INFORMATION SYSTEMS AT THE JOHNS HOPKINS HOSPITAL

Transcription

1 BRUE. BLUM NFORMATON SYSTEMS AT THE JOHNS HOPKNS HOSPTAL omuters are laying an increasingly imortant role in atient care. Ten years ago, systems were limited rimarily to rocessing data; today they are becoming an integral art of the medical decision-making rocess. This article reviews the Johns Hokins Medical nstitutions' exerience with clinical information systems and describes two alications in detail. NTRODUTON The otential benefit of comuters in health care delivery was recognized in the early 960's. t had already been established that about one-fourth of all activity in a hosital was dedicated to information rocessing. learly, the comuter could be used to manage this function, rovide surveillance and reminders, and assume some of the labor-intensive activities. Unfortunately, the technology of the 960's could not economically and reliably suort the concets of the early visionaries. The result was overromise, under-delivery, and a decade of sketicism. By the early 970's, the diligence of researchers and develoers had been rewarded. Systems were in oerational use; in many cases the cost of oeration was fully suorted by atient care revenues. Progress continued, and 975 reresents something of a watershed year in the scientific literature of medical information science. Prior to that year, most aers were either seculative or descritive. Since the oerational half-life of the early systems was shorter than the normal ublication cycle, many aers resented defunct alications. By 975, however, key contributors reorted on fi ve years' exerience with a system. Data were evaluated and system benefits were identified. Although most alications still were suorted by research funds, rogress was made in the develoment of transortable systems that could be justified on the basis of cost or benefit. Building on newer, less exensive equiment, the use of comuters roliferated. The 982 market for hosital' information systems is rojected to be close to $ billion. 2 Over of all hositals of 00 beds or more use or lan to use comuters. 3 Of course, much of this use will be for financial and administrative management. Nevertheless, comuters are laying an increasing role in the rocess of atient care. n what follows, we trace the use of information systems at the Johns Hokins Hosital. This history records both success and failure. Because the rimary focus is the imact on atient care, the discussion is limited to clinical information systems. These are defined as alications that manage clinical data and that also retain an extended data base. Systems in- 04 cluded in this category are the hosital information system, the clinical laboratory system, and surrogate medical record systems. Excluded from consideration are financial information systems, embedded comuters (e.g., the AT scanner), and systems with limited data base facilities (e.g., digital hysiological monitoring systems and EK stations). THE JOHNS HOPKNS HOSPTAL The Johns Hokins Hosital (JHH) is an 00-bed hosital located in the inner city of Baltimore, Md. The hosital annually rovides about 320,000 days of inatient care; there are also over 350,000 outatient visits er year, of which 77,000 are emergency cases. As art of the Johns Hokins Medical nstitutions, the hosital is also involved with teaching, research, and secialized medical care. Hosital nformation Systems n 974, JHH organized itself into a grou of largely indeendent units.4 Each unit (e.g., the Section of Surgical Science or the Deartment of Medicine) was given broad financial resonsibility and authority in the formulation of its individual budget. The reorganization led to dramatic reductions in cost growth and rovided a model for other hositals. With resect to information systems develoment, this reorganization imlied that JHH would rovide central facilities, such as financial management systems and an admission, discharge, and transfer system. ndividual functional units, on the other hand, would be free to urchase or develo local systems if they could be justified. The result was a major commitment to a central system that suorted all JHH units lus the indeendent develoment of smaller systems that would meet secialized needs. Figure is a diagram of the clinical comuter comlex at the East Baltimore camus. The JHH central comuters are located in the basement of the Rutland Avenue garage. They consist of one BM 303 and two BM 434 comuters with 3.5 gigabytes of mass storage. This system suorts 350 video terminals throughout the hosital; it also rovides batch su- Johns Hokins A PL Technical Diges(

2 Outatient Research Deartment of Deartment enter Oncology Laboratory Medicine Pediatric comuter comuter comuter comuter comuter Professional Oncology Laboratory fee LiNFO linical Medicine omrehensive system system nformation System child care System JHH central comuters i L L J Outatient identification Outatient registration ore Record System Radiology reorting Laboratory results reorting Anesthesiology comuter Oerating room scheduling nterfaces to _J other JHH systems..j r--=-= ~-~---patient identification Administrative systems Admission, discharge, transfer Outatient natient systems systems Hokins Patient System Laboratory results reorting Ancillary service systems Radiology reorting --- On- line terminal access Hokins Patient System ancillary suort ---omuter-to-comuter interface Radiology Tae data exchange Pharmacy Figure - Johns Hokins Hosital clinical information systems. ort for all of the hosital's administrative systems and some of the School of Medicine's research needs. The clinical functions suorted by the JHH central comuters are divided into four general categories. Administrative systems. These include the identification of atient history numbers from the name and demograhic information; the recording of such general information as address, next of kin, and insurance coverage; and atient location within the hosital. 2. natient systems. All hosital-wide inatient functions are now being merged into a single integrated unit called the Hokins Patient System. 3. Outatient systems. As in most large hositals, the large number of outatient visits limits the scoe of the automated suort. Access to some current information is available through the Hokins Patient System. 4. Ancillary services. A hosital is normally divided into atient care services (e.g., surgery, ediatrics) and ancillary services (e.g., harmacy, radiology, clinical laboratory). Each system that rovides a atient care service also includes functions that are uniquely directed to the management of the ancillary service units (e.g., work lists for the clinical laboratory). n addition to the JHH central comuter comlex, there are several smaller comuter facilities that are dedicated to the local needs of secific functional units. Deartment of Laboratory Medicine. A network of comuters manages the rocessing of information within the clinical laboratory. This involves interfacing with automatic analyzers, rocessing requisitions, rearing laboratory Volume 4, Number2, 983 reorts, and communicating with the JHH central comuter. 5,6 2. Oncology enter. A air of comuters suorts a 56-bed tertiary care facility for cancer atients that also handles 500 outatient visits a week. The comuters are also used to oerate the rototye ore Record System. 3. Deartment of Pediatric Medicine. This system is dedicated to the omrehensive hild are rogram. t is used to manage administrative data, resent clinical information, and suort quality assurance and research analysis. 7,8 4. Deartment of Anesthesiology and ritical are Medicine. A small comuter is being used to manage scheduling for the oerating rooms and critical resources. There are lans to integrate this system into a much larger resource scheduling system. n addition to these atient care systems, other systems throughout the Medical nstitutions are used for research and administration. Of articular note are the following. LNFO. This system has been secially designed to allow individual investigators to manage and analyze data from clinical trials. 9,O 2. Professional Fee Billing. A commercial system is used for rofessional fee billing. This system can suort new clinical alications for the outatient units. Develoment of the Systems Under the leadershi of Richard Johns and Donald Simborg, the Deartment of Biomedical Engineering was instrumental in develoing several major clinical systems in the late 960's and early 970's. A rototye ward management system was imlemented in the early 970's. t was designed to organize and assist in carrying out hysicians' written 05

