Electronic Medical Records: Lessons from Small Physician …€¦ · Electronic Medical Records: Lessons from Small Physician Practices Pre p a r ed for: CALIFORNIA HEALTHCARE FOUNDATION

E l e c t ronic Medical Record s :Lessons from Small Physician Practices

Prepared by University of California, San FranciscoOctober 2003

The iHealth Re p o rts series includes:

• Im p roving Drug Prescribing Practices in the Outpatient Setting: A Ma rket An a l y s i s

• Achieving Tangible IT Benefits in Small Physician Pra c t i c e s

• Crossing the Chasm with In f o rmation Te c h n o l o gy :Bridging the Quality Gap in Health Ca re

• The Nursing Sh o rtage: Can Te c h n o l o gy He l p ?

• Diffusion of In n ovation in Health Ca re

• Genetics and Pr i vacy: A Pa t c h w o rk of Pro t e c t i o n s

• Implementing the Fe d e ral Health Pr i vacy Rule in Ca l i f o rn i a

• E - Pre s c r i b i n g

• E - En c o u n t e r s

• E-Disease Ma n a g e m e n t

• Wi reless and Mobile Computing

Electronic MedicalRecords: Lessons fromSmall Physician Practices

Pre p a red for:CALIFORNIA HEALTHCARE FOUNDATION

Pre p a red by :Un i versity of California, San Fr a n c i s c o

Au t h o r s :Ro b e rt H. Mi l l e r, Ph.D., Institute for Health and Aging;

Ida Sim, M.D., Ph.D., De p a rtment of Medicine;

and Jeff Newman, M.D., M.P.H., Di re c t o r, Sutter Health Institute for Re s e a rch and Ed u c a t i o n

October 2003

A c k n o w l e d g m e n t sThe authors sincerely thank the many interv i ewees that spenttime providing a wealth of information and insights on theirEMR use. This project was funded by the California He a l t h C a reFoundation; the Ro b e rt Wood Johnson Foundation funded aprior project that generated initial interv i ews for some of therespondents outside of California.

About the FoundationThe California He a l t h C a re Foundation, based in Oakland, is anindependent philanthropy committed to improving California’shealth care delive ry and financing systems. Formed in 19 9 6, ourgoal is to ensure that all Californians have access to afford a b l e ,quality health care.

The information included in this re p o rt is provided as generalinformation and is not intended as medical or legal advice. The California He a l t h C a re Foundation does not, and does notintend to, whether through this publication or otherw i s e ,endorse any product, entity, or person, and re s e rves the right to alter the content of this re p o rt at any time.

Additional copies of this re p o rt and other publications in thei Health Re p o rt series can be obtained by visiting us online atw w w. c h c f . o r g .

ISBN 1-9 3 2 0 6 4- 3 7-0

Copyright © 2003 California He a l t h C a re Fo u n d a t i o n

http://www.chcf.org

C o n t e n t s5 Executive Summary

8 I. I n t r o d u c t i o n

9 II. M e t h o d o l o g y

11 III. EMR Capabilities and Their Use

13 I V. Lessons Learned from Implementing the EMRLesson One: Initial EMR Financial Costs Are Substantial, While Benefits Va r y

Lesson Two: The Five types of Physician EMR Users Differed in Benefits that They Reaped; Successful Users Documented Electronically and Made Many Complementary Changes

Lesson Three: Technology Differences Explain Only Someof the Variation in Benefits

2 3 V. Recommendations for Small GroupsIdentify an EMR Champion—Or Don’t Implement

Obtain Physician Commitments to Use the EMR

Maximize Electronic Data Exchange

Arrange Comprehensive Support

Motivate Physicians to Use the EMR

2 5 VI. Suggestions for Purchasers, Public Policymakers, and Funding Agencies

2 7 E n d n o t e s

El e c t ronic Medical Re c o rds: Lessons from Small Physician Pra c t i c e s | 5

THIS PROJECT WAS DESIGNED TO PROVIDEsolo/small group physicians with practical information on elec-t ronic medical re c o rd (EMR) implementation and use. Ab o u t70 percent of active, practicing physicians in California work insolo/small groups of ten physicians or fewe r, yet little has beenpublished on their experience using EMRs. Un d e r s t a n d i n gEMR use in solo/small groups also can help policymakers ing overnment, employer coalitions, and public and private fund-ing agencies to better craft policies to hasten EMR adoption.

M e t h o d o l o g yA diverse group of EMR physician champions in 20 solo/smallpractices we re interv i ewed between May and December of2 0 0 2. Half the sample used seven different EMR pro d u c t s ,while the other half used the same EMR. Almost all practiceshad ten primary care physicians or fewe r, and most (13) we re in California.

The interv i ew protocol contained numerous open-ended questions about EMR use. Categories of EMR physician userswe re created based on characteristics of users, including extentof EMR use, benefits, and time spent making changes to complement the EMR, as well as current time spent at work.

The study had several important limitations. Many cost and benefits estimates are imprecise for a variety of re a s o n s .Mo re ove r, interv i ewees we re not re p re s e n t a t i ve of EMR users,or the entire physician population, since they we re earlyadopters of the EMR. Ne ve rtheless several patterns of findingsemerged that we re striking enough to be of use to EMR c o n s i d e rers and policymakers.

Background: EMR Capabilities and Their UseThe electronic v i e w i n g capability was a core feature, but whatphysicians could actually view varied among practices. Mo s tphysicians we re electronically d o c u m e n t i n g in the exam ro o mduring the patient visit, and many we re using electronic forms(templates) that we re specific to the type of visit or to thep a t i e n t’s disease or condition. Clinician electronic pre s c r i b i n gor o rd e r i n g was pre valent and popular. El e c t ronic m e s s a g i n gusually was limited to interoffice messages. Clinicians used c a re m a n a g e m e n t / f o l l ow - u p capabilities, including health

Executive Summary

maintenance reminders or disease-specific tem-plates with embedded clinical practice guidelines.

Other EMR capabilities less often used includedanalysis and re p o rt i n g functions, patient-dire c t e dcapabilities (such as ability for patients to secure l ycommunicate with their providers or view theirdata), and billing and scheduling capabilities thatwe re an integral part of EMR use.

Lessons Learned FromImplementing EMRs

Initial EMR Financial Costs AreSubstantial, While Benefits VaryInitial costs ranged from $15,000 to $50,000 per physician, with a median cost of $30,000 perphysician, plus re venue losses due to fewer thannormal patient visits for weeks or months afterthe EMR implementation. Ongoing financialcosts we re much lowe r.

