High-Level Managers’ Considerations for RFID Adoption in Hospitals: An Empirical Study in Taiwan

ORIGINAL PAPER

High-Level Managers’ Considerations for RFID Adoptionin Hospitals: An Empirical Study in Taiwan

Hui-Min Lai & I-Chun Lin & Ling-Tzu Tseng

Received: 12 September 2013 /Accepted: 23 December 2013 /Published online: 21 January 2014# Springer Science+Business Media New York 2014

Abstract Prior researches have indicated that an appropriateadoption of information technology (IT) can help hospitalssignificantly improve services and operations. RadioFrequency Identification (RFID) is believed to be the nextgeneration innovation technology for automatic data collec-tion and asset/people tracking. Based on the Technology–Organization–Environment (TOE) framework, this study in-vestigated high-level managers’ considerations for RFIDadoption in hospitals. This research reviewed literature relatedIT adoption in business and followed the results of a prelim-inary survey with 37 practical experts in hospitals to theorize amodel for the RFID adoption in hospitals. Through a fieldsurvey of 102 hospitals and hypotheses testing, this researchidentified key factors influencing RFID adoption. Follow-upin-depth interviews with three high-level managers of ISdepartment from three case hospitals respectively also pre-sented an insight into the decision of RFID’s adoption. Basedon the research findings, cost, ubiquity, compatibility, securityand privacy risk, top management support, hospital scale,financial readiness and government policy were concludedto be the key factors influencing RFID adoption in hospitals.For practitioners, this study provided a comprehensive over-view of government policies able to promote the technology,while helping the RFID solution providers understand how toreduce the IT barriers in order to enhance hospitals’ willing-ness to adopt RFID.

Keywords Radio Frequency Identification (RFID) .

Technology adoption .

Technology-Organization-Environment (TOE)model .

Hospitals

Introduction

National Health Insurance (NHI) was established by theTaiwan Bureau of National Health Insurance (BNHI) in1995, initially as a fee-for-service reimbursement system. Ithas brought changes to hospitals’ operation ever since itsinception and continued to weighted heavily on hospitals’daily management. Recently, as an attempt to reduce BNHI’shuge deficit, the system was modified into a fixed reimburse-ment, meaning global budget. This change put hospital man-ager under greater pressure to keep the cost down. Whilekeeping cost benefits in check, hospitals have to face a differ-ent challenge of delivering the best possible services broughtabout by the increase awareness of patient rights. By nature,hospitals are in an information-intensive industry and thusrequire to invest in new technologies to maintain or improveits performance. An increasing number of researches hasindicated that an appropriate adoption of information technol-ogy (IT) can significantly improve quality and outcome [1–3].However, Chang et al. [1] pointed out that not all hospitalsadopt IT without hesitation. The issue of “what factorsinfluencing the adoption of IT in a healthcare setting” becomesan important question for all healthcare administrators.

Radio Frequency Identification (RFID) is a fast developingand emerging technology that uses radio waves for data collec-tion, information transfer and patient identification/tracking. Itcan capture data efficiently and automatically without humanintervention [4]. The Institute for Information Industry reportedthat themajor RFID adopters in Taiwan are businesses relating tothe retail, logistics, and transportation companies. However, thereare fast growing demands for RFID in the healthcare industry [5].Although there are extensive applications of RFID suitable forthe healthcare industry, only few hospitals have used RFID. The

H.<M. Lai : L.<T. TsengDepartment of Information Management, Chienkuo TechnologyUniversity, No.1, Chiehshou North Road, Changhua 500, Taiwan,Republic of China

I.<C. Lin (*)Department of Industrial Management and Institute of HealthIndustry Management, National Yunlin University of Science andTechnology, No. 123 University Road, Sec. 3, Douliou,Yunlin 64002, Taiwan, Republic of Chinae-mail: [email protected]

I.-C. Line-mail: [email protected]

J Med Syst (2014) 38:3DOI 10.1007/s10916-013-0003-z

key reason for hospitals hesitation over RFID adoption is its highinitial investment and the difficulty to envision the benefits. Thepotential benefits of RFID adoption in hospitals include improv-ing the quality of care, patient/customer satisfaction, andsearching for the efficiency of equipments [6]. More specifically,it can be applied to medical personnel identification, such aspatient and neonatal management, and critical itemmanagement,such as blood management, medical waste management, anddrug management. Furthermore, it can be used in personalinformation management, such as sample management, medicalrecord management, and patient bedding/clothing management.RFID is believed to be the next generation innovation technologyfor automatic data collection and asset/people tracking. Most ofthe research related to RFID focused on the technology itself,paying little emphasis on related management issues. “Whatpromotes the adoption of RFID in hospitals” is also an importantresearch topic. The purpose of this study is to identify the salientdeterminants of RFID adoption in the healthcare industry. Forresearchers, this study contributes to a theoretical understandingof the adoption of RFID in hospitals. For practitioners, this studyprovides a comprehensive overview for the government, helpingto shape the policies that promote the technology and in additionfor the RFID solution providers, enabling them to reduce the ITbarriers, so as to enhance hospitals’ willingness to adopt RFID.

Literature review

RFID and healthcare industry

RFID technology is a wireless system that uses radio-frequencyelectromagnetic fields to transfer data for identification pur-poses. Three major components of a RFID system are asfollow: (1) Tag: also known as a transponder or a contactlessdata carrier, it is planted to objects for identifying a set ofcorresponding object and usually classified into two types:active or passive depending on the existence of battery. Anactive tag usually contains a battery and is able to can send outinformation at any time, while enjoying a longer communica-tion distance and a larger storage memory. It however has thedisadvantages of bulkiness, relatively short service life, the needfor stricter environmental requirements and higher costs. Apassive tag receives the electromagnetic waves transmitted bythe reader, using microcurrent through the induction process,and then transmitting the information back to the reader. As aresult, it has a shorter communication distance. The fact that itdoes not an external battery gives a passive tag several advan-tages, such as compactness, a relatively long service life andcheaper prices. (2) Reader: communicates with tags to enablewireless data transfer. (3) Software Application: enables a read-er to read or write the electronic tags. In theory, informationstored in a RFID tag is sent to a reader via contactless trans-mission, allowing the reader to read the data before sending

them to a backend application by means of wired or wirelesstransmission for further data analysis or other processing.

There are two main streams of previous studies regarding theRFID adoption in hospitals with one investigating the underlyingmotivations and driving forces behind the adoption of RFID [7,8] and the other looking at the potential utilization of RFID andits effectiveness [6, 9]. The first stream, for example, was derivedfrom the perspectives of technology-push and need-pull, whereinLee & Shim [7] conducted a survey to investigate the likelihoodof adopting RFID in U.S. hospitals. In this survey, they foundone technology push factor (perceived benefits), two need pullfactors (performance gap andmarket uncertainty), and concludedthat the presence of champions (decision makers) was the mostimportant factor influencing the RFID adoption in the healthcareindustry. As a representative of the second stream, Zhou &Piramuthu [9] confirmed that RFID-enabled real-time medicalprocess and labor management provided a marginal improve-ment for the premium medical service providers, meaning that itcould generate appreciable improvement in terms of both effi-ciency and service quality of public health care institutions.

A health care industry is an information-intensive industryand a proper adoption of IT can have a significant impact on ahospital’s medical services quality [1]. However, many hos-pitals are struggling with the ever-growing operating costs,contributed by a combination of factors, such as rising wages,necessary equipment purchases and malpractice lawsuits. Forthe sake of efficiency, hospitals need to introduce a highlydynamic operational process which reduces error rates andimproves equipment management. To this end, RFID adop-tion offers a way. Currently, RFID applications in Taiwan’shospitals are including operation roommanagement; tracking/identification/locating assets/patients; staff attendance man-agement; blood bags/drugs/medical waste management;, neo-natal management, patient escort management etc.

The factors affecting RFID adoption in various contexts

Several previous studies had focused on the different perspec-tives of RFID adoption in various contexts. For example, Lee &Shim [7] predicted the feasibility of RFID adoption in the healthcare industry using the theory of technology-push and need-pull.They surveyed the senior executives of 126 U.S. hospitals inorder to investigate the possible drivers. They assumed therewere three dimensions affecting RFID adoptions in organiza-tions, which included: (1) technology-push: performance gapand market uncertainty; (2) need-pull: vendor pressure and per-ceived benefits; and (3) presence of champions. Their researchresult showed that, except vendor pressure, the otherfour variables–performance gap, market uncertainty, per-ceived benefits and presence of champions–had signifi-cant impact on the possibility of RFID adoption. Amongthese factors, the presence of champions was the most impor-tant factor. This study also proposed that a successful RFID

3, Page 2 of 17 J Med Syst (2014) 38:3

adoption should include technology-push, need-pull, and thepresence of champions for RFID.

