Application of change point analysis to daily influenza-like illness emergency department visits

Investigation of spontaneous abortion and stillbirth adverse events in epilepsy patients treated with levetiracetam: A pharmacovigilance study

BACKGROUND

The prescription of levetiracetam during pregnancy has become more common due to its lower teratogenic risk profile. However, due to a lack of data about its association with stillbirth and spontaneous abortion, worries remain.

OBJECTIVE

To investigate information on any possible association of spontaneous abortion and stillbirth adverse events with levetiracetam in women with epilepsy.

METHODS

This retrospective pharmacovigilance study used disproportionality analysis to detect signals of adverse reaction of interest reported with Levetiracetam in FAERS, the FDA Adverse Event Reporting System. The ratio of reporting odds (ROR) and information component (IC) indices were used to undertake disproportionality analyses, and change point analyses were carried out to identify variations in the frequency of reporting of relevant adverse events. Sensitivity analyses included subgroup analyses by indication, treatment regimen, and reporting region.

RESULTS

Overall, 2870 cases of spontaneous abortion and stillbirth with commonly used antiseizure medications were analyzed. A total of 65.5 % of these cases had epilepsy as the indication. In the entire dataset, we observed disproportionality signals of spontaneous abortion for 6 ASMs (levetiracetam, carbamazepine, lamotrigine, oxcarbazepine, topiramate, valproic acid) and disproportionality signals of stillbirth for 4 ASMs (levetiracetam, carbamazepine, lamotrigine, oxcarbazepine). In the epileptic population, disproportionality signals for stillbirth (ROR0.25 = 4.60; IC0.25 = 1.30) and spontaneous abortion (ROR0.25 = 3.98; IC0.25 = 1.20) in levetiracetam was identified. These disproportionality signals have been consistently robust over the past years, according to a temporal assessment of them. Sensitivity studies proved how reliable the findings were.

CONCLUSION

Using validated pharmacovigilance methods, we found significant disproportional signals for spontaneous abortion and stillbirth associated with levetiracetam. Of these, the signals for spontaneous abortion were observed after 2011 and for stillbirth after 2014, which may be related to the rise in levetiracetam prescriptions during pregnancy in recent years. The association of spontaneous abortion and stillbirth adverse events with levetiracetam and potential biases confounding this association merit further investigation.

Tracking changes in opioid prescriptions dispensed following the enactment of a prescription drug monitoring program use mandate

BACKGROUND

Prescription drug monitoring programs (PDMPs) are state-based surveillance tools used to track controlled substances dispensed to patients and identify patients at-risk of misuse. Starting April 2017, Wisconsin required all prescribers access PDMP to review patient information before issuing a controlled substance prescription order for more than a 3-day supply. A primary goal of PDMP use mandates is to reduce avoidable prescribing and mitigate opioid related mortality and morbidity. Current literature has not evaluated the existence of a time point post-policy implementation, at which the trend in opioid dispensing changes, reflecting normalization/maintenance of opioid prescribing.

OBJECTIVE

We sought to evaluate the impact of the PDMP use mandate on trends in opioid prescriptions dispensed and test a hypothesis that a change or inflection in opioid prescriptions dispensed occurred post-mandate implementation.

METHODS

Interrupted Time Series Analysis (ITSA) design was used to examine whether the level (immediate impact) and trend in opioid prescribing changed significantly after the PDMP use mandate was implemented. We used a novel Change Point Analysis (CPA) approach to test the hypothesis i.e., identify if and when a change or inflection in opioid dispensing trend occurred after implementation of the PDMP use mandate.

RESULTS

ITSA model results showed a significant drop in opioid prescriptions dispensed (p < 0.05) immediately after the mandate implementation (i.e., April 2017). Results of the CPA identified a significant inflection in opioid prescriptions dispensed starting January 2019 (21-months post-policy implementation). An ITSA model using the inflection point as an interruption showed that the trend in opioid prescriptions dispensed became flatter after the inflection point, suggesting normalization.

CONCLUSION

Using a novel CPA approach, the findings showed an inflection in the trend in opioid prescriptions dispensed post-PDMP use mandate implementation, implying that most of the avoidable prescribing likely was curtailed. The results suggest that the patient information presumably accessed from the WI PDMP interface was useful in helping prescribers to make an informed clinical decision about opioid prescribing.

Multiple change point analysis of hepatitis B reports in Xinjiang, China from 2006 to 2021

Objective

Hepatitis B (HB) is a major global challenge, but there has been a lack of epidemiological studies on HB incidence in Xinjiang from a change-point perspective. This study aims to bridge this gap by identifying significant change points and trends.

Method

The datasets were obtained from the Xinjiang Information System for Disease Control and Prevention. Change points were identified using binary segmentation for full datasets and a segmented regression model for five age groups.

Results

The results showed four change points for the quarterly HB time series, with the period between the first change point (March 2007) and the second change point (March 2010) having the highest mean number of HB reports. In the subsequent segments, there was a clear downward trend in reported cases. The segmented regression model showed different numbers of change points for each age group, with the 30-50, 51-80, and 15-29 age groups having higher growth rates.

Conclusion

Change point analysis has valuable applications in epidemiology. These findings provide important information for future epidemiological studies and early warning systems for HB.

Efficient algorithms for real-time syndromic surveillance

OBJECTIVE

Outbreaks of influenza-like diseases often cause spikes in the demand for hospital beds. Early detection of these outbreaks can enable improved management of hospital resources. The objective of this study was to test whether surveillance algorithms designed to be responsive to a wide range of anomalous decreases in the time between emergency department (ED) presentations with influenza-like illnesses provide efficient early detection of these outbreaks.

METHODS

Our study used data on ED presentations to major public hospitals in Queensland, Australia across 2017-2020. We developed surveillance algorithms for each hospital that flag potential outbreaks when the average time between successive ED presentations with influenza-like illnesses becomes anomalously small. We designed one set of algorithms to be responsive to a wide range of anomalous decreases in the time between presentations. These algorithms concurrently monitor three exponentially weighted moving averages (EWMAs) of the time between presentations and flag an outbreak when at least one EWMA falls below its control limit. We designed another set of algorithms to be highly responsive to narrower ranges of anomalous decreases in the time between presentations. These algorithms monitor one EWMA of the time between presentations and flag an outbreak when the EWMA falls below its control limit. Our algorithms use dynamic control limits to reflect that the average time between presentations depends on the time of year, time of day, and day of the week.

RESULTS

We compared the performance of the algorithms in detecting the start of two epidemic events at the hospital-level: the 2019 seasonal influenza outbreak and the early-2020 COVID-19 outbreak. The algorithm that concurrently monitors three EWMAs provided significantly earlier detection of these outbreaks than the algorithms that monitor one EWMA.

CONCLUSION

Surveillance algorithms designed to be responsive to a wide range of anomalous decreases in the time between ED presentations are highly efficient at detecting outbreaks of influenza-like diseases at the hospital level.

