| CJEB | Surveillance is Lunacy: Benchmark testing the Alberta Real-Time Syndromic Surveillance Net using urgent care requests for mental health concerns and full moons. |
Emergency Department Visits and the Synodic Cycle
Authors
Sikora CA1,4, Talbot J2,4, Fan S3 and Predy G3
1Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada
2Alberta Health and Wellness, Edmonton, Alberta, Canada
3Alberta Health Services, Edmonton, Alberta, Canada
4School of Public Health, University of Alberta, Edmonton, Alberta, Canada
Introduction
Individuals with mental health concerns often seek assistance from urgent care and emergency departments. There is an underlying belief amongst health care workers that there is a linkage between such behaviours and the lunar cycle. The authors demonstrate the ability of a syndromic surveillance system, built on patient-generated ‘reason for visit’ data, to discern an association between the phase of the moon and number of visits for mental health reasons.
Methods
The Alberta Real Time Syndromic Surveillance Net (ARTSSN) was used to analyse visits to the Emergency Department (ED) and calls to a 24-hour telephone-based health information and advice service (Health Link Alberta - HLA). The data collection period was from January 1, 2005 to December 31, 2009 in the catchment area of the City of Edmonton, Alberta. Visits to the ED were classified by the individual’s reported mental health concern and calls to HLA were concerning suicide. These visits and calls were then correlated to the synodic (lunar) cycle.
Results
No association was found between the phase of the moon and visit to the ED for mental health concern either by category of complaints or by overall visits (P>0.73), or between phase of moon and calls to HLA for suicidal intent (P>0.17).
Discussion
This study successfully demonstrated the lack of association between the synodic cycle and mental health visits reporting to the ED and telephone-based health information line using a syndromic surveillance system. It is also the first study, using patient-generated complaint data, to investigate this association. It also generated a benchmark number of visits/calls per day for each syndrome type. This demonstration serves as an example of the potential factors that a syndromic surveillance system can monitor to aid Public Health departments in directing responses to signals generated in the system.
KeyWords
Surveillance, Syndromic Surveillance, Mental Illness, Emergency Department
Introduction
Individuals with pressing mental health concerns often seek assistance from emergency departments and other urgent care sources. Amongst workers in these centres, the following conversation can be heard on an all-too-frequent basis:
Person 1: “It’s been an unusually busy day….”
Person 2: “You’re right…it must be a full moon.”
The first statement is often uttered out of frustration and may be used as a mechanism to relieve stress after emotionally-exhausting shifts. In the latter part of this somewhat typical interchange, probability would dictate that on a daily basis, the second statement is correct 3.4% of the time (1/29.53 days). But linking the two statements together is a more complex task. Studies have shown that many health care professionals are indeed superstitious and have a belief that the lunar cycle impacts behaviours. (1) Does the state of the lunar cycle influence the health seeking behaviours of patients? The impact of the moon on health has been previously discussed by Raison et al. (2) and Iosif & Ballon (3) with an exploration of the history and various hypotheses as to this mechanism of action. Previous studies, using retrospective linked health outcome data, have not been successful in demonstrating the association between the lunar cycle and trauma admissions, violent behaviour, completed suicide or psychiatric admissions. (4-7) Even with all evidence suggesting a lack of association, there remains a continued belief amongst health care professionals that there is a link between individual behaviours and the synodic (lunar) cycle. (8)
The hypothesis that there is no association between synodic cycles and health seeking behaviours is the basis for this article demonstrating the benchmark testing of the Alberta Real-Time Syndromic Surveillance Net (ARTSSN) within Alberta Health Services – Edmonton Zone region. This surveillance system, unique in its ability to use patient-generated symptoms, monitors health-seeking behaviours on a real-time basis and may be sensitive and specific enough to identify differences for such behaviours according to the lunar cycle. Given the previous data reporting no association, is it lunacy to consider that a real-time syndromic surveillance system built on generalized patient-generated ‘reason for visit’ (as opposed to a retrospective health outcome process) can show an association between the phase of the moon and number of visits for mental health reasons? The information presented here will be used to generate a benchmark level of visits for the mental health visits in order to develop automated action/response cues within the real-time surveillance system for population health response.
Methods
The Alberta Real Time Syndromic Surveillance Net (ARTSSN) is an automated, electronic, real time syndromic surveillance system that is designed for public health surveillance of injuries, toxic exposures, notifiable diseases and chronic conditions. This system has been in pilot operation since August 2008 in Alberta Health Services – Edmonton Zone. This zone comprises a catchment population of slightly more than one million individuals. Data streams collected and analyzed in this surveillance system include: visits to Emergency Departments (ED), calls to Health Link Alberta (HLA, a 24- hour nurse-staffed telephone health information and advice service), laboratory tests and primary school absenteeism in the City of Edmonton and surrounding area. Information is collected, analyzed, interpreted and disseminated in real-time to individuals for public health decision-making. Real-time data can be presented in pre-configured ‘dashboards’ on desktop screens. Individual level data is available to decision makers, but for the purposes of this analysis, the anonymous, unlinked number of visits/calls for a specified health concern per day was utilized.
