@article{Li_Pererita_Johnson_Johnson_2017, title={Integrated spatiotemporal surveillance system: Data, Analysis and Visualization}, volume={9}, url={https://ojphi.org/ojs/index.php/ojphi/article/view/7641}, DOI={10.5210/ojphi.v9i1.7641}, abstractNote={<div style="left: 82px; top: 292.246px; font-size: 12.9074px; font-family: sans-serif; transform: scaleX(1.08211);" data-canvas-width="58.10914814814815">Objective</div><div style="left: 95.6667px; top: 305.868px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.988685);" data-canvas-width="346.81429259259255">To build an open source spatiotemporal system that integrates</div><div style="left: 82px; top: 321.054px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00146);" data-canvas-width="257.02520370370365">analysis and visualization for disease surveillance</div><div style="left: 82px; top: 349.95px; font-size: 12.9074px; font-family: sans-serif; transform: scaleX(1.11815);" data-canvas-width="75.28890740740741">Introduction</div><div style="left: 95.6667px; top: 363.572px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.992862);" data-canvas-width="347.00532222222205">Most surveillance methods in the literature focus on temporal</div><div style="left: 82px; top: 378.757px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.976726);" data-canvas-width="360.1515166666666">aberration detections with data aggregated to certain geographical</div><div style="left: 82px; top: 393.942px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.03684);" data-canvas-width="361.94435555555543">boundaries. SaTScan has been widely used for spatiotemporal</div><div style="left: 82px; top: 409.128px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.02275);" data-canvas-width="361.40740740740756">aberration detection due to its user friendly software interface.</div><div style="left: 82px; top: 424.313px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01031);" data-canvas-width="358.7949481481481">However, the software is limited to spatial scan statistics and suffers</div><div style="left: 82px; top: 439.498px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.0474);" data-canvas-width="361.9624259259259">from location imprecision and heterogeneity of population. R</div><div style="left: 82px; top: 454.683px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.04102);" data-canvas-width="359.65458148148133">Surveillance has a collection of spatiotemporal methods that focus</div><div style="left: 82px; top: 469.868px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00151);" data-canvas-width="208.2481111111112">more on research instead of surveillance</div><div style="left: 82px; top: 498.765px; font-size: 12.9074px; font-family: sans-serif; transform: scaleX(1.07341);" data-canvas-width="53.062351851851844">Methods</div><div style="left: 95.6667px; top: 512.387px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.03433);" data-canvas-width="345.8643074074073">Based in Ontario, Canada, we used postal codes for determining</div><div style="left: 82px; top: 527.572px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01877);" data-canvas-width="359.2002407407407">the location of cases of reportable infectious diseases. The variation</div><div style="left: 82px; top: 542.757px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.04284);" data-canvas-width="359.520344444444">in geographic sizes and shapes of the case and census geographies</div><div style="left: 82px; top: 557.943px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.05881);" data-canvas-width="362.41160370370375">created challenges for developing a uniform temporal spatial</div><div style="left: 82px; top: 573.128px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00141);" data-canvas-width="160.2583703703704">surveillance system, including:</div><div style="left: 95.6667px; top: 588.313px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00353);" data-canvas-width="333.03692592592563">Linking case and population data due to misclassification errors,</div><div style="left: 95.6667px; top: 603.498px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.04841);" data-canvas-width="348.28315555555554">Distance based correlations due to irregularly shaped areas</div><div style="left: 82px; top: 618.683px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.02621);" data-canvas-width="88.54481481481481">(e.g. FSA’s), and</div><div style="left: 95.6667px; top: 633.868px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00082);" data-canvas-width="344.7155481481481">Visualization bias due to variation in population density, e.g. large</div><div style="left: 82px; top: 649.054px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.0015);" data-canvas-width="136.23768518518523">area with little population.