Relationship of physical symptoms to posttraumatic stress disorder among veterans seeking care for gulf war-related health concerns.

OBJECTIVES: Studies of the relationship of posttraumatic stress disorder (PTSD) to physical symptoms in war veterans consistently show a positive relationship. However, traumatic experiences causing PTSD may correlate with other war exposures and medical illnesses potentially accounting for those symptoms. METHODS: We analyzed data obtained from 21,244 Gulf War veterans seeking care for war-related health concerns to assess the relationship of PTSD to physical symptoms independent of environmental exposure reports and medical illness. At assessment, veterans provided demographic information and checklists of 15 common physical symptoms and 20 wartime environmental exposures. Up to seven ICD-9 provider diagnoses were ranked in order of estimated clinical significance. The relationship of provider-diagnosed PTSD to various physical symptoms and to the total symptom count was then determined in bivariate and multivariate analyses. RESULTS: Veterans diagnosed with PTSD endorsed an average of 6.7 (SD = 3.9) physical symptoms, those with a non-PTSD psychological condition endorsed 5.3 (3.5), those with medical illness endorsed 4.3 (3.4), and a group diagnosed as "healthy" endorsed 1.2 (2.2). For every symptom, the proportion of veterans reporting the symptom was highest in those with PTSD, second highest in those with any psychological condition, third highest in those with any medical illness, and lowest in those labeled as healthy. The PTSD-symptom count relationship was independent of demographic characteristics, veteran-reported environmental exposures, and comorbid medical conditions, even when symptoms overlapping with those of PTSD were excluded. CONCLUSIONS: PTSD diminishes the general health perceptions of care-seeking Gulf War veterans. Clinicians should carefully consider PTSD when evaluating Gulf War veterans with vague, multiple, or medically unexplained physical symptoms.

Race and differences in breast cancer survival in a managed care population.

BACKGROUND: African-American women with breast cancer have poorer survival than European-American women. After adjustment for socioeconomic variables, survival differences diminish but do not disappear, possibly because of residual differences in health care access, biology, or behavior. This study compared breast cancer survival in African-American and European-American women with similar health care access. METHODS: We measured survival in women with breast cancer who are served by a large medical group and a metropolitan Detroit health maintenance organization where screening, diagnosis, treatment, and follow-up are based on standard practices and mammography is a covered benefit. We abstracted data on African-American and European-American women who had been diagnosed with breast cancer from January 1986 through April 1996 (n = 886) and followed these women for survival through April 1997 (137 deaths). RESULTS: African-American women were diagnosed at a later stage than were European-American women. Median follow-up was 50 months. Five-year survival was 77% for African-American and 84% for European-American women. The crude hazard ratio for African-American women relative to European-American women was 1.6 (95% confidence interval [CI] = 1.1-2.2). Adjusting only for stage, the hazard ratio was 1.3 (95% CI = 0.9-1.9). Adjusting only for sociodemographic factors (age, marital status, and income), the hazard ratio was 1.2 (95% CI = 0.8-1.9). After adjusting for age, marital status, income, and stage, the hazard ratio was 1.0 (95% CI = 0.7-1.5). CONCLUSION: Among women with similar medical care access since before their diagnoses, we found ethnic differences in stage of breast cancer at diagnosis. Adjustment for this difference and for income, age, and marital status resulted in a negligible effect of race on survival.

Biological effects of genes in the Grc and EC region of the rat major histocompatibility complex.