Kim & Garrison’s [10] study of 278 organizations in SouthKorean retailers showed that organizational needs (ubiquity,performance gaps, and job relevance), perceived factors (ben-efits and cost Savings), and organizational readiness (financialresources and technological knowledge) as the key factorsdetermining RFID Evaluation; and evaluation impacts itsadoption, and integration.

Another study conducted by Wang et al. [11], according tothe Technology-Organization-Environment (TOE) framework,proposed that there were nine key variables affecting the RFIDadoption in the manufacturing industry, which comprised rel-ative advantage, compatibility, complexity, top managementsupport, firm size, technology competence, information inten-sity, competitive pressure, and trading partner pressure. Theysurveyed IT executives of 133 manufacturers in Taiwan andthe results showed that complexity, compatibility, firm size,competitive pressure, trading partner pressure and informationintensity could affect the RFID adopted by the manufacturers.

A further review of relevant literature helped us decide on thekey factors affecting RFID adoption in hospitals. Table 1 showsthe summary of these factors and their comparison with theresults of prior studies. These factors were categorized into threedimensions: technological, organizational and environmental.

Method and materials

Research process and research model

The research process comprised three steps. First, this researchreviewed literature on IT adoption in business and followed theresults of a preliminary survey with 37 practical experts in

hospitals to theorize a model for the RFID adoption in hospitals.According to ScottMorton [12], key factors affecting technologyadoption in business can be classified into three dimensions,namely organizational dimension, environmental dimensionand characteristics of IT itself [12]. Grover & Goslar [13] pre-sented a similar concept and the use of empirical testing. Basedon these, we formed the Technology-Organization-Environment(TOE) framework in this study. In the preliminary survey (seeAppendix A), a total of 37 high-level mangers from hospitals’ ISor nursing departments were invited to participate in a question-naire survey. They were asked to rank the factors from high tolow according to relative importance. The variables mentionedmore than 20% in the questionnaires were selected (see Table 2).Such selection criteria and suggestion was adopted by [14].Afterward, our research model and hypotheses (H1–H14) wereproposed as seen in Fig. 1.

Second, a field survey involving 102 hospitals wasconducted to test the research model during the periodof February to November 2011. Third, follow-up in-depth interviews with three directors of IS departmentin three case hospitals individually were conducted fromFebruary to April, 2012, in which they were asked toprovide further qualitative data able to contribute to adeeper understanding of the factors that determined theadoption of RFID.

Hypotheses

Technology-RFID characteristics: H1 To H6

Cost Cost includes all kinds of cost associated with RFIDadoption, including cost of tags, readers, installation, systemintegration, education and training, implementation, develop-ment and operation [15]. Adopting innovative technology can

Table 1 Literature review and comparison

Study Research context Technology-RFID characteristics Organizational dimension Environmental dimension

A B C D E F G H I J K L M N O P Q R S T

Brown & Russell [55] Retail sector x x x x x x x x x x x

Lee & Shim [7] Healthcare x x x x x x x

Krasnova et al. [29] Automotive x

Madlberger [34] Supply chain x x x

Kim & Garrison [10] Supply chain x x x x x x x

Tsai et al. [26] Retail chains x x x x x

Wang et al. [11] Manufacturing x x x x x x x x

Hossain & Quaddus [41] Livestock x

Chong & Chan [8] Healthcare x x x x x x x x

A–Perceived benefits (or Relative advantage), B–Compatibility, C–Complexity, D–Cost, E–Ubiquity, F–Job-related, G–Security, F–Performance gap, I–Financial resources (or Financial readiness), J–Technological knowledge (or Technological readiness), K–Organizational size, L–Top managementsupport,M–Presence of champions, N–Vender pressure, O–Market uncertainty (or Competitive pressure), P–Standards uncertainty (or External initiatorsfor changes), Q–External support, R–Government policies, S–Information intensity, T–Expectation of market trends

J Med Syst (2014) 38:3 Page 3 of 17, 3

bring relative benefits to organizations. However, relativebenefits are total benefits deducted by cost and cost is usuallytaken into consideration when adopting an innovative tech-nology [16, 17]. Prior research indicated that the main con-straint of extensive use of RFID is the cost of tags [18].

Although the cost of RFID tags has decreased, the numberof tags required will be enormous if RFID is applied to patientand drug management in hospitals. Coupled with other costs,including hardware (PDA hand held devices and wirelessinternet), software, development and maintenance, RFID

Table 2 The results of a prelimi-nary survey (N=37) Dimension Predictors Frequency Percentage

RFID characteristics Cost 33 89.19 %

Perceived benefits 32 86.49 %

Ubiquity 27 72.97 %

Complexity 15 40.54 %

Compatibility (System integration) 11 29.73 %

Perceived risk (Security and privacy) 8 21.62 %

Job-related 5 13.51 %

Organizational dimension Top management support 34 91.89 %

Hospital scale 32 86.49 %

Financial readiness 26 70.27 %

Technological readiness 22 59.46 %

User support 8 21.62 %

Presence of champions 5 13.51 %

Headcounts of IT department 5 13.51 %

Performance gap 2 5.41 %

Environmental dimension Government policies 31 83.78 %

External support (consultants, software suppliers) 27 72.97 %

Market uncertainty (Competitive pressure) 15 40.54 %

Interference of materials 6 16.22 %

Information intensity 4 10.81 %

Standards uncertainty 3 8.11 %

Vender pressure 2 5.41 %

Technology-RFID Characteristics

Cost ( )Perceived Benefits (+)Ubiquity (+)

Complexity ( )

Compatibility (+)

Security and Privacy Risk ( )

Organizational Dimension

Top Management Support (+)Hospital Scale (+)Financial Readiness (+)Technological Readiness (+)Users Support (+)

Environmental Dimension

Government Policies (+)External Support (+)Market Uncertainty (+)

Decision of RFID Adoption in Hospitals

H1~H6 H7~H11

H12~H14

Fig. 1 Research framework

3, Page 4 of 17 J Med Syst (2014) 38:3

adoption brings a huge financial burden to hospitals. Priorresearch confirmed that when the cost associated with inno-vative technology adoption is too high, users will have diffi-culties in the adoption [19, 20]. We therefore proposed hy-pothesis 1:

H1: Cost has a negative effect on RFID adoption inhospitals.

Perceived benefits Perceived benefits are also called relativeadvantage. When individuals perceive higher relative advan-tage from RFID adoption, the adoption speed becomes faster[21]. Perceived benefits is seen a key factor affecting RFIDadoption [7, 22]. When innovative technology can bring bene-fits such as improving customer service quality and enablingtimely decision-making, organizations are motivated to adoptsuch innovative technology [23]. Therefore, our study infersthat with higher perceived benefits of RFID such as improvingpatient satisfaction, improving service quality and increasingoperating efficiency, hospitals are more likely to adopt RFID.Here we proposed hypothesis 2:

H2: Perceived benefits have a positive effect on RFIDadoption in hospitals.

Ubiquity Ubiquity means that RFID systems can transmitcommunicating, monitoring, and control signals to individualsor objects to perform various functions, regardless of users’whereabouts. RFID can provide personalized and continuousconnection and communication [10] thanks to its light weight,small size and easy connection with mobile communicationdevices. In a vast and hectic workplace like a hospital, tolocate a person or search for an object can become an unwor-thy waste of medical personnel’s capacity and time [24]. RFIDsystems can read location or environment information fromRFID tags. Because of the computing power of mobiledevices and wireless local area network (WLAN), RFIDsystems can be used in ubiquitous computing environ-ments by reading RFID tags with mobile devices andsending data to the database. Medical information andhistory can be accessed and retrieved anytime and any-where [25] to help reduce the rate of medical errors, extendthe coverage of medical services, and improve service quality.Therefore, the ubiquitous nature of RFID makes it even moresuitable for medical management systems. Our study inferredthat hospitals are more motivated to adopt RFID when theyregard such a ubiquitous nature as a convenient feature. Wetherefore propose hypothesis 3:

H3: Ubiquity has a positive effect on RFID adoption inhospitals.

