Transcript Dia 1

The Radboud Hospital affair
Facts against blame free reporting and plea for
public reporting of medical errors.
Wall of silence caused unnecessary victims of medical errors, for years.
Contents:
Huub Wollersheim, M.D. Ph.D.
Radboud Medical Centre, Nijmegen
Title and :
Publication
Sophie Hankes, LL.M.
SIN-NL, Netherlands
Excellence Models and Safety in Health Care
EFQM conference Friday, 12 October 2007
Domus Medica, Utrecht, Netherlands.
The collapse and redesign of the cardio surgical department of the
Radboud Academic Medical Centre: A case study
Speaker: Huub Wollersheim, MD, PhD,
Radboud Medical Centre, Nijmegen, Netherlands.
In 2005 major problems in the cardio surgical treatment chain of the
Academic Medical Centre in Nijmegen were signalled. The average mortality
and complication rate exceeded the national standard. An auditing
commission was installed and confirmed the low performance of the
treatment chain. The problems had a great impact on the hospital. The
cardio surgical department was closed. The staff and the board of the
hospital resigned. The image of the hospital was threatened. An
improvement project was started and carried out to redesign the treatment
chain. Two years later the benchmark shows that the mortality and
complication rate is below the Dutch average.
Radboud case: Dr Hub Wollersheim, internist.
Centre for Quality of Care Research,
Radboud University Medical Centre Nijmegen
The Radboud case:
The fall and rise of cardiac surgery.
Bristol revisited
Do hospitals have memory?
Do hospitals have learning capability?
Radboud case

The facts

The stories

The questions

The resemblance with the Bristol case

What can we learn?
The Radboud case: the facts
On September the ninth in 2005 a multi-disciplinary management
meeting was organised of health care professionals that work in
the Heart Centre at the Radboud University Hospitals in Nijmegen,
the Netherlands.
The main topic on the agenda was to discuss a safety improvement
plan for cardiothoracic surgery patients.
The problematic data that were presented showed a more than
doubled mortality rate.
The day after the meeting one of the recently appointed professors
mailed the data and remarks regarding his concern to his coworkers. His intention was to create a sense of urgency.
The Radboud case: the facts
Somebody anonymously forwarded the E-mail to the press and the
Health Care Inspectorate.
The data were first denied and subsequently ascribed to a high risk
patient case mix.
The unintended messenger was suspended by the board of directors.
The Radboud case was born.
An inadequate care process; report April 2006; www.IGZ.nl




All who know history recognize this classical, yet disastrous
reaction.
The positive consequence was that immediately an indepth
analysis was started.
An internal and external committee started to analyse the mortility
and complication data and the records of all patients died.
They interviewed around 50 persons that (had) worked within this
patients care process.
The conclusions of the inspectorate regarding the analysis were
devastating.
Conclusions of the Inspectorate
The numer of deaths and complications were importantly increased,
due to a dysfunctioning process of care delivery.
Instead of a more, a less complex case load was found.
Failing leadership and multiprofessional cooporation became apparant.
Protocols were absent.
The process and outcome of care were not evaluated systematically.
Janssen D. The development of a CABG database, a never
ending story. A risk analysis of morbidity and mortality in CABG
surgery. Thesis. Pasmans BV Den Haag. Nijmegen; 2006.




Unjustified is the conclusion that the outcome of care was not
evaluated systematically: the conclusion of the Inspectorate
should be: It did not result in effective preventive action.
In 1987 the department started to implement a descriptive
database regarding clinical outcomes of CABG patients (“Corrad”
data base, referring to CORonary StRADboud), that gradually grew
into a more predictive outcome set.
Besides data from isolated myocardial revascularisations also data
from valve surgery and other adult heart procedures were added.
In the past years the data regarding morbidity and mortality were
more refined and corrected for risk factors in several statistical
ways.
Gogbashian A, et al. EuroSCORE: a systematic review of
international performance. Eur J Cardio-thoracic Surg 2004; 25:
965-700.