PROBLEM: To study the mechanism of action of major histocompatibility complex (MHC)-linked genes affecting reproduction, growth, and susceptibility to chemical carcinogens. METHOD OF STUDY: Tumors derived from rat embryonic fibroblasts were transfected with cosmids from the Grc and its linked regions, the unrelated A region, and a nonMHC region, or with genes from the Grc, Grc-linked, and nonMHC regions, to determine whether they could suppress tumor growth as determined by in vitro (soft agar) and in vivo assays. RESULTS: Tumor fibroblasts transfected with cosmids from the Grc or from the EC region decreased tumor growth in both the in vitro and in vivo assays. Transfection with individual genes from the Grc had no effect on tumor growth in either assay. CONCLUSIONS: The effects of the Grc on reproduction, growth, and tumorigenesis are mediated by extended genetic effects, i.e., by the conformation of the DNA in this region. Similar effects were seen following transfection with cosmids from the Grc-linked EC region, and this finding strengthens the hypothesis that the conformation of the DNA in this general region is critical for its function. A similar effect has been described for the locus control region (LCR) in the beta-globin gene family in the human.

Patterns and characteristics of repeat mammography among women 50 years and older.

Whereas efforts encouraging women to obtain initial mammograms are laudable, the importance of returning for subsequent routine mammograms cannot be minimized. The purpose of this study was to measure the timing, patterns, and characteristics of repeat screening mammography over time in a defined population of health maintenance organization members for whom mammography was a fully covered benefit. We identified all women ages 50-74 years who were enrolled in a southeastern Michigan health maintenance organization, assigned to a large medical group, and received at least one screening mammogram with a normal result between January 1, 1989 and December 31, 1996. Using administrative and radiology data, we calculated the proportion of women who received a subsequent mammogram within 2 years and the time to subsequent screening, both overall and stratified by demographic characteristics. We also examined screening patterns over a 5-year period. Of the 8749 women included in this study, 66.0% [95% confidence interval (CI), 65.0-67.0%] were subsequently screened within 2 years. We found slightly higher rates among Caucasians and married women. The proportion of women who received repeat mammography increased with estimated household income [9.5% difference between the highest and lowest categories (95% CI, 6.5-12.5%)]. The median time to subsequent screening was 17.7 months, and the probability of repeat screening was higher for women whose initial mammogram was between January 1992 and December 1994 compared to those receiving an initial mammogram between January 1989 and December 1991 (9.6% difference; 95% CI, 7.5-11.7%). Repeat mammography has improved over time; however, socioeconomic status could contribute to longer-than-intended intervals between screening when translated into real-world clinical practice. In a setting where most physicians recommended annual screening, we found that the median time to subsequent screening was delayed by 6 months. If annual mammography is the goal, recommendations should be made with the understanding of how the timing of repeat screening occurs in clinical practice.

The Henry Ford Health System.

The Henry Ford Health System is a large, vertically integrated health care delivery system, with its core service delivery team comprised of a 1000-person salaried medical group. Oncology services are coordinated through its Cancer Center, which organizes regional distribution of oncology services for residents of southeast Michigan. Oncology services are delivered through five regionally distributed sites, affording a vast majority of cancer services to be provided for patients a short distance from their homes. The System owns the largest health maintenance organization in Michigan (Health Alliance Plan) and its organizational structure affords the opportunity of offering purchasers specific oncology service contracting opportunities. Advantages of providing comprehensive oncology services through an integrated health system include: 1) standardized cancer care guidelines, 2) medical information exchange through an electronic medical record, 3) interdisciplinary cancer care provided by salaried physicians, minimizing potentially conflicting financial issues in treatment decisions, 4) state-of-the-art care afforded through availability of involvement in a large number of National Institutes of Health-sponsored clinical trials, 5) high standards of credentialing for oncology physicians, and 6) integrative managed care perspectives and continuous attention to cost and quality of care issues.

Noninvasive venous testing in the diagnosis of pulmonary embolism: the impact on decisionmaking.

