Difference between revisions of "Regenstrief Medical Record System (RMRS)"

From Clinfowiki
Jump to: navigation, search
m (→‎Nascence: 1972–1990)
 
(48 intermediate revisions by 10 users not shown)
Line 1: Line 1:
−
Orginally developed in 1972 by Clem Mcdonald, Regenstrief Medical Record System (RMRS) is a complex system that involves manipulation, transfer, and storage of data in and associated with electronic medical records. It is the first EMR to generate reminders to physicians about its own content. Using an internally-developed decision support language - CARE, providers could develop protocol-specific reminders to streamline clinical decision making. One study found that physician compliance with preventative care rose to 51% from 22% when RMRS suggestions were used. Physicians did not seem to have learned from the suggestions, because compliance dropped back down to baseline when the suggestions from RMRS was removed. This finding is particularly important because it indicates that physician knowledge is less important for patient care than providing the physician with the appropriate intellectual artifacts.
+
The '''Regenstrief Medical Record System (RMRS)''' is a comprehensive homegrown data model for [[EHR|electronic health records (EHR)]] developed by the Regenstrief Institute of Indiana University<ref name=mcdonald-1999>McDonald CJ, Overhage JM, Tierney WM, et al. The Regenstrief Medical Record System: a quarter century experience. Int J Med Inform. 1999 Jun;54(3):225-53. doi: 10.1016/s1386-5056(99)00009-x.</ref><ref name=colicchio-2020>Colicchio TK, Cimino JJ. Twilighted Homegrown Systems: The Experience of Six Traditional Electronic Health Record Developers in the Post-Meaningful Use Era. Appl Clin Inform. 2020 Mar;11(2):356-365. doi: 10.1055/s-0040-1710310. Epub 2020 May 20.</ref><ref name=apathy-2017>Apathy N, Vest J, Menachemi N, Harle C. Changes in User Perceptions of Their Electronic Health Record Following the Replacement of a Legacy System. AcademyHealth Annual Research Meeting. 2017. Retrieved 2021 Oct 24. https://academyhealth.confex.com/academyhealth/2017arm/meetingapp.cgi/Paper/16422</ref>, as well as an associated database-management system (DBMS), communications system, and a suite of clinician-facing client-side applications called the '''Medical Gopher'''<ref name=mcdonald-1984-gopher>McDonald CJ. The medical gopher—a microcomputer based physician work station. Proc Annu Symp Comput Appl Med Care 1984:453–459.</ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>. The system pioneered [[Indiana Health Information Exchange]]—one of the first [[HIE|health-information exchanges]]<ref name=colicchio-2020></ref>—as well as one of the first computer-assisted order-entry workstations<ref name=mcdonald-1984-gopher></ref><ref name=colicchio-2020></ref>.
  
−
The goal was to:
+
Between 1972<ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref><ref name=colicchio-2020></ref> and 2016<ref name=apathy-2017></ref>, RMRS stored and managed patient records across multiple institutions of the greater Indianapolis area; at its peak, it contained more than 200 million coded observations and 3.25 million narrative reports, for more than 1.5 million patients, across at least fifteen independent healthcare organizations<ref name=mcdonald-1999></ref>. However, its usage was largely retired in 2016<ref name=apathy-2017></ref><ref name=colicchio-2020></ref>. It is succeeded today by the Indiana Network for Patient Care (INPC), the [[Indiana Health Information Exchange]], the Regenstrief Teaching EMR, and [[Arden Syntax|Arden syntax]]—all of which were originally based upon the RMRS data model.
  
−
  • eliminate the logistic problems of the paper record by making clinical data immediately available to authorized users wherever they are—no more unavailable orundecipherable clinical records;
+
== Nascence: 1972–1990 ==
  
−
  2. to reduce the work of clinical book keeping required to manage patients—no more missed diagnoses when laboratory evidence shouts its existence,no more forgetting about required preventive care;  
+
1972 was a year from the age of the 16-bit minicomputer<ref name=16-bit-microcomputer>Bob Supnik. Simulators: Virtual Machines of the Past (and Future). 2004. ACM Queue. 2(5).</ref>. That year, Clement McDonald and Charles Clark, two physicians at Sidney & Lois Eskenazi Hospital (then called Marion County General Hospital<ref name=mcdonald-1999></ref><ref name=eskenazi-history>Eskenazi Health. History. 2020. Retrieved 2021 Oct 24. https://www.eskenazihealth.edu/about/history</ref>) used a PDP-11 minicomputer<ref name=mcdonald-1999></ref><ref name=16-bit-microcomputer></ref> to create a computer-stored medical record for 35 patients with diabetes, with the intent of automating diagnosis and management after manually capturing patient data, using a hardcoded but generic data structure<ref name=mcdonald-1973>McDonald CJ. Computer applications to ambulatory care. Proceedings of the IEEE Conference on Systems, Man, and Cybernetics; Boston, Massachusetts, November 5–6, 1973.</ref><ref name=mcdonald-1974-tree>McDonald CJ, Bhargava B. Tree systems for medical information processing. Proceedings of the Eleventh Annual Rocky Mountain Bioengineering Symposium, USAF Academy, Colorado, March 1974.</ref><ref name=mcdonald-1999></ref>. Their goals included the following<ref name=mcdonald-1981>McDonald CJ. Action-Oriented Decisions in Ambulatory Medicine. Yearbook Medical Publishers, Chicago, 1981.</ref>:
  
−
  3. to make the informational ‘gold’ in the medical record accessible to clinical, epidemiologic, outcomes and management research. The system was to complement, not replace, the paper medical record.
+
# Eliminating logistic problems of paper records, such as availability and indecipherability.
 +
# Reducing the work of clinical bookkeeping that is incidental and nonessential to patient management: “no more missed diagnoses when laboratory evidence shouts its existence, no more forgetting about required preventive care”.
 +
# Making informational “gold” in medical records available to clinical, epidemiological, and managerial research.
  