Financial benefits varied gre a t l y, ranging from nobenefits to gains of more than $20,000 per ye a rper physician. Quality benefits we re common but also varied gre a t l y. Almost all users re p o rt e di n c reased quality of patient care due to betterdata legibility, accessibility, and organization, aswell as prescription ordering, and pre vention anddisease management care decision support.

De c reased staff costs we re common and va r i e dg re a t l y. Mo re successful users decreased transcrip-tionist, medical re c o rds, data entry, billing, andreceptionist costs. In c reased re venue was lesscommon; some re p o rted increased re venue fro mhigher coding levels, more complete capture ofs e rvices provided, and more services per visit.

Physician EMR Users Differ in Benefits Reaped Clinicians differed greatly in how they used the EMR, the amount of effort invested inmaking changes, and the benefits generated.C o m p l e m e n t a ry changes we re essential for gener-ating EMR benefits and reducing extra time costsdue to using the EMR. Changes included sys-

tematically entering patient data from paperc h a rts, customizing electronic forms (templates)that came with the EMR software, creating documenting shortcuts, arranging extra supportfor technical problems, reorganizing work f l ow inthe exam room, and reorganizing processes (e.g.,who does what and how) in the office. Mo s ti n t e rv i ewees emphasized that it took extra time—often several months or years—to effect thesechanges and learn to use the software .

Five types of EMR users were identified. Vi e we r s minimally interacted with the computer,obtained few benefits, and had invested littletime in making complementary changes toi n c rease benefits; at the time of the interv i ew sthey spent little extra time at work. Vi ewers didnot electronically enter any data, but rather dic-tated or hand-wrote notes and prescriptions.

Basic Us e r s e n t e red a limited amount of datainto the EMR, obtained few benefits, had inve s t-ed limited time in customizing electronic forms,entering past data, and making other changes tocomplement the EMR, and at the time of thei n t e rv i ews spent the same or more time at work .In t e rv i ewees we re concerned that some basic usercolleagues would become “s t u c k” at a low level ofEMR use and benefits.

St r i ve r s still invested substantial additional timein customizing software, entering past patientdata, reorganizing work f l ows, and generallylearning to more efficiently use the EMR soft-w a re; they still reaped only moderate financialbenefits, and spent more time at work. T h e ye l e c t ronically documented, using templates withdocumentation shortcuts, and thus generatedsome savings from reduced transcription andmedical re c o rds staff.

Ar r i ve r s had been “s t r i ve r s” for some period oftime, but had already invested substantial addi-tional time in activities that complemented theEMR implementation. As a result, the arrive r swe re reaping sizable benefits and spent the sameor less time at work than before the EMR.No t a b l y, most expected even more benefits, both

6 | CA L I F O R N I A HE A LT HCA R E FO U NDAT I O N

financial and quality-related. All arrivers docu-mented electronically extensive l y, and most had re o r g a n i zed their exam room and officew o rk f l ows.

System Changers we re similar to arrivers, butwe re characterized by even more benefits andtime savings per patient, use of numerous cus-t o m i zed electronic forms, and more extensivechanges in work f l ow—especially delegatingn u m e rous tasks to other clinical staff. Sy s t e mchangers focused on quality improvement anda d vocated stronger internal and health plani n c e n t i ves to encourage EMR use.

Technology Differences Explain Only Some Differences in BenefitsC l e a r l y, differences among EMR software tech-nologies had some effect on benefits achieve d .Ne ve rtheless, the data suggest that the wide va r i-ety of EMRs examined have sufficient capability,u s a b i l i t y, and flexibility to enable the earlyadopters to succeed, at least to some extent, give nthat the most successful users—arrivers and sys-tem changers—used f i ve differe n t EMRs. At thesame time, ten interv i ewees in four of the fived i f f e rent user types used the s a m e EMR. That is,despite almost identical software, users of thesame EMR had a wide range of benefits and timecosts. Clearly, much more than EMR softwaredetermines EMR-related benefits and costs.

Recommendations for Small GroupsIdentify an EMR champion—or don’t i m p l e m e n t . One or more physician EMR cham-pions must be willing to lead in purchasing andimplementing the EMR. Potential EMR champi-ons need to assess whether or not they have thepersonal characteristics, including determination,needed to succeed with an EMR.

Obtain physician commitments to use theE M R . Physicians in the practice must make specific time commitments to document elec-t ronically and learn to use the EMR effectively to generate benefits.

Ma x i m i ze electronic data exc h a n g e . This iscritical for reducing paper and data entry andthus costs. Practices need specific commitmentsf rom the labs and vendors that they will set upefficient electronic data exchange and adequatedata exchange between the practice’s EMR andthe billing and scheduling software.

Ar range compre h e n s i ve support . Su p p o rtshould address all technical and process issues.Although some vendors provide good support, ittends to be less compre h e n s i ve than needed. Itmay be ve ry difficult to arrange truly compre h e n-s i ve support in most areas.

In c e n t i v i ze physicians to use the EMR.Practices should rew a rd those physicians whogenerate benefits from reduced medical re c o rd s ,transcriptionist, and data entry staff time.

Suggestions for Purchasers, Public Policymakers, and Funding AgenciesPu rchasers, government policymakers, and priva t eand public funding agencies have an import a n trole to play by promoting quality perf o r m a n c ere p o rting and financial incentives that dispro p o r-tionately benefit EMR users, and by encouragingthe development of community-wide electro n i cclinical data exchange and new support entities.Funding demonstrations and evaluations of thesen ew initiatives can help determine what helpsEMR users achieve success more quickly.

Also needed is increased qualitative and quantita-t i ve re s e a rch on users of EMRs, includingre s e a rch into what can move a viewe r, basic user,or striver to an arriver status more quickly. Mo reb a s i c a l l y, it is important to know whether it isrealistic to expect that a majority of physicianscould become arrivers in the foreseeable future ,g i ven their characteristics, or whether it will takea long time for even the best policies to improvequality of care by transforming most physiciansinto EMR users.



THIS PROJECT PROVIDES SOLO/SMALL GROUPphysicians with practical information on EMR implementationand use. This group was the focus for two reasons. First, EMRsa re clearly important to at least a substantial minority of physi-cians. Although estimates of actual EMR use rates are re l a t i ve l yl ow—likely substantially less than 13 perc e n t1—estimates ofphysician i n t e re s t in EMRs are substantially higher, rangingf rom 31 percent to more than 65 percent of all physiciansn a t i o n w i d e .2 Second, about 70 percent of active, practicingphysicians in California work in solo/small groups of ten physi-cians or fewe r,3 yet little has been published on their experienceusing EMRs.

