42nd and Dewey Avenue, Omaha, Nebraska 68105. Funded by NCHSR Grant .... Rotating people at the checkout desk. 6. Staffmorale. 1. Supervision of ...
ClnicFunctdon and Computered Ambulatr Records: AConamrent Study with Conventional Records Nathaniel Givner, PhD
James R. Campbell, MD
Kash
CariesB Seelg MD Robest Wigton, MD
th
Pa4 PhD
Thomas G.Tape, MD
Department of Internal Medicine, University of Nebraska Medical Center 42nd and Dewey Avenue, Omaha, Nebraska 68105 Funded by NCHSR Grant #HS-04949-01
Specific objectives of the study were: (1) To assess the evolution of staff attitudes toward the use of the computerized record during installation, and
ABSTRACT Controlled studies of computerized ambulatory information systems are rare. As part of an evaluation of the effects of COSTAR on clinic function, we divided our resident teaching clinic into a study group with access to COSTAR and a control group allowed access only to conventional medical records. We sampled staff attitudes toward use of the computer and did detailed time studies of clinic patient flow. Staff attitudes reflected a high degree of acceptance, favoring COSTAR over conventional records. This was primarily related to improvement in telephone management and demand care. House staff never became facile users of COSTAR because of infrequent clinic sessions. Clinics assigned to COSTAR experienced somewhat longer waiting times due to an increased workload and training effects. Installation of computerized records should prompt a careful evaluation of expected benefits.
(2) To determine the impact of computerized records on the clinic encounter vis4a-vis the duration and conduct of the patient visit. BACKGROUND Utilizing an enhanced version of COSTAR V, computerized medical records were installed in the teaching clinic of the Internal Medicine Department of the University of Nebraska College of Medicine. The installation of COSTAR was in early 1985. Complete medical records features were functioning by September of 1985. Resident physicians were informed of the study goals and selected for access to COSTAR. House staff have a clinic at this site for a half day every other week. Access to clinic records depended on their clinic week. House staff meeting on one week were designated as a control group. These house staff were allowed access only to the paper record. Alternate week residents were designated as the study group, and could access both the paper and computer records.
INTRODUCIION Although automated ambulatory records systems have been installed in increasing numbers of sites throughout the United States, formal evaluations of their function are rare. A review of systems in use sponsored by the National Center for Health Services Research in 1981 [1] made this observation. Comparing their data to those from a similar study six years earlier [2], the investigators observed that many technical problems had been resolved, maldng these systems more reliable and acceptable. However, issues such as acceptability to physicians and effects on clinical practice remained unaddressed.
In contrast, clinic staff, including nurses and administrative personnel, had partial access to the computer record from the very beginning. More specifically, clinic staff were allowed use of the computer record only for patients of study group residents. For all other patient problems, staff ordered the chart from medical records and used the conventional hospital record for patient care. After a patient was seen in clinic, progress notes were typed into COSTAR and a duplicate was filed in the hospital chart. A program interlock prevented access to computer information for patients of the control physicians. This design continued until December of 1986 when vociferous complaints by the clinic staff caused us to relax this study restriction. We then allowed nursing and clerical staff to read computer notes for all patients if the patient was not physically in clinic seeing a 'control' physician. Through the completion of the study period in June of 1987, house staff continued with their initial study assignment and records access.
Contrasting with a poor understanding of how computerized records change medical practice is a growing sense that summarizations and reminders generated by computers can favorably affect the practice of ambulatory medicine [3,4]. This dichotomy emphasizes the need for study of how the computer changes ambulatory medicine. As a part of a larger study of the CQmputer STored Ambulatory Record (COSTAR)*, we thought it important to assess the changes induced by the computer in our practice.
745
0195-4210/88/0000/0745$01.00 X 1988 SCAMC, Inc.
METHODS
Data Analysis Group differences on measures of clinic efficiency were analyzed by t-test comparisons. Analysis of the staff questionnaire focused on the 15 staff members who completed a questionnaire at all three time periods. On all items, a onefactor analysis of variance with repeated measures was used to examine whether staff attitudes significantly varied over time. Using the error term from the repeated measures analysis, comparisons were made to specifically contrast attitudes at each of the three time periods. RESULTS Marked preferences and opinions expressed by clinic staff in questionnaire responses are summarized in Table 1. In regard to general opinion of clinic function, significant time effects were found on questions relating to a) the busyness of the clinic and b) the disruptive influences of nearby clinics on the function of General Medicine. Results reflected an increase in patient volume and a move towards shared clinic resources which were occurring throughout the period of the study. They serve to highlight the changing nature of ambulatory care through this time, and emphasize the desirability of controlled studies. Contrasts on other items bearing on staff attitudes toward the clinic were not statistically significant.