3 B.. Blum - JHM nformation Systems nonharmaceutical orders in a 30-bed acute inatient ward. The system rocessed all hysician orders and roduced action sheets for each nursing team and atient. Secial action sheets dealt with diets, utility rooms, weight, secimens, intake and outut, and vital signs. Standing order sheets were also rinted. During its trial eriod, the system was well received by the users. A detailed analysis that comared automated and nonautomated 30-bed units demonstrated that the system was able to reduce errors. 2 A summary of the findings is given in Table. Unfortunately, the cost of equiment was too great to justify fullscale imlementation, and the system was abandoned after the concet had been demonstrated. Table - Summary of ward errors with and without a rototye ward management system. Orders examined (total no.) Transcrition errors (0J0) ommunication errors (total no.) Uninterruted communication errors (no.) Failure to carry out order exactly (070) Automated Units Unautomated Units A different alication of a comuterized clinical system was develoed for the inatient harmacy in 970.3,4 ts goals were to reduce errors in the administration of medication, to lower the number of nursing hours required for medication-related activities, and to eliminate waste of medication. The system uses unit-dose drugs (that is, drugs that are reackaged in the unit of use) and works as follows 06. Physician drug orders are entered into the data base at a video terminal. This may be done by a medical rofessional or a clerk at the nursing station, or in the harmacy. At JHH, all orders are entered in the harmacy. The order entry rogram will not accet inaroriate doses, methods of administration, etc. 2. After the order is entered, a harmacist reviews the order against the atient's current drug rofile (i.e., a list of current orders and drugs administered). f there are questions, he will call the hysician. 3. Periodically, the system rocesses all active orders and rints dose enveloes that contain the name and location of the atient, the time of scheduled administration, the drug descrition, and any secial instructions. These enveloes have a clear window on the reverse side so that a technician can check to see that the contents match the instructions rinted on the front side. 4. After the enveloes are filled and checked, they are laced in a tray for delivery to the nursing unit. Each tray tyically contains all drugs required for a two-hour eriod. One hour rior to the time of administration, a list of drugs is rinted, and the harmacy technicians adjust the trays to reflect any changes made since the enveloes were first rinted. The tray and list are then delivered to the floor for administration. By having the comuter manage the current orders and by using multile checks for accuracy, the number of medication errors was reduced by about 750/0 (see Table 2). Because the comuter is used to rocess all orders, automatic billing and rearation of daily and discharge drug rofiles are ossible. The availability of the rofiles reduces clerical activity and makes timely data available in an easy-to-read format. Table 2 - Summary of harmacy errors with and without a comuterized unit dose system. Automated Units Unautomated Units Observations (total) Wrong route 2 Wrong dose 7 76 Extra dose 3 8 Unordered drug 5 8 Wrong form 4 The automated unit-dose system was in oerational use from 970 to 983. ts scoe was limited to the deartments of Medicine, Pediatrics, and Oncology. The decision to discontinue its use was based uon its oerating cost and the need to imlement a new harmacy system for the Hokins Patient System. An equivalent harmacy system with an identical flow was written for the Oncology enter and now oerates on the Oncology enter's comuter. A manual, unit-dose cart system is used in all other units; an automated harmacy system is scheduled as art of the imlementation lan for the Hokins Patient System. A third system, the Johns Hokins radiology reorting system, was develoed in the late 960' s to seed and facilitate the generation of radiology reorts. 5 t is used throughout the hosital and follows this flow. On comletion of filming, each study is given an accession number, and the atient identifier Johns Hokins APL Technical Digest

4 B.. Blum - JHM n/ormation Systems information is entered by a clerk into the system's data base. n the reading room, the radiologist signs on to the reorting terminal. All reorts generated after this sign-on will contain his name at the bottom. To begin reorting, the radiologist enters the accession number of the case, and the terminal dislays a frame with logically arranged lists of words and anatomic diagrams secific to each tye of examination (see Fig. 2). The reort is generated by robing (touching) selected words or hrases; colored robe oints indicate the availability of more detailed frames. Once the reort has been generated, it is reviewed in text form at a video terminal. f it is acceted, END EXAM is robed and the rocess continues. Periodically, the newly generated reorts are rinted, and the rinted reorts are distributed to the ordering hysicians. On-line access to the reort, however, is available at all designated video terminals in the hosital as soon as END EXAM is robed. This system has been in routine use since 972. The cost of oeration is comarable to the traditional method of dictation and stenograhy. The system is easy to use. The radiologists generate reorts almost as quickly as they can dictate, and the final reort is immediately available for verification and retrieval. Finally, the availability of comuter-stored reorts is useful for research, medical audit, teaching, or regeneration of lost medical records. (A commercial version of the system was marketed by Siemens; because of limited sales, it is no longer available.) n arallel with the develoment of these standalone systems, there has been a continuous effort to roduce an integrated hosital system. One key comonent of such a system is the atient locator called the admission/ discharge/transfer system. n the mid970's, such a system was develoed at JHH, one that both maintained the current inatient census and linked the various ancillary systems. The success of this aroach led the hosital to evaluate its needs in the area of hosital information management. After erforming an intensive analysis, JHH decided to install a comrehensive hosital information system. The system chosen was the BM Patient are System, which was originally develoed by the Duke Medical enter and is marketed by BM. 6 The needs of each hosital are different. A considerable eriod of time is necessary before the requirements are established and the system can be in- E ROSON(S~ FOREN MAnER FRATURE - HP HEALED Figure 2 Volume 4, N umber 2, 983 TRAUMA Samle radiology dislay from the commercially available system. 07

