Methotrexate and histologic hepatic abnormalities: a meta-analysis.

STUDY OBJECTIVE: To determine the risk of liver toxicity from the long-term administration of methotrexate in patients with rheumatoid arthritis or psoriatic arthritis. DESIGN: A meta-analysis of 15 studies examining the relationship between long-term, low-dose methotrexate administration and biopsy evidence of liver fibrosis. PATIENTS: A total of 636 patients from 15 studies. RESULTS: The incidence of progression of liver disease (defined as worsening of at least one grade on the histologic classification of Roenigk) among 636 patients was 27.9% (95% confidence intervals 24.3 to 31.6). The rate of progression of liver disease in the 15 studies was associated with the cumulative dose of methotrexate (p = 0.01). Patients on average had a 6.7% (95% confidence intervals 2.1 to 11.4) chance of progressing at least one histologic grade on liver biopsy for each gram of methotrexate taken. The overall incidence of advanced pathologic changes on liver biopsy (grades IIIB or IV) among 636 patients was 5.0% (95% confidence intervals 3.5 to 7.0). The development of advanced histologic changes was not associated with the cumulative dose of methotrexate (p = 0.08). Patients who according to their history were heavy drinkers (at least 100 g of alcohol per week) were more likely to have advanced changes on liver biopsy (17.8% versus 4.5%, p = 0.0003) and to show histologic progression (73.3% versus 25.9%, p = 0.0002). Patients with psoriasis were more likely than patients with rheumatoid arthritis to have advanced changes (7.7% versus 2.7%, p = 0.003) and histologic progression (33.1% versus 24.3%, p = 0.02). CONCLUSIONS: The risk of liver toxicity in patients undergoing long-term, low-dose methotrexate therapy is substantial, and that risk increases with the total cumulative dose and with heavy consumption of alcohol. Heavy users of alcohol should not receive long-term methotrexate therapy. For most patients who are not heavy users of alcohol, liver biopsies should be done periodically to monitor for the occurrence of liver toxicity.

Evaluation methodology for ambulatory care information systems.

The central purpose of an ambulatory care information system is to communicate information to the practitioner to facilitate clinical decision making. The clinical decision can be considered the dependent output variable in a process in which the information system, the patient, clinician characteristics and the environment are the independent input variables. Evaluation methodologies must consider there relationships. Approaches using patients outcomes are problematic because of indirect relationship between the information system and patient outcomes, which limits both sensitivity and validity. A process measure technique that focuses on the clinical decision directly as the measure of output could be appropriate if the represented a generic sampling of clinical decisions made in ambulatory care. A new method under development based on an information theory concept may be more widely applicable than currently available methods.

Choosing the correct unit of analysis in Medical Care experiments.

The statistical methodology of health research experiments published in Lancet, the New England Journal of Medicine, and Medical Care between 1975 and 1980 for the presence or absence of an error of experimental design and analysis was examined. The error is the result of inappropriately using patient-related observations as the unit of analysis to form conclusions about provider behavior or outcomes determined jointly by patients and providers. The error was present in 20 of 28 (71%) health care experiments addressing an issue of health provider professional performance. Its usual effect is to increase erroneously the power of an experiment to detect differences between experimental and control groups. It is likely that this type of error could be avoided by the explicit and prospective definition of hypotheses and the populations to which they are intended to pertain.

A controlled experiment to evaluate the use of a time-oriented summary medical record.

A randomized single-blind experiment was done in a medical subspecialty clinic in order to determine whether a flow-sheet type of summary medical record could validly serve as a means to communicate clinical information in the absence of the traditional medical record. Two groups of outpatient physician-patient encounters were compared: In the 68 study encounters (Group S), physicians were given a flow-sheet summary record with the option to receive the standard medical record if they desired; in the 27 control encounters (Group C), physicians were given the standard medical record plus the flow-sheet summary record. Fifty-nine per cent of study-group physicians did not choose to receive the full medical record. The study group was found not to differ (p = 0.013) from controls significantly with regard to the follow-up of clinical information as measured by pre- and post-encounter chart review. Physician providers in the study group were unable to detect by retrospective chart review overlooked clinical information with greater frequency than control group providers. We conclude that a flow-sheet type of summary medical record can serve as the sole source of clinical information in a substantial number of outpatient follow-up encounters in a medical subspecialty clinic without deterioration in the communication of clinical information.

