14 April 2010 Court of Arnhem
Acquittal in case Lucy de Berk
The “Committee Lucia de B.” (whose members are not related in any way to the family of Mrs. Lucia de Berk) was set up to convince both the Netherlands Judiciary, and the general public, that Lucia’s conviction is an appalling miscarriage of justice.
In June 2004, Lucia was convicted of 7 murders and 3 attempted murders by the Court of Appeal in The Hague. She was given a life sentence; in view of the lack of evidence, a perplexing sentence. There are no eye witnesses, there is no direct incriminating evidence. Lucia was never seen in a suspicious situation. She was never found in possession of any of the poisons she was alleged to have used.
Everything started with an at first glance striking number of incidents (deaths or resuscitations) during Lucia’s shifts at the Juliana Children’s Hospital in the Hague: the JKZ. The run drew attention to her. Seven incidents in a row all in the shifts of one nurse could not possibly be a matter of chance! The services of a former statistician, now professor of Psychology of Law, Henk Elffers, were called in, and the number he came up with must have wiped out all remaining doubt. He figured that the probability that all of seven incidents could have happened during Lucia’s shifts by pure chance was 1 in 6,000,000,000. Later, at the insistent request of the court, he extended his analysis to two wards of another hospital where Lucia had worked; but not the wards of two other hospitals where she was also charged with murder, and not in an earlier period during which she worked at JKZ, but when nothing happened. Making an elementary statistical error in the way he combined the information for the three wards, he finally came up with the chance (1 in 342,000,000) that such an extreme concentration of the incidents on all three wards would occur, simultaneously, purely by chance, in the shifts of one JKZ nurse. He stated categorically that the coincidence was not due to chance. The court accepted this conclusion and so did medical experts whose supposed medical judgements were strongly influenced by the statistical conclusion.
In the heat of the moment, no one took account of the fact that the selection of alleged incidents was extremely biased. A natural death changed into an unnatural death as soon as Lucia could be associated with it, just by her having been on the ward at the time. Conversely, an initially “suspicious” death was removed from the list as soon as it turned out that Lucia had not actually been present, after all. Concerning one alleged murder Lucia was found guilty of at the first trial, it later transpired that she had not been in the hospital for three days around the supposed event. This fact was only uncovered at the appeal. That particular “murder” immediately reverted to being a natural death; exactly what it had been, at the actual time it occurred.
As we said, the selection of incidents was biased. On studying all available records we have found three new incidents outside of Lucia’s shifts. Besides that, there is a number of cases in which we could not decide whether the incident occurred inside Lucia's shift or outside. Further, two of the seven “old” incidents fell outside her shifts. Using Elffers’ naive image of “pure chance”, working with his selected three wards and time periods, and using his statistical methodology, but correcting his combination-blunder, we did a recalculation of the probability that the coincidence of Lucia's shifts and incidents was pure chance. The outcome was 1 in 44 rather than 1 in 342,000,000. That is, on average 1 in 44 Dutch nurses may expect to meet such a coincidence. A chance of 1 in 44 is hardly shocking – unless we decide, following the lead of the court of appeal, that all of those nurses should be sentenced to life imprisonment.
In December 2006, statistics professor Richard Gill, allowing for possible time variation in the rate of incidents, and variation between nurses (a fully registerd nurse does not get the same kind of shifts as a trainee), came up with a chance of 1 in 9. His answer has so far not been disputed by anyone, and in particular, not by any statistician.
Thus, the Incredible Coincidence, which had such disastrous consequences, was built on a flawed statistical calculation using flawed data and flawed assumptions.
One might wonder if there is more evidence to convict Lucia. Well, there are two incidents which are “the locomotives”. They are supposed to pull the other eight cases, which admittedly could never reach the terminus “beyond reasonable doubt” on their own. This is not our metaphor! This is what the prosecution itself declared to the court.
So, what about the two locomotives that have to do the heavy work? These two cases were both on the JKZ medium care ward. They were the last two events in Lucia’s run of bad luck. Both children were in a very bad way. The first locomotive (chronologically the last, and the trigger of the police investigation) is an alleged acute and fatal digoxin poisoning of a baby girl. Digoxin had been used in her recent heart operation. The second locomotive is the intoxication of a baby boy by an overdose of chloral hydrate (a powerful sedative; he went into a coma, but came out again in a short time). Neither of these locomotives is even able to move itself.
The alleged digoxin intoxication is refuted by the following facts. (1) No digoxin was found in the liver. An acute digoxin poisoning without any digoxin in the liver is inconceivable. (2) The amount of digoxin found in the kidney is much too low; this was not noticed because the court’s expert did not know the literature on this topic. (3) The heart of the baby was not contracted as it would have been in the event of acute digoxin intoxication. The expert told the court about this relation but did not inquire whether the heart was contracted. It was not, and the prosecution knew that already, but did not inform the expert or the court. (4) The clinical situation is incompatible with acute digoxin intoxication. For example, the potassium level was low rather than high, and the crisis was very abrupt – six minutes from a "healthy" baby to death – while a digoxin intoxication can be seen coming for some time.
