Keywords
electronic medical record - technology acceptance model - physician - gender - clinical
specialty
Introduction
It has greatly consented that the use of information technology (IT) in the health
care sector offers great potential for improving the quality, efficiency, and effectiveness
of the provided services, personnel, above all reduces the organizational expenses.[1]
[2] Previous researchers even confirmed if hospitals did not adopt contemporary information
systems (IS), they could lose the trust of their patients.[3]
[4] Thus, hospital information systems (HIS) have gradually taken over traditional hospital
operation procedures.[4]
Electronic medical record (EMR) is regarded as the integration of several information
tools (such as emergency information, electronic prescription, decision support system,
digital imagery, and telemedicine) that might improve the uptake of evidence into
clinical decisions. Using such evidence in everyday clinical practices might validate
a safer and more effective health care system.[5]
[6] Previous studies recommended several benefits of EMR for patients[7]
[8] and one of the main benefits is to improve the quality of care resulting from accessing
the patients' essential health records from different health care providers, which
significantly can improve the coordination of the patients' care[7]
[8] and the efficiency of health care practices.[7]
[9]
To determine the benefits of the EMR system, user adoption plays a major role. The
existing literature indicated that many previous clinical system implementations had
failed due to lack of user adoption.[10]
[11] As the significant coordinator and provider of patient care, physicians' intention
to use EMR determines the overall success of its implementation.[12] So, the EMR acceptance by physicians is a primary condition to ensure that the expected
benefits will be materialized.[13] Although prior researches indicated that physicians would not consider a system
that interferes with the way they care for patients or places limitations on the way
they practice.[14] Thus, understanding the factors influencing physicians' intention to use EMR is
one of the critical elements in ensuring its optimal integration and ultimately measurable
benefits within the health care system and patients.
To better cope with physicians' IS usage concerns, it is also significant to consider
the gender and clinical specialty issues as previous researches explored that females
are less interested in using new technology than males, a “gender gap,”[15]
[16]
[17]
[18] as well as clinical differences, affect the physician's intention to use EMR.[19]
[20] But, the outcomes reported in these studies remain inconsistent and inconclusive.
Consequently, the present study includes gender and clinical specialty, for better
understanding and predicting physicians' intention to use EMR. Thus, the primary aim
of the current research is to explore the differences of gender and clinical specialty
in perceptions and relationships among the factors influencing physician's intention
to use EMR, and second to test the applicability and effectiveness of the extended
technology acceptance model (TAM) for evaluating physicians' intention to use EMR
in Taiwan. We are positive that the results of the present study can improve our current
understanding regarding physicians' intention to use EMR in the hospital, and it will
support both government and health care providers to determine the factors that may
contribute to physicians' intention to use EMR more clearly.
Literature Review
Electronic Medical Record in Taiwan
The Taiwanese government has endorsed the following six dimensions of EMR: sharing,
standards, infrastructure, applications, legislation, and security,[21] alongside has taken several measures to promote EMR comprehensively after the initial
startup had been successful in 2000 as a trial in a single hospital,[22] that is, the government has propagated the Electronic Signatures Act to legitimately
authorize the use of paperless signatures in 2001. In 2004, Article 69 of the Medical
Care Act was amended to specify that medical care institutions that record and store
medical records electronically are exempted from keeping a written copy.[23] This provision is the formal legal base of EMR. In 2005, the department of health
(DOH) introduced the “Regulations Governing the Production and Management of Electronic
Medical Records in Medical Care Institutions,” instructing that medical institutions
may give back traditional written signatures with e-signatures while turning out EMR.[23]
In 2009, the DOH relaxed restrictions on the regulation of the time stamp to allow
medical institutions to set up a trustworthy time-stamp management mechanism by themselves.
Subsequently, receiving a time stamp from the Health care Certification Authority
(HCA) of the DOH is no longer required which has significantly improved the delivery
speed and convenience of e-signatures.[23] By the end of 2011, 108 EMR reference templates for different types of medical records
(such as outpatient physician order entry records, admission notes, and children's
physical therapy assessment records) had been recommended on the web site.[24] As stated by the reference model, real-world interhospital EMR mechanisms can be
implemented.[23] Thus, these developments support to promote the EMR system nationwide.
Physicians' Intention to use EMR
Understanding an individual's acceptance or rejection of IT is considered one of the
most challenging concerns in IS research.[25]
[26] Successful implementation of an IS depends immensely on the degree of consideration
given to human concerns which have a certain effect on the process[12] and one of the critical human concerns is resistance to change. Klaus and Blanton[27] stated that the critical interference toward implementing successful IT projects
within the organization is employees' resistance to change.
From the IT perspective, different groups have different cognitive biases. Hu et al[28] reported that physicians differed significantly from other professional users in
general competence, intellectual, and cognitive capacity, as well as work arrangement
and nature. However, focusing on the potential adoption of IT by physicians is important
for successfully running medical institutions.[29]
[30]
Yi et al[31] stated that perceived improvement in performance from using IT intensely influenced
physicians' intention to use the system in the health care sector. Chau and Hu[29] explored that the significant role of perceived usefulness (PU) among physicians
in fostering their intention toward using a new technology might be centered on physicians'
utility-based perspective about using technology. In other words, physicians consider
using new HIT while they perceive it has anticipated utility and might be contributory
in their practices. According to Chang et al,[32] PU exerted the most significant impact on physicians' intention to use clinical
decision support systems (CDSS). Based on the study of Kijsanayotin et al,[33] PU was the main determinant of behavioral intention to use HIT. Therefore, as long
as health care professionals perceive the EMR system as an instrument to improve their
patient care performance, they will be more influenced to use the system.