5 B.. Blum - JHM n/ormation Systems stalled. At the resent time, the Hokins Patient System rovides access to clinical reorts in the nursing units and selected outatient areas. New and relacement functions are being installed; several years will be required before the comlete system will be imlemented. Along with the creation of a strong centralized system, several functional units develoed systems for their local needs. Two of these systems rovide an interesting contrast in requirements and scoe. The first is the Oncology linical nformation System, designed to manage data and to suort decisionmaking for cancer theray. The second is the ore Record System, designed to roduce a low-cost summary medical record for use in ambulatory care. Both systems were cost justified on the basis of their contribution to atient care; neither develoment was suorted by any government grants. THE ONOLOY LNAL NFORMATON SYSTEM The Johns Hokins Oncology enter is one of 22 omrehensive ancer enters established as art of the National ancer Plan initiated by the ancer Act of 97. The enter has major rograms in laboratory and clinical research, education, and collaborative activities with community hysicians. are of adult atients is centered in a 56-bed facility that can also serve 500 outatients er week. At any one time, 2000 atients are being treated under one or more of several hundred formally established treatment lans called rotocols. mary atient records, current census, and aointments. 3. Schedule atient care activities and theraies as determined by clinical algorithms and research studies. Since most cancer theraies involve sequences of actions over long eriods of time, the ability of the comuter to rovide reminders was considered an imortant asset in dealing with the roblem of information overload. 4. Maintain clinical data for retrosective analysis. n July 975, one year before the enter was to oen, work on a rototye began.? The first task in develoing a rototye was to convert the existing automated tumor registry file into a format that could be exanded to meet the needs of the enter. The data base was transferred and rograms were develoed to roduce atient abstracts such as that shown in Fig. 3. At the same time, work began on the develoment of formats that could dislay clinical data effectively. Flow sheets and rinter lots were designed, and, after many iterations, the formats resented in the following sections were established. 8 The rototye system oerated on a comuter at the Alied Physics Laboratory. The tumor registry was a functional system with a current data base. t ran in the batch mode and was used for searches, reorts, abstract rearation, and quality assurance. The remainder of the rototye system was used on a ~~ T 0 RY lotins -topkns u Oc, P TA L QilrJL 'Y ENTF'R ),Q,TE Develoment of the System The develoment of the system was begun in the mid-970's, when lans were made to consolidate various oncology activities into a single facility. Although there was no funding for a comuter or for software, the facility's architecture and staffing rojections were redicated on the assumtion that a comuter system would be available. Over 250 wall jacks were set into lace, with cables terminating in a comuter room. During the eriod of lanning, it was decided to develo a rototye system. At that time, outatients were being treated in a Johns Hokins clinic, and inatients were being treated in facilities at Baltimore ity Hosital. At each site, rocedures had been established for atient management based on the careful evaluation and monitoring of clinical data. Based on the Oncology enter's clinical ractices, it was determined that an information system was required that would 'ljq "laue '/2/f<2 PHnN L A.5 5 SP(lUSE A RT.j O ATE ~n "' TTT n SEX ~ 0 2 /02/ ir fl"'jl.low s ro. T US li P OTSl-itlR,EO QAf W- TE 09/0 5/8 O'\/ OQ/30 'N OR 0 220~ n, (n o 0 3 / S / A MAP. STAT U S SN L r;" DE " A SE D -_ _ _....-_ _ _ _ _... -_.. -_.. -_ _ _... Ot A(NnSTS PRTo AQY A RU NDEL EN J~H ~ 2' OU TS DE O AN(SS ANNE HnSD. D ANOS S U PP~P LO BF. LUN" / 3) OAT,EL L AReN O. A ( T- 62) R An E P OORL Y DFFER EN TlAT f. D EXT E NT DSTA NT LATF RL LT OR.. AN N VOL VEM ENT ONLY ;B AS S OF DA NOS S P AT ~OLOY. YTOL O Y. or LUL LUN WTH MET. X- RAY. OT HER OJ HH REP ORT S. ARN OHA. ; ' 8- - BONE.AP. OW BX. HETlSTl T SMA L L ELL A P ~ O M' ONS ST ENT WTH L UN PRM> RY. OSL MED AST NAL L y.ph NODE ( OSL). E TA S TAT SOA LL E' LL AR TN OO. O NSSTENT WT H LU - PRUPY. ;;;;--; _ _ _.. ~ -- ; ~ --- ~-; ~ ;~ -~ -; ~;; -~ - ~ ~ ~-~-~~;~;; ;;; ~ ; ~ ~ ;;~;;;~~; -~;; ~ - - ;;~ ;--~ BOP BTOP 06 4 /8 RO P t~o T E S J ''9 H E" BTOP... Y T O '(&,>.J.,l ORT A,"' Y N, OTH BPONHO SO P Y BX. OT H " EDT ASTN OS OPY BX. JHH NO T E J ~ H NOT E 2 JHH MNE.A PR!)" ( P OS ) A MO VP _ _---- -_... -_ _... H T5T OPY. Organize and dislay the clinical data as an integrated reort that combined inatient and outatient data in ways that could indicate trends and facilitate decision-making. 2. Manage the basic atient record functions, including a hosital-wide tumor registry, sum08 ~ E n A L H STORY SO ll L H S TORY ON SET nf' sy li!. PT n ~ s ' UPAT QN ALLERY OVJFNT S 05 0 P E P T UL ER. 965 T B Str,(n(E 2 OD X 2 } YRS. 3-4 PD D X 8 YRS. 2 / " PT. n, ANK P ol N TEL E PH ON < n rr AT n R PEN LL N DE'. tr.ln~ BAPTTST Figure 3 - Samle atient abstract, Oncology linical nformation System. Johns Hokins APL Technical Digest

6 B.. Blum - JHM nformation Systems limited basis. Because of the difficulty in establishing a current data base, the dislay system was most effective for retrosective studies; its rimary contribution was the establishment of format requirements for oututs. n 976, in art as the result of a gift from the Educational Foundation of America, sufficient funds were available to urchase a comuter and suort the develoment of a system for the Oncology enter. alled the Phase system, it would be a conversion of the rototye to an interactive system on a dedicated comuter. A PDP- comuter was-8elected with the MUMPS oerating system. Work on the rototye stoed, and develoment of the new system began. 9 Figure 4 resents a summary of the develoment activity. As can be seen, the rototye system was converted to an on-line system with the rototye's data base in nine months. The reasons that the system could be raidly imlemented were (a) the existence of a rototye so that the design requirements were well known, (b) the use of MUMPS with its owerful data access tools and interactive debugging caability, and (c) the fact that all activity was devoted to new alications with no effort sent on system maintenance. 20 Develoment of the system continued and new functions were added bacteriology reorts sorted by date, organism, or secimen; aointment system, including scheduling of tests; heresis system to manage blood roduct donation and transfusion; administrative functions such as charge cature and reorting of level of care; and rotocol-directed daily care lans. 2,22 The success of the Phase Oncology linical nformation System began to limit its effectiveness. The system became saturated at 20 concurrent jobs, roduction rocessing required 20 hours a day, and it was difficult to add new functions to the existing system. Further, during eriods of comuter downtime Prototye..t. enter oens t omuter delivered L..-.,. On-line access L-A Daily care lans Second comuter delivered ~, ~ew rograms n use. converslon comlete Analytic tools t------tedum TM develoment t'-- _ L Figure 4 - History of the develoment of the Oncology linical nformation System. Volume 4, N umber 2, 983 there was no automated backu. t was therefore decided to add a second comuter to ermit adding functions and to rovide backu. The second comuter was installed in 980. Standard MUMPS was now available; the revious version was no longer used. Because the two versions were sufficiently different, it was decided to rerogram the system rather than convert it. This became known as the Phase system. t also was decided to use a new rogramming tool, TEDUM, which would rovide data base management system functions and imrove system maintainability U nfortunately, TEDUM was only in its initial stage of develoment, and the staff was faced with the following roblems (a) conversion of what was, by now, a comlex system, (b) maintenance of the old MUMPS system, and (c) use of a new and oorly documented tool. The result was a major sli in the schedule and some user frustration. However, the conversion was successful. Parts of the new system were ut into oerational use in 98, and the old MUMPS system was retired in 982. The two comuters are now linked with distributed data base software; continuous