Association between Patient Portal use and Office Visits.

Extant research on the relationship between portal use and office visits is mixed. Some researchers have stated that there is no correlation between the two events, others have found a positive correlation, and still others have found a negative correlation between portal use and office visits. Through the use of system-generated data from two different portal systems, we demonstrate the correlation between portal visits and office visits. We also demonstrate how this correlation differs between two institutions and across three demographic attributes. We performed a retrospective bivariate correlation analysis between portal visits and office visits. The correlation analysis was followed by an application of Fisher's z transformation of the correlation coefficients to determine significant differences in correlation across ethnicity, gender, and age. We found a positive and statistically significant correlation between portal visits and office visits among both hospital patients (n = 2,594, r = .239, p < .000) and university health service patients (n = 1,233, r = .596, p < .000). This correlation varies significantly across the dimensions of ethnicity, gender, and age. Our findings support the importance of portal use in the care continuum. Future research should aim to better understand the nuance of personal characteristics on the relationship between portal visits and office visits. Knowing these nuances can assist practitioners with further promoting patient self-engagement through portal use.

The Impact of Portal Satisfaction on Portal Use and Health-Seeking Behavior: Structural Equation Analysis.

BACKGROUND: Our study addresses a gap in the modern information systems (IS) use literature by investigating factors that explain patient portal satisfaction (SWP) and perceptions about health-seeking behavior (HSB). A novel feature of our study is the incorporation of actual portal use data rather than the perceptions of use intention, which prevails in the modern IS literature. OBJECTIVE: This study aimed to empirically validate factors that influence SWP as an influencing agent on portal use and HSB. Our population segment was comprised of college students with active patient portal accounts. METHODS: Using web-based survey data from a population of portal users (n=1142) in a university health center, we proposed a theoretical model that adapts constructs from the Technology Acceptance Model by Davis, the revised Technology Adoption Model by Venkatesh, the Unified Theory of the Acceptance and Use of Technology 2, and the Health Belief Model by Rosenstock et al. We validated our model using structural equation modeling techniques. RESULTS: Our model explained nearly 65% of the variance in SWP (R2=0.6499), nearly 33% of the variance in portal use (R2=0.3250), and 29% of the variance in HSB (R2=0.2900). Statistically significant antecedents of SWP included social influence (beta=.160, t499=6.145), habit (beta=.114, t499=4.89), facilitating conditions (beta=.062, t499=2.401), effort expectancy (beta=.311, t499=11.149), and performance expectancy (beta=.359, t499=11.588). SWP influenced HSB (beta=.505, t499=19.705) and portal use (beta=.050, t499=2.031). We did not find a statistically significant association between portal use and HSB (beta=.015, t499=0.513). Perceived severity significantly influenced HSB (beta=.129, t499=4.675) but not portal use (beta=.012, t499=.488). CONCLUSIONS: Understanding the importance of SWP and the role it plays in influencing HSB may point to future technology design considerations for information technology developers and health care providers. We extend current Expectancy Confirmation Theory research by finding a positive association between SWP and portal use.

A state-level analysis of macro-level factors associated with hospital readmissions.

Investigation of the factors that contribute to hospital readmissions has focused largely on individual level factors. We extend the knowledge base by exploring macrolevel factors that may contribute to readmissions. We point to environmental, behavioral, and socioeconomic factors that are emerging as correlates to readmissions. Data were taken from publicly available reports provided by multiple agencies. Partial Least Squares-Structural Equation Modeling was used to test the association between economic stability and environmental factors on opioid use which was in turn tested for a direct association with hospital readmissions. We also tested whether hospital access as measured by the proportion of people per hospital moderates the relationship between opioid use and hospital readmissions. We found significant associations between Negative Economic Factors and Opioid Use, between Environmental Factors and Opioid Use, and between Opioid Use and Hospital Readmissions. We found that Hospital Access positively moderates the relationship between Opioid Use and Readmissions. A priori assumptions about factors that influence hospital readmissions must extend beyond just individualistic factors and must incorporate a holistic approach that also considers the impact of macrolevel environmental factors.

Health Information Seeking Behavior on Social Networking Sites and Self-Treatment: Pilot Survey Study.

Background: Social networking site use and social network-based health information seeking behavior have proliferated to the point that the lines between seeking health information from credible social network-based sources and the decision to seek medical care or attempt to treat oneself have become blurred. Objective: We contribute to emerging research on health information seeking behavior by investigating demographic factors, social media use for health information seeking purposes, and the relationship between health information seeking and occurrences of self-treatment. Methods: Data were collected from an online survey in which participants were asked to describe sociodemographic factors about themselves, social media use patterns, perceptions about their motivations for health information seeking on social media platforms, and whether or not they attempted self-treatment after their social media-related health information seeking. We conducted a binomial logistic regression with self-treatment as a dichotomous categorical dependent variable. Results: Results indicate that significant predictors of self-treatment based on information obtained from social networking sites include race, exercise frequency, and degree of trust in the health-related information received. Conclusions: With an understanding of how sociodemographic factors might influence the decision to self-treat based on information obtained from social networking sites, health care providers can assist patients by educating them on credible social network-based sources of health information and discussing the importance of seeking medical advice from a health care provider.