Evaluation of the application of sequence data to the identification of outbreaks of disease using anomaly detection methods

Anomaly detection methods have a great potential to assist the detection of diseases in animal production systems. We used sequence data of Porcine Reproductive and Respiratory Syndrome (PRRS) to define the emergence of new strains at the farm level. We evaluated the performance of 24 anomaly detection methods based on machine learning, regression, time series techniques and control charts to identify outbreaks in time series of new strains and compared the best methods using different time series: PCR positives, PCR requests and laboratory requests. We introduced synthetic outbreaks of different size and calculated the probability of detection of outbreaks (POD), sensitivity (Se), probability of detection of outbreaks in the first week of appearance (POD1w) and background alarm rate (BAR). The use of time series of new strains from sequence data outperformed the other types of data but POD, Se, POD1w were only high when outbreaks were large. The methods based on Long Short-Term Memory (LSTM) and Bayesian approaches presented the best performance. Using anomaly detection methods with sequence data may help to identify the emergency of cases in multiple farms, but more work is required to improve the detection with time series of high variability. Our results suggest a promising application of sequence data for early detection of diseases at a production system level. This may provide a simple way to extract additional value from routine laboratory analysis. Next steps should include validation of this approach in different settings and with different diseases.

Uncovering COVID-19 infection determinants in Portugal: towards an evidence-based spatial susceptibility index to support epidemiological containment policies

BACKGROUND

COVID-19 caused the largest pandemic of the twenty-first century forcing the adoption of containment policies all over the world. Many studies on COVID-19 health determinants have been conducted, mainly using multivariate methods and geographic information systems (GIS), but few attempted to demonstrate how knowing social, economic, mobility, behavioural, and other spatial determinants and their effects can help to contain the disease. For example, in mainland Portugal, non-pharmacological interventions (NPI) were primarily dependent on epidemiological indicators and ignored the spatial variation of susceptibility to infection.

METHODS

We present a data-driven GIS-multicriteria analysis to derive a spatial-based susceptibility index to COVID-19 infection in Portugal. The cumulative incidence over 14 days was used in a stepwise multiple linear regression as the target variable along potential determinants at the municipal scale. To infer the existence of thresholds in the relationships between determinants and incidence the most relevant factors were examined using a bivariate Bayesian change point analysis. The susceptibility index was mapped based on these thresholds using a weighted linear combination.

RESULTS

Regression results support that COVID-19 spread in mainland Portugal had strong associations with factors related to socio-territorial specificities, namely sociodemographic, economic and mobility. Change point analysis revealed evidence of nonlinearity, and the susceptibility classes reflect spatial dependency. The spatial index of susceptibility to infection explains with accuracy previous and posterior infections. Assessing the NPI levels in relation to the susceptibility map points towards a disagreement between the severity of restrictions and the actual propensity for transmission, highlighting the need for more tailored interventions.

CONCLUSIONS

This article argues that NPI to contain COVID-19 spread should consider the spatial variation of the susceptibility to infection. The findings highlight the importance of customising interventions to specific geographical contexts due to the uneven distribution of COVID-19 infection determinants. The methodology has the potential for replication at other geographical scales and regions to better understand the role of health determinants in explaining spatiotemporal patterns of diseases and promoting evidence-based public health policies.

COVID-19 case-fatality variations with application to the Middle East countries

During a pandemic outbreak, it is important for health officials to know the proportions of deaths among infected individuals and to understand how these proportions change overtime, to accurately predict the impact of the pandemic and to implement effectively new intervention policies and health protocols and to adjust them accordingly. However, most studies where efforts have been made to estimate accurately the case fatality rates did not address the issue of measuring the dynamics of the pandemic deadliness during its course. Daily data on COVID-19 cases and deaths were collected from selected MENA countries. In this paper, two new measures of the pandemic fatality are developed based on the estimated time it takes hospitalized infected patients to eventually die from the disease. The first measure assigns COVID-19 deaths to its most significant lagged number of cases based on a fixed-effects panel data model. The second fatality measure relates pandemic deaths and cases based on their respective change points. The results find notable variations of the pandemic lethality between the Middle East countries, likely due to the difference in the quality of health care. Although crude case-fatality rate does not identify the pandemic lethality variations during the ongoing of the disease, this paper develops two novel measures for COVID-19 case fatality which can identify the dynamics and the variations of the pandemic deadliness.

Multimodal driver state modeling through unsupervised learning

Naturalistic driving data (NDD) can help understand drivers' reactions to each driving scenario and provide personalized context to driving behavior. However, NDD requires a high amount of manual labor to label certain driver's state and behavioral patterns. Unsupervised analysis of NDD can be used to automatically detect different patterns from the driver and vehicle data. In this paper, we propose a methodology to understand changes in driver's physiological responses within different driving patterns. Our methodology first decomposes a driving scenario by using a Bayesian Change Point detection model. We then apply the Latent Dirichlet Allocation method on both driver state and behavior data to detect patterns. We present two case studies in which vehicles were equipped to collect exterior, interior, and driver behavioral data. Four patterns of driving behaviors (i.e., harsh brake, normal brake, curved driving, and highway driving), as well as two patterns of driver's heart rate (HR) (i.e., normal vs. abnormal high HR), and gaze entropy (i.e., low versus high), were detected in these two case studies. The findings of these case studies indicated that among our participants, the drivers' HR had a higher fraction of abnormal patterns during harsh brakes, accelerating and curved driving. Additionally, free-flow driving with close to zero accelerations on the highway was accompanied by more fraction of normal HR as well as a lower gaze entropy pattern. With the proposed methodology we can better understand variations in driver's psychophysiological states within different driving scenarios. The findings of this work, has the potential to guide future autonomous vehicles to take actions that are fit to each specific driver.

Real-Time Detection of Flu Season Onset: A Novel Approach to Flu Surveillance

The current gold standard for detection of flu season onset in the USA is done retrospectively, where flu season is detected after it has already started. We aimed to create a new surveillance strategy capable of detecting flu season onset prior to its starting. We used an established data generation method that combines Google search volume and historical flu activity data to simulate real-time estimates of flu activity. We then applied a method known as change-point detection to the generated data to determine the point in time that identifies the initial uptick in flu activity which indicates the imminent onset of flu season. Our strategy exhibits a high level of accuracy in predicting the onset of flu season at 86%. Additionally, on average, we detected the onset three weeks prior to the official start of flu season. The results provide evidence to support both the feasibility and efficacy of our strategy to improve the current standard of flu surveillance. The improvement may provide valuable support and lead time for public health officials to take appropriate actions to prevent and control the spread of the flu.

A distribution-free control charting technique based on change-point analysis for detection of epidemics

Worldwide, the detection of epidemics has been recognized as a continuing problem of crucial importance to public health surveillance. Various approaches for detecting and quantifying epidemics of infectious diseases in the recent literature are directly influenced by methods of Statistical Process Control (SPC). However, implementing SPC quality tools directly to the general health care monitoring problem, in a similar manner as in industrial quality control, is not feasible since many assumptions such as stationarity, known asymptotic distribution etc. are not met. Toward this end, in this paper, some of the open statistical research issues involved in this field are discussed, and a distribution-free control charting technique based on change-point analysis is applied and evaluated for detection of epidemics. The main tool in this methodology is the detection of unusual trends, in the sense that the beginning of an unusual trend marks a switch from a control state to an epidemic state. The in-control and out-of-control performance of the adapted control scheme from SPC is thoroughly investigated using Monte Carlo simulations, and the applied scheme is found to outperform its parametric and nonparametric competitors in many cases. Moreover, the empirical comparative study provides evidence that the adapted change-point detection scheme has several appealing properties compared to the current practice for detection of epidemics.

