Optimizing the Current Executed NFiDENT Process
A New Process for Illicit Drug Identification in the Netherlands
Lydia Jochems, University of Amsterdam, Master Research Project
Name Lydia Jochems
Student ID 11387971
EC 36
Period 02.01.2018 – 13.07.2018
Education University of Amsterdam
Program MSc Forensic Science
Supervisors MSc Jerien Koopman
MSc Saskia Verheij
Examiner Dhr. dr. Radboud Winkels
Institute Netherlands Forensic Institute
(NFI) Date of submission 13.07.2018
2
AbstractThe NFiDENT process was developed to accelerate the conventional process of illicit drug identification in the Netherlands for the four most common drug types (MDMA, amphetamine, heroin and cocaine). This novel process was developed by the NFI in cooperation with the National Police and the Public Prosecution Service. It is designated to accelerate the turnaround time of the total process for illicit drug cases. Currently the process is running at four police units where the process is performed in practice for the first time. This study analyzes the process as it is executed currently to determine the bottlenecks in the process. Based on the identified bottlenecks, recommendations were provided for optimization. The performance was determined qualitatively, by interviewing the involved personnel about the execution of the process and the experiences. The performance was determined quantitatively by analyzing the turnaround time of different intervals of the process. The quantitative and qualitative data was combined to gain insight into the full process. It was observed that the new steps, introduced by the NFiDENT process, were executed very well at the forensic units of the police. However, the more general processes at the forensic units obstructed the performance of the NFiDENT process as a whole. The NFI showed a great performance concerning the turnaround time. However, the compliance of agreements was lower than expected. This was already optimized during the study but should be further optimized to form the right basis for future deployment of the NFiDENT process. The current executed NFiDENT process can be optimized by improving the connection of the more general processes to the NFiDENT part of the process. This improvement involves better personnel planning, more regular delivery of goods and suitable laboratory settings. But also, more communication between the three fields and sharing experiences to create awareness.
Keywords: Forensic Science, NFiDENT, Illicit Drugs Analysis, Process Optimization
Introduction
The conventional process for illicit drug identification in the Netherlands is inefficient and requires a great deal of time, transport and administration. Therefore, the National Police (NP), the Netherlands Forensic Institute (NFI) and the Public Prosecution Service (PPS) started working together on a so-called ‘co-creation’ project. The aim of this project was to develop a new process that is more efficient and accelerates the total process for illicit drug cases.
The process resulting from this project is called ‘NFiDENT’ and is currently in use at four (out of 10) police units in the Netherlands. In this process the major part of the investigation, the GC-MS analysis, has shifted (from the NFI) to the forensic unit of the police. This reduces the amount of transport and administration drastically. By using specially developed software and a secure digital connection, the experts at the NFI are able to access the analysis data remotely. In this way the experts can still assess the data and deliver an expert report. Due to the decrease in transport and administration and partial digitalization, a lot of time should be saved in the process. The aim of NFiDENT is to deliver an expert report within a day. Currently this process is only made available for cases involving the four most common drugs in the Netherlands, namely MDMA, amphetamine, heroin and cocaine [1]. In this way it should be possible to use the NFiDENT process in 80% of the total drugs cases. This results in a much faster turnaround time (TAT) for most drug cases by a more efficient process.
As previously mentioned there are currently four police units working with the NFiDENT process. These units are Den Haag, Noord-Holland, Noord-Nederland and Oost-Brabant. Den Haag and Noord-Nederland started at the beginning of 2016 (end of January) and Noord-Holland and Oost-Brabant started in 2018 (beginning January and end of March). This means they do not all have the same experience with the new process yet and an adjustment period is necessary. It is intended to be used at all police units in the Netherlands and will be implemented at the remaining units in the coming years.
The NFiDENT process is a pioneer in the field of forensic drug identification and shows a new manner of cooperation in the forensic chain. This chain cooperation is very important in the forensic field, but is also challenging due to its multidisciplinary nature. The NFiDENT process is an innovative process which makes use of new technologies. The process shows a way for the forensic field to improve itself and keep up with the times by making use of more rapid technologies. Nowadays rapid results are becoming increasingly important and the so called ‘CSI-effect’ has raised societies’ expectations of the forensic field [2]. Additionally, a faster process has shown to result in a higher chance of behavioral change in delinquents, due to the greater impact of fast punishment after the act [3]. A more efficient process also makes it possible to create more capacity for (other) investigations, which is necessary due to the increasing amount of requested analyses at the NFI. In the conventional process, the time spent on administrative and logistic steps is not in proportion to the time spent on the actual analysis of the illicit drugs [4]. Additionally, the process is not in proportion with the sentence in these cases. Therefore, there was a demand of a faster and more efficient process. A process like NFiDENT makes it possible for forensic investigations to meet the expectations and the demands of the field more and more. Therefore, it is a progressive process which sets an example for the worldwide forensic field to keep improving itself.
