1 Australia’s Biggest Healthcare IT Initiative in Trouble: Kill It or Keep It?* TIENDUNG LE Introduction In February 2012, David Davis — Victoria's Health Minister — was still considering what he would recommend to the new premier Ted Baillieu with regard to the HealthSMART program. The program was one of the biggest IT initiatives ever taken in Australia, developed to modernize and replace management systems throughout the Victorian public healthcare sector. The Labor Government approved the program in 2003 with a $323.4 million budget and to be completed in four years. Nine years on with almost 50% cost overrun, the main component — the Clinical System — had been installed at only four out of ten originally planned health services. Davis had to decide what he would recommend, continuing funding the program until completion or dumping it in entirety and using the public money for other projects. The Victorian Health System and Technology The Victorian health sector was very large and complex with the public hospital system alone employing more than 73,000 full-time equivalent (FTE) staff and having a turnover of more than $10 billion per annum. It had 21 major health networks, 22 sub-regional health service providers and 43 small rural services. The community health program was provided by approximately 100 organizations in Victoria operating from more than 350 sites in every local government area. Information technology networks in Victoria's public hospitals in the early 1990s had been considered poor in a supposedly world-class health program. Paperbased records, people queuing to use the available computer terminals and difficulty in sharing information with offsite colleagues are the norm. Patients in public hospitals suffered serious problems due to poor IT systems. If a person was treated in the emergency department, and doctors and nurses were not able to get access to the person's history of care with their general practitioner. Without a central health database storing the history of patients' illnesses, treatments and medications, medical staff would have to piece together this information from the patient's own memory. Once the patient was admitted, staff on the wards would have to queue to use a computer so they could access the patient's hospital records and diagnostic information. When staff were finally able to get on a computer, the system was slow and clunky and crashes were common. The lack of connectivity between different areas of the health system meant medication lists, tests, scans and other diagnostic tools were often repeated. Health dollars and clinicians' time were thought to be wasted chasing results and duplicating services in a stretched public hospital system. The quality of care was compromised and patients were at increased risk of mistakes being made in their treatment, diagnosis and prescription of medication. While not common, prescribing the wrong medication could be extremely serious. In the five years from 2006 to 2010, nine Victorians had died in public hospitals after having been given the wrong medication, and there had been a further 25 nearmisses. Medication mistakes also led to higher Dr. Tiendung Le prepared this case solely for class discussion purposes. It is not intended to illustrate effective or ineffective handling of a situation or to serve as a source of data. Although the case is based on a real situation, the information contained is not necessarily correct and true. No part of this case may be reproduced, stored, or distributed without permission from the author. Dr. Tiendung Le is Lecturer in Project Management at the School of Property, Construction and Project Management, RMIT University, Melbourne Australia. Email: [email protected]. Copyright © 2016 by Tiendung Le. Version: July 28, 2016 2 hospital readmission rates because the patient was not taking the medication needed to control their illness, or the medication had made their condition worse. In a report by the Auditor-General in 2008, it was noted that medication error was a leading cause of adverse events in the healthcare sector in Australia ‘with 10-20 per cent of adverse events being drug related, and up to half of these preventable.’ In 2000, Department of Human Service (DHS) (now Department of Health – DOH) estimated the annual cost of these adverse events in Victoria to total $378 million. Other research in 2003-04 estimated the cost to be $460 million. To address patient safety issue across the state, Victoria developed Whole-of-Health Information and Communications Strategy 2003-2007. The Strategy aimed to modernize and replace management systems throughout the Victorian public healthcare sector. The Birth of the HealthSMART Program The HealthSMART Program was developed to implement the Strategy. A funding proposal requesting a budget of $323.4 million was submitted by DHS to the Government in February 2003. In the proposal, two options were presented: 1) ‘do nothing;’ or 2) ‘build a consistent ICT foundation across half of the Victorian public health service.’ This original funding submission, which was based on a high-level strategy document containing a 14-page implementation plan, was deemed to be ‘Cabinet-in-confidence’ and could not be accessed by the public and the Parliament. (For information on the last four state elections in Victoria, see Exhibit 1.) The proposal was approved in the 2003-04 budget. The purpose of the HealthSMART Program was to: • Reduce risk through replacement of a number of obsolete core systems with ‘off-the-shelf’ products; • Reduce risk through the establishment of a robust shared ICT service for hospitals and community health centers with appropriate disaster recovery; and • Improve patient care, in particular reducing medication errors, through the introduction of electronic medical records management. The Program had ten components, of which three main ICT systems were: • The Clinical System, which would include health records, electronic scheduling, diagnostic services, results reporting and e-prescribing capabilities; and • The Patient and Client Management System, which would comprise an integrated patient and client management system to store patient records throughout their care in hospitals and/or community health settings. • The Finance and Supply Management System, which would support core health agency financial management, including general ledger, assets and materials management. Another component of the system related to an integrated human resource and payroll management system was delivered by a separate vendor; The Clinical System component of the Program had the potential to deliver the most benefit to health agencies and patients. HealthSMART Services was also established within the DHS to plan, build, run and support the infrastructure on which the applications were delivered. Exhibit 2 shows the list of the components and applications with their corresponding budgets. Of these components, the Clinical System had the biggest budget and was considered most important and complicated. It was envisioned that the system could allow clinicians to control a patient's medications, generate prescriptions, and importantly, be alerted to potential medication mistakes when patients are allergic to a particular drug, or an incorrect dose has been prescribed. 3 The list of the applications, the number of tenderers, the number of tenderers shortlisted and the names of the final contractors is shown in Exhibit 3. The Business Case According to the Public Accounts and Estimates Committee of the Parliament Victoria, there was not enough detail in relation to whole-of-life costs and benefits, expected outcomes and value-for-money considerations for the Program. CSC Corporation, the contractor that was delivering the Patient and Client Management System, stated: “Typically, a business plan would set out the objectives of the project. I am not aware that there has been any of that or any benefits realization subsequent to the project...what we are talking about here in HealthSMART is a standardization project that has potential IT benefits if it is done well...one of the benefits of standardization is that you have a core platform so that when you have an innovation you can do it once and apply it to many... I do not think the customer is getting as much out of the software as they can, I truly believe the client cannot point to why they did it or what benefit they are getting from it.” Cerner Corporation, the contractor of the Clinical System, believed that the implementation could have been improved by the government committing to the establishment of “baseline measurements of required metrics” to ensure that benefits and outcomes could be measured. Commenting on the implementation of the Clinical System component, the CEO of Austin Health stated: “This project will deliver long term economic and environmental benefits for the Victorian community. Patient safety will be improved through timely access to electronic information, appropriate alerts and decision support. Medical error related to illegible orders and missed test results will be reduced. This will be to the benefit of patients treated in hospitals using this system.” The Implementation of the Program The scope of works under the HealthSMART Program was to replace approximately half of the existing ICT systems across hospitals and almost all metropolitan community health centers with standardized systems. Funding and Costs The Finance and Supply Management System was originally implemented in acute metropolitan health agencies and Bendigo Health. As a result of its success, rural health alliances requested that it be implemented at all five Victorian rural health alliances. In total, the Finance and Supply Management System was implemented at 12 sites at a cost of $25.9 million, which was lower than the original budget of $26.3 million. The original contract for the Patient and Client Management System was for $50 million in 2003 with a local software company, iSOFT Australia Pty Ltd. This company ran into some difficulties overseas and was subsequently acquired by CSC Corporation Australia. The system had been implemented at 10 heath service agencies by the June 2010 deadline and $11.9 million under budget. The Clinical System had the potential to deliver the most benefit in terms of the way health agencies record patient treatment. The system offered end-to-end ordering of pathology and radiology tests, reporting of results, inpatient medication management, discharge prescribing and discharge summaries with support for clinical decision-making and clinical alerts designed to enhance patient management and safety. Prior to this all these processes were hand written and paper-based. 4 The original contract had been signed with the vendor in March 2006 for $79 million. The contract amount rose to $96 million later that year. Towards the end of the 2011-12 fiscal year, the cost for the Clinical System was close to $135 million. By the end of the 2011-12 fiscal year, the project would have been running for nine years, more than double the time it had been planned for originally. The major component, the Clinical System, was rolled out to only four of the ten planned health services. These four services were: • Austin Health, • Peninsula Health, • Eastern Health, and • The Royal Victorian Eye and Ear Hospital. In the 2008-09 state budget the government allocated an another $104 million in operating costs. Therefore, by the end of the 2011-2012 fiscal year, the government had expected to have spent $427 million implementing HealthSMART. If the further $34 million funding to support HealthSMART Services was included, the total estimated project