Behaviur & Infrmatin Technlgy Vl. 28, N. 1, January February 2009, 5 20 Implementatin f an electrnic health recrds system in a small clinic: the viewpint f clinic staff Pascale Carayn a,b *, Paul Smith c,d, Ann Schfs Hundt a, Vipat Kuruchittham e and Qian Li f a Center fr Quality and Prductivity Imprvement, University f Wiscnsin-Madisn, Madisn, WI, USA; b Department f Industrial and Systems Engineering, University f Wiscnsin-Madisn, Madisn, WI, USA; c Department f Family Medicine, University f Wiscnsin Medical Schl, WI, USA; d University f Wiscnsin Medical Fundatin, WI, USA; e Cllege f Public Health, Chulalngkrn University, Bangkk, Thailand; f Center fr Quality and Prductivity Imprvement, University f Wiscnsin- Madisn, WI, USA (Received Nvember 2005; final versin received August 2007) In this study, we examined the implementatin f an electrnic health recrds (EHR) system in a small family practice clinic. We used three data cllectin instruments t evaluate user experience, wrk pattern changes, and rganisatinal changes related t the implementatin and use f the EHR system: (1) an EHR user survey, (2) interviews with key persnnel invlved in the EHR implementatin prject, and (3) a wrk analysis f clinic staff. A lngitudinal design with tw data-cllectin runds was emplyed: data were cllected prir t EHR implementatin and after EHR implementatin. Bth quantitative and qualitative data were cllected and analysed. Emplyees f the small clinic perceived few changes in their wrk after the implementatin f the EHR system, except fr increased dependency n cmputers and a small increase in perceived wrklad. The wrk analysis shwed a dramatic increase in the amunt f time spent n cmputers by the varius jb categries. The EHR implementatin did nt change the amunt f time spent by physicians with patients. On the ther hand, the wrk f clinical and ffice staff changed significantly, and included decreases in time spent distributing charts, transcriptin and ther clerical tasks. The interviews prvided imprtant cntextual infrmatin regarding EHR implementatin, and shwed sme psitive elements (e.g., planning f training), but als sme negative elements (e.g., unclear structure f the prject) that wuld have deserved additinal attentin. Keywrds: technlgy implementatin; healthcare; electrnic health recrds system (EHR) 1. Intrductin The imprtance f implementing and using health infrmatin technlgy (HIT) t imprve the delivery f health care has been increasingly recgnised (Institute f Medicine 2000, 2001, Thmpsn and Brailer 2004, Ash and Bates 2005, Berner et al. 2005, Middletn et al. 2005). The Institute f Medicine (2001) highlighted the central rle f HIT in the redesign f the health care system: Autmatin f clinical, financial, and administrative transactins (thrugh infrmatin technlgy) is essential t imprving quality, preventing errrs, enhancing cnsumer cnfidence in the health system, and imprving efficiency (p. 16). In the United States, federal and reginal effrts are under way t accelerate the adptin and use f electrnic health recrds as a means f facilitating clinical data sharing, prtect health infrmatin privacy and security, and quickly identify emerging public health threats (Thmpsn and Brailer 2004, Overhage et al. 2005). Driven by the needs t facilitate clinical and administrative prcesses, t reduce medical errrs, and t reduce healthcare csts, many healthcare institutins are deciding t implement electrnic health recrds (EHR) systems t allw clinical infrmatin gathering and access at the pint f patient care. An EHR system can access prgress ntes r prcedures data, and may supprt ther functins such as CPOE (cmputerised prvider rder entry) and CDSS (clinical decisin supprt systems). Tls t supprt administrative prcedures, such as billing and scheduling, are als becming cmmn EHR features. The use f EHR can facilitate clinical decisin-making and minimise the ptential fr mistakes due t the inaccuracy and incmpleteness f paper recrds (Institute f Medicine 2001, Thmpsn and Brailer 2004, Kawamt et al. 2005, Ohsfeldt et al. 2005). Hwever, the effects f EHR use n quality f care are nt necessarily autmatic (Linder et al. 2007); they very much depend n the specific characteristics f the EHR system and its impact n the wrk f healthcare prviders and ther staff. Recently, the need t adpt and adapt methds and techniques t understanding human factrs and *Crrespnding authr. Email: carayn@engr.wisc.edu ISSN 0144-929X print/issn 1362-3001 nline Ó 2009 Taylr & Francis DOI: 10.1080/01449290701628178 http://www.infrmawrld.cm
6 P. Carayn et al. rganisatinal issues f the technlgy implementatin prcess has been increasingly recgnised (Smith and Carayn 1995, Carayn and Karsh 2000, Carayn and Haims 2001, Karsh 2004). Regarding EHR implementatin, these include pr usability f EHR user interfaces, clinicians resistance t EHR acceptance, and patients reactin t EHR (Ash and Bates, 2005). The key t a successful EHR implementatin prject is hw well the technlgy is implemented and hw the technlgy can be used t imprve clinician perfrmance and prduce psitive individual and rganisatinal utcmes (Smith and Carayn 1995, Berner et al. 2005). Increased efficiency in healthcare delivery and imprvements in patient infrmatin cllectin, administrative prcessing, wrking cnditins, and user acceptance shuld lead t imprvements in safety, efficiency, and quality. Withut a cmprehensive understanding f end user experience and the rganisatinal changes prduced by the EHR technlgy, we are missing pprtunities t develp better appraches t designing and implementing EHR technlgy. Accrding t the Center fr Disease Cntrl and Preventin, general and family practices represent abut 24% f all physician ffice visits (Centers fr Disease Cntrl and Preventin 2000). It is therefre imprtant t understand the barriers t effective and successful implementatin f EHR technlgy in family practice clinics as a substantial prtin f ambulatry health care ccurs in these settings. EHR has been estimated t be used by abut 24% f physicians in ambulatry settings in the United States in 2005 (Jha et al. 2006). Challenges in dealing with human and rganisatinal factrs can partially explain why the majrity f small family practice clinics are still unwilling r unable t cnsider the use f EHR in their patient care. In additin, as cmpared t large hspitals, small clinics face further challenges due t limited financial and human resurces (Middletn et al. 2005). Healthcare prfessinals and administrative staff f small clinics frequently have t share jb respnsibilities and cver fr their cwrkers because f high wrklad, patient emergencies, and staffing issues such as emplyee vacatin and illness. In this study, we evaluated the implementatin f Practice Partner Patient Recrds, by Physician s Micrsystems, Inc., in a small family practice clinic; befre the EHR implementatin, health recrds were cmpletely in paper recrds. This EHR system is a vendr sftware intended t replace paper-based patient health recrds. We evaluated the rganisatinal aspects f the EHR implementatin prcess and the human factrs issues resulting frm the EHR implementatin. A systematic evaluatin apprach was emplyed: bth quantitative and qualitative data were cllected. This allwed us t evaluate hw emplyees in the clinic perceived their wrk as it related t the EHR technlgy and the changes in wrk patterns due t the EHR implementatin. The direct impact f EHR technlgy n clinical perfrmance and patient care (e.g., quality and safety f patient care) was nt examined in this study. 2. Cnceptual framewrk The mst cmmn reasn fr failure f technlgy implementatin is that the implementatin prcess is treated as a technlgical prblem, and the human factrs and rganisatinal issues are nt fully addressed (Easn 1988). In reactin t this prblem, Carayn and Karsh (2000) have prpsed a cnceptual mdel that specifies the human and rganisatinal issues related t technlgy implementatin (see Figure 1). The intrductin f a new technlgy is likely t change jbs and wrk prcesses. It can create bth psitive and negative impacts n jb characteristics (Carayn-Sainfrt 1992); therefre, it is imprtant t understand the impact f the technlgy n multiple dimensins f the jbs and wrk prcesses. Technlgy characteristics can als impact jb characteristics and quality f wrking life in bth psitive and negative manners (Carayn-Sainfrt 1992). Fr instance, a technlgy with usability deficiencies can increase the wrklad f the users, and affect their frustratin at wrk and ther attitudes tward their rganisatin. This cnceptual framewrk is used as the basis fr selecting measures t assess EHR implementatin in a small family practice clinic. 3. Study design In this study, three data cllectin instruments were used t assess user experience and rganisatinal changes related t the implementatin and use f EHR: a user survey, interviews with key persnnel invlved in the EHR prject, and a wrk analysis. A lngitudinal design with tw data cllectin runds befre and after the EHR implementatin was emplyed. The study site is a University f Wiscnsin family medicine residency clinic in a small cmmunity with a ppulatin f abut 1800, lcated 18 miles suthwest f Madisn, Wiscnsin. At the time f study, it had 6 family medicine faculty, 7 resident physicians, and 12 medical supprt and ffice staff. It had apprximately 11 000 patient visits annually. Participatin in the study by the clinic persnnel was vluntary. Each data cllectin is described separately. The results f each data cllectin are reprted after the descriptin f the data cllectin methd.