−
It began in 1972, in a diabetes clinic with only 35 patients. As of 1999, the RMRS carried 200 million separate coded observations,3.25 million narrative reports, 15 million prescriptions
+
After encountering difficulty with capturing comprehensive data for the pilot population of 35 patients at their diabetes clinic, McDonald, Clark, and other colleagues would build several systems for capturing data at laboratory and pharmaceutical sources<ref name=mcdonald-1974-lab>McDonald CJ, Martin G. A model lab information system. Proceedings of the Fifth Annual Pittsburgh Conference on Modeling and Simulation, April 1974.</ref><ref name=carlstedt-1977>Carlstedt B, Jeris DW, Kramer W, Griefenhage R, McDonald CJ. A computer-based pharmacy system for ambulatory patient care. The Indiana Pharmacist 1977; 58:92–98.</ref><ref name=mcdonald-1999></ref>.
−
and 212,000 electrocardiographic (EKG) tracings. The RMRS carried records for 1.4 million patients in addition to all data generated from several thousand ambulatory and inpatient encounters per year,
+
  
−
As of now, the RMRS is still operating, which carries 660 million distinct observation. This makes RMRS one of the longest operating EMRs in the world. The RMRS is being accessed more than 10 million times every year by Wishard Health Services and 20 million times per year at Clarian Health in Indianapolis.
+
=== Data model and RDBMS ===
  
−
The RMRS has been best studies and it is well known nationally and internationally to be the model for a number of commercial and academic EMR systems.
+
Over 1972–1974, it was found that maintaining a hard-coded data model was not sustainable, due to having to change multiple programs whenever any change was made to their data model’s file structure; the decision was therefore made to create a [[databases|relational-database management system (RDBMS)]], due to no commercial DBMS being available for the PDP-11 microcomputer being used<ref name=mcdonald-1975>McDonald CJ, Bhargava B, Jeris DW. A clinical information system (CIS) for ambulatory care. Proceedings of the AFIPS National Computer Conference, Anaheim, California 1975.</ref>.
  
−
This system serves four hospitals on the Indiana University Medical Center campus and forty outreach practices in the city of Indianapolis.
+
RDBMS, by 1989, ran on several VAX superminicomputer workstations<ref name=mcdonald-1989>McDonald CJ, Day Z, Martin DK, Tierney WM. The Regenstrief medical record: 1989, a campus-wide system. Proceedings of the Symposium on Computer Applications in Medical Care, Washington, DC 1989.</ref> and stored the RMRS relational data model in several large files. Records were physically sorted by patient ID, observation ID, and date/time (reverse chronologically). A single term dictionary controlled the system’s concept vocabulary, including all diagnoses, findings, variables (e.g., Glasgow coma score, diastolic blood pressure), and sets of terms. Questionnaires were defined as structured lists of questions and variables stored in the RMRS; these questionnaires were flexible and accepted codes from many different code systems (e.g., [[ICD-9]], [[CPT|CPT-4]], local Regenstrief codes) that were associated with each variable’s term. The physical RDBMS and its data model resembled the following:<ref name=mcdonald-1988>McDonald CJ, Blevins L, Tierney WM, Martin DK. The Regenstrief Medical Records. MD Comput 1988; 5:34–47.</ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>
  
−
The systems success can be attributed to it's strong foundation in 3 areas.  
+
* Dictionary/master files containing:
−
<nowiki>
+
** Terms, each with a name, synonyms, external references (e.g., ICD-9 and SNOMED), etc.
−
1. Physician leadership in the informatics effort is vital. It is their intelligence, self-confidence, high energy, and clinical knowledge.
+
** Structured-observation types, e.g., “microbiology”.
−
2. Commitment to the mission and vision of high quality and excellence in health care driven and believed in from the highest echelon within the hospital, and
+
** Numeric-observation types, each with a scale factor, reference ranges, etc.
−
3. Continuous quality improvement and incorporation of user feedback to guide this improvement.</nowiki>
+
** Coded-observation types, each with ordinals, nominals, etc.
−
  Quickly fixed mistakes are tolerated by Physicians, and this had driven an evolutionary approach. The software is updated incrementally and feedback from the users is sought early and often. Feature's are added based upon which is easiest to fix or implement.
+
** Doctors, each with a name, a license, a speciality, etc.
 +
** Relationships between these relations:
 +
*** Structured-observation types, numeric-observation types, and coded-observation types were subclasses of terms.
 +
*** Each term could be an aggregate of at least two structured-observation types as its “elements”.
 +
*** Each term could have any number of coded-observation types as its value set.
 +
*** Each term could have one numeric-observation type as its unit.
 +
* Operational data containing:
 +
** Encounters, each with a date, billing data, etc.
 +
** Orders, each with a date/time, a priority, notes, etc.
 +
** Patients, each with a name, address, etc.
 +
** Medical-record items, each with a date/time, etc.
 +
** Narrative reports, each with a header, text, etc.
 +
** Numeric observations, each with a value, etc.
 +
** Binary data, e.g., images, voice recordings, electrocardiograms, etc.
 +
** Coded observations.
 +
** Relationships between these relations:
 +
*** Each encounter was associated with one patient.
 +
*** Each order was associated with one encounter.
 +
*** Each medical-record item was the result of one order.
 +
*** Each medical-record item was associated with one patient.
 +
*** Each medical-record item had one term as its parameter.
 +
*** Narrative reports, numeric observations, binary data, and coded observations were subclasses of medical-record items.
 +
*** Each coded observation could have any number of terms for its value(s).
  
−
The data repository is the key component of the system. It originally called for no data entry for physicians' observations. Creator Clement McDonald, M.D. and his co-authors note that physicians were reluctant to perform data entry. The system developers first enlisted physicians for data entry with a physician orders system because orders are more easily structured for data input than observations.
+
=== Data capture, research, and display ===
  
−
Significant effort is required to create mechanisms to capture clinical data.  Universal or standardized codes such as LOINC and SNOMED through HL7 interfaces have eased the strain. 
+
By 1984, comprehensive multisite data capture by RMRS had improved, and new data standards were being developed to reduce the necessity of ancillary service systems<ref name=mcdonald-1984-lab-exchange>McDonald CJ, Wiederhold G, Simborg DW. A discussion of the draft proposal for data exchange standards for clinical laboratory results. Proceedings of Symposium on Computer Applications in Medical Care 1984; 406413.</ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>. And even as early as the 1980s, several retrospective clinical studies would be performed using the RMRS dataset<ref name=mcdonald-1986>McDonald CJ, Hui SL, Tierney WM. Diuretic-induced laboratory abnormalities that predict ventricular ectopy. J Chron Dis 1986; 39:127–135.</ref><ref name=psaty-1987>Psaty BM, Tierney WM, Martin DK, McDonald CJ. The value of serum iron studies as a test for iron-deficiency anemia in a county hospital. J Gen Intern Med 1987; 2:160–7.</ref><ref name=tierney-1989-predict>Tierney WM, Martin DK, Hui SL, McDonald CJ. Using clinical data to predict abnormal serum electrolytes and blood cell profiles. J Gen Intern Med 1989; 4:375–383.</ref><ref name=tierney-1989-renal>Tierney WM, McDonald CJ, Luft FC. Renal disease in hypertensive adults: Effect of race and type-II diabetes mellitus. Am J Kid Dis 1989; XIII:485–493.</ref>.
−
What data to capture is an important decision.  Multiple systems, types of data, formats, and the large number and variety of sources require a great deal of effort and thus the need for planning and prioritizing.
+
  
