Decade-Long Trends and Factors Associated With Time to Hospital Presentation in Patients With Acute Myocardial Infarction:  The Worcester Heart Attack Study FREE

SEVERAL STUDIES have suggested that recent improvements in short-term survival associated with hospitalization for acute myocardial infarction (AMI%) may, in part, reflect the early application of myocardial reperfusion strategies.^1- 4 Patient-associated delay from the onset of symptoms suggestive of AMI to hospital arrival affects the in-hospital and, potentially, postdischarge outcomes associated with AMI, and likely impacts the timely receipt of these and other therapeutic interventions.⁵

Factors associated with time to hospital presentation, and components of delay, were initially studied before the clinical use of clot-lysing agents.^6- 11 These studies suggested that patient indecision and delay in seeking care promptly after the onset of coronary symptoms were attributed to a variety of demographic, behavioral, clinical, and situational reasons. Relatively limited recent information exists, however, particularly from a well-defined population-based perspective, about the extent of, and secular trends in, duration of patient-associated delay in seeking medical care after the onset of acute coronary symptoms. Limited data are also available about the relationship between extent of prehospital delay and patient-related characteristics, and whether the factors associated with delay have changed over time.

The purpose of the present study was to examine overall and temporal (1986-1997) trends in duration of prehospital delay in residents of the greater Worcester (Mass) area hospitalized with confirmed AMI and the relationship of time to hospital presentation and patient characteristics and situational factors.

The study population consisted of patients hospitalized with validated AMI in 16 university-affiliated and community hospitals in the Worcester metropolitan area (1990 census estimate, 437,000) during 1986, 1988, 1990, 1991, 1993, 1995, and 1997 as part of the Worcester Heart Attack Study.^12- 14 Fewer hospitals were included in recent years because of hospital closures, mergers, or conversion to long-term care facilities. Because data concerning prehospital delay were not collected in this population-based study before 1986, data from the most recent years of abstraction constitute the focus of this report. The methods of sample identification and diagnostic criteria used in this longitudinal study of greater Worcester residents hospitalized with AMI have been described in detail previously.^12- 14 To be considered for study inclusion, patients had to reside in the Worcester metropolitan area and have at least 2 of 3 predefined criteria consistent with AMI. These criteria included a typical history of prolonged chest pain not relieved by rest or use of nitrates, serum enzyme level elevations, and serial electrocardiographic tracings during hospitalization showing evolutionary changes in the ST segment or Q-waves typical of AMI.

Although a variety of factors (eg, patient, provider, and emergency medical services system) contribute to time to hospital presentation, the term prehospital delay was used to represent the various components of delay. This term was used because information about each of the components of delay was not collected. The time during which the patient reported becoming acutely or severely ill was used as the working definition of symptom onset time. Delay time was defined as the interval between acute symptom onset and arrival time in the emergency department. Information was abstracted from the medical records of patients who satisfied the study inclusion criteria concerning demographic, medical history, and clinical data that might be associated with time to hospital presentation.

Differences in the distribution of characteristics in patients classified according to extent of prehospital delay were examined by χ² tests for discrete variables; t tests were used to examine differences between groups for continuous variables. Analysis of variance was used to examine the significance of changes over time in average duration of delay. The Wilcoxon rank sum test was used to examine trends in median delay times. A logistic regression analysis was used to examine the association of patient-related factors and time with extent of delay while controlling for potentially confounding factors. The factors controlled for in this analysis were patient age; sex; race; history of angina, diabetes, hypertension, heart failure, or stroke; AMI order (initial vs prior); AMI type (Q-wave vs non–Q-wave); AMI location (anterior vs inferior or posterior); hospital survival status; time and day of the week of acute symptom onset; and period of hospitalization.

During the 7 study periods spanning 11 calendar years, 5998 residents of the Worcester metropolitan area were admitted to all greater Worcester hospitals with confirmed AMI. Of these, data in the medical record concerning extent of patient delay could not be identified or sufficiently characterized in 2161 patients (36%); these patients were not included in the present analyses. The proportion of patients in whom information on delay time could be calculated did not change over time (60% of information available for patients hospitalized in 1986/1988; 63% of information available in 1995/1997). Significant differences were observed between patients with, as compared with those without, available information related to time to presentation. Patients with available data on delay time were significantly younger, more likely to be male, and more likely to include those without a history of stroke or heart failure, but with a history of angina. Patients with an initial Q-wave AMI, those whose AMI occurred on weekends, and those who survived the acute hospitalization were significantly more likely to have information on delay time available than respective comparison groups. The present report is based on 3837 patients for whom information concerning the time interval between the onset of symptoms suggestive of AMI and hospital arrival could be determined.

The average delays between onset of symptoms suggestive of AMI and presentation to area-wide hospitals did not change significantly during the periods examined (Figure 1). Average delay times were 4.1 hours in 1986 and 4.3 hours in 1997. Median delay times did not change during the periods under study (2.0 hours in 1986 and 1997). When patients with prolonged (>24 hours) delay were redefined as having delayed 24 hours, the average delay was 3.9 hours in 1986 and 4.1 hours in 1997 and the median delays remained unchanged.

