Joseph Colorafi, MD, is a board-cer­ti­fied pul­mo­nolo­gist. Early in his ca­reer, he of­ten worked with sep­sis pa­tients, seek­ing to pro­vide the care that would help them re­turn to health. In his cur­rent role as vice pres­i­dent and chief med­i­cal in­for­ma­tion of­fi­cer at Dig­nity Health, how­ever, Colorafi doesn’t pro­vide much di­rect pa­tient care. Cu­ri­ously, though, he now feels as if he is hav­ing a greater im­pact on pa­tients than ever be­fore.

“In the past, I would see the sep­sis pa­tients after they had cas­caded down­stream and the dis­ease had pro­gressed some. Here I am years later and I am hav­ing a greater im­pact by help­ing to pre­vent sep­sis from de­vel­op­ing in the first place,” Co­larafi said.

How so? Colorafi is lead­ing ef­forts to use big data and an­a­lyt­ics to first iden­tify pa­tients who show a high risk for de­vel­op­ing sep­sis and then get them the treat­ment and ser­vices re­quired to stop the in­fec­tion in its tracks. More specif­i­cally, the San Fran­cisco-based health sys­tem is us­ing Hadoop — an open­source soft­ware frame­work that stores and runs ap­pli­ca­tions on large clus­ters of data — to sup­port a bio sur­veil­lance pro­gram that fo­cuses on sep­sis pre­ven­tion. With this sys­tem in place, the health sys­tem is iden­ti­fy­ing sep­sis as soon as pos­si­ble, mak­ing it pos­si­ble to have a pos­i­tive im­pact on sur­vival rates, in­fec­tion and in­ten­sive care unit length of stay. In­deed, when clin­i­cians are alerted to a sus­pected in­fec­tion, they are able to take key ac­tions out­lined in a “sep­sis bun­dle” within three hours, greatly im­prov­ing care. In fact, for ev­ery 10% in­crease in com­pli­ance with the sep­sis bun­dle, mor­tal­ity is re­duced by 3%.

In ad­di­tion to re­ly­ing on data an­a­lyt­ics to im­prove sep­sis care, Dig­nity is lever­ag­ing data an­a­lyt­ics to pre­vent harm­ful events such as over se­da­tion with opi­oids and to re­duce read­mis­sions.

Jump­ing on the band­wagon

Dig­nity is just one of many health­care or­ga­ni­za­tions seek­ing to move the clin­i­cal care nee­dle for­ward through the ex­panded use of data and an­a­lyt­ics. Cen­tura Health, for ex­am­ple, is lever­ag­ing data to sup­port more ef­fec­tive em­ployee well­ness ini­tia­tives.

“We are us­ing claims data and con­sumer data to pre­dict and model how we can en­cour­age em­ploy­ees to get needed pre­ven­ta­tive screen­ings – mam­mo­grams, cer­vi­cal can­cer screen­ings and col­orec­tal screen­ings. Ba­si­cally, we are look­ing to do what the air­lines and other re­tail or­ga­ni­za­tions do and meet con­sumers where they are at. Our hope is to drill down and get more per­sonal by us­ing data and an­a­lyt­ics to of­fer ser­vices that are rel­e­vant to in­di­vid­ual em­ploy­ees,” said Angie Vil­la­maria, direc­tor of as­so­ci­ate well­ness.

For ex­am­ple, the Colorado-based health sys­tem is an­a­lyz­ing claims, bio­met­ric screen­ing and health as­sess­ment data to iden­tify em­ploy­ees who have a high body mass in­dex and who would be re­cep­tive to par­tic­i­pat­ing in a free weight loss coach­ing pro­gram. Cen­tura’s pop­u­la­tion data is com­bined with con­sumer data that in­cludes in­for­ma­tion on 270 mil­lion Amer­i­cans to create con­sumer in­sights and build pre­dic­tive mod­els to strat­ify pop­u­la­tions based on in­di­vid­u­als’ needs, re­cep­tiv­ity and like­li­hood of par­tic­i­pat­ing in a tar­geted pro­gram.

“In health­care, we have tra­di­tion­ally man­aged peo­ple when they are sick. But it is im­por­tant to shift think­ing to the con­sumer to buy into their health and take some own­er­ship. By pro­vid­ing ser­vices that are ap­peal­ing to them we are look­ing to bend the curve and im­prove health out­comes,” Vil­la­maria said.

While many health­care or­ga­ni­za­tions are lever­ag­ing data an­a­lyt­ics to im­prove com­pli­ance with cur­rent best prac­tices, some pro­fes­sion­als are look­ing to dis­cover more ef­fec­tive treat­ments. For ex­am­ple, rep­re­sen­ta­tives from a va­ri­ety of health­care or­ga­ni­za­tions re­cently came to­gether at a con­fer­ence con­vened by the Su­san G. Komen breast can­cer or­ga­ni­za­tion to ex­plore breast can­cer re­search and im­prove treat­ment through the use of big data.

“We are work­ing for mul­ti­dis­ci­plinary col­lab­o­ra­tions that can help us ap­ply big data to im­prove breast can­cer treat­ment and out­comes,” Ge­orge Sledge, MD, Komen chief sci­en­tific ad­vi­sor and a pro­fes­sor of medicine

“Our hope is to drill down and get more per­sonal by us­ing data and an­a­lyt­ics to of­fer ser­vices that are rel­e­vant to in­di­vid­ual em­ploy­ees.”

at Stan­ford Univer­sity said in a press re­lease. “We plan to iden­tify op­por­tu­ni­ties to work to­gether on the very promis­ing, but com­pli­cated, use of big data.”

In fact, con­fer­ence lead­ers were look­ing to lever­age big data ap­pli­ca­tions in an ef­fort to im­prove the fo­cus and, hope­fully, speed of re­search, mak­ing it pos­si­ble to more quickly de­velop treat­ments that will im­prove how pa­tients feel, func­tion and sur­vive.

In­vest­ing in so­lu­tions

With all of this in­ter­est in big data and an­a­lyt­ics, or­ga­ni­za­tions are in­creas­ingly grav­i­tat­ing to­ward more so­phis­ti­cated tech­nolo­gies, said David Devine, a manag­ing direc­tor with Huron, Chicago. “Big data and an­a­lyt­ics typ­i­cally fall into two cat­e­gories – present the cur­rent data on hand and to pre­dict events. The ma­jor­ity of the tech­nol­ogy that ex­ists now is fo­cused on amass­ing and ag­gre­gat­ing data and pre­sent­ing it in dash­boards to show how it mea­sures up against pre-mapped out per­for­mance in­di­ca­tors. But we are start­ing to see the move to­ward us­ing the data to pre­dict what will hap­pen. So, we are work­ing with more or­ga­ni­za­tions to put the tech­nol­ogy in place that will help them pre­dict what will hap­pen with pa­tients,” Devine said.

In­deed, many or­ga­ni­za­tions are look­ing to turn their elec­tronic health records sys­tems from “sys­tems of record” into “sys­tems of ac­tion,” Devine said.

“We are start­ing to see the move to­ward us­ing the data to pre­dict what will hap­pen.”

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.