clinical and administrative suort is rovided. There is a direct link to the Deartment of Laboratory Medicine's comuter system, and all test results are automatically transferred to the Oncology linical nformation System. The comuter also suorts an Oncology enter harmacy with a common data base. Because the system manages a very large data base, attention is now being directed to the imlementation of tools to maniulate the data base for retrosective analysis. alled the Phase system, arts are in oerational use; all basic tools will be in lace by the end of 983. Data Dislay and Management The data base contains information on about 37,000 atients, including 9000 who have been treated in the Oncology enter. For all atients in the data base, it is essential to have the basic identification (name and history number), demograhic data (e.g., age, race, sex, lace of birth), and diagnosis (e.g., location of disease, tye of disease, date diagnosed). For the atients treated at the Oncology enter, additional information is required for theray, administration, and evaluation. The abstract, shown in Fig. 3, rovides on-line access to much of the summary data required by the health care team. The abstract contains identification and administrative data. For each rimary tumor site, there is a block that gives the diagnosis and a summary of treatment. These blocks combine both codes and text. For examle, for the atient whose abstract is shown in Fig. 3, the disease is coded as 62.3, Uer Lobe, Lung. While this is effective for searches, there is a less recise text descrition that is more meaningful to the hysician, i.e., "arcinoma of LUL lung with met." A summary of the athology reort rovides more detailed information. 09

7 B.. Blum - JHM nformation Systems The resentation and searching of abstract data are generally straightforward. The management of clinical data, however, is more comlex. Some atients are treated for many years, and all of their inatient and outatient test results, medications, and vital signs are retained. The data base now contains over 4.5 million data oints for those atients treated at the Oncology enter. t is not uncommon for atients to have over 00 data oints recorded in a single day. learly, secial tools are required to resent these data in formats that facilitate medical decision-making. 26 The most common format for data dislay is a chronological tabulation called a flow sheet. The system rovides a caability for ordering flow sheets (see Fig. 5) containing secific data for a given time eriod. This resentation is clearly suerior to the collection of laboratory slis that usually confronts a hysician; in fact, in most settings a hysician generally transcribes the information on the laboratory slis onto his own flow sheet. A major deficiency of the flow sheet is its inability to indicate the time axis. This deficiency is effectively overcome in a lot such as that shown in Fig. 6, which shows the first 36 days of theray for a leukemia atient treated with a combination of two antitumor drugs. The latelet and white blood cell counts are lotted on a logarithmic scale. Below the lot are shown the chemotheray administered, the antibi- h ls lllwi NO v4.!o/ _....lqrlah8 -_ Orl -- AR MAa2 0 APkH2 0 APH A PR82 STAN [) A~ [) & t-.lafio;l v A'E. "l.no ~P~ S2 J7Rl l P.y 3~llAY lo na Y J7 DAY 38 DAY 3 9 DAY 4 0 DAY 4, J7 '!l4.~y [) AY 2 )AY 23J Ai 2 4 UAY l5 DAY 2 6 D ~Y 2 ' PAT Sl AT tis _.. TE'.P Df; Pll[.S~. Hf,SP ~H' army... f A" F ~ K,", ". KALfl. ~. ) ~4L.[. % ML ~L ~. ~ PH 'lfl. V.! KA.(P.. ) KA. PP0( P.. ) T OT AL KA, TlT ~" fj r (, Ht.~A ~0 (, ARE Lf n l, f' LU 0 flv), LUln fp ;J) TO fl. [ " lak ll'! '' '; l llll URHS'lOL. i! STnLlL F"rS nlfr UTL Oll l 'U l it ' [f 37., Q ~ 4~ l l276 9~ 54~ 22~ [ b bl. ~ 3"/ lo L~ 24 ~05 25~5 l HO 0 J~ O 550 nso l425 - HoO 8~ ~ bl ~ r( L ' y WB HT Rf;ll'L("YTf. PLAH, Lf. S Mnl.fltH.; o/ U " \ ~~T~ "{(.Y ; ~.l < l].6. 4 ijlooo ' 3000 < ~000 - b OO * < by OOO * noae * \ lbooo< ~ <A N O S " l 5 * f.ljsnflp /ill \ ~ HASflPhl[.; " LY'Pll fl /f>. ; 0 ' 0 * 3 * 9< Mu'o'yns ; Y < b 9 > NETROPl '; ' 62 k O 6 4 t Hf" S,.~ S, " A ~f.v/l Sf.R ~ ~~ V /L S;;R l'l Mf ~ /L StR c n 2 '!' V/. AN /.), ~P S~R l~ H "(; / J L ;f~ kea," ';/ll SlN/"EA SfR (,Life f 'J T Sf- P Vl "(;/ ) L Sf. R AL U'MN TUT kll. O l~ ". Sfl T SPl - / OL i'\ / LJ L "'(, / ('\[, lun l UlL v~ Y5' 29 7 O. Y 3 44* 33 <. 7< Y5 * lo 2 ) ' 5 < 4. 2 ys * 30 D ' l6 4 ~ 0. 7 ll yy < O.b > 0. 3> 37 54' Ol * 2. 9< ' 48 < _ Figure 5-0 Samle flowsheet otics given, the blood roducts transfused, and the temerature in degrees above normal. To illustrate how such lots can be used in atient care, note how this single lot combines information to rovide an overview of the atient's rogress. Theray for the leukemia with cytosine arabinoside and daunomycin was begun on February 6. With the convention used by this rotocol, this date is renumbered treatment day (TX DAY). Based on a laboratory research model of cell kinetics, a second cycle of cytosine arabinoside chemotheray is given beginning on day 8. The effect of this treatment is readily seen. A large number of malignant leukemic cells (W) are eliminated, falling from a retreatment level of 80,000 to a level of 00 by day 5. At the bottom of the lot, the atient's maximum daily temerature is indicated in degrees above normal (centigrade). The temerature is elevated at the outset of treatment. As a toxic reaction to the chemotheray, bone marrow alasia (signified by a white blood cell count of less than 000) is added as an additional medical roblem. Two antibiotics are begun; the temerature falls to normal. Protocols for the management of fever in the absence of normal white blood cells are a routine art of the organized aroach to clinical management that must be taken into consideration in the design of this comuter dislay To continue with this illustration, the imact of a standard rotocol for administration of blood latelets is also seen. Platelet levels, lotted as "P" on the grah, are monitored daily. When the number of latelets in the blood falls below 20,000 cubic millimeter, there is a danger of hemorrhaging, and latelet transfusions are given. This lot uses a horizontal line drawn at 20,000 to rovide a visual guide to the medical management team. Transfusions of latelets and the resonding rise to higher levels on the following day are indicated along the bottom of the lot and in uward deflections of the lotted line. n this atient, it is clear that latelet resonsiveness exists and continues throughout the treatment. By day 9, normal white blood cells are seen to be returning, and by day 32 the atient is shown to be in remission from his leukemia and ready for discharge. The use of daily lots combined with comosite lots of the same variables (mean and confidence limits for ten or more atients treated with this cytotoxic regimen) hel in assessing this atient's status and allow the hysician to comare the atient's resonse to the combined exerience of the total grou under treatment. The availability of actual data to accurately describe ast and current clinical exerience is considered an imortant basis for rational clinical decisions. n addition, these features of the system suort the enter's clinical research rograms. n this examle, more than a month of theray has been summarized in a grah to show the basic rocesses of tumor treatment, control of infectious disease, and blood roduct suort. During this J ohns Hokins A PL Technical Digest

