This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
The future of health care delivery is becoming more citizen centered, as today’s user is more active, better informed, and more demanding. Worldwide governments are promoting online health services, such as electronic health record (EHR) patient portals and, as a result, the deployment and use of these services. Overall, this makes the adoption of patient-accessible EHR portals an important field to study and understand.
The aim of this study is to understand the factors that drive individuals to adopt EHR portals.
We applied a new adoption model using, as a starting point, Ventkatesh's Unified Theory of Acceptance and Use of Technology in a consumer context (UTAUT2) by integrating a new construct specific to health care, a new moderator, and new relationships. To test the research model, we used the partial least squares (PLS) causal modelling approach. An online questionnaire was administrated. We collected 360 valid responses.
The statistically significant drivers of behavioral intention are performance expectancy (beta=.200;
Our research helps to understand the desired technology characteristics of EHR portals. By testing an information technology acceptance model, we are able to determine what is more valued by patients when it comes to deciding whether to adopt EHR portals or not. The inclusion of specific constructs and relationships related to the health care consumer area also had a significant impact on understanding the adoption of EHR portals.
Our study focuses on a specific type of eHealth technology, the patient-accessible electronic health record (EHR) portals [
EHR portals may help patients carry out self-management activities, thereby making the use of the health care system more effective and sustainable, not only from the patient care standpoint, but also from a financial perspective due to rising health care costs and budgets in many countries [
This concept of a national-level patient portal progressed into a trans-European initiative, the European Patients Smart Open Services (epSOS). epSOS concentrates on developing a practical eHealth framework, and an information and communication technology (ICT) infrastructure that enables secure access to patient health information among different European health care systems [
The aim of this study is to identify a set of determinants in the adoption of electronic health record portals by health care consumers. In our study, we examine these determinants in the field of eHealth technology use and acceptance by health care consumers. We then propose a new research model based on Venkatesh's Unified Theory of Acceptance and Use of Technology in a consumer context (UTAUT2) by integrating a new construct from the health care area, self-perception (SP), and a new moderator, chronic disability (CD) [
In this paper, we first review the literature concerning information technology (IT) adoption models regarding consumer health care. We then present a research model to analyze EHR portals for the health care consumer. Finally, we discuss the issue and present conclusions.
There have been several theoretical models developed from theories in psychology, sociology, and consumer behavior employed to explain technology acceptance and use [
When studying eHealth and health care adoption by health care professionals, the most common adoption models used are the technology acceptance model (TAM) [
Ideally, we need a model tailored to the consumer use context, and in this specific field, UTAUT2 was developed with this goal, obtaining very good results [
Unified Theory of Acceptance and use of Technology (UTAUT) model adapted from Venkatesh et al [
Unified Theory of Acceptance and use of Technology in a consumer context (UTAUT2) model adapted from Venkatesh et al [
eHealth adoption models.
Theory | Dependent variable | Findings | Reference |
TAMa, motivational model (MM), integrated model (IM) | eHealth behavioral intention | Users’ perceived ease of use (PEOU), users’ perceived technology usefulness (PU), intrinsic motivation (IM), and extrinsic motivation (MM) have a significant positive influence on behavioral Intention. |
[ |
Elaboration likelihood model (ELM), concern for information privacy (CFIP) | EHRb behavioral intention | Positively framed arguments and issue involvement generate more favorable attitudes toward EHR behavioral intention. |
[ |
TAM (qualitative study) | eHealth services behavioral Intention | PU seemed to be important. |
[ |
TAM, plus several other constructs | Internet use behavior as a source of information | PU, importance given to written media in searches for health information, concern for personal health, importance given to the opinions of physicians and other health professionals, and the trust placed in the information available are the best predictors to use behavior. | [ |
Personal empowerment | Internet use behavior as a source of information | There are three types of attitudes encouraging Internet use to seek health information: professional, consumer, and community logic. | [ |
Extended TAM in health information technology (HIT) | HIT behavioral intention | PU, PEOU, and perceived threat significantly impacted health consumers’ behavioral intention. | [ |
aTAM: technology acceptance model.
bEHR: electronic health record.
