Abstract
Objective
The study sought to examine whether provider encouragement is associated with improvements in engaging patients with their healthcare processes using online portals.
Materials and Methods
Using the Health Information National Trends Survey 2017 (N = 2, 670), we conducted an exploratory factor analysis with varimax orthogonal rotation and derived 3 outcome variables on patient engagement: (1) information access score, (2) care convenience score, and (3) patient engagement score. Multivariable linear regression on each outcome variable was conducted with provider encouragement as the main predictor, controlling for patient demographics.
Results
Women (60%), white participants (69%), and those with a college degree (49%) were more likely to report receiving provider encouragement. Those who were encouraged to use patient portals scored higher on all 3 outcome measures compared with those who were not encouraged (B = 0 .80 vs B = 0.11 for information access, B = 1.13 vs B = 0.13 for care convenience, and B = 0.44 vs B = 0.05 for patient engagement; all P < .001). For every additional 100 patients receiving encouragement, 65 more information access tasks, 94 more care convenience tasks, and 40 more patient engagement tasks would be performed.
Discussion
Provider encouragement was most influential concerning care convenience tasks and least influential on complex decision-making tasks. This may be due to portal design and the content available to patients, which merit consideration in future studies.
Conclusions
Provider encouragement is associated with more patient engagement, as indicated by significantly higher utilization of patient portals for accessing information, participating in routine care processes, and making complex healthcare decisions.
Keywords: patient engagement, patient portals, provider encouragement
INTRODUCTION
Patient engagement is the cornerstone of the patient-centered healthcare paradigm.1,2 Recognizing this important domain, accountable care organizations and other value-based payment and delivery models are increasingly requiring proactive engagement between patients and their healthcare providers.1,3 Given the ubiquity of electronic media and specifically, electronic medical records, most patient engagement efforts are taking place via electronic means, including the provision of patient portals and online access to electronic health records (EHRs).4–7 These types of access to medical records have been found to address several of the Institute of Medicine’s quality domains, including effectiveness, patient-centeredness, and efficiency, by facilitating patient engagement.8 This trend is further facilitated by “Meaningful Use” incentives to medical providers, which requires that recipients provide their patients with electronic access to their health records.9 As information resources and policies promoting the use of health information technology proliferate, it is imperative to understand the factors that might prompt use of such resources by patients to improve their care processes, health outcomes, and overall engagement with their healthcare.
Patient online medical record systems or patient portals (hereafter referred to simply as portals) in their most rudimentary form provide patients access to their medical information including recent doctor visits, discharge summaries, medications, immunizations, allergies, and laboratory test results. More advanced portals enable patients to request prescription refills, schedule nonurgent appointments, and exchange secure messaging with their providers.9 Use of portals is associated with reductions in no-show rates, especially among disadvantaged groups including the uninsured and those covered by Medicaid.10 Interactive portals are found to have a better impact on patient health outcomes and disease management.11 However, the use of portals varies widely among various population subgroups and emerging research indicates that engaging patients via portals is influenced by both patient and provider factors.12–15 While patient attitudes are generally positive toward such portals, there are substantial differences in utilization by race, ethnicity, and literacy levels.16,17 A variety of patient factors such as having complex healthcare needs, having higher health literacy and education levels, and being non-Hispanic white are associated with a higher likelihood of adopting the use of online portals.14,17–20 Studies also show that portal use is higher among those with a higher disease burden due to multiple chronic conditions or among those with access to specific information such as laboratory test results or clinical summaries.14,21,22 A few studies indicate that provider acceptance and promotion may have the potential to encourage overall use of portals.14,15 While there is considerable heterogeneity in portal characteristics, most of the current literature includes studies conducted in integrated health systems or in single-provider settings that provide an in-depth picture of the use of specific portals in their patient populations. Building on these studies, we examine use of portals in a nationally representative sample that presumably includes patient population with access to a variety of portals. Several previous studies have examined 1 or more types of portal use such as viewing laboratory test results, refilling medications, or securely messaging providers.14,20,23 In the present study, we build on this literature by grouping healthcare tasks performed by patients by their levels of complexity to examine their utilization of such portals. Using a nationally representative survey sample, we examine whether encouragement from providers would improve online patient engagement with various healthcare processes including information access and decision-making tasks.
OBJECTIVE
Use of patient portals can be classified into different levels of patient engagement as defined by the Health Information and Management Systems Society framework.24 The first level of engagement is to provide patients electronic access to their medical records so that they can easily view details of their visits to healthcare providers including their medical history, problem lists, diagnosis, laboratory test results, pharmacy orders, discharge notes, and visit summaries. The second level of patient engagement is to make it easier for patients to perform routine healthcare tasks such as making appointments and refilling medications via electronic access to portals. The third level of patient engagement includes engaging patients actively in making decisions and taking action based on the information provided to them. This level of engagement includes 2-way interactions in which patients utilize information to engage with providers and make well-informed decisions.