Behaviur & Infrmatin Technlgy 7 Tcrhnlgy characlcristics: type ftcrhnlgy. funclinality, and usabilily issucs Jb characterislics: jb cnlrl. wrklad. uncertainly/clarily. challenge. and rle ambiguily Quality fwrking lifc: jb salisfactin. Slress, sclf-reprt~-d health, and perceived perfrmance Tcrhnlgical changc prcess: emplyee participalin. feedback, prjccl management. infrmatin and cmmunicatin, and lraining and leaming Figure 1. Impact f EHR technlgy n quality f wrking life and perfrmance (adapted frm Carayn and Karsh 2000). 4. Survey f EHR users 4.1. Pre- and pst-implementatin survey Based n the cnceptual framewrk (see Figure 1) f Carayn and Karsh (2000), the pre-implementatin survey examined the fllwing human and rganisatinal factrs: (a) Jb infrmatin: jb psitin (e.g., ffice staff, nurse, dctr), jb experience, and cmputer experience (b) Jb characteristics: rle ambiguity (Caplan et al. 1975), quantitative wrklad (Caplan et al. 1975), uncertainty (Seashre et al. 1983), challenge (Seashre et al. 1983), task cntrl (McLaney and Hurrell 1988, Greenberger et al. 1989), decisin cntrl (McLaney and Hurrell 1988, Greenberger et al. 1989), resurce cntrl (McLaney and Hurrell 1988, Greenberger et al. 1989), and general jb cntrl (McLaney and Hurrell 1988, Greenberger et al. 1989) (c) Quality f wrking life: rganisatinal identificatin (Ck and Wall 1980), rganisatinal invlvement (Ck and Wall 1980), daily life stress (Reeder et al. 1973), jb satisfactin (Quinn et al. 1971), musculskeletal discmfrt (Sainfrt and Carayn 1994), and anxiety (Sainfrt and Carayn 1994) (d) Technlgy characteristics: dependency n cmputers (Carayn 1994), infrmatin received abut EHR system (Bailey and Pearsn 1983), input regarding design and implementatin f the EHR system (Bailey and Pearsn 1983), attitude tward EHR system (Bailey and Pearsn 1983), EHR effect n perfrmance (Davis 1989), verall user acceptance, learning, and EHR system capabilities (Chin et al. 1988) (e) Self-rated perfrmance (Carayn 1994) (f) Demgraphics: gender, age, educatinal level, and marital status. The first five sectins (a, b, c, d, and e) f the preimplementatin survey were als included in the pstimplementatin survey. Twelve questins n technlgy characteristics were added t the pst-implementatin survey. These questins were derived frm the Questinnaire fr User Interface Satisfactin (QUIS) (Chin et al. 1988). As a usability evaluatin tl, QUIS (Chin et al. 1988) cnsists f five categries f questins n user experience with sftware user interface: verall reactins t the sftware, learning, system capabilities, terminlgy and system infrmatin, and screen. We used the first three sectins (reactins t the sftware, learning, and system capabilities) in the pstimplementatin survey. The demgraphics sectin (Sectin f) was excluded frm the pst-implementatin survey based n a recmmendatin by the University Institutinal Review Bard. 4.2. Participants Twenty-ne ut f 25 clinic emplyees cmpleted the pre-implementatin survey, while 20 ut f
8 P. Carayn et al. 25 emplyees cmpleted the pst-implementatin survey. Respnse rates were 84% and 80%, respectively. 4.3. Prcedures The pre-implementatin survey, alng with the cnsent frm, was distributed t all 25 clinic emplyees in the spring f 2000, six mnths befre the EHR implementatin. Clinic emplyees wh agreed t participate in the study signed the cnsent frm and then cmpleted the survey. They left the signed cnsent frms and cmpleted surveys in a secured mailbx that was accessed nly by the researchers. In February 2002, 15 mnths after the implementatin, the pst-implementatin survey was administered using the same prcedure as in the pre-implementatin survey. 4.4. Data analysis Data frm the survey were manually entered int an SPSS database and duble-checked by anther researcher fr quality cntrl. The first step f the data analysis prduced nrmalised scres, frm 0 (lw) t 100 (high), fr each measure included in the fllwing fur sectins f the survey: jb characteristics, quality f wrking life, technlgy issues, and self-rated perfrmance. Descriptive statistics were calculated. Survey data were cllected at tw different pints in time: befre and after the EHR implementatin; hwever, because f the small sample size and the threat t annymity, individual respnses were nt tracked ver time. Therefre, Mann-Whitney tests were perfrmed t cmpare the grup respnses f the pre- and the pst-implementatin surveys. The sectin n quality f wrking life cnsisted f twenty-tw fur-pint health questins with answers ranging frm 1 ( never ) t 4 ( cnstantly ) and cncerning three dimensins: (1) back, neck, shulder discmfrt, (2) ther musculskeletal discmfrt, and (3) anxiety (Sainfrt and Carayn 1994). Respnses t each f the 22 questins were gruped int never, ccasinally, and frequently and cnstantly in rder t examine participants with n, sme, and a lt f perceived discmfrt and anxiety. Kruskal Wallis tests were perfrmed t cmpare the results f the pre- and the pst-implementatin surveys. 4.5. Results Descriptive statistics f jb characteristics, quality f wrking life, technlgy issues, and self-rated perfrmance, alng with the results f Mann-Whitney tests, are reprted in Table 1. Tw measures, resurce cntrl and dependence n cmputer, were significantly different between the pre- and the pst-implementatin surveys (p 5 0.05). Participants reprted less resurce cntrl and mre dependence n cmputers after EHR implementatin. Perceived quantitative wrklad increased slightly after EHR implementatin, cmpared t befre EHR implementatin (p 5 0.10). Descriptive statistics f the 22 health questins as well as the results f Kruskal Wallis tests can be fund in Table 2. The measure f tight feeling in stmach was fund t be significantly different between the pre- and the pst-implementatin surveys (p 5 0.05). Fewer participants reprted tight stmach feeling after EHR implementatin. There was a slight increase in the percentage f participants wh reprted back pain and pain r stiffness in arms r legs (p 5 0.10), and a slight decrease in terms f swllen r painful muscles and jints (p 5 0.10), after EHR implementatin. 5. Interviews f key persnnel invlved in EHR implementatin prject 5.1. Pre- and pst-implementatin interview guide Structured interviews were cnducted using an interview guide based n the IT prject management interview guide (Krunka et al. 1997, Krunka and Carayn 1999). The timeframe f the questins was mdified t reflect the pre- and pst-implementatin times. Fr example, a questin n training activities planned fr the clinic staff was asked in the pre-implementatin interview, while a questin n the actual training activities that had taken place was asked in the pst-implementatin interview. The interview guide was structured as fllws: (a) Implementatin backgrund (b) Prject identity (c) Prject team (d) Prject manager (e) Steering cmmittee (f) Implementatin prcess, including gals, prcesses, schedule, budget, infrmatin diffusin, evaluatin, prblems/difficulties, prject crises, feedback/cmplaints (nly in the pst-implementatin interviews), and user participatin (g) Training (h) EHR supprt (i) Changes in the wrking envirnment (j) Interviewee prfile.