The findings of Hoque[34] on mobile-health (mhealth) adoption based on TAM reported that perceived ease of
use (PEOU), subjective norm, and PU had a significant positive influence on the users'
intention to adopt mhealth services. Hu et al[28] studied the applicability of the TAM in explaining physicians' decision to adopt
telemedicine technology, and their findings supported that TAM was capable of providing
a reasonable interpretation of physicians' intention to use telemedicine technology.
Kim et al[35] explored physicians' attitudes toward EMR adoption based on TAM and the Unified
Theory of Acceptance and Use of Technology (UTAUT) model and their results found all
the variables (performance expectancy, effort expectancy, social influences, facilitating
conditions, and attitude) positively influenced the physician's intention to adopt
EMR. Al-Adwan and Berger[36] studied physicians' behavioral intention based on TAM and self-perceived variables
(perceived threat [PT], PU, PEOU, and social influences), and they found all the variables
had a positive effect on physicians' intention. Liu and Cheng[37] explored physician's intention to adopt mobile-EMR based on the dual-factor model
and they found PU, PEOU, and PT had the significant influence on physicians' behavioral
intention to adopt mobile EMR. The previous study by Price[38] found that the most significant factors influencing physicians' intention to use
EMR were PEOU, PU, and perceived patient-record privacy.
Overall, previous researches have shown some support to use TAM as a theoretical model
to examine physicians' intention to use EMR. Therefore, to better understanding of
the factors influencing physicians' attitude and intention to use EMR system in Taiwan,
a modified TAM with two antecedents, gender, clinical specialty and one context-specific
factor, and financial incentive was proposed and tested.
Gender
Investigating individual's intention to use technology is yet another area in which
gender differences have been overlooked.[30] However, the present study additionally identifies the influence of gender on the
relationship between determinants and users' behavioral intention to use. Podolny
and James[39] explored the demographic variable, called gender, had been proved to influence the
use of new technologies. Hing and Burt[40] found that the United States suffered a gender gap with regard to IT usage which
might discourage women from entering into IT as a profession. Although prior studies
investigated the gender differences in computer-related attitudes and its use, limited
researches had integrated gender as a personality trait factor in evaluating the physicians'
intention to use the new IS. Gillard et al[41] dealt directly with gender differences as they related to the resistance to the
EMR systems and, more generally, HIS in general. Sykes et al[16] stated that gender predicted the EMR system use. Moffat et al,[42] in their study, found that 41% of female general physicians (GP) were nonusers of
the EMR system compared with 28% of male GP. According to the findings of Schwirian
et al,[43] gender differences influenced physicians' attitude to HIT use. Menon et al[44] explored that several physicians were consenting to use HIT but differed by gender.
Although it was anticipated from previous literature that gender was a significant
factor in association with IT, Chan et al[45] did not find any difference in their study. Lai et al[46] also did not observe any influence of gender on the translation of intention to
the actual implementation of a system. According to the study by Loomis et al,[47] there were no statistically significant differences in gender between users and
nonusers of the EMR system. Djalali et al[48] stated that male primary care physicians (PCP) had shown higher levels of EMR adoption,
while the opposite was spot-on for the remainder.[49]
[50]
Therefore, we can conclude that the outcomes reported in the previous studies concerning
the gender differences in the context of physicians' intention for HIT adoption, especially
the EMR system adoption, remain inconsistent and inconclusive. Chu[51] stated that gender differences in the use of the technology must be cautiously examined,
instead of only representing differences. Thus, making out the gender differences
in the strength of the path coefficients could bring additional perception into conventional
theories regarding gender concerns.
Clinical Specialty
The significant differences have been reported among different medical disciplines
in terms of EMR adoption. The previous literature highlighted concerns that different
clinical disciplines might have a differing level of compatibility with the current
state of EMR.[20] Previous studies supported the need for medical specialty-specific systems to come
across the unique requirements of corresponding specialties.[31]
[32]
[33]
[34] These unique needs come out variations in a standard workflow, information collection
requirements, and clinical documentation methods along with disparities in standard
clinical volume, billing, compliance necessities, and specialty-specific terminology.[20]
The previous study recommended that the ratio of information review of information
entry was likely to vary by specialty.[52] Grinspan et al[53] explored the association between physician specialty and EHR adoption, using a retrospective
serial cross-sectional study over time, and their findings reported that physician
specialty was significantly associated with EHR adoption. Whitacre[54] found that, generally, some types of practices had remarkably high-EMR adoption
rates (such as multispecialty and radiology). Grove and Patel[55] stated in their study that physicians in large multispecialty practices reported
the lowest rates of adoption of EHR. On the other hand, Bhargava and Mishra[52] did not find any significant result among different specialty.
