Temporal trends in prehospital delay in patients with acute myocardial infarction from the second National Registry of Myocardial Infarction.
Goldberg RJ, Gurwitz JH, Gore JM. Duration of, and Temporal Trends (1994-1997) in, Prehospital Delay in Patients With Acute Myocardial InfarctionThe Second National Registry of Myocardial Infarction. Arch Intern Med. 1999;159(18):2141-2147. doi:10.1001/archinte.159.18.2141
Copyright 1999 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1999
Extent of delay in seeking medical care in persons with acute myocardial infarction (AMI) is receiving increasing attention, given the time-dependent benefits associated with early administration of coronary reperfusion therapy.
To examine recent data, and temporal trends therein, about duration of prehospital delay in a large (N=364,131) cross-sectional sample of patients included in the second National Registry of Myocardial Infarction.
The medical records of patients hospitalized with AMI in 1624 US hospitals from June 1, 1994, to October 31, 1997, were reviewed for information about duration of prehospital delay.
There was evidence of a slight decline in average delay times in patients hospitalized in 1997 (5.5 hours) compared with those hospitalized in 1994 (5.7 hours). Median delay times (2.1 hours) did not change. Approximately 20% of patients presented to the hospital within 1 hour of acute symptom onset, and slightly more than two thirds presented within 4 hours. Delay times were more prolonged for older patients, women, nonwhite patients, and patients with a history of diabetes or hypertension vs respective comparison groups. Patients in cardiogenic shock exhibited shorter delay times than less severely ill patients. Patients with previous AMI or who had undergone previous coronary angioplasty presented to the hospital with shorter delay times, as did individuals hospitalized in the Mountain and Pacific regions.
These results provide insights into recent delay times and into groups at risk for prolonged delay.
EXTENT OF patient-associated delay in persons with signs and symptoms of acute coronary disease is receiving increasing attention, given the time-dependent benefits associated with early administration of reperfusion therapy after acute myocardial infarction (AMI).1- 5 Early administration of thrombolytic therapy and other myocardial reperfusion strategies results in improved survival after AMI and may have additional favorable effects on postdischarge survival and functional status.
Patient indecision and delay in seeking care quickly after the onset of acute coronary symptoms have been attributed to a variety of reasons, including failure to recognize these symptoms as serious in nature, denial, misattribution, and fear.6- 8 Elderly, female, and nonwhite patients have been shown to be at particularly high risk for delay, as have patients with a history of coronary disease.6- 9 Although several, albeit uncontrolled, community studies have attempted to reduce the extent of prehospital delay in patients with AMI, relatively little current data are available to describe the extent of prehospital delay, and time trends therein, in patients hospitalized with AMI.
The purpose of our study is to examine temporal and recent trends (June 1, 1994, to October 31, 1997) in duration of prehospital delay after AMI using data from the second National Registry of Myocardial Infarction (NRMI-2). In addition, clinical and demographic factors associated with extent of delay in seeking medical care are described.
The NRMI is a cross-sectional observational database of patients hospitalized with AMI at US hospitals that has been implemented since 1990. This voluntary registry of AMI is sponsored by Genentech, Inc, San Francisco, Calif. The data collection process at each site and quality control features of the initial study registry (NRMI-1) have been described previously.10 The primary purpose of this registry is to monitor temporal trends in institutional and national treatment practices in patients hospitalized with AMI. This observational database is also used to monitor the in-hospital safety experience of alteplase (Activase; Genentech, Inc). This includes information on 364,131 patients with AMI treated and discharged during the study period in whom data were available concerning symptom onset time and who were not transferred into participating hospitals. Since data about duration of prehospital delay were not collected in a similar manner in the NRMI-1 and NRMI-2, and since larger numbers of patients with AMI and with more complete and detailed data are available in the NRMI-2, data from the second NRMI (NRMI-2) are presented herein.
At the time of our study, 1624 US hospitals were participating and contributing data, representing more than one quarter of all acute-care medical-surgical hospitals. Participation in the registry is voluntary, and hospitals receive a small fee for each patient enrolled as reimbursement for data collection–associated costs. Registry hospitals are larger than nonparticipating hospitals. In addition, registry hospitals are more likely to be certified by the Joint Commission on Accreditation of Health Care Organizations, to be affiliated with a medical school, and to have facilities for cardiac catheterization, coronary angioplasty, and cardiac surgery.10
To be included in the NRMI-2, patients must have a documented AMI before discharge using local hospital criteria, such as patient history and serial enzyme or electrocardiographic findings, or an International Classification of Diseases-Ninth Revision discharge diagnosis of AMI (code 410). Participating hospitals are encouraged to enter consecutive patients with AMI irrespective of treatment strategy and outcome. Patients transferred into a registry hospital from another acute-care facility were excluded from our study.
