PRE­DICT­ING SELF HARM

The East African - - OUTLOOK -

De­spite be­ing the 10th lead­ing cause of death in the US, sui­cide re­mains ex­tremely dif­fi­cult to pre­dict. Ex­perts say that’s be­cause many peo­ple’s risk for self-harm is paired with an­other men­tal ill­ness and fluc­tu­ates ac­cord­ing to var­i­ous stres­sors in their life, all of which in­ter­act uniquely within each in­di­vid­ual. Com­pli­cat­ing mat­ters is that sui­ci­dal ideation — which can sig­nal a grow­ing risk for self

‘happy,’ ‘love,’ and ‘fun.’”

Dan­forth said men­tal health pro­fes­sion­als are still de­pen­dent on the Di­ag­nos­tic and Sta­tis­ti­cal Man­ual of Men­tal Dis­or­ders (the DSM) and one-on-one in­ter­views, but he be­lieves the data be­ing amassed by smart­phones means men­tal health is on the verge of a “dig­i­tal rev­o­lu­tion.”

“We’re al­ready talk­ing with doc­tors at the Univer­sity of Ver­mont who want to build a screen­ing tool for the emer­gency room that would ask peo­ple whether they’d be will­ing to have an al­go­rithm look at their so­cial me­dia his­tory,” Dan­forth noted.

The tools, how­ever, will only be as good as the data that is used to cre­ate ma­chine learn­ing al­go­rithms, Harkavy-fried­man said, not­ing that there is a lack of lon­gi­tu­di­nal stud­ies on sui­cide. So­cial me­dia will of­fer im­por­tant in­for­ma­tion, she said, but any pop­u­la­tion of peo­ple be­ing stud­ied will al­ways in­clude false pos­i­tives — peo­ple who ex­hibit sui­ci­dal be­hav­iours but don’t go on to end their own lives.

“The more we can learn about both the fac­tors lead­ing to sui­cide, the bet­ter off we’ll be,” she said.

Ex­perts said it could take an­other five to 10 years to cre­ate al­go­rithms pre­dic­tive enough to be re­li­ably de­ployed in­side hos­pi­tals, schools and ther­a­pists’ of­fices. Ques­tions will have to be re­solved as well, ex­perts said, such as whether pre­dic­tive al­go­rithms will af­fect health in­sur­ance pre­mi­ums or what hap­pens if drug com­pa­nies man­age to ac­cess peo­ple’s pre­dic­tive data?

John Pes­tian, a physi­cian and pro­fes­sor in the divi­sions of Biomed­i­cal In­for­mat­ics and psy­chi­a­try at Cincin­nati Chil­dren’s Hospi­tal Med­i­cal Cen­tre within the Univer­sity of Cincin­nati, said it’s too early to an­swer those ques­tions. When he cre­ated SAM he was only fo­cused on one thing: Alle­vi­at­ing suf­fer­ing with tech­nol­ogy.

“You go into the emer­gency depart­ment and you go to the in­ten­sive care unit and you see tech­nol­ogy ev­ery­where, but you go into a psy­chi­a­trist’s of­fice and you see a couch,” Pes­tian said.

“Why?” he added. “Be­cause folks like me haven’t built any­thing for them un­til now.”

Newspapers in English

Newspapers from Kenya

© PressReader. All rights reserved.