Concepts, knowledge, and language in health-care information systems.
Explore the source record for details and available documents.
Artificial Intelligence↗
Biomedical subjects
C Safran
Publications and source records attributed to C Safran.
Searching for answers on a clinical information system.
We examined observational data on the use of a clinical computing system in an effort to develop an empirical model of clinicians' information needs. Clinicians turn to information systems most often to review the results of diagnostic studies. After that, clinicians turn to information systems most often for communication. Bibliographic retrieval is the third most frequent reason for use, but is an order of magnitude less frequent than either results inquiry or electronic mail. Secondary retrieval of aggregate data from a clinical database is two orders of magnitude less common than primary retrieval. In a study at our hospital, clinicians gained access to an information resource during 16% of all general medical visits of patients with human immunodeficiency virus (HIV) infection while the patient was present. Insofar as our narrow focus on the primary care of HIV infection models the larger context of clinical medicine, clinicians are most often looking at information relating to therapeutics. Whatever knowledge structures are adopted, they must accommodate this observed need for therapeutic information.
Diagnosis, Computer-Assisted↗
Effect of computer-based alerts on the treatment and outcomes of hospitalized patients.
BACKGROUND: Hospital computing systems play an important part in the communication of clinical information to physicians. We sought to determine whether computer-based alerts for hospitalized patients can affect physicians' behavior and improve patients' outcomes. METHODS: We performed a prospective time-series study to determine whether computerized alerts to physicians about rising creatinine levels in hospitalized patients receiving nephrotoxic or renally excreted medications led to more rapid adjustment or discontinuation of those medications, and to determine whether such alerts protected renal function. RESULTS: Laboratory data were observed for 20,228 hospitalizations, with documentation of 1573 events (instances of rising creatinine levels during treatment with a nephrotoxic or renally excreted drug). During the intervention period, doses were adjusted or medications discontinued an average of 21.6 hours sooner after such an event (P < .0001). For patients receiving nephrotoxic medications during the intervention period, the relative risk of serious renal impairment was 0.45 (95% confidence interval, 0.22 to 0.94) as compared with the control period, and the mean serum creatinine level was 14.1 mumol/L (0.16 mg/dL) lower on day 3 (P < .01) and 25.6 mumol/L (0.29 mg/dL) lower on day 7 (P < .05) after an event. Forty-four percent of physicians who responded to a questionnaire said that the alerts had been helpful in the care of their patients, whereas 28% found them annoying. Sixty-five percent wished to continue receiving alerts. CONCLUSIONS: Computer-based alerts regarding patients with rising creatinine levels affect physician behavior, prevent serious renal impairment, preserve renal function, and are accepted by clinicians.
Acute Kidney Injury↗
Design and processing issues for the health care professional workstation: summary and recommendations.
The health care professional workstation is a window to a heterogeneous set of functions that need to be ported to various types of hardware for different types of users. Processing aspects, which belong with the interface, storage and communication, to the architectural dimension of the workstation should not be considered outside of a more comprehensive model including the user, function and time dimensions. Several recommendations, which could provide the appropriate environment for development and help maintain a viable approach as the variables change along the four dimensions of the workstation model, are proposed. They concern the adaptation of the enterprise infrastructure to integrate workstations, the need for a reference architecture for health care workstation development and recommendations for the evaluation and dissemination of results.
Computer Communication Networks↗
Defining clinical 'workstation'.
Interest in the physician's workstation has increased, yet often seems to focus on technological issues. At Boston's Beth Israel Hospital, the Center for Clinical Computing includes heavily used clinical workstations. Their evolution over the past 20 years suggests design criteria: the workstation must be patient-centered, the interface must be uniform, and data acquisition must be addressed at a system level. However, it is clinical function that really defines a workstation. The workstation should do the following: display patient information rapidly and flexibly; assist with administrative tasks; facilitate communication; and provide four important types of decision support: access to literature, access to databases, clinical calculation, and 'synthetic vision,' or different views of patient data. The solutions to our healthcare problems are not in 'workboxes' we can buy, but in creative approaches we can imagine. We need a patient-centered infrastructure and a reduced workload for the clinician-perhaps a 'worklesstation'.
Computer Communication Networks↗
Ribotyping differentiates relapse from reinfection in the treatment failures of Escherichia coli urinary tract infections in children.