This re p o rt should also be of interest to employers and employ-er coalitions that provide health care benefits, policymakers ing overnment that craft legislation, managers at the Centers forMe d i c a re and Medicaid Se rvices, and re s e a rch and demonstra-tion funding agencies. These stakeholders increasingly re c o g n i zethat IT generally—and the EMR specifically—is critical toi m p roving quality of care, including in solo/small gro u p s .4, 5

Understanding EMR use in solo/small groups can help policy-makers craft policies that can hasten EMR adoption.

I. Introduction


Practices and Selection Me t h o d s . A diverse group of EMRphysician champions in 20 solo/small practices we re inter-v i ewed. By design, half the sample used a wide variety of EMRp roducts, while the other half used a single EMR. The designenabled re s e a rchers to examine the diversity of use and benefitspossible from a variety of products, as well as one single pro d-uct. All practices we re small: eight practices had solo physi-cians, nine others had two to ten physicians; two practices with17 and 25 physicians we re included because the practice leftEMR adoption up to semi-autonomous small sites within thepractice. One small practice joined a large group (of 140 physi-cians) during the interv i ew period.

Of the 20 practices selected, a majority we re in California andmost had primary care physicians. Of the thirteen Californiapractices, nine we re primary care (family physicians or internalmedicine); the others we re card i o l o g y, uro l o g y, endocrinology,and pulmonology practices. Of the seven non-California prac-tice interv i ewees, all we re family physicians. The AmericanCollege of Physicians and the American Academy of Fa m i l yPhysicians helped identify volunteers, as did two vendors that provided a list of EMR re f e rence sites (six of the cases).Data were collected in May through December 2002 for thenon-California practices, interview data also were obtained in2000 and 2001 from a previously funded study.

EMRs Used and Years of User Ex p e r i e n c e . Ten practices used Practice Pa rt n e r, while the other half used seven otherEMRs: QD Clinical (four practices), Alteer, Ne x t Gen, A4Healthmatics, Au t o C h a rt, So a pw a re, and Cliniflow. Mo s tphysicians we re satisfied with their EMR and EMR ve n d o r.Users had between six months and ten ye a r’s experience withEMRs; the median EMR use was five ye a r s .

In t e rv i ewees we re not re p re s e n t a t i ve of all physicians or eve nEMR users. Using Ro g e r s’ terminology for innova t i o na d o p t e r s ,6, 7 i n t e rv i ewees fell into the early adopters or innova-tors categories. This sample ove r re p resents the high-achievingend of the spectrum of innovation adopters, in large partbecause these we re physician EMR champions who had beenusing the EMR for several years, and had vo l u n t e e red for thestudy—less successful physicians who we re not EMR champi-ons we re less likely to vo l u n t e e r. Ne ve rtheless, more adva n c e dusers described their own experiences when they we re less

II. Methodology

a d vanced in their use, and also described theexperiences of some of their colleagues who we reless advanced in EMR use.

In t e rview Pro c e s s .The interv i ew protocol con-tained numerous open-ended questions aboutEMR use. Be t ween one and four telephone inter-v i ews we re conducted for each practice. T h ere s e a rchers observed EMR use and conductedadditional in-person interv i ews in three practices.The investigators conducted all interv i ew s .

Data An a l y s i s . The re s e a rchers used a qualitativere s e a rch software program (QSR Nvivo) to codetranscribed interv i ew data to several dozen con-cept categories (such as “implementation activi-t i e s” and “w o rk f l ow changes”), manage the data,and help analyze responses across interv i ew s ,including relationships among variables. For theanalysis, the re s e a rchers used explanation build-ing and pattern matching qualitative re s e a rc ht e c h n i q u e s .8 Categories of EMR physician userswe re created based on characteristics of users,including extent of EMR use, benefits, and timespent making changes to complement the EMR,as well as whether or not they we re now spend-ing more or less time at work.

Li m i t a t i o n s . This study has several import a n tlimitations. Many cost and benefits estimates arei m p recise for various reasons—some EMRimplementations had taken place years ago, someclinicians did not have estimates or did not havethem readily available, and practice characteristics(e.g., practice size and patient volumes) changedover time, making pre-EMR/post-EMR compar-isons difficult. Data we re obtained from thephysician EMR champion or lead physician, whoalso provided an ove rv i ew of use by others in thepractice. Data we re not obtained from non-champions in practices that we re non-solo, otherthan in one site where the re s e a rchers conductedan observational study. As a result, interv i ewe e swe re not re p re s e n t a t i ve of EMR users, let alone

of the entire physician population. Mo re ove r,practices could not be systematically followe dover time.

Ne ve rtheless several patterns of findings emergedthat we re striking enough to be of use to thoseconsidering EMR implementation as well to policymakers developing programs to hasten itsadoption. At the same time, further re s e a rch isneeded to provide more indepth, precise, andc o m p re h e n s i ve information to these audiences.



EMRS HAVE SEVEN MAIN TYPES OF CLINICALcapabilities that correspond to seven sets of clinical activitieswithin ambulatory physician practices. These capabilities are :v i ewing, documenting, ordering (e.g., prescriptions and labs),messaging, care management, analysis and re p o rting, andp a t i e n t - d i rected (e.g., patients ordering prescription re f i l l sonline). All eight EMR products had the first five capabilities;each EMR had somewhat different features for each capability.Some EMRs also included integrated billing and schedulingcapabilities. This re p o rt focuses on physician users, althoughclinical support staff also used EMR capabilities.

The most often used capabilities of the EMR are the following:

Vi e w i n g . The electronic viewing capability was a core feature ,but what physicians could actually view varied among prac-tices. Vi rtually all interv i ewees viewed past pro g ress notes,p roblem lists (chief complaints), past medications, and aller-gies. Some could view lab results, consultant re p o rts, hospitalinpatient data and other related clinical data, but only if theyhad arranged data exchange interfaces between their EMR andoutside information systems, or if their staff entered data onpaper (e.g., lab results) into the EMR.

Do c u m e n t i n g .The electronic documenting capability enabledusers to re c o rd pro g ress notes, chief complaints and diagnoses,allergies, prescriptions, and other data electro n i c a l l y. Mo s tphysicians entered data in the exam room during the patientvisit. Most used electronic forms (templates) that we re specificto the type of visit (e.g., routine return visit) or to the patient’sdisease or condition (e.g., diabetes or low back pain) or a combination of both. Fe a t u res of the electronic forms va r i e damong EMRs. Even users of the same EMR product va r i e dg reatly in their use of electronic forms and their features. T h ee l e c t ronic forms acted to prompt physicians to guide the clini-c i a n s’ exam and discussion with the patient. Physicians usedsome combination of typing in free text, clicking on stru c t u re db oxes (or “picker lists”) that we re embedded in the electro n i cforms, and clicking on “m a c ro” placeholders in the electro n i cforms in order to generate standard text phrases or sentencesfor that pro m p t .