Behaviur & Infrmatin Technlgy 9 Table 1. Survey f EHR users: descriptive statistics and results f Mann-Whitney tests. Pre { (nrmalised scres: 0 100) Pst { (nrmalised scres: 0 100) Mean S.D. Mean S.D. Mann-Whitney tests (p values) Jb characteristics Rle ambiguity 13 13.9 14 13.4 0.666 Quantitative wrklad 72 12.1 77 15.6 0.091* Uncertainty 70 17.3 64 19.6 0.341 Challenge 76 17.3 81 15.5 0.440 Task cntrl 50 18.8 47 17.0 0.801 Decisin cntrl 32 26.9 27 23.5 0.713 Resurce cntrl 47 25.6 29 25.8 0.028** General jb cntrl 44 18.2 38 16.3 0.325 Quality f wrking life Organisatinal identificatin 87 12.0 83 14.6 0.597 Organisatinal invlvement 88 12.5 85 12.5 0.477 Daily life stress 44 11.6 41 11.3 0.247 Jb satisfactin 72 25.6 78 24.3 0.436 Musculskeletal discmfrt 15 12.0 17 12.1 0.530 Back, neck, shulder 24 23.8 29 21.2 0.159 Other musculskeletal 11 9.3 11 9.9 0.916 Anxiety 18 14.1 15 11.3 0.453 Self-reprted perfrmance 77 12.4 76 11.3 0.800 Technlgy characteristics Dependence n cmputers 73 27.5 89 26.8 0.018** Infrmatin received abut system 53 27.6 63 25.1 0.295 Design input 46 33.2 54 28.2 0.271 Implementatin input 49 32.8 56 27.4 0.569 Effect n perfrmance 67 23.9 53 25.7 0.117 Attitude tward system 68 24.0 66 24.5 0.790 Overall reactins N/A N/A 54 16.9 N/A Learning N/A N/A 51 17.4 N/A System capabilities N/A N/A 63 17.1 N/A { Pre-implementatin survey. { Pst-implementatin survey. *p 5 0.10; **p 5 0.05. 5.2. Interviewees Fur key persnnel wh were directly invlved in the implementatin prcess f the EHR system were interviewed: the prject directr, the prject manager, the clinic manager, and the infrmatin system manager. 5.4. Data analysis Data cllected during the interviews were entered int an Access database. Descriptive infrmatin abut the EHR implementatin prcess is prvided in the next sectin. 5.3. Prcedures The face-t-face pre-implementatin interviews were cnducted with the fur interviewees in April 2000. They were prvided with a cpy f the interview guide befre the interviews. An interviewer asked questins ne by ne fllwing the structure f the interview guide. In additin t answering the questins, interviewees were encuraged t give feedback abut the questins and t prvide additinal infrmatin that may be helpful t understand the EHR implementatin prcess. The individual interviews lasted 60 120 min. The pst-implementatin interviews were again cnducted with the same fur peple by telephne between March and April 2001. 5.5. Results 5.5.1. Implementatin backgrund The primary factrs driving EHR implementatin were the need fr imprving medical care and the trend in the industry. Secndary factrs included the desire fr wrk reductin, adjustment t market demands, and the reductin f emplyee wrklad. These resulted in the intrductin f an EHR system t replace an existing paper-based medical recrd system. It tk fur mnths t cmplete the actual implementatin. The wrkstatin selectin was based n vendr recmmendatins. Selectin criteria fr the sftware included capability, serviceability, user friendliness, ppularity, and recmmendatins f the prduct. Decisins n prject scpe
10 P. Carayn et al. Table 2. Survey f EHR Users Descriptive Statistics f Health Questins and Results f Kruskal-Wallis Tests Pre { Pst { Kruskal-Wallis N x O { F/C {{ N x O { F/C {{ test (p values) Back, neck, shulder discmfrt 1. Back pain 7 12 2 3 12 5 0.095* 2. Pain r stiffness in yur neck and shulders 4 14 2 3 12 5 0.291 3. Feeling f pressure in the neck 13 4 4 8 7 5 0.231 4. Shulder sreness 11 7 3 9 9 2 0.796 5. Neck pain that radiates int shulders, arms r hands 15 3 3 13 3 4 0.634 Other musculskeletal discmfrt 6. Swllen r painful muscles and jints 3 15 3 7 12 1 0.092* 7. Pain r stiffness in yur arms r legs 12 8 1 6 12 2 0.085* 8. Persistent numbness r tingling in any part f yur bdy 18 3 0 15 4 1 0.363 9. Pain dwn yur arms 16 5 0 16 3 1 0.842 10. Leg cramps 16 4 1 15 5 0 1.000 11. Difficulty with feet and legs when standing fr prlnged perids 12 8 1 14 5 1 0.431 12. Lss f feeling in the fingers r wrists 17 3 1 16 3 1 0.940 13. Cramps in hands/fingers relieved nly when nt wrking 18 3 0 17 3 0 0.949 14. Lss f strength in arms r hands 18 3 0 18 2 0 0.679 15. Stiff r sre wrists 17 4 0 16 3 1 0.880 Anxiety 16. Occasins f easy irritability 7 12 2 3 17 0 0.481 17. Difficulty sleeping 11 7 3 8 12 0 0.816 18. Perids f depressin 9 10 2 12 8 0 0.194 19. Times f severe fatigue r exhaustin 10 10 1 10 10 0 0.766 20. Tight feeling in stmach 15 6 0 20 0 0 0.011** 21. Perids f extreme anxiety 15 6 0 17 3 0 0.300 22. High levels f tensin 9 11 1 8 12 0 1.000 { Pre-implementatin survey. { Pst-implementatin survey. x Never. { Occasinally. {{ Frequently and cnstantly. *p 5 0.10; **p 5 0.05. and hardware/sftware selectin were made jintly by the steering cmmittee, the prject team, expert end users, and the infrmatin systems department. 5.5.2. Prject identity During the pre-implementatin interviews, all fur interviewees said that the prject was given a special identity using the prject name. Three interviewees reprted that the prject had n special identity when interviewed after the implementatin. 5.5.3. Prject team All fur interviewees agreed that the prject team had an infrmally defined scpe, authrity, and respnsibility. Hwever, their understanding f the prject team cmpsitin diverged. Fr example, their answers t the questin n hw many expert end users were n the prject team varied frm 1 t 5. The fur interviewees agreed that the prject team members were chsen based n prfessinal expertise and EHR knwledge by tp management. Besides their regular jb duties, team members were granted time t wrk n prject-related activities, including 30 weekly prject meetings thrughut the EHR implementatin prcess. The fur interviewees agreed that the verall attitude f the prject team was gd. 5.5.4. Prject manager The prject manager was hired externally and temprarily fr the EHR implementatin prject and reprted t the prject directr. Tp management made the hiring decisin based n criteria such as experience as prject manager, prfessinal expertise, and persnality. The prject manager did nt receive extra training n prject management. It was unclear whether the prject manager had authrity t make decisins in cases f diverging pinins: tw interviewees said that the prject manager did nt have this authrity, while the ther tw cnsidered that the prject manager had infrmal authrity. 5.5.5. Steering cmmittee There was nt an fficially designated steering cmmittee specifically fr this prject. The prject team
Behaviur & Infrmatin Technlgy 11 and the prject directr reprted t the department s standing executive cmmittee. regular jb duties were cvered by ther emplyees. N extra wrk hurs were explicitly needed. 5.5.6. Implementatin prcess The gals f EHR implementatin were t enhance healthcare quality and patient safety, t imprve wrk quality and reliability, t imprve infrmatin sharing and cmmunicatin, and t reduce wrk steps and errrs. These gals were frmulated by the prject team and lcal tp management thrugh preliminary wrk dne befre implementatin, including gal setting, cst-benefit analyses, and risk assessment activities. Tw interviewees indicated that technical difficulties during EHR implementatin were significant, the ther tw reprted nticeable but slight difficulties. Critical issues included hw the EHR system culd interface with billing functins. Prblems with the vendr were reprted, such as crrupted cnfiguratin with lab data (tk six weeks t get it crrected), and system upgrade crash befre ging live (lst days f data). One interviewee rated the prblems as significant, ne as nticeable, and the ther tw as slight. Underestimatin f the amunt f wrk required fr EHR implementatin was anther majr difficulty reprted by three interviewees, in additin t the cncern regarding the authrity f the prject manager, the lack f interest and resistance frm end users, the disagreement within the prject team, the resistance frm middle management, and the lack f pririty fr the prject. Accrding t the interviewees, end users cmplained f an increase f wrk due t the implementatin, technical interruptin, and time pressure during EHR implementatin. The prject manager cmplained abut sftware bugs, while lcal tp management was cncerned with decreased prductivity during the implementatin and the cst. Patients were reprted t have cncern regarding privacy f their medical data. User acceptance f the EHR was evaluated thrugh infrmal discussin. 