Our study objective was to examine the role of provider encouragement in promoting patient engagement using online portals based on the model described previously. Specifically, we hypothesized that encouragement from providers would improve the likelihood of patients participating actively via online portals at all 3 levels of engagement, controlling for patient demographics.
MATERIALS AND METHODS
Data source
The study data were from the National Cancer Institute’s Health Information National Trends Survey–5, Cycle 1, collected in 2017. The Health Information National Trends Survey–5 is a nationally representative survey that includes U.S. adults 18 years of age and older, in civilian, noninstitutionalized settings. This survey collects data on need, access, and use of information related to health and healthcare, and related behaviors, perceptions, and knowledge. More details on the survey and survey methodology can be found elsewhere.25 Specifically, the survey asks a set of 10 questions on how patients use their portals (Figure 1). Over 80% of the survey participants responded to at least some of those 10 questions and thus were included in the analysis (n = 2670).
Figure 1.
Number of patients who utilized portals by task and its distribution by provider encouragement.34
Variables
Outcome variables: exploratory factor analysis
The survey included 10 dichotomous questions (Figure 1) on how patients use their portals. As the research objective is to understand the levels of patient engagement using portals, we utilized these 10 questions to construct our outcome variables. Analysis of missing data among these 10 questions showed that the missingness did not exceed 0.9%. Little’s missing completely at random test was nonsignificant (χ297 = 94.10, P = .56), indicating that the missingness was completely at random.26,27 Based on this result, we chose to utilize listwise deletion of missing values. We performed an exploratory factor analysis (EFA) with varimax orthogonal rotation to identify common themes among these 10 different tasks. The decision on the number of factors to rotate was based on (1) the number of factors with positive eigenvalues and (2) a visual inspection of the scree plot. We conducted several preliminary tests to ensure the adequacy of sample size for the EFA. Bartlett’s test for sphericity was significant (χ245 = 1451.54, P < .001) and the Kaiser-Meyer-Olkin measure of sampling adequacy was 0.82, both of which indicated the appropriateness of the current sample for an EFA. The rotated factors resulted in 3 factors (Table 1). Factor 3 (Table 1) had loadings on whether the patients look up test results and whether they monitor health online. These questions indicated whether patients access information online. Hence, we grouped these questions as “Information Access Score.” While the question on monitoring health was loaded almost equally on the first and the third factors (Table 1), we decided to group it with the question on viewing results, as those questions were closely aligned conceptually. Factor 2 (Table 1) included questions on whether patients can securely message their healthcare provider, make appointments, request refill of medications, and fill out forms or paperwork related to their healthcare using portals. These questions indicated whether patients utilize portals for taking care of their healthcare process needs in a convenient manner and, hence, we referred this group as the care convenience score. Factor 1 (Table 1) included questions on whether patients download health information to other devices, make healthcare decisions, add information to their own records, and request correction of inaccurate information. These questions indicated the highest level of patient engagement by being proactive about information access, accuracy, and completeness, as well using information for decision-making tasks, and, hence, we grouped these variables as the patient engagement score. We added the individual variable values (0/1) within each factor grouping to construct these 3 outcome variables. Thus, information access score ranged from 0 to 2, care convenience score ranged from 0 to 4, and patient engagement score ranged from 0 to 4.
Table 1.
Exploratory factor analysis results (n = 955)
| Factors and Factor Loadings |
|||
|---|---|---|---|
| Variable | Factor 1 | Factor 2 | Factor 3 |
| View results | 0.32 | ||
| Monitor health | 0.38 | 0.31 | |
| Make appointments | 0.57 | ||
| Refill medications | 0.49 | ||
| Complete paperwork | 0.38 | ||
| Message providers | 0.57 | ||
| Download health records | 0.43 | ||
| Make healthcare decisions | 0.56 | ||
| Add info to records | 0.53 | ||
| Request corrections | 0.40 | ||
| Eigenvalue | 2.28 | 0.44 | 0.13 |
Predictor and control variables
Our main predictor variable was a dichotomous indicator based on the question “Have your HCP/doctor/nurse/office staff ever encouraged you to use an online medical record?” We included age, sex, education level, income level, race and ethnicity, self-reported health status, and health insurance type as control variables.
Descriptive analysis and multivariable regression analysis
We first conducted descriptive analysis for the data sample. We ran Pearson’s chi-square tests for categorical variables by whether the providers encouraged them to use portals. We examined the associations between medical care provider’s encouragement for portal use and utilization of portals by conducting multivariable linear regression analyses.
Analysis of missing data showed that the missingness did not exceed 1% for any variable except for race and ethnicity (5%). We tested the assumption of covariance-dependent missingness, which is an extension of Little’s missing completely at random test,26,27 with our outcome and predictor variables. The test was nonsignificant (χ272 = 70.75, P = .52), indicating that the missingness was completely at random with the covariates.26 Based on this result, we used listwise deletion of missing values to conduct complete case hierarchical regressions. First, we conducted multivariable linear regressions with provider encouragement as the predictor variable. Next, we conducted the same regressions with provider encouragement as the main predictor variable and other demographic variables as control variables. All analyses used jackknife replicate weights to adjust for survey characteristics. All analyses were conducted using Stata 14.28
Sensitivity analyses
We conducted the following 3 sensitivity analyses for each of the multivariable regressions.