Although forensic investigations are important for the criminal justice system and provide a societal service, it’s performance is rarely measured [5, 6, 7]. In a novel process such as NFiDENT, is especially important to evaluate performance to compare this to the aim of the initial development. In this way the process can be improved targeted and substantiated. After some experience with and habituation to the NFiDENT process, it is important to be able to
3
optimize the current processes and in addition be prepared for further implementations. Therefore, this study evaluates the whole process and provides recommendations for optimization.The NFiDENT process can be seen as a business process in which different parties work together to deliver the service of criminal justice. Therefore, it can be assessed using business and change management approaches [8, 9, 10]. This kind of approach has been used in the forensic field before and it is coming up more and more [10, 11]. Assessments of processes in the forensic field are mentioned in the literature to be important but surprisingly not often performed [5, 6]. Whereas, by the importance of the delivered service by the forensic processes, one would expect more effort into assessment and improvement in the forensic field.
In clinical laboratories such assessments have been much more common for some time already [12, 13, 14, 15, 16]. Assessments in clinical laboratories can serve as example for the field of forensics as their aim is similar. In the medical field, as well as in the forensic field, the ‘products’ resulting from the investigations are a requirement to deliver a societal service.
In forensic context, laboratories are usually more concerned with the scientific methods and the quality of their evidence product [11]. For the NFiDENT process the quality of the analysis technique has already been examined and compared to the previously used method during a pilot. In this pilot the new process and the old process were run at the same time to find out if the resulting illicit drug identifications were identical in order to guarantee the same quality of evidence as before, as a requirement before implementation. This yielded the same results in 100% of the cases [17]. Therefore, the quality of the technology has already been established and it is now time for the performance of the process to be determined and optimized. Therefore, the main research question addressed in this study is as follows:
How can the current executed NFiDENT process be optimized?
To answer the main research question, sub questions have been established to gain more insight into the current executed process. The sub questions are a mean to ultimately answer the main research question. In this study the following sub questions are addressed:
• In what manner is the process executed?
• What is the current turnaround time performance of the process?
By answering the first sub question, insight will be provided into the current executed process. This will help to determine the bottlenecks in the process. The answer will also show if practical differences occur between the forensic units of the police. This qualitative analysis is also required to be able to interpret the quantitative results, under the second sub question correctly.
The second sub question can be answered by a quantitative analysis of turnaround times (TATs) in the process. The new process was developed mainly to reduce the TAT of the process, which is why it is important to analyze the current TAT. Additionally, the data can be used as reference in future analyses of improvement. It is hypothesized that the TAT analysis will show that the NFiDENT process is significantly faster than the conventional process in which the expert report is delivered within a day and according to the agreements. Additionally, it is hypothesized that the TAT improved over time due to habituation to the process.
The results provide qualitative as well as quantitative information on the current executed NFiDENT process. This combination is used to ultimately answer the main research question and provide recommendations for the future. The initial approach was to investigate the process at all three parties involved (police, NFI and PPS). However, after some initial results the focus was more set to the process at the police and the NFI, which will be explained in the Results & Discussion section. It was also complicated to find out what processes an NFiDENT case undergoes ones it arrives at the PPS as there are a lot of different types of cases handled by different personnel. These initial results have led to an approach which is mainly focused on the processes at the forensic unit of the police and the NFI. These parties have also undergone the greatest changes due to the arrival of the new process.
Materials & Methods
The NFiDENT process is a multidisciplinary cooperation which can be roughly divided into three main steps performed by the three involved parties as shown in Figure 1. In between the steps, the product or service delivered by each party is shown.
Figure 1. Three main steps of the NFiDENT process.
Police Conducts analysis and requests expert report NFI Assesses the results and delivers the expert report. PPS Uses report in court and finishes case Analysis data Expert report Verdict
4
These steps each consist of several smaller steps. Interviews and observation days were scheduled to gain a realistic overview of the current execution of all steps and to see the process in action. This was performed at all involved parties, but mostly focused on the processes at the Police and the NFI.The information obtained during the interviews and observations was used to create stream maps showing the current execution (AS-IS) of each step in the process. Similar structured stream maps were created showing the ideal process. This ideal process was generated using a combination of the input from all three parties and the initial aim of the NFiDENT project. These stream maps visually indicated which parts of the process were not ideal yet. All interviews and observation data were documented in reports for future reference.