A computerized summary medical record system can provide more information than the standard medical record.

We assessed the ability of a computerized outpatient medical record (MR) system, the Summary Time-Oriented Record (STOR), to communicate information to clinicians in two randomized single-blind studies. In the first study, physicians were better able to predict their patients' future symptom changes and laboratory test results from outpatient visits to an arthritis clinic when STOR was added to the standard MR than when the standard MR was used alone. In a separate study, the removal of the standard MR did not result in important decrease in the physicians' ability to predict their patients' symptoms and laboratory test results if they had the option of using the full paper record when they thought they needed it. In 134 (26%) of 514 visits, the physicians exercised this option. We conclude that for outpatient visits, the computerized record system STOR operationally added information to that supplied by the full paper MR. This improved flow of information could improve the clinical decision process.

Validity of a vasculitis activity index for systemic necrotizing vasculitis.

OBJECTIVE: To establish the validity of an index designed to measure activity in systemic necrotizing vasculitis (SNV). METHODS: The Vasculitis Activity Index (VAI) was designed to incorporate appropriately weighted clinical measurements that reflect disease activity in SNV. We performed a pilot study to guide the modification and subsequent testing of the initial design. The data necessary to calculate the VAI are direct ratings by a clinical observer of the degree of activity in 9 organ systems and 3 indirect measures of vasculitis activity. These data are recorded on 0-4 visual analog scales. Physician's global assessment (PGA) is used as the "gold standard" measurement of disease activity. The VAI was validated using 2 independent data sets: the questionnaire data set, derived from test case histories ("paper cases") sent to 100 practicing rheumatologists, and the clinic data set, obtained from use of the VAI in 204 regular care visits of 74 patients with SNV. RESULTS: The VAI correlated highly with the PGA: Pearson's correlation coefficient R = 0.84 (95% confidence interval [95% CI] 0.80-0.88) for the questionnaire data set, and R = 0.92 (95% CI 0.90-0.94) for the clinic data set. The mean of the interobserver coefficients of variation for the test case histories was lower for the VAI than for the PGA (mean difference 0.45; P = 0.002), indicating that the VAI has less interobserver variation than does the PGA. The change in VAI between clinic visits for individual patients correlated highly with the change in PGA (R = 0.88, 95% CI = 0.83-0.91). The VAI data collection form requires about 1 minute to complete, including computation of the score. CONCLUSION: The VAI is a valid measure of vasculitis activity that correlates highly with the PGA. In addition, the VAI has less interobserver variation than the PGA and has a high level of sensitivity to change over time. Additional testing of the VAI appears warranted.

Local area networks and the hospital.

Hospital information systems are characterized by their complexity of individual functions, heterogeneity of functions, and dependence upon integration. A distributed computerized information system is well suited to meeting the needs of hospitals. A local area communications network (LACN) removes a major impediment to the use of distributed systems. An advanced microprocessor-based LACN using fiberoptic communications has been developed by the Applied Physics Laboratory of The Johns Hopkins University and has been implemented at the University of California, San Francisco Hospital.

A prototype generalized network technology for hospitals: initial implementation.

A demonstration implementation of a distributed data-processing hospital information system using an intelligent local area communications network (LACN) technology is described. This system is operational at the UCSF Medical Center and integrates four heterogeneous, stand-alone minicomputers. The applications systems are PID/Registration, Outpatient Pharmacy, Clinical Laboratory, and Radiology/Medical Records. Functional autonomy of these systems has been maintained, and no operating system changes have been required. The LACN uses a fiber-optic communications medium and provides extensive communications protocol support within the network, based on the ISO/OSI Model. The architecture is reconfigurable and expandable. This paper describes system architectural issues, the applications environment, and the local area network.