</div><div style="left: 95.6667px; top: 664.239px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.02581);" data-canvas-width="345.6990925925925">To overcome these challenges, we developed the Ontario Hybrid</div><div style="left: 82px; top: 679.424px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00781);" data-canvas-width="361.0976296296296">Information Map (OHIM) boundary, which is a combination of</div><div style="left: 82px; top: 694.609px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00196);" data-canvas-width="360.5555185185185">Public Health Unit boundaries (rural areas), census subdivisions</div><div style="left: 82px; top: 709.794px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.964521);" data-canvas-width="360.16829629629615">(rural urban mixed) and regular grid cells (urban). The goal is to</div><div style="left: 82px; top: 724.979px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01656);" data-canvas-width="359.1679722222222">capture population details in urban areas without losing information</div><div style="left: 82px; top: 740.165px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.03678);" data-canvas-width="359.22605555555555">in rural areas. OHIM has around 4600 geographies with more than</div><div style="left: 82px; top: 755.35px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.02455);" data-canvas-width="359.1202148148148">half located in urban centers. Population distribution by gender and</div><div style="left: 82px; top: 770.535px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.03607);" data-canvas-width="359.5332518518519">age group was calculated for each OHIM geography. A lookup file</div><div style="left: 82px; top: 785.72px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.994953);" data-canvas-width="355.3473796296296">was also created to link all Ontario postal codes to OHIM geography.</div><div style="left: 95.6667px; top: 800.905px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.048);" data-canvas-width="345.6293925925926">To create baselines, historical data for influenza A were used to</div><div style="left: 82px; top: 816.091px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.974519);" data-canvas-width="360.0521296296295">model the seasonality and calculate expected case count for each</div><div style="left: 82px; top: 831.276px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01817);" data-canvas-width="359.21056666666675">OHIM geography for each week. Standardized incident ratios (SIR)</div><div style="left: 82px; top: 846.461px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.993847);" data-canvas-width="355.34996111111104">were calculated as exploratory statistics, and a spatiotemporal Besag-</div><div style="left: 82px; top: 861.646px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00306);" data-canvas-width="358.50969444444445">York-Mollie (BYM) model was used to calculate the probability that</div><div style="left: 82px; top: 876.831px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.958963);" data-canvas-width="359.467424074074">the risk is higher than a pre-specified threshold. Integrated Nested</div><div style="left: 82px; top: 892.017px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.998069);" data-canvas-width="358.43999444444444">Laplace Approximation (R-INLA) was used in R to explore different</div><div style="left: 82px; top: 907.202px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01313);" data-canvas-width="359.035025925926">types of spatiotemporal interactions and for fast Bayesian inference.</div><div style="left: 82px; top: 922.387px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.02253);" data-canvas-width="359.19765925925924">The ability to apply the models was verified by examining previous</div><div style="left: 82px; top: 937.572px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.977166);" data-canvas-width="360.21605370370344">outbreaks and seeking the opinion of staff that routinely perform</div><div style="left: 82px; top: 952.757px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00695);" data-canvas-width="132.98501851851853">surveillance on influenza.</div><div style="left: 95.6667px; top: 967.943px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.998411);" data-canvas-width="344.70135000000005">To ensure the visualization integrates with the analysis, R package</div><div style="left: 82px; top: 983.128px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.03195);" data-canvas-width="359.49711111111134">Shiny was used to build an interactive spatiotemporal visualization</div><div style="left: 82px; top: 998.313px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01114);" data-canvas-width="361.0627796296295">on OHIM boundary utilizing Open Street Map and html5. The</div><div style="left: 82px; top: 1013.5px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01106);" data-canvas-width="358.85819444444445">application not only allows users to pan and zoom in space and time</div><div style="left: 82px; top: 1028.68px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.989772);" data-canvas-width="358.26445370370374">to explore the results and locate high risk areas, it also gives users the</div><div style="left: 82px; top: 1043.87px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.981175);" data-canvas-width="358.17797407407386">flexibility to change algorithm parameters for instant feedback. Figure</div><div style="left: 82px; top: 1059.05px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.968222);" data-canvas-width="359.8882055555554">1 demonstrates a zoomed-in OHIM boundary with pointers signal</div><div style="left: 