−
Direct capture by electronic interface based on HL7 can be facilitated from beside electronic instruments, patient registration system, laboratory, pharmacy, appointment scheduling, dictation/transcription radiology, nurse telephone triage and billing systems.
+
In addition, [[CPOE|computerized physician order entry (CPOE)]] in the outpatient setting was initiated in 1984, which was later extended to the inpatient setting in 1990. These CPOE systems would evolve into elaborate “Medical Gopher” client-side workstations.<ref name=mcdonald-1992>McDonald CJ, Tierney WM, Overhage JM. H.I.S. and the physician: Direct inpatient order entry by physicians through medical gopher workstationsproblems and promises. Proceedings of the IMIA Working Conference on Trends in Modern Hospital Information Systems. (Gottingen, Germany, September 7–11, 1991). North-Holland: Elsevier Science Publishers B.V., 1992 IMIA.</ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref> Although RMRS’s developers noted initial physician unhappiness in reaction to computerized order entry, this unhappiness would later improve as the Gopher system would improve<ref name=mcdonald-1999></ref>.
  
−
Clerk data entry is used to code the impressions of most diagnostic reports not already coded by the system. This allows the system can understand the diagnostic content for patient retrievals and reminders. They enter standardized phrases and abbreviations instead of numeric codes that would add to the training. The data entry clerks also enter the number values of predefined questions, such as blood pressure or finger stick glucose.
+
Even at these early stages, RMRS generated much research about [[CDS|clinical decision support]] in the outpatient setting, especially regarding rule-based computer reminders<ref name=mcdonald-1976>McDonald CJ. Use of a computer to detect and respond to clinical events: its effect on clinical behavior. Ann Intern Med 1976; 84:162–167.
 +
McDonald J. Computer reminders, the quality of care and the nonperfectability of man. N Engl J Med 1976; 295: 1351–1355.</ref><ref name=prokosch-1995>Prokosch HU, McDonald CJ. The effect of computer reminders on the quality of care and resource use. Chapter in hospital information systems; design and development characteristics; impact and future architecture. Edited by Prokosch HU, Dudeck J. 1995; III(13):221–239. Elsevier Publishing Co. Amsterdam.</ref><ref name=mcdonald-1984-introspective>McDonald CJ, Hui SL, Smith DM, Tierney WM, Cohen SJ, Weinberger M. Reminders to physicians from an introspective computer medical record. Annals of Internal Medicine 1984; 27, 100: 130–138.</ref>. Furthermore, in the 1980s, RMRS’s reminder-rule systems and [[HELP|Health Evaluation through Logical Processing (HELP)]] would give rise in 1989 to [[Arden Syntax|Arden syntax]], a markup language for representing shared medical knowledge, which would later become an [[HL7]] standard in 1999 and an American National Standards Institute (ANSI) standard in 2002.<ref name=hripcsak-2018>Hripcsak G, Wigertz OB, Clayton P. Origins of the Arden Syntax. Artif Intell Med. 2018 Nov;92:7–9. doi: 10.1016/j.artmed.2015.05.006. Epub 2015 Jul 2.</ref>.
  
−
Physician workstation entry eliminates delays, costs and potential errors due to transcription.  Direct entry by physicians also validates the person who is most knowledgeable of the information and is in the best position to act on “smart” computer feedback.
+
== Maturity: 1990–2000 ==
  
−
Today, most nursing home orders, as well as ED release and inpatient orders are entered directly into the computer system by physicians using the Medical Gopher CPOE workstation. Dictated discharge summaries have largely been replaced by physician-generated discharge notes entered directly into the computer. Similarly, most of the outpatient clinic notes are entered directly into the computer using the Medical Gopher system.
+
By the 1990s, RMRS continued to grow and RMRS was gradually expanded from the outpatient clinics to inpatient settings. It would facilitate the care of more than one million patients while generating dozens of studies. At its peak circa 1999, it contained more than 200 million coded observations, 3.25 million narrative reports, 15 million prescriptions, and 212,000 electrocardiograms, for more than 1.5 million patients—across at least fifteen independent healthcare organizations, with access by more than 2500 employees.<ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>
  
−
The Gopher workstation also allows the printing of personalized patient information handouts, permits doctors and nurses to communicate via confidential email and can even display satellite weather photos. Physicians can also access past issues of leading medical journals and the American Hospital Formulary drug monographs to research specific topics or learn more about a certain medication.
+
=== Core system: three-legged stool ===
  
−
A 2001 study from Wishard Memorial Hospital demonstrated that computerized reminders significantly improved the use of preventative measures in eligible patients admitted to the hospital. Compared to a control group, those patients whose physicians received reminders received more influenza and pneumococcal vaccinations, subcutaneous heparin prophylaxis and aspirin at the time of discharge. However, even with the reminders, compliance with the recommendations was still far from being universally accepted by the physicians.
+
RMRS’s flagship institutions continued to be Sidney & Lois Eskenazi Hospital (which by this period had been renamed from Marion County General Hospital to Wishard Hospital)<ref name=eskenazi-history></ref>, as well as the three hospitals of Indiana University Health (formerly Clarian Health)<ref name=iu-2010> Indiana University Media Relations. Clarian Health to become Indiana University Health in spring 2011. 2010 May 5. Retrieved 2021 Oct 24. https://newsinfo.iu.edu/news/page/normal/14377.html</ref>—together with Regenstrief, these together were characterized as RMRS’s “three-legged stool”<ref name=mcdonald-1997>McDonald CJ, Tierney WM, Overhage JM, Dexter P, Takesue BY, Abernathy G. The three-legged stool: Regenstrief Institute for Health Care. Proc 3rd Annual N.E. Davis CPR Recognition Symp. 1997; 101–147.</ref>.
  
