Issuu on Google+

Loving  the  Older  People  in  1mes  of  Cholera     Preliminary  findings  from  a  2008-­‐12  study  to  analyse  care  and  outcomes     for  cholera  pa1ents  treated  by  MSF-­‐OCA  in  Haï1  and  Zimbabwe   Jonathan  Kwok  

Todd  Swarthout  

j.kwok1@uni.bsms.ac.uk  

toddswarth@yahoo.com  

Brighton  &  Sussex  Medical  School  

1  

Pascale  Fritsch  

MSF-­‐OCA  

2  

Conclusions  

3  

Background  

–  Officially  reported  cases  represent  5-­‐10%  of  the  actual  total   1.2-­‐4.3  million  cases;  28,000  to  142,000  deaths  (1)   –  There  is  a  lack  of  published  material  invesHgaHng  the   cholera  epidemiology  of  old  people,  and  their  specific   outcome  risk  factors  

–  Age  is  significantly  associated  with  poor  dehydraHon   status  on  arrival  (2-­‐3x  odds  for  >60yo  vs  adults)  

•  Dehydra1on  status  at  presenta1on  is  strongly   associated  with  mortality   –  58x  odds  of  death  for  Severe  versus  Mild  

•  First  Hai1an  case  of  cholera  in  more  than  a  century,   confirmed  20th  October  2010  

•  Older  pa1ents  are  not  more  likely  to  receive   ac1ve  treatment    

•  More  than  half  a  million  people  affected  by  cholera   in  Haï1  since  (to  20th  April  2012)   –  536,943  cases  noHfied   –  288,839  paHents  hospitalised   –  7,112  deaths  (1.3%  CFR)  (2)  

–  Age  is  not  correlated  with  acHve  treatment     (either  ORS,  iv  Ringers,  or  a  combinaHon  of  both)  

•  Older  pa1ents  are  more  likely  to  die  

–  Belief  that  older  paHents  have  been  more  likely  to  die   –  Desire  to  idenHfy  potenHal  protocol  improvements    

Older  persons  present   with  worse  dehydraHon   DehydraHon  status  of  older     paHents  on  arrival  at  CTC  

250   C  (severe,  iv)   B  (moderate,  ORS)   A  (mild)  

50  

1,000  

%  Old  (>60  yo)   10%  

Jul'11  

Nov'11  

Jun'11  

Oct'11  

Sep'11  

Aug'11  

May'11  

Jan'11  

Apr'11  

Feb'11  

Dec'10  

Nov'10  

Mar'11  

Jul'11  

Nov'11  

Jun'11  

Oct'11  

Sep'11  

Aug'11  

May'11  

Jan'11  

Apr'11  

Feb'11  

Dec'10  

15-­‐24yo   Children   Infants  

0  

•  Poor  Dehydra1on  Status  is  associated  with  Age  

No v'1 De 0   c'1 Ja 0   n'1 Fe 1   b'1 M 1   ar' 1 Ap 1   r'1 M 1   ay '1 Ju 1   n'1 1 Ju   l'1 Au 1   g'1 Se 1   p'1 Oc 1   t'1 No 1   v'1 1  

–  For  Moderate  dehydraHon  vs  Mild,  at  presentaHon  to  CTC  

350   300   250   200   150  

>80yo  

100  

60-­‐79yo  

•  1.4x  odds  raHo  for  60-­‐79yo  (versus  adult,  p<0.001,  95%  CI  1.23-­‐1.67)     1.9x  odds  raHo  for  >80yo  (p<0.001,  95%  CI  1.31-­‐2.83)   –  For  Severe  dehydra1on  vs  Mild   •  1.9x  odds  ra1o  for  60-­‐79yo  (versus  adult,  p<0.001,  95%  CI  1.64-­‐2.20),     2.9x  odds  ra1o  for  >80yo  (p<0.001,  95%  CI  2.01-­‐4.18)   •  Males  enjoy  independent  protecHve  effect  of  0.91  odds  raHo   (p=0.002,  95%  CI  0.85-­‐0.97)  

•  For  Moderate  dehydra1on  versus  Mild  

Age  is  weakly  correlated   with  acHve  treatment   •  Either  ORS,  iv  Ringers  or  both  during   inpa1ent  stay  