UTAUT2 was developed as an adoption model providing the general factors of IT adoption in consumer use. However, according to Venkatesh et al [
Published studies suggest that patients with chronic illness, severe illness, or disability are more likely to use eHealth technologies if they have the resources and support available [
The research model. Notes: 1. Moderated by age or gender; 2. Moderated by age; 3. Moderated by chronic disability on use.
Performance expectancy is defined as the degree to which using a technology will provide benefits to consumers in carrying out certain activities [
Effort expectancy is the degree of ease related to consumers’ use of technology [
Social influence is the extent to which consumers perceive that others who are important to them (eg, friends and family) believe they should use a particular technology [
The construct, facilitating conditions, is defined as the individual perception of the support available for using a technology activity [
Chronic disability is an incapacitating situation (eg, chronic illness) that affects a patient permanently or for long-term periods. Our literature review reveals that patients with chronic illness or disability are more likely to use eHealth technologies if they have the resources and support available (ie, facilitating conditions) [
Hedonic motivation is defined as intrinsic motivation (eg, enjoyment) and has been included as a key predictor in much of the reported consumer behavior research [
Price value in a consumer use environment is also a relevant factor as, unlike workplace technologies, consumers must bear the costs related with the purchase of devices and services [
Habit can be defined as the extent to which people tend to execute behaviors automatically because of learning [
Behind the concept, self-perception, is the health belief model. The model assumes that subjective health considerations determine whether people perform a health-related action, such as consulting their physician [
Studies about patients that look for information online seem to confirm the concept of the health belief model; the results show that a larger proportion of respondents who described their health as poor indicated that they looked for health-related information online “often” compared with those who described their health as fair or better [
The role of intention as a predictor of usage is critical and has been well established not only in IS in general, but also in health care and eHealth, with the literature suggesting that the driver of using specific eHealth platforms is preceded by the intention to use them [
All of the items were adopted from Venkatesh et al [
The scales’ items were measured on a 7-point Likert scale, ranging from
Before the respondents could see any of the questions, an introduction was made explaining the concept of EHR portals (see
A pilot survey was conducted to validate the questions and the scale of the survey. From the pilot survey, we had 30 responses demonstrating that all of the items were reliable and valid. The data from the pilot survey were not included in the main survey.
According to the literature, the technology that we are studying (EHR portals) is being used by less than 7% of the total number of health care consumers or patients [
The survey, via hyperlink, was sent by email in October 2013 to a total of 1618 people at three institutions that provide educational services, from which we obtained 350 responses. NOVA Information Management School (IMS) approved and verified the ethical compliance of the questionnaire before its use. All participants were informed by email about the study purpose, confidentiality protection, and the anonymity of the information collected. A reminder was sent 2 weeks after the first email, only to those who had not responded to the first email, in order to improve the response rate. Following the reminder, we had a total of 465 respondents out of 1618 (28.74% response rate). After removing the invalid responses, the final sample consisted of 360 respondents. A questionnaire was considered invalid if not all questions were answered. According to our statistical modelling method, we cannot use incomplete questionnaires [
To test the research model, we used the partial least squares (PLS) method, which is a causal modelling approach that represents a variance-based technique of path modelling [
Our sample characteristics are shown in
The literature mentions that users of EHR portals are younger than the population average and have significantly higher education [
Use was measured on a scale that ranges from
Sample characteristics (n=360).
Variable and category | Frequency, n (%) | |
|
|
|
|
18-20 | 69 (19.2) |
|
21-24 | 75 (20.8) |
|
25-30 | 76 (21.1) |
|
31-40 | 89 (24.7) |
|
>40 | 51 (14.2) |
|
|
|
|
Male | 142 (39.4) |
|
Female | 218 (60.6) |
|
|
|
|
No | 308 (85.6) |
Yes | 52 (14.4) | |
|
|
|
|
Undergraduate | 132 (36.7) |
Bachelor’s degree | 87 (24.2) | |
|
Postgraduate | 70 (19.4) |
Master’s degree or more | 71 (19.7) |
Types of usage patterns of electronic health record (EHR) portals. UB: use behavior; UB1: management of personal information and communication with health providers; UB2: medical appointment schedule; UB3: check their own EHR; UB4: request for medical prescription renewals.
The results of the measurement model are shown in
Cronbach alpha, composite reliability, and average variance extracted.