The three main steps shown in Figure 1, all have a specific output. The analysis results and the expert reports are outputs in the form of products (data and a report respectively), which are required to proceed the process. The verdict in court is an output in the form of a service, a societal service of criminal justice [18]. The two products are required to ultimately achieve the service of criminal justice and are therefore critical to the quality total process.
To quantify the performance of the process, a measurable indicator was determined [6]. TAT was shown to be a measure of performance of a process, which was mostly used in the clinical setting and could be applied to the forensic field as well [10, 12, 13, 14, 15, 16]. The TAT is especially important in the NFiDENT process as it was designed to accelerate the conventional process. Its current performance could therefore be indicated by measuring its current TAT.
The main steps were elaborated into six time steps which are shown in Table 1. The process starts when the police confiscates the illicit drugs and sends it to the forensic research facility of the police (T1). At the research facility a
preliminary investigation is then conducted using two indicative tests (T2). These tests indicate if the confiscated drugs
are indeed one of the four drugs that can be analyzed via the NFiDENT process. The administrative steps, between T1 and T2, are more general for all items of evidence coming in at the forensic unit of the police. These steps are not
specified in this analysis. If proceeded with the NFiDENT process, a GC-MS analysis is performed (T3). If the analysis
results are suited for an expert report request, a request is send to the NFI (T4). When the NFI receives a request, they
interpret the conclusion made by the software and the police investigator. If this meets the requirements, an expert report is send to the police (T5). When the expert report is received, the police merges the report with the case file and
sends it to the requesting authority (T6). These six time steps were used to calculate TAT intervals, which are shown
in Table 2.
Available data of these time steps, over a certain amount of time, were obtained. Unfortunately, the police had no registration system that recorded these time steps. Most police units registered steps of the process manually, which is why the available data differed per police unit. This data contained dates at which the time steps were performed. At the NFI there was an automated registration system, which made the time steps T4 and T5 easily accessible for all
request in the NFiDENT process that were ever received. This data contained dates and times at which the time steps were performed.
Time steps T3, T4 and T5 were the new steps introduced by the NFiDENT process, here the biggest changes were
implemented. However, the other time steps were still important for the process’ performance as they connected to the new steps and formed the total process together.
Table 2. Turnaround time intervals used for data analysis. Table 1. Chronological time steps of the NFiDENT process,
used to calculate the turnaround time intervals. TAT interval Description Formula
TTAT
Total turnaround time, from the date of the confiscation by the police until the case file is finalized and sent to the requesting authority
T6 – T1
TAT1
Turnaround time interval one, from the date of the confiscation by the police until the start of the preliminary investigation
T2 – T1
TAT2
Turnaround time interval two, from the date of the start of the preliminary investigation until the start of the GC-MS analysis
T3 – T2
TAT3
Turnaround time interval three, from the date of the start of the GC-MS analysis until the expert report is received.
T5 – T3
TAT4
(NFI only)
Turnaround time interval four, from the time of receiving the report request at the NFI until the expert report is sent to the police
T5 – T4
TAT5
Turnaround time interval five, the time from receiving the expert report until the case file is finalized and sent to the requesting authority
T6 – T5 Time steps in
the process Description
T1 Confiscation by the police
T2 Start of preliminary investigation
T3 Start of GC-MS analysis
T4
Digital expert report request by the police
T5
Digital expert report sent to the police
T6 Case file finalized and sent to the
5
The intervals TAT1, TAT2, TAT3, and TAT5 took place at the forensic units of the police and TAT4 only took place at theNFI. Not all intervals were available from every police unit and for some units no data was available at all. All available data was used, to at least determine a part of the performance. When no quantitative data was available, the qualitative data was used to provide recommendations.
The data obtained from the forensic units of the police showed the submission dates for time steps per analyzed sample. This meant that there were cases registered with only one sample but also cases with 10 samples. The data was analyzed and filtered on the case number and duplicates in TAT of the same case number were taken out. This was performed to prevent duplicate TATs to weigh heavier on the average TAT. Otherwise the data would not be suitable for a realistic interpretation. The resulting data was used to determine the intervals. The intervals were used for analyses in the following manner:
1. Compare the intervals between 2017 to 2018 per unit.
The same intervals were compared between 2017 and 2018 to determine if the performance differed over time. In 2017 the process was just implemented at most units which is why it was interesting to determine if the units were getting more accustomed to the process and if this resulted in a more constant and faster process. Descriptive statistics were used to gain insight into the current situation of each of the intervals. The intervals were compared using the nonparametric Wilcoxon median difference test for related samples (at a significance level of 0,01) [19]. This test was performed to determine if significant differences occurred between the intervals at different moments.
