The objective of this research was to determine

Transcription

1 FEATURE: CPOE Time-Saver A Time-Motio Study to Evaluate the Impact of EMR ad CPOE Implemetatio o Physicia Efficiecy By Ame A. Amusa, PMP, MS; Scott Toge, MD; Stuart M Speedie, PhD; ad Adrew Melli, MD, MBA Keywords EMR, CPOE, time motio study. Abstract The objective of this research is to determie the impact of EMR ad CPOE implemetatio o physicia efficiecy. A time series observatioal study was coducted withi a hospital settig at six weeks pre-implemetatio, six weeks post-implemetatio ad five moths post-implemetatio. All 19 subjects were observed twice with oe patiet per observatio. Physicia follow-up roudig times per patiet were measured. Physicias demostrated a mea total roudig time of miutes (pre-implemetatio); miutes (six weeks post-implemetatio); ad miutes (five moths post). Overall, the results showed a statistically sigificat F value=8.26>1 (p=0.0011) that sigifies a reductio i physicia roudig time withi the hospital settig followig EMR implemetatio. Results also showed overall stadard deviatios of 6.96 miutes (preimplemetatio); 5.13 miutes (six weeks post); ad 3.69 miutes (five moths post), possibly sigifyig a reductio i variability ad a arrower distributio of roudig times with icreased similarity i physicias roudig patters. The objective of this research was to determie the impact of electroic medical record (EMR) ad computerized physicia order etry (CPOE) implemetatio physicia efficiecy at Uited Hospital. The focus of this research is to determie if there is reductio i physicia rouds time from pre- to post-implemetatio of the system. Backgroud Physicia resistace to EMR adoptio is a major barrier to successful EMR implemetatio. Poissat et al., i a systematic literature review, states, the importace of evaluatig time efficiecy i documetatio is also related to the observatio that icreased time for documetatio is oe of the most commoly stated barriers to successful implemetatio of a EMR. 1 Movig from a paper-based system to a EMR creates challeges, such as chages to the physicia s workflow, processes, etc. that caot be overlooked. The goal of this research is to examie the effects of these chages affect providers daily activities. Curret studies show that healthcare providers have several importat eeds improved access to iformatio, elimiatio of chart pulls (Bates et al.), access to decisio support that are ot beig met. 2 Quality improvemet through EMR depeds greatly o physicias use of the system to complete day-to-day activities. 3 It is thought that the adoptio ad better use of IT is essetial for improved quality of care. 4 Other studies have cited the importace of EMR i reducig documetatio time ad ehacig efficiecy A survey coducted by O Brie at the JKL Healthcare System oted that 90 percet of the participatig physicias affirmed that EMR made their work much easier. 11 Galvi, et al. state that [it] was clear from the begiig that CPOE was a iitial leap i the larger revolutio i health iformatio techology ecessary to achieve improved quality ad efficiecy. 12 Studies show that CPOE improves efficiecy ad medicatio volume 22 / umber 4 FALL 2008 jhim 31

2 ad order turaroud Some studies show a o-sigificat differece i time whe usig EMR vs. a paper-based system. 18 Though some studies show icreased time spet i the CPOE system due to chages i workflow processes, 10,19 this reduced as kowledge of ad experiece with the system improved. 20 Puffer, et al., state, Perceived efficiecy has bee show to have a sigificat impact o overall satisfactio with electroic cliical systems that, i tur, supports wider adoptio of the techology. 21 We eed to perform more evaluatios of curret systems to determie how well they eable healthcare orgaizatios to reap the beefits of efficiecy. A pre- ad post-implemetatio evaluatio is oe way of achievig this goal ad fidig the best ways to resolve outstadig problems ad/or cocers. 22 Several studies performed i the past have focused o the impact of EMR or CPOE withi the primary care, specialty cliics or outpatiet settigs, but very few studies have bee coducted i ipatiet settigs. 