5.5.7. Training The fur interviewees cnsidered that lcal tp management had been very psitive twards training. The amunt f training that users received was decided jintly by the prject team and the EHR vendr. The training scheduling was established by the prject team. All clinic emplyees were infrmed that they wuld need t be trained. Training schedules and training materials were prvided. Grups f users with similar needs were trained tgether thrugh hands-n practice. Expert users were trained fr 8 h, while thers were trained fr 4 h. The training cnsisted f tw sessins n basic Windws and the EHR system. When a user was attending a training sessin, his (r her) 5.5.8. EHR supprt There were supprt staff present frm the EHR vendr n the day the EHR system went live. In the fllwing tw weeks, at least ne expert end user was present at the clinic. The sftware maintenance was dne internally. In additin, there were plans fr imprving the EHR system by crrecting sftware bugs, adding new applicatins, upgrading new releases, as well as upgrading hardware cmpnents. 5.5.9. Changes in wrking envirnment All 4 interviewees agreed that clinic emplyees experienced changes in skills and wrk flw, and increased wrklad due t the implementatin and use f EHR. The use f EHR did nt result in reductin f persnnel. There was an increase in time spent using the cmputer, althugh it varied depending n the jb categry (e.g., nursing staff and physicians experienced mre changes than thers). Tw interviewees bserved a slight change in scial climate as a result f the implementatin, while the ther tw bserved n change r did nt knw. In general, all interviewees agreed that the climate f the entire clinic was psitive after EHR implementatin. 6. Wrk analysis f clinic staff 6.1. Wrk analysis frm Pre- and pst-implementatin wrk analyses were cnducted using the multidimensinal wrk sampling technique (Sittig 1993, Murray et al. 1999). The multidimensinal wrk sampling technique was used t determine time spent n a variety f predefined activities ( activity ), the purpse f the activity ( functin ), and with whm the persn was in cntact while perfrming the activity ( cntact ). The wrk analysis frm and the definitins fr the activities, functins, and cntacts were first created using infrmatin frm the psitin descriptins prvided by the clinic manager. After creating the frm and the definitins, the researchers met with the medical directr f the clinic and the clinic manager t discuss and revise the data cllectin frm. The frequency, duratin, and timing f the wrk analysis were als discussed. The same frm was used in bth the preand pst-implementatin studies. It included 13 activities, 22 functins, and 14 cntacts. Fr each entry n the frm, study participants culd als recrd cmments when they were unsure what activity, functin, r cntact t recrd (see Appendix).
12 P. Carayn et al. 6.2. Participants All clinic emplyees were invited t participate in the wrk analysis in the pre- and pst-implementatin phases. Twenty-seven clinic emplyees participated in the pre-implementatin study, while 26 emplyees participated in the pst-implementatin study. Unlike the emplyee survey, where primarily full time emplyees were recruited t participate, the wrk analysis recruitment included everyne wh wrked at the clinic, be they part-time, full-time, permanent, r temprary. Fr that reasn there were mre participants in the wrk analysis than there were in the survey questinnaire. 6.3. Prcedures The pre-implementatin wrk analysis was cnducted in April 2000 fr a perid f 10 wrking days. It began n a Tuesday due t the hectic nature f Mndays fllwing a weekend. It was believed that staff wuld have mre time t adjust and be familiar with the wrk analysis tl by beginning n Tuesday. The wrk analysis frms were distributed t all emplyees, including the medical staff, at the clinic. Participants indicated their psitin and the beginning and ending times f their wrkday. During each day, participants were asked t recrd activities, functins, and cntacts every 30 min. They were encuraged t write dwn cmments when they were uncertain abut what t recrd. An annuncement via verhead speaker was made apprximately every 30 min t remind participants t cmplete the frm. At the end f each day, participants drpped their frm in a lcked mailbx, which nly researchers had access t. The pstimplementatin wrk analysis was cnducted in June and July 2002 using the same prcess as in the preimplementatin study, except fr a slight variatin in the recrding time. An bservatin was made by the medical directr that sme staff, anticipating the recrding time, remained at their desk rather than leave r initiate a different task when the 30-min interval apprached. Thus, in the pst-implementatin study, an annuncement was nt made at the exact 30- min interval, but rather at an apprximate time (e.g., plus r minus five minutes f the half-hur). Participants again drpped their frm in a lcked mailbx at the end f each day. 6.4. Data analysis One hundred and frty-five frms were cllected with 1960 entries fr the pre-implementatin study and 122 frms with 1825 entries fr the pst-implementatin study. Data frm the wrk analysis were entered int Excel wrksheets. In case there were tw activities, functins, r cntacts recrded fr a time perid, a research scientist (A.S.H.) wrking n the study chse what appeared t be the mst apprpriate cde based n cmments prvided by the respndent r the patterns f ther entries r bth. This judgment was made by the research scientist t ensure cnsistency in cding. Once all data were entered, frequencies were cmputed fr the activities, functins and cntacts fr each f the three jb categries: physicians, nnphysician clinical staff (e.g., nurses, labratry technicians, radilgy technicians), and ffice staff. We identified sme cnfusin the participants experienced in chsing the activity, functin, r cntact t recrd. Fr instance, cmputer entry (A3) shuld be recrded when entering infrmatin int the cmputer, and nt typing/writing/signing (A12). T address the cnfusin, a standard prcedure fr recding was develped and all data f the pre- and the pstimplementatin studies were reviewed by the same research scientist, again, t ensure cnsistency in the recding. After cmpleting the recding, all frequencies were recmputed. Frequencies fr each activity, functin, cntact, and their task cmbinatin (activity/ functin/cntact) were calculated. Cmparisns f the pre- and the pst-implementatin wrk analysis data were perfrmed t examine changes in the distributin f time spent n varius activities, functins, and cntacts fr each f the three jb categries (see Tables 4 6). w 2 tests were run t cmpare the distributin f frequencies fr each f the three jb categries separately; this same analysis was dne fr the data n activities, functin, and cntact. Because sme f the percentages were small, we cmbined the data fr the categries f activity, functin, r cntact whenever the pre and the pst percentages were smaller than 5% (see Figures 2 4). 6.5. Results The numbers f tasks and entries are prvided in Table 3. Physicians had the least number f tasks and Table 3. entries. Wrk analysis f EHR users: number f tasks and Number f tasks Original Recded Number f entries Pre { Pst { Pre { Pst { Pre { Pst { Physician 101 94 79 81 545 458 Clinical staff 138 83 122 76 576 393 Office staff 204 210 186 173 839 974 All 361 326 305 266 1960 1825 { Pre-implementatin survey. { Pst-implementatin survey.
