TY - JOUR AU - Loschen, Wayne AU - Burkom, Howard AU - Atrubin, David PY - 2017/05/02 Y2 - 2024/03/29 TI - Jurisdictional Usage of the New ESSENCE Word Alert Feature JF - Online Journal of Public Health Informatics JA - OJPHI VL - 9 IS - 1 SE - Non-Infectious Disease Surveillance Use Cases DO - 10.5210/ojphi.v9i1.7718 UR - https://ojphi.org/ojs/index.php/ojphi/article/view/7718 SP - AB - <div style="left: 78.2727px; top: 278.962px; font-size: 12.3207px; font-family: sans-serif; transform: scaleX(1.0823);" data-canvas-width="55.467823232323234">Objective</div><div style="left: 91.3182px; top: 291.965px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.994629);" data-canvas-width="329.0565161616161">The objective of this presentation is to describe the new word alert</div><div style="left: 78.2727px; top: 306.46px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.964851);" data-canvas-width="343.6959803030303">capability in ESSENCE and how it has been used by the Florida</div><div style="left: 78.2727px; top: 320.955px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.974143);" data-canvas-width="343.48283207070676">Department of Health (FDOH). Specifically, this presentation will</div><div style="left: 78.2727px; top: 335.45px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.955276);" data-canvas-width="343.5333469696968">describe how the word alert feature works to find individual chief</div><div style="left: 78.2727px; top: 349.945px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0137);" data-canvas-width="344.7937553030298">complaint terms that are occurring at an abnormal rate. It will</div><div style="left: 78.2727px; top: 364.44px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.97005);" data-canvas-width="343.52718661616154">then provide usage statistics and first-person accounts of how the</div><div style="left: 78.2727px; top: 378.935px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.978251);" data-canvas-width="343.6269843434343">alerts have impacted public health practice for the users. Finally,</div><div style="left: 78.2727px; top: 393.43px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.98323);" data-canvas-width="343.9337699494947">the presentation will offer future enhancement possibilities and a</div><div style="left: 78.2727px; top: 407.925px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00354);" data-canvas-width="307.9683939393938">summary of the benefits and shortcomings of this new feature.</div><div style="left: 78.2727px; top: 435.508px; font-size: 12.3207px; font-family: sans-serif; transform: scaleX(1.11884);" data-canvas-width="71.86668434343434">Introduction</div><div style="left: 91.3182px; top: 448.511px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.02241);" data-canvas-width="331.9752916666665">Syndromic surveillance systems have historically focused on</div><div style="left: 78.2727px; top: 463.006px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.988668);" data-canvas-width="341.9144060606058">aggregating data into syndromes for analysis and visualization. These</div><div style="left: 78.2727px; top: 477.501px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.964936);" data-canvas-width="343.5013131313132">syndromes provide users a way to quickly filter large amounts of</div><div style="left: 78.2727px; top: 491.996px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0245);" data-canvas-width="342.98630757575756">data into a manageable number of streams to analyze. Additionally,</div><div style="left: 78.2727px; top: 506.491px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0273);" data-canvas-width="343.0146452020202">ESSENCE users have the ability to build their own case definitions</div><div style="left: 78.2727px; top: 520.985px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.03836);" data-canvas-width="339.3369141414142">to look for records matching particular sets of criteria. Those user-</div><div style="left: 78.2727px; top: 535.48px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00263);" data-canvas-width="342.18053333333336">defined queries can be stored and analyzed automatically, along with</div><div style="left: 78.2727px; top: 549.975px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.973021);" data-canvas-width="339.3085765151513">the pre-defined syndromes. Aside from these predefined and user-</div><div style="left: 78.2727px; top: 564.47px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00711);" data-canvas-width="342.4343398989899">defined syndromic categories, ESSENCE did not previously provide</div><div style="left: 78.2727px; top: 578.965px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.01702);" data-canvas-width="342.7189482323232">alerts based on individual words in the chief complaint text that had</div><div style="left: 78.2727px; top: 593.46px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.963073);" data-canvas-width="343.68858787878776">not been specified a priori. Thus, an interesting cluster of records</div><div style="left: 78.2727px; top: 607.955px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.999255);" data-canvas-width="342.17806919191923">linked only by non-syndromic keywords would likely not be brought</div><div style="left: 78.2727px; top: 622.45px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00373);" data-canvas-width="99.92093434343435">to a user’s attention.