Overall, 44% of patients with AMI arrived at area-wide hospitals in less than 2 hours of the onset of symptoms, while 78% of patients presented for care in less than 6 hours. No significant differences in these patterns were observed when these data were examined for the 2 earliest and the 2 most recent study periods (Figure 2).

We initially examined the association between demographic and clinical factors and duration of prehospital delay (Table 1). Average and median delay times increased with advancing age, and women delayed slightly longer than men. Patients with diabetes exhibited more prolonged delay than patients without diabetes, and patients without a history of hypertension, heart failure, or stroke exhibited longer mean delays but, in general, shorter median delays than patients with these conditions. Patients with an initial AMI delayed longer than those with previous AMI, as did patients with a Q-wave or anterior AMI as compared with respective comparison groups. Patients experiencing their heart attack symptoms on the weekend, afternoon, evening, or early morning delayed longer.

Given our interest in trends in delay times, we examined changes in mean and median delay times between patients hospitalized in the 2 initial study years in comparison with those hospitalized in the 2 most recent study periods according to selected demographic, medical history, and clinical characteristics (Table 2). Previously observed findings about the association of selected variables to extent of delay remained. Of concern, however, was that the average and median delay times, in general, tended to either remain the same or increase in each subgroup examined during the approximate 10-year study period.

We next examined differences in a variety of characteristics between those arriving at the hospital within 2 hours after the onset of acute symptoms in comparison with those arriving after this period, since the likelihood of receiving clot-lysing agents and benefits to be gained are greatest within the first few hours of AMI onset. Patients who delayed their arrival to area-wide hospitals by more than 2 hours were older and included a significantly greater proportion of women, those with a non–Q-wave AMI, and patients with a history of diabetes, hypertension, and angina (Table 3). These patients were also more likely to include those who developed heart attack symptoms between 6 PM and 6 AM but were less likely to include those who died during hospitalization. Because data exist to suggest that the benefit associated with thrombolytic agents may be extended to patients presenting for treatment within 6 hours of the onset of AMI symptoms, a second analysis was carried out to compare differences between patients arriving at the hospital less than 6 hours after the onset of acute symptoms and those arriving after more than 6 hours (Table 3). Significant differences were seen between these 2 groups with respect to AMI order (initial vs prior), AMI type (Q-wave vs non–Q-wave), and time of presentation for medical care.

Two multivariable regression analyses were performed to examine the association between patient characteristics, situational factors, and period of hospitalization to prehospital delay. The first of these regression analyses examined factors associated with the likelihood of delaying more than 2 hours, while the second analysis examined factors associated with delay of more than 6 hours before hospital arrival (Table 4). The results of the first analysis showed that advancing age; history of angina, MI, or diabetes; failure to survive the acute hospitalization; occurrence of acute symptoms between noon and midnight; and hospitalization in 1997 were significantly associated with delays of more than 2 hours. Greater age (≥75 years), history of diabetes and AMI, absence of previous hypertension, time of symptom occurrence between noon and midnight, and hospitalization in 1990 and beyond were significantly associated with delays of more than 6 hours.

This population-based study included residents with validated AMI from a defined metropolitan area whose sociodemographic characteristics generally reflect those of the US population with the exception of race, because the majority of metropolitan Worcester residents are white. The results suggest no significant differences during the periods under study in time to hospital presentation after the development of symptoms suggestive of AMI. On the other hand, after controlling for potentially confounding variables that might affect delay, patients hospitalized in the more recent study periods were significantly more likely to delay seeking care than were patients hospitalized in earlier periods. Our data also suggest that patients who are at increased risk for developing and dying of AMI are the ones most likely to delay seeking care promptly.

A review of data from a number of US and European studies published between 1969 and 1987 found that median prehospital delay times for patients with AMI ranged from 2½ hours to 7 hours.¹⁵ Median delay of 2.5 hours was observed for patients with AMI coming to 6 Minnesota hospitals in 1991.¹⁶ The Physicians' Health Study demonstrated a median delay time of 1.9 hours for physicians experiencing an AMI.¹⁷ The median prehospital delay in patients hospitalized for evaluation of heart attack symptoms during the baseline period (December 1995 through March 1996) of the large multicenter Rapid Early Action for Coronary Treatment trial was 2.0 hours.¹⁸ Using data from the Second National Registry of Myocardial Infarction, we found that the majority of patients with AMI delayed seeking hospital care by more than 2 hours and median delay times (2.1 hours) did not change between 1994 and 1997.¹⁹ It is evident in reviewing these data that prehospital delay is substantial in the setting of AMI.