For international comparisons a European based data system was
introduced: the ‘EuroSCORE’.
This includes a uniform stratification of 17 deaths predicting risk
factors.
Risk factors: age, sex, previous heart surgery, pulmonary disease,
extra cardiac arteriopathy, neurologic dysfunction, serum
creatinine concentration, left ventricular function, unstable angina
or recent myocardial infarction, urgency and a critical preoperative
state such as mechanical ventilation, inotropic support, intra-aortic
balloon counterpulsation or acute renal faillure.
Feedback of EuroSCORE data




These local data were analyzed each year and at least once a year
shown to all thoracic surgeons in the Department’s annual report.
Eversince 2000 the mortality and complication data seemed higher
than those of other European centres. A peak (6,7% mortality: 2-3
times higher than other comparable European centres) was
observed in 2004.
Although some members of the team had expressed there worries,
no concrete actions were taken by the management or by the
surgeons themselves.
The data were looked upon, found interesting and the next day
work continued as usual.
Feeback of EuroSCORE data: question 1


One of the most fascinating questions is, why were these data not
regarded as sufficiently urgent to start to improve patient safety
upon knowledge?
Sit back and try to answer this question, discuss it with your
neighbour and be prepared to present your analyses.
The local situation




Radboud University Nijmegen Medical Centre.
One of 8 University hospitals in the Netherlands.
900 beds; 8.000 employees; 2.500 students; budget 500 million
Euro’s.
Organised in 4 levels: Board of directors; Cluster; Business unit
(Department and staff units and nursing and out-patient wards);
Work station.
Question 2: Why are hospitals so inadequately organised?
Centres, like the Heart Centre, are alliances of several business units:




Thoracic and cardiac surgery
Cardiology and pediatric cardiology
Pulmonary disease
Also anesthesiology, intensive care, the Clinical Perfusion Unit and
Operating Theatres are involved.
These business units are part of 3 clusters!
A department without leadership and internal cohesion






Department of Cardiothoracic Surgery.
Head: Prof. Dr. Rene Brouwer; 7 other thoracic surgeons.
Around 800 surgical procedures in adults and 250 in children each
year.
In adults: CABG: around 350; Aortic/mitral valve surgery: 140/40
(alone or in combination with CABG).
80% of patients are referred by other hospitals.
The Department is part of the Heart Centre (6 departments; 150
caregivers).
The whistleblower; Johan Damen
Johan Damen; Professor of Cardiac Anesthesiology;
appointed June 2005:



Unintended whistleblower due to an E-mail leaked to a local newspaper
and the Health Care Inspectorate by a still unknown colleague.
Tried to create a sense of urgency among co-workers; the day after a
meeting (the first multi disciplinary group discussion about the high
mortality and morbidity rates!) during which outcome data for cardiac
surgery data collected for the NICE (National Intensive Care
Evaluation) system were presented.
The key message of the E-mail:
‘An unacceptable high level of deaths in patients having heart valve
surgery.’
‘I would not allow the centre to operate on me.’
The furious reaction
Of the hospital board, of the physicians and of his colleagues:






The message was denied (not true; later: due to case mix)
Johan Damen was banished from the staff, gained wide disrespect and
was forced to stop patient care.
He struggled and became ill.
In the mean time he has been appointed as professor of
Anesthesiology and Perioperative Patient Safety, but he still has to be
rehabilitated by his colleagues.
He considers his misery now as a necessary side effect of the very
needed restoration process he started.
Note of SIN-NL: we had one meeting with Prof. Damen and
unfortunately he had not the courage to develop improvements for the
victims of medical errors.
If surveillant bodies fall to protect patients, should not every
honest doctor become a whistleblower?
Questions:
1. Does the Dutch way of surveillance (board of trustees,
Inspectorate, visitation, NIAZ/INK accreditation/certification) fail?
Obviously.
2. Why are the whistleblowers blamed and shamed?
People that save lives should be honoured, respected and legally protected.
The reaction of the hospital