Complexity Complexity refers to the degree to which aRFID technology is perceived as difficult to use RFID

[21]. Although organizations can benefit from adoptinginnovative technology, they may also encounter difficul-ties if innovative technology is too complicated. Priorresearch have shown that complexity is one of theobstacles to adopting RFID [11, 20, 26]. RFID has tobe installed and set up according to specific work en-vironment and application purposes. Its hardware has towithstand the heat and humidity in Taiwan in particular,and be operated with different materials and in different workenvironments [26]. Therefore, in order to have better datatransmission, the interference between backend systems ofRFID and existing IT systems has to be adjusted effectivelyand this increases the level of complexity of adopting RFID[26]. Here we proposed hypothesis 4:

H4: Complexity has a negative effect on RFID adoptionin hospitals.

Compatibility Compatibility is the degree to which a RFIDtechnology is perceived as consistent with the existing values,needs, and past experiences of the potential adopter [21].Technical compatibility measures whether such innovativetechnology matches existing IT systems. With low compati-bility, organizations are more resistant to changes. Studieshave shown that compatibility is one of the key factors ofRFID adoption [11]. Because changes in enterprises’ workprocesses are involved in the implementation of RFID sys-tems, users resistance to changes has significant impacton RFID implementation [27]. We therefore proposedhypothesis 5:

H5: Compatibility has a positive effect on RFID adoptionin hospitals.

Security and privacy risk Security and privacy risk prob-lems are formed when organizations perceive uncertain-ty and possible risks associated with RFID usage [28].Although organizations can improve productivity byadopting RFID, this also means a ubiquitous monitoring[27] likely to expose organizations and individuals tothe threat of security and privacy breaches [18]. Forexample, if RFID systems are not well-secured, theremight be unauthorized data access; or if hackers initiateattacks against RFID systems, hospitals might incurhuge loss due to system malfunctioning. Here we proposedhypothesis 6:

H6: Security and privacy risk has a negative effect onRFID adoption in hospitals.

Organization dimension: H7 to H11

Topmanagement support Topmanagement usually refers to thedecision makers of innovative technology implementation who

J Med Syst (2014) 38:3 Page 5 of 17, 3

have greater influence on the adoption. RFID implementationinvolves significant changes in financial investment and costlyprocesses; therefore can be a strategic decision requiring topmanagement support [29]. Top management support will affecthospitals’ new IT adoptions [30]. Their support can effectivelymitigate users’ resistance against adopting new IT systems [17].With the same IT investment in systems, the stronger the topmanagement commitment, the better the firm performance [31].Top management are key to RFID adoption in an inter-organizational system [22]. We therefore propose hypothesis 7:

H7: Top management support has a positive effect onRFID adoption in hospitals.

Hospital scale In general, large-scale hospitals are more likelyto adopt innovative technology than small-scale hospitals[32–34], because compared to small ones, large ones usuallyhave more resources, more budget, better IT infrastructure, bettertechnological environments and the ability to bear larger risks[32–34]. While smaller organizations are more likely to beconstrained by lack of resources, larger organizations can stilladopt innovative technology [35]. Prior research has indicatedthat organization scale affects RFID adoption [11]. With moreresources, large-scale hospitals are able to assign internal expertsto deal with such tasks. Here we proposed hypothesis 8:

H8: Hospital scale has a positive effect on RFID adoptionin hospitals.

Financial readiness Financial readiness refers to the level offinancial resources available in hospitals for RFID adoptionincluding installation costs, implementation and maintenance[10]. Prior research have shown that due to financial con-straints and lack of knowledge in IT systems, the growth ofIT adoption in small organizations is limited [36]. Only whensmall organizations are financially ready and have sufficientresources, adopting innovative IT technology is consideredfeasible [35]. Iacovou et al. [37] believed that organizationalreadiness includes both financial readiness and technologicalreadiness. Prior research have shown that financial readinessaffects the willingness for the automobile industry to adoptRFID [29]. When hospitals are more financially ready, theyare more willing to adopt RFID. We therefore proposed hy-pothesis 9:

H9: Financial readiness has a positive effect on RFIDadoption in hospitals.

Technological readiness Technological readiness refers to thelevel of sophistication regarding IT usage and IT management inan organization [37]. Small enterprises lack financial and tech-nological resources; therefore, providing financial and techno-logical support is one of the key factors promoting IT adoption[37]. RFID being a radical innovative technology, users have to

learn new skills and establish new infrastructures to supportoperations in RFID environment [22]. Therefore, we infer thatwhen hospitals are more technologically ready, they are morelikely to adopt RFID. Here we proposed hypothesis 10:

H10: Technological readiness has a positive effect onRFID adoption in hospitals.

User support User support refers to the change in users’ psy-chological state, caused by using new systems and performingtheir tasks with the systems [38]. Prior research has shown thatwhen users are not psychologically ready to accept new IT, theirattitude and behavior make them refuse to receive new informa-tion from consulting firms [39]. Lin et al. [3] also indicated thatuser resistance is a critical barrier for healthcare informationtechnology (HIT) adoption, as user resistance often leads to anincrease of cost in HIT implementation, and waste of resourceswithin a hospital. Lack of user support might lead to unoptimizedperformance or failures [40]. When implementing RFID, hospi-tals might have to change their entire work process. Whenmedical personnel provide low level of support for RFID imple-mentation, the risk of failure implementation is increased. Wethus propose hypothesis 11:

H11: User support has a positive effect on RFID adoptionin hospitals.

Environmental dimension: H12 to H14

Government policy Government policy includes govern-ment’s financial support, training curriculum, specificationand policy stability [1]. When adopting RFID, organizationsexpect to receive support from government with respect topolicies, incentives and subsidies to accelerate the rate ofadoption [41]. RFID planning promoted by government helpsreduce hospital’s financial pressure by offering subsidies andtheir continuance of the implementation plan. Here we pro-posed hypothesis 12:

H12: Government policies positively affect RFID adop-tion in hospitals.

External support Hospitals might lack RFID experts, but canlook for other support such as RFID consultants or venders’help. Thong et al. [42] indicates that the efficacy of consultingfirms and suppliers’ support affect the successful IT imple-mentation. This is especially the case for small organizationsas compared to large enterprises due to the lack of internalexperts [35, 43, 44]. Consulting firms can provide profession-al advice, analyze information needs, look for external con-sultants with comprehensive experience helps organizations todraw a complete picture of possible problems faced duringimplementation [40, 45]. Suppliers, on the other hand, canprovide support such as hardware/software testing

3, Page 6 of 17 J Med Syst (2014) 38:3

environment, technical support, education and trainings [42].Therefore, this study infers that support from consultants orsuppliers can help RFID adoption in hospitals. We thereforepropose hypothesis 13:

H13: External support positively affects RFID adoptionin hospitals.

Market uncertainty Market uncertainty is defined as the mo-tivation of adopting RFID resulting from pressure from exter-nal market [46]. The level of competition intensity betweenorganizations is positive related to their adoption of new IT[13, 34, 47] and RFID technology (e.g.,[7, 11, 20]). Therefore,this study infers that higher the market uncertainty, the morelikely hospitals will look for new technology. Decision mak-ing of RFID adoption is also affected. Here we proposedhypothesis 14:

H14: Market uncertainty positively affects RFID adop-tion in hospitals.

Measurement and data collection

A total of 15 variables were included in our research frame-work, as seen in Fig. 1. Most of themweremeasured by a five-point Likert scale anchored between strongly disagree tostrongly agree, with hospital scale as an exception. The de-pendent variable is whether hospitals adopt RFID or not. Allrespondents were asked to choose from the following optionan answer that best describes the current status of RFIDadoption in their hospitals: adopted RFID, started usingRFID, planned to adopt RFID, and no RFID adoption plan.According to the innovation diffusion theory, a hospital iscategorized as an adopter when its respondents identify it ashaving “adopted RFID” and “started using RFID”. In con-trary, when a hospital’s respondents select “planned to adoptRFID” and “no RFID adoption plan” as their answers, it iscategorized as a non-adopter.

According to Taiwan’s Bureau of National Health Insurance,there are 510 district-level and above contract hospitals in 2010,ranging frommedical center, regional hospital to district hospital.To cover them as our survey target, we sent out 510 question-naires via e-mail. In order to increase response rate, each partic-ipant was awarded $7 (US) for participation. There were 102valid responses (20 % valid response rate)–a result similar toother nationwide hospital surveys in [33]. Regarding the phasesand current status of hospitals’ RFID adoption (see Table 3), oursample showed that 37 hospitals (36.3 %) were adopter hospitalsand 65 hospitals (63.7 %) were non-adopter hospitals. Table 4provided respondent characteristics in detail.

