Section 2: Decision Support 2000+

Chapter 3. Decision Support 2000+

A New Information System for Mental Health

Marilyn J. Henderson, M.P.A.,* Sarah L. Minden, M.D.,* and Ronald W. Manderscheid, Ph.D.*

*Center for Mental Health Services, Substance Abuse and Mental Health Services Administration; *Abt Associates, Inc.

Evolving Need for Information

The need for good information has expanded exponentially in the managed care era. Dramatic changes in the roles and types of stakeholders in the mental health care system are taking place. These changes have created a need to expand and improve information and to provide support for decisions made on a daily basis. The quality of information will determine the quality of care: without good data, stakeholders cannot make good decisions and without good decisions, the system cannot continue to operate.

Such information should be available quickly, electronically, and in an easily accessible format. Currently, this situation generally does not prevail in the mental health field because of dramatic under-investment in modern information systems and lack of application of modern information technology to mental health problems.

Such information also should be confidential, protect personal privacy, be available for consumer review and correction, and be used only for medical purposes to improve personal well-being. Currently, this situation generally does not prevail in the mental health field because medical records are fragmented, maintained on paper, transmitted through facsimile machines, sent electronically over the Internet without protection, and available for commercial exploitation.

The mental health field (and, indeed, the human service system as a whole) needs standardized data to manage care effectively. The field also requires measures to evaluate the quality of the care provided, with respect to both practices and outcomes. Clinical and system guidelines exist but are not widely accepted. As a result, they are not used to standardize practice or to provide criteria for judging provider and system performance. Availability of data systems for collecting this information in a uniform and comparable way will enable communication among participants and across systems of care.

Today's technology makes possible a revolution in information: multiple users can participate in what is virtually a single information system that will enable them to share data and communicate effectively. If they adhere to established standards for data collection, this virtual system can be used to meet their information needs, whether they are consumers or providers making choices about treatments, payers deciding among health plans, managers allocating financial and human resources, or researchers determining the need for services in a community. Collection of necessary data can be accomplished while protecting the privacy and confidentiality of personal medical records.

To be useful in the current environment, mental health information needs to span a range from population characteristics through the effects of services. The Survey and Analysis Branch within the Center for Mental Health Services (CMHS) is currently supporting work to develop the framework for such a system. Support and buy-in from all major stakeholders in the system is critical to the success of these projects. To this end, CMHS is working with a broad array of expert consultants from major stakeholder groups, such as mental health consumers, family members, the managed behavioral health care industry, individual service providers, payers, researchers, and experts in mental health electronic records and information technology.

Purpose of Decision Support 2000+

To respond to the mental health field's lack of standardized data, uniform measures, and an accessible and effective information system, the CMHS project team is developing data standards, minimum data recording requirements, procedures, and an information system for mental health. These activities build on what the field has already accomplished, using resources currently in place and focusing on areas that need further work. Decision Support 2000+ is being designed to—

• Improve Decisions. Decisions made by consumers and family members, providers, payers, managers, and researchers will be enhanced by an information system that provides all the data needed quickly, accurately, and efficiently.
• Improve Services. An information system that makes available to stakeholders reliable data on a community's mental health needs, services, service users, costs, revenues, performance, and outcomes is critical to improving care.
• Improve Accountability. To be most beneficial, information on accountability needs to be readily available within the framework of continuous quality improvement.
• Improve Communications. Effective communication within the mental health system and between it and other human service systems is essential for delivering quality care.

A group of experts and stakeholders was convened to guide the development of Decision Support 2000+ and to address the goals identified above. This group recommended that the information system should be able to—

• span the entire mental health system, from epidemiology to service delivery to outcomes;
• link with information systems in a broad range of agencies, locations, programs, and organizations;
• meet the needs of all relevant groups, including consumers, families, providers, payers, managed care organizations, State mental health agencies, administrators, researchers, policy makers, and advocates;
• make use of modern technology while ensuring privacy and confidentiality of data;
• be flexible enough to incorporate information and assessment tools that measure the cultural competence of services; and
• facilitate clinical and organizational decision-making and enhance the quality of care.

Description of Decision Support 2000+

Decision Support 2000+ contains data of four different types: descriptive, prescriptive, evaluative, and corrective. Each type of information has its value for addressing particular types of questions:

Descriptive Information: What are we doing?

Prescriptive Information: What should we be doing?

Evaluative Information: How well are we doing?

Corrective Information: How do we improve?