Table 4. Wrk analysis f EHR users: frequencies f activity (%). Behaviur & Infrmatin Technlgy 13 Physicians Clinical staff Office staff All Activity Pre { Pst { Pre { Pst { Pre { Pst { Pre { Pst { A1 Absent 0 0 0.7 0 1.0 0.1 0.6 0.1 A2 Caring fr patient 49.7 50.7 27.8 39.2 7.2 4.0 25.1 23.3 A3 Cmputer entry 1.7 21.2 4.7 10.4 19.0 30.3 9.9 23.7 A4 Dictatin 9.7 1.5 0.7 0 0 0 2.9 0.4 A5 Meeting 1þ 4.8 3.3 11.1 1.0 5.6 4.6 7.0 3.5 A6 Meeting 3þ 0.7 1.1 1.0 3.1 1.9 2.3 1.3 2.1 A7 Perfrming lab wrk 0.2 0.2 18.4 11.7 0 0.1 5.5 2.6 A8 Phne 11.4 5.5 9.0 13.0 16.2 16.9 12.8 13.2 A9 Preparing 2.4 2.0 14.1 17.8 22.6 14.7 14.5 12.2 A10 Reviewing check 5.0 4.6 3.1 2.0 1.4 0.9 2.9 2.1 A11 Supervising 4.0 4.4 0.9 0 0.5 0.5 1.6 1.4 A12 Typing/writing/signing 9.0 3.5 4.5 1.0 16.2 18.3 10.8 10.8 A13 Other 1.5 2.2 4.0 0.8 8.3 7.2 5.2 4.5 { Pre-implementatin survey. { Pst-implementatin survey. ffice staff had the mst number f tasks. The frmer handled abut 80 tasks, while the latter handled apprximately 180 tasks. The number f tasks perfrmed by physicians and ffice staff did nt change much in the pst-implementatin study as cmpared with that in the pre-implementatin study. There was a 38% decrease in the number f tasks perfrmed by the clinical staff: frm 122 tasks in the preimplementatin study t 76 tasks in the pstimplementatin study. 6.5.1. Activity Frequencies f activity cmparing pre- and pstimplementatin are shwn in Table 4 and Figure 2. Physicians spent abut half f their time caring fr patients in bth the pre- and pst-implementatin studies. The EHR implementatin did nt affect the amunt f time physicians spent with patients, but increased the amunt f time spent by physicians n cmputer entry and decreased time spent n dictatin, phne, and typing/writing/signing. Fr clinical staff, the main differences between the pre- and the pst- EHR implementatin were the fllwing: increases in time spent n patient care, cmputer entry, phne and preparing, and decreases in time spent in meeting, perfrming lab wrk, and typing/writing/signing. The difference between the pre- and pst-ehr frequencies f activity fr the ffice staff was nt statistically significant. Physicians spent almst half f their time examining r treating patients befre and after the EHR implementatin (frm 42.4% t 48.7%). The frequencies f functin fr the physicians did nt significantly change after the EHR implementatin. Clinical staff spent mre time n the fllwing functins: accmpanying patients (frm 17.5% t 22.9%), examining patient (frm 4.3% t 12.7%), and maintaining medical infrmatin system (frm 5.6% t 10.9%). Clinical staff spent less time n distributing chart/ master file/mail (frm 6.6% t 0.3%), and perfrming tests (frm 18.6% t 11.2%). The functins f ffice staff als changed significantly: they spent abut ne-half less time fr distributing chart/master file/mail (frm 5.6% t 2.5%), general clerical assistance/ffice tasks (frm 27.8% t 13.1%), and transcriptin (frm 13.3% t 7.5%), but mre time n maintaining the medical infrmatin system (0% t 19.8%). 6.5.3. Cntact Frequencies f cntact can be fund in Table 6 and Figure 4. Physicians had abut the same distributin f cntact befre and after the EHR implementatin. Clinical staff spent less time in cntact with dctrs and nurses, but mre time with patients and patient representatives, and ding tasks by themselves. On the cntrary, ffice staff spent mre time with nurses, but less time ding tasks n their wn. 6.5.2. Functin Frequencies f functin cmparing the pre- and pstimplementatin are shwn in Table 5 and Figure 3. 6.5.4. Task cmbinatin Task cmbinatins represent cmbinatins f an activity, a functin, and a cntact. Sme f the task
14 P. Carayn et al. 3. Physicians (1=027.22; df=6: p<.ool) 60,---------------, 50 " 30 20 10 0 ~ J; B 0 i ~ 2 ~ f:~ <[.[ i u " b. Clinical stall (x?~18.21; df;6: p<.oi) ~ '~.i! 6Or------------~ 501------------1 " 30 20 10 c. Ollicc stall(;(=5.3: df5: nt significant) 60,-------------- 501------------ "t--------------- 301-- 201-----r 10 t 0 5 '0 B " 0 I :~ ~ 0,, " i,~.~.~ i ~ u Figure 2. Wrk analysis: cmparisn f pre-ehr and pst- EHR frequencies f activity. (a) Physicians (w 2 ¼ 27.22; df ¼ 6; p 5 0.001); (b) clinical staff (w 2 ¼ 18.21; df ¼ 6; p 5 0.01); (c) ffice staff (w 2 ¼ 5.3; df ¼ 5; nt significant). cmbinatins perfrmed by physicians are listed belw:. Cmputer entry/maintaining medical infrmatin system/self increased frm 0.2% t 15.7%. Caring fr patient/examinatin r treatment f patient/patient and patient representative slightly increased frm 40.4% t 45.85%. Dictatin/maintaining medical infrmatin system/self decreased frm 9.7% t 1.5% " 5 lop_i Pst. Typing, writing, signing/maintaining medical infrmatin system/self decreased frm 7.7% t 1.5%. Phne/prviding instructin, infrmatin/patient and patient representative decreased frm 7.3% t 1.5%. Sme f the task cmbinatins perfrmed by clinical staff are listed belw.. Caring fr patient/examinatin r treatment f patient/patient and patient representative increased frm 3.5% t 11.7%. Perfrming lab wrk/perfrming tests/dctr decreased frm 10.4% t 1%. Meeting 1þ/training/nurse drpped ut cmpletely frm 7% t 0%. Sme f the task cmbinatins perfrmed by ffice staff are listed belw.. Cmputer entry/maintaining medical infrmatin system/self increased frm 0% t 7.2%. Typing, writing, signing/transcriptin/self decreased frm 11.7% t 6.5%. Overall, the mst significant change was fund in the cmbinatin f Cmputer entry/maintaining medical infrmatin system/self with an increase frm 0.6% in the pre-implementatin study t 9.1% in the pstimplementatin study. 7. Discussin 7.1. Survey f EHR users Overall, clinic emplyees experienced lw rle ambiguity, high wrklad, high uncertainty, high challenge, mderate task cntrl, and lw decisin cntrl. They felt that they had high rganisatinal identificatin and invlvement, mderate daily life stress, lw musculskeletal discmfrt, lw anxiety, high jb satisfactin, and high self-rated perfrmance. As expected, dependency n cmputers significantly increased with EHR implementatin. Unexpectedly, clinic emplyees felt that they had less resurce cntrl after EHR implementatin. The medical directr f the clinic prvided a pssible explanatin t this unexpected finding: the clinic emplyees were under budgetary cntrl at the time f the pstimplementatin survey. Therefre, they may have reprted deceased cntrl ver resurces such as supplies and materials. Other interesting results f the survey included increases in perceived wrklad, back pain and pain/stiffness in arms/legs, and
