Why not keep the old archive?

In most cases, it is technically feasible to simply keep the old system and query it for images when needed. However, a number of factors speak against this option:

Complexity

• Technical
  Aside from practical configuration management for a two-vendor two-archive system, the split in archive between the "old" and "new" must usually be maintained permanently in the institution’s workflow and staff training.
• Administrative
  Retaining the legacy system also retains one more vendor to deal with, one more set of maintenance procedures to maintain, one more staffing consideration, and several more headaches that administrators would just as well avoid.

Economics

Because of plummeting costs of disk storage in recent years, retaining the legacy system generally entails costs that generally exceed the expense of additional storage on the new system.

• Legacy system support costs.
  The legacy PACS must be maintained, either through annual support agreements with the legacy vendor, or by repairs performed as needed on a time-and-materials basis. Both these options are costly, and the latter should be considered only if the healthcare provider organization has a staff person familiar enough with the legacy system to make appropriate support decisions for the old system.
• Space, power and cooling.
  Machine room space is at a premium in many provider facilities, and legacy PACS archives typically occupy a large footprint. Providing uninterruptible power and cooling for these archives is also a non-trivial expense.

Reliability

• Regulatory Compliance.
  Many, if not most, old systems are not compliant with the HIPAA Security Rule, for which the compliance date was April 20, 2005 for most covered entities (April 20, 2006 for small health plans). Legacy data must in most states be retained for 5 or 7 years. However, most PACS archives also contain pediatric cases mixed in with the adult cases, and the retention of pediatric image data to 5-7 years past age 18 is required in most jurisdictions. Mammography data, present in some archives, must in the USA be retained indefinitely. This creates an extremely long support tail for at least some of the legacy data.
• Legacy vendor support uncertain.
  The availability of support for an old archive may be poor, and is certain to become worse. Legacy support is not a key business goal at any vendor, and those support operations become harder to maintain profitably as the installed base dwindles. And finally, many vendors have been acquired, left the business or discontinued operation.
• Hardware getting old, support getting thin.
  Even with the best vendor support, reliability will become more problematic as hardware and recording media (and support personnel) age. Even if the vendor maintains the support team, the development teams that they rely on for problem escalation support are long gone.
• Migration becoming the Standard of Practice.
  Most provider organizations are electing to migrate their data to their new systems, enough so that it is becoming a standard of practice. As a result, it is becoming increasingly difficult to defend a "plan" to keep the data on the legacy archive until it fails and cannot be fixed, and then declare the loss an unforeseen casualty.