A significant difference in TAT could indicate improvement or deterioration of the process over time, depending on the decrease or increase of the TAT respectively.
2. Analyze the correlation between TAT1, TAT2, TAT3 and TAT5 intervals per unit for 2018.
For all intervals the correlation was analyzed using Spearman’s Rho to determine if there was a relation between the intervals (at a significance level of 0,01). This showed if the process steps influenced each other and if so, to what extent. If the intervals showed a high positive correlation this meant that, when the one interval took long during one case, the other interval took long as well. This could indicate that the intervals influence each other’s turnaround time in some way. If the correlation was negative this meant that when the one interval took long, the other interval took short time. If the correlation was low this indicated that the intervals did not in influence each other and were individual processes.
3. Compare the intervals between the units for 2018.
The same intervals were compared between the units to determine if their performance differed. This was compared for 2018 to compare the most recent performance. The TAT intervals were compared using the nonparametric Mann-Whitney U test for independent samples (at a significance level of 0,01) [19]. This test was performed to determine if significant differences occurred between the intervals of the different units.
A significant difference would indicate a difference in performance which could be caused by a difference in the executed process. If a significant difference occurred the effect size was calculated.
The resulting performances of all units were also summarized in one overview table (Table 15). 4. Compare the TAT4 interval for 2017 and 2018.
The TAT4 was compared between 2017 and 2018 to determine if the performance differed over time. This was
especially interesting due to the NFiDENT software update at the start of 2018. This made it possible to produce the expert report digitally and automatically for NFiDENT. This update was supposed to decrease the TAT at the NFI and make the process less intensive. Descriptive statistics were used to gain insight into the current situation of each of the intervals. The TAT4 of 2017 and 2018 were compared using the nonparametric Wilcoxon median difference test
for related samples (at a significance level of 0,01) [19]. This test was performed to determine if a significant difference occurred between 2017 and 2018.
A significant difference would indicate improvement or deterioration of the process step, depending on the decrease or increase of TAT4 respectively.
5. Compare the TAT4 interval for two time intervals in 2018.
The TAT4 interval was compared for two time intervals in 2018 to determine if the performance improved over time.
The performance of delivering the expert reports according to the agreements for NFiDENT (between the police and the NFI) was also determined. The initial analysis results were discussed with the involved personnel and some ideas for optimization were discussed. After three months TAT4 was analyzed again to see if improvement was made. TI1 is
6
and ends at 12.06.2018. The two time intervals were compared using the nonparametric Wilcoxon median difference test for related samples (at a significance level of 0,01) [19]. This test was performed to determine if significant differences occurred between the two time intervals.A significant difference between the two intervals would indicate a difference in performance and in case of a decreased median TAT4 a succeeded optimization. If a significant difference occurred the effect size was calculated.
6. Compare the TAT4 interval for VDM017 to VDM210
The same interval was compared for the two different products of the NFI. VDM017 (NFiDENT) is supposed to replace the conventional product VDM210 for at least 80% and additionally accelerate the process of TAT4. This comparison
was performed to determine if VDM017 really is an improvement of VDM210 for the process at the NFI (not the total NFiDENT process). Because there are still police units making use of the VDM210 product, both VDM210 and VDM017 could be measured over the same period of time. This data was also used to find the ratio between the two products for the TAT4 interval, to indicate with what factor VDM017 could replace VDM210 at the NFI. The TAT4 of both products
was also compared using the nonparametric Wilcoxon median difference test for related samples (at a significance level of 0,01) [19]. This was performed to determine if a significant difference occurred between the two products. A significant difference would indicate improvement or deterioration, depending on the decrease or increase of the median TAT4 respectively.
The statistical tests were performed using IBM SPSS Statistics 25.
In case of significant difference, the effect size (r) was calculated according to Rosenthal (1991) by dividing the statistical test’s resulting Z value by the square root of N [20]. This was performed to also show the magnitude of the difference instead of solely the significance, as this tells more about the effects on the process. An effect size of 0.5 or higher was considered a large effect, between 0.5 and 0.3 a medium effect and between 0.1 and 0.3 a small effect. This scale mostly used in effect size interpretation, but arbitrarily specified by Cohen (1988) [21]. A negative effect size showed the effect of a decreased TAT and a positive effect size shows the effect of an increased TAT.