Possibly the second baby did go into coma from chloral hydrate intoxication, but there is nothing that connects Lucia to this intoxication, while the records of the hospital indicate that the specialist in charge prescribed the normal maximum dose and allowed that this be tripled in case of unrest. The medical records suggest that twice a double shot was given in the morning a few days before the intoxication. Actually the little boy survived this intoxication, but died three months later when (notice that we do not write because) the hospital is known to have given overmedication. On that later, fatal, day, the boy had been given other medication because of an operation. Incidentally, this child was being treated under a “not to be resuscitated” regime.
Let it be said that Lucia has never admitted any guilt, notwithstanding gruelling and denigrating interrogations. She was respected by many colleagues and superiors as a good and caring nurse.
We suspect that the Incredible Coincidence took control of the minds of the medical doctors and judicial system. They could not imagine that the situation was just a coincidence. So the incidents must be murders and murder attempts and Lucia must have done them, no other explanation came to their minds.
Since her arrest (which aroused great media attention), Lucia has been, for the general public, the prototype of a new kind of serial murderer, who killed patients in hospitals. The public knew vaguely that the court was short on evidence. But with so many convictions, almost everyone supposed that she must have killed some of her patients.
In September 2006, professor Ton Derksen, philosopher of science and writer of the book “Lucia de B.: Reconstruction of a Miscarriage of Justice”, presented Lucia’s case to a special, non-permanent, Commission for the Review of Closed Criminal Cases. A scientist who had written a study on a case was allowed to do this, the convict herself not. The application was accepted, and a triumvirate (“three wise men”) started work on the case. They will report to the Public Prosecution Service. Nearly a year on (July 2007), they have not yet communicated any findings or conclusions.
In the Netherlands there is a growing number of people who believe that the case has to be reopened. Even some judges have said so in public. Medical doctors and specialists will not speak on the case in public (unless they are retired). One told us that speaking out would be the end of anyone’s medical career. And we would not want that to happen, would we?
Internationally there is also growing support. Nature (January 2007) published an article on the flawed statistics. So did The New York Times and The Guardian.
Lucia de Berk has been behind bars since 2001. On March 5 2006, after hearing the disconcerting verdict of the Supreme Court, Lucia had a stroke. For 10 hours Lucia lay in her cell, paralyzed, before she was transported to a hospital. Her family was only informed a day later, about her transfer to the Penitentiary Hospital. Slowly, with minimal therapy, Lucia is recovering. Her speech is almost normal, but she walks with a slight limp, and she has no use of her right hand. Her punishment is now even more severe. Not only for her, but also for her loved ones. This is a nightmare and sadly it is a nightmare that could happen again.
By drawing attention to the case, we realize that we will hurt the feelings of members of the families of the children and older patients who died. We feel deeply sorry for these families, but the old “evidence” has to be scrutinzed in order to convince the judicial authorities that Lucia’s conviction must be overturned.
This committee was founded to undo the injustice done to Lucia. We ask for your support.
Lucia de Berk (born 1961), in the Dutch media generally called Lucy de B. or Lucia de B., is a Dutch nurse who was sentenced to life imprisonment in 2003 for four murders and three attempted murders on patients in to her care. After an appeal she was convicted in 2004 of seven murders and three attempts. Her conviction is controversial among the media and scientists and has been questioned by Peter R. de Vries.
As a result of an unexpected and unexplained death of a baby in the Juliana Children's Hospital (JKZ) in The Hague, The Netherlands, on 4 September 2001, earlier deaths and resuscitations were scrutinized. Between September 2000 and September 2001 there appeared to have been nine incidents, which earlier had all been thought unremarkable but now were considered medically suspicious. Lucia de Berk had been on duty at the time of those incidents, responsible for patient care and delivery of medication. The hospital decided to press charges against the licensed child nurse.
On 24 March 2003 de Berk was sentenced by the court in The Hague to life imprisonment for the murder of four patients and for the attempted murder of three others. The verdict depended in part on a statistical calculation, according to which, allegedly, the probability is only 1 in 342 million that a nurse's shifts would coincide with so many of the deaths and resuscitation purely by chance. De Berk was however only sentenced in cases where, according to some medical expert, other evidence was present or in which, again, according to some medical expert, no natural causes could explain the incident.