Therefore, previous studies regarding the clinical specialty differences in the context
of physicians' intention to HIT use, especially the EMR system usage, remain inconsistent
and inconclusive. Gagnon et al[13] asserted that specialty should be taken into account, while developing EHR implementation
strategies for targeting physicians. As different specialties had expressed different
ideas of how the EHR should function and what it should provide.[56]
As to meet the purpose of HIT requirements of all physician specialties in a better
way, and determining whether such interdisciplinary differences exist, identifying
the physicians' practice nature is imperative. Thus, it is reasonable that physicians'
specialty might influence the strength of the relationship between the criterion and
predictor variables of physicians' intention to EMR use.
Financial Incentives
Employing financial incentives is a renowned practice to improve individual's efficiency
and performance in all kinds of work settings of which the health care system is no
exemption.[57] Due to the physicians' crucial role in coordinating patients' care, any impact employed
on their function has a substantial consequence of the entire health care system.[58] The exercise of remuneration and reward policies are increasingly recognized as
influencing the productivity of health care. However, the use of financial incentives
is one way by which health care organizations attempt to influence the physicians'
behavior. But, thus far, the literature provided little evidence regarding the influence
of financial incentives on the quality of care.[59]
Both Roski et al[60] and Gosden et al[61] supported the argument that financial incentives inevitably contributed intended
behavior changes. Baron et al[62] indicated that the lack of financial incentive was a crucial barrier to EHR adoption
among physicians in the community in their study. Marshall and Harrison[63] pointed out the over-reliance on financial incentives for improving quality of care
and reported that financial incentives did not always result in behavior change as
intended. They further claimed that the attractiveness of financial incentives is
always not based on inclusive empirical evidence. Miller et al[64] reported that lack of financial incentive was one of the factors contributing to
the low adoption rate of EHR in spite of the intensified interest among physicians.
Programs such as “Ontario MD's EMR Adoption Program” provided financial support for
family physicians, and physicians working in primary care to assist with converting
from paper charts and records to EMRs,[65] which boosted up physicians' adoption of EMRs in Canada. Burt and Sisk[66] explored that financial incentive might reward practices that adopted the technology
and improved the physicians' patient care for their use. Conrad and Perry [67] indicated that financial incentives had an impact on clinical quality. As reimbursement
policy might impact medical innovation through their influence on technology adoption,[68]
[69] we explore its influence on physicians' intention to use EMR.
Research Hypothesis
Model Development for Factors Influencing Health Care Technology Acceptance
Davis[70] introduced the TAM and proposed a theoretical framework that explains the relationship
between users' attitude and behavioral intentions. Based on this model, PU and PEOU
are hypothesized to be the principal factors of users' acceptance. Both Yarbrough
and Smith[71] and Holden and Karsh[72] found that PU of IS was positively associated with physicians' attitudes toward
IS tools if the tangible benefits of the IS were reasonably appreciated. In fact,
PU was found to be a strong motivator for predicting health care professionals' attitudes
toward IS in the patient care perspective.[72]
[73]
[74] Conversely, previous studies also concluded that physicians, who perceived the system
as easy to operate and useful, generally developed a positive attitude toward such
a system.[71]
[72]
[73]
Davis[70] stated that there is a direct relationship between attitude and the use of IT, implying
that users will intend to use an IT system when they evaluate it positively. Venkatesh
et al[25] reported that attitude toward using IT, as the second most significant predictor
of the physician's intention to use telemedicine services and this finding supported
the finding of the other studies investigating physicians' IS acceptance.[25]
[75]
[76] Thus, the following three hypotheses based on the above discussion were proposed
in this study:
-
H1
PU positively influences physicians' attitude toward using EMR.
-
H2
PEOU positively influences physicians' attitude toward using EMR.
-
H3
Physician's attitude positively influences his/her intention toward using EMR.
EMR has the potential ability to improve the quality of health care. But, unless physicians
perceive some personal benefits from using EMR, they might not be motivated to switch
and stick to their traditional working processes. Both Miller and Sim[77] and Vishwanath and Scamurra[78] indicated that unless physicians perceived some personal incentives during the implementation
of EMR, the adoption of EMR would not reach the expected level. Significantly, the
incentives considered in the stated studies are mainly financial ones. Patel et al[79] explored that financial concerns were the key barrier to adopting or using HIT by
physicians. Yarbrough and Smith[71] reported that the lack of financial incentives was a critical barrier to adopt and
use EMR by physicians in their systematic review about technology acceptance in health
care. Thus, the following hypothesis based on the above discussion was proposed in
this study:
-
H4
Financial incentives positively influence the physician's attitude toward using EMR.
Previous studies reported differences in the adoption of technology and related application
between men and women.[80]
[81] Research on technology usage between men and women revealed that men tended to exhibit
a task-oriented attitude to show that they understood the usefulness of technology
more effortlessly than women do.[80] Existing literature also revealed that men tended to have more access to technologies
than women.[82] The findings of Dutta et al[83] on the personal health record (PHR) acceptance found that both PU and PEOU had a
stronger influence on female respondents than male respondents.
