Perceptual, Social, and Behavioral Factors Associated With Delays in Seeking Medical Care in Patients With Symptoms of Acute Stroke
Background and Purpose— Despite availability of reperfusion therapy for acute ischemic stroke, most patients remain ineligible mainly because of late hospital arrival. We hypothesized that perceptual, social, and behavioral factors affect delays in seeking help after symptom onset.
Methods— Patients presenting with stroke symptoms were interviewed about symptom experiences, interpretations, and reactions. Odds ratios (95% CI) for risk of delay >3 hours were estimated, and variables associated with increased risk and representing demographic, clinical, perceptual, social, and behavioral factors were included in an assessment of the effect of combined risk factors on delay.
Results— Among 209 patients (mean age 61.8±12 years, 69% men) the median time interval from symptom awareness to seeking help was 2 (0.5 to 9) hours and to hospital arrival, 4.2 (1.3 to 14.5) hours. On multivariate adjustment, perceiving symptoms as severe (odds ratio [OR]: 0.42; 0.17 to 0.95), advice from others to seek help (OR: 0.18; 0.05 to 0.63), and contacting an ambulance (OR: 0.26; 0.10 to 0.63) were associated with decreased risks of delay, whereas perceived control of symptoms (OR: 2.45; 1.08 to 5.71) increased risk of delay in seeking help. Risk of delay in hospital arrival was 3 times greater in women than in men. Increasing proportions of patients who delayed seeking help were observed with increasing numbers of combined risk factors, ranging from 17% to 94% for 0 to 1 and 6 to 7 factors, respectively.
Conclusions— Perceptual, social, and behavioral factors contribute to delay in seeking medical care in acute ischemic stroke beyond demographic and clinical variables, and, when combined, further increase risk of delay. These findings may be important for designing programs to reduce delay.
Despite availability of reperfusion therapy for treatment of acute ischemic stroke, most stroke patients remain ineligible, mainly because of delays in seeking help after symptom onset.1,2 Attempts to increase public awareness of the need to respond quickly to stroke symptoms through educational programs have had limited success. Knowledge of stroke symptoms has not been consistently predictive of early hospital arrival.3,4 Therefore, providing the public with information about stroke symptoms may be insufficient, and a more comprehensive understanding of the dynamics involved in seeking help is needed. We hypothesized that perceptual, social, and behavioral factors affect delays in seeking help following symptom onset.
Subjects and Methods
The study was conducted at a university affiliated public hospital that serves as a tertiary care center, with an immediate catchment area of about 500 000 people. The hospital’s review board approved the conduction of the study. All patients hospitalized in the neurology department between September 2000 and December 2002 with a diagnosis of ischemic stroke according to the World Health Organization (WHO) definition5 were candidates for inclusion. Ischemic stroke was differentiated from intracerebral hemorrhage based on the brain CT performed at the acute stage. Exclusion criteria were inability to estimate the time of symptom awareness and the time of first contact with medical personnel, inability to communicate or dementia, with no family member available to provide information, and refusal to provide informed consent. There were also patients who were not interviewed for logistic reasons. Included patients participated in structured interviews, conducted 2 to 10 days after hospital admission. The interview was designed to collect data on demographic and clinical characteristics, cardiovascular risk factors, the context in which the stroke occurred, cognitive, affective, and behavioral responses to the stroke, the response of others to the event, prior knowledge about stroke, and several personality traits. For patients whose neurological condition did not allow personal participation, but for whom a family member was available, partial interviews (ie, demographics, medical history, context in which symptom onset occurred, and type of medical personnel contacted) were conducted (n=49). Hospital charts were used to complete data and verify the information reported by the participants. Instruments and classification systems used included the National Institutes of Health Stroke Scale (NIHSS)6 for assessment of stroke severity, the Oxford classification of clinical subtypes,6 the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification for stroke etiology,6 and an adaptation of the Revised NEO (Neuroticism, Extraversion, and Openness) Personality Inventory7 for a brief assessment of several facets of personality.
Time of stroke awareness was defined as the moment when either the patient or someone else first noticed the symptoms, because symptom awareness represents the first time when help could be sought, which might be the most appropriate starting point when examining the patient’s reaction.
Statistical analysis was performed using SAS statistical software.8 The end points analyzed in the study included the time interval from symptom awareness to seeking medical help (the time of the patient’s first contact with any type of medical personnel) and to hospital arrival. Because the distribution of time was skewed, nonparametric testing was performed, using the Wilcoxon rank-sum test and the Kruskal-Wallis test. All tests were 2-tailed, and a value of P<0.05 was considered statistically significant. Multivariate analysis was performed using stepwise logistic regressions, which estimated the association of study variables with the risk of delaying the seeking of help and arriving at hospital for >3 hours. Variables chosen for the analyses included age and sex, those for which statistically significant differences were observed in univariate analysis, and those recognized in the literature as being associated with the end points under investigation. The extent to which the models were able to discriminate between patients who delayed seeking help and those who didn’t was assessed by the area under the receiver operating characteristic plot (c statistic). The Hosmer and Lemeshow goodness-of-fit test evaluated model fit. Multivariate analysis of delay in seeking help included patients who had been able to seek help on their own (n=179), whereas multivariate models for delay in hospital arrival included all enrolled patients. In view of a 3-fold risk of delay in hospital arrival observed among women, gender-specific analyses were conducted. Furthermore, a group of variables identified as “risk factors” for delay were studied in combination to determine the effect of increasing numbers of risk factors on the proportion of patients delaying the seeking of help.
