Delayed Hospital Presentation in Patients Who Have Had Acute Myocardial Infarction

Methods

Setting

The Minnesota Clinical Comparison and Assessment Program is a quality improvement program of the Healthcare Education and Research Foundation that involves hospitals throughout Minnesota; the 45 participating hospitals account for 60% of all hospital admissions statewide. Of these hospitals, 37 participated in our study. Fifty-four percent of the study hospitals were located in urban areas. Fifty-one percent had fewer than 100 beds; 43% had 100 to 500 beds; and 5% had 500 or more beds. Two hospitals were academic medical centers, and the rest were community hospitals.

Study Sample

To define a sample of patients in whom acute myocardial infarction was suspected at the time of hospital presentation, all admissions to the study hospitals from October 1992 to July 1993 were retrospectively screened for the following admission diagnoses: acute myocardial infarction, rule-out acute myocardial infarction, and suspected acute myocardial infarction [7]. Information on clinical symptoms at the time of hospital presentation, as well as electrocardiographic results and serum enzyme levels during the first 24 hours of hospitalization, was obtained from medical records. Patients were included if they met two of the following clinical criteria: 1) typical symptoms of acute myocardial infarction [chest discomfort, arm or shoulder pain, diaphoresis, dyspnea, nausea or vomiting, and neck or jaw pain]; 2) explicit medical record documentation by a physician that electrocardiographic findings were considered compatible with acute myocardial infarction; and 3) elevated serum creatine kinase and MB isoenzyme levels that were above the upper limit of normal (as specified by the laboratory at each participating hospital). Of 4968 persons with the specified admission diagnoses, 2559 were excluded for the following reasons: failure to meet clinical criteria (44.4%); transfer from a nonstudy hospital (30.5%); inability to access the medical record (22.0%); history of myocardial infarction during the 2 weeks before the hospital admission of interest (2.7%); and death by the time of arrival at the hospital (0.4%).

Length of Delay

The amount of time that elapsed between the onset of symptoms suggestive of an acute coronary event and hospital presentation was categorized as follows: less than 2 hours, 2 to 6 hours, more than 6 hours to 12 hours, more than 12 hours to 24 hours, more than 24 hours to 72 hours, and more than 72 hours. Delay categories were defined a priori to be clinically relevant to making decisions about the management of acute myocardial infarction, particularly the use of thrombolytic therapy. Patients whose medical records indicated that they had developed symptoms within the day of admission but gave no additional information on timing were placed in a separate category. For analyses of the relation between various demographic and clinical factors and delay, we considered these patients to have presented to the hospital more than 12 hours to 24 hours after symptom onset; this time category was outside the window during which benefit could be achieved from thrombolytic therapy [2].

Patient Characteristics

Demographic and socioeconomic variables of interest were age (<55, 55 to 64, 65 to 74, 75 to 84, or ≥ 85 years), sex, marital status (married, never married, widowed, or divorced), living arrangement (not alone, alone, or residing in a long-term care facility), location of residence based on ZIP code (urban or rural), employment status (employed, unemployed, or retired), and health insurance status (Medicare, health maintenance organization, or other commercial coverage; Medicaid; or no insurance). Each patient was also characterized according to the median income of the ZIP code of residence for 1993, obtained from U.S. census data. Use of emergency medical transport was determined from information in the medical record.

Clinical characteristics noted at the time of admission included a history of hypercholesterolemia, hypertension, diabetes mellitus, angina, myocardial infarction, congestive heart failure, mechanical revascularization (percutaneous transluminal coronary angioplasty or coronary artery bypass graft surgery), family history of coronary artery disease, and previous use of sublingual nitroglycerin. Patients were characterized according to the presence or absence of chest discomfort as a symptom of acute myocardial infarction.

We used the Index of Co-existent Disease to characterize the burden of comorbid disease among the study patients. Details on the development of this index and scoring guidelines have been described elsewhere [8, 9]. The index consists of two dimensions: 1) severity of coexisting medical conditions and 2) degree of physical impairment. These dimensions were determined on the basis of information contained in the medical record at the time of hospital admission. The assessments for both dimensions were condensed into a single composite index consisting of four levels that ranged from no coexisting disease or no significant physical impairment to severe coexisting disease or severe physical impairment (none, mild, moderate, severe).