</div><div style="left: 78.2727px; top: 650.033px; font-size: 12.3207px; font-family: sans-serif; transform: scaleX(1.07424);" data-canvas-width="50.65042676767676">Methods</div><div style="left: 91.3182px; top: 663.036px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.997228);" data-canvas-width="331.3457035353535">In the FDOH ESSENCE system a new detection feature was</div><div style="left: 78.2727px; top: 677.531px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.02198);" data-canvas-width="342.97768308080805">developed to trigger alerts based on anomalous occurrence of terms</div><div style="left: 78.2727px; top: 692.026px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.984943);" data-canvas-width="101.99081313131315">in chief complaints.</div><div style="left: 180.271px; top: 692.352px; font-size: 7.39242px; font-family: serif;">1</div><div style="left: 184.027px; top: 692.026px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.957123);" data-canvas-width="237.81305580808086">This feature used Fisher’s Exact Test to test</div><div style="left: 78.2727px; top: 706.521px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.02434);" data-canvas-width="343.07008838383825">frequencies of individual chief complaint terms relative to all terms</div><div style="left: 78.2727px; top: 721.016px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.970207);" data-canvas-width="343.58262979797985">in a 1-month baseline. The feature used a 7-day guard-band, and</div><div style="left: 78.2727px; top: 735.511px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.01367);" data-canvas-width="342.6043656565656">automatically switched to an efficient chi-square test for sufficiently</div><div style="left: 78.2727px; top: 750.006px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.04735);" data-canvas-width="294.7236338383838">large term counts. A term triggered an alert if its p-value</div><div style="left: 373.142px; top: 752.346px; font-size: 12.3207px; font-family: sans-serif;">≤</div><div style="left: 379.97px; top: 750.006px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.05431);" data-canvas-width="41.43453787878787">10E-4.</div><div style="left: 78.2727px; top: 764.501px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00375);" data-canvas-width="342.3641118686869">This algorithm was then run on chief complaint sets both by hospital</div><div style="left: 78.2727px; top: 778.996px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00716);" data-canvas-width="342.4663737373737">and by region, with region assignment according to patient zip code.</div><div style="left: 78.2727px; top: 793.491px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.02426);" data-canvas-width="343.04298282828285">Results were then displayed in new visualizations showing alerts in</div><div style="left: 78.2727px; top: 807.985px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.01852);" data-canvas-width="342.73250101010103">word cloud and line listing form. Additionally, users were given the</div><div style="left: 78.2727px; top: 822.48px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00563);" data-canvas-width="342.3838249999998">option to ignore stop words, syndromic terms, and a user-created list</div><div style="left: 78.2727px; top: 836.975px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00119);" data-canvas-width="318.8968611111109">of ignorable words in order to focus on words of greater interest.</div><div style="left: 78.2727px; top: 864.558px; font-size: 12.3207px; font-family: sans-serif; transform: scaleX(1.08542);" data-canvas-width="44.50239393939395">Results</div><div style="left: 91.3182px; top: 877.561px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0256);" data-canvas-width="329.8499696969695">The result of using the tool since June 2016 has seen three major</div><div style="left: 78.2727px; top: 892.056px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.953307);" data-canvas-width="343.4766717171718">benefits. First, the original intent for the system to notify users of</div><div style="left: 78.2727px; top: 906.551px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.975664);" data-canvas-width="341.775182070707">abnormal word clusters has proven useful. Users have been able to see</div><div style="left: 78.2727px; top: 921.046px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.992108);" data-canvas-width="68.8357904040404">terms such as</div><div style="left: 147.105px; top: 921.046px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.02219);" data-canvas-width="82.61034090909091">Disaster, Shelter</div><div style="left: 229.712px; top: 921.046px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.983254);" data-canvas-width="23.5325505050505">and</div><div style="left: 253.243px; top: 