Ord e r i n g . The electronic ordering capability enabled users to enter prescriptions into electronic forms, where they could

III. EMR Capabilities and Their Use

12 | CA L I F O R N I A HE A LT HCA R E FO U NDAT I O N12 | CA L I F O R N I A HE A LT HCA R E FO U NDAT I O N

select from various ordering possibilities, and tore c e i ve decision support (alerts) on drug/drugand drug/allergy interactions. Clinicians usuallyhanded the printed prescription to the patientand/or faxed prescriptions directly to the phar-macy. Pharmacies used fax or phone to commu-nicate. Only a handful of practices had drugformulary information, or used lab, referral, orother ordering capabilities.

Clinician electronic prescribing was both pre va-lent and popular. Although initial pre s c r i p t i o n stook time to enter, refilling prescriptions save dtime for physicians and staff:

“ One of the most daunting tasks is [whena] Me d i c a re patient comes in and needsrefills of ten medications. My writing isa t rocious, and the more I write the worseit gets. The prescription writing [feature ]makes it so neat and legible. No questionabout the pharmacist misinterpreting a n y t h i n g .” —Family physician, solo practice

Me s s a g i n g . El e c t ronic messaging usually waslimited to interoffice messages—for example, ap a t i e n t’s telephone request for a prescription re f i l lmight go from the receptionist to the physicianto the receptionist. No interv i ewees had In t e r n e temail capabilities integrated into the EMR.

Ca re Ma n a g e m e n t / Fo l l ow - u p .EMR preventionand disease management capabilities overlappedwith the documentation capability. For healthmaintenance, most interviewees had to pullhealth maintenance data—that is, had toremember to seek information on health main-tenance status for the patient, although othershad customized their templates to automaticallyimport reminders. Clinicians who used disease-or condition-specific templates with embeddedclinical practice guidelines felt that careimproved as a result.

“When I look through the template as I’mtalking with the [diabetic] patient…thetemplate asks whether the patient has vis-ited an eye doctor in the last ye a r. It asksif the patient is examining his feet daily.Or it asks if the patient is on an AC Ei n h i b i t o r. You can say ‘n o’ or ‘advised todo so.’ Under the diabetic plan, you putd own what you advise them to do, whenyou want them to come back, what thelabs are. I can…bring up the diabetict reatment medication module and printthat out. T h a t’s a lot better [than beforethe EMR].”

— Family physician, small group practice.

Other EMR capabilities less often used includedthe follow i n g :

Analysis and Re p o rt i n g . Most interv i ewe e scould do simple searches of patients—for exam-ple, to identify female patients on hormonereplacement therapy. Howe ve r, few used mostEMR analysis and re p o rting functions, in partbecause coded data we re limited.

Pa t i e n t - d i re c t e d . The EMRs usually had nocapabilities that patients could use, such as abilityto access a practice Web site in order to secure l ycommunicate with a provider or view their data;capabilities that we re available we re rarely used.Most interv i ewees discouraged patients fro msending email and replied to email messages by phone.

Billing and Scheduling. Some practices usedEMRs with integrated billing and schedulingmodules; other practices created interfaces toe xchange data between the EMR and practicemanagement system software. Many practices had no integration at all between the EMR andbilling/scheduling software. Greater integrationreduces duplicate data entry and permits automat-ed service capture and higher visit level coding.


Lesson One: Initial EMR Financial Costs Are Substantial, While Benefits Va r yInitial costs are high.Those considering implementing anEMR tend to focus on highly visible financial purchase costswith good reason, since they are substantial. Initial costsranged from $15,000 to $50,000 per physician. For the ninecases that provided rough estimates and had implemented theEMR fewer than four years before, the median cost was$ 3 0,000 per physician. The amounts depended on various fac-tors. For example, better pre-existing hard w a re or leased EMRs o f t w a re / h a rd w a re decreased the initial outlay, while datae xchange interfaces, use of more nurse practitioners and physician assistants per physician, and use of more notebookcomputers increased the initial outlay. T h e re was considerablevariation in software costs among vendors. Some cost datawe re sketchy, as this project depended on interv i ewe e s’ re p o rted costs. Mo re ove r, hard w a re costs have been falling over time.

Most interv i ewees re p o rted that they substantially re d u c e dpatient visits (and thus lost revenues) for a couple of daysduring the installation period, and scheduled fewer than normal patient visits for a period of weeks or months afterthe EMR implementation because getting used to the EMRtook extra time. Unfortunately, interviewees could not quantify such revenue losses.

Ongoing financial costs we re substantially lower than initialcosts. They included 15 percent to 20 percent of initial soft-w a re costs for vendor support and upgrades, and re p l a c e m e n t sf o r, and enhancements to, hard w a re .

Financial benefits varied gre a t l y. Many interv i ewees re p o rt-ed similar types of benefits from implementing the EMR,although the amounts varied gre a t l y, ranging from no benefitsto gains of more than $20,000 per year for a couple of physi-cians in one practice.

Quality benefits we re common but va r i e d . Almost all usersre p o rted increased patient care quality due to such improve-ments as better data legibility, accessibility, and organization,p rescription ordering, and pre vention and disease managementc a re decision support.

I V. Lessons Learned From Implementing the EMR


These re p o rts are consistent with a growing bodyof evidence on the benefits of IT and EMRs onquality of patient care. In addition to stating thebenefits of decision support and reminders, i n t e rv i ewees also stressed more basic changes:

“ … t h e re is a tremendous amount of pooror lost information that really impactspatient care….doctors don’t remember allthe illnesses or all the medicines and can’tread anybody else’s notes in the chart. Ilooked through all the charts at [an aca-demic health center]—most are illegibleand that goes for eve ry institutiona round the country. So you can’t read thei n f o rmation, can’t find it, and then ifyou become suddenly sick, you can’t dealwith it. You can fix all that with anEMR. Also if something bad comes up[e.g., a recalled medicine], you can findpatients that might be affected….there isa better way to practice medicine.”

— Pulmonologist, small group practice

Se veral interv i ewees stressed the benefits of pro-viding patients and consultants with better print-ed information, thus improving the coord i n a t i o nof care :

“The biggest benefit is to patient care .Patient care charts are legible and dru gi n t e ractions can be seen. One of thebiggest problems is that patients are onmultiple medications and go to multiplespecialists and pharmacies, so nobodyk n ows who’s taking what. Now, eve rytime they come in, they get a print-out ofall their medicines and they’re told ‘t a k e

this to all your different specialists and ifthey change your medicine or dose, markit down and the next time yo u’re here ,bring it in’ and we print out a new one.So all the specialists know exactly whatthe patient is taking.”

