Influencing factors of AF detection in cryptogenic stroke: A survey among neurologists in China


 Background and Purpose— Atrial fibrillation (AF) detection is a cornerstone of cardioembolic stroke prevention and is especially recommended in patients initially diagnosed cryptogenic stroke, but few data exist about the extent of AF detection. This study aimed to investigate influencing factors of AF detection for cryptogenic stroke patients from neurologists’ perspective. Methods— A questionnaire survey was conducted from October 2016 to March 2018 and included neurologists from 42 hospitals in China. Respondents’ demographic characteristics, AF-related knowledge and education, attitude to AF detection and daily AF detection practice were surveyed. Pearson chi-square tests and logistic regression were used to identify neurologists characteristics independently associated with AF detection. Results— 611 neurologists were surveyed in this study and 53.0% reported always detecting AF among cryptogenic stroke patients. The main obstacles of AF detection were insufficient knowledge, high expense and inactive attitude. Those who received continuing medical education (training program) about AF detection (OR=2.25, CI[1.57, 3.23], p<0.001), who had positive attitude (OR=3.25,CI [2.24, 4.71], p<0.001) and who acquired graduate degree (OR=2.48, CI[1.21, 5.07], p=0.013) were more likely to detect AF in cryptogenic stroke patients. Neurologists who had worked in neurology over 20 years (OR=3.59, CI[1.61, 8.05], p=0.002), and who received continuing medical education (OR=1.82, CI[1.21, 2.58], p=0.001) were more likely to have an positive attitude to AF detection. Conclusions— Our survey suggested inadequate AF detection among neurologists, and continuing medical education was associated with neurologists’ attitude and AF detection behavior. More continuing medical education is required to increase the effective detection of AF and improve the quality of clinical practice.

Indeed, effective detection of AF is crucial for secondary prevention of stroke, because the detection of AF will lead to anticoagulant therapy which is superior to antiplatelet therapy in cryptogenic stroke [9]. AF related stroke is of higher risk of recurrence and bad outcomes [10][11][12], and increasing evidence from randomized trials supports that AF detection after cryptogenic stroke would decrease risk of stroke recurrent and increase anticoagulant treatment for secondary stroke prevention [13]. Thus, current guidelines emphasize prolonged cardiac rhythm monitoring for AF in cryptogenic stroke [14][15][16][17]. However, prolonged cardiac rhythm monitoring (a minimum of 24 hours) for AF are not usually used as recommended in clinical practice. Ntaios, G.et al found that only 51.6% of patients with cryptogenic stroke had a 24-hour ambulatory Holter monitoring, and 70.9% had continuous electrocardiographic (ECG) monitoring for 1 week or until discharge in Athens Stroke Registry [4]. While a survey among leads of stroke units in Germany revealed that prolonged cardiac rhythm monitoring was recommended by 90.2% in patients initially diagnosed cryptogenic stroke, but merely 13.8% of them have an 5 established pathway for routine follow-up visits [18]. The use of prolonged cardiac rhythm monitoring to detect paroxysmal AF after cryptogenic stroke is currently left to neurologists discretion.
Here, we conducted this survey to investigate AF detection status in cryptogenic stroke and its influencing factors among neurologists in China.

Methods
This cross-sectional study was an investigator-initiated survey among neurologists, which was performed from October 2016 to March 2018 in 42 hospitals in Hubei province in China. A standard-structured anonymous questionnaire was conducted face to face by trained investigators to the neurologists. Hospitals were selected and invited to participate if they met the following eligibility criteria: (1) Level III or Level II hospital with >300 beds; (2) at least 5 neurologists in the hospital. The criteria for participating neurologists were the following: (1) had acquired physician certifications and legal medical qualifications; (2) had diagnosed and treated patients with stroke; (3) agreed to complete the questionnaire.

Statistical Analyses
Descriptive analysis was performed for demographic data and obstacles to AF detection in patients with cryptogenic stroke. Pearson chi-square tests were conducted to compare influencing factor about the rate of AF detection and the attitude to AF detection between groups. Logistic regression analysis was performed to identify the predictors associated with AF detection. Adjusted odds ratios (ORs) and 95% confidence intervals (CI) for each variable were calculated, a 2-sided P<0.05 was considered statistically significant. All analyses used SPSS 23.0 for Windows. 6

Respondent demographics
A total of 611 neurologists from 38 hospitals in Hubei province in China responded to the survey for 4 hospitals declined and 99 neurologists from the participating hospitals refused the questionnaire (due to the busy clinic task or not in general clinic work). The response rate of hospital was 90.5%, and the response rate of neurologist was 86.1%. The surveyed neurologists had a mean age of 36.0 (SD=8.0) years, and 56.5% were men. 64.1% (392) of them had worked more than 5 years in the neurology and 58.3% (356) had acquired Master's degree or beyond.
Respondents demographic characteristics are summarized in Table 1. 99.0% (605) of respondents thought AF would increase the recurrence risk of stroke and 59.6% (364) of them thought paroxysmal AF was associated with higher risk of stroke than persistent AF. However, only 63.0% of them thought it necessary to detect AF and 53.0% of them always detect AF as recommended in cryptogenic stroke patients.