23 To evaluate the impact a ew EMR/CPOE system will have o physicias efficiecy at Uited Hospital i St. Paul, Mi., a beefit measuremet aalysis was performed. Part of this aalysis measured the time it took providers to perform various patiet care activities, particularly rouds. The fial aalysis would determie if the use of EMR/CPOE reduced time spet o patiet roudig. Curret workflow. The physicias at Uited Hospital is a private group with o residets or medical studets assistig i the patiet care. Geerally, they work idividually except whe they assiged a list of patiets ad their locatio each morig. (They roud a average of 15 patiets per day.) Upo arrival to a patiet s uit, the physicia visits the urses statio to review the patiet s paper chart, which icludes progress otes, lab results, curret medicatios ad other pertiet iformatio. Leavig the chart at the statio, the physicia visits the patiet. After the visit, the physicia walks back to the urses uit to update the chart. Notes are dictated via the Dictaphoe system. Orders are placed withi the STAR applicatio or had writte i the otes for the urses or health uit coordiators to eter ito the STAR system. Patiet-related istructios are verbally discussed with urses, peers or caregivers. The physicia the moves o to the ext uit ad the ext patiet o the assiged list. The existig system ad its limitatios. Allia s decisio to implemet a EMR system stems from its goal to have a sigle itegrated system betwee the hospitals ad cliics, as well as betwee reveue cycle ad cliical systems described at Allia as Oe patiet. Oe record. 24,25 The system at Uited prior to Excellia was a traditioal, paper-based health record, with the exceptio of electroic ursig documetatio i the ICU. Shortliffe (pg. 415) described the paper-based system as highly iadequate for meetig the eeds of moder medicie. 26 It is ot a very efficiet system for patiet iformatio storage ad retrieval; chart documetatio is error-proe; ad collectig patiet-related iformatio is time cosumig. (For example, the medicatio admiistratio record, lab results ad progress otes may all be i differet locatios.) New Workflow Physicias are added to the treatmet team of each assiged patiet withi the applicatio. Each provider has a provider list that cotais all patiets associated with the physicia through the treatmet team relatioship ad ca be viewed through the system. The order of seeig patiets depeds o the physicia s discretio ad the acuity of the patiet. Upo arrival to the patiet uit, the physicia logs ito ay available workstatio to retrieve ad review the patiet s electroic chart. He/she logs out of the system ad visits the patiet. After the visit, the physicia walks back to the urses uit, logs ito the system, locates the patiet, ad opes the chart ad roudig avigator. Progress otes are writte usig available templates (smart text), smart list ad smart phrases whe ecessary ad filed. Orders/medicatio are reviewed from the same avigator Physicia resistace to EMR adoptio is a major barrier to successful EMR implemetatio. ad are either discotiued, modified or remai uchaged. The physicia may commuicate verbally with the urse or place a commuicatio order i the system. The physicia the closes the patiet chart ad logs out of the system. With EMR patiet care is doe through a computer, icludig CPOE, cliical messagig, physicia documetatio, results review, deficiecy maagemet ad ursig documetatio review. For the aalysis, physicias were allowed to cotiue to dictate iformatio, but all other physicia documetatio, icludig progress otes, were doe through the EMR. Uited Hospital madated complete adoptio of the EMR/ CPOE by all providers, except those who had 12 or fewer patiet cotacts per year. Workflow chaged for a umber of idividuals. For example, some physicias ow roud large group of patiets ad place orders, the complete the otes off the patiet uit. Other physicias complete the H&P ote directly i the computer at the patiet s bedside. Some providers ow also access the chart from home to follow up o iformatio ad place ew orders. Method Settig. This research was performed at Uited Hospital. Uited Hospital is oe of the largest hospitals i the Twi Cities East metro area, providig a complete variety of health care services to more tha 200,000 people each year. 27 This 426-bed healthcare istitutio has more tha 1,300 healthcare professioals with a broad rage of specialties. I 2006, there were 28,165 ipatiet admissios ad 101,196 outpatiet visits. It is owed by Allia Hospitals ad Cliics (a ot-for-profit etwork of hospitals, cliics ad other healthcare services, providig care throughout Miesota ad wester Wiscosi). Allia has 11 hospitals, 42 cliics, 22 hospital based cliics, 14 commuity pharmacies ad four ambulatory care ceters. Excellia is Allia s bradig of the EMR implemeted at Uited Hospital. Excellia iteds to provide a quick ad easy access to patiet charts ad improve commuicatio betwee physicias treatig the same patiet. It also provides timelier turaroud o 32 jhim FALL 2008 volume 22 / Number 4

