by Daniël P Melters
It is Tuesday morning December 8th, 2009. I’m working on the final report for a class I’m taking. Once this is done, I will get in my car and head over to Davis for my last meeting of the year with my bosses Ian Korf and Simon Chan. I’m in a good mood as my sister had just delivered a child and on Thursday my wife and I were going to get on a plane to Amsterdam to surprise my parents for their 30th anniversary. I had just reheated some left-over lasagna from the day before and about to cut my second bite when my cell phone rings.
It is 11.22 PT. The caller ID says it is a +31 number. I answer and I hear the voice of my sister. At first I thought she is calling me back after I briefly talked to her the day before to congratulate her with her newborn. But the call turns ugly quick. She says she has some bad news. My first thoughts go to my grandmother who was told she would not make the New Year due to a bad heart.
Again I am wrong. Eline, my other sister, was found dead. She tells me Eline had committed suicide. This is what the police told her. Not sure what to think, I ask her if she is sure Eline is dead. She confirmed. We talk for a bit, but my head is already spinning: what could have made her commit suicide?
After we hung up I texted my wife to call me ASAP, which she did. I brought her the news: Eline is dead. As we were talking, my parents call my wife. They weren’t sure if I was already at the university or still working at home as I wasn’t answering my phone. Not surpring as I am on the phone with my wife. This was the earliest moment my parents had the chance to call me, as they just came back from the police station in Sittard to hear the news.
Next up are all the administrative emails and phone calls. I informed everyone I thought needed to know. My wife did the same and also called the airlines to change our tickets to the earliest possible flight. This will be tomorrow, Wednesday December 9th. We’ll take it.
Once the basic administrative issues were dealt with, I turned to the internet. Searching for “dode vrouw Urmond”, I quickly found a link to De Limburger and it was apparent it described Eline’s suicide. But this report made me question if Eline actually did commit suicide. Why was she running over the streets of Urmond? Why was she also screaming? Why were the wounds to be suspected to be stab-wounds? Why was Mart S. note 1, Eline’s boyfriend, arrested? Not as a suspect, but for his own protection? Too many questions I could not put into place for a clear-cut suicide.
The next day, December 9th, we prepared to fetch the overnight flight from San Francisco International Airport to Amsterdam Schiphol Airport, but before we left our house, I checked the internet again to see if there was any news on Eline’s suicide case. There was. Mart was released and the case was closed. Mart was not involved because it was a suicide. We would be arriving at Amsterdam Schiphol Airport at the same time my parents were officially identifying Eline’s body in Maastricht. From the moment we landed in Amsterdam, I entered what was to become a 5-week long non-stop jet-lag. A jet-lag I wished would turn out to be just a bad dream.
How did Eline commit suicide?
Eline became agitated about something and started to undress herself, while at the same time forcing Mart to undress himself. This all happened on the ground-floor of the house they lived in together. Mart got scared and fled into the bathroom. After a count-down threat Eline stabbed herself the moment Mart came out of the bathroom. Following this stab, she ran out the house, tried to enter a church which was up a long and uneven flight of stairs. When she failed to enter this church, she walked on the sidewalk and slowly turned to a grass-field which was in front of a second church. When she tried to get over a small hedge she fell to the ground and died. The path she walked on the streets were marked with blood thatEline left behind. The events of what happened in their house was based on Mart’s testimony / statement.
The suicide scenario described above is not your average suicide, but in short this is what my sister did according to the Limburg-Zuid police’s investigation led by assistant public attorney (hulpofficier van justitie) and detective Bert Spoelstra note 1 and the public attorney of Maastricht Emma Oelmeijer-Naus note 1 in the Netherlands.
The question that popped up in my head shortly after I heard Eline was dead and from what I learned through the media was the following:
The few pieces of information I had gathered through the media made me question if
What is the expected likelihood for a 23 year old woman to commit suicide by stabbing herself in the heart in the presence of her boyfriend after which she runs away over homicide?
To answer this question, we need gather statistical information about suicides and homicides and compare them to each other.Than we can deduce if the conclusion of the police and public attorney are expected or an anomaly.
The crux of this case is that by questioning the conclusion of the police and public attorney, I’m insinuating that is must be homicide and there is only one suspect in that case: Mart S. note 1, Eline’s boyfriend. Saying that someone committed a homicide is a grave accusation and I do not intend to make one. Instead, my disagreement is with the speed at which the police of Limburg-Zuid and the public attorney of Maastricht came to conclude that Eline committed suicide. An important piece of information is that Eline’s body was found by the police at 10.02 CET. At the noon news bulletin broadcast on the radio channel L1 the police spokeswoman Renske Hamming stated that Eline most likely committed suicide. This means that there was a two hour window to uncover near conclusive evidence for Eline’s suicide. Thus, what evidence exists that could warrant such a conclusion in two hours after Eline’s body was discovered? How conclusive does this evidence have to be to be considered such evidence? In other words, this evidence has make suicide very likely and homicide very unlikely to have such gravitas.
