PACS upgrades require careful advance planning

A particular problem facing large teaching hospitals is that even as health information systems such as PACS and RIS require regular updating, the staff members using these systems are constantly changing. Interns and registrars will be moving through the hospital on fixed-term training placements. From a training point of view, the hospital IT support staff would prefer to have stable systems. This would allow training modules to be available online to a wide range of clinical staff who could learn the systems with a minimum of time, effort, and cost.

PACS and RIS vendors tend to provide a major new system release every few years, despite hospitals' requirement for IT stability. Software vendors prefer their major customers to upgrade to the new version once it's released. This simplifies their task of providing worldwide system support. There are also practical reasons why a hospital should undertake a major software upgrade. The latest package is likely to have additional features and resolve any existing problems with the previous version.

Hospitals consequently have to make a series of important decisions when given the opportunity to move to a newer version of an existing PACS or RIS. These decisions involve financial and nonfinancial costs. A decision framework is required to ensure that all relevant aspects of a software upgrade are considered. This should cover the medical imaging department, which has responsibility for the day-today operation of the PACS and RIS; the IT department, which will cover the impact of any software changes on the hospital enterprise IT environment; and the implications for training existing clinical users.

Vendors often indicate that major PACS and RIS software upgrades are "free." In other words, the new version of the software is provided at no additional cost beyond existing monthly maintenance contract fees. There may, however, be costs for additional hardware required to support the new version of the system. Hospitals may also be billed for "professional services," that is, assistance from the vendor to install the new system. These service costs are often based on an hourly rate.

Planning for a major upgrade to a well-used PACS or RIS requires careful planning and timing. Ideally, the upgrade should occur when hospital clinicians and radiologists are least busy. The PACS upgrade plan would need to include an augmented help desk within the hospital enterprise to provide support. Resources such as this should be viewed as a financial cost to the hospital, over and above any professional services fees paid to the vendor.

The assessment of upgrade costs should be realistic. A business case outlining costs and benefits must be prepared and submitted for approval as part of the hospital's annual budget cycle. Inherent delay in the budget approval process means that the timing issues of any proposed upgrade are important; the time for budget approval must simply be factored into the timeline.

PERFECT TIMING?

A major upgrade to a teaching hospital's software system will have an impact on the daily work practices of new and existing clinical users. A detailed analysis should be performed of staff training needs on the new PACS or RIS. This analysis should cover training methodology, development of new online training modules, timing of the rollout in a teaching hospital environment, training resources, and costs.

An enterprise-wide PACS and RIS may serve some 2000 to 3000 clinical users. It may be necessary to develop a pretraining strategy based on the priorities of certain hospital departments and major PACS and RIS users. This should ensure that users' disruption is kept to a minimum once the new system becomes operational. Training activities will clearly place great demands on PACS and RIS administrative support staff and will require other PACS- and RIS-related activities to be rescheduled.

Few major teaching hospital enterprises have RIS and PACS from the same vendor. The upgrade status of the RIS may need to be considered separately. If the hospital's RIS vendor is planning to release a new version, it may be appropriate to undertake this upgrade before considering the PACS upgrade. Retraining RIS users should also then become a priority.

A software system upgrade will require a great deal of planning and will affect clinical staff of all grades. The effect on external hospital referrers, such as specialist doctors and GPs, should not be forgotten, either. These practitioners may have become accustomed to receiving radiology reports from the RIS by secure Internet connections. They may also receive digital medical images on CDs or by accessing secure areas of the hospital's website. System testing of the proposed PACS or RIS upgrade should ensure that at least the same level of service will be provided to these users. Any necessary retraining should be provided as well.

IN-HOUSE SUPPORT

The hospital IT department will have invested a significant amount of time and effort in establishing a satisfactory working linkage between the existing RIS and PACS. It will be essential to prepare a realistic estimate of the timelines and programming resources required to support the implementation of a new RIS or PACS interface. This task should be given priority over other hospital IT activities. Issues to be addressed by the IT department include system interface testing, data storage upgrades, and communications network analysis.

Major public teaching hospitals operate in complex IT environments that support a large number of interacting computer systems. It would not be unusual for a number of systems from a variety of vendors to all be running on as many as 2000 standard PCs. The impact of the new version of the PACS on other hospital IT systems needs to be examined. For example, the hospital's financial system may use a different version of Java from the version on the proposed PACS. This may lead to problems when both systems are operating on the same hospital PC.

The upgraded PACS may support more advanced data media. This could provide an opportunity for the hospital enterprise to migrate old radiology images to a new storage system. Such infrastructure data storage costs should be considered on an enterprise basis rather than as a system cost to the PACS. Current PACS storage maintenance costs incurred by the medical imaging department would need to be compared with future data storage costs.

Moving to a newer version of a PACS may have implications for network traffic. This is especially likely if the proposed system uses a single image database, whereby all images and associated prior studies are sent to or retrieved from a central storage point. The hospital communications network may previously have been based on multiple local image databases at individual sites. This would have minimized enterprise-wide traffic.