For our purposes, time of AMI symptom onset is defined as the time when the onset of ischemic symptoms appeared or became constant in quality and intensity such that the patient decided to seek treatment. Time of hospital presentation is defined as the time the patient arrived at the participating registry hospital.
Measures of central tendency and frequency distributions of duration of prehospital delay were calculated in patients overall as well as according to time of hospitalization. Formal tests of statistical significance were not applied in the examination of delay times or in temporal trends and factors associated with such, given the large size of the database and the descriptive purposes of our study. Multivariable regression analyses were performed to examine demographic data, medical history, and clinical characteristics associated with prolonged hospital delay (>3 hours or >6 hours) while controlling for potentially confounding factors. Multivariable adjusted odds ratios and accompanying 95% confidence intervals were calculated in a standard manner.
The mean age of the study sample was 66.4 years, and nearly two thirds (63%) were men. Racial distribution was as follows: 86%, white; 6%, black; and 3%, Hispanic. A history of MI was present in approximately one quarter (26%) of patients, diabetes was present in a similar proportion, and nearly one half (49%) had a history of hypertension. Coronary angioplasty and bypass surgery had been performed in approximately 8% and 12% of study subjects, respectively, before the index hospitalization. Each of the major geographic regions in the United States were represented, with the greatest number and percentage of patients included from the eastern north-central, southern, and mid-Atlantic states and those in the Pacific region. There were no marked differences in the demographic and clinical characteristics of the study sample over time (Table 1).
The mean and median prehospital delay times for the study period are shown in Figure 1. There was a slight suggestion of a decline in the average delay times over time, with no change in median delay time observed. The mean delay times were 5.7 hours in 1994 and 5.5 hours in 1997. The median delay time was 2.1 hours in 1994 and 1997. After redefining duration of prehospital delay to 24 hours for patients who delayed seeking hospital care by 1 day or longer (5.4% of the study sample), the adjusted mean delay times were 4.9, 5.0, 4.9, and 4.8 hours for patients hospitalized in 1994, 1995, 1996, and 1997, respectively. The median delay times were unchanged. Relatively similar trends were seen when duration of prehospital delay was categorized according to discrete time intervals (Table 2). Approximately 1 in 5 patients with AMI presented to the hospital within 1 hour of the onset of acute symptoms, whereas slightly less than one quarter of patients delayed seeking care by 6 or more hours, with no change in these distributions observed over time.
Temporal and overall trends in duration of prehospital delay in various demographic, geographic, and clinically defined subgroups were examined (Table 3). Elderly patients were more likely than younger patients to delay seeking medical care, as were women compared with men. Sex differences in duration of prehospital delay continued to be observed when the data were examined in 4 age strata. African Americans, Hispanics, and Asians were more likely to delay seeking hospital care than white patients. Despite their relatively small numbers, American Indians delayed seeking care by an average of 2.5 hours longer than white patients and had the highest median delay times of the various racial and ethnic groups examined. In terms of clinical characteristics, persons without a history of MI or angina and those in whom coronary angioplasty or coronary bypass surgery had not been performed previously were more likely to exhibit prolonged delay in the setting of AMI. On the other hand, persons with a history of hypertension or diabetes were, on average, more likely to delay seeking medical care than were patients without these histories. Inconsistent trends were observed according to Killip class at the time of hospital presentation, although patients in cardiogenic shock experienced the shortest delays. There was no marked difference in extent of delay according to geographic region of hospitalization.
Lastly, we examined demographic and clinical factors associated with prehospital delays of longer than 3 hours and longer than 6 hours after the onset of acute coronary symptoms while controlling for the influence of other factors that might affect delay (Table 4). The 3-hour time point was chosen, since most patients with AMI presenting within this window should be eligible to receive thrombolytic therapy based on present treatment recommendations and the relatively short (<1 hour) door to needle times typically seen in hospital emergency departments; patients presenting more than 6 hours after the onset of acute coronary symptoms are less likely to receive survival benefits associated with administration of clot-lysing agents. Women, minorities, and elderly individuals were significantly more likely to exhibit prolonged delay at each delay cut point. Persons with a history of diabetes and hypertension were more likely to delay seeking medical care, whereas those with a history of MI or percutaneous transluminal coronary angioplasty or who were sicker as reflected by advanced Killip class were significantly more likely to present to the hospital after a shorter delay. Persons with AMI in the Mountain and Pacific states were significantly more likely to present to the hospital soon after symptom presentation compared with individuals hospitalized in New England.
