- Email: [email protected]
A record linkage study on former police detainees who died in Amsterdam between 2013 and 2015
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Journal of Forensic and Legal Medicine journal homepage: www.elsevier.com/locate/yjflm
A record linkage study on former police detainees who died in Amsterdam between 2013 and 2015 Daniël B. Smidta,∗, Tina Dornb, Udo J.L. Reijndersc a
Parnassia Groep, The Netherlands Amsterdam Public Health Service, Department of Epidemiology and Health Promotion, Amsterdam, The Netherlands c Amsterdam Public Health Service, Department of Forensic Medicine, Amsterdam, The Netherlands b
A R T I C L E I N F O
A B S T R A C T
Keywords: Forensic medicine Forensic physician Post-mortem examination Police detainees Unnatural death
Background: Police detainees are known to have inferior health. This study identifies the number of former police detainees who received medical care among deaths examined by forensic physicians and presents their death characteristics. Methods: We included all deaths that were examined by forensic physicians of the Public Health Service Amsterdam from 2013 to 2015. Patient files of subjects were scanned for the presence of a prior medical consultation in the police cell and death characteristics were collected from post-mortem examination reports. We performed statistical analyses to discover what characteristics at post-mortem examination were associated with a prior consultation in the police cell. Results: We identified n = 2618 subjects that met the inclusion criteria. Eight percent of subjects had one or more medical consultation(s) in the police cell in a mean follow up time of 4.8 ( ± 3.0) years. No difference was found in the share of unnatural deaths between subjects with and without a prior consultation (68%), but distribution of death causes differed significantly. Male gender OR 2.3 (p < 0.001), age OR 0.98 (p < 0.001), unspecified unnatural dead OR 1.8 (p = 0.002), crime related dead OR 2.2 (p = 0.012) and accidental drowning and submerging death OR 4.6 (p < 0.001) were independently associated with the presence of an earlier consultation in the police cell. Conclusion: Our data suggest that a small percentage of police detainees seen by forensic physicians for provision of medical care are also examined after death by these physicians, typically young males who seem to display risk-taking and criminal behavior resulting in unnatural dead.
1. Introduction The health status of police detainees has been subject of international research in the past decade indicating high prevalence of psychiatric symptoms and substance use.1–6 In Amsterdam, approximately half of all police detainees who had received medical care while in custody were diagnosed with psychological problems and a large share of prescribed medication in this setting was for treatment of addictive disorders.7 Nearly 40% of Amsterdam police detainees submitted to screening qualified for a mental health care assessment.8 Furthermore, up to one third of young violent offenders studied in Amsterdam were eligible for public mental health care.9 Next to poor mental health, more chronic somatic conditions were seen among police detainees in Amsterdam when compared to the general population.10 Although there is a growing body of research on health issues in police detainees, studies on mortality in former police detainees are
∗
absent to date. More extensive research on this subject has been done in prisoner populations. Prisoners, however, differ from police detainees because a large part of police detainees will not be sentenced to prison, but will be released following initial police contact. Moreover, in many countries, people with mental illness who are in need of psychiatric emergency care are temporarily brought to a police station, awaiting further assessment by mental healthcare professionals. Studies in prisoner populations, mostly executed in the US, have shown that prisoners alike police detainees, have significantly more health problems (psychiatric and somatic) when compared to the general population.11–13 A Dutch review, largely based on Dutch studies confirms the higher prevalence of psychiatric and somatic complaints in (released) inmates.14 Studies that assess the risk of death in (released) prisoners compared to the general population are widely available.15–25 A systematic review by Kinner, Forsyth and Williams of 29 record linkage studies on mortality in ex-prisoners revealed a pooled all-cause standardized mortality
Corresponding author. E-mail address: [email protected] (D.B. Smidt).
https://doi.org/10.1016/j.jflm.2018.02.004 Received 5 March 2017; Received in revised form 2 February 2018; Accepted 5 February 2018 1752-928X/ © 2018 Published by Elsevier Ltd.
Please cite this article as: Smidt, D.B., Journal of Forensic and Legal Medicine (2018), https://doi.org/10.1016/j.jflm.2018.02.004
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
examination or after death. The following variables were automatically retrieved from the postmortem examination reports of all subjects: Age at death, gender, manner of death, cause of death, reporting caller, and site of the body on examination. All data were anonymized and imported into the IBM SPSS statistics® program.
ratio (SMR) of 1.7 (95%CI 0.6–5.0) (25 studies). Pooled SMRs were 8.2 for suicide deaths (3 studies) and 8.4 for unnatural deaths (2 studies) in a mean follow up time varying from 12 weeks to 16.1 years. Follow up studies of 25–50 years in the Netherlands, Sweden and the USA in (ex) crime and/or violent offenders revealed comparable higher risks of death (1.6–2.0) when compared to the general population.15,18,20 This risk increased further with recidivistic criminal offences and alcohol abuse.20 Suicide deaths were more common among repeat violence offenders, violence offenders with a history of receiving psychiatric inpatient care and violence offenders with a history of substance abuse.20,26 As both prisoner and police detainee populations overlap, both populations might share comparable risk factors for ill-health and premature death. We performed the current study in order to shed more light on mortality in former police detainees to explore the possible role for forensic physician guided preventive interventions. In this study we attempt to answer the following questions:
3. Statistical analysis Statistical analyses were executed with the IBM SPSS Statistics® program. Comparative analyses were made to assess differences in findings at post-mortem examination between both groups. For categorical variables effect sizes were calculated as odds ratios. For age (continuous variable) the mean difference was calculated as effect size. Statistical tests were used to assess significance of differences found. For age, an independent sample's t-test was used. For categorical variables we used Pearson chi-squared tests. Assumptions for valid testing were confirmed before analyses. Variables that differed significantly in the univariate analysis were assessed in a binary logistic regression model (enter method), with prior consultation (yes/no) as outcome variable. Effects sizes were calculated as odds ratio's.
1. What percentage of all individuals examined post-mortem by forensic physicians of the Public Health Service Amsterdam have a history of receiving medical care in the police cell? 2. Do individuals examined post-mortem by forensic physicians with a history of receiving medical care in the police cell differ from individuals without such a history with regards to manner of death, cause of death, age, and gender as reported at the post-mortem examination?
4. Results We identified 2650 post-mortem examination reports that met all criteria by using the electronic patient register to search for eligible subjects. We excluded 1575 cases of physician-assisted death. In our final analyses, we included 2618 reports. Reasons for exclusion were: the presence of a second post-mortem examination report on the same case or not being able to identify a full post-mortem examination report.
2. Methods 2.1. Setting In the Netherlands, forensic physicians provide medial care in the police cell. This care is comparable to the care provided by a general practitioner. The forensic physician needs to decide whether a detainee can stay in the cell from a medical point of view. If not, hospital referral or referrals to psychiatric emergency services will be organized by the forensic physician. Another major task of forensic physicians in the Netherlands is the external examination of all (potentially) unnatural deaths. If the forensic physician concludes an unnatural death, the case is discussed with the prosecutor who might order further police investigations and/or a medicolegal autopsy by a forensic pathologist.