Construct | Cronbach alpha | Composite reliability coefficient (CR) | Average variance extracted (AVE) |
Performance expectancy | .90 | .94 | .83 |
Effort expectancy | .91 | .94 | .79 |
Social influence | .98 | .98 | .96 |
Facilitating conditions | .80 | .87 | .63 |
Hedonic motivation | .93 | .96 | .88 |
Price value | .93 | .96 | .88 |
Habit | .74 | .85 | .66 |
Self-perception | .67 | .81 | .52 |
Behavior intention | .90 | .94 | .83 |
In order to have good indicator reliability, it is desired that the latent variable explain more than half of the indicators’ variances. The correlation between the constructs and their indicators should ideally be greater than .70 (√.50 ≈.70) [
In order to assess the convergent validity, we used average variance extracted (AVE). The AVE should be greater than .50, so that the latent variable explains, on average, more than 50% of its own indicators [
Correlationsa and square root of average variance extractedb.
PEc | EEd | SIe | FCf | HMg | PVh | HTi | SPj | BIk | UBl | Age | Gender | CDm | |
PE | .91 | ||||||||||||
EE | .47 | .89 |
|
|
|
|
|
|
|
|
|
|
|
SI | .31 | .24 | .98 |
|
|
|
|
|
|
|
|
|
|
FC | .25 | .57 | .23 | .79 |
|
|
|
|
|
|
|
|
|
HM | .47 | .44 | .31 | .32 | .94 |
|
|
|
|
|
|
|
|
PV | .42 | .33 | .34 | .26 | .42 | .94 |
|
|
|
|
|
|
|
HT | .43 | .29 | .55 | .26 | .48 | .46 | .81 |
|
|
|
|
|
|
SP | .04 | -.08 | .15 | -.06 | .08 | .08 | .16 | .72 |
|
|
|
|
|
BI | .50 | .43 | .43 | .29 | .44 | .35 | .61 | .17 | .91 |
|
|
|
|
UB | .23 | .18 | .39 | .24 | .17 | .23 | .41 | .10 | .44 | N/An |
|
|
|
Age | -.01 | -.04 | .13 | -.03 | -.01 | .08 | .09 | .08 | .08 | .20 | N/A |
|
|
Gender | -.02 | -.04 | .05 | 0 | -.08 | .05 | 0 | .05 | -.03 | 0 | .11 | N/A |
|
CD | -.08 | -.10 | .02 | -.08 | -.06 | -.02 | .03 | .24 | .01 | .13 | .18 | .09 | N/A |
aOff-diagonal elements are correlations.
bDiagonal elements are square roots of average variance extracted.
cPE: performance expectancy.
dEE: effort expectancy.
eSI: social influence.
fFC: facilitating conditions.
gHM: hedonic motivation.
hPV: price value.
iHT: habit.
jSP: self-perception.
kBI: behavioral intention.
lUB: use behavior.
mCD: chronic disability.
nN/A: not applicable, because they are not reflective constructs.
Use, which was modelled using four formative indicators, is evaluated by specific quality criteria related to formative indicators. As seen in
Formative indicators’ quality criteria.
Indicators | VIFa | Weights |
|
Outer loadings |
|
UB1b | 2.609 | .861 | 4.70c | .949 | 21.08c |
UB2 | 1.707 | .354 | 2.27d | .746 | 8.41c |
UB3 | 3.237 | .127 | 0.57 | .741 | 8.46c |
UB4 | 2.472 | -.329 | 1.66 | .543 | 4.50c |
aVIF: variance inflation factor.
bUB: use behavior.
c
d
In sum, all assessments are satisfactory. This means that the constructs can be used to test the conceptual model.
The structural model path significance levels were estimated using a bootstrap with 5000 iterations of resampling to obtain the highest possible consistency in the results. The
Structural model results. Notes: 1. Moderated by age or gender; 2. Moderated by age; 3. Moderated by chronic disability on use; *
We also tested the mediating role of behavioral intention between the independent variables and use behavior (see
Summary of findings regarding hypotheses.