The results of this large observational study using data from the NRMI-2 suggest slight to no declines in duration of prehospital delay in patients with AMI during the study period. Average delay times were approximately 5.5 hours in 1997. Approximately one fifth of patients presented to the hospital within 1 hour of acute coronary symptoms, and slightly more than two thirds, within 4 hours of acute symptom onset. Advancing age, female sex, nonwhite race, and a number of comorbidities and clinical factors also were associated with prolonged delay overall in each of the time periods examined.
Extremely limited data, particularly from a large multiple-hospital perspective, are available to describe recent findings, as well as possible temporal trends therein, concerning duration of prehospital delay associated with AMI.
In a recent review of more than a dozen studies that have examined extent of prehospital delay in patients with acute coronary disease,11 average delay times in the setting of AMI ranged from 1.6 to 42.4 hours, whereas median delay times ranged from 1.7 to 8 hours. These wide differences in mean and median delay times are likely attributed to the varying time periods under study; to differing size and demographic, socioeconomic, and clinical characteristics of the samples under study; to differing definitions of AMI and prehospital delay used; and to other contributing factors. Few studies used a standardized questioning of patients to determine time of acute symptom onset, and difficulties and inconsistencies in medical chart documentation and patient recall may be encountered in reviewing medical records to determine time of acute symptom onset from more premonitory symptoms of acute coronary disease.
Although no consistent patterns in duration of prehospital delay could be discerned from reviewing previous studies in this area, the delay times observed in our study are consistent with those of more recent investigations and tend to be lower than those of studies performed in the late 1960s and early to middle 1980s. In a recent study of more than 2400 persons hospitalized with AMI in 37 hospitals in Minnesota from 1992 and 1993, 40% of patients delayed presentation to the hospital by more than 6 hours after the onset of AMI.12 Our results also suggest that, even in the thrombolytic era, a considerable percentage of patients present to the hospital after prolonged delay and outside of the accepted window for the greatest benefit to be gained by the administration of reperfusion therapy. Given the relatively brief 4-year study period examined, it is unclear whether changes in duration of prehospital delay might have been expected.
In one of the few studies that have examined temporal trends in prehospital delay in patients with AMI, data from the multiple-hospital population-based Worcester Heart Attack Study suggested no change in median delay times (2.0 hours) in patients hospitalized with AMI in 1986, 1988, and 1990, whereas average delay times increased slightly during these periods.11 Approximately one half of patients with AMI presented to area hospitals within 2 hours of acute symptom onset. Older age and a history of diabetes or AMI were associated with prolonged hospital presentation in the Worcester Heart Attack Study. In a previous report from that study, advancing age, AMI-associated characteristics, and in particular, duration of prehospital delay were significantly associated with failure to receive thrombolytic therapy.13
A limited number, albeit uncontrolled, of public education campaigns using a variety of educational approaches have been performed in an attempt to reduce the extent of delay in seeking medical care in persons with symptoms of acute coronary disease.14- 23 Three of these studies have been conducted in the United States and 6 in other countries. Most of these studies measured changes in delay times from before the intervention to after its implementation and did not include a concurrent comparison group. As examples of these studies, a public education campaign in Sweden resulted in reduced delay times,19,20 whereas public education campaigns in the United States and Australia have failed to alter appreciably patient delay times compared with periods before the institution of these programs.16,18,21,23 The greatest reductions in delay times were seen when a multiple-component intervention was used. The multiple-community randomized controlled Rapid Early Action for Coronary Treatment trial is examining the effects of public and health care provider educational programs on reducing the duration of prehospital delay in individuals with suspected coronary disease from 10 pair-matched communities throughout the United States.24 One half of this national cross-sectional sample of communities are receiving diverse educational intervention strategies for an 18-month period, whereas the other half are serving as reference sites. The median prehospital delay times during the 4-month baseline period of that trial, which was performed from December 1, 1995, through March 31, 1996, were 2.4 and 2.3 hours in the respective pair-matched communities, findings that are consistent with our results.
A NUMBER OF FACTORS have been associated with delay in seeking care among individuals with symptoms of acute coronary disease and have been the focus of several reviews.6- 8 These factors include sociodemographic characteristics, those associated with the context in which the acute symptoms occur and the influence that others might have on the decision to seek medical care, availability and perceived quality of medical care, cost-related issues, medical history, recognition of the pain as cardiac in origin, severity of the pain, denial-coping mechanisms, and other psychological factors.
