Edi­tor’s Note

Health Data Management - - INSIDE DEPARTMENTS - Fred Baz­zoli Edi­tor

AI moves from hype to re­al­ity.

HEALTH­CARE IS NO STRANGER TO THE in­fa­mous Gart­ner hype cy­cle. As a jour­nal­ist cov­er­ing the in­dus­try, it’s some­times in­trigu­ing to fig­ure out where an emerg­ing tech­nol­ogy is on this graph, which charts the typ­i­cal cy­cle from mo­ment of in­no­va­tion for an new tech­nol­ogy un­til it starts pro­duc­ing re­sults.

Ar­ti­fi­cial in­tel­li­gence is mak­ing its way along the ups and downs of the hype cy­cle. Right about now, AI ap­pears to be near the peak of in­flated ex­pec­ta­tions, per­haps slip­ping over the precipice onto the down­ward slope to­ward the trough of dis­il­lu­sion­ment.

Most HIT ex­ec­u­tives know the pat­tern— they’ve seen over­hyped ex­pec­ta­tions for new tech­nolo­gies, where hopes and dreams ob­scure the hard work that lies ahead. For ex­am­ple, high ex­pec­ta­tions sur­rounded the In­ter­net and elec­tronic pre­scrib­ing in the late 1990s, but it took years be­fore those tech­nolo­gies ma­trured. Only af­ter early dreams are dashed does the real work of in­no­va­tion, and achiev­ing ben­e­fits, set in.

I’d like to sug­gest that there’s a con­cur­rent cy­cle of ex­ag­ger­ated hys­te­ria with new tech­nol­ogy. I saw this dur­ing the past two RSNA con­fer­ences re­gard­ing AI. In 2016, AI was the mon­ster to be dreaded, be­cause ra­di­ol­o­gists feared it would ob­vi­ate the need for them to in­ter­pret im­ages. Last year, how­ever, those fears were nearly gone— and nearly ev­ery vendor in­sisted that it had in­cor­po­rated AI into its prod­ucts.

Ex­pec­ta­tions for AI in health­care are high, par­tic­u­larly when it comes to ex­pec­ta­tions for achiev­ing mea­sur­able im­prove­ments in pa­tient care and fi­nan­cial per­for­mance. How­ever, re­al­ity is hit­ting home—there are many stum­bling blocks to over­come be­fore ad­vanced de­ci­sion-aid­ing tech­nol­ogy is widely adopted, re­ports Linda Wil­son in our cover story for this is­sue.

In her re­port­ing, Wil­son notes that Paul Chang, MD, pro­fes­sor and vice chair­man of ra­di­ol­ogy in­for­mat­ics at the Univer­sity of Chicago School of Medicine, says “We are not at the so­lu­tion level yet; it is a long way be­tween ini­tial promis­ing re­sults in the lab and true clin­i­cally val­i­dated so­lu­tions.” But early work on us­ing and ben­e­fit­ing from AI is un­der way, she finds. Her story be­gins on Page 16.

AI is com­pet­ing for IT re­sources with other IT ini­tia­tives, and one of those in­volves in­for­ma­tion se­cu­rity. At­tacks on providers are com­mon­place, and in­stances of ran­somware at­tacks have been in­creas­ing over the past two years. In this is­sue, start­ing on Page 20, Man­ag­ing Edi­tor Greg Sla­bod­kin takes a mi­cro­scope to ex­am­ine the at­tack last year on Erie County Med­i­cal Cen­ter. Re­cov­ery from the at­tack was ex­ten­sive, time-con­sum­ing and ex­pen­sive, and ECMC used many work­arounds to en­sure care could be de­liv­ered safely to pa­tients.

Sla­bod­kin’s ac­count of the story is in­sight­ful, and we ap­plaud ECMC for al­low­ing us to tell its story. Such open­ness to share in­for­ma­tion with oth­ers in the in­dus­try is cru­cial in al­low­ing provider or­ga­ni­za­tions to in­form oth­ers about hack­ing gam­bits and best prac­tices at de­fense and re­cov­ery.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.