Data were obtained for ED visits for mental health concerns and calls to HLA for suicidal intent. The reason for ED visits was categorized by particular mental health concerns, as reported by the patient (Table 1). These codes correspond to self-reported behaviours that may require urgent attention (Group 1: intent of self harm, homicide, violence, etc…), possible short-term stay/assessment (Group 2: confusion, insomnia, etc…) or outpatient management (Group 3: social concerns, etc…). These categories were used to stratify the severity and urgency of ED visits for mental health and are based on the Canadian Emergency Department Information System (CEDIS) Presenting Complaint List (Version 1.1). (9) Calls to HLA were collected based on the individuals’ reason for calling and, given the seriousness of suicidal intent; this factor was chosen as the indicator to measure for HLA over the study period. The data collection period was from January 1, 2005 to December 31, 2009. For this purposes of this analysis, the data is anonymous and was unlinked from outcome data.
The precise date of the full moon was obtained from the US Naval Observatory and corrected to Edmonton local time. (10) The time of the visit, or call, was classified as occurring in one of five time periods: on the day of a full moon (FMD), three days prior to (waxing Gibbous) or after (waning Gibbous) the full moon, 14 days before the full moon (new moon) and on any day other than the above. The average daily ED visits or HLA calls for each of these time periods were compared using a one-way ANOVA, with an F*-test (11) in concern of sample size differences among the five time periods, using STATA version 9.1 (Stata Corp., College Station, TX, USA). The periodicity of the ED visits and HLA calls was examined on smoothed periodigrams, which were obtained after the raw data were first demeaned and detrended, and the resulting sample periodigrams were then smoothed using modified Daniel smoothers. (12) This time series analysis was done in R. (13)
Results
Throughout the study period, there were 62 full moon days identified. There were a total of 48,314 visits to the ED for mental health reasons and 855 calls to HLA for suicidal intent. This represents a total of 34299 individual patients during the study period. Of the visits, only 381 repeated visits occurred on the same day by the same individual, representing 0.78% of the total visits. No statistically significant difference could be found between phase of moon and visit to the ED for mental health concern either by overall visits (P>0.73, data not shown), or by category of complaints (Table 2), between phase of moon and calls to HLA for suicidal intent (Table 3, P>0.17). The periodigrams of either overall visits for mental health (Figure 1, upper panel) or category of complaints (data not shown) showed no peak near the frequency of 0.0339 (corresponding to 29.53 days). The two non-significant peaks at frequencies 0.1424 and 0.2859 corresponded to 7 and 3.5 days respectively, indicating a day of week effect with Mondays and Wednesdays having slightly more visits than Saturdays and Sundays (Table 4). This table (Table 4) represents the baseline number of visits for mental health concerns expected to the ED by day of week. On the periodigram of HLA calls for suicide intent, no suggestive peak of any sort of a cycle could be identified (Figure 1 bottom panel).
Discussion
As demonstrated in this example, the ARTSSN syndromic surveillance system implemented in Alberta Health Services – Edmonton Zone can use patient-generated reason for visits or calls for mental health concerns as an indicator for events related to the synodic cycle. These self-reported indicators are non-specific in nature, and may lack the sensitivity to capture specific diagnosis, but do identify major syndrome categories requiring assistance/care. There was no identified association between the number of visits to the ED for mental health reasons, or calls to HLA for suicidal intent, and the lunar cycle. The results presented are in keeping with the conclusions from previously identified outcome-linked studies. However, this study is the first which utilizes patient-generated data in a real-time manner.
The strength of this study resides in its rapid reporting capability and ability to capture complex presentations of highly variable illness on continuous basis, as indicated by the various categories of mental illness presented. Such generalized factors serve as a signal for more in-depth study (for instance, a potential for sub-group analysis of not previously identified homicidal behaviour). The authors assume a stable population over the study period with a consistent use of the reported ‘reason for visit’ by ED and HLA staff. Given the anonymous nature of HLA calls, it is beyond the scope of the present study to determine if individuals requesting assistance of HLA report to the ED. Likewise, there is potential for a single individual to provide for multiple signals over the span of one day, although analysis of the surveillance data suggests that this occurs less than 0.8% of the time. The widespread implementation of this syndromic surveillance system could be strengthened with the standardization of visits to age/gender and other, potentially more stable, health seeking syndromic indicators. Accumulating baseline data is important to decrease the probability of false positives after more widespread implementation of the surveillance system. Over the study period, there was a baseline number of visits/calls per day identified by lunar cycle period and by day of week. Monitoring the number of visits to urgent care and emergency departments related to mental health concerns can have ramifications for both health system planning purposes as well as Public Health responses to community events. The methods utilized in creating this study can be applied to study effects of other factors on the health-seeking behaviours of patients in the community. Given the baseline generation, the ARTSSN system could quickly identify changes related to otherwise unknown factors in the community, triggering a Public Health response to initiate mitigation and restore health in the population. Resources may be rapidly allocated in geographic areas noted to be experiencing increased demand. If, in the future, there is a rise in the number of visits to the ED for mental health concerns on a night of a full moon (or any other night for that matter), Public Health will be able to detect the change and develop a response.
Conclusion
This study investigated the association between the synodic cycle and mental health visits reporting to the emergency department and telephone-based health information line. It also generated a benchmark number of visits/calls per day for each syndrome type for the region. There was no association identified between the lunar cycle and identified categories of mental health presentations. This is the first study to investigate this association using patient-generated complaint data. This demonstration serves as an example of the potential factors that a syndromic surveillance system can monitor to aid Public Health departments in directing responses to generated signals in the system and provides for a framework of benchmark development and signal analysis for other indicators. Overall, the authors conclude that is not lunacy to use a syndromic surveillance system to monitor the health of a population.
.jpg)
.jpg)
.jpg)