Dependent variables | Independent variables | Hypotheses (H) | Beta |
|
|
Behavioral intention |
|
49.7% | |||
|
PEa | H1 (supported) | .200 | 3.619l |
|
|
EEb | H2 (supported) | .185 | 2.907l |
|
|
SIc | H3 (not supported) | .081 | 1.544 |
|
|
FCd | H4 (a) (not supported) | .005 | 0.112 |
|
|
HMe | H5 (not supported) | .038 | 0.678 |
|
|
PVf | N/Ag | -.010 | 0.203 |
|
|
PV x age | H6 (not supported) | .026 | 0.563 |
|
|
HTh | N/A | .388 | 7.320l |
|
|
HT x age | H7 (a1) (not supported) | .033 | 0.584 |
|
|
HT x gender | H7 (a2) (not supported) | .010 | 0.183 |
|
|
SPi | H8 (supported) | .098 | 2.285m |
|
|
Age | N/A | .065 | 1.408 |
|
|
Gender | N/A | .052 | 0.454 |
|
|
Gender x age | N/A | -.087 | 0.078 |
|
|
CDj | N/A | -.002 | 0.049 |
|
Use behavior |
|
26.8% | |||
|
FC |
|
.090 | 1.755 |
|
|
FC x CD | H4 (b) (not supported) | .076 | 0.391 |
|
|
HT | N/A | .206 | 2.752l |
|
|
HT x age | H7 (b1) (not supported) | .060 | 0.621 |
|
|
HT x gender | H7 (b2) (not supported) | .066 | 0.704 |
|
|
BIk | H9 (supported) | .258 | 4.036l |
|
|
Age | N/A | .170 | 2.387m |
|
|
Gender | N/A | -.013 | 0.092 |
|
|
Gender x age | N/A | .005 | 0.031 |
|
|
CD | N/A | -.081 | 0.476 |
|
aPE: performance expectancy.
bEE: effort expectancy.
cSI: social influence.
dFC: facilitating conditions.
eHM: hedonic motivation.
fPV: price value.
gN/A: not applicable.
hHT: habit.
iSP: self-perception.
jCD: chronic disability.
kBI: behavioral intention.
l
m
Mediating role of behavior intention on independent variables.
Step 1 | Step 2 | VAFa | ||||
Paths | Beta | |
Paths | Beta | |
|
|
|
|
PEb→BIc | .200 | 3.673l |
|
|
|
|
EEd→BI | .188 | 2.844l |
|
|
|
|
SIe→BI | .082 | 1.616 |
|
|
|
|
FCf→BI | .007 | 0.161 |
|
|
|
|
HMg→BI | .036 | 0.659 |
|
|
|
|
PVh→BI | -.007 | 0.131 |
|
|
|
|
HTi→BI | .392 | 7.313l |
|
|
|
|
SPj→BI | .105 | 2.521m |
|
PE→UBk | .075 | 1.281 | PE→UB | .067 | 1.125 |
|
EE→UB | -.023 | 0.481 | EE→UB | -.026 | 0.451 |
|
SI→UB | .223 | 3.733l | SI→UB | .228 | 3.389l |
|
FC→UB | .124 | 2.609l | FC→UB | .132 | 2.578m |
|
HM→UB | -.107 | 1.617 | HM→UB | -.112 | 1.629 |
|
PV→UB | .012 | 0.192 | PV→UB | .019 | 0.312 |
|
HT→UB | .278 | 3.733l | HT→UB | .276 | 3.801l |
|
SP→UB | .065 | 1.122 | SP→UB | .050 | 0.869 |
|
|
|
|
BI→UB | .271 | 3.746l |
|
|
|
|
(FC→BI)×(BI→UB) | .003 | 0.256 |
|
|
|
|
(SI→ BI)×(BI→UB) | .021 | 1.390 |
|
|
|
|
(HT→BI)×(BI→UB) | .106 | 3.472l | .38 |
aVAF: variance accounted for.
bPE: performance expectancy.
cBI: behavioral intention.
dEE: effort expectancy.
eSI: social influence.
fFC: facilitating conditions.
gHM: hedonic motivation.
hPV: price value.
iHT: habit.
jSP: self-perception.
kUB: use behavior.