PURPOSE: To characterize the use and utility of lower extremity noninvasive venous testing (NIVT) in the diagnosis of pulmonary embolism (PE). METHODS: The study is a retrospective case series of consecutive patients in whom PE was suspected who were referred to a large, urban tertiary care center for NIVT. The main outcome measures of the study were the rate of positive results of NIVT, the amount of new information provided by NIVT, and the frequency of management changes that were attributable to NIVT. RESULTS: Forty-one of 450 patients (9%) had deep venous thrombosis (DVT) by NIVT. The prevalence of DVT by NIVT among patients not evaluated by ventilation/perfusion (V/Q) scanning was 8%. The prevalence of DVT by NIVT among patients with a high-probability V/Q scan result before NIVT was 39%, but no management decisions in this group were based on a positive NIVT result and only two decisions were based on negative NIVT results. The prevalence of DVT according to NIVT among patients who had a negative "diagnostic" (low, or very low probability, or normal) result of V/Q scan before NIVT was 2%. The overall frequency of management changes attributed to NIVT was only 2.5%. In the remaining 97% of patients, management was determined by the result of V/Q scanning or of subsequent pulmonary arteriography. CONCLUSIONS: In patients in whom PE is suspected, results of NIVT are usually negative for acute DVT. Management decisions are almost always based on V/Q scan or results of pulmonary arteriography and not on NIVT. The utility of NIVT to identify DVT in these patients appears limited, and a more selective approach to its application for the diagnosis of PE should be considered.

Redesigning primary care processes to improve the offering of mammography. The use of clinic protocols by nonphysicians.

OBJECTIVE: To develop, within the framework of continuous quality improvement, new processes for offering mammography and determine whether protocols executed completely by nonphysicians would increase mammography utilization. DESIGN: A prospective follow-up study with patients from an intervention clinic and two control clinics. SETTING: Three general internal medicine clinics in a large, urban teaching hospital in Detroit, Michigan. PATIENTS/PARTICIPANTS: A total of 5,934 women, aged 40 through 75 years, making 16,546 visits to one of the clinics during the study period (September 1, 1992, through November 31, 1993). INTERVENTION: Medical assistants and licensed practical nurses in the intervention clinic were trained to identify women due for screening mammography, and to directly offer and order a mammogram if patients agreed. MEASUREMENTS AND MAIN RESULTS: Patients were considered up-to-date with screening if they had a mammogram within 1 year (if age 50-75) or 2 years (if age 40-49) prior to the visit or a mammogram within 60 days after the visit. The proportion of visits each month in which a woman was up-to-date with mammography was calculated using computerized billing records. Prior to the intervention, the proportion of visits in which women were up-to-date was 68% (95% confidence interval [CI] 63%, 73%) in the intervention clinic and 66% (95% CI 61%, 71%) in each of the control clinics. At the end of the evaluation, there was an absolute increase of 9% (95% CI 2%, 16%) in the intervention clinic, and a difference of 1% (95% CI -5%, 7%) in one of the control clinics and -2% (95% CI -3%, 5%) in the other. In the intervention clinic, the proportion of visits in which women were up-to-date with mammography increased over time and was consistent with a linear trend (p = .004). CONCLUSIONS: Redesigning clinic processes to make offering of mammography by medical assistants and licensed practical nurses a routine part of the clinic encounter can lead to mammography rates that are superior to those seen in physicians' usual practice, even when screening levels are already fairly high. Physicians need not be considered the sole, or even the primary, member of the health care team who can effectively deliver some preventive health measures.

Patient-related risk factors that predict poor outcome after total hip replacement.

OBJECTIVE: To identify factors associated with poor outcome after total hip replacement (THR) surgery. DATA SOURCES: This article is the first to present results from the American Medical Group Association (AMGA) THR consortium. STUDY DESIGN: The outcomes evaluated were pain and physical function. Eight patient risk factors were evaluated. These included the age, sex, race, marital status, and education of the patient; whether the patient had polyarticular disease or other comorbid conditions; and the patient's preoperative pain and physical function score. DATA COLLECTION: Data were collected from patients using AMGA-approved, self-administered questionnaires preoperatively and at six weeks, three months, six months, one year, and two years postoperatively. PRINCIPAL FINDINGS: Of the patient risk factors studied, race, education, number of comorbid conditions, and preoperative Health Status Questionnaire (HSQ) score were found to be associated with poor outcome. These risk factors were found to have an effect on both pain and physical function at six months postoperatively. Patients with higher preoperative scores were found to have higher postoperative scores, but substantially fewer of these patients received any benefit from their surgery. For each 10-point increase in preoperative score, patients could expect at least a 6-point decrease in postoperative improvement. CONCLUSIONS: Our study indicates that preoperative status is an important predictor of outcome for THR.