Following Armstrong and Overton’s suggestion [48], wealso examined the sample data for evidence of non-responsebias using t-test. The non-response bias was assessed by

verifying the differences between early respondents (35 %)and late respondents (65 %), as late respondents were almostsimilar to non-respondents. The result indicated that all theindependent variables from the 66 early respondents and the36 late respondents are no significant differences (p<0.05).

Reliability and validity

The validity was examined in terms of content validity, con-vergent validity, and discriminant validity. Content validitywas established from the extant literature, and a pilot testwas performed to improve the validity of the measures. Aconfirmatory factor analysis (CFA) was conducted in order toacquire evidence of convergent and discriminant validity. TheAMOS 18 software with maximum likelihood estimation wasused to perform the CFA. Results of Mardia’s test confirmedthat the data deviated from multivariate normality1.Convergent validity is demonstrated when indicator factorloadings (λ) are significant and exceed the acceptable valueof 0.5 on their corresponding constructs as recommended by[49], and the average variances extracted (AVE) of the constructare larger than 0.5, exceeding the threshold value suggested by[50]. All λ values in the CFA model exceeded 0.5 on theircorresponding constructs and the loadings within con-struct are higher than those across construct (AppendixB), and the AVE for all constructs exceeded the thresh-old value of 0.5; thus, convergent validity was con-firmed. Discriminant validity is demonstrated when thesquare root of the AVE is greater than the inter-constructcorrelations, as suggested by [50]. Table 5 shows that thesquare root of the AVE values is greater than the inter-correlations and thus exhibits acceptable discriminant validity.

Finally, construct reliability was assessed in terms of com-posite reliability and Cronbach alpha value. Table 5 showsthat all composite reliabilities exceeded the minimal reliabilitycriteria of 0.7 recommended by [50]. All Cronbach alphavalues also exceeded 0.60, which is in the acceptablerange [50].

Table 3 Hospitals adoption RFID phase and current status

Adoption Category of adoption status N (%)

Adopter hospitals Already implementedand used RFID

28 (27.5 %)

Started using RFID 9 (8.8 %)

Non-adopter hospitals Planned to use 8 (7.8 %)

Non-adopt 57 (55.9 %)

Overall 102 (100.0 %)

1 The results showed that the Mardia coefficients were 196.298<2915(53×55), the results of Mardia’s test confirmed that the data deviatedfrom multivariate normality.

J Med Syst (2014) 38:3 Page 7 of 17, 3

Hypotheses testing and results

The discriminant analysis was used for hypotheses testing.The hospitals were classified into two groups: 37 adopters and

65 non-adopters. There were only two groups being used,hence displaying only one function. This function providedan index of overall model fit, which was then interpreted as theproportion of variance explained (R2). A canonical correlation

Table 4 Respondent characteris-tics (N=102) Items Categories Frequency Percentage (%)

Hospital Level Medical center 10 9.8

Regional hospital 25 24.5

District hospital 67 65.7

Ownership Type Public hospital 5 4.9

Private hospital 58 56.9

Corporate Hospital 38 37.2

Privately managed public hospital 1 1.0

Number of Hospital Beds Less than 200 beds 31 30.4

201 to 700 beds 33 32.4

701 to 2000 beds 35 34.3

2001 to 4000 beds 3 2.9

Respondents’ Position Director of IS department 44 43.1

Deputy director of IS department 40 39.2

Others 18 17.7

Respondents’ Gender Male 68 66.7

Female 34 33.3

Respondents’ Age Less than 30 18 17.6

31 to 40 28 27.5

41 to 50 53 52.0

Over 50 3 2.9

Length of Service at Current Position Less than 5 years 16 15.7

6 to 10 years 13 12.7

11 to 15 years 48 47.1

16 to 20 years 9 8.8

Over 21 years 16 15.7

Table 5 Discriminant validity and reliability

Variables CR CA 1 2 3 4 5 6 7 8 9 10 11 12 13

1. COST 0.95 0.95 0.93

2. BEN 0.91 0.91 0.53 0.88

3. UBIQ 0.92 0.91 0.67 0.68 0.88

4. COMPLEX 0.82 0.80 −0.28 −0.32 −0.24 0.78

5. COMPAT 0.95 0.95 0.69 0.62 0.62 −0.33 0.91

6. RISK 0.81 0.77 −0.56 −0.33 −0.40 0.36 −0.68 0.83

7. TOPSUP 0.97 0.97 0.69 0.51 0.69 −0.19 0.65 −0.55 0.94

8. FINAN 0.94 0.94 0.53 0.38 0.55 −0.18 0.44 −0.35 0.46 0.94

9. TECH 0.77 0.76 −0.06 −0.04 0.08 0.34 −0.16 0.39 −0.13 0.03 0.73

10. USERSP 0.91 0.91 0.54 0.44 0.50 −0.17 0.54 −0.40 0.50 0.37 0.08 0.92

11. GOVPOL 0.92 0.91 0.62 0.56 0.49 −0.13 0.67 −0.44 0.65 0.32 −0.005 0.47 0.89

12. EXSUP 0.95 0.95 0.62 0.57 0.58 −0.27 0.60 −0.48 0.56 0.44 −0.03 0.53 0.55 0.87

13. MARKET 0.72 0.65 0.51 0.48 0.69 −0.18 0.58 −0.38 0.57 0.49 0.01 0.50 0.50 0.50 0.76

See Appendix B for abbreviations used in Tables 5 and 6

CR composite reliability, CACronbach’s alpha. The bold numbers in the diagonal row are square roots of the Average Variances Extracted (AVE)

3, Page 8 of 17 J Med Syst (2014) 38:3

of 0.906 suggested that the model may explain 82.08 % of thevariation in the grouping variable, whether a respondent wasadopted or not. The Wilks’ lambda value of the dis-criminant function was 0.192 (χ2=153.683, df=14, p=0.000). Wilks’ lambda was used to determine whetherthere was any difference existing between the groups. Asmaller Wilks’ lambda value indicated the higher discrimi-nation power [49]. The result demonstrated that the Wilks’lambda value was 0.1192 (p=0.000), meaning this discrimi-nant function could be used to discriminate between theadopters and the non-adopters.

Three key indicators used in discriminant analysis includeddiscriminant loadings (or structure correlations), standardizeddiscriminant coefficients (or discriminant weight) and partialF values.Discriminant loadings reflected the variance that theindependent variables shared with the discriminant function.Standardized discriminant coefficient reflected the relativecontribution of its associated variables to the discriminantfunction. The F values showed the associated significance ofeach variable and larger F values indicating that it would havegreater discriminatory power [49].

The discriminant loadings for each variable are shown inTable 6. The discriminant loadings were used to measure thesignificance of the variables and the cutoff value was greaterthan 0.3, as suggested by [49]. The variables exhibiting adiscriminant loading greater than 0.3 included costs, ubiquity,compatibility, security and privacy risks, top managementsupport, hospital scale, and financial readiness. These sevenvariables also had a high significance level and thus substan-tially influenced the adoption of RFID. The discriminantloading of government policy was less than 0.3, though thestandardized discriminant coefficient was higher than 0.3.Therefore, it could also be considered as a predictive variableof effective difference. While ranking the relative importanceof the factors based on F value, the list was as follows: securityand privacy risks, hospital scale, investing costs, top manage-ment support, ubiquity, compatibility, financial readiness, andgovernment policy. To summarize, the results of discriminantanalysis supported the hypotheses H1, H3, H5, H6, H7, H8,H9, and H12 (see Table 7). The percentages of classificationaccuracy of the discriminant function for the adopters andnon-adopters groups were 97.3 % and 96.9 % respectively.The overall classification accuracy was estimated to be97.1 %. The histograms in Fig. 2 also showed that the dis-criminant function did well.

Follow-up in-depth interviews

We invite three IS department directors from three caseshospitals which have already implemented and used RFIDfor our follow-up interviews after the field survey. Each inter-view lasted between 60 and 120 min and its full content has

been recorded with the interviewees’ advanced permission.Table 8 summarizes the results of these interviews.