— Family physician, solo practice

De c reased staff costs we re common.Moresuccessful users decreased transcriptionist, medical records, data entry, billing, and evenreceptionist staff FTEs. However, as indicatedbelow, some types of EMR users reaped none ofthese benefits. Several physicians who dictatedpre-EMR were especially enthusiastic about thecost savings:

“We re a l i zed that when we looked ove rthe last two years, the cost to pay for afour- or five - year lease for hard w a re ands o f t w a re was equal to the same amountthat we would have been paying fort ranscription. So by simply waving good-bye to the tra n s c r i p t i o n i s t … we tra n s-f e r red the money we would have beenpaying her and used that to finance our[EMR] computer.”

— Family physician, small group practice

In c reased re venue was less common.Ma n yi n t e rv i ewees felt more comfortable coding tohigher Me d i c a re Evaluation and Ma n a g e m e n t(E&M) levels (visits and consultations prov i d e dby physicians or residents under their superv i-sion) than they had prior to using an EMR, sincethe electronic forms had prompted them to dowhat was needed to justify the higher levels andto document that fact. Asked about whether theEMR increased re venue by increasing the level ofcoding, one respondent re p l i e d :


“I think it does because it makes yo uthink about it. I'm not sure if intern i s t sa re as bad as family practice docs, but wetend to just sort of code the same thingfor eve rything. We tend to under-codebecause it’s easier to under-code thanw o r ry about [Me d i c a re fraud and abusec h a r g e s ] .”


Others re p o rted increased re venue from morecomplete capture of services provided, especiallywith EMR/billing software integration. A fewphysicians provided more services per visit, whichi n c reased re venues. In c reased re venue was lesscommon than decreased staff costs and varied asmuch, from none to ve ry substantial. Re ve n u etended to accrue most to physicians who saw fee-f o r - s e rvice patients, and that had data exc h a n g eintegration between the clinical EMR and thebilling system (or module) capabilities.

Lesson Two: The Five Types ofPhysician EMR Users Differed inBenefits that They Reaped;Successful Users DocumentedElectronically and Made ManyComplementary Changes In order to illustrate the differences among as p e c t rum of users, five different types (categories)of physician users we re created: viewers, basicusers, strivers, arrivers, and system changers.Vi ewers and basic users, the least advanced users,had mostly unchanged clinical and businessp rocesses that still relied heavily on paper orscanned images. In contrast, arrivers and systemchangers had re e n g i n e e red work pro c e s s e senough to virtually eliminate paper-based clinicalp rocesses. Thus, more advanced users used moreEMR capabilities, usually reaped more financialand quality benefits, and had invested more time

in making changes that complemented the EMRthan the less advanced users.

St r i vers we re in the middle of the spectrum. T h e yspent the most extra time at work, since theywe re still investing time in making changes thatcould generate benefits and eventually re d u c etheir time costs. At the point of the interv i ew s ,extra time spent at work was minimal at bothends of the spectrum, as viewers spent little extratime using the system or making changes (andreaped fewer benefits), and arrivers and systemchangers had become more efficient during theirw o rkday as a result of investing extra time inmaking complementary changes.

C o m p l e m e n t a ry changes we re essential for gener-ating EMR benefits and eventually re d u c i n gextra time costs due to implementing the EMR.They included such activities as systematicallyentering patient data from paper charts, cus-tomizing electronic forms (templates) that camewith the EMR software, creating documentings h o rtcuts, arranging extra support for technicalp roblems, reorganizing their work f l ow in theexam room, and rearranging processes in theoffice as a whole.

Many physicians re p o rted that they had to inve s textra time to make the type of complementarychanges described above. For example, physicianchampions spent extra time selecting the EMR,p reparing for EMR installation, and then imple-menting the EMR—they participated in trainingsessions, learned how to use the software, ove r s a winstallation, engaged in trouble-shooting, andw o rked with the EMR installation staff to helpothers learn the basics of EMR use. Su b s e q u e n t l y,many interv i ewees re p o rted that they continuedto invest extra time—several months or even sev-eral years—in such activities as entering pastpatient data, customizing templates, and tro u b l e -shooting technical problems.


C l i n i c i a n s differed greatly in how they used the EMR, the benefits that they generated, theamount of effort invested in making changes,and the changes that they made. Several differ-ent types of EMR users are outlined below. As

Figure 1 indicates, it is important to keep inmind that the EMR user types are from a re l a-t i vely small part of the spectrum of physiciansthat Rogers called (EMR) Early Adopters andIn n ova t o r s .6, 7

Figure 1. Types of Innovation Adopters and EMR Users

Laggards Late Majority Early Majority Early Adopters

Type ofInnovationAdopters(Rodgers)

Viewers Basic Users Strivers ArriversType ofEMR Users

SystemChangers

Non-adopters

Innovators

EMR Adopters


Vi e w e r s

Vi ewers minimally interacted with the computer,obtained few benefits, and had invested littletime in making complementary changes toi n c rease benefits; at the time of the interv i ew sthey spent little extra time at work. Both viewe ri n t e rv i ewees used the EMR to view data and todo little else—they did not electronically enterany data, but rather dictated or hand-wro t ep ro g ress notes and prescriptions.

For example, in one practice, staff imported thetranscribed pro g ress notes into the EMR andstaff (not physicians) electronically entered thehand-written prescriptions. The practice had notre o r g a n i zed work f l ow, maintained problem lists,or made other efforts to use the EMR efficiently,and thus it maintained parallel paper and elec-t ronic processes. As a result, they saw no staffcost savings or re venue increases, and littlechange in physician time spent at work .In t e rv i ewees attributed some of the minimalEMR use to the poor performance of the EMRp roduct—the only interv i ew with negativere v i ews of an EMR.

Basic Users

Basic users entered a limited amount of data intothe EMR, obtained few benefits, had inve s t e dlimited time in customizing electronic forms,entering past data, and making other changes tocomplement the EMR, and at the time of thei n t e rv i ews spent the same or more time at work .While there was only one basic user interv i ewe d ,m o re advanced EMR users described some col-leagues as having these characteristics. The basicuser viewed data, maintained some electronic lists(e.g., chief complaints, past medications, andallergies), and ord e red prescriptions. The basicuser dictated visit notes while viewing visit- ordisease-specific templates, and had trained thetranscriptionist to fill out the template as shetranscribed. Although the practice aimed to elim-inate the paper chart (and medical re c o rds staff) ,transcription costs remained high, and the prac-tice had added costs of scanning tests and paperconsultant re p o rts.