−
One of the keys behind the system is that the numerous forms and reports are not pre-generated.  Instead, they have a generalized template ("schema") which deals with high-level components like "notes in this section" and "vitals in this section". And then based upon the data for a patient, the display is materialized.
+
Its homegrown central [[databases|relational-database management system (RDBMS)]] ran on DEC Alpha servers with an ancillary Novell personal-microcomputer network, both hosted by Eskenazi Hospital.<ref name=mcdonald-1999></ref> By 1999, RMRS also managed the medical records of all outreach/homeless clinics in the Indianapolis area, as well as mental-health clinics. These would connect with the central Eskenazi hub system, via ethernet cables, Transmission System 1 (T1) lines, and dedicated telephone lines, largely with [[HL7]] messages over Internet Protocol (IP).<ref name=mcdonald-1999></ref>
  
−
The general layout of the system is with the problem list on the upper left side with a physician specified and chosen set of observation variables like vitals on the lower left portion of the screen.  On the right side is a section for notes and below that is where orders are entered and reviewed. As far as reports, a great deal of them can be created and customized to the encounter.  Such reports can be as wordy or succinct as the doctor desires and ready prior to the patient encounter.
+
In this core system, RMRS captured clinical data from more than two dozen disparate systems, such as direct capture from bedside measuring devices, direct links to electrocardiography carts, direct links to financial systems, regular technician entry from public health records, nurse radio/computerized dictation and voice recording, clerk-entered questionnaire/form results, and computerized physician entry via “Medical Gopher” client-side microcomputer workstations.<ref name=mcdonald-1992></ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref> Data standards used as of 1999 included [[HL7]] messages, [[DICOM]], NCPDP, [[LOINC]], [[ICD-9]], [[CPT|CPT-4]], and ACR radiology codes; [[UMLS]] and [[SNOMED]] were being explored but not supported at this time<ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>.
  
 +
Physician order entry and clinical-data charts were initially displayed in text-based interfaces from “Medical Gopher” computer terminals. Eventually, graphical interfaces in web pages were also supported; hard copies were also printed on paper—in general, these were designed to yield specialty-specific, single-page, information-dense overviews of each patient, akin to hospital “scut” cards.<ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>
  
−
==References==
+
Informational feedback such as [[alerts]], e.g., for contraindicated medications, occurred during order entry; research was extended from the outpatient setting<ref name=mcphee-1991>McPhee SJ, Bird JA, Fordham D, Rodnick JE, Osborn EH. Promoting cancer prevention activities by primary care physicians: results of a randomized, controlled trial. JAMA 1991; 66:538–544.</ref> to the inpatient setting during the 1990s<ref name=rind-1994>Rind DM, Safran C, Phillips RS, Wang Q, et al. Effect of computer-based alerts on the treatment and outcomes of hospitalized patients. Arch Intern Med 1994; 154:1511–1517.</ref><ref name=pestotnik-1996>Pestotnik SL, Classen DC, Evans RS, Burke JP. Implementing antibiotic practice guidelines through computer-assisted decision support: clinical and financial outcomes. Ann Intern Med 1996; 124:884–890.</ref><ref name=overhage-1996-preventive>Overhage JM, Tierney WM, McDonald CJ. Computer reminders to implement preventive care guidelines for hospital inpatients. Arch Internal Medicine 1996; 156:1551–1556.</ref><ref name=overhage-1997>Overhage JM, Tierney WM, Zhou XH, McDonald CJ. A Randomized Trial of “Corollary Orders” to Prevent Errors of Omission. JAMIA 1997;4:364–375.</ref>. Order suggestions were also given depending on rules expressed in a pair of the homegrown rule languages, CARE and G-CARE; these included including “corollary orders”, as well as formulary restrictions from several health-maintenance organizations (HMOs)<ref name=overhage-1996-g-care>Overhage JM, Tierney WM, Mamlin B, Warvel JA, Warvel JS, McDonald CJ. A tool for provider interaction during patient care: G-CARE. JAMIA 1996; 19th SCAMC Proceedings: 178–182.</ref><ref name=overhage-1997></ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>; for example, entering DVT in a problem list would trigger order sets for heparin and warfarin<ref name=mcdonald-1999></ref><ref name=mamlin-2007>Mamlin BW, Overhage JM, Tierney W, Dexter P, McDonald CJ. Clinical decision support within the Regenstrief Medical Record System. Clinical decision support systems: Theory and practice, Second edition. 2007: 190–214.</ref>. Physicians, however, were “never blocked from completing an order” but were rather made “aware of the potential problems that an order may cause” <ref name=mcdonald-1999></ref>.
−
1. The Three-Legged Stool: Regenstrief Institute for Health Care. Clement McDonald, M.D.
+
  
−
2. [http://www.medicine.indiana.edu/news_releases/archive_00/regenstrief_mrs_00.html Indiana School of Medicine]
+
=== Statewide health-information exchange ===
  
−
3. [http://www.regenstrief.org/medinformatics/rmrs RMRS Introduction]
+
Outside of the four-hospital core system, the RMRS data model was further developed during this period to span multiple healthcare systems: separate medical repositories (data “vaults”) were created for each healthcare systems, but these vaults would share a single standardized term dictionary and a “global patient index”. A personal-name matching algorithm would determine which patient records in different healthcare systems referred to the same patient, assigning them the same “global patient ID”. This “cross-enterprise” normalization occurred at a low level in the data model, and “virtual medical records” from multiple healthcare systems were able to be automatically displayed to clinicians with minimal modification to client applications.<ref name=overhage-1995>Overhage JM, Tierney WM, McDonald CJ. Design and implementation of the Indianapolis network for patient care and research. Bulletin of the Medical Library Association 1995; 83:48–56.</ref><ref name=sideli-1992>Sideli RV, Friedman C. Validating patient names in an integrated clinical information system. SCAMC Proc 1992; 588–592.
 +
</ref><ref name=mcdonald-1997></ref><ref name=mcdonald-1999></ref>
  
−
4. McDonald CJ. The Regenstrief Medical Record System: a quarter century experience. International Journal of Medical Informatics. 1999. 54(1999)225-53.
+
By 1996, the now-“cross-enterprise” RMRS had evolve into one of the first statewide [[HIE|health-information exchanges (HIE)]]: the Indiana Network for Patient Care and Research (INPCR) database<ref name=overhage-1995></ref><ref name=mcdonald-1999></ref>. This would later evolve during the late 1990s into the Indiana Network for Patient Care (INPC) database.
  