Discussion and implications

The objective of this study was to find out what promotes theadoption of RFID in hospitals. When discussing the results ofthis study, we compared our research findings with those ofChang & Chan [8] and Lee & Shim [7] (See Table 9). Both oftheir research topics and contexts are similar to ours, though stillhave some differences. In terms of sample source, Chong &Chan [8] collected data from 183 health companies and hospitals,while we collected data from a total of 102 hospitals. On con-siderations of RFID’s adoption, hospitals’ are different fromsystem suppliers’. In most of the industries suppliers’ RFIDadoption would create vender pressure, forcing them into usingthe same system. In reverse, if a vender is powerful enough andlarge in scale, its RFID adoptionwould create a push force for thesuppliers’ to adopt the system, such as in the case of Walmart. Itis therefore necessary to look at both sides, ie supplier andvender, when exploring factors influencing an industry’s adop-tion of RFID. Despite this, given the current situation of thehealthcare industry, and as our field survey and interviews reveal,the reasons behind a hospital’s decision to adopt RFID aremostlyfrom within itself, such as the hospital’s own technological andorganizational considerations. This research based its sampleonly in hospitals, since it intends only to look at the adoptionof RFID in hospitals. In addition, Lee & Shim [7] focused on themotivations and driving forces behind the adoption of RFID inthe healthcare industry by using the theory of technology-pushand need-pull. Our research provides a more comprehensivemodel that integrates technological, organizational and environ-ment factors in order to understand hospitals considerationsregarding RFID adoption.

Key findings and insights

The major findings from this study are as follows. First, costas a key factor in hospitals’ considerations of RFID adoption(discriminant loading = −0.402, Wilks’ lambda=0.594, F=68.317 and p=0.000). This is consistent with previous re-search, which revealed that the cost was the major constrainton the widespread use of RFID technologies [18]. The ISdepartment director from case hospital A indicated in thefollow-up interview: “the cost of RFID tags is very high.Combining that with the costs of RFID transmitters, readers,encoders and antenna equipments, the initial investment costsare very huge. The patient identification is a relatively lowercost investment due to the use of passive RFID tags. However,the equipment managements, e.g., x-ray equipments, mobilenursing stations, ECG and ultrasound machines, all havehigher relative investment costs due to the use of active

J Med Syst (2014) 38:3 Page 9 of 17, 3

RFID tags. Therefore, it is an extremely high investment costfor the hospital.” And the respondents of hospital B stated:“using passive RFID tags makes the investment cost of accesscontrol management cheaper than that using active RFID tags,meaning it can be more widely used.”

Second, contrary to our expectations, hypothesis 2 (perceivedbenefit factor) was not supported (discriminant loading =

−0.207). Krasnova et al. [29] proposed that RFID technologyhad never been an end in itself, since the benefits of RFID wastoo low or absent compared to its investment costs. In our follow-up interviews, the IS department director from case hospital Aindicated that “a hospital must be operated in the most cost-effective way, since a hospital’s income is not high enough. ROI(return on investment) is key factor. If an investment and its

Table 7 The results of hypotheses testing

Variables Discriminant loading Standardized discriminant coefficient F values Ranka Results

Cost (H1) −0.402 −0.137 68.317 3 Support

Perceived benefits (H2) −0.207 −0.115 17.999 – No support

Ubiquity (H3) −0.349 −0.308 51.541 5 Support

Complexity (H4) 0.094 −0.068 3.695 – No support

Compatibility (H5) −0.347 −0.104 50.826 6 Support

Security and privacy risk (H6) 0.468 0.434 92.301 1 Support

Top management support (H7) −0.378 −0.415 60.435 4 Support

Hospital scale (H8) 0.463 0.827 90.505 2 Support

Financial readiness (H9) −0.326 −0.184 44.727 7 Support

Technological readiness (H10) 0.126 0.081 6.652 – No support

User support (H11) −0.238 −0.067 23.99 – No support

Government policy (H12) −0.179 0.302 13.528 8 Support

External support (H13) −0.269 −0.196 30.646 – No support

Market uncertainty (H14) −0.256 0.174 27.557 – No support

a means there is not enough discrimination power

Table 6 Discriminant analysis

Variables Discriminantloading

Standardizeddiscriminant coefficient

Wilks’ lambda F values Significance level Adopted Non−adopt

Mean S.D Mean S.D.

COST −0.402 −0.137 0.594 68.317 0.000 2.29 0.61 3.74 0.96

BEN −0.207 −0.115 0.847 17.999 0.000 2.31 0.55 2.99 0.89

UBIQ −0.349 −0.308 0.660 51.541 0.000 2.13 0.58 3.10 0.70

COMPLEX 0.094 −0.068 0.964 3.695 0.057 3.08 0.51 2.85 0.63

COMPAT −0.347 −0.104 0.663 50.826 0.000 2.20 0.52 3.43 0.98

RISK 0.468 0.434 0.520 92.301 0.000 3.59 0.54 2.57 0.51

TOPSUP −0.378 −0.415 0.623 60.435 0.000 2.19 0.86 3.61 0.90

SCALE 0.463 0.827 0.525 90.505 0.000 2.89 0.57 1.65 0.67

FINAN −0.326 −0.184 0.691 44.727 0.000 1.85 0.42 2.74 0.74

TECH 0.126 0.081 0.938 6.652 0.011 3.12 0.47 2.81 0.63

USERSP −0.238 −0.067 0.807 23.990 0.000 2.58 1.24 3.62 0.88

GOVPOL −0.179 0.302 0.881 13.528 0.000 2.86 0.75 3.49 0.88

EXSUP −0.269 −0.196 0.765 30.646 0.000 2.47 0.87 3.55 0.99

MARKET −0.256 0.174 0.784 27.557 0.000 2.78 0.81 3.65 0.79

Classification accuracy

Predicted adopter Predicted non-adopter Total

Actual adopter 36 (97.3 %) 1(2.7 %) 37

Actual non-adopter 2 (3.1 %) 63 (96.9 %) 65

Overall accuracy 97.1 %

3, Page 10 of 17 J Med Syst (2014) 38:3

returns cannot be balanced, you will not want to implementRFID at all.” Similarly, the IS department director from casehospital B pointed out: “it seems very hard to get very noticeablebenefits.”However, the IS department director from case hospitalC expressed: “RFID technology has really changed the workflowwithin our hospital. For instance, the hospital’s escort system canmanage the staff locationsmore effectively and has a better controlover patients’waiting time. It can improve the service quality, buteventually it is impossible to reduce the overhead costs.”

Third, contrary to expectations, hypothesis 4 (complexityfactor) was non-supported (discriminant loading=0.094). The

IS department director from case hospital A indicated: “as forthe technology, a team of professors at a university togetherwith the vendors have provided full assistance for our hospital,so that we did not encounter any problems at all.” The ISdepartment director from case hospital C pointed out: “in theprocess of introducing RFID technology the development oftechnical systems was actually the easiest part, but integratingRFID into HIS was not an easy job.”

Fourth, as expected, security and privacy risks were confirmedin this research to be the key factors regarding hospitals’ consid-erations of RFID adoption (discriminant loading=0.468,

Adopter

Predicted Group

Non -adopter

Fig. 2 Histograms showing thedistribution of discriminant scoresfor adopter hospitals and non-adopter hospitals

Table 8 The results of the IS department directors’ interviews

Case Hospital Hospital A Hospital B Hospital C

Hospital level & ownershiptype

Regional & Public hospitals District & Corporate hospitals Medical center & Corporatehospitals

Number of beds 600 280 1600

Number of employees 1100 440 4100

RFID implementation date 2007 2006 2007

RFID applications & scope Access control management, Drugmanagement, Medical equipmentsmanagement, Patient identification

Access control management, Patients’exercise time in the Health PromotionCenter, Meeting attendance

Access control management,Patient escort, Specimendelivery

System Integration HIS HIS HIS

Project leader Director of IS Director of IS Director of IS

Key factors in high-levelmanagers’ considerationsof RFID adoption

Cost, Perceived benefit, Ubiquity, Topmanagement support, Governmentpolicy

Hospital scale, Compatibility, Financialreadiness, User support, Security &privacy risk

Compatibility, Ubiquity,Security & privacy risk, Topmanagement support

Healthcare information system (HIS)

J Med Syst (2014) 38:3 Page 11 of 17, 3

discriminant coefficient=0.434, Wilks’ lambda=0.520, F=92.301 and p=0.000). Patient privacy is becoming a matter ofgreat concern when using RFID tags that send out informationwithout the knowledge of the tagged entity. The relevant devel-opments in RFID cryptography was important [9]. The IS de-partment director from case hospital B pointed out: “RFID helpsmanage patient data more effectively, although it also increasesthe risks of data loss, contributing to patients’ resistance. Hence,patient data should be encrypted in order to protect individualprivacy.” The IS department director from case hospital C indi-cated: “RFID can track the locations of both the escort personneland patients all the time, which may cause their resistance to betagged.”