As chronically ill patients are more inclined to engage with their healthcare,14 they are more likely to utilize all the available tools at their disposal. We calculated the number of chronic conditions for each patient, including diabetes, high blood pressure, heart disease, lung disease, arthritis, depression, or cancer. We then reran the same regressions as described previously by including the number of chronic conditions as an additional control variable.
Some patients have access via portals to their medical records maintained by their providers. Such patients with “tethered” portals are likely to have access to their medical history, visit summaries, clinical notes, laboratory test results, and other relevant information; hence, such patients are more likely to use their portals.14 While the survey did not ask questions on the “tethered” nature of portals, it included questions on whether the respondent had online access to all or some of their laboratory test results, problem lists, allergies, medications, immunizations, visit summaries, and clinical notes. Using these variables, we constructed a measure of the extent of online access and reran the regressions by adding the extent of online access as a control variable.
People who have a higher level of trust in online information accuracy and security are more likely to utilize portals.12,29 Hence, additionally, we reran the regressions by including a variable based on the survey question “In general, how much would you trust information about health or medical topics from the internet?” with answers choices including “not at all,” “a little,” “some,” and “a lot.”
RESULTS
Descriptive analysis
Our analytical sample included 2670 patients who answered all or some of the questions related to portal use. Figure 1 shows the number of patients who answered “yes” for each of the questions included in this analysis and the distribution by provider encouragement, adjusted with sample weights. The comparison indicates a marked difference between the 2 groups, with higher utilization among patients encouraged by their providers. Over 80% of patients who utilized each task were encouraged by their providers, while patients without provider encouragement accounted for <15% of those who utilized the tasks.
The average information access score was 0.38 (SE = 0.02; range = 0-2); average care convenience score was 0.53 (SE =0.03; range =0-4), and average patient engagement score was 0.21 (SE =0.02; range =0-4). However, there was a marked difference in these scores between those who were encouraged by their providers to use portals and those who were not. For information access, the average score for those encouraged was 0.80 compared with 0.11 for those who were not encouraged (P < .001); for care convenience, average scores were 1.13 vs 0.13 (P < .001); and for the most advanced use of portals, patient engagement, the scores were 0.44 versus 0.05 (P < .001). Figure 2 shows the distribution of task utilization within each score by provider encouragement. More than 50% of the population were not encouraged by their providers to use portals and did not perform any of the tasks within each of the scores. Those encouraged by providers were more likely to utilize at least 1 task within each score.
Figure 2.
Distribution of number of tasks utilized (percentage of patients by provider encouragement).34
As seen in Table 2, 41% of respondents reported receiving provider encouragement to use portals. The analytical sample included 49% male and 51% female respondents. A third of individuals were in the age range of 50-64 years (31%) followed by 35-49 years (28%). Women (60%) were more likely to report receiving encouragement from providers to use portals than were men (41%) (P < .001). There was also a notable education and income gradient in which individuals with higher education and higher income levels were more likely to report receiving provider encouragement (P < .001). For example, 49% of those individuals with a college degree or more reported encouragement, compared with 3% of those with less than a high school education. Individuals with private insurance were also more likely to report encouragement than were those with public or no insurance (78% vs 17% and 5%, respectively; P < .001). White patients received more encouragement (69%) compared with Black (10%) or Hispanic (13%) patients.
Table 2.