5. General characteristics Mean age ( ± SD) at death of subjects was 63 ( ± 21) years, but differed significantly (p < 0.001, t-test) between male 59 ( ± 20) and female subjects 70 ( ± 20) (data not shown). The majority of 2618 subjects was male (66%). In 68% of all deaths referred to the forensic physician, the manner of death was confirmed as unnatural after the post-mortem examination. The most common cause of an unnatural death was a falling accident (23%), which showed an association with high age at death (mean age 82 ± 13; data not shown). The majority of natural deaths examined by the forensic physician cannot be assigned a specific cause of death (44%) after the post-mortem examination. Most deaths (53%) were signaled to the forensic physician by the police force. Deaths referred by the police were assigned an unnatural manner of death by the forensic physician in 55% of cases. Deaths referred by general practitioners were assigned an unnatural manner of death in 61% (data not shown).
2.2. Design A retrospective cohort was identified by searching the electronic patient register of the forensic medical department of the PHS Amsterdam. This register contains records of all activities executed by forensic physicians of the department starting 01-01-2005. Record linkage was based on unique identifiers. Data sources that have been linked concern the module on medical care and the module on postmortem examinations (both are two modules within the same database). All individuals who were examined post-mortem by forensic physicians between 01 and 01-2013 and 31-12-2015 were included as subjects to provide the most recent sample of post-mortem examinations executed by forensic physicians in the Amsterdam area. Only cases of physician-assisted death in this period were excluded. In the case of physician assisted death, the forensic physician monitors whether the physician who assisted in the death acted according to the procedures specified by Dutch law. The outcome of interest was the presence of one or more prior care contacts in the police cell in the catchment area of the PHS Amsterdam between 01 and 01-2005 and the date of the post-mortem examination. This design resulted in different retrospective intervals for included subjects. A prior care contact was defined as the presence of at least one contact of care in the electronic patient file, besides the post-mortem examination itself and contacts on the same day of the post-mortem
6. Post-mortem examination findings of ex police detainees Of 2618 individuals, 213 (8%) had one or more prior consultations in the police cell by a forensic physician before death. Age at death was lower in the group that received medical care in the police cell 50 ( ± 14) with a mean difference of 14.5 (11.7–17.4) years (p < 0.001) and subjects with a prior consultation were more often male when compared to subjects without prior history, OR 3.5 (2.4–5.3) (p < 0.001). No difference was found in overall manner of death between individuals with and without a prior consultation (68%). Table 1 shows, however, that distribution of unnatural causes of death differed between both groups. A high incidence of accidental drowning and submerging deaths was found in the prior consultation population 5.2% 2
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
Table 1 General characteristics of all post-mortem examinations executed by forensic physicians of the PHS Amsterdam in a 3 year period (2013–2015). Including a comparative analysis of characteristics at death between subjects with a prior consultation in the police cell (n = 213) and those without (n = 2405).
Denominator of percentages in column Age (mean in years)(SD) Male (%) Unnatural manner of death (%) Unnatural death causes (%) Accidents falling accidentsa traffic accidents accidental drowning and submerging accidental obstruction of airway Occupational accidents rest Suicide Doubt about natural manner of deathb Crime Natural death causes (%) Unknown Cardiovascular diseases Diseases of the digestive tract Rest Reporting caller (%) Police Location of body (%) Home
All subjects
Prior consultation +
Prior consultation -
OR
95% CI
2618 63 ( ± 21) 66% 68%
213 50 ( ± 14) 86% 68%
2405 64 ( ± 21) 64% 68%
* 3.5* 1.0
2.4–5.3 0.7–1.3
35% 23% 4.3% 1.6% 1.1% 0.6% 4.1% 16% 13% 2.9%
18% 2.8% 3.3% 5.2% 0.9% 0.5% 0.5% 17% 24% 8.0%
36% 25% 4.4% 1.2% 1.2% 0.6% 4.0% 16% 12% 2.5%
0.38* 0.1* 0.7 4.3* 0.8 0.8 1.3 1.1 2.3* 3.4*
0.3–0.5 0.0–0.2 0.3–1.6 2.1–8.7 0.2–3.4 0.1–5.7 0.7–2.5 0.8–1.6 1.6–3.2 1.9–5.9
14% 12% 2.2% 4.2%
5% 8.5% 4.2% 5.2%
14% 12% 2.0% 4.1%
1.0 1.5 2.2* 1.3
0.7–1.5 0.4–1.1 1.0–4.5 0.7–2.4
53%
77%
51%
3.1*
2.2–4.3
42%
52%
41%
1.5*
1.2–2.0
*p < 0.05. a Falling accidents: falls caused by slipping and stumbling, falls on stairs and in staircases, falls from ladders, falls from scaffolding, falls from or through a building or construction, falls from trees, other kind of falls (non-occupational accidents). b Doubt about natural manner of death: This category is a conglomerate of the ICD-10 codes belonging to the category ‘Ill-defined and unknown causes of mortality (R95-R99)’, more specifically code R96 ‘Other sudden death, cause unknown’. These ill-defined cases are categorized as unnatural deaths because the physician cannot be ‘convinced of a natural cause’ and the Dutch law states that only in case of no doubt about an unnatural manner of death a natural manner of death can proclaimed.
(11/213), compared to individuals without a prior consultation 1.2% (30/2405), p < 0.001. Nine out of eleven (82%) individuals with a prior consultation who drowned were known substance abusers (mostly alcohol) and strong indications existed at post-mortem examination that substance use played a direct role in the accidental drowning or submerging. In individuals without a prior consultation who drowned, 11 out of 30 (37%) deaths were suspected of having a relation with substance abuse. Crime related deaths and unnatural deaths with an unknown cause were more common in subjects with a prior consultation, OR: 3.4 (1.9–5.9) and 2.3 (1.6–3.2) respectively, both p < 0.001. Mean age of crime related deaths in the prior consultation group was 32 ( ± 9) versus 43 ( ± 18) in the group without prior consultation. No difference in number of suicide deaths was found. The amount of accidents as cause of death was higher in the group without prior consultation (36% versus 18%). This difference was largely caused by the difference in falling accidents, 25% in the group with a prior consultation versus 2.8% in the group with no prior consultation. Distribution of natural causes of death was comparable in both groups, although deaths related to digestive tract disorders were more common in the group with a prior history of consultation, OR 2.2 (1.0–4.5), p = 0.033. In the multivariate analysis (Table 2), accidental drowning showed the largest association with a prior consultation in the police cell (OR 4.6). Odds for dying of an unknown unnatural death or of crime related causes were also correlated with an almost 2 times increase (1.8 and 2.2 respectively) in the prior consultation group. Lower age and male gender remained correlated with a prior history of consultation too. Natural death caused by a digestive tract disorder lost significant association when adjusted for police as referrer, possibly due to the close relationship between the variables referrer and manner of death. Difference in accidents as cause of an unnatural death lost significance when adjusted for age, probably because mostly old people died from accidental causes in our sample (fractures, falls).