3 Table 1: Total miutes spet o each activity ad category pre ad post implemetatio for all physicias. lab or radiology test results; leads to better educatioal opportuities for patiets; ad offers improved security for patiet privacy. 28 To date, eight Allia hospitals ad almost all of the Allia cliics have fully implemeted the system i the ipatiet ad outpatiet eviromet. Subjects. The target populatio was the hospitalists at Uited Hospital. The hospitalists group at Uited is made up of highly traied board-certified physicias. 29 This group collaborates with primary care providers to provide seamless cotiuity of patiet care. They work o a weekly rotatio, which posed some challeges durig the iitial phase of this research as it was difficult to recruit physicia subjects. The itet was to recruit 20 to 30 hospitalists for this research to icrease the statistical power ad precisio of the data aalysis. After obtaiig approval from both the Allia s Health System IRB ad Uiversity of Miesota IRB, the pricipal ivestigator wet to differet hospital uits ad persoally approached providers, explaied the purpose of the study ad haded a copy of the coset forms to these providers. Due to busy ad varyig schedules, oly 11 Uited physicias were recruited. At the ed of the recruitig period, five additioal hospitalists from the ASPEN Medical group agreed to participate i the study. There were also two specialty physicias from the St. Paul Heart Cliic ad oe from the St. Paul Lug Cliic who agreed to participate i the study. All of these providers do rouds o patiets at Uited Hospital withi the same uits, usig the same EMR ad CPOE applicatios. The fial sample size was a total of 19 physicias. Study desig. A time series observatioal time-motio study was performed at six weeks pre-implemetatio; six weeks post-implemetatio; ad five moths post-implemetatio. The observatios focused primarily o the physicias roudig workflow processes per patiet. The time (i miutes) of each of the three workflow tasks was measured withi each observatio sessio. It was measured for the followig tasks: Pre roudig: Iformatio gatherig time Iformatio review time. Seeig patiet This study does ot iclude detailed observatio of processes ivolvig direct patiet care because of privacy cosideratios. However, the time i miutes spet by the physicia withi the patiet s room was recorded ad icluded i this research. Post roudig Time take to write patiet otes Time to place orders (writte ad/or verbal) Commuicatio time (time spet i discussig with the urse or other healthcare providers directly ivolved i the patiet care). For each sessio, a start ad ed time (sigifyig physicia arrival ad departure time) was recorded. Data collectio. Data was collected over a period of two weeks for the pre-implemetatio ad the six-week ad five-moth period post-implemetatio. Data for six weeks pre-implemetatio was collected from April 2 through April 14, 2007, betwee the hours of 8:30 a.m. ad 12:30 p.m. Data for the six weeks post go-live was collected from Jue 16 through Jue , betwee the hours of 8:30 a.m. ad 12:30 p.m. For five moths after go-live, observatio sessios were doe from September 11 to October 8, 2007, betwee the hours of 8:30 a.m. ad 12:30 p.m. A observer followed each provider as they rouded. A miimum of two complete sessios with oe patiet per sessio were recorded i miutes (roudig secods to the earest miute) o a Excel spreadsheet. Zero times were recorded for less tha 30 secods ad i cases where o activity was performed. Times recorded for 30 secods ad above were rouded to the earest miute. Roudig times for admissios ad discharge were ot icluded i this aalysis as it took providers loger times to perform these activities. Data aalysis. The SAS (Statistical Aalysis Software) versio 9.1 TS level 1M3 ad Microsoft 