health records (EHRs) promises to advance clinical research and better inform clinical decision making. Challenges in summarizing and representing patient data prevent widespread practice of predictive modeling using EHRs. Here we present a novel unsupervised deep feature learning method to derive a general-purpose patient representation from EHR data that facilitates clinical predictive modeling. In particular, a three-layer stack of denoising autoencoders was used to capture hierarchical regularities and dependencies in the aggregated EHRs of about 700,000 patients from the Mount Sinai data warehouse. The result is a representation we name "deep patient". We evaluated this representation as broadly predictive of health states by assessing the probability of patients to develop various diseases. We performed evaluation using 76,214 test patients comprising 78 diseases from diverse clinical domains and temporal windows. Our results significantly outperformed those achieved using representations based on raw EHR data and alternative feature learning strategies. Prediction performance for severe diabetes, schizophrenia, and various cancers were among the top performing. These findings indicate that deep learning applied to EHRs can derive patient representations that offer improved clinical predictions, and could provide a machine learning framework for augmenting clinical decision systems.
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
暂无摘要（点击查看原文获取完整内容）
Recently there has been a remarkable upsurge in activity surrounding the adoption of personal health record (PHR) systems for patients and consumers. The biomedical literature does not yet adequately describe the potential capabilities and utility of PHR systems. In addition, the lack of a proven business case for widespread deployment hinders PHR adoption. In a 2005 working symposium, the American Medical Informatics Association's College of Medical Informatics discussed the issues surrounding personal health record systems and developed recommendations for PHR-promoting activities. Personal health record systems are more than just static repositories for patient data; they combine data, knowledge, and software tools, which help patients to become active participants in their own care. When PHRs are integrated with electronic health record systems, they provide greater benefits than would stand-alone systems for consumers. This paper summarizes the College Symposium discussions on PHR systems and provides definitions, system characteristics, technical architectures, benefits, barriers to adoption, and strategies for increasing adoption.