Table 2 Multivariate analysis of characteristics found at death in relation to prior consultation in the police cell. Variablea
Odds ratio
95% CI
Gender (male = 1, female = 0) Age Cause of deatha Unnatural: accidents Unnatural: doubt about natural manner of death Unnatural: crime Unnatural: accidental drowning and submerging Natural: digestive tract Reporting caller: police (police = 1, rest = 0) Location of body: home (home = 1, rest = 0)
2.3* 0.98*
1.5–3.4 0.98–0.99
0.76 1.8* 2.2* 4.6* 2.0 1.6* 1.2
0.5–1.3 1.3–2.7 1.2–4.1 2.0–10.5 0.96–4.3 1.1–2.4 0.84–1.7
*p < 0.05. a Cause of death was modelled using 5 dummy variables, the reference category thus consists of all deaths not belonging to these 5 categories.
7. Discussion We found that 8% of 2618 subjects examined post-mortem by forensic physicians in the Amsterdam region, were known to the forensic medical department as a result of an earlier medical consultation in the police cell. Death in individuals who had received care in the police cell was significantly associated with accidental drowning, death from crime and unnatural death with no specific cause. This is to expect as former detainees have a higher risk of being involved in crime-related activities and crime-related death later in life. Moreover, those with prior care contacts in the police cell were more often male and mean age at death was significantly lower when compared to those without prior contacts. A difference of 14 years in mean age at death was observed between subjects with and without prior consultation in the police cell. Studies that reported age at death in former offenders 3
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
consequence, former police detainees from the Amsterdam region may have been missed as they may have died elsewhere. The other way around goes too, individuals may have died in Amsterdam but may have been seen in a police cell outside the catchment area of the PHS Amsterdam. Third, the retrospective design of this study does not allow for the calculation of (standardized) mortality rates and we cannot conclude that overall mortality in former police detainees is higher when compared to the general population. Fourth, we used a limited retrospective interval for identifying contacts with a forensic physician in the police cell. It is likely that if we had used a longer retrospective interval, the number of subjects with a prior care contact in the police cell would have been higher. This could have resulted in a shift in distribution of death causes in both groups in either way. Fifth, the data used in this study were not primarily recorded for research but for general medical and law related purposes. Information on risk factors such as medical care history, substance use, socio-economic status and lifestyle factors was not available. In previous research, relationships between the nature and amount of crimes committed or the number of imprisonments and the risk of dying of unnatural causes were found,17,18,22,26 we could not adjust for these factors in this study. Finally different physicians do not always arrive at the same conclusions concerning manner and cause of death. The inter-physician variation between forensic physicians can be considerable.33 This inter-physician variation probably applies for both groups (with and without prior consultations), however, so no systematic bias is expected regarding this aspect. A strength of this retrospective cohort study is that we used postmortem examination reports as our main data resource. These reports provide structured and comparable information for all subjects because a standardized format is used for reporting. We included all reports in the catchment area of the PHS Amsterdam and due to the exclusive right of this department to perform the post-mortem examinations, we can provide a complete and representative overview of deaths examined in this area. To conclude, this study adds to the body of current literature on mortality in former police detainees.
showed the same tendency, although observed ages at death were lower.15,20 The large majority of subjects with prior care contacts in the police cell before death (86%) was male. This is comparable with studies in police detainees in Amsterdam were an average 89% was male8,10,27 and studies in prisoner populations demonstrating high male to female ratios.18,21,22 Distribution of unnatural causes of death in both groups was distinctly different. Subjects with a prior consultation died from a crime related death more frequently (OR 3.4). This association is supported by prior research on causes of death in a cohort of convicted Dutch criminals, that showed a risk ratio of 5.9 for dying of crime related causes in a 25 year follow up.18 However, methodological differences exist between both studies limiting comparability. Firstly, our prior consultation group was not referenced to the general population but to those without a prior consultation examined post-mortem by forensic physicians. Secondly, we used a shorter follow up interval and a retrospective view. Thirdly, we used post-mortem examination reports as our primary inclusion method. Furthermore, we found that an accidental drowning death was significantly associated with a prior consultation in the police cell. When reviewing the cases of former police detainees who drowned, there seemed to be a high coincidence with a history of substance abuse. This association is also found by other studies.28,29 Although previous studies did show higher number of suicide deaths in ex-prisoners and law offenders when compared to the general population,16,18,21,25 we found no difference in number of suicides between our groups. Differences in results between our study and previous research might be explained by the fact that former study populations were younger,16,18 suicides were more common among younger subjects21,26 and subjects were convicted criminals instead of (ex) police detainees.16,18,21,26 The risk of suicide in ex-offenders was correlated to the number of offences committed20,26 and the presence of psychosis or a psychiatric hospital admission during imprisonment.16,26 In this regard it is important to notice that police in Amsterdam has the option to present arrestees that appear confused as result of a psychiatric disorder to a psychiatric emergency service without prior consultation of a forensic physician. This policy might have limited the number of police detainees with psychosis seen by forensic physicians. Secondly, no information on the criminal records of included subjects was retrieved for this study. In this study we have compared post-mortem examination findings in former police detainees who received care in the police cell to individuals without such a history. The deaths examined post-mortem by the forensic physicians differ from deaths in the general population, however. In 2015 only 20% of 5425 deaths in the general Amsterdam population were aged under 64 years at death,30 while this was 46% in our sample of 935 deaths from 2015. The share of males was also higher in our sample than in the general population, 66% versus 49%.30 Moreover, the amount of unnatural deaths in our sample is almost 16 times (68/4.3) higher than in the general population.31
9. Implications What remains is the question if there is a group in police custody that is prone to inferior health and premature death because of risk factors that may be influenced, e.g. by offering special care aiming at reducing re-offending or interventions to reduce risk-taking behavior in this group. Reasons for premature death are multifaceted and often interrelated with other vulnerabilities such as low income, low educational level, psychiatric and addiction disorders. It is unknown how many police detainees will be sentenced and incarcerated later in life, but is known that released prisoners are at increased risk for death following release from prison. This emphasizes the need for aftercare planning after release and the need to identify groups that should be targeted to reduce premature death.34 The share of police detainees that commit relatively low grad but often troublesome offences and often do not qualify for imprisonment could possibly benefit from targeted measures. Ideally, forensic physicians systematically aid in (re-)connecting patients with addiction or psychiatric problems with agencies specialized in the treatment of these disorders. Forensic physicians that visit these police detainees might be of value in this process. As our current study is preliminary, it should be exceeded by further research. In order to obtain sight on mortality risks of former police detainees and a proper estimate of risk factors for premature death in this group, it should be explored if police registries could be linked to registries of the general population as done in studies on ex-prisoners.22 Moreover, it should be investigated if an active role of forensic physicians in initializing follow-up care during police custody has the potential to reduce re-offending in former police detainees.