Inadequate follow-up of abnormal mammograms.

Routine mammographic screening increases detection of nonpalpable breast cancer. Timely follow-up of abnormalities is essential because delays may lead to postponement of treatment and decreased survival for women who have cancer. The purpose of this study was to determine the percentage of women with an abnormal mammogram who do not have adequate follow-up and to determine factors associated with inadequate follow-up. The study was conducted in a metropolitan health system that includes a large urban teaching hospital in Detroit and 26 ambulatory care centers. From the radiology database, all women with an abnormal screening mammogram performed between January 1, 1992, and July 31, 1992 were identified. We defined adequate follow-up as follow-up within three months of due date. Follow-up status was determined using medical records and telephone interviews. The percentage of women with inadequate follow-up was calculated. Relative risks compared percentages of women with inadequate follow-up according to demographic and screening-related variables. We calculated adjusted relative risks using multivariate binomial regression. We identified 1,249 women with abnormal screening mammograms. Inadequate follow-up occurred for 226 (18.1%) of the women. Among women with follow-up recommended in 4-6 months, 36.8% had inadequate follow-up. Among women with immediate follow-up recommended (obtain additional views or outside films for comparison, ultrasound, biopsy, or surgical referral), 7.2% had inadequate follow-up. Inadequate follow-up was associated with lower estimated household income and no history of previous mammogram. Among women with inadequate follow-up who were interviewed, 87% reported that they had been notified of their results. We found that the percentage of women with inadequate follow-up of abnormal mammograms is high, especially among women who require six-month follow-up. Women with low income and no history of a previous mammogram were at greatest risk for inadequate follow-up. These results document a previously unrecognized problem with mammography screening and suggest that the implementation of tracking systems to ensure timely follow-up of abnormal screening mammograms is essential. Medical Subject Headings (MeSH): mammography, follow-up, screening.

Evaluation of factors potentially associated with inadequate follow-up of mammographic abnormalities.

BACKGROUND: To increase the proportion of women who receive the recommended follow-up for mammographic abnormalities, factors which inhibit follow-up must be identified. Patient and health care delivery related factors were evaluated, stratified by type of follow-up recommendation, to determine reasons for inadequate follow-up. METHODS: All Caucasian and African American women at the Henry Ford Medical Group, in southeastern Michigan, with an abnormal screening mammogram result between January 1, 1992 and July 31, 1992 were identified. All women with inadequate follow-up, and a random sample of women with adequate follow-up, were invited to participate in a telephone interview that assessed three major dimensions of the Health Belief Model (susceptibility, benefits, and barriers), general health and health behaviors, and related characteristics. The relationship between these factors and inadequate follow-up was evaluated separately for women with immediate and 6-month follow-up recommendations, using univariate and multivariate analyses. RESULTS: A total of 555 women were invited to participate in the study (219 with inadequate follow-up and 336 with adequate follow-up). Interviews were completed for 418 women (75.3%). Women who were not notified of their mammographic abnormality were excluded from this study, leaving 399 women available for analysis. Among the women who had the recommended immediate follow-up, those who reported difficulty in obtaining medical appointments were 4 times more likely to have inadequate follow-up (95% confidence interval [CI] 1.5, 11.3), after adjusting for other variables. Among the women with six-month follow-up recommended, those who received fewer mammograms in the past 5 years were more likely to have inadequate follow-up (odds ratio [OR] = 4.0; 95% CI 1.6, 10.4). In this group, sociodemographic characteristics were not associated with inadequate follow-up, although women with transportation problems were at a higher risk (crude OR = 5.2; 95% CI 1.6, 16.7; adjusted OR = 3. 1; 95% CI 0.5, 18.3). Among women with 6-month follow-up recommended, those who perceived their health as poor or fair (crude OR = 2.4; 95% CI 1.2, 5.1; adjusted OR = 2.3; 95% CI 0.8, 6.8) and those who did not examine their own breasts frequently (crude OR = 2.3; 95% CI 1.0, 5.4; adjusted OR = 2.7; 95% CI 0.5, 18.3) were also more likely to have inadequate follow-up. CONCLUSIONS: Results from this study demonstrate that the relative importance of factors associated with inadequate follow-up of abnormal mammograms differs according to the type of follow-up recommended. For both types of recommendations, the factors identified are noteworthy because they are amenable to intervention. Efforts should be made to facilitate appointment scheduling for follow-up of abnormal mammograms, and information about previous mammography screening should be obtained to identify women who are at a high risk for inadequate follow-up.