What suicide:homicide ratio is expected?
The first step is determining the homicide and suicide rates in the Netherlands around 2009. In the period of 2000 through 2006 an average of 1519 people committed suicide per year, whereas in the period 1992-2001 an average of 229.3 people per year were murdered. The suicide rate only modestly fluctuated and therefore, I can be assumed that the suicide: homicide ratio was the same in 2009.
The ratio suicide:homicide is 6.6:1.
In 2005/2006 the murder rate suddenly and dramatically dropped from about 230 to around 170 a year. It cannot be excluded that this drop is due to the strong reduction in number of autopsies by the national forensic laboratories (NFI = Nederlands Forensisch Instituut), who have the near exclusive rights for forensic autopsies in the Netherlands. In 2005, 617 autopsies were performed whereas in 2012 only 340 were performed. In both cases, these numbers are surprisingly low when compared to countries such as Germany of Zwitserland, where suicides, traffic-accidents, drownings, SIDS, or hospital-deaths are not investigated by a medical pathologist or general practitioners/forensic doctor but by forensic pathologists. This potentially confounding factor has been addressed by various forensic pathologist, but to no avail. There I will use the value from before 2005 for homicide rates in the Netherlands.
The second step would to find out how often a knife or sharp object would be used in either a suicide of homicide. In total 37 to 42 suicides per year involve a knife, whereas 32% of all homicides are committed with a knife. This would mean that 81.5 murders per year are committed with a knife versus up to 42 suicides.
The ratio suicide:homicide has to be recalculated: 1:1.9.
In summary, by focussing on the use of a knife, the ratio suicide:homicide skew went from in favour of suicide to homicide being more favourable.
The third step is to determine the location of stabbing combined with the orientation. To get these statistics the paper by Thore Karlsson (1998) proofed to be very useful. This study was performed in Sweden, which could be a confounding factor. Yet, Sweden is both socially and politically very similar to the Netherlands. Therefore I assume that the data from this Swedish study are expected to be transferable to our case.
6% of all suicides are committed by stabbing vertically in the chest, whereas in 34% of homicides this is the case. Applying these percentages to the numbers from step two the results are the following:
0.06 x 42 = 2.52 suicides per year by vertical stabbing in the chest
0.34 x 81.5 = 27.71 homicides per year by vertical stabbing in the chest
The ratio suicide:homicide by vertical stabbing in the chest is 1:11.
If I had used the number of homicides of 170 (post 2005/2006), this suicide:homicide by vertical stabbing in the chest would be 1:7.3. One of the people, Frank van de Goot, who thinks that the number of detected murders in the Netherlands is artificially low performed Eline’s autopsy.
—start insertion 2013/03/17—
7% of all stabbing suicides are committed in the presence of someone, at least according to a single study (Start et al 1992). Of course, all homicides are committed in the presence of someone, as this is the only way person X can stab person Y. Applying these percentages to the numbers of step two (the study did not distinguish between vertical and horizontal stabbing, nor whether or not this was in the chest of abdomen or head-and-neck regions.
0.07 x 42 = 2.94 suicides by stabbing per year in the presence of someone.
1.00 x 81.5 = 81.5 homicides by stabbing per year in the presence of someone.
The ratio suicide:homicide would be the following: 1:27.7.
For sake of caution I will use the smaller ratio of 1:11 suicide:homicide, even though it is apparent that this number greatly more skewed towards homicide just based on the presence of Mart during Eline’s suicide. No clear data exists to determine how much exactly.
—end insertion 2013/03/17—
Just by taking into account the orientation and location of the wound, the police should have been motivated to treat my sister’s case as a potential homicide. This is what the statistics tell us. These expected values do not take into account the age or sex or depth of the wound or the number of wounds or the absence or presence of defensive wounds or the absence of presence of hesitation marks or other factors. The problem is that these additional parameters cannot be easily incorporated into a straightforward odds ratio.
Women in the Netherlands commit on average 487.9 suicides per year (from 2000-2006) significantly less than men who on average commit 1031.1 suicides per year (from 2000-2006). For those who use a knife, the fraction of women decreases even further. 7 to 10 women use a knife or other sharp object to commit suicide per year (or 2.0%), whereas 30 to 32 men commit suicide using a knife or other sharp object (or 3.1%). This ratio of 3:2 is on the high side and various studies have shown that the male:female ratio is more in the order of 3:1 for suicide by knife or other sharp object (see paragraph below). This does mean that 0.6 women are expected to commit suicide by vertically stabbing themselves in the heart. This is an incidence rate expected over all 25 police-districts.