In the appeal on 18 June 2004 de Berk's conviction for the seven murders and three attempted murders was upheld. The crimes were supposed to have taken place in three hospitals in The Hague; the Juliana Child Hospital (JKZ), the Red Cross Hospital (RKZ) and the Leyenburg Hospital where de Berk had worked earlier. In two cases the court concluded that there was proof that de Berk had poisoned the patients. Concerning the other cases the judges considered that they could not be explained medically, and that they must have been caused by Lucia de Berk, who was unable or unwilling to give a reason why she was present on all those occasions. The idea that only weaker evidence is needed for the subsequent murders after two have been proven beyond reasonable doubt has been dubbed chain-link proof by the prosecution and adopted by the court. At the 2004 trial, de Berk also received TBS (terbeschikkingstelling, psychiatric detention with coerced psychiatric treatment), which is a strange addition to a life sentence, when the suspect did not plead mental illness as cause of her crimes (she denied committing them), nor did the state criminal psychological observation unit find any evidence of this. The Court found this so far unique sentence necessary, because of the likelihood that she would otherwise return to society after the standard release from life imprisonment after 20 years detention. This way de Berk would remain detained for psychiatric treatment and prevented from ever returning to society.
Important evidence at the appeal was to be the statement of a detainee in the Pieter Baan Center, the criminal psychological observation unit, at the same time as Lucia de Berk, that she had said during outdoor-exercise: "I released these 13 people from their suffering". However, during the appeal, the man withdrew his statement, saying that he had made it up. The news service of the Dutch Broadcasting Foundation (NOS), and other media that followed the process, considered the withdrawal of this evidence as a huge setback for Public Prosecution Service (OM). This witness statement was the only hard evidence, and it had vanished. A series of articles appeared over the following years in several newspapers, including Vrij Nederland and the Volkskrant, doubting the safety of the conviction. It was suggested that the prosecution had manipulated evidence and the judges had drawn the wrong conclusions.
The case was next brought to the Supreme Court, which ruled on 14 March 2006 that it was incorrect to combine life imprisonment with subsequent psychiatric detention. Other complaints were not taken into consideration, and the evidence from the Strasburg analysis was not considered relevant. The Supreme Court gave the matter back to the Court in Amsterdam to pass judgement again, on the basis of the same factual conclusions as had been made before. Some days after the ruling of the Supreme Court, de Berk suffered a cerebral infarction and was admitted to the hospital of Scheveningen prison. On 13 July 2006, as expected, de Berk was sentenced by the Court of Appeals in Amsterdam to life imprisonment, with no subsequent detention in psychiatric care.
A few persons continued to express doubts about the conviction of de Berk. Philosopher of science Ton Derksen, and his sister, geriatrician Metta de Noo-Derksen, wrote the (Dutch language) book "Lucia de B: Reconstruction of a Miscarriage of Justice" . The doubts are related to the chaining-evidence argument used by the court and the medical and statistical ingredients.
Of the seven murders and three attempted murders finally attributed to de Berk by the court, only two were considered by the court proven by medical evidence. According to the court de Berk had poisoned these two patients. The court then applies a so-called chaining-evidence argument. This means that if the several (attempted) murders have already been established beyond reasonable doubt, then much weaker evidence than normal is quite sufficient to establish that a subsequent eight “suspicous incidents” are indeed suspicious and in fact are murders or attempted murders carried out by de Berk.
This is controversial, because for the two murders found proven by the court, many experts still did not exclude a natural cause of death. In the case where digoxin poisoning was alleged, and supposedly detected by independent measurements in two Dutch laboratories, the (same) method used in those laboratories did not exclude that the substance found was actually a related substance naturally produced by the body itself (i.e. this could be a false positive, and the two measurements are not better than one). The Strasburg laboratory used a new method, sometimes refered to as the gold standard, a test of high specificity and sensitivity, and it did not reveal presence of any digoxin. In the other case, the intoxication could well have been an overdose caused by a faulty prescription. For both children, the question moreover remained, how and when de Berk was able to administer the poison. Regarding the digoxin case, the prosecution gave a detailed reconstruction of the timing. However, other parts of the evidence, discarded by the prosecution, showed by the timestamp on a certain monitor that at the alleged moment of poisoning Lucia was not with the patient at all, and in fact the specialist and his assistant were with the patient at that time.
The prosecution initially charged de Berk of causing thirteen deaths or medical emergencies. In court, the defense was able to show definitively that de Berk could not have been involved at all in several of these cases. For instance, she had been away for several days; the idea that she was there was due to administrative errors. Furthermore, it should be pointed out that all deaths had been registerd as natural, with the exception of the last event. Even that last event was initially thought natural by the doctors responsible for the child, but within a day, on being connected with Lucia and her repeated presence at recent incidents by other hospital authorities, it became an unnatural death.
