The National Data Infrastructure Improvement Consortium (NDIIC) advocates for government-funded development of electronic health record (EHR) software, often provided free of charge to providers. This open source software is intended to reduce providers' costs while increasing interoperability.
These are definitely noble goals. However, in reviewing similar EHR initiatives, the results are often just the opposite. As the executive director of Austin/Travis County Mental Health Mental Retardation Center (ATCMHMR), David witnessed a significant rise in administrative and clinical costs when such systems were required of his agency. An informal poll of an Anasazi Software users group found similar results: When providers were asked to name their most significant problem with state/federal requirements, they overwhelmingly cited the lack of interoperability with state reporting applications.
EHR systems developed with state and federal funding, such as the Web Infrastructure for Treatment Services (WITS) and Texas' Behavioral Health Integrated Provider System (BHIPS), were created with the goal of significantly reducing costs to providers. Such systems might save money if they were the only EHR system of record for providers within a state. But this one-size-fits-all concept doesn't account for providers' widely diverse needs.
Community centers and counties, which often provide the majority of behavioral healthcare services, have run into several problems with this type of “free” software. It is not robust enough to support their required EHR functionality nor is it comprehensive enough to support their expansive clinical and administrative requirements. Also, they usually are not interoperable with a provider's existing EHR software.
In Texas, the community centers are responsible for mental health, developmental disability, and substance abuse treatment, as well as early childhood intervention, and often contract to provide related services. As a result, these providers require an EHR system that comprehensively addresses all aspects of treatment, contracts, revenue, costs, and reporting for up to 80 purchasers of services contracts. That's why many community centers, including ATCMHMR, maintain an EHR system that can support all the needs for all treatment environments. They use state-provided software, such as BHIPS, but only to enter information contractually required.
Since BHIPS offers no interoperability support, some community centers find it expensive to perform double entry of the information. Using two EHR systems requires twice the training and support for clinicians to accurately record and review clinical records. The increase in clinical and administrative costs can be mitigated somewhat by hiring data-entry personnel, but this is only a partial solution for an unnecessary problem.
Another complication is the additional EHR information required by varying state, county, and city agencies from providers. Texas community centers can have as many as six of these competing systems, including BHIPS. In other states, providers report being required to record information in up to ten additional state, county, and city systems. Some of these applications offer proprietary interoperability support, while most only support double data entry into Web-based applications. Providers also are required to manage the clinical and administrative contract reporting requirements of many other purchasers of services. These complex needs call for a much more robust solution than what any current state-provided system offers.
The additional cost for duplicate data entry is considerable. ATCMHMR budgets more than $100,000 a year for staff exclusively devoted to data reentry and electronic data submission for state systems. This cost could be dramatically reduced if all state applications supported a common interoperability standard.
Currently, the interoperability of state-provided systems is inefficient and sets up a veritable Babel of conflicting interoperability requirements. They employ uniquely defined file transfer methods, different definitions of the events precipitating reporting, and contradictory definitions for data elements reported for each event. For example, the One Family state reporting system for mental health and substance abuse treatment in Florida has a single race code for “Asian,” but the CSI mental health state reporting system in California has 14 distinct Asian race codes.
A consumer's clinical record could potentially be communicated from California to Florida, but there would be problems. For example, if the person was Samoan, he would be recorded as “Asian” in Florida, but his race information would not seamlessly transfer to the California system because that state's reporting requirements do not allow for “Asian.” In California a person treated within the same county for both mental health and substance abuse would face similar problems. He may be coded as Hmong in the mental health CSI system but as “Other Asian” in California's CalOMS substance abuse state reporting system.
In fact, CalOMS has 20 definitions of possible substances to abuse, whereas Illinois' DARTS substance abuse state reporting system has 28. DARTS defines 9 possible reasons for discharge from substance abuse treatment; Florida's One Family system has 11 definitions; and California's CSI system has 13. Given this level of anarchy in data definitions, plus the fact that every state reporting system supports different file structures and formats, it is little wonder that many state-reporting and state-provided EHR systems do not support interoperability.