8 B.. Blum - JHM nformation Systems NA" ~ J OH "S HOPK J ks SfflllLfl, FLlJl UF 'khltf t.ll AN D PLATELET DAT A HS r f) fril OLOY E~Tt (JUP!S 0. -, 000,000) DAT E 0 4 /0 4 / ~ _... - _ _ _ _ ~ ".PP' P P F.... '. P... i' P'" t-? ' P ' ' P. P ' Pi" P... _--_ "' V ' P... P P.. '" 000 "" P P,. W'\ W'.""."' W''' W ' _ l "w w lit '" WW.WW'." W' lit... O't\fro."',.. ~ ""' ~.,..\Url "ff' \ 't TOS t A ~ ---- ~.. -~ -... ~ !YTOS HE A Figure 6 -Samle lot. UAUNU" Y J "' 0 ) U DAUNOMYN bu. ~.L PRt'P ~ ; -- ; ;.. - ~ -... ~... ; _.. ;... ; -- ~.. - ~... ;.. - ;... ; -- ;... ;.. - ; - - ;. - ;.. -;--;--;--;--;--;... ;.. -;--;_... ; -.. ;... ;_... ~ BOW EL PREP t ~ U "l ~ A P ijf.~ll "' KlFl. N <; (, t..! L w.... Y~ P N ll - V Tl~P OAl' f e c e ENTA.JnN A RBENLLN K c c c KEfLN LNDAiYN P[NL-V ~ - ;.. -;-- ~ -- ; -.. ; ;- - ;.. -~-- ~ --- ~ -.. ~ - -;--;--;-- ~ -- ; ~ -.. ;-- ~.. ---~--;.. -;... ~--~-.. ; ;--;---; TEMP PLU Tk u S wac TU~S Ti q~~ ') t ~ ~ -~.. - ; ') h 7 _00 ;.. -; --;.. -;... ;.. - ~ -- ~.. ; ~ -;;- ;;-;;.. ~ 9 2 (} 22 2J 4 2~ 4!b 6;!;;-;;. ;;-;;.. ;;-;;-;;-;;-;;-;;. t l 2 27 Lb 0 (2 u ;;-;;-;;-~~ Daily are Plans Daily care lans were designed to assist the hysician who must treat many atients over long eriods of time using comlex treatment modalities in both an inatient and an ambulatory setting. n a less comlex environment, McDonald has shown 27 that there is a roblem of hysician information overload that can be alleviated by use of an automated system. Automated harmacy systems have illustrated how routine surveillance for drug-drug interactions causes hysicians to modify their theraeutic orders. 28,29 Wirtschafter et al. have demonstrated that automated rocedures used in a community hysician outreach rogram can roduce very high rates of comliance with a research rotocol. 30 The urose of the daily care lan, therefore, was to introduce automated suort for theray decision-making. n order to understand how an automated system may be used in a facility such as the Oncology enter, it is necessary to areciate the general rocess of atient care. Patients are treated for their cancer, as well as for disease or theray-related medical comlications. Much of this theray follows redefined rotocols (clinical algorithms or treatment sequences) that may be groued into four categories. Research rotocols. These define a set of theraeutic actions to be followed for a secific o- PLAT TRA.NS. ;;-;;-;;-;;-;;-;; eriod, other organ systems are under stress, and similar lots may be used to monitor the function of individual organs such as kidneys or liver, selected general functions such as nutrition or fluid balance, or secific medical comlications such as hyercalcemia. Each day aroximately 00 lots are rinted and distributed to the atient areas in time for morning rounds or an outatient visit. Vo{ume4, N umber2, wac TRANS ;;-; ~ 5 X DA. Y DATE ulation in order to roduce a consistent set of data to evaluate a hyothesis. 2. ndividual theray. This involves the use of antitumor drugs generally in multi drug combinations administered using comlex time-sequenced relationshis. Theray may extend for months or years. An examle is the use of a drug sequence demonstrated to be effective by a revious research rotocol. 3. eneral suort and resonse to life-threatening crises. These involve the use of established rocedures to deal with anticiated reactions and roblems associated with the antitumor theray. n many cases, these actions are secified in each research rotocol. Examles are (a) the use of antibiotic drugs to treat infections when it is known that the antitumor drug will severely lower the white blood cell count and (b) management of fluid balance. 4. Disease-secific continuing care. This entails the recommended long- and short-term care and atient monitoring associated with a secific disease indeendent of the theray selected. These actions may also be included in the research rotocol documentation. Examles are (a) routine six-month chest X rays for all atients with breast cancer and (b) routine three-month monitoring of serum rotein electrohoresis results for certain multile myeloma atients. At any given time, an individual atient may be treated by one or more research rotocols (e.g., an antitumor rotocol and an antibiotic rotocol). The atient whose flow sheet is shown in Fig. 5 is being treated by two rotocols. (The day of treatment under each is given in the heading.) These research

9 B.. Blum - JHM nformation Systems rotocols may be sulemented by other suort rotocols (e.g., a hyercalcemia rotocol to hel manage high levels of serum calcium) and long-term follow-u (e.g., annual battery of tests). The situation is further comlicated by the fact that (a) rotocols may have branches for reselected grous or outcomes in which atients are given different doses of the same drug or combinations of different drugs, (b) the Oncology enter is a teaching institution and thus subject to house-staff rotation, (c) at anyone time, there are over 2000 atients being actively treated by the enter, and (d) there are currently more than 25 formal research rotocols active in the enter. The daily care lan system oerates by first breaking down the rotocol into logical theray units called treatment sequences. The system then rovides tools to allow the hysician to assign these sequences to atients starting on a given date. This is called a standing order. Each day the standing orders are examined, redundant orders are removed, and a set of recommended orders is roduced. These are rinted as the lan that the hysician uses to order tests and rocedures. The results of these orders are then entered into the data base so that the next day's lan can be derived from both what the standing order requires and what was actually done. The daily care lan is rinted in several different formats. A hysician lan summarizes information about the atient's status and tests to be ordered, a general lan contains more comlete information, and an order guide is used by the ward clerk to rocess test requisitions and by the hlebotomist to obtain secimens. The general daily care lan is illustrated in Fig. 7. t contains JOHNS HOPK NS ONOLOY ENTER ENERAL DA L Y ARE PLAN HSTORY NO NAME PLAN DATE 08/0680 WEDNES['AY 20 W F HODKN'S D SEASE HT 55 ADM WT 44 flas.4 DEflL WT 45 FLOOD TYPE 0+ HLA PROTOOL E475 FLEO-MOPP-XET DAY 29 STARTED SEOND YLE - BLEO MOPP 8/6/80 PROVDER LENHARD, DR R PROTOOL ENERATED NSTRUTONS ************************************************************************* MAX UMLATVE DOSE OF BLEOMYN S 36 M / M2 * ************************************************************************ ENTER PERFORMMlE STATUS ENTER TUMOR MEASUREMENTS MEASURE LONE ST DAMETER & TS PERPENDULAR DAMETER N M SEE PROTOOL (SETON 5.2> FOR HENO DOSE MODFATON F WB < 4000 OR PLTS < SEE PROTOOL (SErON 5. 2) FOR DOSE MODFATON F TOTAL BLL >.5 MEASURABLE TUMOR STES NTAL D RD OfT M/METHOD TOTAL AREA SZE (M) LAST DATE SZE (M) DATE 7.78 REENT LNAL FNDNS VALUE ESRPT ON 5.6 UMULATVE TOTAL BLEOMYN 44 NTAL BODY WT AM 43.2 LAST BODY WT AM 97 PEFENT BODY WT AM E'ASED ON MAXMUM SER REAT 900 MNMUM WB MNMUM PL.HELETS LAST FNDN 07/6/80 08/05/80 08/05/80 08/04/80 07/6/80 07/6/80 STARTN 07/0/80 07/0/80 THROU H 08/07/80 08/07/80 08/07/80 08/07/80 08/07/80 08/07/80 HEMOTHERAPY DRU H STORY NAME NTR MUST PROARB VNRST PREDNSONE BLEOMYN DOSE 8.3 M 20 M.9 M 56 M 2.8 U 07/22/80 r. MSELNE 00 ;~ OF 6 8S/. OF 40 95;' OF OF 40 07/6/80 loo/. OF 2.8 U VEN 07/6/80 07/22/80 07/6/80 M/M2 M M M/M2 DRUS TO BE ORDERED ON 08/06/ N BLEOMYN t. V., BASED ON.4 M2 x.? M/M N NTROEN MUSTARD J.V., BASED ON.4 M2 X 6 M/M2.96 M VNRSTNE LV. MAX SNLE DOSE 2.0 M., BASED ON. 4 M2 X l. 4 M i M!. 40 H Pf\OARBAZNE P.O. BASED ON.4 M2 X 00 M / M!. rests AND PROEDURES WTH DATE LAST PERFORME[ BLOOD AND SERUM TESTS HEMATORT > HT > PLTS PLT NT > HB HEMOLOBN S) SURVEY SMA-6 M6 SOT >SOT >TBL TOT fll SER REAT > SR RADOLOY PROEDURES HEST XRAY XR FHYSAN r'oedures OB/OS/80 08/05/80 08/05/80 07/04/80 08/04/80 OS/0/80 08/0/80 08/04/80 (> NDATES ORDERED 08/06/80, STi'T OB/06/80) * ORDERED >D FF ; WB M2 SU >SALK >S&F T >SUN SUR w!c DFF Wle SMA- 2 SER OPPER S ALK PHOS SPT S UREA NT S UR AD Figure 7-08/05/80 08/05/80 08/0/80 07/04/80 08/04/80 08/0/80 08/04/8 08/0/80 07/08/80 REPEAT ANY X- R,n TO DOUMENT RESPON SE OR PRORESSON P. A. PROEDURES MUnrMUMPS SKN 07/04/80 TR TRH SUN PPD PPD SKN 07/04/80 DATA OORDNATor, WSTRUTONS HAS DSEASE PRORESSED? HEK WTH M.D. RE. Data about atient status. Examles include height, weight, body surface area, admission weight, and blood tye. 2. Theray summary. This contains rotocols with associated starting dates, names of hysicians, and general comments. 