In the New Year, toss a coin or trust Dr. Oz for medical advice?

by

coin_toss

—

Half of Dr. Oz’s med­ical advice is base­less or wrong, study says (Wash­ing­ton Post):

“…And now, his work has come under even greater scruti­ny in the British Med­ical Jour­nal, which on Wednes­day pub­lished a study ana­lyz­ing Oz’s claims along with those made on anoth­er med­ical talk show. What they found wasn’t reas­sur­ing. The researchers, led by Christi­na Korownyk of the Uni­ver­si­ty of Alber­ta, charged med­ical research either didn’t sub­stan­ti­ate — or flat out con­tra­dict­ed — more than half of Oz’s recommendations…They select­ed 40 episodes from last year, iden­ti­fy­ing 479 sep­a­rate med­ical rec­om­men­da­tions. After pag­ing through the rel­e­vant med­ical research, they found evi­dence only sup­port­ed 46 per­cent of his rec­om­men­da­tions, con­tra­dict­ed 15 per­cent and wasn’t avail­able for 39 percent.”

Study: Tele­vised med­ical talk shows—what they rec­om­mend and the evi­dence to sup­port their rec­om­men­da­tions: a prospec­tive obser­va­tion­al study (British Med­ical Jour­nal). From the abstract:

  • Objec­tive: To deter­mine the qual­i­ty of health rec­om­men­da­tions and claims made on pop­u­lar med­ical talk shows.
  • Sources: Inter­na­tion­al­ly syn­di­cat­ed med­ical tele­vi­sion talk shows that air dai­ly (The Dr Oz Show and The Doctors).
  • Inter­ven­tions: Inves­ti­ga­tors ran­dom­ly select­ed 40 episodes of each of The Dr Oz Show and The Doc­tors from ear­ly 2013 and iden­ti­fied and eval­u­at­ed all rec­om­men­da­tions made on each pro­gram. A group of expe­ri­enced evi­dence review­ers inde­pen­dent­ly searched for, and eval­u­at­ed as a team, evi­dence to sup­port 80 ran­dom­ly select­ed rec­om­men­da­tions from each show.
  • Results: We could find at least a case study or bet­ter evi­dence to sup­port 54% (95% con­fi­dence inter­val 47% to 62%) of the 160 rec­om­men­da­tions (80 from each show). For rec­om­men­da­tions in The Dr Oz Show, evi­dence sup­port­ed 46%, con­tra­dict­ed 15%, and was not found for 39%. For rec­om­men­da­tions in The Doc­tors, evi­dence sup­port­ed 63%, con­tra­dict­ed 14%, and was not found for 24%. Believ­able or some­what believ­able evi­dence sup­port­ed 33% of the rec­om­men­da­tions on The Dr Oz Show and 53% on The Doc­tors. On aver­age, The Dr Oz Show had 12 rec­om­men­da­tions per episode and The Doc­tors 11. The most com­mon rec­om­men­da­tion cat­e­go­ry on The Dr Oz Show was dietary advice (39%) and on The Doc­tors was to con­sult a health­care provider (18%). A spe­cif­ic ben­e­fit was described for 43% and 41% of the rec­om­men­da­tions made on the shows respec­tive­ly. The mag­ni­tude of ben­e­fit was described for 17% of the rec­om­men­da­tions on The Dr Oz Show and 11% on The Doc­tors. Dis­clo­sure of poten­tial con­flicts of inter­est accom­pa­nied 0.4% of recommendations.
  • Con­clu­sions: Rec­om­men­da­tions made on med­ical talk shows often lack ade­quate infor­ma­tion on spe­cif­ic ben­e­fits or the mag­ni­tude of the effects of these ben­e­fits. Approx­i­mate­ly half of the rec­om­men­da­tions have either no evi­dence or are con­tra­dict­ed by the best avail­able evi­dence. Poten­tial con­flicts of inter­est are rarely addressed. The pub­lic should be skep­ti­cal about rec­om­men­da­tions made on med­ical talk shows.