Fifth, contrary to expectations, this research found hypothe-sis 10 (technological readiness factor) non-supported (discrim-inant loading=0.126). The IS department director from casehospital A indicated: “with the assistance from schools, ITvendors and the Ministry of Economic Affairs, we experiencedno technical difficulties. Additionally, thanks to the hospital’sown IT Department, there were few technical problems duringthe process of RFID applications”. The IS department directorfrom case hospital B stated: “with the vendors’ assistance, thetechnology was not a big problem.” In addition, The IS depart-ment director from case hospital C indicated: “since the escortsystem was developed by our Information Department andmanufacturers, we were able to enjoy full technical supports.”

Sixth, contrary to expectations, hypothesis 11 (user supportfactor) was found non-supported in this study (discriminantloading = −0.238). The possible reasons are top managementrequest, meantime, promoting RFID was a hospital policy. Inaddition, the other possible reason is healthcare industry char-acteristic. Medical staff could not choose whether to accept it ornot under a hospital policy and top management request.Doctors enjoy a higher degree of independence, in comparisonto other medical staff that are lower-ranked in a hospital’s chain

of command and trained to follow order either willingly or as acompliancewith policies. That is, theywill use and accept a newsystem if hospital rules dictate. Therefore, hypothesis 11 wasfound non-supported in this research. In our follow-up inter-views, the IS department director from case hospital B indicated:“to improve our service quality is of a high priority, and to meetthis demand medical staff have to cooperate with hospitals’RFID policy.” The IS department director from case hospitalC pointed out: “due to the changes in workflow processescaused by the introduction of RFID, medical staff were resistantat first, but eventually became more acceptant when feeling thebenefits of optimized processes and shorter patient escort time.”

Seventh, contrary to expectations, hypothesis 13 (externalsupport factor) was found in this research non-supported(discriminant loading = −0.269). During our follow-up inter-views, the IS department director from case hospital A pointedout “at the initial stage, we were cooperating with theoutsourcing partners, but after a period of time, our ITDepartment took charge of the subsequent operations.” TheIS department director from case hospital C indicated: “afterimplementing RFID, our IT Department was capable of man-aging the subsequent operations.”

Finally, contrary to expectations, this research found hypoth-esis 14 (market uncertainty) non-supported (discriminant load-ing = −0.256). The IS department director from case hospital Apointed out in the follow-up interview that “when we found outthat our competitors are using RFID, we were, of course,motivated to adopt the technology, though we still need toconsider the financial conditions of the hospital itself beforedetermine whether the benefits outweigh the costs.” Also, theIS department director from case hospital B indicated: “depend-ing on the scale, hospitals have different business strategies thataffect the degree of urgency involving the RFID adoption. Ourcompetitors’ decisions to adopt RFID did not affect ourdecision-making at all.”

Table 9 The comparison of this study and two recent researches relating to RFID adoption in hospitals

Study Sample source & researchcontext

Theoreticalfoundation

Consideration of the RFID adoption

This study 102 hospitals in Taiwan TOE Technology: costa, perceived benefits, ubiquitya, complexity,compatibilitya, security and privacy riska

Organization: top management supporta, hospital scalea, financialreadinessa, technological readiness, user support

Environment: government policya, external support, market uncertainty

Chong & Chan [8] 183 health companies andhospitals in Malaysian

TOE Technology: relative advantage, compatibility, complexitya, costa, securitya

Organization: top management supporta, organization sizea, financialresources, technological knowledgea

Environment: competitive pressurea, expectation of market trendsa

Lee & Shim [7] 126 hospitals in the U.S. technology-pushand need-pull

Technology: perceived benefitsa

Organization: performance gapa, presence of championsa

Environment: vendor pressure, market uncertaintya

a means that hypothesis is supported

3, Page 12 of 17 J Med Syst (2014) 38:3

Academic implications

The academic contributions of this study are mainly three folds.First, we developed a more comprehensive model regardinghospitals’ considerations of RFID adoption based on a well-known TOE framework, experts’ opinions and the results froma preliminary survey, a field survey and follow-up in-depthinterviews. Since every industry is different in its own way, priorstudies on other industries were not able to fully explain thesituation in the healthcare industry. This study fills in the researchgap. Second, this research contributes to the information systemresearches and healthcare researches by confirming several deci-sion factors of the RFID adoption in hospitals, which better theunderstanding of why a hospital decide to adopt RFID or not.Third, our research findings indicated that RFID’s characteristicsand organizational dimension, instead of the environmental char-acteristics, have the relative importance in RFID adoption. Themost importance factor of RFID characteristics is the concernover privacy and security. Consistent with prior research, themost common fear referred to the misuse of data collected byRFID tags, resulting in an undesirable intrusion into the privacyrisk of individuals [28]. This study contributes to the currentunderstanding of the influence on privacy and securityconcerning the adoption of RFID in hospitals by theoreticalanalysis and empirical testing. Hospital scale as the most impor-tant organizational factors, this study looked at a number ofhospitals to see the reasons behind their adoption or non-adoption of RFID.

Practical implications

The practical contributions of this study are four folds. First, theempirical finding of this study showed that RFID in hospitals isstill in an early adoption stage. Of the 102 hospitals surveyed,only 37 hospitals are currently using RFID and 8 hospitals areplanning to use. Hospitals can refer to the findings for theirdecisions over RFID. Second, this study showed that technolog-ical factors, such as security and privacy concerns, costs, ubiq-uity, and compatibility are key considerations regarding RFIDadoption in hospitals. This is especially important in the contextof hospitals where workflow processes are complex and change-able, thus, RFID characteristics is an important factor affectingthe RFID adoption. RFID vendors may need to consider analyz-ing and demonstrating technological benefits based on the expe-rience from other precedent hospital adoption. This will givenon-adopters a better picture of how the technology can benefitthem and encourage their decision to consider RFID adoption.Third, this study also found that hospital scale, top managementsupport and financial readiness are the important factors inhospitals’ decisions regarding RFID adoption. The research find-ings suggest that, to promote the RFID adoption in the hospitals,the financial support is a key factor on the initial adoption in thehospitals. Top management support was shown to be crucial.

Besides, RFID vendors should know that the hospital scale candetermine howmuch it can benefit from adopting the technology,and therefore their decision to do so. Finally, this study showedthat hospitals’ consideration regarding RFID adoption is relyhighly on government policy and grant support. Education train-ing and funding from the government play a critical factor at theinitial stage of RFID adoption. To increase the pace of adoptionand to ensure a continued usage of the adopted technology, thegovernment should communicate the advantages of RFID tohospitals properly and provide necessary supports along the way.

Conclusion and limitations

The goal of this study is to identify the salient determinants ofRFID adoption in the hospital environment. The TOE frameworkderived from the literature related to the RFID adoption, andpreliminary surveys were used to theorize a model for the RFIDadoption in hospitals. The data collected from a field survey of102 hospitals provided an empirical support for the proposedmodel. Our follow-up interviews with senior managers furtheredthe understanding of such complex and dynamic aspects regard-ing the RFID adoption motivations. The findings of this studyenabled a better understanding of factors affecting hospitals’decision on RFID adoption and their relative importance.

There are several limitations in this study. First, the researchsample was collected solely from Taiwan’s hospitals, meaningthat no cultural difference was involved. For this reason, it maylimit the generalizability of our findings. Second, this studyadopted a discriminant analysis to validate its hypotheses.However, the number of participating hospitals, which haveadopted the RFID technologies, was low, mainly because thehealthcare industry was relatively slow in adopting the IT [51].Third, the result of an organization promoting a new technologyis not always “good”. Future researches can try to link a hospi-tal’s decision-making process to the outcomes of such a decisionin term RFID adoption. Finally, we hope this study can contrib-ute to the initial adoption research and help draw attention to it.For future studies there is a variety of post-adoption behaviorssuch as the expansion or long-term continuance of RFID adop-tion that is worthy of examination.

Acknowledgments This study was supported by the Chienkuo Tech-nology University, Taiwan, R.O.C., under Grant No. CTU-101-RP-IM-003-015-A. We are thankful for all participators, in particular Dr. Tien-Cheng Hsu, director of the IS Department and Mr. Tzu-Chia Huangsystem engineer for the IS Department, from Changhua Christia MedcialCenter. The authors would also like to thank the anonymous reviewersand editors for their constructive comments.

Conflict of interest All authors have no financial or non-financialinterests that may be relevant to the submitted work. There is withoutany possibility of favoritism or personal gain conducted via this study. So,the authors declare they have no conflict of interests.

J Med Syst (2014) 38:3 Page 13 of 17, 3

Appendix A

Appendix B

Table 10 Preliminary survey items

RFID characteristics 1. Which of the following characteristics of RFID is the key to RFID adoption decision making? (multiple choices)

• Perceived benefits: Hospitals adopt RFID because it can bring relative advantage against existing information technology.

