The Sick Sense

Some doc­tors say in­tu­ition can help to di­ag­nose pa­tients—but oth­ers are skep­ti­cal

The Walrus - - CONTENTS - by Amitha Kalaichan­dran il­lus­tra­tion by leeay aika wa

Should doc­tors use in­tu­ition to help di­ag­nose pa­tients? by Amitha Kalaichan­dran

Jonathan Sherbino was near­ing the end of his solo, nine-hour shift one De­cem­ber evening in 2011 at an emer­gency room in Hamil­ton, On­tario. He’d spent much of the day peer­ing at ear canals, tak­ing throat swabs, and or­der­ing X-rays in quick suc­ces­sion with few breaks. As the only doc­tor present, he knew he couldn’t linger for too long on any one di­ag­no­sis. A three-year-old with ruddy cheeks was next in line, her nose run­ning pro­fusely and mix­ing with the warm tears stream­ing down her face. When Sherbino ex­am­ined her, she seemed to have all the tell­tale signs. Fever: check. Cough and runny nose: check. Con­tact with some­one who had the flu: check. noth­ing ap­peared un­usual in her med­i­cal his­tory. It was likely she had the flu. But as Sherbino, an emer­gency physi­cian at St. Joseph’s Health­care Hamil­ton, pre­pared to sign a pre­scrip­tion pad for an an­tivi­ral med­i­ca­tion, some­thing forced him to slam the brakes. “It was as though I couldn’t keep writ­ing,” he later told me. “My mind was stop­ping me... it was like a cog­ni­tive block. Some­thing just didn’t fit the clin­i­cal pic­ture.” He re­turned to the mother to ask her a few more ques­tions. The child’s fever, it turned out, was on its fourth day, which was some­what odd — for most chil­dren, in his ex­pe­ri­ence, a flu-re­lated fever lasts two or three days. Her lips were also peel­ing — a symp­tom not usu­ally as­so­ci­ated with the flu. “It could very pos­si­bly be Kawasaki dis­ease,” Sherbino re­called think­ing at the time, re­fer­ring to a child­hood dis­ease of un­known cause. Its symp­toms of­ten in­clude at least five days of fever, red eyes, swollen lymph nodes, and a rash, along with other signs that ap­pear later. It is also rare: in many Western coun­tries, the dis­ease oc­curs in one in ev­ery 10,000 chil­dren un­der the age of five. He knew if his hunch were cor­rect, the tod­dler’s con­di­tion would quickly worsen — she might even de­velop a coro­nary artery aneurysm, which can lead to a heart at­tack and death, in some cases. Once Sherbino ex­plored the symp­tom his­tory fur­ther, Kawasaki more clearly fit the clin­i­cal pic­ture. Sure enough, when the child was trans­ferred to a lo­cal chil­dren’s hospi­tal the fol­low­ing day, Sherbino’s di­ag­no­sis was con­firmed, and the tod­dler re­ceived timely treat­ment that likely saved her life. The in­ci­dent led Sherbino, who is also the as­sis­tant dean of ed­u­ca­tion re­search at Mcmaster Univer­sity’s Fac­ulty of Health Sciences, down a new path in his work. It wasn’t con­scious logic that had ini­tially com­pelled him to re­visit his ini­tial di­ag­no­sis. Some might call it a gut feel­ing. Oth­ers have a med­i­cal term for it. With his col­league and men­tor, Geoffrey Nor­man, Sherbino has been study­ing what’s known as “clin­i­cal in­tu­ition” (or “di­ag­nos­tic rea­son­ing”) for the past eight years. Sherbino and his re­search team at Mcmaster de­scribe clin­i­cal in­tu­ition as a process by which med­i­cal prac­ti­tion­ers form a hy­poth­e­sis quickly and then ver­ify it though more an­a­lyt­i­cal means. Over time, and with ex­pe­ri­ence, clin­i­cal in­tu­ition is thought to be­come more ac­cu­rate, and the an­a­lyt­i­cal process oc­curs more quickly. But it is also a skill that can be taught and honed, Sherbino says. And in time-limited sit­u­a­tions, such as emer­gency-depart­ment ex­am­i­na­tions, he be­lieves, it could help more doc­tors sep­a­rate com­mon coughs and colds from rare, dev­as­tat­ing, and maybe even fa­tal dis­eases. As a physi­cian my­self, I learned anatomy, phys­i­ol­ogy, ev­i­dence-based medicine, how to record a med­i­cal his­tory, and phys­i­cal-exam skills in med­i­cal school. If they’re lucky, med­i­cal stu­dents im­prove

