Hi Ryan,

Very well reasoned and nicely said.  I'm glad you made these points.  The best 
thing about ProCESS is that we will have a long and open discussion about where 
we stand.

For the sake of pro/con, let me make a rejoinder.  :-)

1.  Jones et al. reflected a trial that did not tell us anything much about 
"usual care."  The comparison was head to head EGDT using ScvO2 versus EGDT 
using lactate clearance as a substitute for ScvO2.  Everyone had central lines 
in both arms.  Everyone was resuscitated using CVP as a guide.

2.  Jones et al. and ProCESS reflect care at academic centers.  It may well be 
that the nation's elite centers that have enough interest to sign up 
investigators who care sufficiently to invest years of time into studying 
sepsis, residents and fellows monitoring resuscitation parameters around the 
clock, and well established standards and protocols achieve low mortality rates 
and high compliance rates with SSC guidelines.

3.  Neither Jones nor ProCESS show low compliance with SSC guidelines results 
in great outcomes.  On the contrary, the control arm in ProCESS approached 60% 
compliance with central access, and Jones had 100% compliance with current SSC 
bundles in both arms.

5.  In ProCESS I can't tell you how to be like the control arm because frankly 
we just don't know how care was provided.  A far as we can tell 60% of the 
control patients might have had the same therapy as the intervention.  What 
does this mean?  Nobody can say.

4.  US hospitals just don't look like this care in thousands of hospitals 
across the country.  Compliance with SSC bundles in the US is at best in the 
30-40% range in hospitals that have been actively working on adherence for 2 
years (Levy 2010).  Mortality in that group if community hospitals is in the 
28% range for severe sepsis and septic shock combined (shock is higher).  The 
largest current database if US hospitals that compares mortality of all acute 
care discharges, Medicare and commercial coverage and no coverage combined, is 
the Midas Comparative Database.  This database has 750 US hospitals, roughly 
15% of all hospitals in the country.  As of this month, the P50 shock mortality 
in this swath of hospitals is 32.5%.

 5.  I don't pretend to have all the answers yet.  ProCESS is probably showing 
us that really awesome care short of the SSC bundles, reliably executed is 
pretty darn good.  The public health dilemma and true danger of ProCESS is that 
the huge majority -- thousands of hospitals (there are around 5000 US 
hospitals) -- don't look like ProCESS.

6.  If you can honestly say your own *shock* mortality is 18% -- ask the 
quality department at your system, they know -- and 60% of your patients get 
lines, and you have 24/7 intensivist coverage following these patients, well 
then I guess you look like ProCESS.  The 26 hospitals in my system have none of 
that, each has shock mortality rates greater than 23% and some approaching 34%. 
 One has 24/7 intensivist coverage, central line insertion rates are about 20%.

7.  So, I'm not comfortable saying do whatever you like.  I spend a lot of 
energy visiting hospitals across the country.  For the public health we cannot 
let the message be relax your standards even below where they are now.  I 
continue to say if you want to improve in your current state, aim high, set 
bold targets.  The day you look like ProCESS you can pause and reflect on what 
to do next.

:-)

I think that's the positive message my colleagues and I need to get out there.  
The debate is good to have because it's a chance for us to raise awareness and 
focus people on what they can actually say about their own outcomes thus far.

I really appreciate your intelligent and well made remarks, and I know this 
discussion is far from over!  I've said the best day will come when a neutral 
objective 3rd party can make this statement.  I'm invested as you know.

Sean

Sean R. Townsend, M.D.
Vice President of Quality & Safety
California Pacific Medical Center
2330 Clay Street, #301<x-apple-data-detectors://0/0>
San Francisco, CA  94115<x-apple-data-detectors://0/0>
email [email protected]<mailto:[email protected]>
office (415) 600-5770<tel:(415)%20600-5770>
fax (415) 600-1541<tel:(415)%20600-1541>

On Mar 29, 2014, at 8:38 AM, "Ryan Arnold" 
<[email protected]<mailto:[email protected]>> wrote:

I will suggest a counter to the observation that the PROCESS trial does not 
represent most US hospitals.

We have 2 US-based studies now with ED-centered randomization and similar 
mortality rates. The smaller scale multi-centered RCT (Jones) and the larger 
scale multi-centered RCT (PROCESS) have near identical overall mortality rates 
to each other (Jones: overall mortality 59/300 = 19.7%; PROCESS: overall 
mortality 259/1341 = 19.3%). Each of these used identical inclusion criteria to 
the original EGDT study and identical inclusion criteria to current SSC 
guidelines.

The other 2 studies mentioned in the same NEJM release did have higher 
mortality rates. However, they do not represent the same population as 
addressed in the Emergency Department-centered PROCESS study. The albumin in 
sepsis study (Caironi et al) is an Italian ICU based patient population with 
“severe sepsis or septic shock”, very different population than PROCESS or 
Rivers. The high vs low MAP target study (Asfar et al) is a French ICU-based 
patient population with vasopressor dependent septic shock, again, not the same 
population as the PROCESS or Rivers study.

Even if ARISE and PROMISE reflect a different trend or outcome, from a US 
hospital perspective, PROCESS is as applicable if not more so for a US-based 
hospital system.

From PROCESS, it would seem that there is change supported in current care:
Mandated central line placement for all patients is unsupported (i.e. does not 
improve mortality).
Mandated CVP assessment and ScvO2 assessment for all patients is unsupported 
(i.e. does not improve mortality).

There is a high likelihood of a Hawthorne effect in the usual care arm, where 
the treatment team may have been more attentive and aggressive in resuscitation 
and monitoring knowing they were being “watched", which is the bottom-line 
message we all advocate for and seems to be important across the board: be 
attentive / be aggressive.  The usual care care in Process was great care as 
was pointed out, timely fluids and antibiotics, central lines when indicated, 
and all this continues to support early recognition, aggressive up-front 
fluids, and early antibiotics. If central access is necessary to treat fluid 
refractory hypotension with vasopressors, I think a case can be made for 
assessment of both CVP and ScvO2 as they are valuable measures if low, but much 
less helpful, if at all, when normal.

I think PROCESS will translate to an overall improvement in sepsis recognition 
and care, where clinicians are more likely to participate in sepsis 
resuscitation now that they are supported with a less-invasive alternative than 
is currently suggested.

Ryan


Ryan Arnold, MD
Department of Emergency Medicine
Christiana Care Health System
Newark, DE


_______________________________________________
Sepsisgroups mailing list
[email protected]<mailto:[email protected]>
http://lists.sepsisgroups.org/listinfo.cgi/sepsisgroups-sepsisgroups.org
_______________________________________________
Sepsisgroups mailing list
[email protected]
http://lists.sepsisgroups.org/listinfo.cgi/sepsisgroups-sepsisgroups.org

Reply via email to