EXCERPTED

  from   “ENDING   MEDICAL   REVERSAL,”   by   VINAYAK   PRASAD   and   ADAM   CIFU.  

Published   by   Johns   Hopkins   University   Press  ©   2015.     Reprinted   by   permission   of   the  
publisher.  
 
Chapter  1    
 
What  Is  Medical  Reversal?  
 
We  expect  that  medical  therapy  will  change  and  evolve  with  time.  Good  treatments  will  replace  
bad   ones,   and   then   better   ones   will   replace   those.   Antibiotics   have   replaced   arsenic,   and  
anesthesia  has  replaced  a  bullet  held  bracingly  between  the  teeth.  Recently,  however,  change  
has  occurred  in  surprising  ways.  If  you  have  followed  the  news  about  prostate  cancer  screening,  
mammography   for   women   in   their   forties,   hormone   replacement,   cholesterol-­‐lowering  
medications,   and   stents   for   coronary-­‐artery   disease,   you   might   think   doctors   cannot   get  
anything   straight.   These   common   practices   were   not   replaced   by   better   therapies;   they   were  
found  to  be  ineffective.  In  some  cases,  they  did  more  harm  than  good.  You  might  be  worried  
that  some  medical  practices  are  nothing  more  than  fads.  In  some  cases,  you  would  be  right.  
 
We  call  this  phenomenon  “medical  reversal.”  Instead  of  the  ideal,  which  is  replacement  of  good  
medical   practices   by   better   ones,   medical   reversal   occurs   when   a   currently   accepted   therapy   is  
overturned—found  to  be  no  better  than  the  therapy  it  replaced.  Now,  you  might  argue  that  this  
is   how   science   is   supposed   to   proceed.   In   high   school,   we   learned   that   the   scientific   method  
involves  proposing  a  hypothesis  and  testing  to  see  whether  it  is  right.  This  is  true.  But  what  has  
happened   in   medicine   is   that   the   hypothesized   treatment   is   often   instituted   in   millions   of  
people,   and   billions   of   dollars   are   spent,   before   adequate   research   is   done.   Not   surprisingly,  
sometimes   the   research   demonstrates   that   the   hypothesis   was   incorrect   and   that   the  
treatment,  which  is  already  being  used,  is  ineffective  or  harmful.  
 
Cases  of  Reversal  
 
A   few   people’s   stories   highlight   notorious   examples   of   reversal   in   medicine.   Consider   Samuel  
Jones,   a   57-­‐year-­‐old   gentleman   who   had   a   heart   attack   in   1991.   He   was   admitted   to   the  
hospital  and  treated  with  the  standard  medical  regimen,  which,  at  the  time,  included  flecainide,  
a  drug  used  to  stabilize  his  heart  rhythm.  In  the  late  1980s  and  early  1990s,  flecainide  and  its  
sibling   drugs   were   widely   used   with   the   intention   of   suppressing   extra   heartbeats   (called  
premature   ventricular   contractions,   or   PVCs)   and   preventing   death.   The   logic   behind   the   use   of  
these   drugs   was   ironclad.   PVCs   are   strongly   associated   with   sudden   death.   The   more   you   have,  
the   more   likely   you   are   to   die.   And   there   was   no   better   medicine   for   suppressing   PVCs   than  
flecainide.  How  could  it  not  save  lives?  Some  cardiologists  were  so  confident  in  the  drug  that  
they   would   not   let   their   patients   enter   studies   of   this   medication.   Such   studies   required   that  
some  patients  be  randomized  to  receive  a  placebo  instead  of  flecainide.  In  all  good  conscience,  
how  could  they  risk  having  their  patients  deprived  of  this  lifesaving  drug?  
 
Unfortunately,   flecainide   did   not   work.   In   1992,   a   large   study   called   the   Cardiac   Arrhythmia  
Suppression   Trial   (or   CAST   trial)   showed   that   flecainide,   as   well   as   a   similar   drug,   decreased  
PVCs  as  expected  but  also  increased  patients’  chance  of  dying.1  The  finding  was  devastating.  It  
changed   not   only   how   we   treat   heart-­‐attack   patients   but   how   we   evaluate   medical   therapies  
altogether.  CAST  taught  us  that  even  the  most  careful  reasoning  and  the  best  scientific  models  
do  not  guarantee  an  effective  clinical  treatment.  What  works  in  the  lab,  or  on  a  computer,  or  in  
the  head  of  the  smartest  researcher  does  not  always  work  in  a  patient.    
 