In agreement with the relatively limited number of studies that have examined factors associated with delay, advancing age, female sex, and nonwhite race were associated with prolonged delay.6- 8,12 Most studies have, however, included only white patients, and neither socioeconomic status nor level of education has been shown to be associated with care-seeking behavior. Despite the small percentage of nonwhite patients included in our investigation, these individuals consistently exhibited longer delay patterns than white patients, placing them at increased risk for adverse outcomes in the setting of AMI and reduced likelihood of receiving reperfusion therapy.
Patients with a history of AMI and those who had undergone previous coronary angioplasty were significantly more likely to seek medical care soon after the onset of acute symptoms than were persons without such histories. These findings are in contrast to those of several studies that have shown a history of coronary heart disease or presence of cardiac risk factors to be associated with more prolonged delay times.9 On the other hand, previous MI and a history of mechanical revascularization have been associated with reduced prehospital delay.12 Since these patients should be more aware of the need for prompt medical attention based on their underlying disease state, our results are encouraging. These individuals, and particularly those who sought medical care more rapidly following the onset of acute coronary symptoms, could serve as positive advocates to other groups at risk for prolonged delay in seeking medical attention. Diabetic patients in particular, and hypertensive patients to a lesser extent, delayed seeking medical care longer than respective comparison groups. Of all the medical history characteristics examined, having diabetes was associated with the most prolonged delay in care seeking patterns. Delay patterns may be more extreme in these patients due to differences in pain perception, greater denial, or other reasons. It is unclear why hypertensive patients may delay seeking medical care to a greater extent than persons without such a history; this finding is consistent with the results observed in the study of patients with AMI hospitalized in medical centers throughout Minnesota.12 This may be related to symptom attribution, previous experiences with the medical care system, or other factors.
Irrespective of the reasons observed for the prolonged delays in seeking medical attention in these groups, persons with a history of either medical condition are at increased risk for AMI and for prolonged delay, thereby constituting a high-risk group of patients to whom educational efforts need to be directed. It appears, based on our data, that perceived or actual severity of the acute clinical presentation is associated with extent of prehospital delay; persons with a more compromised clinical state, as reflected by Killip class at the time of hospital presentation, were significantly more likely to present for care earlier after the onset of symptoms, particularly those in cardiogenic shock. This may be due to the perceived severity of the acute symptoms in these patients and/or encouragement from others to seek care quickly based on clinical signs and reported symptoms. Although no marked variations in delay times were observed according to region of hospitalization, after adjusting for other confounding factors, persons in the Mountain and Pacific regions were more likely to seek care quickly following the onset of acute coronary symptoms. This finding may be a statistical artifact based on the multiple comparisons performed or may reflect a greater awareness of the importance of seeking care rapidly following the development of suspect coronary symptoms in persons from these geographic regions.
The strengths of our study include its large sample size, multiple-center nature, diversity of patients represented, and provision of recent insights into the duration of prehospital delay among patients with AMI. Limitations include its hospital- but not population-based approach, inability to query patients directly about acute symptom onset or to ascertain the components of delay in a standardized manner, and failure to collect data on psychological, situational, or other factors that might be associated with duration of prehospital delay. Concerns might be raised about the accuracy of patient self-reports about the time of onset of acute coronary symptoms that led them to seek medical care, and the recording of these times by physicians and nurses at the large number of hospitals participating in this voluntary registry. Despite the large number and proportion of all US hospitals represented in the NRMI-2, there were differences between participating and nonparticipating registry hospitals that might limit the generalizability of our results.
The results of this large observational study provide insights into the present magnitude of prehospital delay in patients with AMI and factors associated with delay. Given the apparent lack of change in the extent of prehospital delay during the relatively brief 4-year study period, continuing population-based approaches to further reducing these delay times remain needed. Moreover, average delay times remain prolonged, and there has been little shift in the distribution of patients presenting for treatment within several hours of acute symptom onset. Also, we may have arrived at a ceiling effect about the duration of prehospital delay in persons with AMI, and it may be difficult to reduce these times appreciably given the complexity of factors involved. It is hoped that educational efforts in this area will translate to fewer out-of-hospital deaths attributed to coronary heart disease, to a greater proportion of patients with AMI receiving thrombolytic therapy and other beneficial cardiac medications, and to an increasing percentage of patients being placed under acute surveillance and care that should result in improved hospital survival.
Accepted for publication February 9, 1999.
Reprints: Robert J. Goldberg, PhD, Division of Cardiovascular Medicine, Department of Medicine, University of Massachusetts Medical School, 55 Lake Ave N, Worcester, MA 01655 (e-mail: email@example.com).