–  InvesHgated  by  creaHng  a  new  variable  for   treatment,  to  run  bivariate  correlaHon  

50  

No v'1 De 0   c'1 0 Ja   n'1 Fe 1   b'1 M 1   ar' 1 Ap 1   r'1 M 1   ay '1 Ju 1   n'1 1 Ju   l'1 Au 1   g'1 Se 1   p'1 Oc 1   t'1 No 1   v'1 1  

0  

•  Very  weak  correlaHon  between   Age  and  Time  to  presentaHon  

•  Ethical  approvals  were  not  required  for  this   retrospecHve  secondary  data  analysis  

–  The  project  is  staffed  with  full  Hme  research  from   Brighton  &  Sussex  Medical  School,  part  Hme  voluntary   support  from  Todd  Swarthout,  part  Hme  technical   advice  from  Help  Age  InternaHonal  and  MSF-­‐OCA,  and   administraHve  support  from  Kathrin  Go3wald  

(4)  Pan  American  Health  OrganizaHon.  Atlas  of  cholera   outbreak  in  La  Hispaniola,  2010-­‐12.  Washington  DC;   2012.  Available  at   h3p://new.paho.org/hq/images/Atlas_IHR/ CholeraHispaniola/atlas.html  

%  Adult  (25-­‐59  yo)  

20%  

0%  

Nov'10  

0  

Mar'11  

2,000  

25-­‐59yo  

Map  data  ©  OpenStreetMap  contributors,  CC  BY-­‐SA  

–  57.8x  odds  of  death  (p<0.001,  95%  CI  26-­‐128)  

30%  

200  

100  

60-­‐79yo  

CumulaHve  cholera  deaths  (4)  

•  For  Severe  dehydra1on  versus  Mild  

40%  

150  

>80yo  

–  Older  pa1ents  present  later   –  Older  pa1ents  present  worse   –  Older  pa1ents  suffer  worse  outcomes  

–  10.2x  odds  of  death  (p<0.001,  95%  CI  4.5-­‐23.3)  

50%  

Age  (years)   Sex  (M  or  F)   District  and  ward   Days  since  first  symptoms   DehydraHon  status  at  presentaHon   Treatments  (ORS  or  iv  Ringers  in  litres)   Outcome  (Cured,  Transferred,  Lost  to  Follow  Up,  Died)   Dates  of  admission  and  discharge  

•  Data  were  cleaned  up  and  charted  in  Microsoq   Excel  (MS  Excel  for  Mac  2008  v12.1.5)  and  then   exported  to  IBM  SPSS  StaHsHcs  (v19)  for  staHsHcal   invesHgaHon  of  the  following  hypotheses  

Dominican   Republic  

MSF-­‐OCA   MSF-­‐OCG   MSF-­‐OCBA  

DehydraHon  is  strongly   associated  with  mortality  

ProporHon  of  older  (>60yo)  versus  adults   presenHng  with  severe  dehydraHon  

350   300  

3,000  

MSF-­‐OCG  

MSF-­‐OCBA   MSF-­‐OCB  

•  MSF  doctors  and  nurses  have  become  concerned   about  outcomes  for  their  older  pa1ents  

Results  

4,000  

MSF-­‐OCP  

HaïH   MSF-­‐OCA  

–  La  Saline,  Delmas,  Carrefour  (urban  districts  N,  E  &  W  of   Port  au  Prince)   –  Barradères,  Bonne  Fin,  Plaisance  du  Sud  (rural  southwest)   –  Treated  >26,881  paHents  (3)  

(1)  Ali  M  et  al.  The  global  burden  of  cholera.  Bull  World  Health   Organ  2012;  90:  209-­‐218A   (2)  Ministère  de  la  Santé  Publique  et  de  la  PopulaHon,  HaïH.   Rapports  journaliers  du  MSPP  sur  l'évoluHon  du  choléra  en   HaïH.  Port  au  Prince;  2012.  Available  at   h3p://www.mspp.gouv.ht/site/downloads/Rapport %20journalier%20MSPP%20du%2023%20avril%202012.pdf   (3)  Médicins  Sans  FronHères.  MSF-­‐OCA  linelist  data  to  April  12th,   2012  