An open repository of real-time COVID-19 indicators

To study the COVID-19 pandemic, its effects on society, and measures for reducing its spread, researchers need detailed data on the course of the pandemic. Standard public health data streams suffer inconsistent reporting and frequent, unexpected revisions. They also miss other aspects of a population’s behavior that are worthy of consideration. We present an open database of COVID signals in the United States, measured at the county level and updated daily. This includes traditionally reported COVID cases and deaths, and many others: measures of mobility, social distancing, internet search trends, self-reported symptoms, and patterns of COVID-related activity in deidentified medical insurance claims. The database provides all signals in a common, easy-to-use format, empowering both public health research and operational decision-making.

The COVID-19 pandemic presented enormous data challenges in the United States. Policy makers, epidemiological modelers, and health researchers all require up-to-date data on the pandemic and relevant public behavior, ideally at fine spatial and temporal resolution. The COVIDcast API is our attempt to fill this need: Operational since April 2020, it provides open access to both traditional public health surveillance signals (cases, deaths, and hospitalizations) and many auxiliary indicators of COVID-19 activity, such as signals extracted from deidentified medical claims data, massive online surveys, cell phone mobility data, and internet search trends. These are available at a fine geographic resolution (mostly at the county level) and are updated daily. The COVIDcast API also tracks all revisions to historical data, allowing modelers to account for the frequent revisions and backfill that are common for many public health data sources. All of the data are available in a common format through the API and accompanying R and Python software packages. This paper describes the data sources and signals, and provides examples demonstrating that the auxiliary signals in the COVIDcast API present information relevant to tracking COVID activity, augmenting traditional public health reporting and empowering research and decision-making.

Suicide attempt and intentional self-harm during the earlier phase of the COVID-19 pandemic in Washtenaw County, Michigan

OBJECTIVE

Determine the early impact of the COVID-19 pandemic on emergency department (ED) encounters for suicide attempt and intentional self-harm at a regional tertiary academic medical centre in Washtenaw County, Michigan, which is one of the wealthier and more diverse counties in the state.

METHODS

Interrupted time series analysis of daily ED encounters from October 2015 through October 2020 for suicide attempt and intentional self-harm (subject n=3002; 62% female; 78% Caucasian) using an autoregressive integrated moving average modelling approach.

RESULTS

There were 39.9% (95% CI 22.9% to 53.1%) fewer ED encounters for suicide attempt and intentional self-harm during the first 7 months of the COVID-19 pandemic (ie, on or after 10 March 2020, when the first cases of COVID-19 were identified in Michigan).

CONCLUSIONS

Fewer individuals sought emergency care for suicide-related behaviour during the earlier phase of the COVID-19 pandemic than expected when compared to prior years. This suggests initial outbreaks of COVID-19 and state of emergency executive orders did not increase suicide-related behaviour in the short term. More work is needed to determine long-term impacts of the COVID-19 pandemic on suicide-related behaviour and whether there are high-risk groups.

Development of a healthcare system COVID Hotspotting Score in California: an observational study with prospective validation

OBJECTIVE

To examine the value of health systems data as indicators of emerging COVID-19 activity.

DESIGN

Observational study of health system indicators for the COVID Hotspotting Score (CHOTS) with prospective validation.

SETTING AND PARTICIPANTS

An integrated healthcare delivery system in Northern California including 21 hospitals and 4.5 million members.

MAIN OUTCOME MEASURES

The CHOTS incorporated 10 variables including four major (cough/cold calls, emails, new positive COVID-19 tests, COVID-19 hospital census) and six minor (COVID-19 calls, respiratory infection and COVID-19 routine and urgent visits, and respiratory viral testing) indicators assessed with change point detection and slope metrics. We quantified cross-correlations lagged by 7-42 days between CHOTS and standardised COVID-19 hospital census using observational data from 1 April to 31 May 2020 and two waves of prospective data through 21 March 2021.

RESULTS

Through 30 September 2020, peak cross-correlation between CHOTS and COVID-19 hospital census occurred with a 28-day lag at 0.78; at 42 days, the correlation was 0.69. Lagged correlation between medical centre CHOTS and their COVID-19 census was highest at 42 days for one facility (0.63), at 35 days for nine facilities (0.52-0.73), at 28 days for eight facilities (0.28-0.74) and at 14 days for two facilities (0.73-0.78). The strongest correlation for individual indicators was 0.94 (COVID-19 census) and 0.90 (new positive COVID-19 tests) lagged 1-14 days and 0.83 for COVID-19 calls and urgent clinic visits lagged 14-28 days. Cross-correlation was similar (0.73) with a 35-day lag using prospective validation from 1 October 2020 to 21 March 2021.

CONCLUSIONS

Passively collected health system indicators were strongly correlated with forthcoming COVID-19 hospital census up to 6 weeks before three successive COVID-19 waves. These tools could inform communities, health systems and public health officials to identify, prepare for and mitigate emerging COVID-19 activity.

Change Point Analysis for Detecting Vaccine Safety Signals

It is important to detect signals of abrupt changes in adverse event reporting in order to notice public safety concerns and take prompt action, especially for vaccines under national immunization programs. In this study, we assessed the applicability of change point analysis (CPA) for signal detection in vaccine safety surveillance. The performances of three CPA methods, namely Bayesian change point analysis, Taylor’s change point analysis (Taylor-CPA), and environmental time series change point detection (EnvCpt), were assessed via simulated data with assumptions for the baseline number of events and degrees of change. The analysis was validated using the Korea Adverse Event Reporting System (KAERS) database. In the simulation study, the Taylor-CPA method exhibited better results for the detection of a change point (accuracy of 96% to 100%, sensitivity of 7% to 100%, specificity of 98% to 100%, positive predictive value of 25% to 85%, negative predictive value of 96% to 100%, and balanced accuracy of 53% to 100%) than the other two CPA methods. When the CPA methods were applied to reports of syncope or dizziness following human papillomavirus (HPV) immunization in the KAERS database, Taylor-CPA and EnvCpt detected a change point (Q2/2013), which was consistent with actual public safety concerns. CPA can be applied as an efficient tool for the early detection of vaccine safety signals.

Two septal-entorhinal GABAergic projections differentially control coding properties of spatially tuned neurons in the medial entorhinal cortex

Septal parvalbumin-expressing (PV+) and calbindin-expressing (CB+) projections inhibit low-threshold and fast-spiking interneurons, respectively, in the medial entorhinal cortex (MEC). We investigate how the two inputs control neuronal activity in the MEC in freely moving mice. Stimulation of PV+ and CB+ terminals causes disinhibition of spatially tuned MEC neurons, but exerts differential effects on temporal coding and burst firing. Thus, recruitment of PV+ projections disrupts theta-rhythmic firing of MEC neurons, while stimulation of CB+ projections increases burst firing of grid cells and enhances phase precession in a cell-type-specific manner. Inactivation of septal PV+ or CB+ neurons differentially affects context, reference, and working memory. Together, our results reveal how specific connectivity of septal GABAergic projections with MEC interneurons translates into differential modulation of MEC neuronal coding.