Sample characteristics: demographic variables
| Provider Encouraged Use of Portals (n = 1437, 59%) | Provider Did Not Encourage Use of Portals (n = 1233, 41%) | Total (N = 2670) | P Value | |
|---|---|---|---|---|
| Age group | .33 | |||
| 18-34 y | 0.2 (143) | 0.26 (183) | 0.24 (326) | |
| (0.02) | (0.03) | (0.02) | ||
| 35-49 y | 0.31 (281) | 0.26 (266) | 0.28 (547) | |
| (0.02) | (0.03) | (0.02) | ||
| 50-64 y | 0.32 (428) | 0.30 (459) | 0.31 (887) | |
| (0.02) | (0.02) | (0.01) | ||
| 65-74 y | 0.11 (247) | 0.10 (288) | 0.11 (535) | |
| (0.01) | (0.01) | (0.00) | ||
| 75+ y | 0.07 (99) | 0.08 (191) | 0.07 (290) | |
| (0.01) | (0.01) | (0.00) | ||
| Sex | <.001 | |||
| Male | 0.41 (431) | 0.55 (631) | 0.49 (1062) | |
| (0.02) | (0.01) | (0.01) | ||
| Female | 0.60 (782) | 0.45 (786) | 0.51 (1568) | |
| (0.02) | (0.01) | (0.01) | ||
| Education | <.001 | |||
| Less than high school | 0.03 (28) | 0.09 (114) | 0.07 (142) | |
| (0.01) | (0.02) | (0.01) | ||
| High school graduate | 0.16 (148) | 0.26 (321) | 0.22 (469) | |
| (0.02) | (0.01) | (0.01) | ||
| Some college | 0.32 (345) | 0.34 (422) | 0.33 (767) | |
| (0.02) | (0.02) | (0.01) | ||
| College graduate or more | 0.49 (689) | 0.3 (541) | 0.38 (1230) | |
| (0.02) | (0.01) | (0.01) | ||
| Income | <.001 | |||
| $0-$9 999 | 0.04 (39) | 0.07 (107) | 0.06 (146) | |
| (0.01) | (0.01) | (0.01) | ||
| $10 000-$14 999 | 0.02 (37) | 0.06 (105) | 0.04 (142) | |
| (0.01) | (0.01) | (0.01) | ||
| $15 000-$19 999 | 0.04 (45) | 0.07 (81) | 0.06 (126) | |
| (0.01) | (0.01) | (0.01) | ||
| $20 000-$34 999 | 0.09 (128) | 0.14 (231) | 0.12 (359) | |
| (0.01) | (0.02) | (0.01) | ||
| $35 000-$49 999 | 0.12 (148) | 0.16 (207) | 0.15 (355) | |
| (0.01) | (0.02) | (0.01) | ||
| $50 000-$74 999 | 0.2 (239) | 0.2 (256) | 0.2 (495) | |
| (0.02) | (0.02) | (0.01) | ||
| $75 000-$99 999 | 0.16 (191) | 0.11 (161) | 0.13 (352) | |
| (0.02) | (0.01) | (0.01) | ||
| $100 000-$199 999 | 0.25 (302) | 0.14 (197) | 0.19 (499) | |
| (0.02) | (0.02) | (0.01) | ||
| $200 000 or more | 0.09 (97) | 0.05 (73) | 0.06 (170) | |
| (0.01) | (0.01) | (0.01) | ||
| Race/ethnicity | .19 | |||
| Non-Hispanic white | 0.69 (796) | 0.65 (815) | 0.66 (1611) | |
| (0.01) | (0.01) | (0.01) | ||
| Non-Hispanic black | 0.10 (145) | 0.10 (163) | 0.10 (308) | |
| (0.01) | (0.01) | (0.01) | ||
| Hispanic | 0.13 (123) | 0.18 (216) | 0.16 (339) | |
| (0.01) | (0.01) | (0.01) | ||
| Non-Hispanic Asian or Pacific Islander | 0.06 (55) | 0.05 (53) | 0.05 (108) | |
| (0.01) | (0.01) | (0.00) | ||
| Non-Hispanic other | 0.02 (40) | 0.03 (51) | 0.03 (91) | |
| (0.01) | (0.01) | (0.00) | ||
| General health status | .07 | |||
| Poor | 0.01 (17) | 0.03 (43) | 0.02 (60) | |
| (0.01) | (0.01) | (0.01) | ||
| Fair | 0.13 (154) | 0.14 (223) | 0.14 (377) | |
| (0.02) | (0.01) | (0.01) | ||
| Good | 0.31 (391) | 0.36 (506) | 0.34 (897) | |
| (0.02) | (0.03) | (0.02) | ||
| Very good | 0.43 (516) | 0.35 (496) | 0.38 (1012) | |
| (0.02) | (0.02) | (0.02) | ||
| Excellent | 0.11 (147) | 0.12 (151) | 0.12 (298) | |
| (0.01) | (0.02) | (0.01) | ||
| Insurance status | <.001 | |||
| Uninsured | 0.05 (38) | 0.12 (121) | 0.09 (159) | |
| (0.02) | (0.01) | (0.01) | ||
| Public | 0.17 (242) | 0.27 (393) | 0.23 (635) | |
| (0.02) | (0.02) | (0.02) | ||
| Private | 0.78 (950) | 0.61 (911) | 0.68 (1861) | |
| (0.02) | (0.02) | (0.02) |
Values are % (observed) (SE).
Multivariable regression analysis
Results from the 3 multivariate linear regression models consistently showed that provider encouragement to use portals positively and statistically significantly predicted higher utilization of those portals at all 3 levels of patient engagement, both with and without the inclusion of demographic variables (Table 3). Under each outcome variable, we show model 1, which included provider encouragement only, and model 2, which included provider encouragement as well as other demographic variables.
Table 3.