8. Limitations and strengths First, in interpreting the results of our study it is important to avoid generalizing the results to the entire police detainee population, because this study focuses solely on detainees that received medical care during police custody. Previous research shows that approximately 28% of all Amsterdam police detainees receive care during detainment,7 in other parts of the country rates of 31–38% were previously reported.32 Systematic differences exist between those receiving medical care during police custody and those who do not. For example, police detainees that are visited by a forensic physician are older than the average police detainee.10 Furthermore, consulting a forensic physician indicates the presence of health problems, which likely is associated with a higher risk of death. Second, only deaths that were examined by the forensic physicians of the PHS Amsterdam were included. As a 4
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
Conflicts of interest
16. Joukamaa M. The mortality of released Finnish prisoners; a 7 year follow-up study of the WATTU project. Forensic Sci Int. 1998;96(1):11–19. 17. Kjelsberg E, Laake P. Is the high mortality risk in sentenced offenders independent of previous imprisonment? Eur J Epidemiol. 2010;25(4):237–243. 18. Nieuwbeerta P, Piquero AR. Mortality rates and causes of death of convicted Dutch criminals 25 years later. J Res Crime Delinquen. 2008;45(3):256–286. 19. Paanila J, Hakola P, Tiihonen J. Mortality among habitually violent offenders. Forensic Sci Int. 1999;100(3):187–191. 20. Stenbacka M, Moberg T, Romelsjo A, Jokinen J. Mortality and causes of death among violent offenders and victims–a Swedish population based longitudinal study. BMC Publ Health. 2012;12:38. 21. Binswanger IA, Stern MF, Deyo RA, et al. Release from prison–a high risk of death for former inmates. N Engl J Med. 2007;356(2):157–165. 22. Graham L, Fischbacher CM, Stockton D, Fraser A, Fleming M, Greig K. Understanding extreme mortality among prisoners: a national cohort study in Scotland using data linkage. Eur J Publ Health. 2015;25(5):879–885. 23. Fazel S, Benning R. Natural deaths in male prisoners: a 20-year mortality study. Eur J Publ Health. 2006;16(4):441–444. 24. Sailas ES, Feodoroff B, Lindberg NC, Virkkunen ME, Sund R, Wahlbeck K. The mortality of young offenders sentenced to prison and its association with psychiatric disorders: a register study. Eur J Publ Health. 2006;16(2):193–197. 25. Kinner SA, Forsyth S, Williams G. Systematic review of record linkage studies of mortality in ex-prisoners: why (good) methods matter. Addiction. 2013;108(1):38–49. 26. Kariminia A, Law MG, Butler TG, et al. Factors associated with mortality in a cohort of Australian prisoners. Eur J Epidemiol. 2007;22(7):417–428. 27. Buster M, Dorn T, Ceelen M, Das K. Detainees in Amsterdam, a target population of the public mental health system? J Forensic Leg Med. 2014;25:55–59. 28. Driscoll TR, Harrison JA, Steenkamp M. Review of the role of alcohol in drowning associated with recreational aquatic activity. Inj Prev. 2004;10(2):107–113. 29. Howland J, Hingson R. Alcohol as a risk factor for drownings: a review of the literature (1950–1985)∗. Accid Anal Prev. 1988;20(1):19–25. 30. Death rates related to gender and age at death 2015 [Internet]. Municipality of Amsterdam; Research, Information and Statistics agency (OIS). Available from: http://www.ois.amsterdam.nl/feiten-en-cijfers/#geboorte-en-sterfte. 31. Death: deathcauses, age and gender. [Internet]. Statistics Netherlands. 2011; 2011. Available from: http://statline.cbs.nl/Statweb/publication/?DM=SLNL&PA=7052_ 95&D1=a&D2=a&D3=0&D4=0,41,51,61,l&HDR=G2,G1,G3&STB=T&VW=T. 32. Blaauw E, Lulf RE. Police custody detainees and forensic medicine: use of medical resources in the cell block. Med Law. 1999;18(4):645–651. 33. Dorn T, Ceelen M, Reijnders U, Das K. Medical death certification by forensic physicians in The Netherlands: validity and interdoctorvariation. J Forensic Leg Med. 2016;43:61–69. 34. Zlodre J, Fazel S. All-cause and external mortality in released prisoners: systematic review and meta-analysis. Am J Publ Health. 2012;102(12):67–75.
None declared. Acknowledgment This study was financed by the Public Health Service Amsterdam. References 1. Ogloff J, Warren L, Tye C, Blaher F, Thomas S. Psychiatric symptoms and histories among people detained in police cells. Soc Psychiatr Psychiatr Epidemiol. 2011;46(9):871–880. 2. Heffernan EB, Finn J, Saunders JB, Byrne G. Substance-use disorders and psychological distress among police arrestees. Med J Aust. 2003;179(8):408–411. 3. Clement R, Gerardin M, Vigneau Victorri C, Guigand G, Wainstein L, Jolliet P. Medical, social, and law characteristics of intoxicant's users medically examined in police custody. J Forensic Leg Med. 2013;20(8):1083–1086. 4. Payne-James JJ, Green PG, Green N, McLachlan GM, Munro MH, Moore TC. Healthcare issues of detainees in police custody in London, UK. J Forensic Leg Med. 2010;17(1):11–17. 5. Payne-James JJ, Wall IJ, Bailey C. Patterns of illicit drug use of prisoners in police custody in London, UK. J Clin Forensic Med. 2005;12(4):196–198. 6. McKinnon I, Grubin D. Health screening in police custody. J Forensic Leg Med. 2010;17(4):209–212. 7. Dorn T, Ceelen M, Buster M, Stirbu I, Donker G, Das K. Mental health and health-care use of detainees in police custody. J Forensic Leg Med. 2014;26:24–28. 8. Dorn T, Ceelen M, Buster M, Das K. Screening for mental illness among persons in Amsterdam police custody. Psychiatr Serv. 2013;64(10):1047–1050. 9. Fassaert T, Segeren M, Grimbergen C, Tuinebreijer W, de Wit M. Violent offenders as a target population for public mental health care. J Forensic Leg Med. 2016;40:54–57. 10. Ceelen M, Dorn T, Buster M, Stirbu I, Donker G, Das K. Health-care issues and healthcare use among detainees in police custody. J Forensic Leg Med. 2012;19(6):324–331. 11. Fazel S, Baillargeon J. The health of prisoners. Lancet. 2011;377(9769):956–965. 12. Fazel S, Danesh J. Serious mental disorder in 23000 prisoners: a systematic review of 62 surveys. Lancet. 2002;359(9306):545–550. 13. Wilper AP, Woolhandler S, Boyd JW, et al. The health and health care of US prisoners: results of a nationwide survey. Am J Publ Health. 2009;99(4):666–672. 14. Dirkzwager A, Nieuwbeerta P, Fiselier J. Onbedoelde gevolgen van vrijheidsstraffen. Een literatuurstudie. Tijdschr Criminol. 2009;51:21. 15. Laub JH, Vaillant GE. Delinquency and mortality: a 50-year follow-up study of 1,000 delinquent and nondelinquent boys. Am J Psychiatr. January 2000;157(1):96–102.