Screening mammography use: the importance of a population perspective.

Mammography has been shown to be effective in decreasing breast cancer mortality. However, not all women receive recommended screening. The purpose of this study was to measure the effect of systemic health care delivery factors and patient demographic factors on the use of mammography among a population of women with insurance coverage for screening mammography. We studied 8,805 women, 50 to 74 years of age, who were members of a health maintenance organization in Michigan during 1992. Data were obtained from computerized patient registration and billing systems. Estimated household income was based on zip code and 1990 U.S. Census information. During the study period, 47% of the entire study population received a mammogram. Among the 4,628 women who did not receive a mammogram, 2,358 (51%) did not have any primary care visits in 1992, or did not have a visit at a time when they were due for mammography. Among women with at least one visit during 1992, 60% received a mammogram. Analysis of this group revealed that Caucasian race, married status, and higher estimated household income were associated with 3%-5% higher mammography rates (in absolute terms); seeing a gynecologist and attending a primary care clinic where mammography was available on site were associated with 10%-11% higher rates; and not having at least one of their primary care visits at a time when due for screening was associated with a 33.5% lower mammography rate. These differences persisted after adjusting for other variables. In this population, factors surrounding health care delivery, such as having an office visit when due for screening, the number of office visits, and having been seen by a gynecologist, had a greater association with mammography use than did demographic factors. A population-based perspective that includes developing outreach strategies for women overdue for screening, who have not visited their health care provider, is as important as further improvements in office-based strategies to increase breast cancer screening.

Patient notification and follow-up of abnormal test results. A physician survey.

OBJECTIVE: To investigate physician practices in the handling of patients' test results from the time the test was ordered until the time any required follow-up was completed. METHODS: Survey of 161 attending physicians and 101 residents in family practice and internal medicine practicing at a large urban teaching hospital and 21 suburban primary care practices in Southeastern Michigan. The survey included questions about physician demographics, and whether physicians have methods for ensuring that (1) the results of all tests ordered are received, (2) all patients are notified of results, (3) all patient notification is documented, and (4) all required follow-up is done. Physicians were also asked to self-rate the reliability of their methods and the importance of various steps in the handling of patients' test results. RESULTS: The response rate was 79% for both attending physicians and residents. Approximately 17% to 32% of physicians reported having no reliable method to make sure that the results of all tests ordered are received. One third of physicians do not always notify patients of abnormal results. The most common reasons reported for not notifying patients were that the results were trivial and that the patient was expected to return to the clinic soon. Residents were significantly less likely to document notifying patients of abnormal results (P < .001). Only 23% of physicians reported having a reliable method for identifying patients overdue for follow-up. CONCLUSIONS: Lack of methods to ensure that the results of tests ordered were received, dependence on follow-up visits to inform patients of results, and lack of documentation were relatively common among physicians surveyed. These could lead to an increased risk of malpractice litigation and suboptimal patient care.

