Patient-generated secure messages and eVisits on a patient portal: are patients at risk?

North, Frederick; Crane, Sarah J.; Stroebel, Robert J.; Stephen, S.; Edell, Eric S.; Tulledge–Scheitel, Sidna M.

doi:10.1136/amiajnl-2012-001208

Cited by 65 publications

(64 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The physicians in this system are not compensated for providing e‐visit access and are expected to respond within 48 h. While we do not analyze the patient characteristics in this study, prior literature shows that e‐visit adopters are more likely to be male, White (as opposed to African‐American or Hispanic), high income, and have more comorbidities (Bavafa et al, ; Yamin et al, ). The most frequently discussed e‐visit topics are medication, lab results, and patient questions about symptoms (Byrne, Elliott, & Firek, ; North et al, ). In our data, physicians start using e‐visits at different points in time, a feature that we leverage in a robustness check using instrumental variable analysis.…”

Section: Empirical Strategymentioning

confidence: 99%

Work after work: The impact of new service delivery models on work hours

Bavafa

Terwiesch

2019

J of Ops Management

View full text Add to dashboard Cite

In many professional service organizations (PSOs), a single server such as a physician or lawyer delivers services to customers across multiple channels. In these settings, there exists a risk that work obligations encroach on the personal lives of the servers. We empirically examine this concern in the primary care setting of physicians providing care to patients following the introduction of a new service delivery channel, e‐visits. Our data cover an 8.5‐year timespan and include 3.3 million patient encounters, including about one million e‐visits. We find that e‐visit adoption nearly doubles the number of hours containing work (defined as office visits or e‐visits) each week. Of the additional hours containing such work, only 49% are in weekday business hours; the rest are in weekday evenings or early mornings (37%) and on weekends (14%). We also show that increases in office or e‐visit workloads lead to more post‐workday activity in the e‐visit channel. We conclude the analysis by examining the impact of these workloads on e‐visit work content, as measured by the number of words and amount of time spent on each e‐visit; we estimate precise but practically small effects of workload on these work content outcomes.

show abstract

Section: Empirical Strategymentioning

confidence: 99%

Work after work: The impact of new service delivery models on work hours

Bavafa

Terwiesch

2019

J of Ops Management

View full text Add to dashboard Cite

show abstract

“…Together, these limitations could lead to misdiagnosis or poor quality of care. [4][5][6] Although there have been several studies of telemedicine services between primary care providers (PCPs) and their existing patients, 5,7 there has been limited prior work on the quality of DTC telemedicine outside of the medical home and whether DTC telemedicine actually increases access for the underserved. In prior work we have examined antibiotic prescribing in DTC telemedicine 8 but did not look at established quality measures or access to care.…”

Section: Introductionmentioning

confidence: 99%

Access and Quality of Care in Direct-to-Consumer Telemedicine

Uscher‐Pines

Mulcahy

Cowling³

et al. 2016

Telemedicine and e-Health

140

145

View full text Add to dashboard Cite

Background: Direct-to-consumer (DTC) telemedicine serves millions of patients; however, there is limited research on the care provided. This study compared the quality of care at Teladoc (www.teladoc.com), a large DTC telemedicine company, with that at physician offices and compared access to care for Teladoc users and nonusers. Materials and Methods: Claims from all enrollees 18-64 years of age in the California Public Employees' Retirement System health maintenance organization between April 2012 and October 2013 were analyzed. We compared the performance of Teladoc and physician offices on applicable Healthcare Effectiveness Data and Information Set measures. Using geographic information system analyses, we compared Teladoc users and nonusers with respect to rural location and available primary care physicians. Results: Of enrollees offered Teladoc (n = 233,915), 3,043 adults had a total of 4,657 Teladoc visits. For the pharyngitis performance measure (ordering strep test), Teladoc performed worse than physician offices (3% versus 50%, p < 0.01). For the back pain measure (not ordering imaging), Teladoc and physician offices had similar performance (88% versus 79%, p = 0.20). For the bronchitis measure (not ordering antibiotics), Teladoc performed worse than physician offices (16.7 versus 27.9%, p < 0.01). In adjusted models, Teladoc users were not more likely to be located within a healthcare professional shortage area (odds ratio = 1.12, p = 0.10) or rural location (odds ratio = 1.0, p = 0.10). Conclusions: Teladoc providers were less likely to order diagnostic testing and had poorer performance on appropriate antibiotic prescribing for bronchitis. Teladoc users were not preferentially located in underserved communities. Short-term needs include ongoing monitoring of quality and additional marketing and education to increase telemedicine use among underserved patients.