Figures 1 and 2 illustrate the Decision Support 2000+ model. Figure 1 summarizes the key information modules (see descriptions below) and shows how they can be linked together and transformed to answer a range of critical stakeholder questions. The key information modules are—

• population, plan enrollment, encounters with service providers, and the financial, organizational, and human resource characteristics of clinical and administrative entities within the care system;
• measures that reflect adherence to system and clinical guidelines; and
• results reported through system performance measures, consumer outcome measures, and surveys of consumers, providers, and others.

Figure 2, by contrast, shows how both the mental health care system and Decision Support 2000+ are linked to the care and information systems of other key groups and human service agencies. The stakeholders in the mental health care system provide data for and receive information from Decision Support 2000+. Stakeholder queries can range from questions about plan quality to questions about adherence to practice guidelines.

The information system will record data from various sources that are needed to manage mental health systems effectively. Population data will describe demographic characteristics, medical and mental health status and level of functioning, and quality of life of community members. Enrollment data will describe demographic, insurance, and baseline health and mental health status of enrollees and their family members. Encounter data will characterize all users of services (e. g., health and mental health status, diagnosis, symptoms, functional status), types of services used, and frequency of use. Financial data will reflect costs of services, administrative costs, other expenditures, and revenues. Human resource data will describe the characteristics of providers of care, support staff, and other personnel. Organizational data will reflect information about organizational structure and process.

Clinical guideline data have the potential to serve three primary functions: clinical decision support (selection of the most effective treatments for conditions), treatment process tracking (a detailed and standardized record of clinical interventions), and guideline variance tracking (the congruence between guideline-recommended treatment and actual treatment delivered). While significant progress has been made in establishing the importance of clinical guidelines and their measures, guidelines are currently unavailable for many disorders; there is no consensus on which guidelines are the best; it is recognized that few clinicians have been trained in the use of guidelines; and clinical guidelines software has only recently become available. Implementation of measures for treatment process and guideline variance tracking systems awaits a standard terminology of treatments with associated definitions and codes that can be integrated into routinely used software. Clinical decision support, in turn, depends upon building interfaces with treatment process tracking and consumer characteristics. As we develop this component of Decision Support 2000+, we will involve end-users in the development of guidelines, taxonomies, measures, and software so that they are meaningful, reputable, and user-friendly.

Even though system guideline data are essential for improving the quality of care and efficiency of operations, they are only in the earliest stages of development. They specify measures with respect to infrastructure, executive, and management functions; service components directly operated; and service functions outside of mental health that support clinical programs. Prototypical system guidelines and measures exist in the National Alliance for the Mentally Ill's recently published manual on the Program for Assertive Community Treatment (PACT) (Allness and Knoedler, 1998); in operational manuals prescribing organizational practices (accreditation, credentialing, personnel and financial management, buildings maintenance) and clinical interventions (involuntary commitment, seclusion and restraint); and in the quality improvement tools used by some State mental health agencies for assessing provider and organizational performance. Through the work of the CMHS project team, the area of system guidelines is being defined and clarified for the first time. As minimum data sets are developed, we also will clarify the measurement of system guidelines.

Performance indicator, report card, and consumer outcome data are critical for the accountability, quality improvement, and management of mental health systems. Although the field lacks uniform sets of performance indicators and outcome measures, there is an emerging consensus on the critical components for each, and steady progress toward standardization. Several initiatives are under way to standardize measures and definitions across systems; to develop methodological and implementation guidelines; and to analyze, interpret, and present results in comparable ways.

Key Features of Decision Support 2000+

Decision Support 2000+ has several hallmark features that deserve mention. The first is protection of privacy and confidentiality of personal medical records. The second is evolution of field-wide standards for data recording. The third is reliance upon existing information whenever possible in order to reduce the cost of implementing the new system. The fourth is the linkage of data from different sources using Internet-based query technology. Each of these features is discussed below.

Protecting Privacy and Confidentiality. Decision Support 2000+ is being designed to protect privacy and confidentiality of personal medical records using modern information technology. An overarching concern in conceptualizing this new system was an awareness of the need to specifically address these issues throughout the development and implementation process. In preparing the requirements analysis for Decision Support 2000+ (Minden et al., 2000), a document was commissioned on the issue of privacy from the consumer point of view. This document is available as part of the requirements analysis on www.mhsip.org.

Privacy and confidentiality are of concern to most people. This concern becomes magnified when considering medical records and particularly acute when considering mental health medical records. Stigma, loss of control, exploitation, discrimination, and potential negative consequences all combine to exacerbate these concerns. Such considerations have provided strong motivation for efforts to pass a health care bill of rights empowering the consumer community to gain access to medical records and correct errors in them, and bringing forces together to promote ways that human rights can be preserved and enhanced through better privacy and confidentiality protections.