Understanding COVID-19 transmission through Bayesian probabilistic modeling and GIS-based Voronoi approach: a policy perspective

Originating from Wuhan, China, COVID-19 is spreading rapidly throughout the world. The transmission rate is reported to be high for this novel strain of coronavirus, called severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), as compared to its predecessors. Major strategies in terms of clinical trials of medicines and vaccines, social distancing, use of personal protective equipment (PPE), and so on are being implemented in order to control the spread. The current study concentrates on lockdown and social distancing policy followed by the Indian Government and evaluates its effectiveness using Bayesian probability model (BPM). The change point analysis (CPA) done through the above approach suggests that the states which implemented the lockdown before the exponential rise of cases are able to control the spread of the disease in a much better and efficient way. The analysis has been done for states of Maharashtra, Gujarat, Madhya Pradesh, Rajasthan, Tamil Nadu, West Bengal, Uttar Pradesh, and Delhi as union territory. The highest value of Δ (delta) is reported for Gujarat and Madhya Pradesh with a value of 9.6 weeks, while the lowest value is 4.7, evidently for Maharashtra which is the worst affected. All of the states indicate a significant correlation (p < 0.05, tstat > tcritical) for Δ, i.e., the difference in the time period of CPA and lockdown with cases per population (CPP) and cases per unit area (CPUA), while weak correlation (p < 0.1 and tstat < tcritical) is exhibited by delta and cases per unit population density (CPD). For both CPP and CPUA, tstat > tcritical indicating a significant correlation, while Pearson's correlation indicates the direction to be negative. Further analysis in terms of identification of high-risk areas has been studied from the Voronoi approach of GIS based on the inputs from BPM. All the states follow the above pattern of high population, high case scenario, and the boundaries of risk zones can be identified by Thiessen polygon (TP) constructed therein. The findings of the study help draw strategic and policy-driven response for India, toward tackling COVID-19 pandemic.

Emergency department syndromic surveillance systems: a systematic review

BACKGROUND

Syndromic surveillance provides public health intelligence to aid in early warning and monitoring of public health impacts (e.g. seasonal influenza), or reassurance when an impact has not occurred. Using information collected during routine patient care, syndromic surveillance can be based on signs/symptoms/preliminary diagnoses. This approach makes syndromic surveillance much timelier than surveillance requiring laboratory confirmed diagnoses. The provision of healthcare services and patient access to them varies globally. However, emergency departments (EDs) exist worldwide, providing unscheduled urgent care to people in acute need. This provision of care makes ED syndromic surveillance (EDSyS) a potentially valuable tool for public health surveillance internationally. The objective of this study was to identify and describe the key characteristics of EDSyS systems that have been established and used globally.

METHODS

We systematically reviewed studies published in peer review journals and presented at International Society of Infectious Disease Surveillance conferences (up to and including 2017) to identify EDSyS systems which have been created and used for public health purposes. Search criteria developed to identify "emergency department" and "syndromic surveillance" were applied to NICE healthcare, Global Health and Scopus databases.

RESULTS

In total, 559 studies were identified as eligible for inclusion in the review, comprising 136 journal articles and 423 conference abstracts/papers. From these studies we identified 115 EDSyS systems in 15 different countries/territories across North America, Europe, Asia and Australasia. Systems ranged from local surveillance based on a single ED, to comprehensive national systems. National EDSyS systems were identified in 8 countries/territories: 2 reported inclusion of ≥85% of ED visits nationally (France and Taiwan).

CONCLUSIONS

EDSyS provides a valuable tool for the identification and monitoring of trends in severe illness. Technological advances, particularly in the emergency care patient record, have enabled the evolution of EDSyS over time. EDSyS reporting has become closer to 'real-time', with automated, secure electronic extraction and analysis possible on a daily, or more frequent basis. The dissemination of methods employed and evidence of successful application to public health practice should be encouraged to support learning from best practice, enabling future improvement, harmonisation and collaboration between systems in future.

PROSPERO NUMBER

CRD42017069150 .

Estimating a breakpoint in the pattern of spread of COVID-19 in South Korea

•

We applied the SIR (susceptible–infected–removed) model with a breakpoint to allow for a change in transmission rate.

•

The model accurately estimated the trajectory of COVID-19 spread in South Korea.

•

The breakpoint may reflect potential effects of preventive strategies on transmission rates.

•

Counterfactual experiments illustrated potential impacts of the breakpoint on the spread of infection.

Objectives

Amid the global coronavirus disease 2019 (COVID-19) crisis, South Korea has been lauded for successfully preventing the spread of this infectious disease, which may be due to the aggressive implementation of preventive policies. This study was performed to evaluate the pattern of spread of COVID-19 in South Korea considering the potential impact of policy interventions on transmission rates.

Methods

A SIR (susceptible–infected–removed) model with a breakpoint that allows a change in transmission rate at an unknown point was established. Estimated trajectories of COVID-19 from SIR models with and without a breakpoint were compared.

Results

The proposed model with a break fitted the actual series of infection cases much better than the classic model. The estimated breakpoint was March 7, 2020 and the transmission rate dropped by 0.23 after the breakpoint. A counterfactual study based on our estimate indicated that the number of infected could have reached 2 500 000 compared to the peak of 8000 in the observed series.

Conclusions

It is critical to consider a change in the transmission rate to evaluate the trajectory of spread of COVID-19 in South Korea. Our estimation and counterfactual experiments indicate that public health interventions may play a role in determining the pattern of spread of infectious diseases.

A first estimation of the impact of public health actions against COVID-19 in Veneto (Italy)

Background

Veneto is one of the first Italian regions where the COVID-19 outbreak started spreading. Containment measures were approved soon thereafter. The present study aims at providing a first look at the impact of the containment measures on the outbreak progression in the Veneto region, Italy.

Methods

A Bayesian changepoint analysis was used to identify the changing speed of the epidemic curve. Then, a piecewise polynomial model was considered to fit the data in the first period before the detected changepoint. In this time interval, that is, the weeks from 27 February to 12 March, a quadratic growth was identified by a generalised additive model (GAM). Finally, the model was used to generate the projection of the expected number of hospitalisations at 2 weeks based on the epidemic speed before the changepoint. Such estimates were then compared with the actual outbreak behaviour.

Results

The comparison between the observed and predicted hospitalisation curves highlights a slowdown on the total COVID-19 hospitalisations after the onset of containment measures. The estimated daily slowdown effect of the epidemic growth is estimated as 78 hospitalisations per day as of 27 March (95% CI 75 to 81).

Conclusions

The containment strategies seem to have positively impacted the progression of the COVID-19 epidemic outbreak in Veneto.