expenditure would have been $471 million. DOH estimated that a further $95 million would be needed to implement the Clinical System at the remaining six sites. This would bring the final project cost to around $566 million. The Clinical System alone would have cost approximately $230 million. The clinical application had been more challenging than DHS had estimated. The HealthSMART Program Director once commented that DOH had overestimated the standard of local infrastructure and the skills of local ICT staff while underestimating the full costs associated with training doctors and nurses, including backfilling their roles. Functionality and the Business Processes The Program was to deliver a “State-wide footprint” by standardizing systems in the health sector in Victoria. The systems would be built as a common platform that would be implemented multiple times. However, the governance and autonomy of separate health services in Victoria had resulted in significant differences in business processes and practices. There were significant variations in ICT capability and maturity among the health services. One of the first four hospitals having the Clinical System implemented believed that standardizing information and reporting processes and ensuring cross-systems compatibility and information sharing would have been better than standardizing IT systems within the hospitals. CSC Corporation stated: “The reality is while we all like to think that hospitals and agencies do processes the same way, they do not... therefore, when we started to implement it in agencies, we started to come across issues that they do not admit patients that way, they do not discharge patients that way, they do not transfer patients that way.” Austin Health viewed the decision to adopt a State build approach to the HealthSMART clinical system as good because it was aimed at ensuring some degree of consistency and commonality across health services which would provide benefits for staff and patients moving between health services. However, the hospital believed that the development should have been undertaken by a lead hospital rather than DOH and based on an existing build from elsewhere in Australia. This comment was supported by Cerner Corporation, who said that there would have been less usability issues and delays if the State-wide footprint had been designed and built within a lead health service rather than in a ‘lab’ type environment. Contract Agreements and Management A tripartite contractual arrangement was created among vendor companies, DOH and the health service agencies. There was a central heads agreement between DOH and vendors in addition to 5 implementation contracts between health services and vendors. This complex nature of the contractual relationships was believed to have made the management and control of the program bureaucratic and slower than necessary. DOH stood between the vendor and health agencies that were accountable to their own boards of management and to DOH. Health services were required to prepare their own individual business cases for the system components and seek approval from their own boards. The vendor of the Clinical System viewed this arrangement as a delay factor in the implementation of systems. The tripartite arrangement was also cumbersome in process change requests that needed to go through DOH and then to the vendor. The CEO of Austin Health said “the change request process established by the DOH was initially onerous, overly bureaucratic and slow and was hampered by the fact that all implementing health services had to agree to every change.” However, in 2011, a number of initiatives were implemented to streamline the process and enable changes to be made locally. These initiatives were seen to have improved turnaround times dramatically. The Secretary of the Department was the Chair of the Board of Health Information Systems, DTF and the Department of Premier and Cabinet (DPC) were members of this Board and also system users were strongly represented. Staffing and Capacity The project experienced difficulties in recruiting people with proper sets of skills and experience in both ICT and project management due to salary band and headcount restrictions. Contract staff and secondments from health agencies were used to fill key positions. The lack of technical and project management expertise added further burden to the timely delivery of the program. For example, in 2008, at the time of the Auditor-General’s report, key positions including the portfolio manager and account manager for clinical systems were vacant. Furthermore, the combined pressure of crosstraining by vendor on the new system and services that needed to be delivered to customers during the roll-out resulted in high turnover of staff. One of the vendors commented: “It was evident that there was a high turnover of staff in the HealthSMART services, and I think there are probably a couple of reasons for that. One was that it was a high-pressure environment in HealthSMART services, because the time lines were aggressive... So generally those guys were working very long hours, because when you are trying to upgrade in a production environment you can only do it after hours, so there was a lot of night work and some weekend work... Obviously those guys were coming under increasing pressure, because they were getting their brains filled with us trying to train them on the product as well, plus all the internal pressures.” Perspective of Austin Health Austin Health believed that the Clinical System solution offered an integrated platform for the hospital and would enable the delivery of a full electronic health record. It was committed to continuing its relationship with the vendor and purchasing additional modules to support its activities. Austin Health was probably one of the leaders of clinical