{"id":20855,"date":"2020-11-04T07:17:31","date_gmt":"2020-11-04T12:17:31","guid":{"rendered":"https:\/\/michigan.it.umich.edu\/news\/?p=20855"},"modified":"2024-07-08T06:04:43","modified_gmt":"2024-07-08T10:04:43","slug":"patient-data-tells-a-story-to-help-michigan-medicine-clinicians-read-ahead","status":"publish","type":"post","link":"https:\/\/michigan.it.umich.edu\/news\/2020\/11\/04\/patient-data-tells-a-story-to-help-michigan-medicine-clinicians-read-ahead\/","title":{"rendered":"Patient data tells a story to help Michigan Medicine clinicians read ahead"},"content":{"rendered":"\n
\""
(Michigan Medicine\/Shutterstock)<\/figcaption><\/figure><\/div>\n\n\n\n

Hospital patients produce a tremendous amount of data. Throughout Michigan Medicine, MCIRCC\u2019s PICTURE analytic is reading the stories within this data to help clinicians act on potential emergencies chapters in advance.<\/p>\n\n\n\n

In the summer of 2020, the Michigan Center for Integrative Research in Critical Care (MCIRCC) detailed its predictive analytic \u201cPICTURE,\u201d<\/a> which crunches an array of patient data\u2014including health records, labs, and vital signs\u2014to predict ICU transfer or death as a proxy for patient deterioration. Since the original article, PICTURE and its suite of specialized analytics have been moving full steam ahead in both the research sphere and at Michigan Medicine. The team behind PICTURE-General (a version of the analytic tuned to general patients) recently published two major papers: the first, available in the JMIR: Medical Informatics, demonstrates the consequences of learning “missingness” patterns in early warning systems for preventative health care<\/a>; the second, a comparison of PICTURE-General and the Epic Deterioration Index in a setting of COVID-19, is published as a preprint in MedRXiv<\/em><\/a> and has also been submitted to the Journal of Medical Internet Research.<\/p>\n\n\n\n

At Michigan Medicine, the Clinical Intelligence Committee green-lit PICTURE-General for real-time implementation, and the software is in the process of being uploaded to the Epic Cloud platform. The team is planning an implementation study to learn how the analytic fits into the workflows of the ICU. \u201cIt\u2019s important to make sure we are providing new and useful
information in a timely manner without interrupting them,\u201d notes Brandon Cummings, a data scientist and machine learning specialist at MCIRCC.<\/p>\n\n\n\n

\"An
\u201cThe PICTURE product line is another great example of MCIRCC using multidisciplinary
innovation and integration to develop scalable platforms that span the spectrum of
critical illness and injury.\u201d – Kevin Ward, MCIRCC Executive Director<\/figcaption><\/figure>\n\n\n\n

Of course, the analytic\u2019s capabilities don\u2019t end at the ICU. Under the PICTURE umbrella is a full suite of specialized versions adapted for various patient populations. Along with PICTURE-General, these include PICTURE-Rehab\u2014currently running in Michigan Medicine\u2019s Physical Medicine and Rehabilitation Unit\u2014and PICTURE-COVID-19. Both have also seen major
developments in recent months.<\/p>\n\n\n\n

PICTURE-Rehab\u2019s addition to the Epic Cloud platform recently contributed towards a funding award from Epic\u2019s Honor Roll Program. With Michigan Medicine expanding its Physical Medicine & Rehabilitation unit to St. Joseph Mercy Chelsea, the PICTURE-Rehab team is aiming to have the analytic implemented there, as well. <\/p>\n\n\n\n

PICTURE-COVID-19 was awarded a MIDAS grant in May. To further hone its performance, the team is now training the network through transfer learning\u2014a technique in machine learning through which the analytic takes knowledge it gained while solving one problem and applies it to a different but related problem. <\/p>\n\n\n\n

Versions of PICTURE tuned for pediatric patients and patients with sepsis are also currently in development.<\/p>\n\n\n\n

Learn more<\/h2>\n\n\n\n