5
Contents lists available at ScienceDirect
Journal of Forensic and Legal Medicine journal homepage: www.elsevier.com/locate/yjflm
A record linkage study on former police detainees who died in Amsterdam between 2013 and 2015 Daniël B. Smidta,∗, Tina Dornb, Udo J.L. Reijndersc a
Parnassia Groep, The Netherlands Amsterdam Public Health Service, Department of Epidemiology and Health Promotion, Amsterdam, The Netherlands c Amsterdam Public Health Service, Department of Forensic Medicine, Amsterdam, The Netherlands b
A R T I C L E I N F O
A B S T R A C T
Keywords: Forensic medicine Forensic physician Post-mortem examination Police detainees Unnatural death
Background: Police detainees are known to have inferior health. This study identifies the number of former police detainees who received medical care among deaths examined by forensic physicians and presents their death characteristics. Methods: We included all deaths that were examined by forensic physicians of the Public Health Service Amsterdam from 2013 to 2015. Patient files of subjects were scanned for the presence of a prior medical consultation in the police cell and death characteristics were collected from post-mortem examination reports. We performed statistical analyses to discover what characteristics at post-mortem examination were associated with a prior consultation in the police cell. Results: We identified n = 2618 subjects that met the inclusion criteria. Eight percent of subjects had one or more medical consultation(s) in the police cell in a mean follow up time of 4.8 ( ± 3.0) years. No difference was found in the share of unnatural deaths between subjects with and without a prior consultation (68%), but distribution of death causes differed significantly. Male gender OR 2.3 (p < 0.001), age OR 0.98 (p < 0.001), unspecified unnatural dead OR 1.8 (p = 0.002), crime related dead OR 2.2 (p = 0.012) and accidental drowning and submerging death OR 4.6 (p < 0.001) were independently associated with the presence of an earlier consultation in the police cell. Conclusion: Our data suggest that a small percentage of police detainees seen by forensic physicians for provision of medical care are also examined after death by these physicians, typically young males who seem to display risk-taking and criminal behavior resulting in unnatural dead.
1. Introduction The health status of police detainees has been subject of international research in the past decade indicating high prevalence of psychiatric symptoms and substance use.1–6 In Amsterdam, approximately half of all police detainees who had received medical care while in custody were diagnosed with psychological problems and a large share of prescribed medication in this setting was for treatment of addictive disorders.7 Nearly 40% of Amsterdam police detainees submitted to screening qualified for a mental health care assessment.8 Furthermore, up to one third of young violent offenders studied in Amsterdam were eligible for public mental health care.9 Next to poor mental health, more chronic somatic conditions were seen among police detainees in Amsterdam when compared to the general population.10 Although there is a growing body of research on health issues in police detainees, studies on mortality in former police detainees are
∗
absent to date. More extensive research on this subject has been done in prisoner populations. Prisoners, however, differ from police detainees because a large part of police detainees will not be sentenced to prison, but will be released following initial police contact. Moreover, in many countries, people with mental illness who are in need of psychiatric emergency care are temporarily brought to a police station, awaiting further assessment by mental healthcare professionals. Studies in prisoner populations, mostly executed in the US, have shown that prisoners alike police detainees, have significantly more health problems (psychiatric and somatic) when compared to the general population.11–13 A Dutch review, largely based on Dutch studies confirms the higher prevalence of psychiatric and somatic complaints in (released) inmates.14 Studies that assess the risk of death in (released) prisoners compared to the general population are widely available.15–25 A systematic review by Kinner, Forsyth and Williams of 29 record linkage studies on mortality in ex-prisoners revealed a pooled all-cause standardized mortality
Corresponding author. E-mail address: [email protected] (D.B. Smidt).
https://doi.org/10.1016/j.jflm.2018.02.004 Received 5 March 2017; Received in revised form 2 February 2018; Accepted 5 February 2018 1752-928X/ © 2018 Published by Elsevier Ltd.
Please cite this article as: Smidt, D.B., Journal of Forensic and Legal Medicine (2018), https://doi.org/10.1016/j.jflm.2018.02.004
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
examination or after death. The following variables were automatically retrieved from the postmortem examination reports of all subjects: Age at death, gender, manner of death, cause of death, reporting caller, and site of the body on examination. All data were anonymized and imported into the IBM SPSS statistics® program.
ratio (SMR) of 1.7 (95%CI 0.6–5.0) (25 studies). Pooled SMRs were 8.2 for suicide deaths (3 studies) and 8.4 for unnatural deaths (2 studies) in a mean follow up time varying from 12 weeks to 16.1 years. Follow up studies of 25–50 years in the Netherlands, Sweden and the USA in (ex) crime and/or violent offenders revealed comparable higher risks of death (1.6–2.0) when compared to the general population.15,18,20 This risk increased further with recidivistic criminal offences and alcohol abuse.20 Suicide deaths were more common among repeat violence offenders, violence offenders with a history of receiving psychiatric inpatient care and violence offenders with a history of substance abuse.20,26 As both prisoner and police detainee populations overlap, both populations might share comparable risk factors for ill-health and premature death. We performed the current study in order to shed more light on mortality in former police detainees to explore the possible role for forensic physician guided preventive interventions. In this study we attempt to answer the following questions:
3. Statistical analysis Statistical analyses were executed with the IBM SPSS Statistics® program. Comparative analyses were made to assess differences in findings at post-mortem examination between both groups. For categorical variables effect sizes were calculated as odds ratios. For age (continuous variable) the mean difference was calculated as effect size. Statistical tests were used to assess significance of differences found. For age, an independent sample's t-test was used. For categorical variables we used Pearson chi-squared tests. Assumptions for valid testing were confirmed before analyses. Variables that differed significantly in the univariate analysis were assessed in a binary logistic regression model (enter method), with prior consultation (yes/no) as outcome variable. Effects sizes were calculated as odds ratio's.
1. What percentage of all individuals examined post-mortem by forensic physicians of the Public Health Service Amsterdam have a history of receiving medical care in the police cell? 2. Do individuals examined post-mortem by forensic physicians with a history of receiving medical care in the police cell differ from individuals without such a history with regards to manner of death, cause of death, age, and gender as reported at the post-mortem examination?
4. Results We identified 2650 post-mortem examination reports that met all criteria by using the electronic patient register to search for eligible subjects. We excluded 1575 cases of physician-assisted death. In our final analyses, we included 2618 reports. Reasons for exclusion were: the presence of a second post-mortem examination report on the same case or not being able to identify a full post-mortem examination report.
2. Methods 2.1. Setting In the Netherlands, forensic physicians provide medial care in the police cell. This care is comparable to the care provided by a general practitioner. The forensic physician needs to decide whether a detainee can stay in the cell from a medical point of view. If not, hospital referral or referrals to psychiatric emergency services will be organized by the forensic physician. Another major task of forensic physicians in the Netherlands is the external examination of all (potentially) unnatural deaths. If the forensic physician concludes an unnatural death, the case is discussed with the prosecutor who might order further police investigations and/or a medicolegal autopsy by a forensic pathologist.