Table 5. Wrk analysis f EHR users: frequencies f functin (%). Behaviur & Infrmatin Technlgy 15 Functin Physicians Clinical staff Office staff All Pre { Pst { Pre { Pst { Pre { Pst { Pre { Pst { F1 Accmpanying patients 0.2 1.1 17.5 22.9 0.5 1.1 5.4 5.8 F2 Assisting physician, dctrs r medical technician 0.4 0.9 1.9 3.1 0.1 0 0.7 0.9 F3 Billing activities 0.4 0.4 0 0.3 16.2 16.8 7.0 9.2 F4 Checking message 0.9 3.5 0.7 0.5 1.9 2.2 1.3 2.1 F5 Data review and retrieval 3.9 3.3 4.0 4.1 3.1 1.6 3.6 2.6 F6 Distributing chart/master file/mail 0 0.2 6.6 0.3 5.6 2.5 4.3 1.4 F7 Examinatin r treatment f patient 42.4 48.7 4.3 12.7 0 0.4 13.1 15.2 F8 General clerical assistance/ffice task 0.4 0.2 3.5 1.5 27.8 13.1 13.0 7.4 F9 Maintaining equipments, instruments, supplies and medicatins 0.2 0 3.0 7.1 0.5 0.3 1.1 1.7 F10 Maintaining lab reprts 0.2 0.2 4.9 6.4 0 0.1 1.5 1.5 F11 Maintaining medical infrmatin system 23.3 24.7 5.6 10.9 0 19.8 8.1 19.1 F12 Maintaining patient s infrmatin recrd 0.4 0.7 1.0 0.3 2.1 2.1 1.3 1.3 F13 Participating in resident and student educatin 7.3 5.5 0.7 0 0.5 0 2.4 1.4 F14 Perfrming tests 0 0.2 18.6 11.2 0.1 0.2 5.5 2.6 F15 Preparing fr examinatins and surgical prcedures 0.4 0 1.4 2.0 0 0.3 0.5 0.6 F16 Prviding instructin/infrmatin 13.0 4.8 5.6 6.6 1.0 1.2 5.7 3.3 F17 Purchasing and making inventry arrangement 0 0 0.3 0.5 0.2 0.2 0.2 0.2 F18 Reprting prblem 0.4 0.9 0.9 0.3 0.2 1.2 0.5 0.9 F19 Scheduling 0.2 0.2 4.9 3.8 13.8 14.9 7.4 8.8 F20 Training 0.6 0.9 6.9 0.5 1.0 1.8 2.6 1.3 F21 Transcriptin 0 0 0.2 0 13.3 7.5 5.8 4.0 F22 Other 5.7 3.7 7.6 5.1 11.9 12.3 8.9 8.6 { Pre-implementatin survey. { Pst-implementatin survey. decreases in swllen/painful muscles and jints and reprt f tight feeling in stmach. Befre EHR implementatin, clinic emplyees reprted that they received mderate infrmatin abut the EHR system, had mderate design input and mderate input int the implementatin prcess. After EHR implementatin, their perceptins n these issues were mre psitive: they reprted that they received better infrmatin and that their inputs were mre widely cnsidered. In general, clinic emplyees attitude tward EHR system was psitive. They reprted that the EHR system had sme psitive effect n their perfrmance. They felt that the EHR system was mderately easy t learn and that the EHR system had mderate capabilities in terms f technical perfrmance. 7.2. EHR implementatin prcess: interviews The fur interviewees were the key persnnel invlved in the EHR implementatin prcess. They agreed that the verall attitude f the prject team was gd and the climate f the entire clinic during the implementatin was psitive. On the ther hand, they reprted that clinic emplyees cmplained f the increased wrklad due t the technical prblems and time pressure assciated with EHR implementatin. Technical difficulties, user resistance, and ther prblems are relatively cmmn with this type f technlgy implementatin prject (Ash and Bates 2005). The prject team was able t identify thse prblems and crrect them during weekly meetings. In additin, the interviewees reprted that user training was thrughly planned and delivered and technical supprt was available t end users as they needed. With regard t sme issues such as the prject identity, the authrity f prject manager, and the severity f technical difficulties experienced during the implementatin, the fur interviewees perceptins varied. This variatin in perceptins highlights the need fr clarifying and specifying the structure f the EHR implementatin prcess. 7.3. Wrk analysis Overall, the wrk analysis shwed many differences in the wrk f clinical staff and ffice staff, and few changes fr the wrk f physicians. There was n difference in physician time spent caring fr patients befre and after EHR implementatin. Physicians spent abut half f their time n the activity f patient care and abut 55% f their time n the tw functins f examinatin r treatment f patient and prviding instructin-infrmatin. This result is similar t
16 P. Carayn et al. a. Physicians (l:4.68; df=4; nt significant) 60,------------------, 50 40 30 20 10 ----------------j I~ 0 ~" ~ " 0 E ~ g,, " 0 ~.~ i "..E ~ ", ~ E " i " " >,.,.~~ i: b. Clinical slaff (/=22.29: df=9; p<.o I) 0 60~----------------------~ "t------------------------1 4Ot-----------------------1 30 IOPffil Pst 20 10 c. Officc staff (/=28.Q3: df=6: p<. I) Figure 3. Wrk analysis: cmparisn f pre-ehr and pst-ehr frequencies f functin. (a) Physicians (w 2 ¼ 4.68; df ¼ 4; nt significant); (b) clinical staff (w 2 ¼ 22.29; df ¼ 9; p 5 0.01); (c) ffice staff (w 2 ¼ 28.03; df ¼ 6; p 5 0.001). H,. ph I. ~~tl 60,------------------, 5Ot-----------------j 40+-----------------1 3O+-----~=._----------_I 20+-::=----1 10.;,,~ I~.1, ~I ], ~.~ t 0 E~ := ~ f, ~ l!;""".,.- il.,! thse f ther studies. Fr example, a study f physicians in an utpatient nclgy clinic fund that physicians spend abut 43% f their time in patient care (Fntaine et al. 2000). As expected, cmputer entry activity by physicians increased in place f dictatin, phne, and typing/writing/signing activities, which decreased. Clinical staff spent mre time caring fr patients after EHR implementatin. A pssible explanatin t this might be that there was less lab wrk t be perfrmed n patients (summer appintments as ccurred during the pst-implementatin study frequently tend t include mre well-patient wrkups and physical exams that d nt require lab wrk). Meetings with ne r tw persns drpped frm 11% t 1%, prbably because f use f the EHR internal messaging system. In additin, clinical staff spent mre time in maintaining the medical infrmatin system instead f distributing chart/master file/mail. Office staff used EHR by spending mre time n cmputer entry and less time n preparing activities (e.g., filling, retrieving, and distributing charts). After
Table 6. Wrk analysis f EHR users: frequencies f cntact (%). Behaviur & Infrmatin Technlgy 17 Physicians Clinical staff Office staff All Cntact Pre { Pst { Pre { Pst { Pre { Pst { Pre { Pst { C1 Billing crdinatr 0 0 0 0 1.0 0.7 0.4 0.4 C2 Dctr 2.6 2.6 14.6 9.2 2.4 1.2 6.0 3.3 C3 Manager 0.7 0 0.5 0.3 0.8 0.7 0.7 0.4 C4 Medical student 1.5 0.2 0.2 0 0 0 0.5 0.1 C5 Medical technician 0.9 0.4 0.9 0 2.6 0.1 1.6 0.2 C6 Nurse 0.2 0.9 9.5 1.0 0.5 7.5 3.1 4.4 C7 Office staff 0 1.3 1.2 0.3 6.2 4.3 3.0 2.7 C8 Other student 0 0 0 0 0.4 0.2 0.2 0.1 C9 Patient and patient representative 56.1 53.7 29.2 40.7 18.7 19.2 32.2 32.5 C10 Resident 4.8 5.2 1.0 0.5 0.8 0.2 2.0 1.5 C11 Self 28.6 31.2 33.9 41.0 55.8 48.6 41.8 42.6 C12 Supervisr 0.7 1.1 0.3 0 1.2 2.0 0.8 1.3 C13 Other 3.9 3.3 7.8 7.1 9.3 14.9 7.3 10.3 C14 Radigrapher 0 0 0.9 0 0.2 0.2 0.4 0.1 { Pre-implementatin survey. { Pst-implementatin survey. EHR implementatin, mre f their time was spent in maintaining the medical infrmatin system rather than n a general clerical assistance/ffice task. The ffice staff spent mre time with nurses and less time ding tasks n their wn after EHR implementatin. 