3. omments generated by treatment sequence. These are groued into several categories, with the most imortant rinted in a box to command attention. 4. Tumor measurements. Where there are solid tumors, a section of the lan records initial and current dimensions and comutes total crosssectional area and ercent of change since the start of treatment. 5. linical findings. These are functions of the clinical data, such as ercent of change in the atient's weight since admission, cumulative drug dose, or a maximum recorded laboratory value. 6. hemotheray orders. Where a treatment sequence indicates that a drug is required, this fact is listed. The notice may also include the WT(OB/05/80) 07/04/80 STOPPN TX SEQUENES Samle general daily care lan (see Note 3). dose comuted as a function of weight, body surface area, or ideal body weight. 7. hemotheray history. This is a flow sheet containing the date and dose of the last administration of each antitumor drug. f a recommended dose has been defined, the ercent of the recommendation is comuted. 8. Tests and rocedures. Deending on the format, this ortion of the lan will contain tests and rocedures to be ordered as determined by the active treatment sequences. STAT (immediate) tests are flagged, and the date the test was last ordered is given. Daily care lans have been used to order tests and rocedures for all bone marrow translant and leukemia atients since Aril 980. They have recently been extended to cover all inatients and outatients treated at the enter. 32,33 Johns Hokins APL Technical Diges(

10 B.. Blum - JHM n/ormation Systems Evaluation of the Oncology linical nformation System Table 3 - Weekly utilization statistics (98) for the Oncol ogy linical nformation System. Three criteria are used to evaluate a clinical information system. Utilization. This is a measure of how and where the system is used. learly, extensive use indicates that the system is an integral art of the health care rocess. Furthermore, extensive use would suggest that the system is erceived to be beneficial. 2. ost. This is the cost of the system both for one-time develoment and for oerations. t includes all costs associated with the information system within the organization. 3. ost benefit. This is a measure of the net cost imact of the system, i.e., the difference between the oerating costs with the nformation system and the estimated cost of oerations if no comuter system existed. With resect to utilization, the system oerates seven days a week and is the rimary suort for clinical data dislays, clinic scheduling, atient admission, tumor registry oeration, the heresis rogram, and test ordering. One asect of use is how well the system fits into the Oncology enter's oeration; Table 3 resents some summary statistics. A second measure of utilization is the number of new rojects. During the ast 2 months, a arallel system was established for Pediatric Oncology, all radiation theray activity is being rocessed by the system for both charge cature and reorting, a unit-dose harmacy system was imlemented, and forms for starting new rotocols have been develoed. Suort for the system and its nformation enter is taken from different cost centers within the Oncology enter. With the excetion of some searately funded research activities, all oerating costs are derived from atient care revenues. The nformation enter assessment is based on a guideline of 3 % of the charges for the services suorted. (The 3070 figure includes a five-year write-off of the caital investment.) Units suorting the nformation enter include Medical Oncology, Pediatric Oncology, Radiation Oncology, the Pheresis enter, the Oncology Pharmacy enter, and the Administrative Offices. The measurement of cost benefit is a more difficult task. n evaluating the system, it is clear that no manual system could rovide the caabilities of the automated system. Thus, a cost benefit analysis would be comaring different caabilities. Furthermore, it must be recognized that the traditional methods for cost benefit analysis consider only those items aearing in the institution's budget. ost savings to other segments of the economy usually are not tabulated. Yet, some of the most dramatic cost savfngs benefit atients and third-arty ayors and may have a negative cost imact on the Oncology enter. By way of examle, the system comonent that was used Volume 4, Number 2, 983 Printed Plot s Daily clinical sets Flow sheets Abstracts Bacteriology reorts Other reorts (searches, queries, secial requests) On-Line Requests *Not alicable. to suort the Pheresis enter enabled the Oncology enter to reduce morbidity and mortality; at the same time, it roduced a cost avoidance to atients of $240,000 a year. 22 Yet, in this case at least, the comuter's contribution to relaized cost savings within the enter's budget cannot be isolated. THE ORE REORD SYSTEM The ore Record System is a rototye automated ambulatory medical record system designed for multiclinic use in outatient deartments. Like many other large urban hositals, JHH frequently serves as a rimary source of health care for a large segment of the local oulation. The visit atterns of these atients are characterized by reeated use of walk-in facilities (articularly the Emergency Deartment), continuing care for the chronically ill, uncoordinated and overlaing use of multile secialty clinics, and occasional inatient admissions. Each year, 00,000 atients are seen in 350,000 visits. n an institution of this size and comlexity, the continuity of care for ambulatory atients is frequently comlicated by the difficulty of finding and extracting information from the medical records. There are aroximately 500 requests for medical records made each day by the outatient clinics. While many requests are satisfied in a timely manner, it is clear that the raid delivery of records on short notice resents a considerable challenge and exense. n addition to the roblem of medical record availability, there is also the need to coordinate atient care within the institution. Dulication of tests should be avoided, rescrition of medication by secialty clinics must be integrated, and atient follow-u and referral must be coordinated. The ore Record System is designed to meet many of the information needs of an ambulatory setting at a cost that is low enough to be accetable by an outatient deartment unit. This is in marked contrast to the Oncology linical nformation System, where the high cost of tertiary care rovides an economic justification for a comlex (and costly) information system. 3

11 B.. Blum - JHM nformation Systems Develoment of the System The initial model for the ore Record System was the Minirecord System. This system was initiated as the result of a grant from the Alied Physics Laboratory to roduce a rototye demonstration of how medical care could benefit from the use of information systems. Begun in 974, the Minirecord System was designed to suort the needs of the Hamman-Baker Medical linic, which rovided long-term care for 7000 chronically ill atients. t had already been demonstrated that the availability of atient roblem lists imroved follow-u and continuity of care, 34 and it was hyothesized that the availability of an automated roblem list would have a beneficial effect on atient care. onsequently, a system was secified that would rovide. Better treatment and follow-u of roblems identified within the Medical linic through the use of a roblem and medication summary; 2. mroved follow-u of roblems identified and treated outside the Medical linic; 3. mmediate access to a minimal medical record to exedite medical evaluations during visits to the Emergency Deartment or Primary are Walk -n linic; 4. Availability of on-line data to evaluate abnormal laboratory test results and unscheduled requests for rescrition refills; 5. An aointment system for the clinic's 00 health-care roviders. Following the imlementation of a rototye system on the Alied Physics Laboratory comuter, a data base was develoed and the system was modified to oerate on the JHH central comuter. Access to the Minirecord was available in the Medical linic, the Emergency Deartment, and selected inatient units. The system was well received and remained in lace until 982, when the Hamman-Baker Medical linic was closed and relaced by the Johns Hokins nternal Medicine Associates. 