2003 Excel was used for data aalysis. Alpha (type I error) was set to The SAS GLM (Geeral Liear Model) procedure was used to aalyze the roudig time (miutes). The decisio to use the SAS GLM procedure was due to the logitudial (multiple measures o oe perso) ature of the data collected durig this study. Proc UNIVARIATE was also used to perform aalysis o the data collected ad the followig statistical results: mea ad stadard deviatio was reported. Table 1 shows the total time spet o each activity pre- ad post-implemetatio by all physicias. The iformatio gatherig activity showed a 3.5-miute reductio i time at six weeks post-implemetatio ad a 100 percet time reductio five moths post-implemetatio. The 100 percet reductio was due to zero times recorded for periods less tha 30 volume 22 / umber 4 FALL 2008 jhim 33

4 Table 2: The average total times i miutes ad stadard deviatios for all categories. secods ad i cases where o task was performed. The iformatio review activity had a four-miute reductio i time six weeks post, ad the had a 18.5-miute reductio five moths post. Although a reductio i the time a provider spet with a patiet was ot aticipated, there was a 16-miute reductio six weeks after, ad a 24.5-miute reductio five moths post. I otes activity, there was a 14-miute reductio six weeks post ad a 28.5-miute reductio five moths post. Time spet i the orders activity icreased from three to 39 miutes six weeks post ad the approximately 27 miutes five moths post. Commuicatio activity declied drastically six weeks post by 33 miutes ad 42 miutes five moths post. A possible reaso for this is a reductio i physicia dictatio time ad icreased use of electroic commuicatio. This is the total time spet o roudig activities (pre-roudig, seeig patiet ad post-roudig). Pre-roudig had a 7.5-miute reductio i total time at six weeks ad a 40-miute reductio five moths after go-live. Seeig patiets had a 16-miute reductio i total time at six weeks ad a 24.5-miute reductio i total time five moths after go-live. Table 1 also shows that post-roudig had a 11-miute reductio i total time six weeks after go-live ad a 46-miute reductio i total time five moths after go-live. The GLM Procedure Repeated Measures Aalysis of Variace simply called Repeated Measures ANOVA aalysis was performed o the time motio dataset for pre-roudig, seeig patiets ad post-roudig. 30 There were a total of 38 observatios with the 19 subjects. Aalysis also was performed o the average total time spet roudig o patiets per provider for the periods of pre-implemetatio, six weeks post ad 5 moths post implemetatio. A test for ormality was performed to verify that the data was ormally distributed. Shapiro-Wilk test is the most commoly used test for this aalysis. A p-value greater tha alpha sigifies a ormal distributio of data. SAS recommeds usig Shapiro-Wilk test for small sample size ot greater tha Before proceedig with a GLM procedure usig uivariate or multivariate aalysis, a test for sphericity was performed. The Mauchly s sphericity assumptio (spherical matrix equals zero with equal variaces ad covariace) must be met to perform uivariate aalysis. I order to evaluate the impact of roudig to zero miutes, we tried the alterative of roudig to 0.5 miutes ad did the aalysis. Results Results from the ANOVA aalysis showed that the data from six weeks post-implemetatio was ormally distributed while data from six weeks pre-implemetatio ad five moths post-implemetatio did ot appear to be ormally distributed but rather skewed to the right. Results from the Shapiro-Wilk s test of ormality for six weeks pre-implemetatio was o-sigificat (p=0.081) meaig that the dataset is ormally distributed. Similarly results were observed for the period of six weeks post-implemetatio (p-value=0.72) ad five moths post-implemetatio (p-value=0.56). Table 2 shows average total times per physicia ad stadard deviatios for the pre-roudig, seeig patiets ad the postroudig categories for each of the three time periods. The aalysis shows that the sphericity assumptio was violated with a chi-square of with 2 degrees of freedom which has a p-value= This meas that the uivariate test would likely yield ivalid results. Accordigly, the repeated measures aalysis of variace was employed. The multivariate F-test for 34 jhim FALL 2008 volume 22 / Number 4