Regression analysis: provider encouragement and use of portals
| Information Access Score |
Care Convenience Score |
Patient Engagement Score |
||||
|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 1 | Model 2 | Model 1 | Model 2 | |
| Provider encouraged portal use | 0.69c | 0.65c | 1.00c | 0.94c | 0.39c | 0.40c |
| (0.62 to 0.76) | (0.57 to 0.74) | (0.89 to 1.10) | (0.81 to 1.07) | (0.33 to 0.46) | (0.32 to 0.47) | |
| Age group (reference group: 18-34 y) | ||||||
| 35-49 y | 0.05 | 0.07 | –0.04 | |||
| (–0.09 to 0.18) | (–0.14 to 0.28) | (–0.18 to 0.09) | ||||
| 50-64 y | –0.02 | –0.01 | –0.12 | |||
| (–0.14 to 0.09) | (–0.26 to 0.24) | (–0.25 to 0.00) | ||||
| 65-74 y | –0.03 | –0.03 | –0.18b | |||
| (–0.15 to 0.10) | (–0.23 to 0.17) | (–0.31 to –0.05) | ||||
| 75+ y | –0.09 | –0.16 | –0.19b | |||
| (–0.24 to 0.07) | (–0.41 to 0.10) | (–0.32 to –0.06) | ||||
| Female | 0.02 | 0.01 | –0.06 | |||
| (–0.06 to 0.10) | (–0.14 to 0.15) | (–0.14 to 0.01) | ||||
| Education (reference group: less than high school) | ||||||
| High school | –0.08 | 0.16 | 0.03 | |||
| (–0.22 to 0.07) | (–0.04 to 0.37) | (–0.08 to 0.14) | ||||
| Some college | –0.05 | 0.16 | 0.03 | |||
| (–0.20 to 0.10) | (–0.02 to 0.33) | (–0.08 to 0.15) | ||||
| College or more | 0.01 | 0.28b | 0.08 | |||
| (–0.15 to 0.17) | (0.08 to 0.47) | (–0.04 to 0.20) | ||||
| Income (reference group: <$10 000) | ||||||
| $10 000-$14 999 | –0.03 | –0.08 | 0.05 | |||
| (–0.19 to 0.12) | (–0.31 to 0.15) | (–0.10 to 0.20) | ||||
| $15 000-$19 999 | –0.06 | –0.10 | 0.09 | |||
| (–0.24 to 0.12) | (–0.35 to 0.15) | (–0.10 to 0.27) | ||||
| $20 000-$34 999 | 0.01 | 0.02 | 0.05 | |||
| (–0.14 to 0.16) | (–0.27 to 0.30) | (–0.11 to 0.20) | ||||
| $35 000-$49 999 | 0.03 | 0.01 | 0.10 | |||
| (–0.13 to 0.18) | (–0.21 to 0.23) | (–0.09 to 0.30) | ||||
| $50 000-$74 999 | 0.03 | 0.05 | 0.05 | |||
| (–0.14 to 0.19) | (–0.20 to 0.30) | (–0.12 to 0.22) | ||||
| $75 000-$99 999 | –0.01 | 0.11 | 0.00 | |||
| (–0.20 to 0.17) | (–0.18 to 0.41) | (–0.16 to 0.17) | ||||
| $100 000-$199 000 | 0.11 | 0.20 | 0.07 | |||
| (–0.05 to 0.27) | (–0.06 to 0.45) | (–0.14 to 0.27) | ||||
| $200 000 or more | 0.14 | 0.20 | 0.00 | |||
| (–0.06 to 0.34) | (–0.17 to 0.58) | (–0.21 to 0.21) | ||||
| Race/ethnicity (reference group: non-Hispanic white) | ||||||
| Non-Hispanic black | 0.00 | –0.05 | 0.01 | |||
| (–0.13 to 0.13) | (–0.25 to 0.14) | (–0.13 to 0.15) | ||||
| Hispanic | –0.07 | –0.04 | –0.02 | |||
| (–0.18 to 0.05) | (–0.18 to 0.09) | (–0.14 to 0.11) | ||||
| Non-Hispanic Asian | 0.09 | 0.12 | 0.04 | |||
| (–0.10 to 0.29) | (–0.17 to 0.41) | (–0.18 to 0.25) | ||||
| Non-Hispanic other | 0.14 | 0.13 | –0.04 | |||
| (–0.06 to 0.33) | (–0.27 to 0.53) | (–0.21 to 0.14) | ||||
| Health status (reference group: poor) | ||||||
| Fair | –0.06 | 0.10 | –0.06 | |||
| (–0.24 to 0.13) | (–0.22 to 0.41) | (–0.39 to 0.26) | ||||
| Good | –0.09 | –0.04 | –0.17 | |||
| (–0.24 to 0.06) | (–0.30 to 0.23) | (–0.49 to 0.14) | ||||
| Very good | –0.04 | –0.05 | –0.19 | |||
| (–0.20 to 0.12) | (–0.33 to 0.23) | (–0.50 to 0.12) | ||||
| Excellent | –0.05 | –0.08 | –0.20 | |||
| (–0.22 to 0.12) | (–0.40 to 0.24) | (–0.52 to 0.13) | ||||
| Insurance type (reference group: uninsured) | ||||||
| Public | 0.05 | 0.08 | 0.06 | |||
| (–0.08 to 0.17) | (–0.12 to 0.28) | (–0.03 to 0.14) | ||||
| Private | 0.10 | 0.06 | 0.09a | |||
| (–0.02 to 0.21) | (–0.13 to 0.25) | (0.00 to 0.18) | ||||
| Constant | 0.10 | –0.15 | 0.14 | |||
| (–0.18 to 0.38) | (–0.61 to 0.32) | (–0.27 to 0.55) | ||||
| n | 2657 | 2354 | 2629 | 2339 | 2635 | 2341 |
| R2 | 0.26 | 0.29 | 0.19 | 0.22 | 0.09 | 0.13 |
| F value | 372.33 | 17.27 | 360.69 | 15.20 | 150.86 | 8.21 |
Values are: B (95% confidence interval).