systems implementation in Australia, hosting visitors from interstate and from other Victorian hospitals. In 2012, the hospital was working towards a fully paperless environment and hoped to achieve a rating as an ‘Electronic Medical Record Adoption Model, Level 6-7’ hospital within the next two to four years. An internal audit at Austin Health covering various implementation aspects (e.g., governance structures, project controls, stakeholder management and communication, implementation plans, quality checkpoints and criteria for each phase of the product implementation) gave the project an overall rating of “Good.” The hospital did not agree that the system had had negative impact on patient safety as stated by a report by the Ombudsman in 2011. As the project progressed, Austin Health developed the required knowledge, skills and experience to manage the system locally and therefore provide quicker 6 response to user requests for changes. The hospital recognized numerous benefits and functionality that the Clinical System could bring about. According to the hospital, the system was flexible and could support various configurations and workflows. It managed more than 4,000 orders for pathology and radiology each day. The hospital had a Benefits Realization Plan and a program to collect data on benefits post-implementation. From the preliminary observations, efficiency of the junior doctors in surgical and medical units seemed to be improved (e.g., reduced time in ordering pathology and radiology). A Pharmacy-led evaluation to measure the impact of electronic prescribing for discharge indicated a number of benefits and improvements, such as: decrease in the use of error prone abbreviations, improved consistency between the discharge medication regimen and the discharge summary sent to the GP, significant reduction in corrections and additions to prescriptions to ensure reimbursement by Medicare Australia, and increase in ward pharmacist time spent on discharge process. In terms of workflows, the impact of electronic orders within the pathology department had been positive. The number of activities in the workflow was reduced by 17; this improvement led to improved productivity. The time needed to process a specimen was cut by three times while the average number of episodes that could be processed in an hour increased by 39. Also, the staff required to process electronic requests has reduced by two FTE (full-time equivalent) people. The nursing staff at the hospital had reported that the implementation of the new system had been a positive experience and that it assisted them in managing the wards more efficiently and would lead to increased patient safety. Austin Health also realized that the implementation of the system had indicated that some clinical practices might not have been ideal previously, or might not have been consistent across the health service. According to the hospital, the vendor had committed to working to develop solutions in response to a number of changes requested by Australian clients. It was developing a business case for the chemotherapy module but funding had not been identified. Austin Health had independently purchased the Cerner Advanced Clinical Documentation module for progressive implementation commencing in 2012. The Cerner Emergency Department system would also be implemented in 2013 and the Cerner Surgical system was a high priority for the future. The CEO of Austin Health said: Despite the delays and problems encountered it is the view of Austin Health that this project is a success and will deliver the predicated benefits and more. Perspective of the Royal Victorian Eye and Ear Hospital (RVEEH) The RVEEH had responsibility for local delivery of the Clinical System as well as coordinating its activities with other health services, DOH and the vendor. The Chief Executive of RVEEH believed that the original project timeline had been quite ambitious. DOH intended the system to provide a ‘Statewide footprint’ and as such the system was broadly based on the requirements of a general tertiary hospital and there was limited flexibility to adapt to local hospital requirements. This was especially an issue for the RVEEH which is a smaller scale specialist hospital. In December 2010, the RVEEH evaluated its initial implementation of the system through a Post- Implementation Review to assess performance against the original objectives. The hospital also commissioned a review from a chartered accounting firm in April 2011. The firm provided a Remediation Plan that was later implemented. The hospital advised that it has developed an ICT Strategy that is aligned with HealthSMART and the state-wide strategy. It was also committed to the progressive development and implementation of an electronic medical record using the standard International Healthcare Information and Management Systems Society (HIMSS) model of adoption. 7 RVEEH recognized “improved prescription legibility' as a result of the implementation of the Clinical application. However, it recorded an initial loss of 13.5 per cent in outpatients’ medical productivity, which later fell to 8 per cent. While the impact was considered typical following the implementation of an electronic medical record, additional staff had to be employed to maintain the same throughput of patients. Main dissatisfaction with the system mostly came from senior medical staff who found the system difficult to use. The dissatisfaction level seemed to reduce gradually with familiarity. However, the hospital was faced with a number of significant challenges. The ICT infrastructure required to run the sophisticated electronic Clinical System was underestimated and needed to be upgraded. Medicare stilling required a paper prescription for reimbursement. A significant factor in outpatient productivity at the RVEEH was related to