The frequency distributions were visualized in histograms as well. The same scales for the X and Y axes were used per unit, to be able to also visually compare the histograms of the different time intervals.
The results of the above described analyses, combined with the qualitative knowledge were used to provide recommendations for optimization.
7
Results & DiscussionInitial results showed that the PPS could only anticipate optimally during the NFiDENT process, if the preceding processes were more continuous and provided security of delivery. Therefore, the focus was put on the performance of the police and the NFI. The main results will now be discussed per unit.
Den Haag
At the unit den Haag there was an overall backlog which influenced the NFiDENT process as well. This caused cases to be just waiting for a long time prior to the analysis. This is a waste, as the NFiDENT process itself was going very well. If there was no backlog, the whole NFiDENT process would be able to perform better. This seems to be a real bottleneck at this unit.
The TTAT data shows a long TAT, as shown in Table 3. The TTAT significantly increased from 2017 to 2018. The variance of both intervals increased as well, showing that the TAT is more widely spread and less constant. This shows a deterioration of the process as a whole. The frequency distribution shown in Figure 2 makes this visible as well. It can be seen that the TTAT is more spread out in 2018. This shows the backlog seriously obstructs the total process of NFiDENT.
Table 3. Results of the data analysis of TTAT at the unit Den Haag for 2017 and 2018 (till 12.06.2018).
Interval Year N Median
(days) Variance (σ2) Comparison 2017 – 2018 (P value) Effect size (r) TTAT (confiscation – finish) 2017 191 42,0 426,4 0,000 +0,273 2018 135 59,5 731,2
8
The backlog also causes cases to be waiting a long time before they can be analyzed. The data results of TAT1correspond to this as well. The results are shown in Table 4. It is shown that the TAT1 interval increased from 22
days in 2017 to 56 days in 2018. This means that the time spent on processes before the start of the analysis increased. This increase is significant with a medium effect size. This shows that this interval did not improve and in fact deteriorated in 2018. This is alarming, as this causes the process to start with a great delay. As discussed previously this can be explained by the backlog at this unit. Additionally, the variance was increased, showing that TAT1 was widely spread. This influences the security of supply negatively and therefore deteriorates the
performance of the total process. The deterioration becomes visible in the frequency distribution shown in Figure 3.
Table 4. Results of the data analysis of TAT1 at the unit Den Haag for 2017 and 2018 (till 12.06.2018).
Interval Year N Median
(days) Variance (σ2) Comparison 2017 – 2018 (P value) Effect size (r) TAT1 (confiscation – preliminary investigation) 2017 176 22,0 1108,2 0,000 +0,479 2018 145 56,0 1477,4
9
The analysis consists of a preliminary investigation and a GC-MS analysis. At the unit Den Haag these processes were separated, as these analyses were not performed consecutively. This obstructs the continuity of the process by separating the two process steps T2 and T3. Fortunately, the unit is building a more suitable laboratory, this willimprove the continuity.
The TAT2 data corresponds to this as well. Table 5 shows that there were five days between the start of the
preliminary investigation and the start of the GC-MS analysis in 2017. The preliminary investigation itself is fairly easy and should not take longer than one day. Therefore, the TAT2 mostly shows the waiting time between the
steps. This waiting time was caused by the separation of the process steps. Therefore, a consecutive process would improve the process and decrease the turnaround time of TAT2. The TAT2 could then be reduced to the
same time spent on TAT3 as the steps will be combined.
Unfortunately, no data of this interval was available for 2018. Therefore, no recent data could be analyzed nor compared to 2017.
Table 5. Results of the data analysis of TAT2 at the unit Den Haag for 2017.
Interval Year N Median (days) Variance (σ2)
Comparison 2017 – 2018 (P value) Effect size (r) TAT2 (preliminary investigation – GC-MS analysis) 2017 176 5,0 18,3 N.A. N.A. 2018 N.A. N.A. N.A.