82px; top: 1074.24px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.954985);" data-canvas-width="359.7268629629628">for “high risk” area at user specified statistics exceeds a threshold</div><div style="left: 82px; top: 1089.42px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.04525);" data-canvas-width="362.1431296296293">(e.g., SIR > 2). Using the algorithms and visualization tools,</div><div style="left: 82px; top: 1104.61px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.950763);" data-canvas-width="359.6261851851851">surveillance experts pick the optimal time and place to be notified</div><div style="left: 464.667px; top: 290.683px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.03644);" data-canvas-width="359.5229259259255">based on historical data and therefore the optimal threshold, which</div><div style="left: 464.667px; top: 305.868px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00381);" data-canvas-width="291.8235740740739">will be verified by prospectively running the algorithms.</div><div style="left: 464.667px; top: 334.765px; font-size: 12.9074px; font-family: sans-serif; transform: scaleX(1.08548);" data-canvas-width="46.62155555555556">Results</div><div style="left: 478.333px; top: 348.387px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.96657);" data-canvas-width="346.09664074074084">The OHIM boundaries build the foundation for efficient spatial</div><div style="left: 464.667px; top: 363.572px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.997029);" data-canvas-width="358.3986907407408">modelling and visualization for public health surveillance in Ontario.</div><div style="left: 464.667px; top: 378.757px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.991664);" data-canvas-width="360.27155555555595">Together with the integrated modelling and visualization system,</div><div style="left: 464.667px; top: 393.942px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.02277);" data-canvas-width="359.3164074074074">staff are able to interactively optimize the aberration thresholds and</div><div style="left: 464.667px; top: 409.128px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.0082);" data-canvas-width="358.78591296296287">identify potential outbreaks in real time. Staff reported preference of</div><div style="left: 464.667px; top: 424.313px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00135);" data-canvas-width="306.51220370370356">SIR due to its faster computations and easier interpretation.</div><div style="left: 478.333px; top: 439.498px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.966578);" data-canvas-width="346.16375925925917">One major challenge was scalability: the ability to handle high</div><div style="left: 464.667px; top: 454.683px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.0139);" data-canvas-width="359.02211851851854">resolutions of spatiotemporal data. When the system was applied on</div><div style="left: 464.667px; top: 469.868px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00257);" data-canvas-width="358.55099814814804">4600 polygons by 200 weeks, significant delays were encountered in</div><div style="left: 464.667px; top: 485.054px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.04011);" data-canvas-width="359.7552592592591">both analysis and visualization. Difficulties in computational time,</div><div style="left: 464.667px; top: 500.239px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.979731);" data-canvas-width="359.87142592592585">memory requirement and visualization interactivity created delays</div><div style="left: 464.667px; top: 515.424px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.990644);" data-canvas-width="360.36319814814794">and freezing, thereby limited user experience. This problem was</div><div style="left: 464.667px; top: 530.609px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00127);" data-canvas-width="345.2215185185183">partially addressed by optimizing parameters for fast computations</div><div style="left: 464.667px; top: 559.505px; font-size: 12.9074px; font-family: sans-serif; transform: scaleX(1.10391);" data-canvas-width="77.45735185185185">Conclusions</div><div style="left: 478.333px; top: 573.128px; font-size: 12.9074px; font-family: serif; transform: scaleX(0.999342);" data-canvas-width="347.4544999999999">This work shows the “proof of concept” for an open source,</div><div style="left: 464.667px; top: 588.313px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.01174);" data-canvas-width="360.803340740741">customizable spatiotemporal surveillance system that overcomes</div><div style="left: 464.667px; top: 603.498px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00772);" data-canvas-width="358.74977222222213">existing data challenges in Ontario. However, more work is required</div><div style="left: 464.667px; top: 618.683px; font-size: 12.9074px; font-family: serif; transform: scaleX(1.00377);" data-canvas-width="295.76033333333334">to make this fully operational and efficient in production.</div>}, number={1}, journal={Online Journal of Public Health Informatics}, author={Li, Lennon and Pererita, Reuben and Johnson, Steven and Johnson, Ian}, year={2017}, month={May} }