show abstract

“…[15, 16] North et al manually classified 323 messages, demonstrating 37% were medication related, 23% were symptom related, 20% were test related, 7% were medical questions, 6% were acknowledgements, and 9% had more than one issue [16]. Haun et al asked senders classify their messages in predefined categories and observed the following distribution, although user-assigned categories were not consistently applied accurately: 59% general (i.e., condition management/report, specialty/procedure request, correspondence request, medication refill request,), 24% appointments (i.e., confirmations, cancellations, specialty appointment requests), and 16% refill and medication inquiries [15].…”

Section: Introductionmentioning

confidence: 99%

A comparison of rule-based and machine learning approaches for classifying patient portal messages

Cronin

Fabbri

Denny

et al. 2017

International Journal of Medical Informatics

View full text Add to dashboard Cite

Objective Secure messaging through patient portals is an increasingly popular way that consumers interact with healthcare providers. The increasing burden of secure messaging can affect clinic staffing and workflows. Manual management of portal messages is costly and time consuming. Automated classification of portal messages could potentially expedite message triage and delivery of care. Materials and Methods We developed automated patient portal message classifiers with rule-based and machine learning techniques using bag of words and natural language processing (NLP) approaches. To evaluate classifier performance, we used a gold standard of 3253 portal messages manually categorized using a taxonomy of communication types (i.e., main categories of informational, medical, logistical, social, and other communications, and subcategories including prescriptions, appointments, problems, tests, follow-up, contact information, and acknowledgement). We evaluated our classifiers’ accuracies in identifying individual communication types within portal messages with area under the receiver-operator curve (AUC). Portal messages often contain more than one type of communication. To predict all communication types within single messages, we used the Jaccard Index. We extracted the variables of importance for the random forest classifiers. Results The best performing approaches to classification for the major communication types were: logistic regression for medical communications (AUC: 0.899); basic (rule-based) for informational communications (AUC: 0.842); and random forests for social communications and logistical communications (AUCs: 0.875 and 0.925, respectively). The best performing classification approach of classifiers for individual communication subtypes was random forests for Logistical-Contact Information (AUC: 0.963). The Jaccard Indices by approach were: basic classifier, Jaccard Index: 0.674; Naïve Bayes, Jaccard Index: 0.799; random forests, Jaccard Index: 0.859; and logistic regression, Jaccard Index: 0.861. For medical communications, the most predictive variables were NLP concepts (e.g. Temporal_Concept, which maps to ‘morning’, ‘evening’ and Idea_or_Concept which maps to ‘appointment’ and ‘refill’). For logistical communications, the most predictive variables contained similar numbers of NLP variables and words (e.g. Telephone mapping to ‘phone’, ‘insurance’). For social and informational communications, the most predictive variables were words (e.g. social: ‘thanks’, ‘much’, informational: ‘question’, ‘mean’). Conclusions This study applies automated classification methods to the content of patient portal messages and evaluates the application of NLP techniques on consumer communications in patient portal messages. We demonstrated that random forest and logistic regression approaches accurately classified the content of portal messages, although the best approach to classification varied by communication type. Words were the most predictive variables for classification of most communication types...

show abstract

Patient-generated secure messages and eVisits on a patient portal: are patients at risk?

Cited by 65 publications

References 14 publications

Work after work: The impact of new service delivery models on work hours

Work after work: The impact of new service delivery models on work hours

Access and Quality of Care in Direct-to-Consumer Telemedicine

A comparison of rule-based and machine learning approaches for classifying patient portal messages

Contact Info

Product

Resources

About