5. General characteristics Mean age ( ± SD) at death of subjects was 63 ( ± 21) years, but differed significantly (p < 0.001, t-test) between male 59 ( ± 20) and female subjects 70 ( ± 20) (data not shown). The majority of 2618 subjects was male (66%). In 68% of all deaths referred to the forensic physician, the manner of death was confirmed as unnatural after the post-mortem examination. The most common cause of an unnatural death was a falling accident (23%), which showed an association with high age at death (mean age 82 ± 13; data not shown). The majority of natural deaths examined by the forensic physician cannot be assigned a specific cause of death (44%) after the post-mortem examination. Most deaths (53%) were signaled to the forensic physician by the police force. Deaths referred by the police were assigned an unnatural manner of death by the forensic physician in 55% of cases. Deaths referred by general practitioners were assigned an unnatural manner of death in 61% (data not shown).
2.2. Design A retrospective cohort was identified by searching the electronic patient register of the forensic medical department of the PHS Amsterdam. This register contains records of all activities executed by forensic physicians of the department starting 01-01-2005. Record linkage was based on unique identifiers. Data sources that have been linked concern the module on medical care and the module on postmortem examinations (both are two modules within the same database). All individuals who were examined post-mortem by forensic physicians between 01 and 01-2013 and 31-12-2015 were included as subjects to provide the most recent sample of post-mortem examinations executed by forensic physicians in the Amsterdam area. Only cases of physician-assisted death in this period were excluded. In the case of physician assisted death, the forensic physician monitors whether the physician who assisted in the death acted according to the procedures specified by Dutch law. The outcome of interest was the presence of one or more prior care contacts in the police cell in the catchment area of the PHS Amsterdam between 01 and 01-2005 and the date of the post-mortem examination. This design resulted in different retrospective intervals for included subjects. A prior care contact was defined as the presence of at least one contact of care in the electronic patient file, besides the post-mortem examination itself and contacts on the same day of the post-mortem
6. Post-mortem examination findings of ex police detainees Of 2618 individuals, 213 (8%) had one or more prior consultations in the police cell by a forensic physician before death. Age at death was lower in the group that received medical care in the police cell 50 ( ± 14) with a mean difference of 14.5 (11.7–17.4) years (p < 0.001) and subjects with a prior consultation were more often male when compared to subjects without prior history, OR 3.5 (2.4–5.3) (p < 0.001). No difference was found in overall manner of death between individuals with and without a prior consultation (68%). Table 1 shows, however, that distribution of unnatural causes of death differed between both groups. A high incidence of accidental drowning and submerging deaths was found in the prior consultation population 5.2% 2
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
Table 1 General characteristics of all post-mortem examinations executed by forensic physicians of the PHS Amsterdam in a 3 year period (2013–2015). Including a comparative analysis of characteristics at death between subjects with a prior consultation in the police cell (n = 213) and those without (n = 2405).
Denominator of percentages in column Age (mean in years)(SD) Male (%) Unnatural manner of death (%) Unnatural death causes (%) Accidents falling accidentsa traffic accidents accidental drowning and submerging accidental obstruction of airway Occupational accidents rest Suicide Doubt about natural manner of deathb Crime Natural death causes (%) Unknown Cardiovascular diseases Diseases of the digestive tract Rest Reporting caller (%) Police Location of body (%) Home
All subjects
Prior consultation +
Prior consultation -
OR
95% CI
2618 63 ( ± 21) 66% 68%
213 50 ( ± 14) 86% 68%
2405 64 ( ± 21) 64% 68%
* 3.5* 1.0
2.4–5.3 0.7–1.3
35% 23% 4.3% 1.6% 1.1% 0.6% 4.1% 16% 13% 2.9%
18% 2.8% 3.3% 5.2% 0.9% 0.5% 0.5% 17% 24% 8.0%
36% 25% 4.4% 1.2% 1.2% 0.6% 4.0% 16% 12% 2.5%
0.38* 0.1* 0.7 4.3* 0.8 0.8 1.3 1.1 2.3* 3.4*
0.3–0.5 0.0–0.2 0.3–1.6 2.1–8.7 0.2–3.4 0.1–5.7 0.7–2.5 0.8–1.6 1.6–3.2 1.9–5.9
14% 12% 2.2% 4.2%
5% 8.5% 4.2% 5.2%
14% 12% 2.0% 4.1%
1.0 1.5 2.2* 1.3
0.7–1.5 0.4–1.1 1.0–4.5 0.7–2.4
53%
77%
51%
3.1*
2.2–4.3
42%
52%
41%
1.5*
1.2–2.0
*p < 0.05. a Falling accidents: falls caused by slipping and stumbling, falls on stairs and in staircases, falls from ladders, falls from scaffolding, falls from or through a building or construction, falls from trees, other kind of falls (non-occupational accidents). b Doubt about natural manner of death: This category is a conglomerate of the ICD-10 codes belonging to the category ‘Ill-defined and unknown causes of mortality (R95-R99)’, more specifically code R96 ‘Other sudden death, cause unknown’. These ill-defined cases are categorized as unnatural deaths because the physician cannot be ‘convinced of a natural cause’ and the Dutch law states that only in case of no doubt about an unnatural manner of death a natural manner of death can proclaimed.
(11/213), compared to individuals without a prior consultation 1.2% (30/2405), p < 0.001. Nine out of eleven (82%) individuals with a prior consultation who drowned were known substance abusers (mostly alcohol) and strong indications existed at post-mortem examination that substance use played a direct role in the accidental drowning or submerging. In individuals without a prior consultation who drowned, 11 out of 30 (37%) deaths were suspected of having a relation with substance abuse. Crime related deaths and unnatural deaths with an unknown cause were more common in subjects with a prior consultation, OR: 3.4 (1.9–5.9) and 2.3 (1.6–3.2) respectively, both p < 0.001. Mean age of crime related deaths in the prior consultation group was 32 ( ± 9) versus 43 ( ± 18) in the group without prior consultation. No difference in number of suicide deaths was found. The amount of accidents as cause of death was higher in the group without prior consultation (36% versus 18%). This difference was largely caused by the difference in falling accidents, 25% in the group with a prior consultation versus 2.8% in the group with no prior consultation. Distribution of natural causes of death was comparable in both groups, although deaths related to digestive tract disorders were more common in the group with a prior history of consultation, OR 2.2 (1.0–4.5), p = 0.033. In the multivariate analysis (Table 2), accidental drowning showed the largest association with a prior consultation in the police cell (OR 4.6). Odds for dying of an unknown unnatural death or of crime related causes were also correlated with an almost 2 times increase (1.8 and 2.2 respectively) in the prior consultation group. Lower age and male gender remained correlated with a prior history of consultation too. Natural death caused by a digestive tract disorder lost significant association when adjusted for police as referrer, possibly due to the close relationship between the variables referrer and manner of death. Difference in accidents as cause of an unnatural death lost significance when adjusted for age, probably because mostly old people died from accidental causes in our sample (fractures, falls).