Any effort to address privacy and confidentiality must start with human values and ethics. In mental health, human rights and dignity are basic values. Hence, these values must provide a foundation for any work undertaken in this area. In recognition of this, CMHS has supported the Workgroup for the Computerization of Behavioral Health and Human Services Records which has designed a virtual medical record for behavioral health care in which the key to the medical record is controlled by the consumer (The Workgroup, 1998). This proposed virtual record is also based upon technology that makes it feasible to protect privacy and to control confidentiality. Decision Support 2000+ will incorporate the fundamental concepts elaborated by the Workgroup.

The U.S. Department of Health and Human Services has recently issued Federal regulations to protect privacy and confidentiality of medical records. A need exists to monitor developments in these regulations with respect to their potential impact on behavioral health care in general, and mental health care in particular. Thus, the regulations ultimately released by the Department will provide another element of the foundation for Decision Support 2000+.

Establishing Standards.Decision Support 2000+ recommends standards for data recording— including minimum data sets, measures and instruments, and procedures for collecting and analyzing data— that permit information reporting at the person, plan, local, State, and national levels. It builds on the work of MHSIP in developing standards for mental health. In the late 1980's, MHSIP created a Task Force to consider existing data standards and recommend revisions. In its 1989 report, Data Standards for Mental Health Decision Support Systems (commonly known as FN– 10), the Task Force presented minimum data sets for patient/client data; event/encounter data; human resources data; financial data; and organization data (Leginski et al., 1989); subsequently, recommendations were made in regard to data elements relevant to children (MHSIP, 1992). Owing to the quality of MHSIP's work, all States now have voluntarily adopted many of these standards. A MHSIP workgroup began the process of updating and refining FN– 10 (MHSIP, 1997); this work is being continued through development of Decision Support 2000+ and elaboration of minimum data sets for each of its components.

Using Existing Data. Decision Support 2000+ makes use of existing information technology and data collection activities, and allows users to bring their current practices closer to the ideal without major overhauls and massive investments. It would obviously be impossible— de novo— to build, implement, and finance Decision Support 2000+. Most components of the system already exist in one form or another. The Federal and some State governments collect population-level data; managed behavioral health care organizations and providers collect enrollment, encounter, and outcome data, use financial and human resource data, and report on performance indicators; and measures are currently being developed for clinical and system guidelines because of the rapid evolution of this field. Certainly, we need to expand and standardize these data collection efforts, but we must not minimize how much exists. The issue is one of improving what we have and reaching consensus on how to do so, not on totally rebuilding.

The same is true for information systems. Clearly, problems exist with incompatibility in hardware and software— systems that cannot talk to one another cannot share information. But the Internet is an untapped resource and advances in data warehouse and object-oriented technologies are enabling us to overcome local differences. Other technical issues, of course, must be resolved: we need unique identifiers before we can link data on persons, programs, or plans from different databases; we need dependable ways to ensure privacy and confidentiality; and we need to be able to collect comparable information in an efficient and affordable way. Again, the issue is one of improvement and consensus, not starting over.

Linking Data. Part of the enormous potential of Decision Support 2000+ lies in its capacity to link data from different sources, both within the mental health system and between mental health and other service systems. By drawing from several different data sets through an Internet-based query system, it is possible to answer key questions ranging from the outcome of a single individual's treatment to projections of service needs and financing requirements for entire populations.

By linking data sets virtually, information about persons can be used to improve the quality of care and to evaluate plans and programs. For example, quality of care could be greatly enhanced through the implementation of a virtual integrated patient record spanning the mental health, health, and human services delivery systems (The Workgroup, 1998). Linking enrollment and encounter data aggregated for all persons served by a plan can be used to show whether standards within a contract have been met, such as requirements to provide mental health services to certain percentages and categories of a State's population. Similarly, linking data from consumer satisfaction surveys and other performance measures with aggregated enrollment and encounter data can show the relationship between such factors as satisfaction, availability of specialists, denials of services, and rates of plan enrollment and disenrollment. Linking data virtually will meet many needs in mental health, including:

• The need to coordinate care more efficiently and effectively. A primary barrier to effective and efficient delivery of mental health and human services is the lack of a coordinated communication system that would allow for the sharing of timely, accurate, and appropriate information among all the agencies and service systems involved in care.
• The need to meet reporting requirements. Most mental health organizations are accountable to public or private payers and are required to report routinely. Exchange of core data sets, agreement on data exchange protocols, and use of Web-based Internet and intranet applications would increase the efficiency and cost-effectiveness of data collection and reporting.
• The need for research. Mental health phenomena at both the personal and the service levels are enormously complex. Our ability to understand current circumstances and predict future trends depends on being able to look at how the many factors affect outcomes and performance; this, in turn, depends on being able to link data.