Different physician groups evaluate the usefulness of IS based on the assessment of
how their needs are satisfied and how easy to use it. Except IS has an extensive attractiveness
to the different clinical specialties (e.g., a diagnostic innovation), there will
be resistance from specialties to the adoption of the system, as the system's perceived
value is limited to a solitary clinical specialty.[1] Melas et al[19] used physician specialty as a moderator to examine the acceptance of clinical information
systems among the hospital medical staff and reported the significant moderating effect
of physicians' specialty. But it is logical that physicians' specialty may directly
influence the strength of the relationship between criterion and predictor variables
in the modified TAM. Based on the above discussion about relationships between personality
traits, PU, and PEOU, the following hypotheses were proposed in this study:
-
H5
Gender influences physician's PU of EMR.
-
H6
Gender influences physician's PEOU of EMR.
-
H7
Clinical specialty positively influences physician's PU of EMR.
-
H8
Clinical specialty positively influences physician's PEOU of EMR.
[Fig. 1] presents the research framework used for the current study.
Fig. 1 Research framework.
Materials and Methods
Questionnaire Design and Data Collection
A questionnaire survey was employed to investigate the proposed theoretical framework.
A questionnaire was developed with a range of items intended to evaluate each construct
of the current study. A preliminary list of measurement items was primarily developed
after reviewing the literature regarding TAM, gender, clinical specialty, and EMR.
The instruments used for the current study comprised three sections. In the first
section, the cover page, the purpose of the study, and the definition of EMR was provided.
The second section considered the respondent's demographic information, including
their gender and clinical specialty. The third section contained indicators concerning
TAM (21 items) and financial incentives (five items). All the items were determined
on a five-point Likert's scales, ranging from 1 for strongly disagree to 5 for strongly
agree. Additionally, the content of the items was revised based on the results of
a pretest and pilot study to enhance the reliability and validity of the items.
Both a pretest and a pilot study were conducted to validate the instrument. The pretest
involved the following five experts: a professor of information management (IM) department,
three doctoral scholars in the medical information field, and an employee who has
been working in the health informatics department in the hospital for more than 10
years. Respondents were asked to explore the appropriateness of items, the format,
and the wording of the scales. The pilot study involved twelve physicians self-selected
from the study population. Based on the respondents' response at the pretest and pilot
study, some items were modified to exhibit the survey's purpose more rationally and
summarized in [Appendix A]. The reliability of all items was acceptable (Cronbach's α is above 0.80) and items loaded in the confirmatory factor analysis were 0.70 or
more. Thus, the instrument has endorsed reliability and content validity. [Appendix B] presents the result of the pilot study.
Appendix A
Measurement items
|
Construct
|
Item no.
|
Item
|
References
|
|
Perceived usefulness (PU)
|
PU1
|
I expect using EMR will improve the quality of my job to provide better patient care
|
Davis[70]
|
|
PU2
|
I believe using EMR will allow me to better control over my work schedule
|
|
PU3
|
I expect using EMR would allow me to finish task more quickly
|
|
PU4
|
I expect using EMR would allow me to finish more task within my work schedule than
before
|
|
PU5
|
I believe using EMR would improve my overall usefulness in my job.
|
|
PU6
|
I expect using the EMR would make my job easier to complete.
|
|
PU7
|
Overall, practicing EMR would be a useful tool in my profession
|
|
Perceived ease of use (PEOU)
|
PEOU1
|
I think that my interaction with EMR would be clear and understandable
|
Davis[70]
|
|
PEOU2
|
I expect learning of EMR would be easy for me
|
|
PEOU3
|
I believe that I would be skillful of using EMR
|
|
PEOU4
|
Overall, I expect that use of EMR will be easy for physician
|
|
Attitude toward using EMR (ATT)
|
ATT1
|
The implementation of the EMR will support the physician in providing better patient
care
|
Davis[70]
|
|
ATT2
|
I will encourage my colleague to use the EMR
|
|
ATT3
|
I need EMR system to provide effective patient care
|
|
ATT4
|
I am not satisfied with using the paper-based patient record in my job
|
|
ATT5
|
All physicians should learn to use EMR successfully
|
|
ATT6
|
Overall, my attitude about EMR has been positive
|
|
Financial incentives (FI)
|
FI1
|
The size of the financial incentive
|
National Health Care Purchasing Institute[108]
|
|
FI2
|
The incentive and need for change recognized among physicians
|
|
FI3
|
The level of support for the incentive program in the medical leadership
|
|
FI4
|
The practicing physicians' knowledge and understanding of the performance incentives
|
|
FI5
|
Overall, financial incentive encourages physician's decision to use EMR
|
|
Intention to use EMR (INT)
|
INT1
|
When it is available in my clinical practice, I intend to use EMR for all my clinical
activities
|
Davis[70]
|
|
INT2
|
When it is available in my organization, I intend to adopt EMR for all my clinical
activities
|
|
INT3
|
The probabilities that I use EMR for all my clinical activities when available in
my organization are very high
|
|
INT4
|
Whatsoever the environments, I do not intend to use EMR when it becomes available
in my organization
|
Abbreviation: EMR, electronic medical record.