35,36 The ore Record System grew out of a series of task forces organized in 978 and 979. These grous were directed to identify areas in which automation might imrove oerations in the Outatient Deartment and the Medical Records Deartment. A solution roosed by these grous was the ore Record System. Designed as a multiclinic exansion of the Minirecord, the new system would suort the following functions.. Maintenance of a minimal automated ambulatory medical record. This would be available on-line at strategically located terminals and be rinted as art of the hysician's visit record. t would contain the minimal information necessary for roviding atient care when a more comlete record is not available. The 4 minimal record would also aid in coordinating care among clinics. 2. On-line atient registration and rocessing of charges. Terminals and rinters would be installed in each articiating clinic. 3. Oeration of a clinic-oriented aointment system. Management reorts would be reared and a data base would be maintained for retrosective analysis. 4. ntegration of the system with existing hosital administrative systems. Since the hosital is starting to install an inatient hosital information system, the ore Record System must be caable of being integrated with it. Develoment of the system began in 979, and the system was in roduction use in 98. An evaluation of the system has been comleted, and the hosital is considering exanding it to suort all Outatient Deartment clinics. 37,38 Descrition of the System The rincial comonent of the ore Record System is its medical record; a samle record is shown in Fig. 8. The record contains four basic sets of data. Standard identification and registration information (e.g., age, address). This is maintained in a format comatible with the hosital administrative and business systems. 2. Active roblem list, inactive roblem list, and medication summary. This information is stored in free text, but there are lans to code the text by a comuterized coding scheme. 3. nformation extracted from other hosital systems. This includes lists of outatient visits and current aointments with any outatient clinic. 4. linic-secific data. This includes atient instructions, work release information, and other clinic-defined data. Data also are listed on a comuter-rinted encounter form. This form contains the basic registration data, a summary of the current roblem list and medications, and sace to record the rogress note and the tests and rocedures erformed. The encounter form is either rerinted (for clinics using an aointment system) or rinted on demand (e.g., for the Emergency Deartment). The registrar and rovider each uses the encounter form to record his actions. The original becomes the rogress note that is stored in the medical record; a coy is used for the entry of the charges and to udate the summary of the roblem list and medications. The system is currently in use in the Emergency Deartment (60,000 visits/year), the Primary are Walk-n linic (23,000 visits/year), the Orthoedic linics (4500 visits/ year), and the Oncology enter's Radiation Theray and Outatient units (35,000 Johns Hokins APL Technical Digest

12 JHH ORE REORD E RENT DEPARTlENT RESTRATON NFORlATlON BAl T!ORE B.. Blum -JHM nformation Systems 09/08/8 BRTHDATE V27f27 LN RAE B SEX F Ml STAT "D 228 PHONE 30_ "EDAD DATA HU"BER HA"E FRO!! OV7b TO V8 flu NO ALLER ES REORDED ff ATVE PROBms ---- Z-A BLATERAL ARrAL TUNNELS ("MRE, 04/80 4-A HYPERTENSON ("NRE, 04//80 5-A "ORBD OBESTT ("NRE, 04//80 b-a ARTHRTS L KMEE ("NRE, m05/80 -A TORN l KNEE LMENT WRTH, 06/04/8 2-A DEENERATVE JONT DSEASE LEFT LE (ERNT, 07/3/8 3-A RT FLAN< PAN (ERNT, OB/WBll NATVE AND OTHER PROBLE~ -----l DSH FRO" WL -3 ON 028S ("NRE, OS/W8 (ONTlNl OR (QlUT.JlH ORE REORD E"ERENY DEPART"ENT 3- SP R PAROTD EXSON - 09/0Bt8 ("MRE, OSZb/Sl 9- een "fd LN PROV BASO", REBEA ("MRE, 06/98 0- MEXT EN "ED APPOMmNT ON 08/07/8 ("NlRE, 06/9/ "EDATJONS ---- ALDO"ET (m!re, 04/ / 80 Z. HTl ("NRE, 04//80) 3 ELAYL (PlHRE, 04/80) 4 KL ("NRE, 04/80) 5 NALTON ("HRE, m05/s0 6 DARAYON (ERMT, 07/3/S) REENT PA TlEMT OPD YS TS een PED 0926/80 ORTH 003/80 PLAS SUR 0/780 ER 0mlSO ORTH 030/80 ER/P 05/80 ORTH V04/80 een "ED PLAS SUR V05/80 PLAS SUR VZ2SO PLAS SUR OUZ3/S EM PED 0Zl27/S ORTH 04/09/8 P. T. 05/06/8 ORTH Obl04/8 PLAS SUR 0626/8 ERNT 07/3/8 PLAS SUR 08/78 ERNT 08ZbS ffff NO APPONTPENTS fill END OF ORE REORD (0) lsplat (PRlMT (NEl PATENT Figure 8 - Samle core record. visits/year). linic-secific data have been defined by individual clinics. For the Emergency Deartment, there are instructions to the atient and information regarding the atient's disosition. This disosition information is articularly useful to the Emergency Deartment in that an automated "locator" file is maintained and the status of a atient can be determined by querying the system. For the Orthoedic linic, work release information and instructions to the atient reresent the clinic-secific comonents. n addition to the medical record, the system rovides two other services an aointment system and a charge cature system. The aointment system is designed as an inter- and intra-clinic system with online caabilities for entry, udate, and dislay of aointments. Aointments can be dislayed by atient, health care rovider, or clinic. The aointment system allows scheduling of ancillary tests and rocedures that need to be erformed at the time of the next aointment. The charge cature function allows each clinic to rocess all of its charges. The Outatient Business Office can then rocess the clinic's bills and create a tae that is inut to the billing system on the main hosital comuter. Volume 4, N umber2, 983 Evaluation The ore Record was evaluated in 982 based on data from 32,500 encounters in four clinics over a nine-month eriod. 39 (The actual eriod varied among the clinics, deending on the date a clinic came on the system.) Four factors were considered medical information, rocess integration, new functions, and cost. With resect to the medical data, it was shown that the ore Record contained information about visits to more than one clinic in over 300/0 of the cases. Such information normally would be available only through the comlete medical record. The availability of ore Records was next considered. The results were biased by the fact that there was only a limited eriod of data collection rior to the evaluation. Nevertheless, within the Emergency Deartment and Primary are Walk-n enter, over one-third of all atients entering the clinic were already in the system. (Over half the Emergency Deartment cases are walk-in atients with no current record.) Of those identified, two-thirds had a ore Record. An evaluation of the ore Records showed that they tended to contain more than one roblem and deending on the Outatient Deartment unit-less than one medication. Physician comliance in comleting the records ranged from 75% in the Emergency Deartment to 96% in the Orthoedic linic. Availability of information was a function of tye of atient and treatment; reeating atients with chronic roblems required-and had-better documentation. A brief comarison of the ore Record with the comlete medical record showed that the ore Record tended to contain more summary information. Where data were available from the medical record, of course, the documentation was more comlete. The evaluation of both rocess integration and new functions was anecdotal. t was demonstrated that charge cature and medical summary rocessing could be integrated. New functions such as a atient locator, aointment system, and management reorts could be rovided. The cost analysis focused on the cost to oerate the system. A target of $.00 er encounter was considered reasonable. The actual cost was $0.9 er encounter lus a charge of $0.5 to $0.20 for forms. Potential cost savings were identified, but realization of the savings would be deendent on reorganization within the hosital. Finally, a subjective evaluation by 2 clinicians, administrators, and suort staff in two clinics found that. The ore Record concet is easy to understand (20 of 20 resonding); 2. Provider use of the system is not time-consuming (roviders only, 0 of resonding); 3. Desite the added burden to users, the ore Record System was worth the effort (9 of 20 resonding); 4. f such a system were imlemented, there would 5