In t e rv i ewees we re concerned that some basic usercolleagues—who had invested minimal time inmaking the changes needed to have the EMRp roduce more benefits—would remain at a lowl e vel of EMR use and benefits, and continue tospend the same or more time at work, whichwould eventually undermine their willingness touse the EMR.


Most interv i ewees fell into the striver or arrive rcategories of users who had already embraced theconcept of the EMR as an integral interactivetool in their work .

S t r i v e r s

St r i vers still invested substantial additional timein creating changes that complemented the EMRwith the hope of generating financial benefitsand reducing their time costs. The five strive r scontinued to customize software, enter pastpatient data, re o r g a n i ze work f l ows, and generallylearn to use the EMR software more efficiently.Although they made more use of EMR capabili-ties than did viewers and basic users, they stillwe re reaping only moderate financial benefits.

St r i vers electronically documented using tem-plates with documentation shortcuts, and thusgenerated some savings from reducing transcrip-tion and medical records staff. For example, oneinterviewee estimated that he eliminated $600in transcription costs per month. However, thephysician still had to spend an extra hour a dayat work, more than three years after implement-ing the EMR. In part, he believed that he wasleaving work later due to time needed to pro-vide higher quality of care, including docu-menting visits and communicating results morethoroughly, and sending better information to

specialists. The same interv i ewee also had identi-fied a series of efficiency improvements (includ-ing improved templates) that he thought wouldhelp in reducing time at work.

One striver—a self-described “sophisticated user”who recently switched to a new EMR—discussedthe time demands of changing work f l ow, learningto use the software, and customizing templates:

“[The EMR] disrupts your work f l ow a loti n i t i a l l y. T h e re is a big learning curve toactually use it in the room with apatient, or to document using some com-bination of handwriting, notes, and[data entry or dictation]. It takes a cou-ple of months to really become a compe-tent user. Classes or practice or tutorialsa re better than doing nothing, but onceyo u’ve done them, you feel…at the barebones level of being able to [use theE M R ] .”


The same interv i ewee also discussed the timere q u i red in addressing technical pro b l e m s ,including support for hard w a re breaks:

“I am the support system….That’s re a l l yone of the biggest problems I see rightn ow, for little offices at least—there is noi n f o rmation systems department. Sowhen things go down, when a printer orcomputer breaks, when things fre e ze, yo uh a ve to stop and go deal with it becauseothers really can’t .”



A r r i v e r s

A r r i vers had been strivers for some period oftime; unlike strivers, they had already inve s t e dsubstantial additional time in activities that com-plemented the EMR implementation—they hade n t e red past patient data, customized templates,c reated interfaces, developed stable technical sup-p o rt stru c t u res, and generally ascended the EMRlearning curve. As a result, the ten arriver inter-v i ewees we re reaping sizable benefits and spentthe same or less time at work than before theEMR. No t a b l y, most expected even more bene-fits, both financial and quality-re l a t e d .

One arriver discussed the time spent (during thes t r i ver phase) to pre p a re for the EMR and to cus-t o m i ze templates—for example, so that vitalsigns, lab, or other data could be imported auto-matically into the pro g ress note or key phrasescould be generated quickly, sometimes at theclick of a mouse button.

“That champion that I was talkingabout? Not only do you need to spende x t ra time with the other prov i d e r sb u t . . . . b e f o re you start the program there’s a lot of setup that has to be done,including setting up your health caremaintenance reminders and labtables....massaging existing templates foryour practice and creating new templates,and new quick text [documentations h o rtcuts]. I remember being here late inthe evening, almost eve ry day work i n ganother hour or two on the computer...[for the first ye a r ] .”


Entering past patient data into the EMR wastime-consuming.

“As we went along, we tried to re t i repaper charts. Initially we dictated pro b-lem lists and social histories, and used at ranscriber trained by the vendor toinput it into the system. She quit whenher volume was going way down becausewe we re n’t dictating our regular notes.Now we’re manually entering three chart sa day as patients come in. We tra n s f e r[past information] into the problem listsand social, family, past medical history,into the health maintenance sections ofthe EMR. We scan key re p o rts that arerecent. All this takes extra time.”



Most arrivers had re o r g a n i zed their exam ro o mand office work f l ows—especially import a n tbecause, as one interv i ewee put it, “the EMRchanges how you do things.” Changing work f l owmeant changing tasks performed by each staffm e m b e r. For example, since receptionists andnurses spent less time tracking down charts, theycould spend more time talking with patients andentering patient data into electronic forms priorto the patient entering the exam room, whichreduced the time physicians needed to documentthe visit. Using new EMR-related capabilities tokeep track of patients’ care, staff also could iden-tify and then contact patients who had not ye ttaken ord e red tests, we re ove rdue for pre ve n t i vetests, or needed some chronic care follow - u p. Ineach case, the physicians first created protocols orrules for identifying patients that needed follow -up care, and pro c e d u res for when and how staffwould follow-up with patients.

Sometimes the extra time arrivers spent at work(during the striver phase) was coupled withreduced re venues for a limited period of time:

“The first three months I cut back on thenumber of patients I saw by around 25p e rcent. And then when I found I wasn’tstaying really late at night, then I’d add a few more patients. So I’d cut dow npatients by maybe 10 to 15 percent. Ata round six months, I was no longer stay-ing late at night but I knew it was tak-ing me more time. I had to work thro u g hlunches or stay a little late. But by thetime I got to t h ree ye a r s, the time I spentat work was the same as before the EMR.When you get past three years, then s u d d e n l y, yo u’re going home earl y.”

— Family physician, small practice

System Changers

System changers we re similar to arrivers, but we rec h a r a c t e r i zed by even more benefits and time sav-ings per patient, use of numerous customize de l e c t ronic forms (templates), and changes inw o rk f l ow—such as those discussed above — e s p e-cially delegating numerous tasks to other clinicals t a f f. They also attempted to change the externale n v i ronment by encouraging health plans torew a rd practices for producing higher quality ofc a re due to the EMR.

In contrast to arrivers, system changers went fur-ther in the type of changes they made, in partbecause both we re strong leaders within theirpractices. No t a b l y, both had focused on qualityi m p rovement:

“I can incorporate sophisticated clinicalpathways into my practice that aresophisticated. For example, if we decideto screen for depression, I put those ques-tions in the template; I can’t help buts c reen for it—takes no time at all.



Both system changers advocated stronger internali n c e n t i ves to encourage EMR use:

“ Eventually you want to incentivize. Weneed to cut down on the dictation, so oneof the redesigns we’re currently work i n gon is to give docs the cost of the dictationas part of their pro d u c t i v i t y, so they havean incentive to say, ‘I want to decre a s emy dictation; how do I do that?’ And we say, ‘He re, here’s a template.’ ”


The system changers we re local leaders, attempt-ing to change the external environment bypersuading payers and independent practiceassociations (IPAs) to reward quality of care.The experience of one interviewee illustratesthat such rewards would disproportionatelybenefit physicians using EMRs.