−
5. Dexter PR, et al. A Computerized Reminder System to Increase The Use of Preventative Care for Hospitalized Patients. NEJM 2001; 345 (13):965-70.
+
== Twilight: 2000–2016 ==
  
−
6. Friedlin J., et al. Details of a Successful Clinical Decision Support System. AMIA Annu Symp Proc 2007; 2007:254-258.
+
A 1999 review of RMRS characterized its clinician users across its several core system as “happy”; it identified several challenges for the future, including further accommodating dictation-based workflows, further developing its then-new web-based graphical user interface, exploring wireless and remote clinician access, and maintaining secure authentication with biometric technology<ref name=mcdonald-1999></ref>.
−
+
 
−
[[Category: EHR]]
+
However, by 2014, the three hospitals of Indiana University Health (formerly Clarian Health)<ref name=iu-2010></ref> had switched from to Cerner’s EHR suite<ref name=cerner-2018>Park S. The Buck Stops Here. Cerner Perspectives. 2018 Sep 19. Retrieved 2021 Oct 24. https://www.cerner.com/perspectives/the-buck-stops-here</ref>, and Sidney & Lois Eskenazi Hospital remained the only major healthcare institution directly using RMRS and its Gopher system for order entry.
−
[[Category: UT-SHIS SP09]]
+
 
 +
In the meantime, the Regenstrief Institute in 2004 gave control of INPC to a new nonprofit corporation, the [[IHIE|Indiana Health Information Exchange (IHIE)]], which was formed from an alliance between thirteen Indiana healthcare institutions and incorporated in 2004<ref name=ihie-2009>Overhage JM. Indiana Health Information Exchange, Nationwide Health Information Network (NHIN) Trial Implementations, Task 9 Deliverable: Jurisdiction-Specific Business Plan, Contract No. HHSP23320074102EC. 2009 Jan 12. Retrieved 2021 Oct 24. Retrieved 2021 Oct 24. https://www.healthit.gov/sites/default/files/indiana_business_plan_nhin_final.pdf</ref>, with the goal to create sustainable business models and provide commercial-level support for INPC, while the Regenstrief Institute would continue managing research access to its information<ref name=regenstrief-2021>Regenstrief Institute Data Services. How we bring the data to you. 2020. Retrieved 2021 Oct 24. https://www.regenstrief.org/rds/data/</ref>.
 +
 
 +
RMRS’s Gopher system continued to serve Sidney & Lois Eskenazi Hospital’s patients until 2016, when Eskenazi Hospital switched from RMRS to Epic<ref name=apathy-2017></ref><ref name=colicchio-2020></ref>. This was a part of a greater trend of the replacement of homegrown EHRs in the United States by vended products that were government-certified under [[meaningful use|meaningful-use]] programs<ref name=colicchio-2020></ref>.
 +
 
 +
== Legacy ==
 +
 
 +
As of 2021, the [[IHIE|Indiana Health Information Exchange (IHIE)]] and the Regenstrief continue to manage INPC and provide statewide health-information exchange via several client-side applications<ref name=ihie-2021>Indiana Health Information Exchange. 2020. https://www.ihie.org/</ref><ref name=regenstrief-2021></ref>, which were originally derived from RMRS’s homegrown data model from the 1970s, and which and continue to use the cross-enterprise global patient index that RMRS pioneered in the 1990s<ref name=overhage-1995></ref>.
 +
 
 +
Regenstrief, the Indiana School of Medicine, Sidney & Lois Eskenazi Hospital, and the American Medical Association jointly created the Regenstrief Teaching EMR (tEMR) as a “copy” of the RMRS data model and other technology. The Regenstrief tEMR has been successfully deployed in at twelve health-professional schools with over 11,000 unique student users as of 2021.<ref name=takesue-2021>Takesue BY, Tierney WM, Embí PJ, Mamlin BW, Warvel JA, Litzelman DK. Regenstrief Teaching Electronic Medical Record (tEMR) platform: a novel tool for teaching and evaluating applied health information technology. 2021 Mar 17;4(1):ooab010. doi: 10.1093/jamiaopen/ooab010. eCollection 2021 Jan.</ref>
 +
 
 +
RMRS’s [[CDS|clinical decision support]] in the 1980s, along with [[HELP|Health Evaluation through Logical Processing (HELP)]], would give rise in 1989 to [[Arden Syntax|Arden syntax]], a markup language for representing shared medical knowledge. These continue on as [[HL7]] and American National Standards Institute (ANSI) standards.<ref name=hripcsak-2018>Hripcsak G, Wigertz OB, Clayton P. Origins of the Arden Syntax. Artif Intell Med. 2018 Nov;92:7–9. doi: 10.1016/j.artmed.2015.05.006. Epub 2015 Jul 2.</ref>.
 +
 
 +
== References ==
 +
<references/>
 +
 
 +
Submitted by J. S. Choi, MD.
 +
[[Category:BMI512-FALL-21]]

Latest revision as of 00:22, 8 December 2021

The Regenstrief Medical Record System (RMRS) is a comprehensive homegrown data model for electronic health records (EHR) developed by the Regenstrief Institute of Indiana University[1][2][3], as well as an associated database-management system (DBMS), communications system, and a suite of clinician-facing client-side applications called the Medical Gopher[4][5][1]. The system pioneered Indiana Health Information Exchange—one of the first health-information exchanges[2]—as well as one of the first computer-assisted order-entry workstations[4][2].

Between 1972[5][1][2] and 2016[3], RMRS stored and managed patient records across multiple institutions of the greater Indianapolis area; at its peak, it contained more than 200 million coded observations and 3.25 million narrative reports, for more than 1.5 million patients, across at least fifteen independent healthcare organizations[1]. However, its usage was largely retired in 2016[3][2]. It is succeeded today by the Indiana Network for Patient Care (INPC), the Indiana Health Information Exchange, the Regenstrief Teaching EMR, and Arden syntax—all of which were originally based upon the RMRS data model.

Nascence: 1972–1990

1972 was a year from the age of the 16-bit minicomputer[6]. That year, Clement McDonald and Charles Clark, two physicians at Sidney & Lois Eskenazi Hospital (then called Marion County General Hospital[1][7]) used a PDP-11 minicomputer[1][6] to create a computer-stored medical record for 35 patients with diabetes, with the intent of automating diagnosis and management after manually capturing patient data, using a hardcoded but generic data structure[8][9][1]. Their goals included the following[10]:

  1. Eliminating logistic problems of paper records, such as availability and indecipherability.
  2. Reducing the work of clinical bookkeeping that is incidental and nonessential to patient management: “no more missed diagnoses when laboratory evidence shouts its existence, no more forgetting about required preventive care”.
  3. Making informational “gold” in medical records available to clinical, epidemiological, and managerial research.