Denial: nothing is wrong (September 2005).
The data are due to an unfavorable case mix: for example 22%
diabetes patients (September 2005).
There could be something wrong:we are going to analyse it,sept05
There is something wrong: we are going to restore it (Oct. 2005);
protocols, clinical pathways.
Rene Brouwer (Head of the Department) resigns (March 2006).
Stop of all cardiothoracic operations (April 2006 after an order
from the Inspectorate).
Resignation of the chairman of the hospital board or directors
(April 2006).
Resignation of the other two members of the hospital board or
directors (May 2006).
The reaction of the hospital




Resignation of the medical board (May 2006).
Debate in the House of Parliament (23 May 2006).
New temporary hospital board (June 2006) and a new head of the
Cardiothoracic Department: prof. dr. Leon Eijsmans: start of the
restoration process (June 2006).
3 of 7 left surgeons resign (June 2006).
Many more questions

Signs of failing surveillance (the Supervisory board, the
Inspectorate, the visitation by the Dutch Cardiothoracic Society,
the NIAZ accreditation in 2002 and 2006)
Did they not know?
Did they not react?

Is it so difficult to say sorry when you fail?

Why not show compassion with or make excuses to the patients,
their relatives or other damaged people?
A focus on the outcome data: mortality rates of CBAG

European norm: around 4%
mean in US 2%; mean in the Netherlands 2,7%

From 2000 it was above the norm with a peak in 2004: 6.7%

From 203-2005: 83 of 1725 patients died.
This was 20% above the European norm;
In complex surgery it was 50% above the norm
After the restoration process from October 2006-March 2007:
2 death in 310 operatons: 0,6%.

No public reporting of Dutch reference data (except one)!
Why are we not like New York State?


Except for Nijmegen and the St.Antonius hospital in Nieuwegein
the Dutch departments of cariodthoracic surgery did not collect
and publish data in public regarding their mortality and
comlications until 2006, when most due to the Radboud case
started to gather datas. Some had shown selective data in their
annual report.
In 2006 the surgeons from Nieuwegein published their data by
comparing 2002 with 1992. The overall mortality increased form
1,7% in 1992( 1537 procedures) tot 3,6% in 2002 (1742)
procedures). The number of re-operations decreased from 9,8% to
5,5%, the number of complicated procedures increased as the age
of the patients did. (Schoenmakers MCJ et al. Cardiac surgery and
operative mortality in 1992 and 2002; the St.Antonius hospital
experience Neth Heart J. 2006: 14: 132-138).
Do you understand data and statistics?

The external comittee extrapolated from the EuroSCORE sysytem
the CUSUM mortality analyses (CUmulative SUMmation).
The CUSUM is an analytical technique that presents a risk score of
every treated patient over time. If a patient survives a given
procedure, the care chain involved is awarded a performance score
corresponding to the risk predicted by the EuroSCORE.
The CUSUM should be zero (the expected normal death rate).

Above zeromeans less than expected and under zero more than
expected deaths.
The CUSUM score in Radboud Nijmegen.
CUSUM

2003
-3,7

2004
-17,9

2005
-12,6
How difficult are outcome data?




External Investigation Committee
At first an internal analyses using the EroSCORE system showed
data comparable to European dat, except for high risk patients
undergoing mitral ( and to some extent aortic) valve surgery.
There are several problems with the EuroSCORE
Relative uncommon risk factors and the socio-demographic
setting are not taken into account.
What about the relationship between outcome dat and QI?
(Should not we focus on process data?) EINDE PAG.4
Some data are simple
Complications:

Resternotomies: normal: 3%:
Radboud after CABG 8%
after aortic/mitral valve surgery 16%/18%

Renal failure :normal: 2,5%
Radboud after aortic/mitral valve surgery: 9%/15%

Bleeding: relationship with resternomies
The problem with outcome data
Many influencing factors:
1.
Time