Analysis of restriction fragment-length polymorphism of ribosomal DNA regions (ribotypes) was used as an epidemiologic tool to compare 25 pre- and posttreatment strains obtained from 12 patients treated with either cefpodoxime proxetil or amoxicillin-clavulanic acid. Ribotyping is a promising method to differentiate relapse from reinfection in the treatment failures of Escherichia coli urinary tract infections.
Amoxicillin↗
Five versus ten days treatment of streptococcal pharyngotonsillitis: a randomized controlled trial comparing cefpodoxime proxetil and phenoxymethyl penicillin.
A total of 220 adults and children > 10 years old (mean 29.5 +/- 11.7 years) with pharyngitis/tonsillitis were randomized to receive either cefpodoxime proxetil 100 mg bid for 5 days (n = 113) or phenoxymethyl penicillin, 600 mg tid for 10 days (n = 107). At the end of treatment of the 166 evaluable patients, a satisfactory clinical response was obtained in 85/88 (96.6%) patients treated with cefpodoxime proxetil and in 75/78 (96.1%) treated with phenoxymethyl penicillin. Group A beta-hemolytic streptococci (GABHS) eradication was similar in both groups: 79/82 (96.3%) patients in the cefpodoxime proxetil group and 64/68 (94.1%) patients in the phenoxymethyl penicillin group. At follow-up (20-30 days after the end of treatment) the GABHS eradication persisted in 67/72 (93.1%) patients treated with cefpodoxime proxetil and in 56/61 (91.8%) patients treated with phenoxymethyl penicillin. Significantly better compliance (p < 0.01) was noticed with the cefpodoxime proxetil regimen compared with the phenoxymethyl penicillin regimen, with only 2/110 (2%) poorly compliant patients in the cefpodoxime proxetil group vs 17/104 (16%) in the phenoxymethyl penicillin group. Thus, the shorter duration of therapy, in conjunction with demonstrated clinical and bacteriological efficacy that is equivalent to standard therapy, makes cefpodoxime proxetil an acceptable alternative for the treatment of GABHS pharyngitis/tonsillitis.
Adolescent↗
Twice daily cefpodoxime proxetil compared with thrice daily amoxicillin/clavulanic acid for treatment of acute otitis media in children.
A total of 260 children, 3 months to 11 years old (median age 24 months), with acute otitis media (AOM) received either cefpodoxime proxetil (CP) 8 mg/kg/d b.i.d. or amoxicillin/clavulanic acid (ACA) 40/10 mg/kg/d t.i.d. for 8 days. A significant difference in clinical cure rates was observed between the CP group 71/118 (60%) and the ACA group 42/105 (40%), p = 0.003. At the follow-up visit (20-30 days after the start of treatment), significant advantages were recorded with the CP vs. ACA therapy, in terms of satisfactory clinical response [90/111 (81%) vs 60/94 (63.8%), p = 0.005] residual middle ear effusion (14.4% vs 28.7%, p = 0.01) and normal tympanometry (78% vs 61.4%, p = 0.017). Compliance and adverse event frequency were the same in both treatment groups. The higher clinical cure rate and equivalent safety profile of CP indicates that it is an acceptable alternative to ACA for the treatment of AOM in children.
Acute Disease↗
Online medical records save time & energy.
Explore the source record for details and available documents.
Ambulatory Care Information Systems↗
Closing the loop of patient care--a clinical trial of a computerized discharge medication program.
A frustrating time for hospitalized patients and their primary care providers is after discharge from the hospital, because of changes in patients' medications. We developed a computer program to improve the discharge process, by providing guidance to the physician writing the prescriptions, offering educational material to the patients, and providing electronic notification of medication changes to the primary care providers. During a one-year clinical evaluation of this system, in which use of the program was voluntary, 1000 patients were discharged through the program. House officers tended to use the program more often for patients who were older and in the hospital longer. Both house officers and primary care physicians found the program extremely useful, and the process took no longer than the manual method of creating discharge medication lists. Patients who were discharged using this program may have had better adherence to medication regimens. We conclude that computer-assisted compilation of a discharge medication list is a useful method for improving the discharge process.
Boston↗
Be well!: a computer-based health care interview for hospital personnel.