There is overwhelming evidence of the beneficial effects provided by reperfusion strategies in patients with AMI. Although various contraindications to the administration of thrombolytic therapy presently exist, including prolonged prehospital delay, thrombolytic therapy is underutilized in the management of AMI. We previously reported findings from the Worcester Heart Attack Study that showed that patients who delayed between 4 and 6 hours and those who delayed more than 6 hours were 2.5 and 6.5 times, respectively, less likely to receive thrombolytic agents compared with patients who arrived at area-wide hospitals within 1 hour of the onset of AMI; fewer than 1 in 5 patients with validated AMI in this nonrandomized observational study received thrombolytic therapy.²⁰ In the present study, more than 1 in 5 patients with confirmed AMI delayed their presentation to the hospital by more than 6 hours. These and other data suggest that a considerable percentage of patients with AMI present to the hospital after prolonged delay and outside of the window of opportunity for the greatest benefit to be gained by the administration of reperfusion therapy, even in the thrombolytic era. These findings provide further support for increased efforts through educational approaches to reduce the extent of patient delay and enhance the more widespread use of time-dependent management strategies in patients hospitalized with AMI.

Sociodemographic, behavioral, clinical, and contextual factors have been examined in association with time to hospital presentation after the onset of AMI symptoms.^21- 23 Our findings suggest that older individuals take longer to access medical care than do younger individuals. Reasons for a more prolonged delay among the elderly are unknown but may reflect more limited access to, and use of, medical care; negative previous experiences with hospital treatment; failure to recognize or appreciate the symptoms of AMI (which may be more nonspecific in these persons); or a number of other socioeconomic, attitudinal, or behavioral reasons.

The reasons patients with a history of diabetes delayed their hospital arrival is unclear. It could be speculated that the delay is greater because pain perception is altered in diabetic patients, and greater denial may also be operative in these patients.

Of the components contributing to institution of medical care, time elapsed between the onset of AMI symptoms and the decision to seek care accounts for the largest proportion. When symptoms of acute coronary disease develop, many patients at first either deny or minimize these symptoms or are unable to distinguish them from more common complaints. It is unclear why patients who have had a previous AMI in this study, and who should have been counseled about the importance of seeking hospital care quickly if symptoms of acute coronary heart disease recur, delay longer than those experiencing a first AMI.

Patients who died during their acute hospitalization and those who experienced their symptoms in the afternoon or evening were more likely to delay seeking care. It is unclear why patients who died were more likely to delay seeking care; one might speculate that these patients might have experienced more severe symptoms, thereby encouraging them to seek care more rapidly. Because of the timing of the symptoms, patients may prefer to wait before seeking medical care at different times of the day or evening, being reluctant to contact their provider or ask for the assistance of family members or coworkers at potentially less convenient times. Community-wide educational efforts to reduce delay should be directed not only to previously described subgroups at risk for prolonged delay, who are also at high risk for development of AMI, but to all patients with symptoms suggestive of AMI.

Public education campaigns that have used a variety of educational approaches have been undertaken during the past decade to reduce time to seeking medical care in those with evolving AMI. Some studies demonstrated no significant reductions in delay times,^24- 28 whereas others showed positive results.^29- 31 Some studies were able to show increases in the percentage of patients arriving within several hours of symptom onset or increases in the use of thrombolytic therapy.^29- 30 The Rapid Early Action for Coronary Treatment trial is the largest, most recent, and most methodologically rigorous study completed to date designed to reduce time to hospital presentation in persons with suspect AMI.^18,31 This trial was conducted in 10 pair-matched communities throughout the United States and an 18-month educational campaign was applied. It is apparent that methodologic limitations in previous studies and other concerns make data interpretation difficult, and many of the designed educational interventions have not tailored their approaches to persons most likely to delay seeking medical treatment.

There are several limitations to the present study. Patients who died of AMI outside of the hospital were not studied, nor were patients with silent or unrecognized AMI included, and their approaches to seeking health care are unknown. We were able to obtain information about patient delay in slightly less than two thirds of patients with confirmed AMI in our study. Significant differences in several characteristics were found between those with known and those with unknown delays. Therefore, appropriate caveats need to be placed on the interpretation of the study findings, since different delay patterns may have existed in patients in whom we were unable to define the time of onset of AMI symptoms. Possible explanations for the large percentage of data missing from the medical records concerning delay may be the result of poor or nonspecific recollection by patients or insufficient questioning or recording of this information by health care professionals. Efforts should continue to improve patient interviews and the information recording process to the systematic and standardized assessment of delay. This remains timely, as even in 1997, nearly 40% of patients with confirmed AMI did not have sufficient information noted in medical records to characterize either time of acute symptom onset or duration of delay. Because of our methods of data collection, we were unable to differentiate the collection of information on delay time in patients with vs those without ST-segment elevation. The selection of reperfusion strategies is critical to the treatment of patients with ST-segment elevation, and, correspondingly, accurate collection of information about duration of delay in these patients is essential. Patient interviews during the initial physical examination were carried out by a variety of physicians, nurses, and paramedical personnel at community and teaching hospitals, with varying degrees of skill in obtaining this information. The present results might have been different, and delay patterns longer, had we used more ill-defined, nonspecific symptoms or included patients from different socioeconomic strata and geographic locations.

Accepted for publication June 14, 2000.

This study was supported by grant RO1 HL35434 from the the National Heart, Lung, and Blood Institute, Bethesda, Md.

Reprints: Robert J. Goldberg, PhD, Department of Medicine, University of Massachusetts Medical School, 55 Lake Ave N, Worcester, MA 01655 (e-mail: Goldberr@ummhc.org).