Many challenges exist to linking the components of an information system and linking that system to others. These challenges include creation of privacy-protected unique client and provider identifiers, linking structurally different databases, and collecting and reporting real-time data. When linking data sets, it is critical that data elements and coding be clearly specified to avoid misunderstanding and unwanted variation in coding items. Data collection procedures and databases that serve multiple purposes, such as reimbursement and quality measurement, are more likely to be adopted by users than more limited ones; but they also increase the need for instruments that are straight-forward and transparent, and that minimize additional staff training and development of training materials and documentation.

Status and Next Steps

With guidance from a Technical Expert Work-group, the CMHS project team has completed the requirements analysis for Decision Support 2000+. For each component, this analysis describes the field's achievements and remaining work, in terms of the degree of consensus that exists on domains (issues, categories, or topics of interest), indicators (measurable activities, events, characteristics, or items that represent a domain), and measures (the instruments used to assess, evaluate, and reflect an indicator); whether the measures have been field tested and/ or implemented; and whether the component is fully ready for inclusion in the information system. The components are at different levels of development. For the enrollment and encounter components, for example, there is fairly broad consensus on what to include within the information system, but problems such as specifying unique identifiers remain unresolved. For other components, particularly population, financial, and guideline data, much work remains to be done.

As noted earlier, the complete requirements analysis is posted on the MHSIP website (www.mhsip.org) for broad review and comment by the field. For those who do not have time to review the entire requirements analysis, brief summaries for each component are available on the website.

Currently, no typology is available for organizational and financial arrangements under managed behavioral health care. The team will address this critical gap in our knowledge base and will assess the extent to which the requirements analysis fits each of the major arrangements identified within the typology. This analysis will ensure that Decision Support 2000+, as it is refined, is appropriate for and relevant to the needs of evolving organizational and financial arrangements.

Once the typology is available, the project team will move on to the next phase. Over the next two years, groups of experts will be convened to address outstanding issues such as creating unique identifiers, selecting key performance indicators, and recommending uniform outcome measures. They will also develop core minimum data sets for recommendation to the field. While users should collect any data that meet their particular needs, widespread use of the minimum data sets will provide the field with uniform and comparable data to facilitate com-munication and decisionmaking.

Conclusion

Decision Support 2000+ is an integrated, public-health-oriented information system that is fully congruent with the current and future information needs of the mental health field. Implementation of this information system will facilitate the availability of comparable data to the field for decision support for planning, service design, clinical feedback, and evaluation. Widespread use of the information system will be of tremendous benefit to the entire mental health community.

References

Allness, D. J., & Knoedler, W. H. (1998). The PACT model of community-based treatment for persons with severe and persistent mental illness: A manual for PACT startup. Arlington, VA: NAMI.

Leginski, W. A., Croze, C., Driggers, J., Dumpman, S., Geertson, D., Kamis-Gould, E., Namerow, M. J., Patton, R. E., Wilson, N. Z., & Wurster, C. R. (1989). Data standards for mental health decision support systems. National Institute of Mental Health Series FN No. 10. DHHS Pub. No. (ADM) 89–1589. Washington, DC: Government Printing Office.

MHSIP Ad Hoc Group. (1997). The handbook of mental health data: A reference manual for anyone who wants to collect, find, report, understand or use mental health data. Rockville, MD: Center for Mental Health Services. Manuscript in preparation.

MHSIP Task Force on Enhancing MHSIP to Meet the Needs of Children. (1992). Enhancing MHSIP to meet Decision Support 2000+ A New Information System for Mental Health the needs of children: Final report. Rockville, MD: Center for Mental Health Services.

Minden, S. L. (Ed.), Davis, S., Ganju, V., Guidera, S., Hale, C., Hernandez, M., Kaufman, C., Mazade, N., Noonan, D., Rich, T., Rosenthal, M., Trabin, T., Van Tosh, L., & Webman, D. (2000). Draft requirements analysis for Decision Support 2000+. www.mhsip.org.

The Workgroup for the Computerization of Behavioral Health and Human Services Records. (1998). The virtual consumer and family-focused behavioral health and human services record. Cambridge, MA: The Workgroup for the Computerization of Behavioral Health and Human Services Records, Inc.

Index | Previous | Next