2.
A. spontaneous fluctuations in time; the influence of chancee gets
bigger as the denominator gets smaller (n<200), especially if also the
nominator is small
B. the longer the time between your acivity and death, the more
influences of other factors you cannot influence
C. damage already sustained (complicated patients, tertiary referral;
palliative phase)
Co-morbidity; polypharmacy; age; sociodemographic factors;
Unknown factors (eg genetic profice); data imprefections (not
measured; not mistakenly or faulty registered): risk adjustment
methods cannot compromise them all and all have their
methodological weaknesses.
The problem with outcome data (2)
3.
Medicine is no mathematic sciense. Diseases (atypical presentation
or course) or atypical reaction towards therapy, patients and other
people involved and varying cirucmstances influence the calculated
risk (gain in quality vs increasy in unsafety)
4.
May induce unintended perverse behavior :excessive focus on the
measurement itself: preferring little benefit-low risk interventions:
excluding high risk patients ( referal or refusals:keeping out of the
registrations) and avoidance of well indicated but risky
interventions/treatments; data dredging or fraud
The restoration process:
prof dr L.Eijsman, prof dr van Swieten, C. Nogarede






1.Strong leadership; agreement=agreement; taking
responsibility;, tackle each other if necessary:surgical perfection
2. Team work
3. Standarisation/checks/clinical pathway/process and outcome
performance indication
4. Preoperative screening
5.Personal/experienced cariologic pre and post-operative
management
6.Multidiscipinacry complication and mortality registration, review
and discussion
The restoration process-2-: ambition and determination





Now the best centre in the Netherlands?
Ambition: to become the best and the safest hospital in Europe
The lower you fly, the higher you can reach
The failing Dutch suveillance system (SIN-NL: the Inspectorate of
Health Care) will be serious debated by ‘van Vollenhoven report’
Replace Radboud by Bristol and the stories are alike:
as well in their failure as in their triumph
The restoration process- 3-: 0,6% death rate








New team
Traditional military leadership
Disciplin
Process strictly redesigned
Protocols
Very stricly protocolled transfer; signed check list
Personal ‘experienced’ continuity
Taking responsibility
Some data are simple
Complications:

Resternotomies: normal: 3%:
Radboud after CABG 8%
after aortic/mitral valve surgery 16%/18%

Renal failure :normal: 2,5%
Radboud after aortic/mitral valve surgery: 9%/15%

Bleeding: relationship with resternomies
Prof. Dr. Mark Chassin: internist; Mount Sinai School of
Medicine; president Joint Commission




CABG registration NY State; serving 19 million people
From 1989; oldest patient/intervention related outcome base
First internal use:from 1990-1992 generally in public
Major results:
Decrease in death rates (-24%)
Graduallly more emphasis on data and case mix adjustments
Outlier stories
Outlier stories: individual and group

Dr A (St Peter’s Hospital):
1990: 5,2%!?
young/tlented: analyses: bad team
1995: 0,8%

Erie County Medical Centre
RAMR: 7,1% in 1989; volume 200 (:3)
3,6% in 1994: volume 464 (:2 and BPR)
Mark Chassin; 18 years of public reporting in NY State:
reduces death: no side effects: better risk adjusted data:
no negative media influence





Did we refer high risk patients out of state? No
Did fewer high risk patients receive CABG? No
Was non CABG prefered in high risk patients? No
Did patients select higher quality hospitals? No
Did the media loose initial attention? Yes, after 5 years.
Please be aware that the text of the powerpoint is copied and published
by SIN-NL from the hand-out provided by Dr. Hub Wollersheim
at his lecture on the 12th of October 2007.
After the lecture we asked Dr. Wollersheim whether the victims
and their relatives were informed open and honestly about the
medical errrors and whether they received adequate remedial
medical care. He answered in public: I am ashamed.