Prophylaxis for stress-related gastrointestinal hemorrhage: a cost effectiveness analysis.

OBJECTIVE: To assess the cost-effectiveness of prophylaxis for stress-related gastrointestinal hemorrhage in patients admitted to the intensive care unit. DESIGN: Decision model of the cost and efficacy of sucralfate and cimetidine, two commonly used drugs for prophylaxis of stress-related hemorrhage. Outcome estimates were based on data from published studies. Cost data were based on cost of medications and costs of treatment protocols at our institutions. MEASUREMENTS AND MAIN RESULTS: The marginal cost-effectiveness of prophylaxis, as compare with no prophylaxis, was calculated separately for sucralfate and cimetidine and expressed as cost per bleeding episode averted. An incremental cost-effectiveness analysis was subsequently employed to compare the two agents. Sensitivity analyses of the effects of the major clinical outcomes on the cost per bleeding episode averted were performed. At the base-case assumptions of 6% risk of developing stress-related hemorrhage and 50% risk-reduction due to prophylaxis, the cost of sucralfate was $1,144 per bleeding episode averted. The cost per bleeding episode averted was highly dependent on the risk of hemorrhage and, to a lesser degree, on the efficacy of sucralfate prophylaxis, ranging from a cost per bleeding episode averted of $103,725 for low-risk patients to cost savings for very high-risk patients. The cost per bleeding episode averted increased significantly if the risk of nosocomial pneumonia was included in the analysis. The effect of pneumonia was greater for populations at low risk of hemorrhage. Assuming equal efficacy, the cost per bleeding episode averted of cimetidine was 6.5-fold greater than the cost per bleeding episode averted of sucralfate. CONCLUSIONS: The cost of prophylaxis in patients at low risk of stress-related hemorrhage is substantial, and may be prohibitive. Further research is needed to identify patient populations that are at high risk of developing stress-related hemorrhage, and to determine whether prophylaxis increases the risk of nosocomial pneumonia.

Who should be screened for HIV infection? A cost-effectiveness analysis.

BACKGROUND: The advent of effective prophylactic treatments for asymptomatic persons infected with human immunodeficiency virus has led to interest in widespread screening programs. However, the costs of screening programs and therapy are high, and the prevalence of infection above which screening becomes an appropriate use of scarce health care dollars remains undetermined. METHODS: To examine the cost-effectiveness of screening in populations with differing prevalences of infection, we developed a Markov model to compare costs and life expectancy for two strategies: (1) screening and prophylactic treatment for infected persons who have or who develop low CD4+ (T4) cell counts, and (2) no screening. Based on studies in the literature, we estimated the prevalence of HIV infection, the rate of T4-cell loss, the rates of developing the acquired immunodeficiency syndrome and Pneumocystis pneumonia stratified by T4 cell counts, the life expectancy with the acquired immunodeficiency syndrome, the efficacy of prophylactic therapies, and costs. RESULTS: In populations with a prevalence of infection more than 5%, which includes known risk groups, screening costs less than $11,000 per life-year gained. In populations with a prevalence as low as 0.15%, screening costs only $29,000 per life-year gained. Even when the efficacy of zidovudine is assumed to be limited to 3 years, screening still costs less than $40,000 per life-year gained in populations with a prevalence of 0.5% or greater. However, in populations with a very low prevalence of infection (two to 10/100,000), such as members of the general population without reported risk factors, screening costs rise to between $290,000 and $1,277,400 per life-year gained. CONCLUSIONS: When considering only direct medical benefits, screening for asymptomatic human immunodeficiency virus infection in the general population, without regard to reported risk factors or seroprevalence data, would be expensive. In populations with a prevalence of infection of 0.5% or greater, however, the cost-effectiveness of screening falls within the range of currently accepted medical practices. These results suggest that screening be offered routinely to all persons in defined populations, such as persons receiving care at hospitals or clinics, or residing in geographic areas, where the seroprevalence is 0.5% or more, and underscore the need to conduct seroprevalence studies to identify such populations.