921.046px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.01538);" data-canvas-width="50.65042676767678">Fireworks</div><div style="left: 303.891px; top: 921.046px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.990302);" data-canvas-width="116.44300252525254">which were not part of</div><div style="left: 78.2727px; top: 935.541px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.977866);" data-canvas-width="341.5361603535353">any prior syndromes and use these notifications to investigate possible</div><div style="left: 78.2727px; top: 950.036px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00556);" data-canvas-width="342.2926517676768">issues. The second benefit found by users was the ability to find new</div><div style="left: 78.2727px; top: 964.531px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.986822);" data-canvas-width="341.9501361111111">misspellings or abbreviations commonly used by hospitals. The terms</div><div style="left: 78.2727px; top: 979.026px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.957068);" data-canvas-width="24.197868686868684">Zyka</div><div style="left: 102.479px; top: 979.026px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.08132);" data-canvas-width="25.8796452020202">and</div><div style="left: 128.358px; top: 979.026px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0097);" data-canvas-width="23.458626262626257">GLF</div><div style="left: 151.827px; top: 979.026px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.04499);" data-canvas-width="269.50191439393933">(Ground Level Fall) are examples of these. Finally,</div><div style="left: 78.2727px; top: 993.521px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00598);" data-canvas-width="342.41955505050504">the system has helped discover new trends in hospital processes. For</div><div style="left: 78.2727px; top: 1008.02px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0386);" data-canvas-width="343.2240972222222">example, the tool has helped discover first person and non-English</div><div style="left: 78.2727px; top: 1022.51px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.01047);" data-canvas-width="342.56740353535355">phrases in the chief complaint. This observation led to the discovery</div><div style="left: 78.2727px; top: 1037.01px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.03278);" data-canvas-width="343.08856944444415">that some hospitals are using kiosks or mobile phone apps to allow</div><div style="left: 78.2727px; top: 1051.5px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00136);" data-canvas-width="216.610351010101">patients to enter their own chief complaints.</div><div style="left: 443.545px; top: 278.962px; font-size: 12.3207px; font-family: sans-serif; transform: scaleX(1.10408);" data-canvas-width="73.93656313131314">Conclusions</div><div style="left: 456.591px; top: 291.965px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.03776);" data-canvas-width="329.85243383838383">The word alert feature has provided value to the users of FDOH</div><div style="left: 443.545px; top: 306.46px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.0555);" data-canvas-width="345.9765431818181">ESSENCE. While accomplishing its initial goal of triggering</div><div style="left: 443.545px; top: 320.955px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.968902);" data-canvas-width="343.4273888888891">abnormal non-syndromic term usage, the additional ability to find</div><div style="left: 443.545px; top: 335.45px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.01121);" data-canvas-width="342.5969732323232">new misspellings and abbreviations may have even larger impact by</div><div style="left: 443.545px; top: 349.945px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00133);" data-canvas-width="342.1891578282828">keeping syndrome and subsyndrome definitions up-to-date over time</div><div style="left: 443.545px; top: 364.44px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.992168);" data-canvas-width="343.8524532828282">for traditional syndromic alerting. Beyond these current benefits,</div><div style="left: 443.545px; top: 378.935px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.04283);" data-canvas-width="345.35065126262623">additional visualization enhancements are under consideration.</div><div style="left: 443.545px; top: 393.43px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.992316);" data-canvas-width="344.15184646464627">Additionally, the resources required to perform the detection are</div><div style="left: 443.545px; top: 407.925px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.978588);" data-canvas-width="341.75546893939384">substantial, and implementation improvements are under development</div><div style="left: 443.545px; top: 422.42px; font-size: 12.3207px; font-family: serif; transform: scaleX(0.99363);" data-canvas-width="344.3921002525252">to improve the performance and enable more advanced free-text</div><div style="left: 443.545px; top: 436.915px; font-size: 12.3207px; font-family: serif; transform: scaleX(1.00154);" data-canvas-width="93.76058080808083">anomaly detection.</div> ER -