Memoriai Health System is an eighthospital integrated health care system in the midwestem United States. The health system has two downtown agship tertiagg care hospitals each licensed for more than Till] beds. located in the two major metropolitan areas served by the system. The remaining six hospitals are community-based facilities, ranging in size from EDI] to slim beds. These hospitals are located in the suburban and rural areas served by Memorial Health System. Four years ago, the system's board of directors approved a rnul'timilliondollar initiative to install an enterprisewide clinician provider order entry [CPUEJ system intended to dramatically reduce medical errors. Today, the system is far from fully implementedI and in fact has been removed from all but one of ! a the two tertiary care facilities, where it remains in pilot adopter status. At the time the board approved the SPICE initiative, the project was championed by Fred Dryer. the CEO, and was closely supported by Joe Roberts, the chief information officer {CID} of the health system. Even during its proposal and evaluation by the boardi the project was considered controversial by some of the health system's stakeholders. For exam ple, many of its physicians, who are communitybased independent providers, were adamantiy opposed to the CPOE system. They worried that their workload would increase because DPDE systems replace verizral orders with computerentered orders by doctors. Dr. Mark Allen, a primary care physician commented, \"The hospital is trying to turn me into a $12an-bour secretary, and they aren't even paying me $12 an hour.\" In securing board approvai, Dryer and Roberts presented an aggressive implementation plan that called forthe requirements anatysis, request for proposal {EFF}, vendor selection, and project implementation to be completed in less than 18 months in all eight hospitals. During the discussion with the board, several members questioned the timeline. One noted, "it took you two years to set up email, and everyone wanted e-mail. This will affect every clinician in every hospital. Do you realg think you can do this in 13 months?\" In an effort to demonstrate results. Dryer and Roberts demanded results from the clinical and IT team formed for the project. By this time, a rushed requirements analysis had been completed, an RFP' issued, a vendor selected, and a contract signed. The acquisition process took a little more thanimonths, leaving a year for the implementation. In protest, a number of prominent physicians took their referral business to the other health system in the area that seized on the controversy by promising that they would not use a CPDE. Shortty there after, the two leading champions for lDPDIEDryer and Robertsleft Memorial Health System. The chief medical ofcer, Barbara Lu, who was a vocal opponent of the project, was appointed interim CED. Although Lu opposed the project, many members of the board still supported it. In addition, none of the board members wanted to lose a substantial down payment to the vendor, so Lu was instructed to proceed with implementing the system. Lu appointed a close colleague, Dr. Melvin Sparks, to serve as the interim CID of the system. Sparks was both a practicing radiologist and a degreed computer engineer, so Lu thought he would be an ideal CID torthe system. Sparks hired Saliy Martin as the executive project manager overseeing the implementation. After evaluating the progress made to date and preparing a detailed thousandstep project plan, Martin reported back to Sparks on the status of the project with an exceptionally detailed report. Several key points were noteworthy in her regort. Due to the rushed requirements analysis, several key woddiow and system integration issues were missed. Consequently, to complete the project in the remaining 12 months, the organization would have to do the following: - Double the IT staff assigned to the groiect from 15 to 32 geogle. I Purchase approximately $5,ill]il in integration software not already budgeted. o Aitematively, the scoge of the gro'ect could be reduced from an enterprise deployment to something less than that. o Alternatively, the duration of the groiect could be doubled to 2.4 months, keeping the staff but not avoiding the $5,l]il[l software cost. Dr. Sparks did not resgond well to the news, exhibiting a great deal of anger at Martin, who was not working for the health system when the groiect was scoged and budgeted. Sparks yelled at Martin and told her neverto come back into his ofce with bad news again. Herjob, Sparks screamed, was to \"gure out how to turn bad news into good news or no news.\" As she left Sparks' ofce, Martin resolved never to convey bad news to Sparks again, no matter how serious the issue was. Over the nerd 12 months, the project progressed but got a bit further behind schedule each week. Martin reminded herself that she wasn't conveying E news to Sparks. ln each status review meeting, Martin always gresented a groiect schedule that was on scoge on schedule, and on budget. During this time the health system took on a number of other imgoriant [T initiatives reguiring human ers.ources Each time another grolect fell behind schedule Sgarks took resources from the SPICE groiect. From the 16 geogie originally budgeted, the team was reduced to eight. tThe only positive aspect was that the project, which was costing money even though it was making little or no progress, was expending less cash as it made no progress. As the project went into its 16th month, two months before the scheduled launch, nearly all the gro'ect budget had been consumed, andin an effort to save moneythe enduser training budget was cut to the bare minimum. At the same time, some doctors who had not left the system attended the SPICE vendor's annual user group meeting. They saw the release of the vendor's most recent system and immediately decided they wanted it for Memorial Health System. Upon returning to the hospital, the doctors met with Sparks and gersuaded him that the only hope for enlisting physician support for the changed workow was to adopt the newest version of the software, which wasjust being introduced. The physicians told Sparks they had persuaded the vendor to appoint Memorial Health System as an alpha site for the new software. 1Il'l'hen Sparks informed Martin ofthe change in the scoge ofthe groiect, Martin was concerned, but rememben'ng Sparks' reaction to bad news, she kept her thoughts to herself. She framed her questions in the form of the risks that such a maior change in direction might cause with so little time to recover. Sparks smiled and told Martin, \"Don't wony; it will all work out.\" So two months before the launch Marlin worked with her team to alter the groiect work glan to install the new software. test the software, congure the software and interfaces, and train the usersall in two months even though the same activities had taken almost eight months the first time. The scheduled date for the launch arrived, and all eight hospitals went Live on the new CPDE system on the same day. The new software had aws. The lack of enduser training was apparent, and the many requirements missed during the analysis became immediately obvious. Doctors could not log onto the system, and nurses could no longer enter orders. Patients were kegt waiting for medications and tests. After several days of this, Lu instructed Sparks to decommission the CPCIE system and revert back to the manual procedures. An unknown physician was quoted in a maior health care publicationunder the titie \"CPUE Doesn't Work\"describing the debacle at Memorial Health System. During the proiect postmortem. Sparks expressed surprise the project was not going as planned and asked Martin why she had not been more forthcoming about the problems issues: and n'sks. The vendor took six months to x the aws in the software, and31] months into the projectCPDE was launched again. However, this time it was in one ICU in one of the tertiary care hospitals. Four years after the beginning of the project, this is the only unit in the entire health system in which CPOE is operational