7.4. Implementatin prcess The three data cllectin methds prvided cmplementary infrmatin n EHR implementatin and its impact n the clinic staff and their wrk. Accrding t the questinnaire survey, staff reprted increased dependency n the cmputer, which was cnfirmed by the increased amunt f time spent using the cmputer in the wrk analysis. Perceptins f the staff regarding the EHR implementatin (i.e. infrmatin received abut the EHR implementatin and input int the implementatin prcess) imprved after the EHR implementatin. This infrmatin was supprted by reprts by interviewees f a number f activities fr invlving end users (e.g., planning f training, inquiry by the prject team regarding prblems experienced by the end users). The questinnaire data analysis shwed a slight increase in wrklad. This may have been due t technical prblems and time pressure assciated with the EHR implementatin, issues that were described by the key prject members in the interviews. A number f interesting results emerge frm this case study. First, the EHR implementatin had sme impact n the perceived wrk cntent f clinic staff as measured by the survey, especially regarding increased dependency n cmputers that was related t increasing use f cmputers fr varius tasks. Secnd, the amunt f time spent by physicians n patient care (abut 50% f their time) did nt change with the EHR implementatin. A recent study f physician time use befre and after implementatin f an EHR system prvides a similar finding (Pizziferri et al. 2005). Pizziferri and clleagues (2005) fund that the mean verall time spent by physicians per patient did nt significantly change frm befre implementatin t after EHR implementatin. Third, there were majr changes in the wrk f clinical staff and ffice staff fllwing the EHR implementatin. Clinical staff spent mre time in cmputer entry and maintaining the medical infrmatin system; ffice staff als spent mre time in cmputer entry and maintaining the medical infrmatin system, and less time in distributing chart and transcriptin; hwever, these changes did nt induce an increase in time spent ding tasks n their wn. On the cntrary, ffice staff spent mre time in cntact with nurses after the EHR implementatin. In this case study, the EHR implementatin went relatively smthly, prbably because f a psitive climate existing in the clinic. A few implementatin issues culd have been imprved (e.g., clarifying the structure f the EHR implementatin rganisatin). Hwever, it seemed that the prject implementatin prcess was designed t identify emerging issues (e.g., reasns fr resistance t change) and t prvide slutins just-in-time. 7.5. Study limitatins and future research The data reprted in this paper are based n nly ne small family medicine residence clinic, and therefre cannt be generalisable t ther clinics. Hwever, it
18 P. Carayn et al. 3. Physici3flS (;{=O.20: df=3: flot Sigflilic3flt) 60,-----------------, 60 40 20 " Patil:mt 00'" b. Clinical slaff(l=11.73; dr=4: 1'<.05) Figure 4. Wrk analysis: cmparisn f pre-ehr and pst- EHR frequencies f cntact. (a) Physicians (w 2 ¼ 0.20; df ¼ 3; nt significant); (b) clinical staff (w 2 ¼ 11.73; df ¼ 4; p 5 0.05); (c) ffice staff (w 2 ¼ 7.04; df ¼ 4; p 5 0.05). 60,----------------, 5O+-------------j 40+-------- 30t--------,- 2Ot--------i " /'ijrse Palient c. Office staff(x l =7.04; df=4; 1'<.05) 0"''' IDP'" PstI prvides imprtant lessns regarding EHR implementatin and its evaluatin. First, an EHR implementatin shuld be cnsidered as a prject and shuld therefre utilise prject management cncepts and methds (e.g., prject structure, rles, timeline). This can help with the prcess itself, such as mnitring the implementatin and being aware f prblems with the implementatin. Secnd, attentin t the EHR implementatin as a prject can help anticipate the S' 60,-----------------, 50+----------1 40+----------\ 30t---------1 20 t--------,~,,+---==-----\ I'iJrse Oflice staff Patient S. 0"''' impact f the technlgy n the wrk f prviders and clinic staff and prvide imprtant infrmatin fr training. In the prcess f change, we came t understand several keys t a successful EHR implementatin prject:. Imprtance f analysing needs and preferences f medical prviders and key administratrs. A strng physician leader t champin the prject. Hiring a prject manager with dedicated time t lead the prject. Frming a prject leadership team f key persnnel frm clinical, ffice, and infrmatin system staff. Gathering needs f ther users early in the planning prcess. Obtaining buy-in by clinicians and ffice staff early in the prcess. Our study clearly shws the imprtance f using multiple data cllectin methds in rder t fully appreciate the range f human and rganisatinal factrs invlved in technlgy implementatin. The questinnaire survey prvided infrmatin n the EHR implementatin frm the viewpint f the clinic staff; the interviews with key prject persnnel allwed a better understanding f the EHR implementatin prcess and its characteristics; the wrk analysis allwed an in-depth evaluatin f the impact f the EHR technlgy n the wrk f different jb categries. We wuld like t recmmend that future research n the impact f EHR technlgy implementatin use multiple data cllectin methds, including bth qualitative and quantitative appraches. 8. Cnclusin In this paper, we described a case study f the implementatin f an EHR system in a small family practice clinic. Quantitative and qualitative data cllectin methds prvided cmplementary infrmatin n hw emplyees f the small clinic perceived their wrk and the implementatin f the EHR system. The data shwed few changes in wrk patterns f physicians due t the use f EHR, except fr the increased cmputer entry. On the ther hand, there were majr changes in the wrk f clinical staff and ffice staff. A cmprehensive examinatin f the human and rganisatinal factrs as t EHR implementatin was reprted in this case study. This can prvide valuable inputs fr a successful implementatin f EHR in small clinic settings. The results f ur study highlight the need t cnsider EHR implementatin as a majr
Behaviur & Infrmatin Technlgy 19 scitechnical change prject. Once a healthcare rganisatin has decided t purchase an EHR system, principles f prject management and technlgical change need t be applied t ensure rapid and efficient uptake by end users and t minimise disruptins t wrk flw (Smith and Carayn 1995, Krunka and Carayn 1999). References Ash, J.S. and Bates, D.W., 2005. Factrs and frces affecting EHR system adptin: reprt f a 2004 ACMI discussin. Jurnal f the American Medical Infrmatics Assciatin, 12, 8 12. Bailey, J.E. and Pearsn, S.W., 1983. Develpment f a tl fr measuring and analyzing cmputer user satisfactin. Management Science, 29, 530 545. Berner, E.S., Detmer, D.E., and Simbrg, D., 2005. Will the wave finally break? A brief view f the adptin f electrnic medical recrds in the United States. Jurnal f the American Medical Infrmatics Assciatin, 12, 3 7. Caplan, R.D., Cbb, S., French, J.R.P., Harrisn, R.V., and Pinneau, S.R., 1975. Jb demands and wrker health. Washingtn, D.C.: U.S. Gvernment Printing Office. Carayn, P., 1994. Research n preventin strategies in autmated ffices. In: G.E. Bradley and H.W. Hendrick, eds. Human Factrs in Organizatinal Design and Management IV. Amsterdam: Elsevier, 707 712. Carayn, P. and Haims, M.C., 2001. Infrmatin & cmmunicatin technlgy and wrk rganizatin: achieving a balanced system. In: G. Bradley, ed. Humans n the Net-Infrmatin & Cmmunicatin Technlgy (ICT), Wrk Organizatin