After encountering difficulty with capturing comprehensive data for the pilot population of 35 patients at their diabetes clinic, McDonald, Clark, and other colleagues would build several systems for capturing data at laboratory and pharmaceutical sources[11][12][1].

Data model and RDBMS

Over 1972–1974, it was found that maintaining a hard-coded data model was not sustainable, due to having to change multiple programs whenever any change was made to their data model’s file structure; the decision was therefore made to create a relational-database management system (RDBMS), due to no commercial DBMS being available for the PDP-11 microcomputer being used[13].

RDBMS, by 1989, ran on several VAX superminicomputer workstations[14] and stored the RMRS relational data model in several large files. Records were physically sorted by patient ID, observation ID, and date/time (reverse chronologically). A single term dictionary controlled the system’s concept vocabulary, including all diagnoses, findings, variables (e.g., Glasgow coma score, diastolic blood pressure), and sets of terms. Questionnaires were defined as structured lists of questions and variables stored in the RMRS; these questionnaires were flexible and accepted codes from many different code systems (e.g., ICD-9, CPT-4, local Regenstrief codes) that were associated with each variable’s term. The physical RDBMS and its data model resembled the following:[15][5][1]

  • Dictionary/master files containing:
    • Terms, each with a name, synonyms, external references (e.g., ICD-9 and SNOMED), etc.
    • Structured-observation types, e.g., “microbiology”.
    • Numeric-observation types, each with a scale factor, reference ranges, etc.
    • Coded-observation types, each with ordinals, nominals, etc.
    • Doctors, each with a name, a license, a speciality, etc.
    • Relationships between these relations:
      • Structured-observation types, numeric-observation types, and coded-observation types were subclasses of terms.
      • Each term could be an aggregate of at least two structured-observation types as its “elements”.
      • Each term could have any number of coded-observation types as its value set.
      • Each term could have one numeric-observation type as its unit.
  • Operational data containing:
    • Encounters, each with a date, billing data, etc.
    • Orders, each with a date/time, a priority, notes, etc.
    • Patients, each with a name, address, etc.
    • Medical-record items, each with a date/time, etc.
    • Narrative reports, each with a header, text, etc.
    • Numeric observations, each with a value, etc.
    • Binary data, e.g., images, voice recordings, electrocardiograms, etc.
    • Coded observations.
    • Relationships between these relations:
      • Each encounter was associated with one patient.
      • Each order was associated with one encounter.
      • Each medical-record item was the result of one order.
      • Each medical-record item was associated with one patient.
      • Each medical-record item had one term as its parameter.
      • Narrative reports, numeric observations, binary data, and coded observations were subclasses of medical-record items.
      • Each coded observation could have any number of terms for its value(s).

Data capture, research, and display

By 1984, comprehensive multisite data capture by RMRS had improved, and new data standards were being developed to reduce the necessity of ancillary service systems[16][5][1]. And even as early as the 1980s, several retrospective clinical studies would be performed using the RMRS dataset[17][18][19][20].

In addition, computerized physician order entry (CPOE) in the outpatient setting was initiated in 1984, which was later extended to the inpatient setting in 1990. These CPOE systems would evolve into elaborate “Medical Gopher” client-side workstations.[21][5][1] Although RMRS’s developers noted initial physician unhappiness in reaction to computerized order entry, this unhappiness would later improve as the Gopher system would improve[1].

Even at these early stages, RMRS generated much research about clinical decision support in the outpatient setting, especially regarding rule-based computer reminders[22][23][24]. Furthermore, in the 1980s, RMRS’s reminder-rule systems and Health Evaluation through Logical Processing (HELP) would give rise in 1989 to Arden syntax, a markup language for representing shared medical knowledge, which would later become an HL7 standard in 1999 and an American National Standards Institute (ANSI) standard in 2002.[25].

Maturity: 1990–2000

By the 1990s, RMRS continued to grow and RMRS was gradually expanded from the outpatient clinics to inpatient settings. It would facilitate the care of more than one million patients while generating dozens of studies. At its peak circa 1999, it contained more than 200 million coded observations, 3.25 million narrative reports, 15 million prescriptions, and 212,000 electrocardiograms, for more than 1.5 million patients—across at least fifteen independent healthcare organizations, with access by more than 2500 employees.[5][1]

Core system: three-legged stool

RMRS’s flagship institutions continued to be Sidney & Lois Eskenazi Hospital (which by this period had been renamed from Marion County General Hospital to Wishard Hospital)[7], as well as the three hospitals of Indiana University Health (formerly Clarian Health)[26]—together with Regenstrief, these together were characterized as RMRS’s “three-legged stool”[5].

Its homegrown central relational-database management system (RDBMS) ran on DEC Alpha servers with an ancillary Novell personal-microcomputer network, both hosted by Eskenazi Hospital.[1] By 1999, RMRS also managed the medical records of all outreach/homeless clinics in the Indianapolis area, as well as mental-health clinics. These would connect with the central Eskenazi hub system, via ethernet cables, Transmission System 1 (T1) lines, and dedicated telephone lines, largely with HL7 messages over Internet Protocol (IP).[1]

In this core system, RMRS captured clinical data from more than two dozen disparate systems, such as direct capture from bedside measuring devices, direct links to electrocardiography carts, direct links to financial systems, regular technician entry from public health records, nurse radio/computerized dictation and voice recording, clerk-entered questionnaire/form results, and computerized physician entry via “Medical Gopher” client-side microcomputer workstations.[21][5][1] Data standards used as of 1999 included HL7 messages, DICOM, NCPDP, LOINC, ICD-9, CPT-4, and ACR radiology codes; UMLS and SNOMED were being explored but not supported at this time[5][1].

Physician order entry and clinical-data charts were initially displayed in text-based interfaces from “Medical Gopher” computer terminals. Eventually, graphical interfaces in web pages were also supported; hard copies were also printed on paper—in general, these were designed to yield specialty-specific, single-page, information-dense overviews of each patient, akin to hospital “scut” cards.[5][1]

Informational feedback such as alerts, e.g., for contraindicated medications, occurred during order entry; research was extended from the outpatient setting[27] to the inpatient setting during the 1990s[28][29][30][31]. Order suggestions were also given depending on rules expressed in a pair of the homegrown rule languages, CARE and G-CARE; these included including “corollary orders”, as well as formulary restrictions from several health-maintenance organizations (HMOs)[32][31][5][1]; for example, entering DVT in a problem list would trigger order sets for heparin and warfarin[1][33]. Physicians, however, were “never blocked from completing an order” but were rather made “aware of the potential problems that an order may cause” [1].