their em­pa­thy and lis­ten­ing skills from role mod­els. Now, as a pe­di­atrics res­i­dent, I re­al­ize that clin­i­cal in­tu­ition is not a skill stu­dents are for­mally taught. Med­i­cal prac­ti­tion­ers have his­tor­i­cally favoured a slow, an­a­lyt­i­cal ap­proach to de­ci­sion mak­ing: stu­dents are usu­ally taught to list symp­toms that sup­port a par­tic­u­lar di­ag­no­sis, as well as those that might con­tra­dict it, be­fore com­ing to a con­clu­sion. Al­though some in the med­i­cal com­mu­nity see prom­ise in clin­i­cal-in­tu­ition re­search, to crit­ics, the idea is not es­tab­lished enough to risk al­ter­ing a cen­turies-old canon of med­i­cal teach­ing.

Pat Croskerry, an emer­gency-room doc­tor and direc­tor of the crit­i­cal-think­ing pro­gram at Dal­housie Univer­sity’s divi­sion of med­i­cal ed­u­ca­tion, has penned let­ters to the jour­nal Aca­demic Medicine draw­ing cau­tion to re­search done by Sherbino and his team. Croskerry stud­ies how physi­cians can avoid cog­ni­tive bi­ases — er­rors in think­ing — that can lead them to hold onto the first di­ag­no­sis that comes to mind, even if new in­for­ma­tion con­tra­dicts it. Low­er­ing the rate of mis­di­ag­noses — es­ti­mated to be at about 10 to 15 per­cent of all cases — is a widely held goal in the med­i­cal com­mu­nity. Croskerry be­lieves that what Sherbino and oth­ers call clin­i­cal in­tu­ition could be termed “fast de­ci­sion mak­ing,” which al­lows us to make quick as­so­ci­a­tions be­tween ideas. (A kid with a runny nose, mus­cle aches, and fever around fam­ily mem­bers with the flu? Must be the flu.) He says an over­re­liance on fast de­ci­sion mak­ing can lead to harm­ful de­pen­dence on com­mon cog­ni­tive bi­ases. “We can and do use [fast] de­ci­sion mak­ing ex­ten­sively,” Croskerry says. “But we should al­ways be check­ing it with our slower, me­thod­i­cal, ev­i­dence-based mind.” It’s the lat­ter com­po­nent, he says, that be­comes stronger with in­creased ex­pe­ri­ence and knowl­edge: over time, as they see more cases, doc­tors and nurses are able to elim­i­nate wrong di­ag­noses more quickly. But “tak­ing your time doesn’t change er­ror rates,” Sherbino says, an as­sess­ment he makes based off his own re­search. “It just makes you slower.” He may be right. One night, when I was a ju­nior res­i­dent, an­other res­i­dent and I ad­mit­ted a child with joint pain. The pa­tient’s con­di­tion seemed to both of us most in keep­ing with an au­toim­mune dis­ease, and though we each came up with slightly dif­fer­ent di­ag­noses, we used a slower, an­a­lyt­i­cal ap­proach — the one we’d been taught through­out our train­ing. We knew there was a pos­si­bil­ity it could be a se­ri­ous in­fec­tion, but we didn’t con­sider it likely un­til we re­viewed the case with a more ex­pe­ri­enced res­i­dent. Though the pa­tient didn’t demon­strate a true fever — a sign of an in­fec­tion — our col­league was in­clined to think the pain was caused by a bac­te­rial in­fec­tion, pos­si­bly based on an eerily sim­i­lar case he had seen the year be­fore. He turned out to be right, and the pa­tient was soon given high-dose an­tibi­otics to fight the in­fec­tion. If the di­ag­no­sis had been missed, it would have been dev­as­tat­ing for the pa­tient. It’s hard to tell whether clin­i­cal in­tu­ition prompted my col­league to re­think our di­ag­no­sis or whether his ad­di­tional years of ex­pe­ri­ence sim­ply al­lowed him to draw on more data. But even if clin­i­cal in­tu­ition is ef­fec­tive, so far the avail­able re­search is un­clear to what ex­tent med­i­cal prac­ti­tion­ers should rely on it. De­spite their com­pet­ing pri­or­i­ties, Sherbino and Croskerry seem to be tack­ling the same sys­temic prob­lems: as the de­mands on Canada’s health sys­tem con­tinue to in­crease, man­i­fest­ing as longer wait times and over­flow­ing emer­gency rooms, doc­tors and nurses are hav­ing to make de­ci­sions in high-stress sit­u­a­tions on a daily ba­sis, of­ten with limited re­sources and time. For early ca­reer med­i­cal prac­ti­tion­ers, then, is a shift to­ward a more in­stinc­tive ap­proach — one that de­fies the tra­di­tional method — in­evitable?

