Concept: Electronic health record
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.
Hospitals and clinics are adapting to new technologies and implementing electronic health records, but the efforts need to be aligned explicitly with goals for patient safety. EHRs bring the risks of both technical failures and inappropriate use, but they can also help to monitor and improve patient safety.
The HITECH Act created incentives to encourage adoption of electronic health records. As of May 2012, only 12.2% of 62,226 eligible professionals had attested to meaningful use, including 9.8% of specialists and 17.8% of primary care providers.
Primary care physicians spend nearly 2 hours on electronic health record (EHR) tasks per hour of direct patient care. Demand for non-face-to-face care, such as communication through a patient portal and administrative tasks, is increasing and contributing to burnout. The goal of this study was to assess time allocated by primary care physicians within the EHR as indicated by EHR user-event log data, both during clinic hours (defined as 8:00 am to 6:00 pm Monday through Friday) and outside clinic hours.
Data mining approaches have been increasingly applied to the electronic health record and have led to the discovery of numerous clinical associations. Recent data mining studies have suggested a potential association between cat bites and human depression. To explore this possible association in more detail we first used administrative diagnosis codes to identify patients with either depression or bites, drawn from a population of 1.3 million patients. We then conducted a manual chart review in the electronic health record of all patients with a code for a bite to accurately determine which were from cats or dogs. Overall there were 750 patients with cat bites, 1,108 with dog bites, and approximately 117,000 patients with depression. Depression was found in 41.3% of patients with cat bites and 28.7% of those with dog bites. Furthermore, 85.5% of those with both cat bites and depression were women, compared to 64.5% of those with dog bites and depression. The probability of a woman being diagnosed with depression at some point in her life if she presented to our health system with a cat bite was 47.0%, compared to 24.2% of men presenting with a similar bite. The high proportion of depression in patients who had cat bites, especially among women, suggests that screening for depression could be appropriate in patients who present to a clinical provider with a cat bite. Additionally, while no causative link is known to explain this association, there is growing evidence to suggest that the relationship between cats and human mental illness, such as depression, warrants further investigation.
BACKGROUND: Implementation and use of electronic health records (EHRs) could lead to potential improvements in quality of care. However, the use of EHRs also introduces unique and often unexpected patient safety risks. Proactive assessment of risks and vulnerabilities can help address potential EHR-related safety hazards before harm occurs; however, current risk assessment methods are underdeveloped. The overall objective of this project is to develop and validate proactive assessment tools to ensure that EHR-enabled clinical work systems are safe and effective. METHODS: This work is conceptually grounded in an 8-dimension model of safe and effective health information technology use. Our first aim is to develop self-assessment guides that can be used by health care institutions to evaluate certain high-risk components of their EHR-enabled clinical work systems. We will solicit input from subject matter experts and relevant stakeholders to develop guides focused on 9 specific risk areas and will subsequently pilot test the guides with individuals representative of likely users. The second aim will be to examine the utility of the self-assessment guides by beta testing the guides at selected facilities and conducting on-site evaluations. Our multidisciplinary team will use a variety of methods to assess the content validity and perceived usefulness of the guides, including interviews, naturalistic observations, and document analysis. The anticipated output of this work will be a series of self-administered EHR safety assessment guides with clear, actionable, checklist-type items. DISCUSSION: Proactive assessment of patient safety risks increases the resiliency of health care organizations to unanticipated hazards of EHR use. The resulting products and lessons learned from the development of the assessment guides are expected to be helpful to organizations that are beginning the EHR selection and implementation process as well as those that have already implemented EHRs. Findings from our project, currently underway, will inform future efforts to validate and implement tools that can be used by health care organizations to improve the safety of EHR-enabled clinical work systems.
Purchasing electronic health records (EHRs) typically follows a process in which potential adopters actively seek information, compare alternatives, and form attitudes towards the product. A potential source of information on EHRs that can be used in the process is vendor websites. It is unclear how much product information is presented on EHR vendor websites or the extent of its value during EHR purchasing decisions.