Among 518 patients with ischemic strokes admitted to the neurology department, 209 patients were included in the study. A comparison of included patients with all ischemic stroke patients hospitalized in the neurology department showed that included patients were comparable to all neurology patients with respect to age (62±12 versus 63±13), gender (69% versus 67% men) and stroke severity.
Seventy-five percent of the study patients were married, 86% didn’t live alone, and 50% were employed at the time of the stroke. Regarding education, 24% had completed 8 years of school, 46% had a secondary education, and 30% reported ≥13 years of schooling. The most prevalent risk factor was hypertension (65%), followed by dyslipidemia (50%). Nineteen percent of the patients had experienced a previous stroke, whereas 34% had a history of cardiac disease.
The mean time interval from symptom awareness to seeking help was 8.4±15.7 hours (median 2 hours, interquartile range: 0.5 to 9), and from symptom awareness to hospital arrival 15.3±28.3 hours (median 4.2 hours, interquartile range: 1.3 to 14.5). The majority of events occurred during the day (78%), when the patient was home (70%), and when accompanied (80%). Half of the patients reported knowing at least 1 stroke symptom before the event, and the symptoms most familiar to them were motor weakness (83%) and speech disturbances (58%).
Analysis of the time interval from symptom awareness to seeking help, by demographic characteristics, showed no significant differences with respect to age, sex, family status, work status, education, or living alone versus accompanied. In the analysis of time to hospital arrival, older patients (age >70) arrived earlier than younger patients (age <50).
Table 1 shows the distribution of patients by clinical characteristics. Shorter time intervals were associated with sudden onset of event, motor weakness, speech disturbance, total anterior circulation infarctions, and strokes of cardioembolic etiology, whereas longer time intervals were associated with experiencing dizziness or headaches, and having relatively mild strokes. These patterns characterized the time intervals to both seeking help and hospital arrival.
Table 2 shows the distribution of patients by selected cognitive, affective, social, and behavioral factors. Perceiving control over symptoms (feeling able to control symptoms to some extent), attributing symptoms to problems other than stroke, hesitations about seeking help (verbalization of apprehension about seeking medical help), and an initial reaction of self-treatment or waiting to see what happens were associated with longer time intervals for both end points. Factors associated with shorter time intervals included someone else other than the patient first noticing the symptoms, advice by others to seek help, and initial contact with an emergency medical service (EMS). Perceiving symptoms as severe (patient regarded symptoms to be at least somewhat severe) also shortened the time interval to some extent. No significant differences in median time intervals were observed when comparing those familiar with at least 1 stroke symptom before the event (51%) and counterparts who weren’t.
Analysis of the immediate reaction of the patient to his symptoms showed that patients classified as having “high” general anxiety levels were less likely to self-treat (21%) than those with low and average anxiety levels (36% and 43%, respectively). These patients preferred to seek help rather than wait.
Multivariate analysis (Table 3) indicated that early help seekers more often experienced a sudden onset of symptoms, perceived their symptoms as severe, were advised by others to seek help, and contacted or were transported to the hospital by EMS. The contribution of “high” general anxiety levels to seeking help early was of borderline significance. Delays in seeking help were more likely to occur when patients perceived control over their symptoms.
Being a woman was associated with a 3-fold risk of delay in reaching the hospital. Delay in hospital arrival was also associated with experiencing fatigue, and the patient being the first to notice his symptoms. Early arrival at the hospital was associated with sudden onset of symptoms, attribution of symptoms to stroke, the patient being advised to seek help, and contact with or transport by EMS.
When an additional model, including the NIHSS score and excluding variables related to the score, was tested using 3 subgroups (score ≤5, 6 to 10 and >10), a score of >10 was predictive of early hospital arrival (OR: 0.19; 95% CI: 0.05 to 0.59). None of the other variables included in this model demonstrated statistically significant associations with delay. Examination of the relationship between NIHSS scores and the patient’s perception of the severity of his symptoms showed no association between the 2. However, the proportions of patients perceiving control over their symptoms declined with increasing NIHSS scores (38%, 24%, and 10% for patients with scores ≤5, 6 to 10, and >10, respectively; P=0.02).
We attempted to identify factors that could explain the increased delay among women by conducting a series of gender-specific analyses. Among the gender differences observed was a greater proportion of women who lived alone (10% versus 25%; P=0.004). Among patients who first contacted a physician, a greater proportion of men were referred immediately to the emergency department (77% versus 62%; P=0.02). Data on risk of a delay >3 hours in hospital arrival associated with selected variables by gender are presented in Table 4. Although fatigue was associated with delay in both sexes, perceived control of symptoms was associated with a >5-fold risk of delay in women. In multivariate analysis, the effect on delay of fatigue and perceived control of symptoms was maintained among women, whereas delay in men was associated with symptoms being first noticed by the patient.