Appendix B
Results of confirmatory factor analysis and reliability analysis
|
Constructs
|
Item
|
Loadings
|
Standardized Cronbach's α
|
|
Perceived usefulness (PU)
|
PU1
|
0.908
|
0.965
|
|
PU2
|
0.919
|
|
PU3
|
0.841
|
|
PU4
|
0.915
|
|
PU5
|
0.949
|
|
PU6
|
0.905
|
|
PU7
|
0.934
|
|
Perceived ease of Ease (PEOU)
|
PEOU1
|
0.927
|
0.955
|
|
PEOU2
|
0.932
|
|
PEOU3
|
0.934
|
|
PEOU4
|
0.961
|
|
Attitude toward using EMR (ATT)
|
ATT1
|
0.965
|
0.965
|
|
ATT2
|
0.983
|
|
ATT3
|
0.823
|
|
ATT4
|
0.868
|
|
ATT5
|
0.919
|
|
ATT6
|
0.983
|
|
Financial incentives (FI)
|
FI1
|
0.922
|
0.970
|
|
FI2
|
0.978
|
|
FI3
|
0.902
|
|
FI4
|
0.956
|
|
FI5
|
0.963
|
|
Intention to use EMR (INT)
|
INT1
|
0.850
|
0.893
|
|
INT2
|
0.811
|
|
INT3
|
0.875
|
|
INT4
|
0.823
|
Abbreviation: EMR, electronic medical record.
Research Setting
In terms of hospital attributes, services, as well as the number of beds and physicians,
the hospital employed in this study is able to refer as a typical regional hospital,
located in Southern Taiwan with 654 beds, and 213 physicians available as of 2018.
Prior to commencing the research, Institutional review board (IRB) consent was pursued
and obtained from the hospital. All participants were given a consent form and information
sheet which clearly explained the purpose of the current study. Respondents were also
notified about their rights to withdraw participation at any time during the study.
Moreover, we presented our participants with a short description of how EMR works
in general. This approach was chosen because of the following two reasons. First,
to overwhelm any lack of knowledge about EMR that could have perceived among our participants
reasonably of its continuous technological innovation, and second, to develop a reasonable
conclusion about the prospective applications of EMR.
Study Design and Sampling Distribution
A total of 119 out of 213 physicians working in this hospital responded to this study.
The response rate is around 56% (55.87%). The number of physicians that responded
the questionnaire might be not as large as in other nationwide survey researches.
However, the population of this study is limited on the one site, and furthermore
it is over half of the physicians participate in this study. Additionally, it is representing
that the number of physicians of the current study coincides with the previously published
articles focusing on physicians' intention to adopt health care IT or IS.[77]
[84]
[85]
[86]
[87]
[88]
[89]
[90]
[91]
[92]
[93]
[94] Moreover, to give an indication of the representatives of the study sample, the
personality traits, gender, and clinical specialty, of the current study sample were
compared with the nationwide sample of Taiwan. According to the Taiwan Medical Association
report in 2011, the proportion of total male physicians in Taiwan was 83.7%.[95] However, the majority of the current study sample is male (79.13%), and specialty
divisions correspond with national data of Taiwan.[96] Thus, it is indicating that the distribution of physician in the current study coincides
with the national physician distribution of Taiwan. In other words, the unbiased sample
size of the current study lower down the sample biased and increases statistics power.
Thus, it is indicating that the study hospital and the number of responses are acceptable
compared with prior published researches.
Data Analysis
SPSS and PLS software were used for statistical analysis. Structural equation modeling
(SEM) was used to analyze structural relationships due to three reasons. First, SEM
is a multivariate technique that lets the simultaneous estimation of multiple equations.[97] Second, SEM performs factor analysis and regression analysis in the single step,
as SEM is used to test a structural theory. Third, PLS uses a nonparametric approach
and is not limited by data normality.[98] All constructs were modeled as reflective for the model tested. Data analysis was
conducted on the two-step approaches suggested by Anderson and Gerbing.[99] First, testing convergent validity and discriminant validity of the measurement
model, and subsequently testing research hypotheses and structural model.
Results
Demographic Data
The current study collected 119 responses. Four responses out of 119 were considered
unusable due to missing values. Therefore, we incorporated 115 valid responses for
the final analysis. [Table 1] presents the demographics of respondents, and it points out that the respondents
differ respectively in gender and clinical specialty.
Table 1
Profile of survey respondents
|
Item
|
Option
|
Count
|
Percentage (%)
|
|
Gender
|
Men
|
91
|
79.13
|
|
Women
|
24
|
20.87
|
|
Clinical Specialty
|
Surgery
|
21
|
20.87
|
|
Medicine
|
41
|
33.04
|
|
Obstetrics and gynecology
|
8
|
6.96
|
|
Internal
|
30
|
26.08
|
|
Pediatrics
|
10
|
8.70
|
|
Others
|
5
|
4.35
|
Tests of the Measurement Model
Reliability analysis was tested using Cronbach's α and composite reliability (CR), and [Table 2] shows the results. Cronbach's α of each construct ranged from 0.961 to 0.980 which is above the recommended value
of 0.7 by Hair et al.[97] However, Nunnally[100] suggested 0.90 as the “minimally tolerable estimate” for clinical purposes, with
an ideal of 0.95. But Steiner[101] recommended 0.90 is the most likely indicated unnecessary redundancy. Thus, we consider
CR values of the latent factors to measure the model's internal consistency. The CR
value of each construct is above the recommended value of 0.7,[97] implying acceptable reliability and consistency of the measurement items of each
construct.