But  now,  more  than  20  years  later,  this  is  a  lesson  that  many  physicians  and  leading  researchers  
still  have  not  really  learned.  
 
Let’s   consider   another   medication,   this   one   for   high   blood   pressure   (hypertension).  
Hypertension   remains   one   of   the   most   common   ailments   in   America.   Although   most   people  
with  hypertension  are  entirely  without  symptoms,  there  is  no  doubt  that  hypertension  strongly  
predicts   an   increased   risk   of   stroke,   heart   disease,   and   death.   It   is   truly   a   silent   killer.   People  
with   naturally   lower   blood   pressure   live   longer,   and   people   who   use   certain   medications   to  
lower   their   blood   pressure   can   decrease   their   risk   of   death.   One   of   the   first   drugs   used   to   treat  
hypertension  was  atenolol.  Atenolol,  a  member  of  the  beta-­‐blocker  class  of  antihypertensives,  
dramatically   lowered   blood   pressure,   and   for   decades   it   achieved   the   rarefied   status   of   “trial  
standard,”  meaning  that  you  had  to  show  your  new  drug  was  as  good  as  atenolol  in  order  to  get  
it  on  the  market.  
 
In   2002   something   really   unsettling   (or   maybe   even   terrifying)   happened.   After   atenolol   had  
been  used  to  treat  hypertension  for  nearly  20  years,  the  results  of  the  LIFE  trial  were  published.  
This   trial   compared   atenolol   to   a   newer   drug,   losartan.   People   who   took   losartan   had   fewer  
strokes   and   lived   longer   than   those   that   took   atenolol.   At   first   glance   it   seemed   that   we   had  
come   up   with   a   better   antihypertensive;   this   seemed   like   an   example   of   replacement.  
Somewhat  surprising  was  that  both  drugs  lowered  blood  pressure  the  same  amount.  
 
The  question  was,  did  losartan  beat  atenolol  because  losartan  is  better  or  because  atenolol  is  
actually   ineffective?   The   tantalizing   finding   in   the   LIFE   study   was   extended   in   2004,   when   a  
pooled   analysis   of   all   people   who   took   atenolol   versus   sugar   pills   (placebos)   in   trials   showed  
that  atenolol  was  no  better  than  the  placebo.  Atenolol  did  lower  people’s  blood  pressure,  but  it  
did  not  decrease  people’s  risk  of  dying  or  of  having  a  heart  attack.  Let  that  sink  in.  A  drug  that  
was  widely  used,  accepted  as  the  standard  of  care,  had  made  millions  and  millions  of  dollars  for  
its  manufacturers,  and  had  made  high  blood  pressure  the  subject  of  dinnertime  conversation,  
did   not   make   you   live   a   single   day   longer.   A   recent   study   showed   that   metoprolol   (another  
beta-­‐blocker)   is   no   better   than   atenolol.   If   you   have   taken   a   beta-­‐blocker   for   high   blood  
pressure,   you   may   have   shaved   a   few   percentage   points   off   your   risk   of   stroke,   but   you   did   not  
extend  your  life.  In  the  world  of  blood-­‐pressure  medicines,  you  took  a  pill  that  does  not  work.2  
 
If   the   only   place   modern   medicine   erred   was   with   medications,   we   would   be   fortunate.   But  
there   are   also   medical   procedures,   some   used   for   decades   by   practitioners   who   have   reaped  
huge  financial  rewards,  that  have  been  shown  not  to  work.  
  
Let’s  examine  another  person.  Anthony  Baker  is  a  55-­‐year-­‐old  mechanic  who  thinks  of  himself  
as  active.  His  job  in  a  garage  requires  him  to  be  on  his  feet  most  of  the  day.  In  the  summer  he  
mows  his  lawn  every  week—with  a  push  mower,  no  less—and  in  the  winter  he  shovels  snow  
from  his  walk  after  storms.  Anthony  does  not  have  any  health  problems—well,  at  least  that  he  
knows  of—and  he  sees  his  doctor  yearly.  At  his  last  checkup,  he  was  told  that  he  was  in  fine  
health,  aside  from  the  yearly  advice  to  quit  smoking.  
 