5 time (pre, six weeks post ad five moths post) had a sigificat F value=15.66 with a correspodig p-value of For the pre-roudig aalysis, the sphericity assumptio was met with a chi-square of 1.47 ad a associated p-value of which is greater tha alpha (α =0.05). The uivariate test used for this aalysis showed a sigificat result with a p-value of This shows a cosiderable chage i physicia pre-roudig time across the trials. Similarly for the seeig patiet category, the sphericity assumptio was met (chi-square=0.86 ad a p-value=0.65) ad we proceeded with the Uivariate test that showed a o sigificat F=1.72 with a p-value=0.19. I the post-roudig category, the sphericity assumptio was also met with a chi-square of 4.06 ad a associated p-value=0.13. The uivariate test showed a sigificat F=3.01 with a p-value of This result demostrates a dowward tred that approached sigificace. The overall results from roudig to 0.5 miutes was statistically sigificat with p= This did ot show ay cosiderable differece from the result obtaied from previous aalysis. Discussio The purpose of this study was to determie the impact of a EMR & CPOE implemetatio at Uited Hospital o physicia hospitalists efficiecy. Our fidig shows a statistically sigificat result with a p-value of , which sigifies a reductio i the time spet o roudig by physicias per patiet. There were iitial cocers of possible bias to the study due to the three additioal specialty physicias added to the subjects. The decisio was made to iclude these physicias give that the itet of the aalysis was ot to target the specialty, but rather the tasks beig performed. Hospitalists ad specialists performed similar tasks at the time of observatio ad used the same EMR ad CPOE withi the Uited Hospital. The sigificat differece i pre-roudig compared to the other roudig categories was due to the fact that the providers had to sped less time lookig for charts or patiet iformatio from disparate systems. Some of the providers said that they did all their pre roudig activities cosistig of iformatio gatherig ad review prior to comig to the patiet s uits. This made the process much faster for them. However, i this research, all participats performed all the activities while roudig o patiets ad as a result of that the outcome of this study was ot biased i this respect. The zero times see i Table 1 above for order (pre-implemetatio) ad iformatio review (five moths post) were due to physicias possibly hadwritig patiet orders while writig patiet otes or commuicated this verbally to other cliical staff (e.g., urses). Very few physicias logged i to the Allia STAR system (used pre-excellia implemetatio) at the time of observatios. The observer had o way of kowig what was beig writte. It is likely that some of the order times i the pre implemetatio may have bee added to the ote takig time. As a result of that, the amout of time i miutes recorded for the otes activity may have bee over estimated. The fial data aalysis above also shows o sigificat differece i times through the trials for the category seeig patiet. The time a provider spet i the patiet room was cosidered as time spet seeig patiet. The actual time the physicia spet directly o patiet care while iside the patiet s room may have bee lesser tha the time reported. The observer did ot go ito the patiets rooms ad hece was uable to tell from the observatio the amout of time spet directly o patiet care. This is subject to further study. The results shows that the highest stadard deviatio six weeks post-implemetatio for pre-roudig ad seeig patiets (Table 2). Seeig patiets ad post-roudig saw the highest stadard deviatio durig the pre-implemetatio phase. A possible reaso for that could be icreased variability or wide distributio of roudig times amog physicias. Physicias had dissimilar roudig patters as there was o stadardized method of roudig. The overall stadard deviatios showed a decliig tred sigifyig a reductio i variability ad a arrower distributio of roudig times with icreased similarity i physicias roudig patter. There was o cotrol over the acuity of the cases which could possibly have iflueced how the physicias spet their time. The post-roudig category evideces a decliig tred i average time which approached statistical sigificace. The lack of a sigificat declie may have bee due to some