CMH data is currently positioned behind firewalls and tightly controlled. Access is restricted to insure compliance with regulations and to protect the data. There are presently few personnel exclusively designated as data managers at CMH. Epic Systems software was developed beginning over 20 years ago before the development of databases now widely in use such as Oracle, SQL Server. Epic uses a Caché database. The Epic data is expored to its Clarity Enterprise Reporting system. The Clarity data dictionary describes the tables, their fields, and relationships between tables. Data is also exported to the PFF billing service, Springfield Service Corportation (SSC). However, SSC enters the CMH Medical Record Number into only about 20% of its transactions. Thus, it is not possible to relate most bills to specifice services described in CMH data. Similarly, the PTTF heard from several administrators about the presence of at least six databases for faculty and staff. These are maintained at individual administrative sites.
Clarity Reporting utilizes Crystal Reports, a standard software produce, to extract data. These data can be displayed in reports or exported to data tables where they can be further processed and analyzed offline. Many of the clinical management reports are generated by this means. David A Stumpf, MD, PhD
CMH Clinical Data Repository (CDR)
Children's Memorial Hospital is in the process of establishing a CDR that will eventually house data from laboratory testing, medical imaging, outpatient visit, and inpatient charting. Currently there is full deployment of centralized data for laboratory testing and medical imaging. The use of an ambulatory electronic health record has been begun, and plans are being established to extend its use throughout all Children's clinics. Extending the data repository to the other domains of data is also being planned. Commercial vendors are used for the current data residing in the beginnings of the clinical data repository.
The size of operations at Children's Memorial Hospital does not permit the use of programmers who can code viable applications that can successfully be maintained (and periodically upgraded). Such attempts at other hospitals to use the approach of in house coders have almost uniformly been unsuccessful. Furthremore, such an effort restricted to clinical applications would not address research needs, for reasons outlined below. Therefore, vendors have been selected by Children's Memorial Hospital, following the strategic plan, whenever possible, of selecting a core vendor, to enable integration of data transfer. Therefore, Epic Systems has provided the applications for the ambulatory EMR, and associated functions, such as visit scheduling. Because Epic does not provide applications for digital medical imaging, the Picture Archiving Computer System (PACS) that provides digital medical images and reports to the repository has been provided by GE. Similarly, other vendors are being considered for digital imaging of Cardiology procedures. Epic provides an inpatient EMR, which has been deployed in relatively few hospitals. Therefore, the current strategy is to continue use of CareVue, an application provided by Phillips, for inpatient nursing charting. There is a more modest function in CareVue for physician charting. Eventually, there will be integration of data into the CDR EpicCare outpatient EMR with inpatient data, probably either from CareVue, as that application evolves, or the EpicCare inpatient EMR. (Epic Systems and Phillips have recently begun mutual planning around their products, that will likely have implications for product selection for inpatient EMR charting). Michael Miller, MD
Medicine is already benefiting from these revolutions in managing archived medicical content. We routinely search the medical literature and pharmacologic databases; locate educational materials; and identify resources for managing problems. These external systems will continue to evolve with the Internet. But we have not yet implemented these systems in our clinical data. The introduction of Bayesian and other systems promises to provide a Google-like extraction of key information from routine clinical information.
Email Communications
Spam now consumed a high percentage of email resources. It is wasteful and costly. Most email software now includes decision support components, usually using Bayesian logic, to filter messages into "junk" and non-junk. These systems are now able to filter 90% of undesirable messages. These systems must be dynamic in order to keep up with innovative spammers. Bill Gates argues that this will not be successful in the long term. The is promoting the use of "stamps" which will impose a minimal cost to routine email users, but a prohibitive cost on spammers.
Physicians are inundated with information. Much of it is junk. Much of it is clinical data that is normal; it needs to be available during encounters, but requires no action between encounters. Advanced decision support systems can filter this information, focussing attention where it is required and enhancing the role of the physician in addressing important results. There are lessions from the Internet about how to extract this information in useful form.