13 B.. Blum - JHM nformation Systems be an imrovement in health care delivery (8 of 2 resonding); 5. f the system were not continued, health care delivery would be adversely affected (roviders only, 9 of 2 resonding). ONLUSON The use of comuters to manage clinical information systems has undergone tremendous change'in the ast decade. t is now clear that comuters ca'n be used to control costs 40,4 and imrove atient care. 42 Equiment rices are falling and our exerience base is growing. Five years ago, few major systems could be started without external financial suort; the next five years should bring into the marketlace a broad sectrum of validated clinical information systems. The history of clinical information systems at Johns Hokins reflects the rogress in the field. Starting with some grant-suorted rojects in radiology and the harmacy, systems were ut into oerational use that continue to serve as integrated comonents of the care rocess. n the mid-seventies, new systems were initiated in the clinical laboratory, Pediatrics, Oncology, and the Outatient Deartment. Each continues to lay an essential role in the hosital's oeration. Finally, in the early 980's, the hosital committed itself to roviding a comrehensive inatient system. Future work will most certainly involve the exansion of current caabilities and networking of the systems. (A model for hosital system networking has been develoed at the Alied Physics Laboratory and is in use at the University of alifornia Hosital in San Francisco. 43,44 ) The availability of owerful desk-to comuters suggests that alications will roliferate. Portions of larger systems such as the Oncology linical nformation System will become available to suort the management of other chronic diseases. Finally, the ability to link and merge data and algorithms among systems will dramatically imrove utility. Thus, one can exect even more exciting develoments in the alication of comuters to medical care. REFERENES and NOTES R, A. Jydstru and M. J. ross, "ost of nformation Handling in Hositals," Health ServoRes (Winter 966). 2Reort by the U.S. eneral Accounting Office, omuterized Hosital nformation Systems Need Further Evaluation to Ensure Benefits from Huge nvestments, AFMD-8-3 (Nov 980). 3D. A. B. Lindberg, The rowth of Medical nformation Systems in the United States, Lexington Books, D.. Heath and o., Lexington, Mass.,. 98 (975). 4R. E. ibson and R. J. Johns, A Study of the Management of the Johns Hokins Medical School and the Johns Hokins Hosital, JHU nternal Reort (25 May 97). SR. E. Miller,. L. Steinbach, and R. E. Dayhoff, "A Hierarchical omuter Network An Alternative Aroach to linical Laboratory omuterization in a Large Hosital," in Fourth Ann. Sym. on omuter Alication in Medical are, EEE 80H 570-, (980). 6B.. Blum, R. E. Miller, and R. E. Lenhard, Jr., "Distributed Laboratory Data Processing in a Large Hosital The Johns Hokins Exerience," in nformatics in Health Facilities Vol.. omuterization and Automation, M. Rubin, ed., R Press, nc., Boca Raton, Fla. (in ress). 6 7R. H. Drachman, M. J. O'Neill, and K. V. Ledford, "mlementation of a Medical nformation System in an Ambulatory are Setting," in Third Ann. Sym. on omuter Alications in Medical are, EEE 79H 489-3, (979). 8R. H. Drachman, M. J. O'Neill, and K. V. Ledford, "omuter Review of Ambulatory are - The "ARE " System," in Fifth Ann. Sym. on omuter Alications in Medical are, EEE 8H 696-4, (98). 9J.. Mabry et 0., "A Prototye Data Management and Analysis System (LNFO) System Descrition and User Exerience," in MEDNFO 77, North-Holland Publishing o., New York, (977). 0J. B. Johnson, Jr., et 0., "Five Years' Exerience with the LNFO Data Base Management and Analysis System," in Sixth Ann. Sym. on omuter Alications in Medical are, EEE 82H 805-, (982). D. W. Simborg, "The Develoment of a Ward nformation-management System," Meth. nform. Med. 2, 7-26 (973). 2D. W. Simborg et 0., "Ward-nformation Management System An Evaluation," omut. Biomed. Res. 5, (972). 3H. J. Derewicz and D. D. Zellers, "The omuter-based Unit Dose System in the Johns Hokins Hosital," Am. J. Hos. Pharm. 30, (973). 4D. W. Simborg and H. J. Derewicz, "A Highly Automated Hosital Medication System, Five Years' Exerience and Evaluation," Ann. ntern. Med. 83, (975). S. S. Wheeler and D. W. Simborg, "The Johns Hokins Radiology omuter Reorting System," Electromedia 2-3,04-08 (975). 6BM Health are Suort/ DLl, Patient are System TerminalOerators uide. Program Number 5796-ANY, SH (Nov 977). 7B.. Blum and R. E. Lenhard, Jr., "A Prototye linical Patient Management System," in AM Tech. Sym., Washington, D..,. 5-2 (Jan 977). 8B.. Blum et 0., "A linical nformation Dislay System," in First Sym. on omuter Alications in Medical are, EEE 77H (977); rerinted in V. Sondak and H. Schwartz, omuters and Medicine, Artech House, nc., Dedham, Mass. (979). 9B.. Blum and R. E. Lenhard, Jr., " Design of an Oncology linical nformation System," AM 77,0-07 (Oct 977). 20B.. Blum et 0., "The Johns Hokins linical nformation Systems A Study in MUMPS Productivity," in 978 MUMPS Users' rou, ollege Park, Md.,. 6-7 (Jun 978). 2 B.. Blum and R. E. Lenhard, J r., "An Oncology linical nformation System," omut. Mag., (Nov 979). 22 R. E. Lenhard, Jr., et 0., "The Johns Hokins Oncology linical nformation System," in Sixth Ann. Sym. on omuter Alications in Medical are, EEE 82HB (982). 23 B.. Blum, "A Tool for Develoing nformation Systems," in Automated Tools for nformation Systems Design, H. J. Schneider and A.. Wasserman, eds. North-Holland Publishing o., New York, (982). 24B.. Blum and. W. Brunn, " mlementing an Aointment System with TEDUM," in Fifth Ann. Sym. on omuter Alications in Medical are, EEE 8H (98). 2s B.. Blum, "MUM PS, TEDUM and Productivity," in First EEE omuter Society nt. on on Medical omuter Science/ omutational Medicine (MEDOMP),EEE 82 H (Se 982). 26B.. Blum and R. E. Lenhard, Jr., " Dislaying linical Data for Decision Making,". lin. Eng. 8,6-68 (983). 27c. J. McDonald, "Protocol-Based omuter Reminders, The Quality of are and the Non-Perfectability of Man," N. Engl. J. Med. 295, 24 (976). 28 R. K. Halse et 0., "omuterized Medication Monitoring System," Am. J. Hos. Pharm. 33, (976). 29 A. M. Weissman et 0., "omuter Suort of Pharmaceutical Services for Ambulatory Patients," Am. J. Hos. Pharm. 33, 7-75 (976). 30 D. D. Wirtschafter, J. T. arenter, and E. Mesel, "A onsultant-extender System for Breast ancer Adjuvant hemotheray," Ann. ntern. Med. 90, (979). 3 This lan is for the first da y of the second cycle of theray. The comment in the box warns the hysician that one of the drugs to be administered, Bleomycin, has a maximum cumulative dose. The sections on measurable tumor sites, recent clinical findings, and chemotheray history are automatically udated from the clinical data base each day. The actual drug dose is comuted on the basis of a body surface area (BSA, see to of lan) of.4. The general lan includes a list of all tests and rocedures required for the rotocol. Twelve tests are to be ordered from this lan. eneral hysicians comments are listed at the beginning of the lan. Messages related to a rocedure are li sted in the rocedure ordering section. Most messages tend to be reminders. The next version of the system will be able to generate warning messages based uon data trends. 32 B.. Blum, R. E. Lenhard, Jr., and E. E. Mcolligan, " Protocol Directed Patient are Using a omuter, in Fourth Ann. Sym. on omuter Alications in Medical are, EEE 80H (980). Johns Hokins APL Technical Digest