Computer Stored Ambulatory Record (COSTAR)
In 1968, Octo Barnett and Jerome Grossman of the Laboratory of Computer Science at Massachusetts General Hospital developed COSTAR. Within the first 2 years, the system had records for 20,000 patients resulting from 80,000 visits. COSTAR operates using the MUMPS programming language. MUMPS was widely used in the 1970s and 1980s for medical applications. Both the system and the language are still in use to this current date(2).
COSTAR supports patient registration, scheduling of patient visits, storage, and retrieval of clinical information, and billing accounts receivable.
The first implementation of COSTAR was at Harvard Community Health Plan (HCHP), which adopted and began testing the COSTAR system in 1969 (Grossman et al., 1973). COSTAR was also used in the Navy Occupational Health Information Management System (NOHIMS) as one of two subsystems that provided the medical-information component. In 1986, more than 4.1 million pages of patient records were being printed annually. About 1.3 million patient encounters were documented, more than 2.5 million lines of text were catpured and about 650,000 laboratory test results were entered. By 1987, HCHP had installed 10 minicomputers to support clinical computing in their 9 care facilities, which together serve an active membership of more than 225,000 people.
Characteristics of COSTAR
COSTAR has two characteristics which made it possible to use the system in a variety of sites. The first characteristic is a modular design allowing for extensive reuse of the separate modules at different sites according to each site's needs. The other is its substantial, extensible data dictionary. The core of he data dictionary is a controlled vocabulary of clinical terms, which includes the ability to associate modifier terms with primary name for a clinical concept - such as, synonyms added as modifier terms for a disease or a symptom acting as a clinical concept(2).
Metathesaurus Scope The Metathesaurus representation of COSTAR includes approximately 3,500 terms occurring frequently at three COSTAR implementation sites, which were supplied by MGH
Metathesaurus Update Frequency COSTAR was last updated in the Metathesaurus in 2002.
Some of the major modules in COSTAR and percentage of use:
- Registration (95%)
- Medical Records (100%)
- Accounts Receivable (43%)
- Report Generator (75%)
- Scheduling (52%)
- Medical Query Language (50%)
COSTAR was one of the first programs that used "reminders" that were generated once a patient's profile was used. This aided in proper observance of clinical guidelines by the practitioner (3). It has also had a set of reminders for follow ups for cases of hypertension (3).
COSTAR relied on a structured encounter form and clerical data entry. It also captured free-text entries, primarily to allow specificity of physicians' notes. But the free-text format made summary analysis of medical records impossible. A review of natural-language parsing in 1983 found that, of 233 throat cultures for Beta Hemolytic Strep Group A, 90 were falsely labeled as normal by a natural-language processor that looked for specific strings in the free-text entries.
Researchers in 1983 urged a new system based on a series of menus that limited the data entry in some areas to pre-coded text that could be more accurately analyzed. They noted new developments with graphical interfaces that included "windows" that could be navigated with a new computer accessory, the "mouse."
Take-home from early adoption
Among the early adoptors of COSTAR is Medical Information Management, Inc. (MIM) founded in 1977. Between October 1978 and September 1979, the company sought a stop-gap system to the then, arguable similar, prevailing problems between clinical workflow and the promises of technology advancements. During the time, COSTAR was the only system with full medical record capability and it fit MIM's specifications. There were systems implementation gaps, as one might expect of new systems let alone pioneering endeavors, both in hardware and software components. For example, PDP-11 electrical circuits required improved ventilations. Similarly today, it's not uncommon for implementers to encounter additional out-of-the-box modifications to hardware systems -- additional printers, servers, terminals, etc. -- that may not be obvious during preliminaries.
Digital Equipment Corporation, which later on was acquired by Compaq, now merged into HP, itself had operating system shortcomings in MUMPS that caused the system to crash as a result of increased concurrency, some of which may have been attributable to COSTAR itself being a compeletly new environment altogether. Again, this is typical of what we might encounter nowadays during large scale full-implementation -- one cannot reasonably expect 100% efficiency, optimal delivery in groundbreaking systems implementation. It's commonplace to have additoinal modules developed from scratch, or existing ones modified to fit the new environment. But the takehome from this early implementation is that even mature systems may encounter, in fact should be prepared to come accross, gaps that implementors/vendors and customers should make room for adjustments in the process.
- The database is available for retrospective quality assurance review and to pre-program reminders and alerts within the system for patient follow up and selection of preferred therapies. []
- Detmer, Don E., Dick, Richard S., Steen, Elaine B.. The Computer-Based Patient Record By Institute of Medicine (U.S.). Committee on Improving the Patient Record, Edition 2.
- Hattwick MA. Computer Stored Ambulatory Record (COSTAR) in Real Life Practice. Proc Annu Symp Comput Appl Med Care. 1979 October 17; 761–764.
- Hattwick, Michael A. W., M.D. COmputer-stored Ambulatory Record (COSTAR) In Real-life Practice. Retrieved: September 12, 2010. []
- James R. Campbell. An Ambulatory Information System Serving the Needs of Clinical Practice: COSTAR V. Proc Annu Symp Comput Appl Med Care. 1986 October 26; 141–146. PMCID: PMC2244993
- John McrLatchey, G. Octo Barnett, Geoffrey McDonnell, Judith Piggins,Rita D. Zielstorff, Frances Weidman-Dahl, Ekdward Hoffer, Jon A. Hupp. Proc Annu Symp Comput Appl Med Care. 1983 October 26; 329–332 http://www.pubmedcentral.nih.gov/picrender.fcgi?artid=2578353&blobtype=pdf
- Luo JS. Electronic Medical Records. Primary Psychiatry. 2006; 13(2): pp. 22-30. Available from: http://www.primarypsychiatry.com/aspx/articledetail.aspx?articleid=87
- Navy Occupational Health Information Management System (NOHIMS). Users' reference manual. COSTAR (Computer-Stored Ambulatory Record System). Operators guide. 1987 May 01.