10
There were three personnel members working on NFiDENT which was not ideal, as the personnel also had other work duties besides NFiDENT of which some were in the first line of work as well (at the crime scene). This causes a lack of back up in times of expected and unexpected absence of personnel. Despite of this slight understaffing the unit was capable of performing NFiDENT almost every day. This combined with the constant delivery of cases at this unit forms some basis towards a more continuous NFiDENT process.The TAT3 data also shows that the actual NFiDENT GC-MS analysis was performed very fast, with a median TAT
of zero days in 2018. Table 6 shows that the decrease in TAT from 2017 to 2018 was significant with a large effect size. This means that the GC-MS analysis and receiving the report was performed in a shorter period of time. This is actually ideal already, as it could not be performed in less than zero days. This is a great performance given the fact that this process step was newly (compared to the conventional process) introduced by the NFiDENT process. The variance decreased as well, which shows that the TAT was less spread and became steadier in 2018. This is a sign of improvement and shows habituation to and experience with the process. Figure 5 shows the frequency distribution shifted to the right and was les widely spread in 2018. However, the variance could be more limited to provide a more continuous process, as explained before.
Table 6. Results of the data analysis of TAT3 at the unit Den Haag for 2017 and 2018 (till 12.06.2018).
Interval Year N Median
(days) Variance (σ2) Comparison 2017 – 2018 (P value) Effect size (r) TAT3 (GC-MS analysis – report request) 2017 185 3,0 15,8 0,000 -0,503 2018 132 0 12,9
11
The investigating personnel only handled the reports of the cases they analyzed themselves. This means that if this person was absent for one or more days, in which the report was delivered by the NFI, the report would be waiting to be send to the requesting authority. This is a waste of time, as it is a simple administrative step, and delays the further process unnecessarily. If all NFiDENT personnel would be able to send all incoming reports to the requesting authority, the process would be faster and more efficient. Expected and unexpected absence would then influence the TAT less and would provide more continuity for the process.Table 7 shows a decrease of TAT5 from 13 days in 2017 to one day in 2018. This decrease was significant with a
large effect size. The variance decreased as well, which show the TAT is less spread and more continuous. This shows improvement of the process. Figure 6 shows the frequency distribution shifted to the left and was less widely spread in 2018. However, the variance could still be improved as explained before.
Table 7. Results of the data analysis of TAT5 at the unit Den Haag for 2017 and 2018 (till 12.06.2018).
Interval Year N Median
(days) Variance (σ2) Comparison 2017 – 2018 (P value) Effect size (r) TAT5
(report received – casefile to requesting authority)
2017 162 13,0 451,8
0,000 -0,571
2018 134 1 17,5
12
The correlation, shown in Table 8, shows that the only significant correlation occurred between TTAT and TAT1.This means that when the time between the confiscation and the start of the preliminary investigation was long the TTAT was also long and the same holds for a short time. This corresponds with the fact that TAT1 contributes
most to the TTAT and showed the highest positive effect size. The other intervals do not show any significant correlations and the correlation coefficients are quite low. This indicates that there is no clear relation between the intervals and that these intervals are individual processes.
Table 8. Correlation coefficients between the intervals at the unit Den Haag in 2018 (significant correlations indicated with an *).
Correlation coefficients (2018)
TTAT TAT1 TAT3 TAT5
TTAT 1,000 0,992* 0,139 0,027 TAT1 0,992* 1,000 0,087 0,196
TAT3 0,139 0,087 1,000 0,099
TAT5 0,272 0,196 0,099 1,000
Overall the results of the unit Den Haag show that the specific steps for NFiDENT (TAT3) were performed very
well but that the more general police processes connecting to it, obstructed the total process. This obstruction was mainly caused by the waiting time prior to analysis (TAT1), which was caused by the backlog at this unit.
More detailed information on the practical execution and the consequences is provided in Appendix A-1 and a schematic overview of the practical process at this unit is shown in Appendix B-1.
Noord-Holland
At the unit Noord-Nederland there was a unique cooperation between the police and the PPS. Deadlines were specified per case, as not all cases demand the same turnaround time. This has been working very well for this unit and resulted in a specific prioritization and an effective application of the NFiDENT process.
The forensic unit of the police had four personnel members who worked with the NFiDENT process. NFiDENT was scheduled for three times a week, but this was not always achieved. The personnel also had other work duties, of which some were in the first line of work as well. The current personnel composition does not provide proper back up in times of expected and unexpected absence of personnel.
This unit did have a suitable laboratory setting, as the laboratory was large enough for the preliminary investigation and the GC-MS analysis to be performed consecutively. This makes the process efficient and forms a good basis for the further process. However, the delivery of cases was not constant nor frequent, which obstructs the continuity of the process already from the start.
More detailed information on the practical execution and the consequences is provided in Appendix A-2 and a schematic overview of the process at this unit is shown in the stream map in Appendix B-2.