• Compatibility: Hospitals adopt RFID because it matches existing value, need and past experience.

• Complexity: Hospitals do not adopt RFID because it is difficult to use.

• Cost: Hospitals do not adopt RFID because of the cost associated with RFID adoption including cost of tags, readers,installation, system integration, education and training, implementation, development and operation.

• Ubiquity: Hospitals adopt RFID because it provides personalized and continuous connection and communication.

• Work-related: Hospitals adopt RFID because it is appropriate to use it at work.

2. Besides the aforementioned factors, what are other characteristics of RFID that would affect your hospital’s RFID adoptiondecision making? (open-ended question)

Organizationaldimension

1. Which of the following characteristics of RFID is the key to RFID adoption decision making? (multiple choices)

• Performance gap: Hospitals adopt RFID because there are gaps in performance and satisfaction with respect to informationtechnology.

• Financial readiness: Hospitals adopt RFID because hospitals have financial resources to support RFID purchases,implementation and maintenance.

• Technological readiness: Hospitals adopt RFID because hospitals have high maturity with respect to the use andmanagement of information technology.

• Hospital scale: Hospitals adopt RFID because there are more medical personnel and hospital beds.

• Top management support: Hospitals adopt RFID because of support and commitment from top management.

• Presence of champions: Hospitals adopt RFID because management realizes the usefulness of adopting innovativeinformation technology and provides necessary authorization and resources during development and implementation.

2. Besides the aforementioned factors, what are other characteristics of RFID that would affect your hospital’s RFID adoptiondecision making? (open-ended question)

Environmentaldimension

1. Which of the following characteristics of RFID is the key to RFID adoption decision making? (multiple choices)

• Vender pressure: Hospitals’ motivation to adopt RFID comes from vender pressure.

• Market uncertainty/competitive pressure: Hospitals’ motivation to adopt RFID comes from pressure from external market.

• Standards uncertainty: Hospitals’ obstacle to adopting RFID comes from lack of standards setting.

• External support: Hospitals adopt RFID because of the support received, such as support from suppliers and consultants.

• Government policies: Hospitals adopt RFID because the government provides financial support, training curriculum andpolicy descriptions and stability of government policies.

• Information intensity: Hospitals adopt RFID because hospitals highly rely on information.

2. Besides the aforementioned factors, what are other characteristics of RFID that would affect your hospital’s RFID adoptiondecision making? (open-ended question)

Table 11 Constructs and items

Item Question Factor loadinga

Cost [23, 52]

COST1 The costs of adoption of RFID are far greater than the expected benefits. 0.937

COST2 The costs of maintenance and supports for RFID are very high for our hospital. 0.960

COST3 The amount of money and time invested in training employees to use RFID are very high. 0.883

Perceived benefits [7]

BEN1 RFID overhead costs will be reduced. 0.845

3, Page 14 of 17 J Med Syst (2014) 38:3

Table 11 (continued)

Item Question Factor loadinga

BEN2 RFID will improve the customer service. 0.905

BEN3 RFID will improve the hospital image and expertise. 0.896

Ubiquity [10]

UBIQ1 RFID provides our hospital with communication and connectivity at anytime-and-anywhere. 0.922

UBIQ2 The communication and information accessibility in anytime-and-anywhere provided by RFIDis highly critical for the effectiveness our hospital.

0.925

UBIQ3 My hospital requires personalized and uninterrupted connection and communication. 0.803

Complexity [23, 26]

COMPLEX1 The skills required to use RFID are too complex for the most of our employees. 0.776

COMPLEX2 Integrating RFID into our current work practices is very difficult. 0.922

COMPLEX3 RFID may encounter little or no harmonization between standards, e.g. due to the lack of unifiedstandards for RFID that may increase the complexity of relevant applications or operations.

0.619

Compatibility [53]

COMPAT1 Using RFID technology is compatible with all aspects of my works. 0.866

COMPAT2 Using a RFID technology is completely compatible with my current situations. 0.883

COMPAT3 I think that using RFID technology will fit well with the way I work. 0.919

COMPAT4 Using RFID technology fits into my work style. 0.902

Security and privacy risk (Developed based on Cases [54])

RISK1 Use of RFID may cause my personal information to be stolen. N/A

RISK2 I do not think it is safe to use RFID because of the privacy and security concerns. 0.920

RISK3 I have doubts about the data security of RFID applications. 0.683

Top management support [23]

TOPSUP1 The top management enthusiastically supports the RFID adoption. 0.958

TOPSUP2 The top management has allocated adequate resources to the RFID adoption. 0.948

TOPSUP3 Top management is aware of the benefits of RFID adoption. 0.952

TOPSUP4 Top management actively encourages employees to use RFID technologies in their daily activities. 0.942

Hospital scale

SCALE Number of beds in the hospital.

Financial readiness (Developed based on [7])

FINAN1 Our hospital has the financial resources for adopting RFID. 0.988

FINAN2 The overall information systems budgets are significant enough to support the development andimplementation of RFID applications.

0.896

Technological readiness [10]

TECH1 We use RFID because we know the technology. 0.607

TECH2 We have the technical knowledge and skills to implement RFID. 0.913

TECH3 We know how to integrate RFID with the existing systems of our hospital. 0.635

User support [38]

USERSP1 Employees (and patients) are enthusiastic about the RFID adoption. 0.805

USERSP2 Employees (and patients) have a negative opinion about the RFID adoption. (Reverse) N/A

USERSP3 Employees (and patients) are ready to accept the changes caused by the RFID adoption. 0.915

Government policy [1]

GOVPOL1 Financial aid for the installation will be provided by the government. 0.851

GOVPOL2 Training courses will be provided by the government. 0.821

GOVPOL3 Specification and stability of government policies. 0.989

External support [42]

EXSUP1 RFID suppliers will offer adequate technical supports after the implementation of RFID applications. 0.955

EXSUP2 High quality of technical supports will be provided by the RFID suppliers. 0.965

EXSUP3 High quality of training programs will be provided by the RFID suppliers. 0.950

EXSUP4 Effectiveness in performing information requirements analysis will be provided by the consultants. 0.894

EXSUP5 Effectiveness in recommending suitable solutions will be provided by the consultants. 0.888

J Med Syst (2014) 38:3 Page 15 of 17, 3

References

1. Chang, I. C., Hwang, H. G., Yen, D. C., and Lian, J. W., Criticalfactors for adopting PACS in Taiwan: Views of radiology departmentdirectors. Decis. Support. Syst. 42(2):1042–1053, 2006.

2. Fichman, R. G., Kohli, R., and Krishnan, R., The role of informationsystems in healthcare: Current research and future trends. Inf. Syst.Res. 22(3):419–428, 2011.

3. Lin, C., Lin, I. C., and Roan, J., Barriers to physicians’ adoption ofhealthcare information technology: An empirical study on multiplehospitals. J. Med. Syst. 36(3):1965–1977, 2012.

4. Yao, W., Chu, C. H., and Li, Z., The adoption and implementation ofRFID technologies in healthcare: A literature review. J. Med. Syst.36(6):3507–3525, 2012.

5. Finch, C., Mobile computing in healthcare. Health ManagementTechnology 20(3):64–65, 1999.

6. Qu, X., Simpson, L. T., and Stanfield, P., A model for quantifying thevalue of RFID-enabled equipment tracking in hospitals. Adv. Eng.Inform. 25(1):23–31, 2011.

7. Lee, C. P., and Shim, J. P., An exploratory study of radio frequencyidentification (RFID) adoption in the healthcare industry. Eur. J. Inf.Syst. 16(6):712–724, 2007.

8. Chong, A. Y. L., and Chan, F. T. S., Structural equation modelling formulti-stage analysis on radio frequency identification (RFID) diffusionin the health care industry. Expert Syst. Appl.39(10):8645–8654, 2012.

9. Zhou, W., and Piramuthu, S., Framework, strategy and evaluation ofhealth care processes with RFID. Decis. Support. Syst. 50(1):222–233, 2010.

10. Kim, S., and Garrison, G., Understanding users’ behaviors regardingsupply chain technology: Determinants impacting the adoption andimplementation of RFID technology in South Korea. Int. J. Inf.Manag. 30(5):388–398, 2010.

11. Wang, Y. M., Wang, Y. S., and Yang, Y. F., Understanding thedeterminants of RFID adoption in the manufacturing industry.Technol. Forecast. Soc. Chang. 77(5):803–815, 2010.