A systematic review of aberration detection algorithms used in public health surveillance

The algorithms used for detecting anomalies have evolved substantially over the last decade to take advantage of advances in informatics and to accommodate changes in surveillance data. We identified 145 studies since 2007 that evaluated statistical methods used to detect aberrations in public health surveillance data. For each study, we classified the analytic methods and reviewed the evaluation metrics. We also summarized the practical usage of the detection algorithms in public health surveillance systems worldwide. Traditional methods (e.g., control charts, linear regressions) were the focus of most evaluation studies and continue to be used commonly in practice. There was, however, an increase in the number of studies using forecasting methods and studies applying machine learning methods, hidden Markov models, and Bayesian framework to multivariate datasets. Evaluation studies demonstrated improved accuracy with more sophisticated methods, but these methods do not appear to be used widely in public health practice.

Development and Performance of a Checklist for Initial Triage After an Anthrax Mass Exposure Event

BACKGROUND

Population exposure to Bacillus anthracis spores could cause mass casualties requiring complex medical care. Rapid identification of patients needing anthrax-specific therapies will improve patient outcomes and resource use.

OBJECTIVE

To develop a checklist that rapidly distinguishes most anthrax from nonanthrax illnesses on the basis of clinical presentation and identifies patients requiring diagnostic testing after a population exposure.

DESIGN

Comparison of published anthrax case reports from 1880 through 2013 that included patients seeking anthrax-related care at 2 epicenters of the 2001 U.S. anthrax attacks.

SETTING

Outpatient and inpatient.

PATIENTS

408 case patients with inhalation, ingestion, and cutaneous anthrax and primary anthrax meningitis, and 657 control patients.

MEASUREMENTS

Diagnostic test characteristics, including positive and negative likelihood ratios (LRs) and patient triage assignation.

RESULTS

Checklist-directed triage without diagnostic testing correctly classified 95% (95% CI, 93% to 97%) of 353 adult anthrax case patients and 76% (CI, 73% to 79%) of 647 control patients (positive LR, 3.96 [CI, 3.45 to 4.55]; negative LR, 0.07 [CI, 0.04 to 0.11]; false-negative rate, 5%; false-positive rate, 24%). Diagnostic testing was needed for triage in up to 5% of case patients and 15% of control patients and improved overall test characteristics (positive LR, 8.90 [CI, 7.05 to 11.24]; negative LR, 0.06 [CI, 0.04 to 0.09]; false-negative rate, 5%; false-positive rate, 11%). Checklist sensitivity and specificity were minimally affected by inclusion of pediatric patients. Sensitivity increased to 97% (CI, 94% to 100%) and 98% (CI, 96% to 100%), respectively, when only inhalation anthrax cases or higher-quality case reports were investigated.

LIMITATIONS

Data on case patients were limited to nonstandardized, published observational reports, many of which lacked complete data on symptoms and signs of interest. Reporting bias favoring more severe cases and lack of intercurrent outbreaks (such as influenza) in the control populations may have improved test characteristics.

CONCLUSION

A brief checklist covering symptoms and signs can distinguish anthrax from other conditions with minimal need for diagnostic testing after known or suspected population exposure.

PRIMARY FUNDING SOURCE

U.S. Department of Health and Human Services.

How to manipulate friends and influence practice: Application of complexity science leads to quality improvement in laboratory sample submissions

Background

We sought to reduce healthcare-associated infections (HCAIs) through the application of complexity science.

Objective

To confirm incidental findings that altering the structure of microbiology reports with targeted education led to better utilisation of laboratory resources, while participating in efforts to reduce HCAI.

Methods

We adopted a different approach to laboratory result authorisation, using narrative to engage the clinicians and induce behavioural change. Subsequent educational opportunities emphasised key messages.

Findings/Results

Positive urine means calculated by the analysis tool numbered 2179/month throughout the study period. Negative urines started at 5576/month, reduced to 5134/month in November 2014 and to 4602/month in April 2016, coinciding with our changes. Opportunity costs were saved.

Discussion

The changes in both policy and reporting were contemporaneous with a decline in negative samples. There were no significant changes in the number of positive specimens. The efficiency and effectiveness of the laboratory was improved and resources released: £145,000 ($182,000) for a resident population of 384,000. This suggests an annual release of about £25 million ($31 million) may be possible in the UK and £122 million ($155 million) in the USA.

Period of Measurement in Time-Series Predictions of Disease Counts from 2007 to 2017 in Northern Nevada: Analytics Experiment

BACKGROUND

The literature in statistics presents methods by which autocorrelation can identify the best period of measurement to improve the performance of a time-series prediction. The period of measurement plays an important role in improving the performance of disease-count predictions. However, from the operational perspective in public health surveillance, there is a limitation to the length of the measurement period that can offer meaningful and valuable predictions.

OBJECTIVE

This study aimed to establish a method that identifies the shortest period of measurement without significantly decreasing the prediction performance for time-series analysis of disease counts.

METHODS

The data used in this evaluation include disease counts from 2007 to 2017 in northern Nevada. The disease counts for chlamydia, salmonella, respiratory syncytial virus, gonorrhea, viral meningitis, and influenza A were predicted.

RESULTS

Our results showed that autocorrelation could not guarantee the best performance for prediction of disease counts. However, the proposed method with the change-point analysis suggests a period of measurement that is operationally acceptable and performance that is not significantly different from the best prediction.

CONCLUSIONS

The use of change-point analysis with autocorrelation provides the best and most practical period of measurement.

When to re-order laboratory tests? Learning laboratory test shelf-life

Most laboratory results are valid for only a certain time period (laboratory tests shelf-life), after which they are outdated and the test needs to be re-administered. Currently, laboratory test shelf-lives are not centrally available anywhere but the implicit knowledge of doctors. In this work we propose an automated method to learn laboratory test-specific shelf-life by identifying prevalent laboratory test order patterns in electronic health records. The resulting shelf-lives performed well in the evaluation of internal validity, clinical interpretability, and external validity.

Benefits of combining change-point analysis with disproportionality analysis in pharmacovigilance signal detection

BACKGROUND

Change-point analysis (CPA) is a powerful method to analyse pharmacovigilance data but it has never been used on the disproportionality metric.

OBJECTIVES

To optimize signal detection investigating the interest of time-series analysis in pharmacovigilance and the benefits of combining CPA with the proportional reporting ratio (PRR).

METHODS

We investigated the couple benfluorex and aortic valve incompetence (AVI) using the French National Pharmacovigilance and EudraVigilance databases: CPA was applied on monthly counts of reports and the lower bound of monthly computed PRR (PRR-). We stated a CPA hypothesis that the substance-event combination is more likely to be a signal when the 2 following criteria are fulfilled: PRR- is greater than 1 with at least 5 cases, and CPA method detects at least 2 successive change points of PRR- which made consecutively increasing segments. We tested this hypothesis by 95 test cases identified from a drug safety reference set and 2 validated signals from EudraVigilance database: CPA was applied on PRR-.