Table 2 Multivariate analysis of characteristics found at death in relation to prior consultation in the police cell. Variablea
Odds ratio
95% CI
Gender (male = 1, female = 0) Age Cause of deatha Unnatural: accidents Unnatural: doubt about natural manner of death Unnatural: crime Unnatural: accidental drowning and submerging Natural: digestive tract Reporting caller: police (police = 1, rest = 0) Location of body: home (home = 1, rest = 0)
2.3* 0.98*
1.5–3.4 0.98–0.99
0.76 1.8* 2.2* 4.6* 2.0 1.6* 1.2
0.5–1.3 1.3–2.7 1.2–4.1 2.0–10.5 0.96–4.3 1.1–2.4 0.84–1.7
*p < 0.05. a Cause of death was modelled using 5 dummy variables, the reference category thus consists of all deaths not belonging to these 5 categories.
7. Discussion We found that 8% of 2618 subjects examined post-mortem by forensic physicians in the Amsterdam region, were known to the forensic medical department as a result of an earlier medical consultation in the police cell. Death in individuals who had received care in the police cell was significantly associated with accidental drowning, death from crime and unnatural death with no specific cause. This is to expect as former detainees have a higher risk of being involved in crime-related activities and crime-related death later in life. Moreover, those with prior care contacts in the police cell were more often male and mean age at death was significantly lower when compared to those without prior contacts. A difference of 14 years in mean age at death was observed between subjects with and without prior consultation in the police cell. Studies that reported age at death in former offenders 3
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
consequence, former police detainees from the Amsterdam region may have been missed as they may have died elsewhere. The other way around goes too, individuals may have died in Amsterdam but may have been seen in a police cell outside the catchment area of the PHS Amsterdam. Third, the retrospective design of this study does not allow for the calculation of (standardized) mortality rates and we cannot conclude that overall mortality in former police detainees is higher when compared to the general population. Fourth, we used a limited retrospective interval for identifying contacts with a forensic physician in the police cell. It is likely that if we had used a longer retrospective interval, the number of subjects with a prior care contact in the police cell would have been higher. This could have resulted in a shift in distribution of death causes in both groups in either way. Fifth, the data used in this study were not primarily recorded for research but for general medical and law related purposes. Information on risk factors such as medical care history, substance use, socio-economic status and lifestyle factors was not available. In previous research, relationships between the nature and amount of crimes committed or the number of imprisonments and the risk of dying of unnatural causes were found,17,18,22,26 we could not adjust for these factors in this study. Finally different physicians do not always arrive at the same conclusions concerning manner and cause of death. The inter-physician variation between forensic physicians can be considerable.33 This inter-physician variation probably applies for both groups (with and without prior consultations), however, so no systematic bias is expected regarding this aspect. A strength of this retrospective cohort study is that we used postmortem examination reports as our main data resource. These reports provide structured and comparable information for all subjects because a standardized format is used for reporting. We included all reports in the catchment area of the PHS Amsterdam and due to the exclusive right of this department to perform the post-mortem examinations, we can provide a complete and representative overview of deaths examined in this area. To conclude, this study adds to the body of current literature on mortality in former police detainees.
showed the same tendency, although observed ages at death were lower.15,20 The large majority of subjects with prior care contacts in the police cell before death (86%) was male. This is comparable with studies in police detainees in Amsterdam were an average 89% was male8,10,27 and studies in prisoner populations demonstrating high male to female ratios.18,21,22 Distribution of unnatural causes of death in both groups was distinctly different. Subjects with a prior consultation died from a crime related death more frequently (OR 3.4). This association is supported by prior research on causes of death in a cohort of convicted Dutch criminals, that showed a risk ratio of 5.9 for dying of crime related causes in a 25 year follow up.18 However, methodological differences exist between both studies limiting comparability. Firstly, our prior consultation group was not referenced to the general population but to those without a prior consultation examined post-mortem by forensic physicians. Secondly, we used a shorter follow up interval and a retrospective view. Thirdly, we used post-mortem examination reports as our primary inclusion method. Furthermore, we found that an accidental drowning death was significantly associated with a prior consultation in the police cell. When reviewing the cases of former police detainees who drowned, there seemed to be a high coincidence with a history of substance abuse. This association is also found by other studies.28,29 Although previous studies did show higher number of suicide deaths in ex-prisoners and law offenders when compared to the general population,16,18,21,25 we found no difference in number of suicides between our groups. Differences in results between our study and previous research might be explained by the fact that former study populations were younger,16,18 suicides were more common among younger subjects21,26 and subjects were convicted criminals instead of (ex) police detainees.16,18,21,26 The risk of suicide in ex-offenders was correlated to the number of offences committed20,26 and the presence of psychosis or a psychiatric hospital admission during imprisonment.16,26 In this regard it is important to notice that police in Amsterdam has the option to present arrestees that appear confused as result of a psychiatric disorder to a psychiatric emergency service without prior consultation of a forensic physician. This policy might have limited the number of police detainees with psychosis seen by forensic physicians. Secondly, no information on the criminal records of included subjects was retrieved for this study. In this study we have compared post-mortem examination findings in former police detainees who received care in the police cell to individuals without such a history. The deaths examined post-mortem by the forensic physicians differ from deaths in the general population, however. In 2015 only 20% of 5425 deaths in the general Amsterdam population were aged under 64 years at death,30 while this was 46% in our sample of 935 deaths from 2015. The share of males was also higher in our sample than in the general population, 66% versus 49%.30 Moreover, the amount of unnatural deaths in our sample is almost 16 times (68/4.3) higher than in the general population.31
9. Implications What remains is the question if there is a group in police custody that is prone to inferior health and premature death because of risk factors that may be influenced, e.g. by offering special care aiming at reducing re-offending or interventions to reduce risk-taking behavior in this group. Reasons for premature death are multifaceted and often interrelated with other vulnerabilities such as low income, low educational level, psychiatric and addiction disorders. It is unknown how many police detainees will be sentenced and incarcerated later in life, but is known that released prisoners are at increased risk for death following release from prison. This emphasizes the need for aftercare planning after release and the need to identify groups that should be targeted to reduce premature death.34 The share of police detainees that commit relatively low grad but often troublesome offences and often do not qualify for imprisonment could possibly benefit from targeted measures. Ideally, forensic physicians systematically aid in (re-)connecting patients with addiction or psychiatric problems with agencies specialized in the treatment of these disorders. Forensic physicians that visit these police detainees might be of value in this process. As our current study is preliminary, it should be exceeded by further research. In order to obtain sight on mortality risks of former police detainees and a proper estimate of risk factors for premature death in this group, it should be explored if police registries could be linked to registries of the general population as done in studies on ex-prisoners.22 Moreover, it should be investigated if an active role of forensic physicians in initializing follow-up care during police custody has the potential to reduce re-offending in former police detainees.