On moving to a centralized store-and-retrieve model, peak network traffic loads should be reconsidered and costs provided for network upgrades. Given the increase in imaging studies being used enterprise-wide and the greater use of multislice CT, this may be an appropriate time to recalculate image production, network loads, and associated response times. The introduction of advanced image compression algorithms could be considered as well.

IMPACT ON IMAGING

The medical imaging department is responsible for maintaining a certain level of service for clinical users. This remains true even if the implementation of a new PACS or RIS is unsuccessful. A backup strategy in case of such an eventuality is essential. Smaller scale backup plans should already be in place to cover the breakdown of components during daily use. For example, spare storage capacity should be available to hold patient images on a temporary basis if the network link to the main archive fails. A master backup plan should be developed in a similar manner, enabling the original version of the PACS and/or RIS to be reinstalled in the case of an unsuccessful upgrade.

The medical imaging department may take the opportunity to upgrade its existing services to clinical users. If the hospital enterprise is considering introducing a cardiac PACS to integrate with the radiology PACS, for example, it then makes sense to question whether the newer version of the PACS would support this activity. The business case should cover the various cardiac modalities involved (CT angiography, digital subtraction angiography, and echocardiography) as well as make realistic estimates of image volumes, response times, and data storage requirements.Teleradiology is becoming an important aspect of the service provided by the medical imaging department at a major teaching hospital. This service includes linking to clinicians' consulting rooms and homes, and providing online second opinions to other institutions. Teaching hospitals have previously used a stand-alone teleradiology product in addition to the PACS and RIS. An upgraded PACS should, ideally, include an efficient teleradiology and telecardiology module to support the hospital's consulting activities.

Each department in a major teaching hospital is likely to use key performance indicators to measure staff productivity and response times to customers. Attainment of these indicators may be linked to promotion prospects and financial bonuses. Data required to make these measurements may be extracted from the PACS database, using specially designed programs. If the PACS or RIS is upgraded, this information should still be obtainable. A seamless transition is essential if the measurements are to remain accurate.

In conclusion, a decision framework must be developed when considering a major PACS or RIS upgrade at a teaching hospital. The framework should include associated costs, a statement of the hospital enterprise's resources, and a timetable that accounts for clinical requirements. A number of complex issues must be considered, and existing clinical users must be retrained. Close cooperation is needed among the medical imaging department, the IT department, and clinical users to ensure that the upgrades support clinical work and do not interrupt teaching activities.

ACKNOWLEDGEMENT

This article is based on a paper presented at the 21st International Congress and Exhibition on Computer Assisted Radiology and Surgery (CARS), held in Berlin, June 2007.

Further reading
Ahlers C, Klos G, Kurz K, et al. Open source software tool for vendor independent integration of PACS clients with the teaching database MIRC according to IHE TCE. Presented at the 21st International Congress and Exhibition on Computer Assisted Radiology and Surgery, Berlin; June 2007:S299-S300.
Blado ME, Tomlinson A. Monitoring the accuracy of a PACS image database. J Digit Imaging 2002;15(Suppl 1):87-95.
Caffery L, Manthey K. Implementation of a Web-based teleradiology management system. J Telemed Telecare 2004; 10(Suppl 1):22-25.
Crowe B, Sim L. Implementation of a RIS/PACS and image transfer system at a large public teaching hospital--assessment of success of adoption by clinicians. J Telemed Telecare 2004;10(Suppl 1):25-27.
Crowe BL, Sim L. An assessment of the effect of the introduction of a PACS and RIS on clinical decision making and patient management at Princess Alexandra Hospital Brisbane, Australia. Final report--December 2004. Presented at the 19th International Congress and Exhibition on Computer Assisted Radiology and Surgery, Berlin; June 2005:964-967.
Halsted MJ, Perry LA, et al. Improving patient care: the use of a digital teaching file to enhance clinicians' access to the intellectual capital of interdepartmental conferences. AJR 2004;182(2):307-309.
Krotz D. Efficiencies of PACS make challenges worthwhile. Diagnostic Imaging Asia Pacific 2001;(AOCR Suppl):11-12.
McEnery KW, Suitor CT, Thompson SK, et al. Enterprise utilization of "always on-line" diagnostic study archive. J Digit Imaging 2002;15(Suppl 1):81-86.
Wehrle T, Kunzel U, Reindl P. PACS workflow and economic consequences. Medica Mundi 2000;44,1:52-65.
Wendt G, Peppler W, Edwards W, et al. Clinical implementation of the IHE presentation of grouped procedures integration profile in a multivendor environment-workflow modification and barriers to implementation. J Digit Imaging 2002;15(Suppl 1):64-66.
Wendt G, Peppler W, Edwards W. PACS disaster recovery procedures--evaluation in a clinical environment. J Digit Imaging 2002;15(Suppl 1):112-113.

Mr. Crowe is a partner at Bernard Crowe & Associates. Dr. Sim is PACS support manager at Princess Alexandra Hospital in Brisbane, Australia. Ms. Whitter is PACS project manager at Eastern Health in Melbourne, Australia.