Screening flexible sigmoidoscopy: patient attitudes and compliance.

OBJECTIVE: To measure patients' expectations and attitudes about screening flexible sigmoidoscopy and their discomfort during the procedure, and to identify factors affecting compliance among patients scheduled for sigmoidoscopy. DESIGN: Patient survey at the time sigmoidoscopy was ordered and again one week after the procedure was performed. SETTING: An academic general internal medicine practice. PATIENTS: 105 consecutive patients scheduled for screening flexible sigmoidoscopy. MAIN RESULTS: Seventy-five percent of patients (79/105) scheduled for sigmoidoscopy complied with the procedure. Compliance was higher among men and among patients who had family histories of colon cancer. Although many patients experienced moderate to extreme embarrassment (27%), discomfort (42%), and pain (31%), patients experienced less embarrassment (p = 0.03) and pain (p = 0.02) than they had expected. Patients aged 65 years and older were twice as likely as younger ones (52% versus 25%) to experience moderate to extreme pain (p = 0.04). Only 1.4% of patients reported that they would probably not have the test again. CONCLUSION: Although flexible sigmoidoscopy is an uncomfortable procedure for some patients, especially those aged 65 and older, in general it is not as bad as patients expect and most would have the test again. Therefore, rather than assuming sigmoidoscopy is too uncomfortable for all patients to tolerate as a screening test, clinicians should inform their patients about the potential benefits and risks of sigmoidoscopy and about what the patient can expect during the procedure, thus enabling the patient to make an informed decision about whether to undergo screening sigmoidoscopy.

Missed diagnoses of acute myocardial infarction in the emergency department: results from a multicenter study.

STUDY OBJECTIVE: To determine the rate of missed acute myocardial infarction (AMI) in the emergency department and the factors related to missed diagnoses. STUDY DESIGN: Observational and case-control study. SETTING: Data were analyzed from a multicenter study of coronary care unit admitting practices that included patients who presented to the ED with chest pain or other symptoms suggestive of acute cardiac ischemia (N = 5,773). Patients with missed AMI (cases) were compared with control patients admitted with AMI and to a second control group of patients discharged without AMI. RESULTS: Of 1,050 patients with AMI, 20 (1.9%; 95% confidence interval, 1.2-2.9%) were not admitted. Patients with missed AMI were significantly less likely to have ECG changes and a history of AMI or nitroglycerin use than patients admitted with AMI. However, they were significantly more likely to have ECG changes than patients discharged without AMI. Five patients with missed AMI (25%) had ST-segment elevation, and seven (35%) were discharged with a diagnosis of ischemic heart disease by the physician in the ED. Death or potentially lethal complications occurred in 25% of missed AMI patients. CONCLUSION: The rate of missed AMI in the ED was only 1.9%. However, 25% of these might have been prevented had ST-elevation not been missed, and another 25% might have been prevented had patients who were recognized to have ischemic heart disease by the physician in the ED been admitted.

Design considerations of CareWindows, a Windows 3.0-based graphical front end to a Medical Information Management System using a pass-through-requester architecture.

The Care Windows development project demonstrated the feasibility of an approach designed to add the benefits of an event-driven, graphically-oriented user interface to an existing Medical Information Management System (MIMS) without overstepping economic and logistic constraints. The design solution selected for the Care Windows project incorporates three important design features: (1) the effective de-coupling of severs from requesters, permitting the use of an extensive pre-existing library of MIMS servers, (2) the off-loading of program control functions of the requesters to the workstation processor, reducing the load per transaction on central resources and permitting the use of object-oriented development environments available for microcomputers, (3) the selection of a low end, GUI-capable workstation consisting of a PC-compatible personal computer running Microsoft Windows 3.0, and (4) the development of a highly layered, modular workstation application, permitting the development of interchangeable modules to insure portability and adaptability.