l
m
The results suggest that using our research model in a health-related area—EHR portal acceptance by health care consumers—yields good results, explaining 49.7% of the variance on behavioral intention and 26.8% of the variance in technology use [
Concerning our results, some of our hypotheses were supported and others not; both H1 and H2 were supported. In studies that have addressed similar problems, including those studying patient portals [
Our results were also not able to confirm that patients with chronic illness or disability are more likely to use EHR portals if they have the resources and support available, as stated in H4 (b). This stands at odds with findings reported in the literature [
Hedonic motivation also has no significant impact on behavioral intention (H5). Hedonic motivation is defined as intrinsic motivation (eg, enjoyment) for using EHR portals. Patients seem not to perceive the use of EHR portals as an enjoyment. This is probably because much of the use of portals is driven by the presence of a disease or a health problem, and the need for the portal is associated with that unfortunate fact—something that does not promote enjoyment [
The impact of habit in behavioral intention and use behavior was not moderated by age or gender; H7 (a1), H7 (a2), H7 (b1), and H7 (b2) were therefore not supported. However, the construct habit has a significant impact on both behavioral intention and use behavior, in line with findings from literature that mention habit as a predictor of behavioral intention and use behavior [
Overall, we were able to demonstrate that habit, a construct specific to consumer technology acceptance, and self-perception, which is related to the area of knowledge we are testing, are both very important in understanding the acceptance of EHR portals. Specific tailor-made models that incorporate specific changes related to the study’s topic may be an effective option for studying complex areas of knowledge, such as IT health care.
The findings of this study have valuable managerial implications for the conceptualization, design, and implementation of an EHR portal. We found that performance expectancy and effort expectancy have a significant impact on the adoption of EHR portals. Earlier studies using TAM identified these constructs as being relevant for the adoption of patient portals [
Managerial implications. EHR: electronic health record.
We acknowledge that this research is limited by the geographic location, as it pertains to only one country and to only a sample of educational institutions. According to the literature, the technology that we are studying—EHR portals—is being used by less than 7% of the total number of health care consumers or patients [
Regarding the model tested, the inclusion of a health-related construct with significant positive impact demonstrates that it is relevant and that its inclusion is warranted. It also reveals the value of adding specific constructs related to the area in which the technology is used to existing frameworks. For future studies, it may also be advantageous to include other constructs (eg, confidentiality) that are not specific to health care but which, according to the literature, may be influential in eHealth adoption [
EHR portal adoption is a new and growing field of study that is an important topic in government-level discussions in the European Union and the United States. In our study, we used a new model in which we identified key additional constructs and relationships based on the literature review that are specific to IT health care adoption and integrated them into UTAUT2. The research model was tested and was found to explain 49.7% of the variance in behavioral intention and 26.8% of the variance in EHR portal technology use. Of all the constructs tested, performance expectancy, effort expectancy, self-perception, and habit had the most significant effects on behavioral intention. Habit and behavioral intention had a significant effect on technology use. Two specific constructs—habit (consumer related) and self-perception (health care)—were very significant in explaining the adoption of EHR portals, showing how important it is to use specific adoption models that include constructs specific to the area. The impact of chronic disability as a moderator of facilitating conditions to explain use behavior was not supported in our study. Not only is the adoption of EHR portals still low, but most current users of these platforms use them only infrequently. We used the results obtained in this study to provide managerial insights that may increase the adoption and usage of EHR portals.
Questionnaire items.
Partial least squares (PLS) loadings and cross-loadings.
average variance extracted
behavioral intention
chronic disability
concern for information privacy
Centers for Medicare & Medicaid Services
composite reliability coefficient
effort expectancy
electronic health record
elaboration likelihood model
European Patients Smart Open Services
facilitating conditions
hypothesis 1
hypothesis 2
hypothesis 3
hypothesis 4 (a)
hypothesis 4 (b)
hypothesis 5
hypothesis 6
hypothesis 7 (a1)
hypothesis 7 (a2)
hypothesis 7 (b1)
hypothesis 7 (b2)
hypothesis 8
hypothesis 9
health belief model
health information technology
hedonic motivation
habit
information and communication technology
integrated model
Information Management School
information systems
information technology
motivational model
not applicable
nonsignificant
performance expectancy
perceived ease of use
partial least squares
perceived usefulness
price value
social influence
self-perception
technology acceptance model
use behavior
unified theory of acceptance and use of technology
unified theory of acceptance and use of technology in a consumer context
variance accounted for
variance inflation factor
None declared.