8. Limitations and strengths First, in interpreting the results of our study it is important to avoid generalizing the results to the entire police detainee population, because this study focuses solely on detainees that received medical care during police custody. Previous research shows that approximately 28% of all Amsterdam police detainees receive care during detainment,7 in other parts of the country rates of 31–38% were previously reported.32 Systematic differences exist between those receiving medical care during police custody and those who do not. For example, police detainees that are visited by a forensic physician are older than the average police detainee.10 Furthermore, consulting a forensic physician indicates the presence of health problems, which likely is associated with a higher risk of death. Second, only deaths that were examined by the forensic physicians of the PHS Amsterdam were included. As a 4
Journal of Forensic and Legal Medicine xxx (xxxx) xxx–xxx
D.B. Smidt et al.
Conflicts of interest
16. Joukamaa M. The mortality of released Finnish prisoners; a 7 year follow-up study of the WATTU project. Forensic Sci Int. 1998;96(1):11–19. 17. Kjelsberg E, Laake P. Is the high mortality risk in sentenced offenders independent of previous imprisonment? Eur J Epidemiol. 2010;25(4):237–243. 18. Nieuwbeerta P, Piquero AR. Mortality rates and causes of death of convicted Dutch criminals 25 years later. J Res Crime Delinquen. 2008;45(3):256–286. 19. Paanila J, Hakola P, Tiihonen J. Mortality among habitually violent offenders. Forensic Sci Int. 1999;100(3):187–191. 20. Stenbacka M, Moberg T, Romelsjo A, Jokinen J. Mortality and causes of death among violent offenders and victims–a Swedish population based longitudinal study. BMC Publ Health. 2012;12:38. 21. Binswanger IA, Stern MF, Deyo RA, et al. Release from prison–a high risk of death for former inmates. N Engl J Med. 2007;356(2):157–165. 22. Graham L, Fischbacher CM, Stockton D, Fraser A, Fleming M, Greig K. Understanding extreme mortality among prisoners: a national cohort study in Scotland using data linkage. Eur J Publ Health. 2015;25(5):879–885. 23. Fazel S, Benning R. Natural deaths in male prisoners: a 20-year mortality study. Eur J Publ Health. 2006;16(4):441–444. 24. Sailas ES, Feodoroff B, Lindberg NC, Virkkunen ME, Sund R, Wahlbeck K. The mortality of young offenders sentenced to prison and its association with psychiatric disorders: a register study. Eur J Publ Health. 2006;16(2):193–197. 25. Kinner SA, Forsyth S, Williams G. Systematic review of record linkage studies of mortality in ex-prisoners: why (good) methods matter. Addiction. 2013;108(1):38–49. 26. Kariminia A, Law MG, Butler TG, et al. Factors associated with mortality in a cohort of Australian prisoners. Eur J Epidemiol. 2007;22(7):417–428. 27. Buster M, Dorn T, Ceelen M, Das K. Detainees in Amsterdam, a target population of the public mental health system? J Forensic Leg Med. 2014;25:55–59. 28. Driscoll TR, Harrison JA, Steenkamp M. Review of the role of alcohol in drowning associated with recreational aquatic activity. Inj Prev. 2004;10(2):107–113. 29. Howland J, Hingson R. Alcohol as a risk factor for drownings: a review of the literature (1950–1985)∗. Accid Anal Prev. 1988;20(1):19–25. 30. Death rates related to gender and age at death 2015 [Internet]. Municipality of Amsterdam; Research, Information and Statistics agency (OIS). Available from: http://www.ois.amsterdam.nl/feiten-en-cijfers/#geboorte-en-sterfte. 31. Death: deathcauses, age and gender. [Internet]. Statistics Netherlands. 2011; 2011. Available from: http://statline.cbs.nl/Statweb/publication/?DM=SLNL&PA=7052_ 95&D1=a&D2=a&D3=0&D4=0,41,51,61,l&HDR=G2,G1,G3&STB=T&VW=T. 32. Blaauw E, Lulf RE. Police custody detainees and forensic medicine: use of medical resources in the cell block. Med Law. 1999;18(4):645–651. 33. Dorn T, Ceelen M, Reijnders U, Das K. Medical death certification by forensic physicians in The Netherlands: validity and interdoctorvariation. J Forensic Leg Med. 2016;43:61–69. 34. Zlodre J, Fazel S. All-cause and external mortality in released prisoners: systematic review and meta-analysis. Am J Publ Health. 2012;102(12):67–75.
None declared. Acknowledgment This study was financed by the Public Health Service Amsterdam. References 1. Ogloff J, Warren L, Tye C, Blaher F, Thomas S. Psychiatric symptoms and histories among people detained in police cells. Soc Psychiatr Psychiatr Epidemiol. 2011;46(9):871–880. 2. Heffernan EB, Finn J, Saunders JB, Byrne G. Substance-use disorders and psychological distress among police arrestees. Med J Aust. 2003;179(8):408–411. 3. Clement R, Gerardin M, Vigneau Victorri C, Guigand G, Wainstein L, Jolliet P. Medical, social, and law characteristics of intoxicant's users medically examined in police custody. J Forensic Leg Med. 2013;20(8):1083–1086. 4. Payne-James JJ, Green PG, Green N, McLachlan GM, Munro MH, Moore TC. Healthcare issues of detainees in police custody in London, UK. J Forensic Leg Med. 2010;17(1):11–17. 5. Payne-James JJ, Wall IJ, Bailey C. Patterns of illicit drug use of prisoners in police custody in London, UK. J Clin Forensic Med. 2005;12(4):196–198. 6. McKinnon I, Grubin D. Health screening in police custody. J Forensic Leg Med. 2010;17(4):209–212. 7. Dorn T, Ceelen M, Buster M, Stirbu I, Donker G, Das K. Mental health and health-care use of detainees in police custody. J Forensic Leg Med. 2014;26:24–28. 8. Dorn T, Ceelen M, Buster M, Das K. Screening for mental illness among persons in Amsterdam police custody. Psychiatr Serv. 2013;64(10):1047–1050. 9. Fassaert T, Segeren M, Grimbergen C, Tuinebreijer W, de Wit M. Violent offenders as a target population for public mental health care. J Forensic Leg Med. 2016;40:54–57. 10. Ceelen M, Dorn T, Buster M, Stirbu I, Donker G, Das K. Health-care issues and healthcare use among detainees in police custody. J Forensic Leg Med. 2012;19(6):324–331. 11. Fazel S, Baillargeon J. The health of prisoners. Lancet. 2011;377(9769):956–965. 12. Fazel S, Danesh J. Serious mental disorder in 23000 prisoners: a systematic review of 62 surveys. Lancet. 2002;359(9306):545–550. 13. Wilper AP, Woolhandler S, Boyd JW, et al. The health and health care of US prisoners: results of a nationwide survey. Am J Publ Health. 2009;99(4):666–672. 14. Dirkzwager A, Nieuwbeerta P, Fiselier J. Onbedoelde gevolgen van vrijheidsstraffen. Een literatuurstudie. Tijdschr Criminol. 2009;51:21. 15. Laub JH, Vaillant GE. Delinquency and mortality: a 50-year follow-up study of 1,000 delinquent and nondelinquent boys. Am J Psychiatr. January 2000;157(1):96–102.
5