Over  the  past  few  months,  however,  Anthony  has  noticed  that  when  he  exerts  himself,  he  feels  
an   ache   in   his   chest.   This   “pressure”   resolves   after   about   five   minutes   of   rest.   Anthony   calls   his  
doctor   to   tell   him   about   the   chest   tightness.   His   doc   is   concerned   that   Anthony   might   have  
developed  coronary-­‐artery  disease  and  that  his  symptoms  represent  angina.  
 
Angina  is  cardiac  pain  that  occurs  when  the  heart’s  demand  for  oxygen  outstrips  its  supply.  It  
often   happens   during   exertion   because   that   is   when   the   demand   for   oxygen   is   greatest.  
Anthony’s  doctor   orders   a   stress   test,  a   test   that   examines   the   heart’s   function   during   exercise.  
The  test  shows  that  Anthony  has  a  narrowing  in  his  circumflex  coronary  artery  (one  of  the  three  
coronary  arteries  that  supply  oxygenated  blood  to  the  heart).  Based  on  the  results  of  the  test,  
the   doctor   diagnoses   Anthony   with   coronary-­‐artery   disease,   tells   him   to   stop   smoking,   starts  
him   on   a   few   medications   (metoprolol,   a   beta-­‐blocker,   to   lessen   the   heart’s   oxygen   demand;  
atorvastatin,   a   cholesterol-­‐lowering   medication;   and   aspirin).   He   also   schedules   him   to   see   a  
cardiologist  the  following  week.  
 
When   Anthony   sees   the   cardiologist,   he   has   not   yet   started   his   medications   and   reports   one  
episode   of   chest   pain   that   occurred   while   he   was   walking   his   dog   over   the   weekend.   The  
cardiologist  schedules  Anthony  for  a  coronary  angiogram.  During  an  angiogram  dye  is  injected  
into   the   coronary   arteries   to   assess   their   patency—to   determine   whether   or   not   there   are  
blockages.   The   angiogram   shows   a   90   percent   blockage   of   his   circumflex   artery.   During   the  
procedure   the   cardiologist   places   a   stent,   a   metal   tube,   in   the   artery   and   the   90   percent  
blockage   disappears.   He   gives   Anthony   another   prescription,   this   one   for   clopidogrel,   a   blood  
thinner,  to  add  to  the  rest  of  his  medications.  
 
One   week   later   Anthony   sees   his   regular   doctor   and   is   pleased   to   say   that,   despite   having  
shoveled  six  inches  of  snow  the  previous  day,  he  is  feeling  the  best  he  has  felt  in  years.  He  has  
no  pain  and,  in  fact,  he  recognizes  that  he  was  probably  having  more  chest  symptoms  than  he  
was  aware  of  before  the  procedure.  
 
So,  was  Anthony  cured  by  the  cardiologist?  It  is  undeniable  that  Anthony  feels  better.3  You  may  
also  be  thinking  that  Anthony  narrowly  made  it—that  if  he  had  not  had  the  artery  opened,  he  
would   have   had   a   heart   attack.   That   because   of   the   stent,   Anthony   will   live   longer.   And   though  
the  procedure  cost  more  than  $10,000,  on  the  whole,  it  must  be  worth  it.  
 
But  what  if  we  told  you  that  Anthony  is  no  less  likely  to  have  a  heart  attack.  What  if  we  add  that  
Anthony   will   not   live   a   single   day   longer.   And,   perhaps,   most   concerning,   what   if   we   tell   you  
that  in  12  months,  Anthony’s  chest  pain  will  be  the  same,  even  though  he  had  the  stent.  Did  
Anthony  really  benefit?  
 
There  is  a  long  list  of  medical  procedures  whose  use  is  based  on  scientific  evidence  that  is,  at  
best,   suspect.   Vertebroplasty   is   another   great   example.   For   years,   doctors   struggled   to   treat  
people  (mostly  women)  who  suffered  osteoporotic  fractures  of  the  spine.  These  fractures  can  
lead  to  chronic  back  pain.  In  the  late  1990s,  a  couple  of  radiologists  had  a  brilliant  idea.  Why  
not  insert  a  needle  into  the  fractured  bone  and  inject  medical-­‐grade  cement.  The  theory  was  
that  the  cement  would  plump  up  the  bone,  the  nerves  would  get  some  extra  breathing  room  as  
the   fractured   bone   was   lifted   away,   and   the   pain   would   dissipate.   When   these   pioneers  
performed  the  procedure,  vertebroplasty,  on  a  few  dozen  patients,  they  were  amazed  with  the  
results.  Patients  immediately  felt  better.  Patients  and  doctors  were  convinced.  
 