Table 2
Measurement model
|
Constructs
|
Item
|
Loadings
|
No. of items
|
Composite reliability
|
Standardized Cronbach's α
|
AVE
|
|
PU
|
PU1
|
0.941
|
7
|
0.983
|
0.980
|
0.894
|
|
PU2
|
0.944
|
|
PU3
|
0.933
|
|
PU4
|
0.956
|
|
PU5
|
0.949
|
|
PU6
|
0.963
|
|
PU7
|
0.931
|
|
PEOU
|
PEOU1
|
0.946
|
4
|
0.979
|
0.971
|
0.921
|
|
PEOU2
|
0.971
|
|
PEOU3
|
0.947
|
|
PEOU4
|
0.973
|
|
ATT
|
ATT1
|
0.971
|
6
|
0.975
|
0.968
|
0.868
|
|
ATT2
|
0.961
|
|
ATT3
|
0.956
|
|
ATT4
|
0.772
|
|
ATT5
|
0.929
|
|
ATT6
|
0.982
|
|
FI
|
FI1
|
0.932
|
5
|
0.978
|
0.972
|
0.899
|
|
FI2
|
0.941
|
|
FI3
|
0.962
|
|
FI4
|
0.946
|
|
FI5
|
0.958
|
|
INT
|
INT1
|
0.953
|
4
|
0.971
|
0.961
|
0.894
|
|
INT2
|
0.958
|
|
INT3
|
0.949
|
|
INT4
|
0.922
|
Abbreviations: ATT, attitude toward using EMR; AVE, average variance extracted; EMR,
electronic medical record; FI, financial incentives; INT, intention to use EMR; PEOU,
perceived ease of use; PU, perceived usefulness.
Convergent validity of the scales was tested by using the following three standards
suggested by Bagozzi[102]: (1) loading of each indicator should be above 0.7,[103] (2) CR value of each indicator should be higher than 0.7, and (3) average variance
extracted (AVE) of each construct should be exceeded the variance because of the measurement
error of that construct (i.e., AVE should be exceeded 0.50). As [Table 2] indicates, the factor loading of each item in the measurement model of the current
study exceeded are well above 0.7. The CR values ranged from 0.971 to 0.983. AVE values
of constructs ranged from 0.868 to 0.921, thus meeting each condition for convergent
validity.
To test discriminant validity, Fornell and Larcker[103] recommended that the square root of the AVE of the construct should be higher than
the estimated correlation shared between the construct and other constructs in the
model. [Table 3] shows that the square root of AVE for each construct is higher than the correlation
values of the construct, thus meeting the condition for discriminant validity.
Table 3
AVE and correlation among constructs
|
ATT
|
INT
|
FI
|
PEOU
|
PU
|
AVE
|
|
ATT
|
0.931
|
|
|
|
|
0.868
|
|
INT
|
0.578
|
0.945
|
|
|
|
0.894
|
|
FI
|
0.804
|
0.487
|
0.948
|
|
|
0.899
|
|
PEOU
|
0.867
|
0.520
|
0.760
|
0.959
|
|
0.921
|
|
PU
|
0.892
|
0.537
|
0.807
|
0.918
|
0.945
|
0.894
|
Abbreviations: ATT, attitude toward using EMR; AVE, Average variance extracted; EMR,
electronic medical record; FI, financial incentives; INT, intention to use EMR; PEOU,
perceived ease of use; PU, perceived usefulness.
Tests of the Structural Model
[Fig. 2] displays the standardized path coefficients, path significances, and variance explained
(R
2) by each path, all supported by the path analysis results, except H2. The coefficient
for determination (R
2) points out that the research framework interprets 33.5% of the variance associated
with intention to use EMR and 82.8% of the variance associated with attitude toward
using EMR is explained by PU, PEOU, and financial incentives. [Table 4] reports the results of the hypothesis test.
Table 4
Results of the regression tests
|
Path
|
β
|
Results
|
|
PU → ATT
|
0.454
|
H1: supported
|
|
PEOU → ATT
|
0.277
|
H2: not supported
|
|
ATT → INT
|
0.227
|
H3: supported
|
|
FI → ATT
|
0.578
|
H4: supported
|
|
Gender and clinical specialty differences
|
|
Model 1: F = 14.749, df = 2/112, p = 0.000 < 0.05
|
|
1. Perceived usefulness (R = 0.457, R2
= 0.208, Durbin–Watson = 1.711)
|
|
Gender
|
−0.212
|
H5: supported
|
|
Clinical specialty
|
0.420
|
H6: supported
|
|
Model 2: F = 8.886, df = 2/112, p = 0.000 < 0.05
|
|
2. Perceived ease of use (R = 0.370, R2
= 0.137, Durbin–Watson = 1.858)
|
|
Gender
|
−0.204
|
H7: supported
|
|
Clinical specialty
|
0.463
|
H8: supported
|
Abbreviations: ATT, attitude toward using EMR; EMR, electronic medical record; INT,
intention to use EMR; FI, financial incentives; PEOU, perceived ease of use; PU, perceived
usefulness.