RESULTS

For benfluorex and AVI, change points detected by CPA on PRR- were more meaningful compared with monthly counts of reports: More change points detected and detected earlier. In the reference set, 14 positive controls satisfied CPA hypothesis, 6 positive controls only met first requirements, 3 negative controls only met first requirement, and 2 validated signals satisfied CPA hypothesis.

CONCLUSIONS

The combination of CPA and PRR represents a significant advantage in detecting earlier signals and reducing false-positive signals. This approach should be confirmed in further studies.

A case for revisiting peer review: Implications for professional self-regulation and quality improvement

Background

Quality improvement in healthcare has often been promoted as different from and more valuable than peer review and other professional self-regulation processes. In spite of attempts to harmonize these two approaches, the perception of dichotomous opposition has persisted. A sequence of events in the troubled California prison system fortuitously isolated workforce interventions from more typical quality improvement interventions. Our objectives were to (1) evaluate the relative contributions of professional accountability and quality improvement interventions to an observed decrease in population mortality and (2) explore the organizational dynamics that potentiated positive outcomes.

Methods

Our retrospective mixed-methods case study correlated time-series analysis of mortality with the timing of reform interventions. Quantitative and qualitative evidence was drawn from court documents, public use files, internal databases, and other archival documents.

Results

Change point analysis reveals with 98% confidence that a significant improvement in age-adjusted natural mortality occurred in 2007, decreasing from 138.7 per 100,000 in the 1998–2006 period to 106.4 in the 2007–2009 period. The improvement in mortality occurred after implementation of accountability processes, prior to implementation of quality improvement interventions. Archival evidence supports the positive impact of physician competency assessments, robust peer review, and replacement of problem physicians.

Conclusions

Our analysis suggests that workforce accountability provides a critical quality safeguard, and its neglect in scholarship and practice is unjustified. As with quality improvement, effective professional self-regulation requires systemic implementation of enabling policies, processes, and staff resources. The study adds to evidence that the distribution of physician performance contains a heterogeneous left skew of dyscompetence that is associated with significant harm and suggests that professional self-regulation processes such as peer review can reduce that harm. Beyond their responsibility for direct harm, dyscompetent professionals can have negative impacts on group performance. The optimal integration of professional accountability and quality improvement systems merits further investigation.

Amygdala Adaptation and Temporal Dynamics of the Salience Network in Conditioned Fear: A Single-Trial fMRI Study

Research in rodents has established the role of the amygdaloid complex in defensive responses to conditioned threat. In human imaging studies, however, activation of the amygdala by conditioned threat cues is often not observed. One hypothesis states that this finding reflects adaptation of amygdaloid responses over time. We tested this hypothesis by estimating single-trial neural responses over a large number of conditioning trials. Functional MRI (fMRI) was recorded from 18 participants during classical differential fear conditioning: Participants viewed oriented grayscale grating stimuli (45° or 135°) presented centrally in random order. In the acquisition block, one grating (the CS+) was paired with a noxious noise, the unconditioned stimulus (US), on 25% of trials. The other grating, denoted CS–, was never paired with the US. Consistent with previous reports, BOLD in dorsal anterior cingulate cortex (dACC) and insula, but not the amygdala, was heightened when viewing CS+ stimuli that were not paired with US compared to CS– stimuli. Trial-by-trial analysis showed that over the course of acquisition, activity in the amygdala attenuated. Interestingly, activity in the dACC and insula also declined. Representational similarity analysis (RSA) corroborated these results, indicating that the voxel patterns evoked by CS+ and CS– in these brain regions became less distinguishable over time. Together, the present findings support the hypothesis that the lack of BOLD differences in the amygdaloid complex in many studies of classical conditioning is due to adaptation, and the adaptation effects may reflect changes in large-scale networks mediating aversive conditioning, particularly the salience network.

Evaluating Google, Twitter, and Wikipedia as Tools for Influenza Surveillance Using Bayesian Change Point Analysis: A Comparative Analysis

Background

Traditional influenza surveillance relies on influenza-like illness (ILI) syndrome that is reported by health care providers. It primarily captures individuals who seek medical care and misses those who do not. Recently, Web-based data sources have been studied for application to public health surveillance, as there is a growing number of people who search, post, and tweet about their illnesses before seeking medical care. Existing research has shown some promise of using data from Google, Twitter, and Wikipedia to complement traditional surveillance for ILI. However, past studies have evaluated these Web-based sources individually or dually without comparing all 3 of them, and it would be beneficial to know which of the Web-based sources performs best in order to be considered to complement traditional methods.

Objective

The objective of this study is to comparatively analyze Google, Twitter, and Wikipedia by examining which best corresponds with Centers for Disease Control and Prevention (CDC) ILI data. It was hypothesized that Wikipedia will best correspond with CDC ILI data as previous research found it to be least influenced by high media coverage in comparison with Google and Twitter.

Methods

Publicly available, deidentified data were collected from the CDC, Google Flu Trends, HealthTweets, and Wikipedia for the 2012-2015 influenza seasons. Bayesian change point analysis was used to detect seasonal changes, or change points, in each of the data sources. Change points in Google, Twitter, and Wikipedia that occurred during the exact week, 1 preceding week, or 1 week after the CDC’s change points were compared with the CDC data as the gold standard. All analyses were conducted using the R package “bcp” version 4.0.0 in RStudio version 0.99.484 (RStudio Inc). In addition, sensitivity and positive predictive values (PPV) were calculated for Google, Twitter, and Wikipedia.

Results

During the 2012-2015 influenza seasons, a high sensitivity of 92% was found for Google, whereas the PPV for Google was 85%. A low sensitivity of 50% was calculated for Twitter; a low PPV of 43% was found for Twitter also. Wikipedia had the lowest sensitivity of 33% and lowest PPV of 40%.

Conclusions

Of the 3 Web-based sources, Google had the best combination of sensitivity and PPV in detecting Bayesian change points in influenza-related data streams. Findings demonstrated that change points in Google, Twitter, and Wikipedia data occasionally aligned well with change points captured in CDC ILI data, yet these sources did not detect all changes in CDC data and should be further studied and developed.

Prognostic implications of 2005 Gleason grade modification. Population-based study of biochemical recurrence following radical prostatectomy

OBJECTIVE

To assess the impact of the 2005 modification of the Gleason classification on risk of biochemical recurrence (BCR) after radical prostatectomy (RP).

PATIENTS AND METHODS

In the Prostate Cancer data Base Sweden (PCBaSe), 2,574 men assessed with the original Gleason classification and 1,890 men assessed with the modified Gleason classification, diagnosed between 2003 and 2007, underwent primary RP. Histopathology was reported according to the Gleason Grading Groups (GGG): GGG1 = Gleason score (GS) 6, GGG2 = GS 7(3 + 4), GGG3 = GS 7(4 + 3), GGG4 = GS 8 and GGG5 = GS 9-10. Cumulative incidence and multivariable Cox proportional hazards regression models were used to assess difference in BCR.

RESULTS

The cumulative incidence of BCR was lower using the modified compared to the original classification: GGG2 (16% vs. 23%), GGG3 (21% vs. 35%) and GGG4 (18% vs. 34%), respectively. Risk of BCR was lower for modified versus original classification, GGG2 Hazard ratio (HR) 0.66, (95%CI 0.49-0.88), GGG3 HR 0.57 (95%CI 0.38-0.88) and GGG4 HR 0.53 (95%CI 0.29-0.94).