P < .05.
P < .01.
P < .001.
In both models, for each outcome, provider encouragement was a significant and positive predictor. Inclusion of demographic variables improved the model fit (ΔR2 = 0.03 for the information access score and care convenience score models; ΔR2 = 0.04 for the patient engagement models) and slightly changed the magnitude of the coefficient on provider encouragement. The results remained substantively the same. Among all independent variables included in all 3 model 2 regressions, provider encouragement had the greatest impact in terms of magnitude or coefficient size (Table 3). Specifically, provider encouragement was associated with an increase of 0.65 (P < .001) for information access score, 0.94 (P < .001) for care convenience score, and 0.40 (P < .001) for patient engagement score, holding all else constant. In other words, these results indicated that for every 100 patients encouraged to use portals, 65 more information access tasks, 94 more care convenience tasks, and 40 more engagement tasks would be performed using portals.
Having a college degree or more education was positively associated with the care convenience score, compared with less than high school education (B = 0.28, P < .05). Those who were 65-74 years of age (B = –0.18, P < .05) and 75 years of age and older (B = –0.19, P < .05) were associated with a lower score for patient engagement, compared with those 18–34 years old. In contrast, having private insurance was positively associated with patient engagement score (B = 0.09, P < .05). Sex, race and ethnicity, or general health status did not predict any of the 3 outcome variables.
Sensitivity analysis
While having 3-4 chronic conditions was positively associated with all 3 outcomes (P < .05), including the number of chronic conditions in the regression models did not change the magnitude or the statistical significance of provider encouragement. Having 1 or more aspects of the medical information available online was significantly associated with all 3 outcomes of patient engagement (P < .05). Including this variable in the regression models slightly reduced the magnitude of provider encouragement on the outcomes, but did not change the results substantively. Whether a patient trusted the accuracy of online health information was not significantly associated with any of the 3 outcomes and did not change the regression results.
DISCUSSION
With increasing availability of online health record portals for patients, it is essential to understand how to promote their utilization to enhance patient engagement and improve care experience. To this end, our study has 2 important findings. First, our study indicates that among patients with access to online portals, a large percentage are not using them to conduct any healthcare-related tasks. Figure 2 shows that 71.5% of patients did not use portals for accessing information (ie, had a score of 0 for information access), 77.3% did not use them for making their care processes easier (ie, had a score of 0 for care convenience), and 87.4% did not use them for patient engagement tasks (ie, had a score of 0 for patient engagement). This is in line with past studies that indicate low levels of utilization of portals.30,31 Second, encouragement from healthcare providers has a strong and positive impact on patient engagement with healthcare, including accessing online information, utilizing information to make their care processes easier, and proactively engaging with providers and making healthcare decisions. This effect of encouragement is significant even after controlling for other factors that may influence use of portals such as age, race and ethnicity, education, income, and health status. While the impact of provider encouragement on all 3 levels of portal use is large and statistically significant, the effect is largest for care convenience score, which includes using online portals for routine healthcare tasks such as filling forms, ordering prescription refills, making appointments, or messaging providers. As these processes are similar to activities that are often conducted online such as buying airline tickets, completing application forms or emailing, patients may be more inclined to utilize online portals for similar tasks in the healthcare domain, especially after they learn about their availability from their providers. By comparison, it is relatively harder to motivate patients to make informed medical decisions using portals, a concept captured in the patient engagement score. This may be due to complexities involved in engaging patients to make decisions compared with the relatively easier tasks of accessing their information using portals. Patients may also be more comfortable in utilizing portals for routine tasks such as refilling prescriptions or making appointments than for the more complicated tasks of making healthcare decisions. However, despite differences in healthcare needs of patients with different health status or the variety of information available to patients, encouragement from providers remains a significant factor in patients using online portals to engage with their healthcare at all levels, indicating providers’ influence beyond the medical office.
It is notable that less than half of the patients (41% from Table 3) reported that they received encouragement from their providers to use online portals. It is possible there is a lack of interest or incentives among providers to encourage use because they typically do not receive direct compensation for interacting with patients via online portals. Nevertheless, there are nuances in who receives provider encouragement. Descriptive analysis showed that higher education and income levels are positively correlated with provider encouragement, suggesting a socioeconomic status gradient for such encouragement. In other words, patients who have lower income or education levels might be less likely to learn about using portals offered by their providers. Future studies need to explore the factors examining why certain patients such as women, those with higher education attainment, and those with higher income levels seem to receive more encouragement compared with other population groups that may benefit more from a provider’s encouragement. In addition, while we see a strong connection between provider encouragement and use of portals, the exact mechanism through which patients are motivated and take action is unclear, and warrants future research.