The day of symptom onset was characterized according to whether it was a weekday or occurred on a weekend. The time of day during which symptoms developed was categorized as follows: 6:00 a.m. to 11:59 a.m., noon to 5:59 p.m., 6:00 p.m. to 11:59 p.m., and midnight to 5:59 a.m. To determine the time-of-day category for a particular patient, the number of hours reflecting the midpoint of the delay category for that patient was subtracted from the time of hospital presentation. This measure was determined only for patients with delays of 24 hours or less because the relevance of this factor for patients with delays of more than 1 day was believed to be questionable.

Data Collection and Integrity

All information that was relevant to our study was abstracted from medical records. Trained nurse abstracters collected all data. Abstracters were required to show ongoing inter-rater agreement with a criterion review of 95% or higher. A random sample of 10% of each abstracter's completed cases was retrospectively audited to ensure that this standard was met.

Statistical Analysis

Patients were classified according to whether the delay between onset of symptoms suggestive of acute myocardial infarction and hospital presentation was 6 hours or less or more than 6 hours. Factors potentially associated with a delay of more than 6 hours were first examined by using chi-square statistics. Logistic regression models were used to determine factors that were independently associated with delay of more than 6 hours. Regression models initially included terms for patient age, sex, marital status, living arrangement, location of residence, employment status, health insurance status, median income, presence of chest discomfort as a symptom, medical history (hypercholesterolemia, hypertension, diabetes mellitus, angina, myocardial infarction, congestive heart failure, mechanical revascularization, family history of coronary artery disease, and use of sublingual nitroglycerin), and Index of Co-existent Disease category. A term was also included in the model to indicate the day of the week on which symptoms developed. To examine the independent effect on delay of the time of day at which symptoms occurred, an analysis was done only for patients whose symptoms had lasted 24 hours or less.

Final models were constructed by using the standard stepwise regression procedure available in SAS PROC LOGIST (SAS Institute, Cary, North Carolina) [10]; all dummy variables associated with a unique study characteristic were forced into the final model if any of the dummy variables was associated with a P value of 0.05 or less. Selected variables (age and sex) were forced into the models regardless of the significance level. Odds ratios and 95% CIs were calculated directly from the estimated regression coefficients and their SEs.

Results

Length of delay could be assigned for all but 5 of the 2409 patients (99.8%). The distribution of delay was as follows: less than 2 hours, 35% of 2404 patients; 2 to 6 hours, 25%; more than 6 hours to 12 hours, 9%; more than 12 hours to 24 hours, 8%; more than 24 hours to 72 hours, 12%; more than 72 hours, 7%; and within the day of admission without further information on timing, 4%. Of these 2404 patients, 969 (40%) delayed presentation to the hospital for more than 6 hours after the onset of symptoms. Forty-two percent of all patients used emergency medical transport services.

In univariate analyses, several factors were associated with a delay of more than 6 hours (Table 1 and Table 2). Women were significantly more likely to delay than men. Patients older than 85 years of age were more likely to delay than those younger than 55 years of age. Patients who did not have a living spouse, who lived alone, who resided in a long-term care facility, or who were retired delayed more often than other patients. Patients with a history of hypertension or congestive heart failure were more likely to delay than were those without such a history, as were those with an increased burden of coexisting disease. Patients whose symptoms developed between 6:00 p.m. and 11:59 p.m. and between midnight and 5:59 a.m. had the greatest risk for delay compared with patients whose symptoms developed between 6:00 a.m. and 11:59 a.m. Patients who had a history of coronary angioplasty or bypass graft surgery were less likely to delay than those who had not had these procedures. Patients who had chest discomfort were more likely to delay than patients who had symptoms of acute myocardial infarction other than chest discomfort.