and Human Beings. Stckhlm: Prevent, 119 138. Carayn, P. and Karsh, B., 2000. Scitechnical issues in the implementatin f imaging technlgy. Behaviur and Infrmatin Technlgy, 19, 247 262. Carayn-Sainfrt, P., 1992. The use f cmputers in ffices: impact n task characteristics and wrker stress. Internatinal Jurnal f Human Cmputer Interactin, 4, 245 261. Centers fr Disease Cntrl and Preventin (CDC), 2000. Healthy peple 2010: US Department f Health and Human Services. Available nline at http://www.health. gv/healthy peple/dcument/default.htm. Chin, J.P., Diehl, V.A., and Nrman, K.L., 1988. Develpment f an instrument measuring user satisfactin f the human-cmputer interface. In: Prceedings f ACM SIGCHI (pp. 213 218), New Yrk: ACM/SIGCHI. Ck, J. and Wall, T.D., 1980. New wrk attitudes measures f trust, rganizatinal cmmitment, and persnal need nn-fulfillment. Jurnal f Organizatinal Psychlgy, 53, 39 52. Davis, F.D., 1989. Perceived usefulness, perceived ease f use, and user acceptance f infrmatin technlgy. MIS Quarterly, 13, 319 340. Easn, K., 1988. Infrmatin Technlgy and Organizatinal Change. Lndn: Taylr & Francis. Fntaine, B.R., Speedie, S., Abelsn, D., and Wld, C., 2000. A wrk-sampling tl t measure the effect f electrnic medical recrd implementatin n health care wrkers. The Jurnal f Ambulatry Care Management, 23, 71 85. Greenberger, D.B., Strasser, S., Cummings, L.L., and Dunham, R.B., 1989. The impact f persnal cntrl n perfrmance and satisfactin. Organizatinal Behavir and Human Decisin Prcesses, 43, 29 51. Institute f Medicine, 2000. T Err Is Human: Building a Safety Health System. Washingtn, D.C.: Natinal Academy Press. Institute f Medicine, 2001. Crssing the Quality Chasm: A New Health System fr the 21st Century. Washingtn, D.C.: Natinal Academy Press. Jha, A.K., Ferris, T.G., Dnelan, K., DesRches, C., Shields, A., Rsenbaum, S., and Blumenthal, D., 2006. Hw cmmn are electrnic health recrds in the United States? A summary f the evidence. Health Affairs, 25, w496 w507. Karsh, B., 2004. Beynd usability: designing effective technlgy implementatin systems t prmte patient safety. Quality and Safety in Health Care, 13, 388 394. Kawamt, K., Hulihan, C.A., Balas, E.A., and Lbach, D.F., 2005. Imprving clinical practice using clinical decisin supprt systems: a systematic review f trials t identify features critical t success. British Medical Jurnal, 330, 765. Krunka, C. and Carayn, P., 1999. Cntinuus implementatin f infrmatin technlgy: the develpment f an interview guide and a crss-natinal cmparisn f Austrian and American rganizatins. Human Factrs and Ergnmics in Manufacturing, 9, 165 183. Krunka, C., Weiss, A., and Zauchner, S., 1997. An interview study f cntinuus implementatins f infrmatin technlgy. Behaviur & Infrmatin Technlgy, 16, 3 16. Linder, J.A., Ma, J., Bates, D.W., Middletn, B., and Staffrd, R.S., 2007. Electrnic health recrd use and the quality f ambulatry care in the United States. Archives f Internal Medicine, 167, 1400 1405. McLaney, M.A. and Hurrell, J.J.J., 1988. Cntrl, stress, and jb satisfactin in Canadian nurses. Wrk and Stress, 2, 217 224. Middletn, B., Hammnd, W.E., Brennan, P.F., and Cper, G.F., 2005. Accelerating U. S. EHR adptin: hw t get there frm here. Recmmendatins based n the 2004 ACMI retreat. Jurnal f the American Medical Infrmatics Assciatin, 12, 13 19. Murray, M.D., Ls, B., Tu, W., Eckert, G.J., Zhu, X.H., and Tierney, W.M., 1999. Wrk patterns f ambulatry care pharmacists with access t electrnic guidelinebased treatment suggestins. American Jurnal f Health- System Pharmacy, 56, 225 232. Ohsfeldt, R.L., Ward, M.M., Schneider, J.E., Jaana, M., Miller, T.R., Lei, Y., et al., 2005. Implementatin f hspital cmputerized physician rder entry systems in a rural state: feasibility and financial impact. Jurnal f the American Medical Infrmatics Assciatin, 12, 20 27. Overhage, J.M., Evans, L., and Marchibrda, J., 2005. Cmmunities readiness fr health infrmatin exchange: the natinal landscape in 2004. Jurnal f the American Medical Infrmatics Assciatin, 12, 107 112. Pizziferri, L., Kittler, A.F., Vlk, L.A., Hnur, M.M., Gupta, S., Wang, S., et al., 2005. Primary care physician time utilizatin befre and after implementatin f an eletrnic health recrd: a time-mtin study. Jurnal f Bimedical Infrmatics, 38, 176 188.
20 P. Carayn et al. Quinn, R., Seashre, S., Kahn, R., Mangin, T., Cambell, D., Staines, G., et al., 1971. Survey f wrking cnditins: final reprt n univariate and bivariate tables. Washingtn, D.C.: U.S. Gvernment Printing Office. Dcument N. 2916-0001. Reeder, L.G., Schrama, P.G., and Dirken, J.M., 1973. Stress and cardivascular health: an internatinal cperative study I. Scial Science and Medicine, 7, 573 584. Sainfrt, F. and Carayn, P., 1994. Self-assessment f VDT peratr health: hierarchical structure and validity analysis f a health checklist. Internatinal Jurnal f Human Cmputer Interactin, 6, 235 252. Seashre, S.E., Lawler, E.E., Mirvis, P., and Cammann, C.. eds. 1983. Observing and Measuring Organizatinal Change: A Guide t Field Practice. New Yrk: Jhn Wiley. Sittig, D.F., 1993. Wrk-sampling: a statistical apprach t evaluatin f the effects f cmputers n wrk patterns in healthcare. Methds f Infrmatin in Medicine, 32, 167 174. Smith, M.J. and Carayn, P., 1995. New technlgy, autmatin, and wrk rganizatin stress prblems and imprved technlgy implementatin strategies. Internatinal Jurnal f Human Factrs in Manufacturing, 5, 99 116. Thmpsn, T.G. and Brailer, D.J., 2004. The decade f health infrmatin technlgy: delivering cnsumercentric and infrmatin-rich health care. Washingtn, D.C.: U.S. Department f Health & Human Services. Appendix Time Activity Functin Cntact Ntes 13:30 14:00 14:30 15:00 15:30 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:30 20:00 Time leave: Activity Functin Cntact A1, absent F1, accmpanying patients C1, billing crdinatr A2, caring fr patient F2, assisting physician, dctrs r medical technician C2, dctr A3, cmputer entry F3, billing activities C3, manager A4, dictatin F4, checking message C4, medical student A5, meeting 1þ F5, data review and retrieval C5, medical technician A6, meeting 3þ F6, distributing chart/master file/mail C6, nurse A7, perfrming lab wrk F7, examinatin r treatment f patient C7, ffice staff A8, phne F8, general clerical assistance/ffice task C8, ther student A9, preparing F9, maintaining equipments, instruments, supplies and medicatins C9, patient and patient representative A10, reviewing check F10, maintaining lab reprts C10, resident A11, supervising F11, maintaining medical infrmatin system C11, self A12, typing/writing/signing F12, maintaining patient s infrmatin recrd C12, supervisr A13, ther F13, participating in resident and student educatin C13, ther F14, perfrming tests C14, radigrapher F15, Preparing fr examinatins and surgical prcedures F16, prviding instructin/infrmatin F17, purchasing and making inventry arrangement F18, reprting prblem F19, scheduling F20, training F21, transcriptin F22, ther