Major obstacle to AF detection
When surveyed the major obstacle to AF detection, 457 neurologists chose that patients were not willing or cannot tolerate to finish the detection because of their poor condition, which was an insurmountable obstacle. In the rest 154 neurologists (Figure1), 27.9% (43) respondents ascribed the major obstacle to neurologists' insufficient knowledge, 24.0% (37) attributed it to the high expenses and 18.2% (28) complained of the deficiency of equipment. The followed major obstacle in turn were: neurologists' inactive attitude (13.1%), tense doctor-patient relationship (12.3%), and others (4.5%).

7
Neurologists' characteristics associated with AF detection   AF detection in initially diagnosed cryptogenic stroke patient was well emphasized and recommended by clinical guidelines [17,[19][20][21][22]. However, only 53.0% of the respondents would always detect AF in cryptogenic stroke patients, forming a substantial gap between management recommendations and everyday clinical practice, even though almost all the neurologists in this survey knew the high recurrence risk of stroke associated with AF. Consisting result was found by Giruparajah, M. et al in a survey to physicians all over the world that only 58% of physicians globally and 36% of physician in East Asia and the Pacific region would detect AF in cryptogenic stroke patients as recommended [23]. Comparing to recent study among cardiologists, the neurologists we surveyed had a higher awareness of health threat posed by AF [24], but their perception of the relationship between AF and stroke was not accurate. 59.6% of them thought paroxysmal AF had higher risk of stroke than persistent AF, incorrectly estimating the risk of different types of AF.
Thus, there is an insufficient knowledge of neurologists about AF in developing country, and continuing medical education may be benefit to solve this problem.
The main obstacle to AF detection in patients with cryptogenic stroke was once thought to be the deficiency of detecting equipment. However, this survey revealed that deficient knowledge was the first of all the changeable causes. 27.9% of respondents regarded the insufficient knowledge as the major obstacle, and 13.1% of them attribute it to the inactive attitude, both of which could be overcome by continuing medical education. More education program about AF detection for neurologists is warranted. 12.3% of respondents attributed the major obstacle to tense doctor-patient relationship. The tense relationship may be partly caused by the misunderstanding between doctors and patients about AF screening, especially healthcare professionals, particularly general practitioners are as the most important medical information source of patients [24]. Public continuing medical education about cryptogenic stroke associated with AF will be helpful. 24.0% of them ascribed it to high expense of AF detection. Indeed, medical decision-making are always limited to patient's medical burden and affordability in China, as current government medical input only accounts for 4.6% of gross national product (GNP) for the health of 1.4 billion people, 22% of the world's population. Thus, Chinese government should increase health support, decrease medical economic burden, and improve the efficiency and quality of health insurance [25,26].
The survey revealed that positive attitude, continuing medical education (AF detection training), and master degree were important influencing factors of AF detection behavior, and the most powerful factor was neurologists' attitude (OR = 3.25, CI [2.24, 4.71], p < 0.001). Neurologists having positive attitude would more likely to do AF detection. Furthermore, positive attitude of AF detection was found to be associated with long working duration (over 20 years) and continuing medical education about AF detection. Actually, the survey highlighted that even for those who did not acquire high academic degree and who had a short working duration, continuing medical education about AF detection would promote neurologists' attitude and behavior. Neurologists receiving continuing medical education may be more familiar with current guideline recommendations, more aware of the risk of AF associated stroke, and do AF detection more frequently as recommended.
Physicians' knowledge are cornerstone for evidence based practice, effective treatment and disease prevention [27]. Previous studies also found that continuing medical education for clinicians would improve medical quality and patients outcome [28,29]. Besides, the association between positive attitude and long working duration could be partly explained by more chances to receive continuing medical education about AF. Thus, continuing medical education was most important to improve AF detection in cryptogenic stroke patients, and more continuing education about AF are required.
There are several limitations to this study. First, neurologists from each hospital was voluntary and those who responded to the surveys may be more knowledgeable about current recommendations than those declined, thus response bias may exist.

Availability of data and materials
The data that support the findings of this study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The study was conducted in strict accordance with the Declaration of Helsinki and its later amendments and was approved by the Research Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology. Each neurologists were appropriately informed about the study aims and completed an informed consent form before participating in the study.

Consent for publication
Not applicable.  Figure 1 The Major Obstacle to Detect AF