Further, it would be beneficial to examine the source and content of online records as well as the exact nature of online tasks conducted by patients to understand whether specific tasks can benefit more from provider encouragement.
Limitations
As our analyses examined associations using cross-sectional data, we cannot infer causality. Still, we find a strong correlation between provider encouragement and use of portals with several sensitivity checks on the models.
The survey does not differentiate between encouragement by physicians, nurses or office staff. Nor does the survey assess the extent or exact nature of support offered by providers. Encouragement by physicians and nurses may provide a seal of approval in terms of information quality, which may lead to its use for complex tasks such as healthcare decision making. However, patients who are not very familiar with digital technology may also benefit from helpful staff who may provide logistical support such as help with setting up accounts and utilizing online patient portals.32,33
There is no distinction in the data between whether the portals are “tethered” to an EHR or not. “Tethered” records include data that come directly from the patient EHRs rather than partial or interpreted data available through untethered records that are managed by patients. It may be easier to access the “tethered” information to engage in healthcare as the information is readily available from the electronic record.14 Also, “tethered” records are more complete in terms of healthcare information and, hence, patients may feel more confidence in data quality and accuracy to use such data for complex tasks such as making healthcare decisions. On the other hand, complexity of information may be daunting to average users who may find it difficult to understand clinical language commonly used in such records. While our data are nationally representative and possibly include various types of portal users, knowing the type of information source and the content would be important in understanding use of portals in relation to provider encouragement.
CONCLUSION
Many benefits of patient engagement are undisputed and use of patient portals results in lower costs and higher utility for both patients and providers.9However, as seen by lower percentages of patients who engage in using these portals, we surmise that there remain challenges in engaging patients meaningfully in portal use. One of the challenges that we see from the data in our study is that only about 41% of patients report getting any encouragement to use online portals from their providers, which may be due to a number of reasons. Providers may lack an understanding of the utility of these portals, look at portals as a replacement for traditional aspects of the patient-provider relationships, or not have additional time needed to help patients set up their accounts or educate them in using patient portals. However, as indicated by our results, encouragement from providers might go a long way in improving use of portals at all levels of patient engagement. However, disparate access, limited understanding of the utility of such portals, and questionable quality can limit portal utilization. Future studies should examine the methods of provider encouragement to all patient groups including those with low education attainment, those with low income, and minorities, while taking into account the content, quality, and accessibility of information provided to patients.
AUTHOR CONTRIBUTIONS
Both authors were involved in study conceptualization, study design, data analysis and interpretation of results, and manuscript preparation and have approved the article.
ACKNOWLEDGMENTS
We are grateful to Dr Grace Reynolds-Fisher for her help in editing the article.
CONFLICT OF INTEREST STATEMENT
None declared.
REFERENCES
- 1. Carman KL, Dardess P, Maurer M, et al. Patient and family engagement: A framework for understanding the elements and developing interventions and policies. Health Aff 2013; 322: 223–31. [DOI] [PubMed] [Google Scholar]
- 2. Berwick DM, Nolan TW, Whittington J.. The triple aim: Care, health, and cost. Health Aff 2008; 273: 759–69. [DOI] [PubMed] [Google Scholar]
- 3. Hibbard JH, Greene J.. What the evidence shows about patient activation: better health outcomes and care experiences; fewer data on costs. Health Aff 2013; 322: 207–14. [DOI] [PubMed] [Google Scholar]
- 4. Lupton D. The digitally engaged patient: Self-monitoring and self-care in the digital health era. Soc Theory Health 2013; 113: 256–70. [Google Scholar]
- 5. Hassol A, Walker JM, Kidder D, et al. Patient experiences and attitudes about access to a patient electronic health care record and linked web messaging. J Am Med Inform Assoc 2004; 116: 505–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Mandl KD, Mandel JC, Kohane IS.. Driving innovation in health systems through an apps-based information economy. Cell Syst 2015; 11: 8–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hawn C. Take two aspirin and tweet me in the morning: how Twitter, Facebook, and other social media are reshaping health care. Health Aff 2009; 282: 361–8. [DOI] [PubMed] [Google Scholar]
