Targeted learning with daily EHR data

Sofrygin, Oleg; Zhu, Zheng; Schmittdiel, Julie A.; Adams, Alyce S.; Grant, Richard W.; Laan, Mark J. van der; Neugebauer, Romain

doi:10.1002/sim.8164

Cited by 31 publications

(27 citation statements)

References 67 publications

(77 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(refer Section VI for more detail), thus effectiveness of sequence labeling is limited for extraction of relevant information from EHR data [42]. To sum up, It is usually a difficult task for the traditional sequence labeling algorithms to extract relevant information from EHR data [43] and traditional algorithms also face challenges in analyzing EHR data due as such algorithms are not suitable for dealing with huge columns of data [44].…”

Section: Deep Learning For Ehr Data Analysismentioning

confidence: 99%

Secondary Use of Electronic Health Record: Opportunities and Challenges

Shah¹,

Khan²

2020

IEEE Access

View full text Add to dashboard Cite

In the present technological era, healthcare providers generate huge amounts of clinical data on a daily basis. Generated clinical data is stored digitally in the form of Electronic Health Record (EHR) as a central data repository of hospitals. Data contained in EHR is not only used for the patients' primary care but also for various secondary purposes such as clinical research, automated disease surveillance and clinical audits for quality enhancement. Using EHR data for secondary purposes without consent or in some cases even with consent creates privacy issues. Secondly, EHR data is also made accessible to various stakeholders including different government agencies at various geographical sites through wired or wireless networks. Sharing of EHR across multiple agencies makes it vulnerable to cyber attacks and also makes it difficult to implement strict privacy laws as in some cases data is shared with organization that is governed by specific regional law. Privacy of an individual could be severely affected when their sensitive private information contained in EHR is leaked or exposed to the public. Data leaks can cause financial losses or an individual may encounter social boycott if their medical condition is exposed in public. To protect patients personal data from such threats, there exists different privacy regulations such as General Data Protection Regulation (GDPR), Health Insurance Portability and Accountability Act (HIPAA) and My Health Record (MHR). However, continually evolving state-of-the-art techniques in Machine Learning (ML), Data Analytics (DA) and hacking are making it even more difficult to completely protect an individual's / patient's privacy. In this article, we have systematically examined various secondary uses of EHR with the aim to highlight how these secondary uses affect patients' privacy. Secondly, we have critically analyzed GDPR & HIPAA regulations and highlighted their possible areas of improvement, considering escalating use of technology and different secondary uses of EHR. INDEX TERMS Electronic Health Records (EHR), ethical concerns, General Data Protection Regulation (GDPR), privacy, secondary uses of EHR.

show abstract

Section: Deep Learning For Ehr Data Analysismentioning

confidence: 99%

Secondary Use of Electronic Health Record: Opportunities and Challenges

Shah¹,

Khan²

2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Statistical Analyses-We implemented longitudinal targeted maximum likelihood estimation (longitudinal TMLE) 28,30 to estimate wave-specific associations between having an adult child in the US and outcomes of interest. Briefly, longitudinal TMLE models the expected outcome via a sequence of nested regressions and uses wave-specific inverse probability of treatment and censoring weights to update the resulting estimates.…”

Section: Unmet Needs For Assistancementioning

confidence: 99%

“…Due to computational demands, we only used the gradient boosting machine algorithm for treatment and attrition models for analyses of binary outcomes. Estimation was performed using stremr 30,48 and sl3 49 packages for R. 50 We implemented multiple imputation procedures to address missing data using the Amelia package for R. 51 We created ten multiply imputed datasets for the baseline wave, and incorporated mean values from multiply imputed baseline models into the multiple imputation of time-varying measures at successive follow-up waves. We combined estimates using Rubin's rules.…”

Section: Unmet Needs For Assistancementioning

confidence: 99%

See 1 more Smart Citation

Having an Adult Child in the United States, Physical Functioning, and Unmet Needs for Care Among Older Mexican Adults

et al. 2019

Self Cite

View full text Add to dashboard Cite

Background: Migration of adult children may impact the health of aging parents who remain in low-and middle-income countries. Prior studies have uncovered mixed associations between adult child migration status and physical functioning of older parents; none to our knowledge has examined the impact on unmet caregiving needs. Methods: Data come from a population-based study of Mexican adults >50 years old. We analyzed 11806 respondents surveyed over an 11-year period. We used longitudinal targeted maximum likelihood estimation to estimate associations between having an adult child in the US and lower-body functional limitations, and both needs and unmet needs for assistance with basic or instrumental activities of daily living (ADLs/IADLs). Results: For women, having an adult child in the US at baseline and 2-year follow-up was associated with fewer lower body functional limitations (marginal risk difference (RD): −0.14, 95% CI: −0.26, −0.01) and fewer ADLs/IADLs (RD: −0.08, 95% CI: −0.16, −0.001) at 2-year follow-up. Having an adult child in the US at all three study waves was associated with more lower-body functional limitations at 11-year follow-up (RD: 0.04, 95% CI: 0.01, 0.06). Having an adult child in the US was associated with a higher prevalence of unmet needs for assistance at two (RD: 0.13, 95% CI: 0.04, 0.21) and 11-year follow-up for women (RD: 0.07, 95% CI: −0.02, 0.15) and 11-year follow-up for men (RD: 0.08, 95% CI: 0.00, 0.16).

show abstract

“…This leaves open the question of how to select an appropriate discretization, ideally at a fine enough scale that would capture all time‐dependent confounding while balancing for inflated variance. Even though some concerns and warnings regarding arbitrary discretization have emerged, 11,12 no criteria have yet been proposed to guide an analyst's choice of discretization. We investigate different tools which collectively inform whether there is adequate data support for a given discretization for use with pooled longitudinal targeted maximum likelihood estimation (LTMLE) 13 .…”

Section: Introductionmentioning

confidence: 99%

Impact of discretization of the timeline for longitudinal causal inference methods

Guerra

Schnitzer

Forget

et al. 2020

Statistics in Medicine

View full text Add to dashboard Cite

In longitudinal settings, causal inference methods usually rely on a discretization of the patient timeline that may not reflect the underlying data generation process. This article investigates the estimation of causal parameters under discretized data. It presents the implicit assumptions practitioners make but do not acknowledge when discretizing data to assess longitudinal causal parameters. We illustrate that differences in point estimates under different discretizations are due to the data coarsening resulting in both a modified definition of the parameter of interest and loss of information about time‐dependent confounders. We further investigate several tools to advise analysts in selecting a timeline discretization for use with pooled longitudinal targeted maximum likelihood estimation for the estimation of the parameters of a marginal structural model. We use a simulation study to empirically evaluate bias at different discretizations and assess the use of the cross‐validated variance as a measure of data support to select a discretization under a chosen data coarsening mechanism. We then apply our approach to a study on the relative effect of alternative asthma treatments during pregnancy on pregnancy duration. The results of the simulation study illustrate how coarsening changes the target parameter of interest as well as how it may create bias due to a lack of appropriate control for time‐dependent confounders. We also observe evidence that the cross‐validated variance acts well as a measure of support in the data, by being minimized at finer discretizations as the sample size increases.

show abstract

Targeted learning with daily EHR data

Cited by 31 publications

References 67 publications

Secondary Use of Electronic Health Record: Opportunities and Challenges

Secondary Use of Electronic Health Record: Opportunities and Challenges

Having an Adult Child in the United States, Physical Functioning, and Unmet Needs for Care Among Older Mexican Adults

Impact of discretization of the timeline for longitudinal causal inference methods

Contact Info

Product

Resources

About