In  the  very  early  2000s,  interested  parties  lobbied  Medicare  to  pay  for  vertebroplasty,  and  their  
request   was   granted.   In   a   few   years,   tens   of   thousands   of   people   were   having   the   procedure  
each  year.  By  the  end  of  the  decade,  vertebroplasty  was  a  billion-­‐dollar-­‐a-­‐year  industry.  There  
were  complications—rarely,  the  cement  would  go  somewhere  it  was  not  supposed  to—but  on  
the  whole,  this  procedure  seemed  like  a  real  advance.  In  2009  two  groups  of  brave  investigators  
put   vertebroplasty   to   the   true   test.   They   enrolled   200   patients.   Half   of   the   patients   got  
vertebroplasty;   the   other   half   were   taken   to   the   procedure   room   and   prepped   for   the  
procedure;   the   cement   was   opened,   so   patients   could   smell   it;   and   salt   water   was   injected.  
Both  groups  of  patients,  those  that  got  vertebroplasty  and  those  that  had  a  sham  procedure,  
had  identical  improvements.  Vertebroplasty,  as  it  turned  out,  is  no  better  than  a  placebo.  
 
This  outcome  alone  may  not  convince  everyone  that  vertebroplasty  is  a  bad  thing.  Who  cares  if  
the   effect   is   “in   your   head,”   as   long   as   it   works?   Some   with   expertise   in   the   placebo   effect  
would   hold   this   position.   Although   vertebroplasty   was   no   better   than   a   sham   procedure,   either  
may  very  well  be  better  than  doing  nothing.  In  this  book  we  examine  other  interventions  that  
claim  to  decrease  pain  but  perform  no  better  than  a  sham  procedure.  In  many  cases,  both  are  
better   than   doing   nothing.   What   does   this   mean?   There   is   something   about   the   acts   of   a  
medical   intervention—the   thoughtful   counsel   of   the   doctor,   the   skilled   support   of   the   nurses  
and   staff,   the   psychological   comfort   of   acting—that   collectively   make   us   feel   better.   The  
downside  is  only  that  the  procedure  is  costly,  may  involve  risk  (either  because  of  the  procedure  
itself   or   by   delaying   an   intervention   that   has   intrinsic   benefit),   and   involves   deception.   The  
question   then   becomes,   Can   we   reap   the   benefits   of   placebo   treatments   without   the  
deception,   the   risk,   and   the   cost?   In   chapter   2   we   discuss   the   placebo   effect   in   more   detail.   For  
now,  let  us  at  least  agree  that  the  money  spent  on  the  cement  used  in  vertebroplasty—often  
thousands  of  dollars  per  procedure—was  wasted.  Salt  water  would  have  sufficed.  
 
When  medical  reversal  involves  pills  or  procedures,  it  affects  patients  unlucky  enough  to  have  
an   illness   and   be   prescribed   a   faulty   therapy.   When   reversal   involves   prevention   efforts   and  
screening   campaigns,   millions   of   healthy   people   can   be   affected.   In   the   past   five   years,   two  
major   public   health   efforts,   breast   and   prostate   cancer   screening,   have,   to   a   large   extent,   been  
overturned.  Let’s  consider  mammography.  
  
In  2009  the  U.S.  Preventive  Services  Task  Force,  generally  considered  the  most  impartial  of  the  
hundreds   of   groups   that   produce   guidelines   for   physicians   to   follow   in   their   clinical   practice,  
changed   its   recommendation   on   whether   women   in   their   forties   should   undergo   screening  
mammography.   The   old   recommendation   was   to   do   a   mammogram   in   this   age   group   every  
other  year.  The  new  recommendation  is  not  to  do  it  at  all.  
 