providers havig The hypothesis that the implemetatio of the EMR ad CPOE will result i a icrease i physicia s efficiecy was accepted. to sped more time eterig procedure, lab ad commuicatio orders i the CPOE system as this was madatory with the ew system. I the previous paper world, these orders were mostly commuicated to the urse verbally or through patiet otes. Oe likely but ukow impact o this study is the Hawthor effect (which is the awareess of the presece of the observer by the physicias). However i this research, this effect became miimal with icreased umber of observatios. Coclusio ad Future Directios I coclusio, the hypothesis that the implemetatio of the EMR ad CPOE will result i a icrease i physicia s efficiecy with a reductio i the time spet o roudig patiets was accepted. This research oly observed physicias whe they were roudig o the floor ad the focus was o hospitalists. It did ot however iclude admissios ad/or discharges. The results might be differet if these categories of patiets bee icluded. For future studies, it is recommeded that a larger physicia sample size be eeded ad the umber of trials be icreased i order to improve the precisio ad accuracy of the aalysis. Aother cosideratio for future research is to radomly select a large umber of providers, sed out a survey to selected providers to further group them by work life style. Some providers prefer reviewig patiets iformatio i batches of three or four prior to seeig patiets. Others prefer seeig patiets i batches ad go back to the physicias louge to write all the otes, place orders, volume 22 / umber 4 FALL 2008 jhim 35

6 etc. Varyig work styles ca be a form of bias to the research if ot take ito cosideratio. These physicias atural roudig workflow processes were chaged i order to capture data more accurately. I the future it will be beeficial to track the time physicias sped o roudig over some observatioal period rather tha per patiet to determie the impact the EMR/CPOE will have o atural workflow processes. I also recommed that future research address the followig questios: Would it have bee better to record times to the secod? Were the results skewed by primarily followig heavy users of the EMR ad CPOE like hospitalists? How would surgeos or differet specialists do? Is there a differece i the kid of visit, i.e. regular roudig vs. admissio or discharge? What would the umbers look like after a year? If we icluded time o the computer i the patiet s room, what effect would this have? Will groupig of physicias by geder chage the results? Ad, fially, will geeral computer proficiecy make ay differece i the fial result? Future research should also cosider comparig the results from this aalysis to previous research. Ackowledgemets This research was fuded by the Uited Hospital Foudatio. Our appreciatio goes to Dr. Do Coelly ad Dr. Boie Westra. We would like to ackowledge the Uited Hospital Services hospitalists for participatig i this study. The hospitalists from the ASPEN Medical group, St. Paul Heart Cliic ad St. Paul Lug Cliic for participatig i this research. Thak you Jeifer Raiter for helpig with the physicia schedules. Thaks to Jae Tobias (Allia Excellia Support-Lead) ad Sharo Bresaha (Maager-Allia Excellia support). Thaks to Susa Telke (Istructor Divisio of Biostatistics Uiversity of Miesota) for the statistical advice provided. Thaks to Lida Fletcher ad Beverly Collis for reviewig this research literature. Fially we would like to ackowledge the etire staff ad studet of the Uiversity of Miesota Health Iformatics Divisio. Thaks to Tola, Samuel ad Joh Amusa for all their uderstadig ad support throughout this research. JHIM Ame A Amusa, PMP, MS, is a member of the Health Iformatics Divisio, Uiversity of Miesota. Scott Toge, MD, is Excellia Medical Director, Uited Hospital, St. Paul, Miesota. Stuart M Speedie, PhD, is a member of the Health Iformatics Divisio, Uiversity of Miesota. Adrew Melli, MD, MBA, was Excellia Director (former), Allia Hospitals ad Cliics, Miesota. Refereces 1. Poissat L, Pereira J, Tambly R, Kawasumi Y. The impact of electroic health records o time efficiecy of physicias ad urses: a systematic review. J Am Med Iform Assoc. 2005;12(5): Bates DW, Ebell M, Gotlieb E, Zapp J. A proposal for electroic medical records i US primary care. J Am Med Iform Assoc. 