The suicide:homicide ratio of 1:11 would be further skewed toward homicide when taking into account the age of the person committing suicide or falling victim to homicide. Eline was 23 years of age when she died. How likely is it that a 23 year old commits suicide by stabbing herself vertically in her chest? Little useful age-specific data exists, but we can look what the mean age was of people who used a knife to commit suicide.
- Fukube et al (2008): 55% of these suicides were between the age of 50 and 70; 5 were in their 20s (sex ratio: 16 female vs 49 male).
- Karlsson (1998): mean age = 51.5 years with a StDev of 17.07 (male to female ratio of 3.6:1).
- Scolan et al (2004): mean age = 52.8 years with a StDev of 18.6 (male to female ratio of 1:1)
- Karlsson et al (1988): mean age for females = 47.3 years with a StDev of 15.36 (male to female ratio of 3.3:1).
- Racette et al (2008): mean age for those with hesitation marks = 42 years (male to female ratio is 51:7), mean of those with defensive wounds = 39 years (male to female ratio: 60:59)
- Brunel et al (2010): mean age suicide = 48.2 (male to female ratio: 40:8), mean age homicide = 43.5 (male to female ratio: 47:23).
- Start et al (1992): no females under the age of 26 were identified (male to female ratio: 3:1).
Eline would fall outside one standard deviation of the mean age (where this information is available) to commit suicide by knife, reducing the likelihood she committed suicide by stabbing herself in vertically in the chest.
An additional skewing factor towards homicide would be the depth of the wounds (2 of the 3 stabs wounds exited the heart on the dorsal side of 10cm in depth) and the number stabbing wounds (2 confirmed entry wounds, 3 confirmed wounds in the sternum/ribs and heart). 21% of cases in the Start study had 2-5 stabbing wounds, whereas 64% of cases only had a single wound.
Are there any factors that might skew towards suicide? Indeed there is. Hesitation marks are a strong indicator for suicide (74% in Racette study; 64% in Start study), but postmortem analysis was unable to find hesitation marks. On the other hand, defensive wounds would be a strong indicator for homicide (61% in Racette study); which were found on both her hands. Even though defining these wounds are based on subjective interpretation, little doubt exists in the literature that these wounds are easily distinguishable.
Next, we could consider the presence or absence of a witness for a suicide (7% of stabbing suicides were witnessed in Start study). Of course, this statistic is useless for homicides, as every homicide requires a witness, the perpetrator. Therefore, I will disregard this option, in part as very little data is available for a reasonable indication for either suicide or homicide. Neither have I taken into account the use of threats before committing suicide or the lack of a suicide note (only found in 25-30% of all suicides) or the fact that Eline walked/ran for about 200 meters between being stabbed and dying from exsanguination.
The one factor that is believed to be strongly indicative of suicide was the fact that the stabbing location was undressed. Actually, Eline was completely undressed and according to Mart’s statements, he was forced to undress by Eline as well. If Eline did commit suicide, this could be indicative, but by no means conclusive, as shown by Karger et al (2000).
Finally, a mental disease or disorder was commonly found with people who commit suicide by stabbing themselves in the chest, most notably severe depression and schizophrenia (57% in Start study). Eline was not diagnosed with any psychiatric disease or disorder. Based on the statements of Mart, the public attorney Emma Oelmeijer-Naus and head public attorney Annemarie Penn-te Strake are convinced that Eline had (acute) phycological problems. On the morning (just after 08.00 CET) of her death, Mart called the general practitioner twice. By his own statement to request Eline to be placed on a emergency psych hold. This request never materialized. At the second call Eline talked to a psychotherapist which resulted in an appointment for later that day (13.45 CET). The precise nature of Eline’s psychological problems remain unmentioned and taking into account that no clinical diagnosis was made, it would be interesting to know how the public attorneys came to their medical diagnosis.
Both Mart and Eline took mind-altering substances as they both smoked a joint about 3 to 4 hours before Eline’s death, as they were both habitual smokers. Postmortem toxicology testing confirmed Eline had smoked. Yet, the head district attorney Annemarie Penn-te Strake deemed this information irrelevant for their conclusion of suicide.
Is this not to say that no suicides happen where someone stabs themselves in the chest? It is indeed very uncommon, but it does happen. Most of these cases are described as individual cases, such as a 44-yr old Japanese man who stabbed himself, changed his clothing, are lunch with his aunt before dying of cardiac tamponade. Or a case where a man stabbed himself an impressive 90 times before dying from exsanguination. Or an undetermined case in Poland where a man is found dead in the woods with multiple stab wounds to the chest.