12. Scott Morton, M. S., The corporation of the 1990s: Informationtechnology and organizational transformation. Oxford UniversityPress, New York, 1991.

13. Grover, V., and Goslar, M. D., The initiation, adoption and imple-mentation of telecommunications technologies in U.S.Organizations. J. Manag. Inf. Syst. 10(1):141–163, 1993.

14. Pavlou, P. A., and Fygenson, M., Understanding and predictingelectronic commerce adoption: An extension of the theory of plannedbehavior. MIS Q. 30(1):115–143, 2006.

15. Konsynski, B., and Smith, H. A., Developments in practice x: Radiofrequency identification (RFID) - an internet for physical objects.Commun. Assoc. Inf. Syst. 12(1):301–311, 2003.

16. Bingi, P., Sharma, M. K., and Godla, J. K., Critical issues affecting anERP implementation. Inf. Syst. Manag. 16(3):7–14, 1999.

17. Markus, M. L., Power, politics, and MIS implementation. Commun.ACM 26(2):430–444, 1983.

18. Roberts, C.M., Radio frequency identification (RFID).Computers &Security 25(1):18–26, 2006.

19. Teo, T. S., Lin, S., and Lai, K. H., Adopters and non-adopters of e-procurement in Singapore: An empirical study. Omega-InternationalJournal of Management Science 37(5):972–987, 2009.

20. Chang, S. I., Hung, S. Y., Yen, D. C., and Chen, Y. J., The determi-nants of RFID adoption in the logistics industry - a supply chainmanagement perspective. Commun. Assoc. Inf. Syst. 23(1):197–218,2008.

21. Rogers, E. M., Diffusion of innovations, 3rd edition. The Free Press,New York, 1983.

22. Sharma, A., and Citurs, A., RFID adoption drivers: A radical inno-vation adoption perspective. Proceedings of the 11th AmericasConference on Information Systems (AMCIS), Omaha, 2005.

23. Premkumar, G., and Roberts, M., Adoption of new informationtechnologies in rural small businesses. Omega-InternationalJournal of Management Science 27(4):467–484, 1999.

24. Bardram, J. E., and Christensen, H. B., Pervasive computing supportfor hospitals: An overview of the activity-based computing project.IEEE Pervasive Computing 6(1):44–51, 2007.

25. Varshney, U., Pervasive healthcare: Applications, challenges andwireless solutions. Commun. Assoc. Inf. Syst. 16(1):52–72, 2005.

26. Tsai, M. C., Lee, W., and Wu, S. C., Determinants of RFID adoptionintention: Evidence from Taiwanese retail chains. Information &Management 47(5–6):255–261, 2010.

27. Ngai, E. W. T., Cheng, T. C. E., Au, S., and Lai, K. H., Mobilecommerce integrated with RFID technology in a container depot.Decis. Support. Syst. 43(1):62–76, 2007.

28. Thiesse, F., RFID, privacy and the perception of risk: A strategicframework. J. Strateg. Inf. Syst. 16(2):214–232, 2007.

29. Krasnova, H., Weser, L., and Ivantysynova, L., Drivers of RFIDadoption in the automotive industry. Proceedings of the AMCIS(American Conference on Information Systems), Toronto, 2008.

30. Lian, J.W., Yen, D.C., Wang, Y.T., An exploratory study to under-stand the critical factors affecting the decision to adopt cloud com-puting in Taiwan hospital. International Journal of InformationManagementAvailable online 2013

31. Weill, P., The relationship between investment in information tech-nology and firm performance: A study of the valve manufacturingsector. Inf. Syst. Res. 3(4):307–333, 1992.

32. Chang, I. C., Hwang, H. G., Hung, M. C., Lin,M. H., and Yen, D. C.,Factors affecting the adoption of electronic signature: Executives’perspective of hospital information department. Decis. Support. Syst.44(1):350–359, 2007.

33. Hung, S. Y., Hung, W. H., Tsai, C. A., and Jiang, S. C., Criticalfactors of hospital adoption on CRM system: Organizational andinformation system perspectives. Decis. Support. Syst. 48(4):592–603, 2010.

Table 11 (continued)

Item Question Factor loadinga

EXSUP6 Effectiveness in managing RFID implementation will be provided by the consultants. 0.869

Market uncertainty [7]

MARKET1 The competition among hospitals is very intense. 0.766

MARKET2 The frequency of cost-increase in the healthcare industry. 0.685

a Factor loadings are obtained from confirmatory factor analysis (CFA). RISK1 and USERSP2 were dropped due to poor loadings in the factor analysis

3, Page 16 of 17 J Med Syst (2014) 38:3

34. Madlberger, M., A model of antecedents of RFID adoption intentionin the supply chain. Proceedings of the 42nd Hawaii InternationalConference on System Sciences, Waikoloa, pp. 1–10, 2009.

35. Thong, J. Y. L., An integrated model of information system adoptionin small business. J. Manag. Inf. Syst. 15(4):187–214, 1999.

36. Cragg, P. B., and King, M., Small-firm computing: Motivators andinhibitors. MIS Q. 17(1):47–60, 1993.

37. Iacovou, C. L., Benbasat, I., and Dexter, A. S., Electronic datainterchange and small organizations: Adoption and impact of tech-nology. MIS Q. 19(4):465–485, 1995.

38. Wang, E. T. G., and Chen, J. H. F., Effects of internal support andconsultant quality on the consulting process and ERP system quality.Decis. Support. Syst. 42(2):1029–1041, 2006.

39. McLachlin, R. D., Factors for consulting engagement success.Manag. Decis. 37(5):394–402, 1999.

40. Newpeck, F. F., and Hallbauer, R. C., Some advice for the smallbusiness considering computer acquisition. J. Small Bus. Manag.19(3):17–23, 1981.

41. Hossain, M.A., Quaddus, M., Impact of external environmentalfactors on RFID adoption in Australian livestock industry: An ex-ploratory study, in Proceedings of the Pacific Asia Conference onInformation Systems (PACIS), 2010. pp 1735–1742.

42. Thong, J. Y. L., Yap, C. S., and Raman, K. S., Top managementsupport, external expertise and information systems implementationin small businesses. Inf. Syst. Res. 7(2):248–267, 1996.

43. Senn, J. A., and Gibson, V. R., Risks of investment in microcom-puters for small business management. J. Small Bus. Manag. 19(3):24–32, 1981.

44. Thong, J. Y. L., and Yap, C. S., CEO characteristics, organizationalcharacteristics and information technology adoption in small busi-ness. Omega-International Journal of Management Science 23(4):429–442, 1995.

45. Zhu, Y., Li, Y., Wang, W., and Chen, J., What leads to post-implementation success of ERP? An empirical study of the Chineseretail industry. Int. J. Inf. Manag. 30(3):265–276, 2010.

46. Robertson, T. S., and Gatignon, H., Competitive effects on technol-ogy diffusion. J. Mark. 50(3):1–12, 1986.

47. Premkumar, G., and Ramamurthy, K., The role of interorganizationaland organizational factors on the decision mode for adoption ofinterorganizational systems. Decis. Sci. 26(3):303–336, 1995.

48. Armstrong, J. S., and Overton, T. S., Estimating nonresponse bias inmail surveys. J. Mark. Res. 14(1):396–402, 1977.

49. Hair, J. F., Anderson, R. E., Tatham, R. L., and Black, W. C.,Multivariate data analysis. Pearson Education, Upper Saddle River,1998.

50. Fornell, C., and Larcker, D. F., Evaluating structural equation modelswith unobservable variables and measurement error. J. Mark. Res.18(1):39–50, 1981.

51. Yi, M. Y., Jackson, J. D., Park, J. S., and Probst, J. C., Understandinginformation technology acceptance by individual professionals:Toward an integrative view. Information & Management 43(3):350–363, 2006.

52. Premkumar, G., Ramamurthy, K., and Nilakanta, S., Implementationof electronic data interchange: An innovation diffusion perspective.J. Manag. Inf. Syst. 11(2):157–186, 1994.

53. Moore, G. C., and Benbasat, I., Development of an instrument tomeasure the perceived characteristics of adopting an informationtechnology innovation. Inf. Syst. Res. 2(3):192–222, 1991.

54. Cases, A. S., Perceived risk and risk-reduction strategies in internetshopping. Distribution and Consumer Research 12(4):375–394,2002.

55. Brown, I., and Russell, J., Radio frequency identification technology:An exploratory study on adoption in the south African retail sector.Int. J. Inf. Manag. 27(4):250–265, 2007.

J Med Syst (2014) 38:3 Page 17 of 17, 3