Statewide health-information exchange

Outside of the four-hospital core system, the RMRS data model was further developed during this period to span multiple healthcare systems: separate medical repositories (data “vaults”) were created for each healthcare systems, but these vaults would share a single standardized term dictionary and a “global patient index”. A personal-name matching algorithm would determine which patient records in different healthcare systems referred to the same patient, assigning them the same “global patient ID”. This “cross-enterprise” normalization occurred at a low level in the data model, and “virtual medical records” from multiple healthcare systems were able to be automatically displayed to clinicians with minimal modification to client applications.[34][35][5][1]

By 1996, the now-“cross-enterprise” RMRS had evolve into one of the first statewide health-information exchanges (HIE): the Indiana Network for Patient Care and Research (INPCR) database[34][1]. This would later evolve during the late 1990s into the Indiana Network for Patient Care (INPC) database.

Twilight: 2000–2016

A 1999 review of RMRS characterized its clinician users across its several core system as “happy”; it identified several challenges for the future, including further accommodating dictation-based workflows, further developing its then-new web-based graphical user interface, exploring wireless and remote clinician access, and maintaining secure authentication with biometric technology[1].

However, by 2014, the three hospitals of Indiana University Health (formerly Clarian Health)[26] had switched from to Cerner’s EHR suite[36], and Sidney & Lois Eskenazi Hospital remained the only major healthcare institution directly using RMRS and its Gopher system for order entry.

In the meantime, the Regenstrief Institute in 2004 gave control of INPC to a new nonprofit corporation, the Indiana Health Information Exchange (IHIE), which was formed from an alliance between thirteen Indiana healthcare institutions and incorporated in 2004[37], with the goal to create sustainable business models and provide commercial-level support for INPC, while the Regenstrief Institute would continue managing research access to its information[38].

RMRS’s Gopher system continued to serve Sidney & Lois Eskenazi Hospital’s patients until 2016, when Eskenazi Hospital switched from RMRS to Epic[3][2]. This was a part of a greater trend of the replacement of homegrown EHRs in the United States by vended products that were government-certified under meaningful-use programs[2].

Legacy

As of 2021, the Indiana Health Information Exchange (IHIE) and the Regenstrief continue to manage INPC and provide statewide health-information exchange via several client-side applications[39][38], which were originally derived from RMRS’s homegrown data model from the 1970s, and which and continue to use the cross-enterprise global patient index that RMRS pioneered in the 1990s[34].

Regenstrief, the Indiana School of Medicine, Sidney & Lois Eskenazi Hospital, and the American Medical Association jointly created the Regenstrief Teaching EMR (tEMR) as a “copy” of the RMRS data model and other technology. The Regenstrief tEMR has been successfully deployed in at twelve health-professional schools with over 11,000 unique student users as of 2021.[40]

RMRS’s clinical decision support in the 1980s, along with Health Evaluation through Logical Processing (HELP), would give rise in 1989 to Arden syntax, a markup language for representing shared medical knowledge. These continue on as HL7 and American National Standards Institute (ANSI) standards.[25].