CONCLUSION

Due to grade migration following the 2005 Gleason modification, outcome after RP are more favourable. Consequently, outcomes from historical studies cannot directly be applied to a contemporary setting. J. Surg. Oncol. 2016;114:664-670. © 2016 Wiley Periodicals, Inc.

Ring-Shaped Microlanes and Chemical Barriers as a Platform for Probing Single-Cell Migration

Quantification and discrimination of pharmaceutical and disease-related effects on cell migration requires detailed characterization of single-cell motility. In this context, micropatterned substrates that constrain cells within defined geometries facilitate quantitative readout of locomotion. Here, we study quasi-one-dimensional cell migration in ring-shaped microlanes. We observe bimodal behavior in form of alternating states of directional migration (run state) and reorientation (rest state). Both states show exponential lifetime distributions with characteristic persistence times, which, together with the cell velocity in the run state, provide a set of parameters that succinctly describe cell motion. By introducing PEGylated barriers of different widths into the lane, we extend this description by quantifying the effects of abrupt changes in substrate chemistry on migrating cells. The transit probability decreases exponentially as a function of barrier width, thus specifying a characteristic penetration depth of the leading lamellipodia. Applying this fingerprint-like characterization of cell motion, we compare different cell lines, and demonstrate that the cancer drug candidate salinomycin affects transit probability and resting time, but not run time or run velocity. Hence, the presented assay allows to assess multiple migration-related parameters, permits detailed characterization of cell motility, and has potential applications in cell biology and advanced drug screening.

Outbreak definition by change point analysis: a tool for public health decision?

Background

Most studies of epidemic detection focus on their start and rarely on the whole signal or the end of the epidemic. In some cases, it may be necessary to retrospectively identify outbreak signals from surveillance data. Our study aims at evaluating the ability of change point analysis (CPA) methods to locate the whole disease outbreak signal. We will compare our approach with the results coming from experts’ signal inspections, considered as the gold standard method.

Methods

We simulated 840 time series, each of which includes an epidemic-free baseline (7 options) and a type of epidemic (4 options). We tested the ability of 4 CPA methods (Max-likelihood, Kruskall-Wallis, Kernel, Bayesian) methods and expert inspection to identify the simulated outbreaks. We evaluated the performances using metrics including delay, accuracy, bias, sensitivity, specificity and Bayesian probability of correct classification (PCC).

Results

A minimum of 15 h was required for experts for analyzing the 840 curves and a maximum of 25 min for a CPA algorithm. The Kernel algorithm was the most effective overall in terms of accuracy, bias and global decision (PCC = 0.904), compared to PCC of 0.848 for human expert review.

Conclusions

For the aim of retrospectively identifying the start and end of a disease outbreak, in the absence of human resources available to do this work, we recommend using the Kernel change point model. And in case of experts’ availability, we also suggest to supplement the Human expertise with a CPA, especially when the signal noise difference is below 0.

Electronic supplementary material

The online version of this article (doi:10.1186/s12911-016-0271-x) contains supplementary material, which is available to authorized users.

Dynamic taste responses of parabrachial pontine neurons in awake rats

The parabrachial nuclei of the pons (PbN) receive almost direct input from taste buds on the tongue and control basic taste-driven behaviors. Thus it is reasonable to hypothesize that PbN neurons might respond to tastes in a manner similar to that of peripheral receptors, i.e., that these responses might be narrow and relatively "dynamics free." On the other hand, the majority of the input to PbN descends from forebrain regions such as gustatory cortex (GC), which processes tastes with "temporal codes" in which firing reflects first the presence, then the identity, and finally the desirability of the stimulus. Therefore a reasonable alternative hypothesis is that PbN responses might be dominated by dynamics similar to those observed in GC. Here we examined simultaneously recorded single-neuron PbN (and GC) responses in awake rats receiving exposure to basic taste stimuli. We found that pontine taste responses were almost entirely confined to canonically identified taste-PbN (t-PbN). Taste-specificity was found, furthermore, to be time varying in a larger percentage of these t-PbN responses than in responses recorded from the tissue around PbN (including non-taste-PbN). Finally, these time-varying properties were a good match for those observed in simultaneously recorded GC neurons-taste-specificity appeared after an initial nonspecific burst of action potentials, and palatability emerged several hundred milliseconds later. These results suggest that the pontine taste relay is closely allied with the dynamic taste processing performed in forebrain.

Use of syndromic surveillance at local health departments: movement toward more effective systems

CONTEXT

Syndromic surveillance systems enhance public health practice in both large and small population settings. However, data from these systems are typically monitored by state and federal agencies and less frequently used by small public health agencies, such as local health departments (LHDs). Syndromic surveillance system modifications may facilitate use by LHDs.

OBJECTIVE

To describe syndromic surveillance system modifications and survey LHD staff to assess subsequent changes in system use.

DESIGN

Pre- and postintervention cross-sectional analysis.

SETTING

North Carolina (NC) LHDs, 2009 and 2012.

PARTICIPANTS

LHD nursing and preparedness staff.

MAIN OUTCOME MEASURES

Use of syndromic surveillance data by LHDs for outbreak response, seasonal event response, program management, and stakeholder reports.

RESULTS

In NC, syndromic surveillance system modifications made between 2009 and 2012 included implementation of LHD-specific data "dashboards" and increased distribution of LHD-specific surveillance information by the state public health agency. Users of LHD syndromic surveillance system increased from 99 in 2009 to 175 in 2012. Twenty-seven of 28 (96%) and 62 of 72 (86%) respondents completed the 2009 and 2012 surveys, respectively. Among respondents, 23% used syndromic surveillance data for outbreak response in 2009, compared with 25% in 2012. In 2009, 46% of respondents used these data for seasonal event response, compared with 57% in 2012. Syndromic surveillance data were used for program management by 25% of respondents in 2009 (compared with 30% in 2012) and for stakeholder reports by 23% of respondents in 2009 (compared with 33% in 2012).

CONCLUSIONS

Syndromic surveillance system changes supported modest increases in LHD use of syndromic surveillance information. Because use of syndromic surveillance information at smaller LHD is rare, these modest increases indicate effective modification of the NC syndromic surveillance system.

Use of data mining at the Food and Drug Administration

Objectives This article summarizes past and current data mining activities at the United States Food and Drug Administration (FDA).

Target audience We address data miners in all sectors, anyone interested in the safety of products regulated by the FDA (predominantly medical products, food, veterinary products and nutrition, and tobacco products), and those interested in FDA activities.

Scope Topics include routine and developmental data mining activities, short descriptions of mined FDA data, advantages and challenges of data mining at the FDA, and future directions of data mining at the FDA.

Signal detection using change point analysis in postmarket surveillance

Purpose

Signal detection methods have been used extensively in postmarket surveillance to identify elevated risks of adverse events associated with medical products (drugs, vaccines, and devices). However, current popular disproportionality methods ignore useful information such as trends when the data are aggregated over time for signal detection.