From the unit Noord-Holland no data for a quantitative analysis was received. Therefore, no quantitative analysis could be performed. This would be interesting to be performed in the future, as this would provide more insight into the current performance of the process at this unit.
13
Noord-NederlandThe forensic unit of the police at the unit Noord-Nederland had six personnel members which all work in the second line of work (work at the research facility). This prevents disturbance of the process by first line emergencies. However, due to other work duties the unit performed NFiDENT only one or two times a week. This obstructs the continuity of the total process.
At this unit the laboratory setting was not ideal, as the preliminary investigation and the GC-MS analysis were performed in different laboratories. These laboratories were even one floor apart, which obstructed the connection of these process steps. Additionally, the delivery of cases occurred only once a week. This is less efficient and delays the further process.
Not all time steps mentioned in Table 1 were available for the unit Noord-Nederland, only T2, T3 and T5 were
available. Therefore, the TTAT, TAT1 and TAT5 could not be determined nor analyzed.
The TAT2 data also shows that there was generally more than two weeks between the preliminary investigation
and the GC-MS analysis. Most of this time is waiting time, as the preliminary investigation is fairly simple and should not take longer than one day. Table 9 shows the TAT2 significantly increased in 2018. The effect size is small, but
it does show a deterioration of the process. The variance did decrease, meaning the distribution was less widely spread in 2018 as shown in Figure 7. However, TAT2 generally takes more time meaning the distribution shifted to
the right.
If the process could be performed consecutively by a suited laboratory setting, the TAT2 could be decreased and
could be performed in the same time as TAT3 as they would practically become one process step.
Table 9. Results of the data analysis of TAT2 at the unit Noord-Nederland for 2017 and 2018 (till 27.02.2018).
Interval Year N Median
(days) Variance (σ2) Comparison 2017 – 2018 (P value) Effect size (r) TAT2 (preliminary investigation – GC-MS analysis) 2017 66 14 412,5 0,009 +0,286 2018 17 25 199,3
14
NFiDENT was performed only one or two times a week, but when performed the process step was very fast. The TAT3 data corresponds to this, as Table 10 shows a TAT of zero days. This is already ideal as it could not beperformed in less than zero days. It is also shown that TAT3 did not significantly change from 2017 to 2018.
However, the variance did increase in 2018. This means the frequency distribution was more widely spread in 2018, as show in Figure 8. This shows a small deteriorated by being less constant in TAT, but this is also caused by the smaller sample size. The variance is still important for the continuity of the process and could be improved by a more frequent scheduling of personnel.
Table 10. Results of the data analysis of TAT3 at the unit Noord-Nederland for 2017 and 2018 (till 27.02.2018).
Interval Year N Median
(days) Variance (σ2) Comparison 2017 – 2018 (P value) Effect size (r) TAT3 (GC-MS analysis – report request) 2017 62 0 9,0 0,799 - 2018 16 0 21,5
Figure 9. Frequency distribution of the turnaround time interval 3 (time in days) at the unit Noord-Nederland for 2017 and 2018 (till 27.02.2018).
15
The correlation between the intervals is shown in Table 11. The correlation between TAT2 and TAT3 is notsignificant, showing there is no clear relationship between the two intervals. The correlation that is present is negative. This shows that generally when TAT2 takes long time TAT3 takes short time (and the other way around),
which is an inverse relationship. This was mainly caused by the increase of TAT2, as shown previously, of which
the variance is relatively high. The variance of TAT3 is relatively low, meaning that when TAT2 is finished TAT3 was
mostly performed within the same time. Therefore, this correlation shows no significant relation.
Table 11. Correlation coefficients between the intervals at the unit Noord-Nederland in 2018 (no significant correlations were observed).
Correlation coefficient
(2018) TAT2 TAT3
TAT2 1,000 -0,566
TAT3 -0,566 1,000
The available data indicates TAT2 as the bottleneck in the process at this unit. However, there is much data that is
not available which could be a bottleneck in the process as well. To determine the bottleneck with more certainty more data should be available.
For now it has been shown that TAT2 takes too much time and did not improve in 2018. The long TAT2 seems to
be caused by the unsuitable laboratory setting and its consequences.
More detailed information on the practical execution and the consequences is provided in Appendix A-3 and schematic overview of the process at this unit is shown in the stream map in Appendix B-3.
Oost-Brabant
At the unit Oost-Brabant the NFiDENT process did not start yet, at the time of the interviews. However, interviews were performed to provide some insight into the intentions for the NFiDENT process.