References

  1. ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 1.21 1.22 1.23 McDonald CJ, Overhage JM, Tierney WM, et al. The Regenstrief Medical Record System: a quarter century experience. Int J Med Inform. 1999 Jun;54(3):225-53. doi: 10.1016/s1386-5056(99)00009-x.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 Colicchio TK, Cimino JJ. Twilighted Homegrown Systems: The Experience of Six Traditional Electronic Health Record Developers in the Post-Meaningful Use Era. Appl Clin Inform. 2020 Mar;11(2):356-365. doi: 10.1055/s-0040-1710310. Epub 2020 May 20.
  3. ↑ 3.0 3.1 3.2 3.3 Apathy N, Vest J, Menachemi N, Harle C. Changes in User Perceptions of Their Electronic Health Record Following the Replacement of a Legacy System. AcademyHealth Annual Research Meeting. 2017. Retrieved 2021 Oct 24. https://academyhealth.confex.com/academyhealth/2017arm/meetingapp.cgi/Paper/16422
  4. ↑ 4.0 4.1 McDonald CJ. The medical gopher—a microcomputer based physician work station. Proc Annu Symp Comput Appl Med Care 1984:453–459.
  5. ↑ 5.00 5.01 5.02 5.03 5.04 5.05 5.06 5.07 5.08 5.09 5.10 5.11 McDonald CJ, Tierney WM, Overhage JM, Dexter P, Takesue BY, Abernathy G. The three-legged stool: Regenstrief Institute for Health Care. Proc 3rd Annual N.E. Davis CPR Recognition Symp. 1997; 101–147.
  6. ↑ 6.0 6.1 Bob Supnik. Simulators: Virtual Machines of the Past (and Future). 2004. ACM Queue. 2(5).
  7. ↑ 7.0 7.1 Eskenazi Health. History. 2020. Retrieved 2021 Oct 24. https://www.eskenazihealth.edu/about/history
  8. ↑ McDonald CJ. Computer applications to ambulatory care. Proceedings of the IEEE Conference on Systems, Man, and Cybernetics; Boston, Massachusetts, November 5–6, 1973.
  9. ↑ McDonald CJ, Bhargava B. Tree systems for medical information processing. Proceedings of the Eleventh Annual Rocky Mountain Bioengineering Symposium, USAF Academy, Colorado, March 1974.
  10. ↑ McDonald CJ. Action-Oriented Decisions in Ambulatory Medicine. Yearbook Medical Publishers, Chicago, 1981.
  11. ↑ McDonald CJ, Martin G. A model lab information system. Proceedings of the Fifth Annual Pittsburgh Conference on Modeling and Simulation, April 1974.
  12. ↑ Carlstedt B, Jeris DW, Kramer W, Griefenhage R, McDonald CJ. A computer-based pharmacy system for ambulatory patient care. The Indiana Pharmacist 1977; 58:92–98.
  13. ↑ McDonald CJ, Bhargava B, Jeris DW. A clinical information system (CIS) for ambulatory care. Proceedings of the AFIPS National Computer Conference, Anaheim, California 1975.
  14. ↑ McDonald CJ, Day Z, Martin DK, Tierney WM. The Regenstrief medical record: 1989, a campus-wide system. Proceedings of the Symposium on Computer Applications in Medical Care, Washington, DC 1989.
  15. ↑ McDonald CJ, Blevins L, Tierney WM, Martin DK. The Regenstrief Medical Records. MD Comput 1988; 5:34–47.
  16. ↑ McDonald CJ, Wiederhold G, Simborg DW. A discussion of the draft proposal for data exchange standards for clinical laboratory results. Proceedings of Symposium on Computer Applications in Medical Care 1984; 406413.
  17. ↑ McDonald CJ, Hui SL, Tierney WM. Diuretic-induced laboratory abnormalities that predict ventricular ectopy. J Chron Dis 1986; 39:127–135.
  18. ↑ Psaty BM, Tierney WM, Martin DK, McDonald CJ. The value of serum iron studies as a test for iron-deficiency anemia in a county hospital. J Gen Intern Med 1987; 2:160–7.
  19. ↑ Tierney WM, Martin DK, Hui SL, McDonald CJ. Using clinical data to predict abnormal serum electrolytes and blood cell profiles. J Gen Intern Med 1989; 4:375–383.
  20. ↑ Tierney WM, McDonald CJ, Luft FC. Renal disease in hypertensive adults: Effect of race and type-II diabetes mellitus. Am J Kid Dis 1989; XIII:485–493.
  21. ↑ 21.0 21.1 McDonald CJ, Tierney WM, Overhage JM. H.I.S. and the physician: Direct inpatient order entry by physicians through medical gopher workstationsproblems and promises. Proceedings of the IMIA Working Conference on Trends in Modern Hospital Information Systems. (Gottingen, Germany, September 7–11, 1991). North-Holland: Elsevier Science Publishers B.V., 1992 IMIA.
  22. ↑ McDonald CJ. Use of a computer to detect and respond to clinical events: its effect on clinical behavior. Ann Intern Med 1976; 84:162–167. McDonald J. Computer reminders, the quality of care and the nonperfectability of man. N Engl J Med 1976; 295: 1351–1355.
  23. ↑ Prokosch HU, McDonald CJ. The effect of computer reminders on the quality of care and resource use. Chapter in hospital information systems; design and development characteristics; impact and future architecture. Edited by Prokosch HU, Dudeck J. 1995; III(13):221–239. Elsevier Publishing Co. Amsterdam.
  24. ↑ McDonald CJ, Hui SL, Smith DM, Tierney WM, Cohen SJ, Weinberger M. Reminders to physicians from an introspective computer medical record. Annals of Internal Medicine 1984; 27, 100: 130–138.
  25. ↑ 25.0 25.1 Hripcsak G, Wigertz OB, Clayton P. Origins of the Arden Syntax. Artif Intell Med. 2018 Nov;92:7–9. doi: 10.1016/j.artmed.2015.05.006. Epub 2015 Jul 2.
  26. ↑ 26.0 26.1 Indiana University Media Relations. Clarian Health to become Indiana University Health in spring 2011. 2010 May 5. Retrieved 2021 Oct 24. https://newsinfo.iu.edu/news/page/normal/14377.html
  27. ↑ McPhee SJ, Bird JA, Fordham D, Rodnick JE, Osborn EH. Promoting cancer prevention activities by primary care physicians: results of a randomized, controlled trial. JAMA 1991; 66:538–544.
  28. ↑ Rind DM, Safran C, Phillips RS, Wang Q, et al. Effect of computer-based alerts on the treatment and outcomes of hospitalized patients. Arch Intern Med 1994; 154:1511–1517.
  29. ↑ Pestotnik SL, Classen DC, Evans RS, Burke JP. Implementing antibiotic practice guidelines through computer-assisted decision support: clinical and financial outcomes. Ann Intern Med 1996; 124:884–890.
  30. ↑ Overhage JM, Tierney WM, McDonald CJ. Computer reminders to implement preventive care guidelines for hospital inpatients. Arch Internal Medicine 1996; 156:1551–1556.
  31. ↑ 31.0 31.1 Overhage JM, Tierney WM, Zhou XH, McDonald CJ. A Randomized Trial of “Corollary Orders” to Prevent Errors of Omission. JAMIA 1997;4:364–375.
  32. ↑ Overhage JM, Tierney WM, Mamlin B, Warvel JA, Warvel JS, McDonald CJ. A tool for provider interaction during patient care: G-CARE. JAMIA 1996; 19th SCAMC Proceedings: 178–182.
  33. ↑ Mamlin BW, Overhage JM, Tierney W, Dexter P, McDonald CJ. Clinical decision support within the Regenstrief Medical Record System. Clinical decision support systems: Theory and practice, Second edition. 2007: 190–214.
  34. ↑ 34.0 34.1 34.2 Overhage JM, Tierney WM, McDonald CJ. Design and implementation of the Indianapolis network for patient care and research. Bulletin of the Medical Library Association 1995; 83:48–56.
  35. ↑ Sideli RV, Friedman C. Validating patient names in an integrated clinical information system. SCAMC Proc 1992; 588–592.
  36. ↑ Park S. The Buck Stops Here. Cerner Perspectives. 2018 Sep 19. Retrieved 2021 Oct 24. https://www.cerner.com/perspectives/the-buck-stops-here
  37. ↑ Overhage JM. Indiana Health Information Exchange, Nationwide Health Information Network (NHIN) Trial Implementations, Task 9 Deliverable: Jurisdiction-Specific Business Plan, Contract No. HHSP23320074102EC. 2009 Jan 12. Retrieved 2021 Oct 24. Retrieved 2021 Oct 24. https://www.healthit.gov/sites/default/files/indiana_business_plan_nhin_final.pdf
  38. ↑ 38.0 38.1 Regenstrief Institute Data Services. How we bring the data to you. 2020. Retrieved 2021 Oct 24. https://www.regenstrief.org/rds/data/
  39. ↑ Indiana Health Information Exchange. 2020. https://www.ihie.org/
  40. ↑ Takesue BY, Tierney WM, Embí PJ, Mamlin BW, Warvel JA, Litzelman DK. Regenstrief Teaching Electronic Medical Record (tEMR) platform: a novel tool for teaching and evaluating applied health information technology. 2021 Mar 17;4(1):ooab010. doi: 10.1093/jamiaopen/ooab010. eCollection 2021 Jan.

Submitted by J. S. Choi, MD.

Retrieved from "http://clinfowiki.org/wiki/index.php?title=Regenstrief_Medical_Record_System_(RMRS)&oldid=33123"