“T h e re’s a pot of money [from withheldpayments on HMO patients]…that willbe re t u rned to us based on how well wep ractice medicine. T h i rty-nine percent ofit is how cost effective we are, but the re s tof it…is quality of care and patient satis-faction. Suddenly this EMR is pro b a b l ygoing to generate hundreds of thousandsof dollars. When the HMOs announcewhat parameters they will use, with thisp ro g ram we can make sure that wea c h i e ve close to 100 percent on eve ryselected para m e t e r. A paper re c o rd officewould never be able to even try to dothat; we will be able to do it ve ry easily.That will pay for this system two or thre etimes ove r.”


It is also important to note that b o t h s y s t e mchangers had joined larger groups. They we re inthe process of integrating themselves into thosegroups and becoming EMR champions in alarger arena of action.

Lesson Three: Te c h n o l o g yDifferences Explain Only Some of the Variation in BenefitsC l e a r l y, differences among EMR software pro d-ucts had some affect on benefits achieved. So m eEMRs we re more capable, usable, and flexiblethan others. Cert a i n l y, bad EMR software couldlead to implementation failure, as in the case ofthe practice that found that its EMR was unac-ceptable and was seeking a new one.Un f o rt u n a t e l y, it was impossible to determinewhich EMRs provided the most benefits and thebest value, given the small sample.

Ne ve rtheless, the data suggest that many differe n tEMRs have sufficient capability, usability, andflexibility to enable the early adopters to succeed,at least to some extent, given that the most suc-cessful users—arrivers and system changers—u s e d f i ve different EMRs. Mo re ove r, most inter-v i ewees we re quite satisfied with both EMRp roducts and services. One explanation for thesefindings is that in selecting software, clinicianslooked for different EMR styles that would workfor them (e.g., some EMR software packagese m p h a s i ze typing in free text, whereas otherse m p h a s i ze stru c t u red data entry). Another expla-nation is that many in this extraord i n a ry sampleof early adopters had characteristics that enabledthem to make changes that led to at least somesuccess, despite differences in the underlying usefulness of the technology.

Alternatives to Full EMRs

EMR use is one path to quality improvement.H o w e v e r, EMR use often forces clinicians tomake numerous and sometimes difficultchanges in how they work, leading to extratime costs that compound the already substan-tial financial costs of the EMR. Such potentiallyprohibitive time and financial investments dis-courage EMR adoption.

One alternative to a full-blown EMR is toadopt the component EMR capabilities incre-mentally—e.g., adopt lab viewing followed by electronic prescribing. Take the case ofpurchasers implementing new performancestandards that require decision support forelectronic prescribing, and for prevention and disease management. Such require-ments may be better served by componentcapabilities.

H o w e v e r, most performance standards likelywould require physicians to use several typesof integrated EMR capabilities, not just one ortwo. For example, in order to use decision sup-port to improve quality, physicians would needto electronically view data; order prescriptions;document chief complaints, allergies, andother data; report and analyze performancedata; and likely need care management tem-plates and inter-office messaging. Moreover,physicians would have to make accompanyingdata exchange, workflow, and other changes.Thus a requirement that physicians use deci-sion support to improve quality of care wouldrequire use of key elements of an integratedEMR, as well as complementary changes thatwould require real effort.

There is no data—for or against—to indicatewhether or not an incremental approach is aviable alternative. Especially unclear is theextent to which electronic and paper workflowcan peacefully co-exist—i.e., produce benefitsand yet require minimal cost, data exchanges,and changes in workflow.


The data underscore the fact that h ow the EMRis used is important, as ten interv i ewees in fourof the five different user types used the sameEMR. That is, despite almost identical software ,users of the same EMR had a wide range of ben-efits and time costs. Clearly, much more thanEMR software determines EMR-related benefitsand costs.


Identify an EMR Champion—or Don’t Implement For solo/small groups considering an EMR, one or morephysician EMR champions must be willing to lead in purc h a s-ing and implementing the EMR. Potential EMR championsneed to assess whether or not they have the personal character-istics, including determination, needed to succeed with anEMR. Potential EMR champions should think about the kindof differences between the EMR early adopters that we re inter-v i ewed and most other clinicians. Most early adopter interv i e-wees found it difficult to make needed changes, despite theiroverall IT savvy, work f l ow change skills, and positive attitudet ow a rds change. Other solo/small group physicians may find ite ven more difficult to achieve success, given that they are likelyto have less IT savvy and less enthusiasm tow a rds makingEMR process changes.

At present, becoming an EMR champion can be challenging.While some physicians are willing to spend some time gettingused to EMRs and making needed changes, most physiciansdo not necessarily want to figure out by themselves how tomost efficiently enter past data, fix hard w a re and software ITp roblems, customize software, reengineer their work f l ow andtheir office’s work f l ow, and orchestrate data exchange interf a c e sb e t ween themselves and outside data providers.

Obtain Physician Commitments to Use the EMRIf a practice has one or more EMR champions, other physi-cians in the practice must make specific time commitments ino rder to achieve success. Physicians must understand that theywill have to change their work f l ow in order to generate bene-fits—in part i c u l a r, not write or dictate pro g ress notes butinstead type in text or click on check boxes. Mo re ove r, theyneed to commit extra time to learn to use the EMR e f f e c t i ve l y,including customizing electronic forms and their own documentation shortcuts. Without that commitment, somephysicians will quickly become discouraged, reduce theirpotential EMR use, not generate EMR-related financial orquality benefits, and reduce the benefits generated by others.

V. Recommendations for Small Groups


Maximize Electronic Data ExchangeMaximizing electronic data exchange is criticalfor reducing paper and data entry and thus forreducing costs. First, the practice has to obtainadequate electronic data from outside sources. In part i c u l a r, this means obtaining specific commitments from the labs to set up efficiente l e c t ronic data exchange to enable physicians to view lab results within the EMR. Second, the practice needs to arrrange adequate datae xchange between the EMR and the billing andscheduling software within the practice. Fo rsome, this means purchasing an EMR with integrated scheduling and billing modules, andc o n ve rting data from the old system to the newsystem. For others, this means obtaining contrac-tual commitments from the EMR and practice management software vendors to set up efficiente l e c t ronic data exchange between the two sys-tems. Since most users in the sample had noguarantees, user experience with data exc h a n g ewas highly va r i a b l e .