One hundred forty-seven patients with unknown values for any of the categorical variables were excluded from the multivariate analysis; thus, 2262 patients from the total study sample remained for analysis (Table 3). According to this analysis, women were significantly more likely to delay than men. Patients 85 years of age or older were more likely to delay than patients younger than 55 years of age. Among clinical factors, history of hypertension was associated with delay. The presence of chest discomfort and a history of mechanical revascularization significantly reduced the risk for delay. In a multivariate model that included only patients with symptoms that had lasted 24 hours or less, the onset of symptoms between midnight and 5:59 a.m. posed the greatest risk for delay compared with the reference time period (6:00 a.m. to 11:59 a.m.) (adjusted odds ratio, 3.64 [95% CI, 2.55 to 5.21]).

Discussion

Our community-based study is one of the most extensive evaluations to date of the various factors potentially related to delayed hospital presentation in patients suspected of having had acute myocardial infarction. Advanced age and female sex predisposed patients to an increased delay in hospital presentation. Onset of symptoms during the evening and early morning hours also substantially increased the likelihood of delay; the risk was greatest for patients whose symptoms developed between midnight and 6 a.m. As expected, patients whose symptoms included chest discomfort and patients who had previously had a mechanical revascularization procedure (angioplasty or bypass surgery) were less likely to delay presentation. However, history of cardiac risk factors or cardiac disease (including hypercholesterolemia; diabetes mellitus; family history of coronary artery disease; and personal history of angina, myocardial infarction, or congestive heart failure) had no discernible effect on the occurrence of prolonged delay. A history of hypertension actually increased the likelihood of delaying presentation by more than 6 hours.

Since the 1960s, several investigations have examined patient delay in response to symptoms of acute myocardial infarction [11, 12]. Comparisons across studies are extremely difficult to make because of differences in study setting, study years, definitions and components of delay time, inclusion criteria, numbers of patients studied, and the demographic and clinical characteristics examined in relation to delay [13]. Our study has several strengths in comparison with previous work. The time period of our study fits squarely into the thrombolytic era of cardiovascular care. Our study sample consisted of many patients who presented to a large and diverse group of hospitals with symptoms of acute myocardial infarction. Inclusion criteria were based on information about the clinical assessment of patients limited to the first 24 hours of hospitalization. We believe that our study sample represents the most relevant group of patients with regard to clinical decision making in the initial phase of diagnosing and treating acute myocardial infarction.

In the final model, advanced patient age was significantly related to delayed hospital presentation. Several factors may explain this finding. Older patients are more likely to have atypical or asymptomatic myocardial infarction and are more likely to have an increased burden of comorbid disease than are younger patients [14-16]. The association between older patient age and delay has been noted in many population-based studies as well as in randomized trials in patients with acute myocardial infarction [1, 5, 13, 17-21]. The risk for dying after acute myocardial infarction increases dramatically with advancing age [22, 23]. Data from the Worcester Heart Attack Study [24] indicate that in 1990, the in-hospital case-fatality rate from acute myocardial infarction increased from 3% among persons younger than 65 years of age to 14.2% among persons 65 to 74 years of age and to 29.2% for persons 75 years of age or older. Reducing the time between the onset of symptoms and hospital presentation may be one way to begin to address the poor outcomes currently seen in older patients who have had acute myocardial infarction.

Women suspected of having had acute myocardial infarction were more likely than men to delay presentation. Although various studies have examined differences in patterns of cardiovascular care between women and men, little attention has been paid to the prehospital phase of acute myocardial infarction [25]. Factors that may explain the role of sex in prolonged delay include differences between men and women in age, comorbid conditions, symptoms, social support, and insurance characteristics. We tried to control for many of these factors in our analyses. This finding implies the need for further exploration of the role of women's knowledge and interpretation of symptoms of acute coronary events as well as physicians' knowledge and attitudes about female patients who report symptoms suggestive of cardiac disease.

A link between low socioeconomic status and delay was not clearly established in previous studies [26]. However, such a relation has been suggested by the findings of Ell and colleagues [27], who compared black patients hospitalized with acute chest pain in a large urban public hospital (in which most patients had no insurance and no other source of care) with patients admitted to a large urban private health maintenance organization hospital [27]. Patients admitted to the public hospital had substantially longer delays than did patients admitted to the private hospital. In our study, we did not observe an association between median income (determined on the basis of ZIP code of residence) and prolonged delay.