An analysis focusing on specific risk factors for delay in seeking help (age ≤70, male, NIHSS ≤5, didn’t contact EMS, perceived control over symptoms, onset not sudden, not advised to seek help) showed that as the number of combined risk factors in patients increased, the time from symptom awareness to seeking help also increased, and increasing proportions of patients delayed the seeking of help for >3 hours (Figure).
The main focus of the majority of research on delayed presentation of stroke patients has been the identification of demographic and clinical characteristics of patients who delay seeking treatment. We have found that the risk of delay is affected not only by demographic and clinical variables, but also by perceptual, social, and behavioral factors. This is demonstrated by the association between the perception of symptom severity, advice from others to seek help and perception of control of symptoms and the delay in seeking help.
This approach has also been applied in the study of patients with acute coronary syndrome, suggesting that the patient’s perception of his symptoms, his appreciation of their meaning and severity, and the realization that medical help is required are factors which educational programs could address when attempting to facilitate appropriate reactions to symptom onset.9,10,11
The time taken by the patient to react to his symptoms has been shown to constitute a major proportion of the total out-of-hospital delay,12 and shorter “reaction intervals” have been observed when someone else identified the problem, EMS was contacted, patients were transported to the hospital by EMS, and symptoms were considered “most urgent”. Other studies have shown shorter delays among patients who attributed their symptoms to stroke,13,14 older patients,14 those in whom symptom onset was sudden and remained stable,14 and patients with severe strokes.15
Our data suggest that the patient’s perception of the severity of his symptoms, and even more so, the extent of perceived control over them, influence “reaction times”. The strong association found in our study between advice by others to seek help and shorter reaction times emphasizes the importance of intervention by family members or significant others. Other findings of interest include the 3-fold delay in hospital arrival observed among women, and the possible role of personality traits, such as “high anxiety”, in the decision process.
We found an association between increasing numbers of risk factors representing demographic, clinical, perceptual, social, and behavioral variables and risk of delay. This finding emphasizes the multidimensional nature of the decision-making process with which patients are confronted from the moment of symptom awareness, and should be considered when designing interventions aimed at reducing delay.
A limitation of our study is that patients with intracerebral hemorrhage were not included, but given the emerging introduction of ultra-early hemostatic therapy,16 delays in intracerebral hemorrhage should be examined as well. As the study was not powered to assess gender differences and the number of women is small, the analysis of reasons for the observed increased delay in hospital arrival among women should be viewed as hypothesis generating. Because of the selection criteria and logistic reasons, only a proportion of patients admitted during the whole study period were included, but the characteristics of these patients were comparable to counterparts not included.
The study of perceptual, social, and behavioral factors is obviously complex and deserves further research. The implications of our findings are that educational efforts to reduce the time interval between symptom onset and seeking medical assistance must extend beyond providing information about symptoms and the need to call EMS, and should address the problem of symptom interpretation and perception, and the need for the intervention of others.
- Received October 10, 2005.
- Revision received January 12, 2006.
- Accepted February 2, 2006.
Barber A, Zhang J, Demchuk AM, Hill MD, Buchan AM. Why are stroke patients excluded from tPA therapy? An analysis of patient eligibility. Neurology. 2001; 56: 1015–1020.
Williams LS, Bruno A, Rouch D, Marriott DJ. Stroke patients’ knowledge of stroke: influence on time of presentation. Stroke. 1997; 28: 912–915.
Kothari R, Sauerbeck L, Jauch E, Broderick J, Brott T, Khoury J, Liu T. Patients’ awareness of stroke signs, symptoms and risk factors. Stroke. 1997; 28: 1871–1875.
Thorvaldsen P, Asplund K, Kuulasmaa K, Rajakangas AM, Schroll M. Stroke incidence, case fatality, and mortality in the WHO MONICA project. World Health Organization monitoring trends and determinants in cardiovascular disease. Stroke. 1995; 26: 361–367.
Herndon R. Handbook of neurologic rating scales. New York: Demos Vermande, 1997.
Costa PT, McCrae RR. Revised NEO Personality Inventory (NEO PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual, Odessa, Florida, Psychological Assessment Resources, Inc. 1992.
SAS Institute Inc. SAS/STAT Users’ Guide, Version 8, Cary, NC: SAS Institute Inc, 1999; 3884.
Barch CA. Seeking health care following stroke: public education. Proceedings of a national symposium on rapid identification and treatment of acute stroke. Bethesda, Md: National Institute of Health, 1997.
Feldmann E, Gordon N, Brooks JM, Brass LM, Fayad PB, Sawaya KL, Nazareno F, Levine SR. Factors associated with early presentation of acute stroke. Stroke. 1993; 24: 1805–1810.
Chang K, Tseng M, Tan T. Pre-hospital delay after acute stroke in Kaohsiung, Taiwan. Stroke. 2004; 35: 700–704.