This   reversal   made   a   lot   of   people   very   angry.   An   article   from   November   of   that   year   in   the  
Seattle   Times   had   the   provocative   title   “Mammogram   Mania:   Risking   Lives   or   Dollars?  
Physicians’  Community  Speaks  Up,  against  New  Recommendations.”  Radiologists,  who  had  the  
most  to  lose  (at  least  financially),  protested  the  loudest.  They  accused  the  task  force  of  trying  to  
save  money  instead  of  lives.  
 
Was   any   of   this   a   fair   characterization?   Were   the   changes   to   the   guidelines   a   shock?   Not   to  
those  who  had  followed  the  medical  literature.  The  truth  is  that  the  guideline  change  was  three  
years  in  the  making.  
 
In   2006   a   huge   study   on   exactly   the   question   of   mammographic   screening   for   40-­‐somethings  
was   published   in   Lancet.   More   than   160,000   women   in   their   forties   were   randomized   to  
mammography   screening,   or   not,   and   they   were   followed   for,   on   average,   10.7   years.   It   is  
worth   pausing   to   consider   just   how   impressive   a   feat   this   study   was.   The   study   authors   got  
more  than  100,000  women  to  participate  in  a  study  of  a  medical  treatment  that  required  yearly  
visits  for  over  a  decade.  For  reasons  we  discuss  later,  as  far  as  evidence  in  medicine  goes,  this  is  
about  as  good  as  it  gets.  
 
The   authors   of   the   trial   found   that   there   were   fewer   deaths   from   breast   cancer   among   the  
women   who   received   mammograms,   compared   to   those   who   did   not,   but   this   difference   did  
not   reach   statistical   significance.   In   other   words,   the   difference   was   so   small   that   we   cannot   be  
sure  the  difference  was  not  due  to  chance  alone.  What  this  means  is  that  a  few  women  who  got  
mammograms   died   of   breast   cancer   and   a   few   women   who   did   not   get   mammograms   also  
died.  Neither  group  had  more  women  who  died  of  breast  cancer.  
 
If   you   look   at   the   study   for   what   is   truly   important,   dying,   from   any   cause,   not   just   breast  
cancer,  you  find  that  the  death  rate  was  identical  in  the  two  groups.  Therefore,  we  should  not  
screen  these  women  with  mammography,  because  in  the  history  of  medicine  no  one  has  ever  
shown  that  doing  so  saves  lives.  In  aggregate.  On  the  whole.  Regardless  of  the  financial  impact.  
 
Put  another  way,  what  if  all  the  women  in  their  forties  of  Washington,  D.C.,  got  mammography  
and   all   the   women   in   their   forties   of   Chicago   did   not.   Well,   nearly   all   the   women   in   this   age  
group  in  both  cities  would  be  alive  at  50,  and  a  few  unlucky  ones  in  each  city  would  have  died  
of   breast   cancer.   And   all   the   statisticians   at   the   NIH   in   Bethesda,   Maryland,   and   at   the  
University  of  Chicago  would  be  hard-­‐pressed  to  say  for  certain  which  city  got  mammograms  if  
you  did  not  tell  them.  If  the  statisticians  looked  at  death  rates  for  40-­‐something  women  in  both  
cities,  they  would  be  identical.  
  
What   led   to   all   the   hand-­‐wringing   about   the   guideline   changes   was   not   the   reversal   of  
screening  mammography;  it  was  that  doctors  had  recommended  mammography  for  women  in  
their   forties   for   all   those   years.   A   costly,   ambitious   medical   program   was   conducted   on   a  
national  level  for  years  before  we  had  data  that  said  that  it  worked.  What  we  know  for  sure  is  
that  because  of  false-­‐positive  mammograms  and  a  phenomenon  called  “overdiagnosis”  (more  
in   chapter   4)   the   program   gave   many   women   a   cancer   scare,   a   needle   biopsy,   radiation  
exposure,  surgery,  or  worse.  
 
We  believe  that  reversal  is  the  most  important  problem  in  medicine  today.  When  we  doctors  
flip-­‐flop  on  our  advice  to  patients,  it  usually  is  not  because  the  treatments  stopped  working.  It  
usually  is  not  because  someone  discovered  a  harm  no  one  had  previously  noticed.  It  is  usually  
because  the  practice  never  worked—we  were  wrong  all  along.  We  promoted  it  before  we  had  
properly   studied   it.   We   knew   it   had   some   harms,   sure,   but   we   never   thought   it   lacked   benefits.  
This  problem  underlies  people’s  distrust  of  the  medical  establishment  and  is  a  very  important  
reason  that  health-­‐care  costs  are  soaring  without  any  improvement  in  people’s  health.  
 