Although none of the factors I described are conclusive, combined they do point to an expected odds ratio greater than 1:11 in favour of homicide. As differentiating numbers are missing for most of the described factors, I cannot determine if the odds ratio would have turned out to be 1:12 or 1:20 or 1:50 or 1:100. Even though, my point remains. The police should have been more cautious to treat Eline’s death as a suicide while at the same time been more eager to consider the possibility of homicide.
The Irrefutable Evidence or lack thereof!?
From this analysis there is no support for the swift conclusion of the police of Limburg-Zuid and the public attorney of Maastricht. This swift conclusion would be possible if yet to disclosed conclusive evidence was found on the scene. An additional negating factor is that the discovery of this evidence is time dependent. At the noon news bulletin on the local radio channel L1, police spokesperson Renske Hamming clearly stated that the woman found dead on the streets of Urmond most likely committed suicide. Eline’s body was found at 10.02 CET. This means that the evidence to point to suicide was found within 2 hours after finding her body. This evidence has to be strong enough to be near conclusive, while at the same time also making homicide extremely unlikely.
What evidence can be this conclusive?
- A goodbye note.
- An eyewitness of the suicide itself independent of Mart.
- A video-recording of the suicide.
The first option is unlikely as we were never told of the discovery of such a note, even though it would have fitted with her character, as she was a prolific poet.
The third option is also unlikely as we were never told of the discovery of such a video. If it would have been found, I can imagine that the police would be reluctant to show it, but why would they be reluctant to inform us about finding it? I cannot think of a valid reason.
This leaves the second open as the most viable of the three options. Yet, when we ask for this information, we are denied access to the police-report based on the privacy-sensitive information of Mart and Eline that is contained in this report. We even filed an official request to re-open the case through our lawyer Arthur van der Biezen, but the public attorney of Maastricht denied our request based on the testimonies made by Eline boyfriend Mart, his mother, the family-friend and general practitioner (GP) who was in contact with Mart and Eline earlier that day, and the psychotherapist, as well as some eyewitness(es) note 1. For these latter eyewitness(es) to be provide conclusive evidence should have witnessed the actual suicide, which means they must have been in the house of Mart and Eline. In addition, these witnesses must have made themselves available to the police the second they arrived on the scene. The former is not very likely, but the latter requirement is.
Mart’s mother cannot provide conclusive evidence, as Mart called her at 10.02 CET to talk with her for over 3 minutes and 2 seconds. It would be illogical to call someone who was in the same house as you were just a few minutes earlier.Note of special interest, this happened to be the only phone call Mart made after having calling the GP and before his arrest.
The whereabouts of the GP and the psychotherapist at the time of the stabbing I do not know, but as they both called earlier with Mart and Eline during their regular working hours during a regular working week, I would be surprised if they were present to witness the suicide.
This leaves Mart as a possible eyewitness, which he claims to be. If it was a suicide, this would indeed be correct. As described earlier, you do need a second person to commit a homicide with a knife. Although I do not want to claim this is the case, Mart would be the prime suspect if Eline did not commit suicide. His testimony should be treated as such.
Any evidence gather after noon cannot be considered as this would mean that the police made a tentative conclusion before the evidence was available. This includes any findings by the forensic doctor who arrived around 14.00 CET, the toxicology testing, and the autopsy. The latter two were not performed until the next day, Wednesday December 9th, 2009. Actually, the autopsy was ongoing when Mart was released on Wednesday and he was not considered a suspect anymore. Neither could any information have been gathered from Eline’s family as they weren’t informed of her death until the evening of December 8th.
As you can imagine, I am very curious what the evidence, found within the first 2 hours after Eline’s body was discovered, is and how that proofs that Eline committed suicide beyond a reasonable doubt.
note 1 It is convention in the Netherlands to not use full names of suspects of a crime, instead they use the first name with the initial of the last name. Even after a conviction this convention is continued. The names of victims, and especially of homicide victims are used in full. So rather than using Eline M, the media would use Eline Melters. I will adhere to this convention with the exceptions of assistant public attorney (hulpofficier van justitie) and detective Bert Spoelstra as he is a public figure as an author of crime-novels and officier van justitie Emma Oelmeijer-Naus and hoofd officier van justitie Annemarie Penn-ten-Strake (currently she a procureur generaal) as they are public attorneys and thereby a figures in a public function.
PS. I added a small section where I show that the presence of someone during a suicide greatly skews the suicide:homicide ratio in favour of homicide. See section marked —start/end insertion—