- 8. Davis Giardina T, Menon S, Parrish DE, Sittig DF, Singh H.. Patient access to medical records and healthcare outcomes: a systematic review. J Am Med Inform Assoc 2014; 214: 737–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Kruse CS, Bolton K, Freriks G.. The effect of patient portals on quality outcomes and its implications to meaningful use: a systematic review. J Med Internet Res 2015; 172: e44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Horvath M, Levy J, L'Engle P, Carlson B, Ahmad A, Ferranti J.. Impact of health portal enrollment with email reminders on adherence to clinic appointments: a pilot study. J Med Internet Res 2011; 132: e41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Ammenwerth E, Schnell-Inderst P, Hoerbst A.. The impact of electronic patient portals on patient care: a systematic review of controlled trials. J Med Internet Res 2012; 146: e162.[published Online First: Epub Date]|. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kruse CS, Argueta DA, Lopez L, Nair A.. Patient and provider attitudes toward the use of patient portals for the management of chronic disease: a systematic review. J Med Internet Res 2015; 172: e40.[published Online First: Epub Date]|. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Miller DP Jr, Latulipe C, Melius KA, Quandt SA, Arcury TA.. Primary Care Providers' views of patient portals: Interview study of perceived benefits and consequences. J Med Internet Res 2016; 181: e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Irizarry T, Dabbs AD, Curran CR.. Patient portals and patient engagement: a state of the science review. J Med Internet Res 2015; 176: e148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lyles CR, Harris LT, Jordan L, et al. Patient race/ethnicity and shared medical record use among diabetes patients. Med Care 2012; 505: 434–40. [DOI] [PubMed] [Google Scholar]
- 16. Goldzweig CL, Orshansky G, Paige NM, et al. Electronic patient portals: Evidence on health outcomes, satisfaction, efficiency, and attitudes: A systematic review. Ann Intern Med 2013; 15910: 677–87.[published Online First: Epub Date]|. [DOI] [PubMed] [Google Scholar]
- 17. Tieu L, Schillinger D, Sarkar U, et al. Online patient websites for electronic health record access among vulnerable populations: portals to nowhere? J Am Med Inform Assoc 2017; 24 (e1): e47–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Peacock S, Reddy A, Leveille SG, et al. Patient portals and personal health information online: perception, access, and use by US adults. J Am Med Inform Assoc 2017; 24 (e1): e173–e77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Ancker JS, Hafeez B, Kaushal R.. Socioeconomic disparities in adoption of personal health records over time. Am J Manag Care 2016; 228: 539–40. [PMC free article] [PubMed] [Google Scholar]
- 20. Lyles CR, Sarkar U, Schillinger D, et al. Refilling medications through an online patient portal: consistent improvements in adherence across racial/ethnic groups. J Am Med Inform Assoc 2016; 23 (e1): e28–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Greenberg AJ, Falisi AL, Finney Rutten LJ, et al. Access to electronic personal health records among patients with multiple chronic conditions: a secondary data analysis. J Med Internet Res 2017; 196: e188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sarkar U, Lyles CR, Parker MM, et al. Use of the refill function through an online patient portal is associated with improved adherence to statins in an integrated health system. Med Care 2014; 523: 194–201.[published Online First: Epub Date]|. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Harris LT, Koepsell TD, Haneuse SJ, Martin DP, Ralston JD.. Glycemic control associated with secure patient-provider messaging within a shared electronic medical record. Diabetes Care 2013; 369: 2726–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. HIMSS Foundation. HIMSS Patient Engagement Framework. 2014. https://www.himss.org/himss-patient-engagement-framework. Accessed March 14, 2018.
- 25. National Cancer Institute. Health information national trends Survey 5 Cycle 1 Methodology report In: Bethesda, MD: U.S. Department of Health and Human Services; 2017. [Google Scholar]
- 26. Li C. Little's test of missing completely at random. Stata J 2013; 134: 795–809. [Google Scholar]
- 27. Little RJA. A test of missing completely at random for multivariate data with missing values. J Am Stat Assoc 1988; 83404: 1198–202.[published Online First: Epub Date]|. [Google Scholar]
- 28. Stata Statistical Software: Release 14 [program]. College Station, TX: StataCorp LP, 2014. [Google Scholar]
- 29. Lyles CR, Sarkar U, Ralston JD, et al. Patient-provider communication and trust in relation to use of an online patient portal among diabetes patients: the Diabetes and Aging Study. J Am Med Inform Assoc 2013; 206: 1128–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Turner AM, Osterhage K, Hartzler A, et al. Use of patient portals for personal health information management: The older adult perspective. AMIA Annu Symp Proc 2015; 2015: 1234–41. [PMC free article] [PubMed] [Google Scholar]
- 31. Clark SJ, Costello LE, Gebremariam A, Dombkowski KJ.. A national survey of parent perspectives on use of patient portals for their children's health care. Appl Clin Inform 2015; 61: 110–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Tieu L, Sarkar U, Schillinger D, et al. Barriers and facilitators to online portal use among patients and caregivers in a safety net health care system: a qualitative study. J Med Internet Res 2015; 1712: e275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Ancker JS, Barrón Y, Rockoff ML, et al. Use of an electronic patient portal among disadvantaged populations. J Gen Intern Med 2011; 2610: 1117–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Westat. Health Information National Trends Survey 5 (HINTS 5). Cycle 1 methodology report. July 2017. https://hints.cancer.gov/docs/methodologyreports/HINTS5_Cycle_1_Methodology_Rpt.pdf. Accessed March 15, 2018 .