Now  reversal  is  not  the  only  problem  in  medicine,  but  it  is  a  common  channel  through  which  
many   of   the   problems   run.   Yes,   pharmaceutical   companies   manipulate   data   to   get   drugs   and  
devices  approved  and  used.  Yes,  physicians  overtest  and  overtreat  for  fear  of  malpractice  suits  
and  because  of  the  financial  benefits  that  this  behavior  brings.  Yes,  hospitals  and  clinics  market  
tests   and   treatments   that   offer   little   benefit   because   they   believe   they   will   attract   well-­‐insured  
patients.  However,  if  we  find  a  solution  to  the  problem  of  medical  reversal,  a  way  of  preventing  
the  use  of  therapies  that  do  not  work,  many  of  these  other  problems  will  cease  to  exist.  
 
What  Is  to  Come  
 
In  this  first  part  of  the  book,  we  describe  many  reversals  and  estimate  the  frequency  of  medical  
reversal.  We  describe  the  harms  of  our  broken  system.  Many  are  obvious,  but  there  are  some  
surprising   harms   of   medical   reversal.   The   repercussions   may   go   further   than   you   imagine.   In  
part   2,   we   provide   you   with   an   understanding   of   evidence-­‐based   medicine,   an   understanding  
that  will  prepare  you  for  the  remainder  of  the  book  and  allow  you  to  question  the  next  report  
about   a   “revolutionary   new   treatment.”   Part   3   focuses   on   the   causes   of   medical   reversal.  
Money,  of  course,  is  involved,  but  in  this  case  it  is  not  the  entire  problem.  The  seeds  of  reversal  
are  sown  at  the  start  of  medical  school  and  are  nourished  throughout  the  medical  industry.  
 
In   part   4   we   propose   solutions   for   the   problem   of   medical   reversal.   Our   solutions   are   not   a  
checklist  and  a  vague  call  for  you  to  be  prepared  when  you  see  the  doctor.  We  show  just  where  
changes   can   be   made   to   medical   education,   medical   research,   and   the   process   by   which  
treatments   are   approved,   so   as   to   eliminate   ineffective   therapies.   We   also   offer   suggestions  
about   how   individuals   can   become   better   consumers   of   health   care,   how   to   be   people   who   are  
unlikely  to  want,  or  be  given,  treatments  that  do  not  work.  Because  fundamentalism,  in  all  its  
forms,   is   dangerous,   we   end   by   discussing   those   times   when   it   is   not   necessary   to   demand   that  
practices  be  based  on  ironclad  data.  
  
We   hope   that   through   reading   this   book   you   will   understand   the   scope   of   the   problem   of  
medical   reversal—how   reversal   affects   every   part   of   the   doctor-­‐patient   relationship.   This  
understanding   may   very   well   change   your   opinion   of   “feeling   better.”   It   will   certainly   affect  
what   you   do   when   you   feel   well,   in   order   to   stay   healthy,   and   what   you   do   when   you   feel  
unwell.  
 
NOTES  
1
As  you  will  see  throughout  this  book,  one  of  the  great  sports  in  academic  medicine  is  to  choose  
a  name  for  your  trial  that  can  be  reduced  to  a  catchy,  memorable  acronym.  In  our  careers  we  
have   seen   not   only   CAST   and   LIFE,   discussed   in   this   chapter,   but   also   NICE-­‐SUGAR,   CLEOPATRA,  
and  our  favorite,  CABG  Patch.  
 
2
  We   are   only   talking   about   hypertension.   If   you   have   heart   failure,   you   really   need   your  
carvedilol  or  metoprolol  succinate!  
 
3  
In   later  chapters   we’ll   learn   about   other   procedures   that   make   people   feel   better   but   provide  
no  real  benefit.  These  examples  will  make  you  reconsider  the  whole  “feeling  better”  issue.  
 
EXCERPTED   from   “ENDING   MEDICAL   REVERSAL,”   by   VINAYAK   PRASAD   and   ADAM   CIFU.  
Published   by   Johns   Hopkins   University   Press  ©   2015.     Reprinted   by   permission   of   the  
publisher.