issues associated with printing of prescriptions, which needed to be resolved if electronic hospital prescribing of outpatients was to bring about the full efficiency. Also, RVEEH had a large number of visiting medical officers who would only be in the hospital a few hours each week. That put extra burden and complexity to training people to use the system. Perspective of Cerner Corporation Commenting on the achievements of the Clinical System, Cerner Corporation’s Vitoria General Manager said: “...there appears to be little recognition of the significant asset that has been created from this project. This project has effectively been the commencement of pharmaceutical informatics in this country in implementation of medications management across hospital inpatient and outpatient settings. The assets created and knowledge gained from this project has the potential to significantly and positively impact patient safety and experience across our nation and the resultant solution can be leveraged both further in Victoria and across the rest of Australia. The alignment with NeHTA standards will ensure interoperability of critical patient data between health care venues regardless of jurisdiction.” The NeHTA was the National E-Health Transition Authority, which has been established by the Australian, State and Territory governments to develop better ways of electronically collecting and securely exchanging information between health agencies. ‘System Is Sick, not Dead’ In the article titled ‘System is Sick, not Dead’ published on January 25, 2011 on The Sydney Morning Herald, Dr. Harry Hemley, AMA Victoria president, made a strong statement: The Baillieu government has inherited an enormous problem in hospital IT that Health Minister David Davis has described as the myki of health. However, the solution shouldn't be to abandon this failing project. Quite the opposite. It's time the Baillieu government stepped up to the challenge and turned the HealthSMART mess around. Most importantly, the Premier should commit to the delivery of medication management systems in all Victorian hospitals by the end of 2012. It's an ambitious target, but the benefits for patient care will be immeasurable. We can't make the necessary increases in efficiency and improve patient care without better IT. And, in fact, he was hopeful: ‘I've had discussions with government and they know it will improve efficiency and safety and they want to institute it, they just have to find the money.’ 8 Kill It or Keep It Davis said recently: “… On one hand, there's the recognition that you need a strong and forward IT system, which is critical to high-quality, efficient healthcare. However, HealthSMART has been implemented at a much greater cost than was initially proposed and is well behind time and not fully effective in a number of key locations.” For Davis it was a decision between killing or keeping funding the program. Davis understood the HealthSMART program was delivering various benefits to the health services in Victoria. Austin Health was a good example of how health services could use HealthSMART to be more effective. Also, completing the program would allow for universal IT infrastructure across health services in Victoria. However, Davis knew if the program was to be continued he would have to secure further funding and make sure it would deliver what it had promised. That was nothing to be sure of. On the other hand, a decision to discontinue would mean more money for other pressing projects in the state and individual health services would have more flexibility and autonomy in deciding what IT systems would be suitable for them. It was not a straight forward situation for Davis and he was not sure he should recommend killing it or keeping it. Davis knew the new government faced ‘a genuine dilemma.’ *This case was adopted from the 112th report of The Parliament of Victoria’s Public Accounts and Estimates Committee titled “Inquiry into Effective Decision Making for the Successful Delivery of Significant Infrastructure Projects” and used various other public domain sources. 9 Exhibit 1. Victoria’s Last Four State Elections Date Party Winning Election Comments 18 September, 1999 Labor Kennett Liberal-National Coalition Government defeated. Bracks minority Labor Government takes office. 30 November, 2002 Labor 25 November, 2006 Labor John Brumby took over as Premier in July 2007 after Steve Bracks quit the re-elected Bracks Government. 27 November, 2010 Liberal-National Coalition Brumby Government defeated. The Baillieu Government takes office. Source: Victorian Electoral Commission Exhibit 2. HealthSMART Program Budget Breakdown by Component Program Component 2003 Budget ($ million) Clinical System 79.0 Patient and Client Management System (Integrated) 50.0 Finance and Supply Management System 26.3 Client Management System (Community Health) 13.5 Human Resources Management Systems (including Payroll) 4.0 Sub-Total 172.8 ICT Support 57.9 ICT infrastructure 66.7 Program management 7.2 Related health applications 18.8 Total HealthSMART Program 323.4 10 Exhibit 3. HealthSMART Program Tenderers & Vendors Program Component No. Tenderers No. of Tenderers Shortlisted Awarded Contract Clinical System 9 2 Cerner Corporation Pty Ltd Patient and Client Management System (Integrated) 5 - CSC Corporation Australia (formerly iSOFT Australia Pty Ltd) Finance and Supply Management System 18 3 Oracle Corporation Pty Ltd Client Management System (Community Health) 4 2 Trakhealth Australia Pty Ltd Human Resources Management Systems (including Payroll) 8 3 Frontier Software Pty Ltd Rostering 7 4 Kronos Australia Pty Ltd Managed Services 5 3 Fujitsu Storage Solution 5 2 Hitachi Data Systems Middleware Solution 11 2 SeeBeyond Technology Corporation Picture Archiving & Communication System (Digital x-rays) 8 3 Fujifilm Australia Pty Ltd