The key to avoid­ing mis­di­ag­noses, it turns out, may have less to do with the de­ci­sion-mak­ing process it­self than with con­fi­dence in one’s de­ci­sions. And con­fi­dence and ex­pe­ri­ence don’t nec­es­sar­ily go hand in hand. A re­cent pa­per, co-au­thored by Car­men Sanchez, a PHD can­di­date at Cor­nell Univer­sity, in­cluded sev­eral stud­ies that sur­veyed par­tic­i­pants tasked with mak­ing de­ci­sions about whether a pa­tient was healthy or had one of two “zom­bie dis­eases” — fic­tional ill­nesses with made-up symp­toms that were cre­ated for the pur­poses of the stud­ies. The par­tic­i­pants, none of whom were for­mally trained in clin­i­cal medicine, were asked how con­fi­dent they were in each of their di­ag­noses be­for­ere­ceiv­ing feed­back about whether they were ac­cu­rate. Though the par­tic­i­pants’ ac­cu­racy slowly and in­cre­men­tally in­creased with ex­pe­ri­ence, their con­fi­dence fol­lowed a dif­fer­ent pat­tern. At first, the novices were aware of their in­abil­ity. Af­ter just a lit­tle learn­ing, their con­fi­dence shot up, though their ac­tual per­for­mance hardly changed. To­ward the end of the ex­per­i­ment, their con­fi­dence in­creased again, pos­si­bly be­cause their ac­cu­racy im­proved and they had seen more sub­jects over time. The study shows that clin­i­cal in­tu­ition isn’t nec­es­sar­ily more ef­fec­tive than the tra­di­tional method, or vice versa, Sanchez says: “The trou­ble is when over­con­fi­dence clouds judge­ment, and that can hap­pen with both less ex­pe­ri­enced and more ex­pe­ri­enced doc­tors.” With ris­ing pres­sures in health care, physi­cians at all lev­els of train­ing may be forced to make ef­fi­cient de­ci­sions about any­thing from triag­ing which pa­tients are sick enough to be ad­mit­ted to which are more likely to have symp­toms sug­ges­tive of a dev­as­tat­ing dis­ease, all while min­i­miz­ing the risk of mis­di­ag­no­sis. Clin­i­cal in­tu­ition, if used ap­pro­pri­ately, and when drawn from a vast pool data from adoc­tor’s own ex­pe­ri­ence or that of oth­ers, may be one way to help make ac­cu­rate di­ag­nos­tic de­ci­sions when the stakes are high and time and re­sources are limited. The best cases, I sus­pect, are when de­ci­sions are made with hu­mil­ity, through con­sult­ing with those more ex­pe­ri­enced, to ef­fec­tively crowd­source the di­ag­nos­tic pos­si­bil­i­ties, and tak­ing a more an­a­lytic ap­proach when­ever pos­si­ble.

AMITHA KALAICHAN­DRAN is a res­i­dent physi­cian and jour­nal­ist whose work has ap­peared in the New York Times, the Bos­ton Globe, and The Atavist.

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.