We developed a computer-based health care interview for employees at an urban teaching hospital. The interview is part of the hospital-wide, integrated Center for Clinical Computing (CCC) system at the Beth Israel Hospital in Boston, and is available on 1500 terminals located throughout the hospital. The program emphasizes medical problems and patterns of living for which behavioral change is considered desirable. Conducted in private and with protection of confidentiality, the interview offers information about people and facilities available to help members of the hospital community seek better health. Between May 15, 1990, and May 14, 1992, a total of 1281 people completed the interview, and 84% indicated an interest in one or more of the health-related programs offered by the hospital. That these are difficult times in America was illustrated by the unhappiness reported by 43% of the participants, including 5% who indicated that life sometimes seemed not worth living. We hope they availed themselves of the opportunity offered by the interview to obtain help for their problems, and that other hospitals will do what Beth Israel Hospital is doing to promote well-being among its employees.
Adult↗
An electronic medical record that helps care for patients with HIV infection.
We have built a clinical workstation to help doctors and nurses care for patients with HIV infection. This knowledge-based medical record system provides medication alerts, reminders about primary care, and on-line information to support the care of patients with HIV infection. We are conducting a controlled clinical trial of this computer system in a single practice setting, which consists of 18 staff physicians, 13 nurses, and 113 residents, who cooperatively practice in four teams. Two teams of physicians are assigned to an intervention group and two teams to a control group. This paper reports preliminary results from the first year of study, January 15, 1992, through January 14, 1993. During this period 274 patients with HIV infection were followed by the general medical practice--130 in a control group and 144 in an intervention group. Physicians in the intervention group more rapidly and more completely followed primary care guidelines than did physicians in the control group. Patients in the intervention group had 2476 ambulatory or emergency visits (17.2 visits per patient) compared with 1882 visits (14.5 visits per patient) for the control patients (p < 0.01). There were 101 hospitalizations for 51 patients in the intervention group (an admission rate of 0.7) compared with 104 admissions for 54 patients in the control group (an admission rate of 0.8) (p = NS). There were 8 deaths in the intervention group (5.6%) compared with 13 (10%) in the control group (p = NS).(ABSTRACT TRUNCATED AT 250 WORDS)
Artificial Intelligence↗
Use of electronic mail in a teaching hospital.
Electronic mail has been in use for almost 20 years at Boston's Beth Israel Hospital and is an integral part of the clinical information system. Through a study of usage patterns during a one-week period, we found that 1247 persons sent 7482 messages to 1302 different recipients. Each category of user (attending physician, house officer, nurse, etc.) sent the most e-mail to others of the same user category. Through an electronically administered questionnaire, we found that self-reported usage patterns had a high correlation (r = 0.6) with measured use. Sixty-six percent of respondents used e-mail daily or weekly, and 58% used it for issues of patient care; nearly all users found this useful for communicating about patient care issues. Ninety percent of respondents felt e-mail made their lives easier and 61% felt it had a humanizing influence on their lives. We conclude that the e-mail system is well-utilized by clinical personnel and felt to be useful in both patient care and nonpatient care situations.
Adult↗
Real and imagined barriers to an electronic medical record.
We developed an electronic medical record for ambulatory patients as part of the integrated clinical information system at Beth Israel Hospital. During the four years since it was installed, clinicians have entered 76,060 patient problems, 137,713 medications, and 33,938 notes. Residents, who had to type notes in themselves, entered 49.5% of their notes into OMR. Several factors that we had predicted would be barriers to an electronic medical record, such as clinician reluctance to type or perform data entry, have not proved to be significant problems. Other anticipated barriers, such as difficulties with dual charting on paper during transition to an electronic medical record, have been realized. The major unexpected barrier that has been encountered is increased clinician concern about the privacy and security of full text notes relative to other data elements in the clinical information system. We have attempted to modify the electronic medical record so as to overcome some of these barriers.
Attitude to Computers↗
Computer-based interview for screening blood donors for risk of HIV transmission.