Fig. 2 Path analysis result. *Significant at p < 0.05 level; p < 0.01**, p < 0.001***. EMR, electronic medical record; NS, not significant at p < 0.05 level.
The Effect of the Personality Traits (Gender and Clinical Specialty)
The second section of the study investigates the influences of personality traits
on EMR use decision. Multiple regression analyses were conducted to test the proposed
hypotheses on the constructs: PU, PEOU, and personality traits (gender and clinical
specialty). Both the personality trait variables, gender (H5 and H6) and clinical
specialty (H7 and H8) were consistently found to predict both PU and PEOU. [Table 4] presents the relationships between personality traits and [Fig. 2] indicates β values for each hypothesis.
Discussion
The present study empirically validates the classical theory, TAM in a health care
perspective by going a step further to examine the effect of the gender and clinical
specialty on physicians' intention to use EMR. Physicians' resistance to using the
EMR system appears as the critical barrier to attaining comprehensive interoperability
and achieve the benefits that can be executed from EMR. Thus, identifying the factors
affecting physicians' intention toward EMR use is important. As they are the major
user-group of EMR use and their intention to use is the primary condition to ensure
that the expected benefits will be materialized.
There was a significant positive association between PU and attitudes toward using
EMR, supporting H1. It implies that if physicians believe EMR is useful in patients'
care, then physicians develop a positive attitude toward using EMR. Given this, when
developing the EMR system, the health care providers should focus on strengthening
the usefulness of EMR; considerably improve the functions that must meet physicians'
requirement to carry out the medical practice. This development can motivate physicians
to use the system persistently and further improves their intention to use EMR. However,
this finding is consistent with the finding of Taylor and Todd[104] who also indicated that PU had both direct and indirect influences on the attitude
toward using the system. Conversely, PEOU was negatively correlated with attitude
toward using EMR, consistent with the previous study using the physician as subject.[25] Therefore, H2 was not supported. The EMR is being used by physicians with a specific
purpose to improve the quality of patients' care by increasing care coordination and
eliminating errors; additionally, want to perform their medical practice more efficiently.
So, physicians are mostly concerned about whether the services and contents offered
by the EMR system are beneficial to improve their patients' care performance rather
than feelings of easiness to operate the system. If physicians perceive that despite
the system is easy to use, but did not improve their patients' care performance, then
their attitude toward using EMR is not going to be improved anyway. Therefore, the
difficulties with the system's interface or easiness to operate may possibly not be
a significant consideration in patients' care perspective. Thus, the finding of the
current study recommends that health care system developers should emphasize on the
factors, physicians reasonably expecting from the EMR system, such as well-timed and
necessary information for patients' care, authentic data regarding the patients' health
condition, etc., could improve physicians' intention to use EMR.
In other words, the result from the nonsignificant relationship could be the combination
of factors such as comprehensive computer literacy and small sample sizes. Physicians,
on average, have a higher level of competency, intelligence, reasoning capacity, and
skills to adopt new technologies. As a result of that, they are different from other
user groups. Therefore, the variables of PEOU are still critical, but not significant
(p = 0.07). Additionally, the small sample size limited the ability to detect a difference.
Therefore, it is potential that a difference would have been noticed if the sample
size was bigger. Thus, we suggest conducting more studies, especially enlarging the
number of respondents.
The finding of the study revealed that attitude toward using EMR was positively associated
with intention to use EMR, supporting H3. This finding is consistent with Davis[70] and Chau and Hu.[25] The significant relationship between attitude and the use of EMR indicates that
physicians evaluate the system positively. If physicians perceive a positive attitude
toward using EMR system, their intentions of using EMR also improve. Thus, positive
attitude should be considered as a strong determinant, while acceptance of the EMR
system is concerned.
The results indicated that physicians place positive value regarding financial incentives,
and it favorably affects attitude which indirectly influences physicians' intention
to use EMR, supporting H4. However, financial incentives can positively change the
physicians' attitudinal belief and form a positive attitude toward EMR usage. The
finding of the current study is in line with the previous study finding by Cohen,[105] which indicated that financial reward positively influenced physicians' perception
toward using IS.
Previous studies based on TAM hypothesis stated that personality traits would moderate
an individual's intention to use new Information Communication Technology (ICT) by
persuading individuals' PU and PEOU. These variables might support researchers to
understand the individual's adoption decision behavior. In the present study, it is
theorized that personality traits, gender, and clinical specialty, mediate the relationships
between an individual's belief (PU and PEOU) and future intention to use. Pijpers
et al.[106] also explored that the interpretation of the dynamics between mediating, independent,
and dependent variables in TAM could result in better prediction of variables to intensify
the intention to use emerging technologies.
Gender has recently appealed to the researchers' attention to study its influence
on adoption behavior of individuals.[80] Previous TAM studies addressed gender-based affective differences and their effects.