BACKGROUND: Various problems concerning the introduction of personal health records in everyday healthcare practice are reported to be associated with physicians' unfamiliarity with systematic means of electronically collecting health information about their patients (e.g. electronic health records - EHRs). Such barriers may further prevent the role physicians have in their patient encounters and the influence they can have in accelerating and diffusing personal health records (PHRs) to the patient community. One way to address these problems is through medical education on PHRs in the context of EHR activities within the undergraduate medical curriculum and the medical informatics courses in specific. In this paper, the development of an educational PHR activity based on Google Health is reported. Moreover, student responses on PHR’s use and utility are collected and presented. The collected responses are then modelled to relate the satisfaction level of students in such a setting to the estimation about their attitude towards PHRs in the future. METHODS: The study was conducted by designing an educational scenario about PHRs, which consisted of student instruction on Google Health as a model PHR and followed the guidelines of a protocol that was constructed for this purpose. This scenario was applied to a sample of 338 first-year undergraduate medical students. A questionnaire was distributed to each one of them in order to obtain Likert-like scale data on the sample’s response with respect to the PHR that was used; the data were then further analysed descriptively and in terms of a regression analysis to model hypothesised correlations. RESULTS: Students displayed, in general, satisfaction about the core PHR functions they used and they were optimistic about using them in the future, as they evaluated quite high up the level of their utility. The aspect they valued most in the PHR was its main role as a record-keeping tool, while their main concern was related to the negative effect their own opinion might have on the use of PHRs by patients. Finally, the estimate of their future attitudes towards PHR integration was found positively dependent of the level of PHR satisfaction that they gained through their experience (rho = 0.524, p <0.001). CONCLUSIONS: The results indicate that students support PHRs as medical record keeping helpers and perceive them as beneficial to healthcare. They also underline the importance of achieving good educational experiences in improving PHR perspectives inside such educational activities. Further research is obviously needed to establish the relative long-term effect of education to other methods of exposing future physicians to PHRs.
BACKGROUND: The increasing availability of Electronic Health Record (EHR) data and specifically free-text patient notes presents opportunities for phenotype extraction. Text-mining methods in particular can help disease modeling by mapping named-entities mentions to terminologies and clustering semantically related terms. EHR corpora, however, exhibit specific statistical and linguistic characteristics when compared with corpora in the biomedical literature domain. We focus on copy-and-paste redundancy: clinicians typically copy and paste information from previous notes when documenting a current patient encounter. Thus, within a longitudinal patient record, one expects to observe heavy redundancy. In this paper, we ask three research questions: (i) How can redundancy be quantified in large-scale text corpora? (ii) Conventional wisdom is that larger corpora yield better results in text mining. But how does the observed EHR redundancy affect text mining? Does such redundancy introduce a bias that distorts learned models? Or does the redundancy introduce benefits by highlighting stable and important subsets of the corpus? (iii) How can one mitigate the impact of redundancy on text mining? RESULTS: We analyze a large-scale EHR corpus and quantify redundancy both in terms of word and semantic concept repetition. We observe redundancy levels of about 30% and non-standard distribution of both words and concepts. We measure the impact of redundancy on two standard text-mining applications: collocation identification and topic modeling. We compare the results of these methods on synthetic data with controlled levels of redundancy and observe significant performance variation. Finally, we compare two mitigation strategies to avoid redundancy-induced bias: (i) a baseline strategy, keeping only the last note for each patient in the corpus; (ii) removing redundant notes with an efficient fingerprinting-based algorithm. aFor text mining, preprocessing the EHR corpus with fingerprinting yields significantly better results. CONCLUSIONS: Before applying text-mining techniques, one must pay careful attention to the structure of the analyzed corpora. While the importance of data cleaning has been known for low-level text characteristics (e.g., encoding and spelling), high-level and difficult-to-quantify corpus characteristics, such as naturally occurring redundancy, can also hurt text mining. Fingerprinting enables text-mining techniques to leverage available data in the EHR corpus, while avoiding the bias introduced by redundancy.
Personally controlled health management systems (PCHMS), which include a personal health record (PHR), health management tools, and consumer resources, represent the next stage in consumer eHealth systems. It is still unclear, however, what features contribute to an engaging and efficacious PCHMS.