OBJECTIVE: To test the ability of a computer-based interview to detect factors related to the risk of the human immunodeficiency virus (HIV) among potential blood donors and to determine donor reactions to the use of the interview. DESIGN: A comparison of the rate of detection of HIV-related factors elicited by a computer interview with that obtained by standard American Red Cross procedures for assessment of donor suitability, including a randomized crossover trial in which the order of the two methods was reversed. Information obtained by the computer was not available to influence the use of blood components for transfusion. SETTING: The computer interview was administered to donors at an American Red Cross blood donor center and at a mobile blood drive at a hospital. SUBJECTS: Consecutive sample of 294 male and female blood donors 18 to 75 years of age. MAIN OUTCOME MEASURES: Subjects' responses to the computer-based interview as well as responses to the standard Red Cross written questionnaires and face-to-face interviews were used for donor assessment. RESULTS: The interview took an average of 8 minutes to complete. From among 272 donors who provided complete data, the computer identified 12 donors who reported either behaviors associated with a risk of HIV acquisition or symptoms compatible with the acquired immunodeficiency syndrome; none of these donors had been so identified either by routine written questionnaires or by face-to-face interviews used to screen potential blood donors. Only one of the 12 identified donors used the confidential unit exclusion procedure to prevent use of his donated unit. The rate of elicitation of HIV-related factors by the computer interview was 12 (4.4%) of 272 (95% confidence interval [CI], 2.3% to 7.6%), compared with two (0.13%) of 1536 (95% confidence upper bound, 0.28%) using the standard Red Cross procedure (P less than .0001). Tests for antibodies to HIV were negative in blood samples from all of the 272 subjects studied. The subjects enjoyed the computer interview and judged it to be more private than the standard donor assessment method. They also predicted that donors would be more honest with the computer interview than with a human interviewer. CONCLUSIONS: Computer-based screening elicits more HIV-related factors in the health histories of blood donors than do the standard questionnaire and interviewing methods currently in use. Computer-based screening is also acceptable to blood donors.
Adolescent↗
Serum albumin level on admission as a predictor of death, length of stay, and readmission.
We studied the serum albumin level within 48 hours of hospitalization for acute illness to predict in-hospital death, length of stay, and readmission in 15,511 patients older than 40 years. Patients with low serum albumin levels (less than 34 g/L), who made up 21% of the population, were more likely to die, had longer hospital stays, and were readmitted sooner and more frequently than patients with normal albumin levels. The in-hospital mortality was 14% among patients with low albumin levels, as compared with 4% among patients with normal levels. Although the serum albumin level was a nonspecific marker, it was a stronger predictor of death, length of stay, and readmission than age. We conclude that the serum albumin level on admission is an important variable that should be incorporated in severity-of-illness measures based on physiologic indexes.
Adult↗
Frequency of low serum high-density lipoprotein cholesterol levels in hospitalized patients with "desirable" total cholesterol levels.
Because the National Cholesterol Education Program guidelines suggest that levels of total serum cholesterol less than 5.17 mmol/liter (200 mg/dl) are "desirable," we performed a retrospective observational analysis to determine the prevalence of coronary artery disease (CAD) in patients with total cholesterol less than 5.17 mmol/liter (200 mg/dl) and the prevalence of total cholesterol less than 5.17 mmol/liter (200 mg/dl) in patients with CAD by angiography. Cholesterol levels less than 5.17 mmol/liter (200 mg/dl) were found in 1,084 of 2,535 patients (42%) having cholesterol measured on hospital admission; 690 of these 1,084 (64%) had CAD. These patients were mostly men, had a family history of premature CAD, and 60% (414 of 690) had high-density lipoprotein (HDL) cholesterol less than 0.90 mmol/liter (35 mg/dl). In a separate group of patients defined from the same admission population but having angiographically established CAD, 32% (424 of 1,197) had a total cholesterol less than 5.17 mmol/liter (200 mg/dl), 59% of whom (252 of 424) had HDL less than 0.90 mmol/liter (35 mg/dl). An analysis of persons admitted electively for angiography (to exclude any effects of hospitalization per se on serum lipids) revealed a similar proportion of persons with total cholesterol less than 5.17 mmol/liter (200 mg/dl) (35%), CAD (82%), and HDL less than 0.90 mmol/liter (35 mg/dl).(ABSTRACT TRUNCATED AT 250 WORDS)
Cholesterol↗
Using routinely collected data for clinical research.
Clinical research involving prospective data collection in randomized controlled trials is not always feasible. Increasingly, hospitals are developing large clinical databases that are waiting to be mined. We have developed a computer program, ClinQuery, that facilitates such exploration and analysis. We have also shown in a series of studies that the use of clinical data is a powerful tool in health services research. In some cases, we have shown that coded data are inaccurate and that alternative clinical data are preferable. In other cases, a combination of clinical data and coded discharge diagnoses is preferable.
Boston↗