Hoque[30] thus hypothesized that males had a higher level of mHealth adoption intention than
females. Though, empirical findings of the current study reported significant but
negative, the relationship between gender (i.e., maleness), PU, and PEOU. It indicates
that physicians consider a high degree of usefulness for health care technology, such
as EMR. EMR influences actual usage of services within hospitals. Thus, physicians
should be motivated to make use of EMR since it improves performance by decreasing
errors and time necessary for treatment. Additionally, the finding indicates that
physicians also demonstrate an understanding of EMR usage, along with its convenience,
ease of use, and lack of technical difficulties; until they would like it to be usable.
Physicians primarily have the key responsibility of treating patients and their time
is limited. Thus, they want their association with technology to be uncomplicated.
However, the negative relationship might be moderately attributed to the gender imbalance
in the physician sample in the current study.
The current study found that the clinical specialty was the consistent predictor of
physicians' PU and PEOU which also indirectly influences their intention to use EMR.
The differences reported across specialties may be indicative of the emphasis that
put on a certain type of practices. The potential factor in differing frequencies
of EMR use among different specialties is that, until recently, many of the EMR systems
tended to be “one-for-all” with features and tabs that might be effective for some
specialties but not as efficient for others.[107] In recent times, nevertheless, EMR developers have focused on developing more specialty
specific EMR programs. However, it is still not well defined what influence these
specific specialty programs will have on the implementation rate of EMR in the soon.
Another contributing factor could be considering the perceptible requirement of all
the proposed uses for an adaptable interface, presently used EMR only moderately fulfills
the requirements of the different specialties. Thus, physicians' requirements should
be carefully integrated during the development stages of EMR. The differences in preferences
among different physicians indicate that different specialties have different needs
during the implementation of innovations. Therefore, EMR system developers might use
this finding to design the implementation trajectories to fit the needs of different
physician specialties.
The present research contributes to theory and practice in several ways. First, the
integrated model analyzed in the current study, combined elements of the TAM, context-specific
variable financial incentives, and personality traits, gender, along with clinical
specialty, has overcome the limited applicability of the TAM to study physicians'
intention to use HIT. Additionally, the results of the study improve the current understanding
in the field of technology acceptance and HIT implementation. Second, the study instrument
not only contributes an overall assessment but also has the ability to examine what
characteristics of the EMR (technology, behavioral, or user's specific differences)
adoption are challenging from the physicians' perspective. Third, our extended TAM
explains how differences in usage intention are influenced by EMR perception in physicians.
The adoption theory evolved from the current study could be improved for application
in large-scale services and organizations considering the adoption of EMR. Fourth,
an understanding of the effects of gender differences on the intention to use EMR
is important in overcoming barriers to the diffusion of technology across institutions.
A consideration of the mechanisms through which gender differences impact technology
usage behavior, is significant for trimming down resistance to technology use. Fifth,
an understanding of the effects of clinical differences of physicians on the intention
to use EMR is important in overcoming barriers to the diffusion of technology across
the hospital and reducing the resistance to technology use. Sixth, as this study focuses
on EMR use and unlike, studies examine behavioral intention; any development regarding
the better understanding of phenomena can translate into higher acceptance and usage
of the HIT after implementation. Finally, the results of the present study lead to
better technology usage and can also have a better consideration for health care providers
and policymakers before taking the decision about further spending on new HIT implementation.
Limitations and Future Research
Limitations and Future Research
Despite its significant findings and implications, two potential limitations of this
study require consideration. First, the current study is one of the first employing
gender and clinical specialty as antecedents and included a context-specific factor,
clinical specialty with the classical model TAM to evaluate the physician's behavioral
intention to use and adopt the EMR in Taiwan. In addition, the implications are from
a single study with samples in Taiwan which may have led to institutional bias. Thus,
researchers should be careful while generalizing the results to other health care
circumstances. Future studies should conduct research in a nationwide perspective
to explore and compare the differences in the antecedents to usage intention. Second,
56% of the eligible physicians participated in the survey. However, the response rate
is reasonably acceptable in the perspective of physicians' HIS acceptance, but the
sample size is moderately low. Relatively low sample size may have introduced selection
bias, and we are not conclusive that the results of the study are representative of
the entire population of physicians in Taiwan. Thus, the findings of this study can
be used as the basis for a nationwide survey to enhance the external validity of those
findings.
Conclusions
The current study examines critical factors influencing physicians' intention to use
EMR through an integrated model derived from classical theory TAM with the attitudinal
behavioral factor and financial incentives. Additionally, we explore the effect of
gender and clinical specialty to confirm and expand the EMR system adoption model.
Results from SEM analysis demonstrated that the model provided meaningful intuitions
for perception, interpretation, anticipation, and presented good explanatory power
to predict physicians' intention to use EMR, and providing a new direction for researchers
to contemplate in subsequent research. The current study primarily identifies three
relevant factors, that is, PU, financial incentives, and attitude toward using EMR,
positively influencing physicians' intention to use EMR. Additionally, gender and
clinical specialty differences are consistently found to predict both PU and PEOU
which indirectly influence physicians' intention to use EMR.
Moreover, despite the small sample size, findings of the current study make a significant
contribution in both academic and practical issues. Our study hospital is a typical
regional teaching hospital with 654 beds, 213 physicians from different specialties,
and fairly provides medical services to 650,000 patients annually. Thus, in terms
of the number of beds, physicians, and outpatient services, we consider the findings
of the current study are acceptable and external validity is also validated.