Staff Questionnaire Relying in part on discussion with clinic staff, a questionnaire was developed in order to 1) monitor staff opinions regarding clinic function and 2) assess staff attitudes concerning the COSTAR medical record. Clinic staff were asked to rate, on a 5-point scale, their preference for the COSTAR record on 13 aspects of medical records function. These included: accuracy of information, organization of information, time required to find information, ability to find current medications, ability to find lab and x-ray, ability to find health maintenance items and record availability. On an additional 13 items, they were also asked to give their impression of the extent that the COSTAR record benefited such clinic functions as scheduling of appointments, handling of phone calls, managing prescription reflls, and patient waiting room time. Additional questions were added to assess more personal aspects of working in the clinic. These related to workload, job performance and satisfaction, and definition of responsibilites or role. Questionnaires were distributed to staff in November, 1985 (baseline), October, 1986 (study inception), and June of 1987 (study completion). Analysis of this data focused on the 15 staff members who completed a questionnaire at all three time periods. lVZCUbUIVARZ1LZS
JL %4LLUM. LfLIIUMMY
The impact of the computerized record on efficiency of the patient encounter was examined in the context of patient waiting times and speed of patient flow through clinic. Data relevant to this objective were collected from April through June, 1987, on 684 patient visits. At this time, only residents in the study group had access to the COSTAR record. Clinic staff were prevented from using COSTAR during control group clinics although they had access at other times for telephone management and demand care. Residents in the control group were limited to the paper record. Using a stopwatch, a trained observer who was blinded to group assignments recorded the tming of events occurring during patient visits. Patient arrival time was taken from computer scheduling logs which were kept for all patients.
Table 1 Staff Attitudes with Consistent Assessments at all Tlme Periods I. Strong preference/benefit at all time periods For thceOSA Mecdia Record 1. 2. 3. 4. 5. 6.
FFmding record when needed Fmding most recent patient encounter
Readability
Finding Current Problems Time required to get infomati Time spent answering patient billing questions 7. Handling phone calls in clinic 8. Managng prescription refills
From these data, a series of time intervals were calculated for each patient visit:
For the Pper Meical Recrd
1. Fnding hospitalization summary
* waiting room time * nurse check-in * waiting for house staff * history and examination * waiting for attending review * attending review with patient * nurse preparation for checkout * checkout at discharge desk
II. Staff attitudes/Opinions that were Rated Consistently High or Low at all Time Periods.
Nectivcb Evaluated Across Time 1. 2. 3. 4. 5. 6.
In order to evaluate patient flow in the context of the intensity of care, a chart review was conducted on 896 encounters that occurred during the time study. The number of procedures performed, number ofclinical problems addressed and the intensity of service were tabulated from billing data.
Hurried in clinic duties Patient flow in clinic Having to do thing at the last minute Back-up nursing services Rotating people at the checkout desk Staff morale
1. Supervision of residents 2.. Communication with supervisor
746
Concerning the care provided by the housestaff during the
rs preferences for records use Table 1 additionally su that we repeatedly observed. Most of these attitudes changed in only minor ways over the three observation periods. Staff consistently and strongly preferred all aspects of the COSTAR record that were operational during our project. By way of control, the only aspect of the paper record that was preferred was hospitalization summares-a type of informaton that was not duplicated in the COSTAR record. The only reservation mentioned by the staff was an expressed concern about patient waing times.
conduct of the time study, the study group had a total of 469 encounters during 108 separate clinic sessions, and the control group had 427 during 102 sessions. A physician in clinic for 4 hours was considered to be one session. The study group recorded 790 separate problems in their clinic notes, and the control group 760. Table 2 summarizes the services that were billed for these encounters. Totals of selected clinic procedures are also included in order to contrast the level of clinic activity undertaken by the study and control physicians.
Significant (p