Fewer than half of all patients in our study used emergency medical transport services. Previous studies have suggested that a patient's decision to call a physician first rather than an ambulance substantially increases the time to arrival at the hospital [28-32]. We could not determine the level of physician involvement (if any) in the decision to use emergency transport to the hospital. The process by which patients with chest pain are triaged by telephone varies by physician; even carefully designed chest pain protocols, when used by improperly trained personnel or when applied to patients with atypical symptoms, may not be adequate for making prompt and appropriate decisions. In the managed care setting, increasing pressures to avoid expensive emergency transport costs and emergency department care raise additional concerns about unnecessary delays in hospital presentation.

Our investigation also has several limitations. We relied on retrospective chart review to obtain all data used. We did not have access to patients or their physicians to directly obtain specific information on the components of delay or to more fully explore factors that may have contributed to delay [4, 33]. However, we could ascertain information on delay for almost all study patients; this level of ascertainment is generally seen only in clinical trials of acute myocardial infarction that require specific information on delay to confirm eligibility for participation. Patients participating in such trials are often not representative of the general population of patients who have had acute myocardial infarction [34]. Although our study was limited to one state, we were able to include many predominantly community hospitals that accounted for a substantial proportion of the hospital admissions throughout Minnesota. Furthermore, many of our findings are consistent with those of studies that derived patients from several sites in diverse geographic regions [1, 3].

Among patients with symptoms of acute myocardial infarction, various factors contribute to prolonged delay in seeking hospital care: attempts at self-treatment with rest or medication, an extended process of decision making in which the patient seeks the advice of a physician or family member before obtaining hospital care [31], and underuse of emergency medical services [26]. Continued improvements in 911 systems, emergency medical dispatch systems, the staffing and equipping of emergency medical service systems, and the rapid identification and treatment of patients in emergency departments are essential to improving the care of patients who have had acute myocardial infarction [35, 36]. However, the existing literature suggests that patient indecision about seeking medical help is the most important reason for delay in hospital presentation among patients suspected of having had acute myocardial infarction. In a study of more than 2000 patients hospitalized with acute myocardial infarction in King County, Washington, the most common reasons for delays in seeking care were the belief that the symptoms would go away and the belief that the symptoms were not severe enough to warrant seeking care. The major reasons for not using emergency medical transport services were that the symptoms were not severe enough to warrant their use, the patient had not thought of calling 911, and the patient had believed that self-transport would be faster [37].

No proven method is currently available for educating patients to hasten their presentation to the hospital in response to symptoms suggestive of acute myocardial infarction [38]. Meischke and colleagues [37] have suggested that interventions need to go beyond increasing knowledge of signs and symptoms of acute myocardial infarction and should include components that lead to increased patient self-confidence in dealing with these symptoms in order to rapidly activate the emergency medical system [37]. Our findings have several implications for the design and implementation of such interventions. The elderly and women are two patient populations that require targeted interventional efforts. Efforts are also required to increase the awareness of persons with established cardiac risk factors of the need for prompt response to symptoms. Increasing the appropriate use of emergency medical services requires interventions directed at both patients and physicians. Any reluctance by patients to seek medical care during the evening and early morning hours must be discouraged. Continued efforts to reduce patient delay remain integral to improving the management of acute myocardial infarction.

Drs. McLaughlin and Soumerai and Ms. Gao: Department of Ambulatory Care and Prevention, Harvard Medical School and Harvard Pilgrim Health Care, 126 Brookline Avenue, Suite 200, Boston, MA 02215.

Dr. Willison: System-Linked Research Unit on Health and Social Service Utilization, McMaster University Health Sciences Centre, 1200 Main Street West, Room 3N46, Hamilton, Ontario L8N 3Z5, Canada.

Drs. Guadagnoli and Hauptman: Department of Health Care Policy, Harvard Medical School, 25 Shattuck Street, Parcel B-1st Floor, Boston, MA 02115.