5,000  

Cuba  

•  MSF-­‐OCA  has  operated  cholera  treatment  centres   at  six  loca1ons  

–  Age  is  significantly  associated  with  mortality  outcome,   controlled  for  sex,  dehydraHon  status  at  presentaHon,   duraHon  and  type  of  treatment,  experience  and   acHvity  levels  of  the  CTC  

6,000  

Methodology   –  –  –  –  –  –  –  – 

•  WHO  es1mates  2.8  million  cases  of  cholera   annually  worldwide    

•  Older  pa1ents  present  with  worse  dehydra1on  

Number  of  paHents  admi3ed  to  CTC  by  age  group    

Melanie  Newport  

Brighton  &  Sussex  Medical  School  

•  MSF  clinical  teams  collected  pa1ent  data  using  an   internal  “linelist”  data  tool  

–  Age  is  only  weakly  correlated  with  delays  in  Hme  from   first  symptoms  to  presentaHon  at  a  CTC  

Similar  presentaHon  pa3ern,   over  Hme,  for  all  age  groups  

InsHtute  of  Postgraduate  Medicine  

pfritsch@helpage.org  

•  Older  pa1ents  present  no  later  than  others  

4  

Abid  Raza  

Help  Age  InternaHonal  

–  However,    we    are    cauHous    of    the    accuracy    of     the    ORS    and    iv    Ringers    observaHons    in    our         linelist    data    tool      

•  >20%  of  pa1ents  presen1ng  with  Mild   dehydra1on  go  on  to  receive  iv  fluids  

–  Half  of  the  Moderate,  and  four  fiqhs  of  the   Severe  dehydraHon  paHents  received  both  ORS   and  iv  Ringers   –  We  believe  paHents  receiving  only  iv  Ringers   died  early  before  switching  to  oral  soluHons  

Age  is  associated  with   greater  risk  of  mortality   •  Pa1ents  who  died  have  a  higher  mean  age  (38  yo,   SD  29)  versus  those  who  recovered  (23yo,  SD  18)   p<0.001  

Regression  Model   •  Binary  logisHc  regression  on  Outcome   –  Controlled  for  sex,  age  group,  CTC  experience   (admission  month),  duraHon  of  care,  dehydraHon   status  on  presentaHon,  type  of  treatment,  and   acHvity  levels  of  the  CTC  (paHents  per  week)  

•  Age  60-­‐79yo  associated  with  3.97x  odds  of   death  (95%  CI  2.29-­‐6.90,  p<0.001)   –  Age  >80yo  associated  with  11.4x  odds  for  death   95%  CI  5.13-­‐25.44,  p<0.001  

•  Male  sex  associated  with  0.62  odds  protec1ve   effect  (95%  CI  0.41-­‐0.94,  p=0.025)   –  Longer  dura1on  of  care  was  associated  with  a   significant  protecHve  effect  (odds  0.54,  95%  CI   0.46-­‐0.62,  p<0.001)   –  “Busy”  ness  or  ac1vity  levels  of  the  CTC  had  no   significant  associaHon  with  outcome  (odds  0.999)  

5  

Next  steps   •  Sta1s1cal  inves1ga1on  of  addi1onal  pa1ent   data  from  con1nuing  epidemic  in  Haï1   –  Latest  updates  (Nov’11  to  Apr’12)  

•  A3empt  to  idenHfy  and  control  for  addiHonal   confounders,  eg.   –  Socioeconomic  status   –  NutriHonal  status   –  Co-­‐morbidiHes  

•  Deeper  analysis  of  idenHfied  subgroups,  eg.  

–  Subset  of  severely  dehydrated  paHents  appear  to   die  quickly,  before  ORS  is  started   •  InvesHgaHon  of  paHent  characterisHcs?   Treatment  profiles?  Time  pa3ern  to  mortality?  

•  Examine  data  from  MSF-­‐OCA  mission  to   Zimbabwe   –  Comparison  with  HaïH   •  Poten1al  discussions  with  previous  and  current   MSF  field  teams   •  PotenHal  contribuHon  to  ongoing  MSF  review  of   Cholera  Guidelines   –  AnHcipated  Dec’12  

 

 With  thanks  to  Leslie  Shanks  and  Kathrin  Go3wald,  Emanuele  Sozzi,  MSF’s  country  teams,  Muna  Aljawad,  the  Royal  Sussex  County  Hospital  and  the  people  of  HaïH  and  Zimbabwe


Loving the Older People in Times of Cholera