2003;10(1): Miller RH, Sim I. Physicias use of electroic medical records: barriers ad solutios. ]Health Aff.2004;23(2): Preparig for implemetatio. Available at: sectio.php?sectio=3. Accessed o December 8, Sujasky WV. The beefits ad challeges of a electroic medical record: much more tha a word-processed patiet chart. West J Med. 1998;169(3): Chaudhry B, Wag J, Wu S, Maglioe M, Mojica W, Roth E, Morto SC, Shekelle PG. Systematic review: impact of health iformatio techology o quality, efficiecy ad costs of medical care. A Iter Med. 2006;144(10): Toward a high-performig EMR health system. EMR vs. EHR: is there a differece? Available at: Accessed o December 8, Ford EW, Meachemi N, Phillips MT. Predictig the adoptio of electroic health records by physicias: whe will health care be paperless. J Am Med Iform Assoc. 2006;13(1): Kha SA, Paye PRO, Johso SB, Bigger JT, Kukafka R. Modelig cliical trials workflow i commuity practice settigs. AMIA Au Symp Proc. 2006;2006: Overhage JM, Perkis S, Tierey WM, McDoald CJ. Cotrolled trial of direct physicia order etry: effects o physicias time utilizatio i ambulatory primary care iteral medicie practices. J Am Med Iform Assoc. 2001;8(4): O Brie MS. Implemetatio of the EPIC electroic medical record/physicia order-etry system. JHIM. 2006:FELLOW PROJECT: Galvi RS, Delbaco S, Milstei A, Belde G. Has the Leapfrog Group had a impact o the health care market? Health Aff. 2005;24(1): Foster RA, Atoelli PJ. Computerized physicia-order etry: are we there yet? Otolarygol Cli North Am. 2002;35(6): Ali NA, Mekhjia HS, Kueh PL, Betley TD, Kumar R, Ferketich AK, Hoffma SP. Specificity of computerized physicia order etry has a sigificat effect o the efficiecy of workflow for critically ill patiets. Crit Care Med. 2005;33(1): Cordero L, Kueh L, Kumar RR, Mekhjia HS. Impact of computerized physicia order etry o cliical practice i a ewbor itesive care uit. J Periatol. 2004;24(2): Lehma ML, Brill JH, Skarulis PC, Keller D, Lee C, Physicia Order Etry impact o drug tur-aroud times, Proc AMIA Symp. 2001; La Cour V, Hellster-Hauerslev C. Evaluatig the implemetatio ad use of a computerized physicia order etry system: a case study. Stud Health Techol Iform. 2007;130: Newmarka L, Kittlera A, Lippicotta M, Volk LA, Hoour MM, Gupta S, Wag SJ, Bates DW. Does usig a electroic health record require more time for primary care physicias? J I Med Iform Assoc. 2004; Cheg CH, Goldstei MK, Geller E, Levitt RE. The effects of CPOE o ICU workflow: a observatioal study. AMIA Au Symp Proc. 2003; Holligworth W, Devie EB, Hase RN, Lawless NM, Comstock BA, Wilso- Norto JL, Tharp KL, Sulliva SD. The impact of e-prescribig o prescriber ad staff time i ambulatory care cliics: a time motio study. J Am Med Iform Assoc. 2007;14(6): Puffer MJ, Ferguso JA, Wright BC, Osbor J, Ashus AL, Cahill BP, Kamath J, Rya MJ. Parterig with cliical providers to ehace the efficiecy of a EMR. JHIM. 2007;21(1): Optimizig your EMR (Practice efficiecy & ogoig improvemet). Available at: Accessed o December 8, jhim FALL 2008 volume 22 / Number 4

7 23. Pizziferri L, Kittler AF,Volk LA, Hoour MM,Gupta S, Wag S, Wag T, Lippicott M, Li Q, Bates DW. Primary care physicia time utilizatio before ad after implemetatio of a electroic health record: a time-motio study. J Biomed Iform. 2005;28(3): HIStalk Iterviews Kim Pederso, VP of Excellia, Allia Hospitals & Cliics. Posted May 23, Available at: iterviews_kim_pederso_vp_of_excellia_allia_hos.htm. Accessed December 8, Pederso K, Truitt J, Melli A, Hery S, Heichert S, Nueske K, Buhr K, Aldrich T, Aaress P, Lambert M, Scott L, Boyd T, Hirsch C, Bresaha S, Scheider K. Allia Hospitals & Cliics EMR implemetatio. 2007; HIMSS Davies award: Shortliffe EH. The evolutio of electroic medical records. Acad Med. 1999:74(4): Allia Medical Cliic-Uited Hospital. Available: com/ahs/uited.sf/page/aboutus. Accessed o December 8, Allia lauches electroic medical record system at Abbott Northwester Hospital. MINNEAPOLIS, July 26, Available at: ews.sf/ewspage/aw_7_26_05. Accessed o December 8, Allia Medical Cliic- Uited Hospitalist Service, AMC.sf/page/AMC_UHS. Accessed o December 8, Repeated Measures ANOVA Usig SAS PROC GLM, 31 July 1997, Usage Note: Stat-40, Copyright , ACITS, The Uiversity of Texas at Austi, Accessed o December 8, volume 22 / umber 4 FALL 2008 jhim 37