Methods

In this paper, we applied change point analysis (CPA) to trend analysis of medical products in a spontaneous adverse event reporting system. CPA was used to detect the time point at which statistical properties of a sequence of observations change over time. Two CPA approaches, change in mean and change in variance, were demonstrated by an example using neurostimulator adverse event dataset.

Results

Two significant change points associated with upward trends were detected in June 2008 (n = 20, p < 0.001) and May 2011 (n = 51, p = 0.003). Further investigation confirmed battery issues and expansion of the indication for use could be possible causes for the occurrence of these change points. Two time points showed extremely low number of loss of therapy events, two cases in October 2009 and three in November 2009, which could be the result of reporting issues such as underreporting.

Conclusion

As a complimentary tool to current signal detection efforts at FDA, CPA can be used to detect changes in the association between medical products and adverse events over time. Detecting these changes could be critical for public health regulation, adverse events surveillance, product recalls, and regulators’ understanding of the connection between adverse events and other events regarding regulated products. © 2015 The Authors. Pharmacoepidemiology and Drug Safety published by John Wiley & Sons, Ltd.

Visualizing the quality of partially accruing data for use in decision making

Secondary use of clinical health data for near real-time public health surveillance presents challenges surrounding its utility due to data quality issues. Data used for real-time surveillance must be timely, accurate and complete if it is to be useful; if incomplete data are used for surveillance, understanding the structure of the incompleteness is necessary. Such data are commonly aggregated due to privacy concerns. The Distribute project was a near real-time influenza-like-illness (ILI) surveillance system that relied on aggregated secondary clinical health data. The goal of this work is to disseminate the data quality tools developed to gain insight into the data quality problems associated with these data. These tools apply in general to any system where aggregate data are accrued over time and were created through the end-user-as-developer paradigm. Each tool was developed during the exploratory analysis to gain insight into structural aspects of data quality. Our key finding is that data quality of partially accruing data must be studied in the context of accrual lag-the difference between the time an event occurs and the time data for that event are received, i.e. the time at which data become available to the surveillance system. Our visualization methods therefore revolve around visualizing dimensions of data quality affected by accrual lag, in particular the tradeoff between timeliness and completion, and the effects of accrual lag on accuracy. Accounting for accrual lag in partially accruing data is necessary to avoid misleading or biased conclusions about trends in indicator values and data quality.

Temporal changes in survival after cardiac surgery are associated with the thirty-day mortality benchmark

OBJECTIVE

To assess the hypothesis that postoperative survival exhibits heterogeneity associated with the timing of quality metrics.

DATA SOURCES

Retrospective observational study using the Nationwide Inpatient Sample from 2005 through 2009.

STUDY DESIGN

Survival analysis was performed on all admission records with a procedure code for major cardiac surgery (n = 595,089). The day-by-day hazard function for all-cause in-hospital mortality at 1-day intervals was analyzed using joinpoint regression (a data-driven method of testing for changes in hazard).

DATA EXTRACTION METHODS

A comprehensive analysis of a publicly available national administrative database was performed.

PRINCIPAL FINDINGS

Statistically significant shifts in the pattern of postoperative mortality occurred at day 6 (95 percent CI = day 5-8) and day 30 (95 percent CI = day 20-35).

CONCLUSIONS

While the shift at day 6 plausibly can be attributed to the separation between routine recovery and a complicated postoperative course, the abrupt increase in mortality at day 30 has no clear organic etiology. This analysis raises the possibility that this observed shift may be related to clinician behavior because of the use of 30-day mortality as a quality metric, but further studies will be required to establish causality.

Variation in information needs and quality: implications for public health surveillance and biomedical informatics

Understanding variation among users' information needs and the quality of information in an electronic system is important for informaticians to ensure data are fit-for-use in answering important questions in clinical and public health. To measure variation in satisfaction with currently reported data, as well as perceived importance and need with respect to completeness and timeliness, we surveyed epidemiologists and other public health professionals across multiple jurisdictions. We observed consensus for some data elements, such as county of residence, which respondents perceived as important and felt should always be reported. However information needs differed for many data elements, especially when comparing notifiable diseases such as chlamydia to seasonal (influenza) and chronic (diabetes) diseases. Given the trend towards greater volume and variety of data as inputs to surveillance systems, variation of information needs impacts system design and practice. Systems must be flexible and highly configurable to accommodate variation, and informaticians must measure and improve systems and business processes to accommodate for variation of both users and information.

Syndromic surveillance for influenza in the emergency department-A systematic review

The science of surveillance is rapidly evolving due to changes in public health information and preparedness as national security issues, new information technologies and health reform. As the Emergency Department has become a much more utilized venue for acute care, it has also become a more attractive data source for disease surveillance. In recent years, influenza surveillance from the Emergency Department has increased in scope and breadth and has resulted in innovative and increasingly accepted methods of surveillance for influenza and influenza-like-illness (ILI). We undertook a systematic review of published Emergency Department-based influenza and ILI syndromic surveillance systems. A PubMed search using the keywords "syndromic", "surveillance", "influenza" and "emergency" was performed. Manuscripts were included in the analysis if they described (1) data from an Emergency Department (2) surveillance of influenza or ILI and (3) syndromic or clinical data. Meeting abstracts were excluded. The references of included manuscripts were examined for additional studies. A total of 38 manuscripts met the inclusion criteria, describing 24 discrete syndromic surveillance systems. Emergency Department-based influenza syndromic surveillance has been described worldwide. A wide variety of clinical data was used for surveillance, including chief complaint/presentation, preliminary or discharge diagnosis, free text analysis of the entire medical record, Google flu trends, calls to teletriage and help lines, ambulance dispatch calls, case reports of H1N1 in the media, markers of ED crowding, admission and Left Without Being Seen rates. Syndromes used to capture influenza rates were nearly always related to ILI (i.e. fever +/- a respiratory or constitutional complaint), however, other syndromes used for surveillance included fever alone, "respiratory complaint" and seizure. Two very large surveillance networks, the North American DiSTRIBuTE network and the European Triple S system have collected large-scale Emergency Department-based influenza and ILI syndromic surveillance data. Syndromic surveillance for influenza and ILI from the Emergency Department is becoming more prevalent as a measure of yearly influenza outbreaks.

Structural models used in real-time biosurveillance outbreak detection and outbreak curve isolation from noisy background morbidity levels

OBJECTIVE

We discuss the use of structural models for the analysis of biosurveillance related data.

METHODS AND RESULTS

Using a combination of real and simulated data, we have constructed a data set that represents a plausible time series resulting from surveillance of a large scale bioterrorist anthrax attack in Miami. We discuss the performance of anomaly detection with structural models for these data using receiver operating characteristic (ROC) and activity monitoring operating characteristic (AMOC) analysis. In addition, we show that these techniques provide a method for predicting the level of the outbreak valid for approximately 2 weeks, post-alarm.

CONCLUSIONS

Structural models provide an effective tool for the analysis of biosurveillance data, in particular for time series with noisy, non-stationary background and missing data.