Arrange Comprehensive SupportC l e a r l y, compre h e n s i ve and multifaceted supports e rvices would help many physicians learn to useEMRs more effectively and more quickly.C o m p re h e n s i ve services should address all techni-cal issues, including hard w a re, software, operat-ing systems, telecommunications and pro c e s sissues—past data entry, template customization,w o rk f l ow redesign, and learning efficient use ofthe EMR. Although some vendors provide goods u p p o rt, it tends to be less compre h e n s i ve thanneeded for the many changes that go well beyo n dd i rect use of the EMR software.

In re a l i t y, it may be ve ry difficult to arrange tru l yc o m p re h e n s i ve support, since the market doesnot offer it in most areas. At a minimum, thepractice must have solid technical service supportas a backup to whatever the hard w a re, telecom-munications, and software vendors can prov i d e .

Most EMR vendors supply the names of poten-tial technical support services firms, althoughsome vendors may not know how effective suchfirms really are .

Practices considering adopting an EMR couldalso obtain some support from other practices inthe same area that had already successfully adopt-ed the same EMR and are willing to prov i d eongoing advice on how to use the EMR efficient-l y, change work f l ow, and ove rcome obstacles.

Motivate Physicians to Use the EMRPractices should consider rewarding those physi-cians that electronically document and thusgenerate benefits from reduced medical records,transcriptionist, and data entry staff time.Incentives had a major effect on behavior in the few practices that tried them.


EM P LOY E R H E A LT H C A R E P U RC H A S E R S,g overnment agencies, and foundations could adopt policies tofacilitate more EMR adoption and more e f f e c t i ve EMR use.5, 9

Potential policies include promoting quality performancereporting and financial incentives that disproportionatelybenefit EMR users, encouraging the development of commu-nity-wide electronic clinical data exchange and new supportorganizations, and funding demonstrations and evaluations of alternative initiatives to determine what helps EMR usersachieve success more quickly.

Pr i vate and public funders have a part i c ularly important roleto play in financing and assessing demonstrations of experi-mental support service organizations. Such organizations cancoordinate the EMR support already provided by variousvendors, fill holes in technical support, and provide insightsinto how to reorganize workflows and make other changes to best use EMRs. Funders can also help create easy-to-use,community-wide clinical data exchange systems amonghealth care providers and organizations so that a clinician canelectronically view and use all clinical data on a given patient.Developments in Santa Barbara, Indianapolis, andWashington state point towards the increased feasibility ofsuch systems, and lessons from experiments could be helpfulin other areas.10, 12

Also needed is increased qualitative and quantitative re s e a rc hon users of EMRs. Foundations and the Agency for He a l t h c a reRe s e a rch and Quality (AHRQ) could fund re s e a rch into whatcan move a viewe r, basic user, or striver to an arriver statusm o re quickly. Mo re basically, it is important to know whetherit is realistic to expect that a majority of physicians couldbecome arrivers in the foreseeable future, given their character-istics, or whether it will take a long time for even the best poli-cies to improve quality of care by transforming most physiciansinto EMR users.

V I . Suggestions for Purchasers, Public Policymakers, and Funding Agencies


An “Open-source” EMR as Alternativeto Current EMRs

Another alternative to current full EMRs is totry to change EMR capabilities and price—forexample through the creation of a public EMRwith open-source code, analogous to program-mers world-wide developing Linux as an alter-native to Windows. If successful, this approachmight lower the initial EMR software cost.H o w e v e r, it would not lower other financialcosts, and does not address the time invest-ment and complementary change challengesthat are so daunting for small practices.Furthermore, any lower initial costs for lessexpensive, open-source EMR might be partiallynegated by higher time costs, and possiblyhigher maintenance costs, unless integratedcapabilities and ease-of-use were the same asor better than current products. The open-source approach remains unproven.


1. Miller R.H., J.M. Hillman and R.S. Gi ve n .“ Expected Physician Use of IT: Results Fro mthe Deloitte Re s e a rch Su rve y.” Under re v i e w2 0 0 3.

2. A A F P. AAFP This We e k Jan 29. 2003.

3. C. Dowe r, T. Mc Ree, and K. Grumbach, et al. The Practice of Medicine In Ca l i f o rnia: A Profile of the Physician Wo rk f o rc e .San Francisco, CA: California Wo rk f o rc eIn i t i a t i ve at the UCSF Center for the He a l t hProfessions; 2001.

4. Institute of Medicine. Crossing the Qu a l i t yChasm: A New Health System for the 21stCe n t u ry. Washington, D.C.: Na t i o n a lAcademy Press; 2001.

5. Institute of Medicine. Fostering Ra p i dAd vances in Health Ca re: Learning fro mSystem De m o n s t ra t i o n s . Washington, D.C.:National Academy Press; 2002.

6. Rogers E.M. Diffusion of In n ova t i o n s . 4th ed.New Yo rk, NY: The Free Press; 19 9 5.

7. Cain M. and R. Mittman. Diffusion ofIn n ovation in Health Ca re . Oakland, CA:California He a l t h C a re Foundation; 2002.

8. Yin R.K. Case Study Re s e a rc h : Design andMethods. Applied Social Re s e a rch Me t h o d sSeries. Thousand Oaks, CA: Sa g ePublications, Inc.; 19 9 4.

9. Miller R.H. and I. Sim “Using El e c t ro n i cMedical Re c o rds for Quality Im p rovement in Physician Practices: Barriers and Po t e n t i a lSolutions.” Fo rthcoming in Health Affairs.

E n d n o t e s

10 . Karp S. Santa Barbara County Care Da t aExchange. w w w. m a h e a l t h d a t a . o r g / f o rm s / e ve n t se ve n t s2 0 0 3 / H I T _ 2 0 0 3 0 2 0 7.

11. O verhage J.M. The Indianapolis Ne t w o rk for Patient Care. w w w. e h e a l t h i n i t i a t i ve . o r g /H I PA A / p re s e n t a t i o n s /

12. Walker J. An Ove rv i ew of the Wa s h i n g t o nState Project. w w w. e h e a l t h i n i t i a t i ve . o r g /H I PA A / p re s e n t a t i o n s

http://www.mahealthdata.org/forums/events/events2003/HIT_20030207

http://www.ehealthinitiative.org/HIPAA/presentations/

http://www.ehealthinitiative.org/HIPAA/presentations/

The fastest,

easiest way

to keep up

with the

Internet’s

impact on

health care.

Free Daily Internet Health NewsSign up today at www.ihealthbeat.org

Documents

Electronic Medical Records: Lessons from Small Physician …€¦ · Electronic Medical Records: Lessons from Small Physician Practices Pre p a r ed for: CALIFORNIA HEALTHCARE FOUNDATION