There is no hard proof against de Berk. She was not caught “red-handed”, and has always denied guilt. Among other things, the court made heavy use of statistical calculations to achieve its conviction. The chance of a nurse working at the three hospitals being present at the scene of so many unexplained deaths and resuscitations is estimated at one in 342 million. The relevance of this number is controversial. It seems that the hospital and the OM had decided in advance that the deaths had been caused by a nurse; and still insist exclusively on this line of enquiry.
The multiplication of probabilities concerning three departments was wrong: if one wishes to combine p values (right tail probabilities) of the statistical tests based on data from three separate wards, one must introduce a correction according to the number of tests, as a result of which the chance becomes one in a million (Fisher's method for combination of independent p-values ). In the summing up after the appeal, the Court of Justice no longer reported the results of a probabilistic calculation. However, the judge implicitly continued the use of the statistical data. It was considered clear that so many incidents could not have occurred by chance. Derksen has reported to the Posthumus II committee, chaired by Y. Buruma, that the number of incidents that occurred during de Berk’s shifts were too large, and the numbers for other nurses were too small. Correcting those figures, increases the chance to about one in fifty. Taking account also of some innocent variation in the rates of incidents experienced by different nurses, as suggested earlier by A.F. de Vos [PDF], the chance becomes one in nine. In the judges’s summing-up, evidence is selected from many different medical experts, in order to argue that each of the incidents to be attributed to Lucia was indeed an unnatural death or near-death. Sometimes specialist A is preferred above specialist B, sometimes the other way around, sometimes another, C. In that case, C is arguably the most expert person, because he alone considered the incident unnatural. In another case there is only one expert who considers the death unnatural. He actually stated that he only has this opinion because Lucia was present and she had been present at so many other incidents. This is not mentioned in the judge’ conclusion.
The use of probability arguments in the de Berk case was discussed in a 2007 Nature article by Mark Buchanan
The court needs to weigh up two different explanations: murder or coincidence. The argument that the deaths were unlikely to have occurred by chance (whether 1 in 48 or 1 in 342 million) is not that meaningful on its own – for instance, the probability that ten murders would occur in the same hospital might be even more unlikely. What matters is the relative likelihood of the two explanations. However, the court was given an estimate for only the first scenario.
De Berk's diary also played a role in her conviction. On the day of death of one of her patients (an elderly lady in a terminal stage of cancer) she wrote that she had 'given in to her compulsion.' She wrote on other occasions that she had a 'very great secret,' and wrote that she was concerned about 'her tendency to give in to her compulsion.' De Berk stated that this referred to her passion for reading tarot cards, whereas the court decided it was evidence that she had euthanized the patients. De Berk explains that she did this in secret, because she did not believe it appropriate to the clinical setting of a hospital. According to the court the reading of cards does not accord with a compulsion nor with perhaps an expression of fatigue as she described it at the time. The daughter of de Berk, Fabiënne, explained in an interview on the television program Pauw & Witteman that some of her mother's notes in the diaries are 'pure fiction', which she wanted to use for a thriller.
After the appeal procedings were closed, but before the judges delivered their verdict, the Public Prosecution Service received, via the NFI, a report from a prestigious forensic laboratory in Strasbourg. This had been requested by the NFI who had hoped that it would support the hypothesis of digoxin poisoning. In fact, the NFI wrote to Strasburg that they were hoping for a miracle. The report subsequently lay in a drawer of the NFI for two years, but it did turn up in time for the final evaluation of the case before the Supreme Court. According to the Public Prosecution, the report contained no new facts, but according to de Berk's defence the report proved that there was not a lethal concentration of digoxin in the first case. The Supreme Court accepts the facts reported by the judges at the appeal court, and is concerned only with jurisprudence and correctness of the sentence, given those facts. The report therefore was not admitted to the final considerations of the sentence given to Mrs. de Berk.
Ton Derksen submitted his and Metta de Noo's research on the case, to the Posthumus II Commission. This ad hoc, non permanent, commission examines some selected closed cases and looks for evidence of errors in the police investigation regarding "tunnel-vision" and misunderstanding of scientific evidence. Derksen pointed out that the medical experts who had ruled out the possibility of death by natural causes had not been given all relevant information, that the hypothesis of digoxin poisoning was disproven, in particular by the Strasburg analysis, and that the statistical data was biased and its analysis incorrect, and the conclusions drawn from it invalid. The commission announced on 19 October 2006 that this is one of the few cases it will consider in detail. Three wise men, recruited by the Public Prosecution service from the full Posthumus II committee, are considering the following matters, though they have been instructed to focus on possible blemishes in the criminal investigation:
The Three Wise Men will make no judgement on de Berk's guilt or innocence but could recommend that her case be reopened. It is now expected that the commission will make its recommendation in October 2007.