Five personnel members of the unit Oost-Brabant are expected to work on NFiDENT. Two of them also worked in the first line and could be called to a crime scene at any moment. The other three only worked in the second line, but also had other work duties. It was planned to perform NFiDENT two times a week, which is not ideal for the continuity and security of supply for the whole NFiDENT process.
The laboratory setting was not ideal, as the preliminary investigation and the GC-MS analysis were performed in different laboratories. This could cause a bad connection between the process steps and therefore an inefficient NFiDENT process.
The delivery of cases occurred directly from the police according to the case flow. This is fortunate for the continuity of the NFiDENT process, but not the intention of the police processes. Therefore, this could cause trouble in the future and obstruct the continuity of the NFiDENT process when changed.
More detailed information on the practical execution and the consequences is provided in Appendix A-4. No schematic overview of the process was established yet, as the process was not operational yet.
NFI
At the NFI four personnel members were assigned as reporting experts for NFiDENT. They were scheduled to perform the reporting for one week a month. In the future this might not be enough, as more request will be coming in when more units start using NFiDENT. By that time, it might not be possible to perform all the reporting next to the other work duties anymore. For now, the reporting duty was seen as an easy and fast job.
Software which would automatically interprets the data is in development. This will make the process even more easy and faster. However, there will be a transition phase in which all units are using NFiDENT but the automated interpretation software is not available yet. A proper preparation is needed for this phase in order to maintain at least the current continuity and speed of the process at the NFI.
At the NFI the incoming of requests and outgoing of reports was automatically registered. This data was used to analyze the TAT4 and compare 2017 and 2018.
16
Table 12 shows the TAT4 significantly decreased in 2018 with a large effect size. This shows a great improvementof the process compared to 2017. One of the main factors in this decrease would be the digitalization of the report which was introduced at the start of 2018. This eliminated administration steps at the NFI and further accelerated the process as shown by the TAT4 decrease. The variance decreased as well, which show that the frequency was
less widely distributed in 2018. Overall the TAT4 was largely improved in 2018. Figure 10 shows the frequency
distribution of both years and clearly visualizes the improvement. However, there are still some outliers present for which there was no clear explanation.
Table 12. Results of the data analysis of TAT4 at the NFI in 2017 and 2018 (till 12.06.2018).
Interval Year N Median
(hh:mm:ss) Variance (σ2) Comparison 2017-2018 (P value) Effect size (r) TAT4 (report request – report sent to police)
2017 705 27:38:54 2,07E+10
0,000 -0,529 2018 855 1:57:39 4,08E+09
17
The experts were not completely aware of the exact agreements between the police and the NFI, which was why they did not know exactly if they complied to it. The deadline for NFiDENT cases that was set in the registration system of the NFI, was five days. This provided an unrealistic insight into the compliance of agreements as this did not correspond to actual deadline of NFiDENT cases. The specific agreements between the police and the NFI are shown in Appendix C – Table 1.The data of 2018 was compared for two time intervals (TI) in 2018, namely: 01.01.2018 until 07.03.2018 (TI1) and
08.03.2018 until 12.06.2018 (TI2). This was performed after the analysis of TI1 showed unexpected results, namely
a 71% compliance of agreements. This was lower than expected by the experts and they were not aware of this and not all agreements were not clear to them as mentioned previously. Therefore, the results were discussed and it was agreed that the experts would pay attention to the specific agreements and would schedule more times during the day to check for new requests. This should result in an improvement of the compliance.
TAT4 was analyzed again after three months. The resulting percentages of agreement compliance are shown in
Table 13 along with the TAT4 comparison results. It is shown that the compliance of agreements increased with
22,7% between TI1 and TI2. The TAT4 also significantly decreased. Although the effect size is not that large, the
compliance of agreements does show a great increase. In this case the compliance of agreements is more important as performance measure. A decreased TAT is a mean to comply more to the agreements but does not correlate directly to an increased compliance of agreements. Therefore, the effect size provides just an indication of the change that occurred and the compliance of agreement percentage shows the more relevant performance of TAT4.
The TAT4 showed a great improvement from TI1 to TI2. TI2 shows a percentage of compliance of 93,8% which is
almost the ideal compliance. The compliance goal would be 95%, because 100% would not be realistic as there will always be exceptions and outliers in real life. To accomplish the 95%, the process needs to improve just a little more. The results show that the NFI is on good way and it is assumed that that the same approach will eventually lead to the 95%. Especially if the deadline is set correctly in their registration system, in this way the experts will be aware of their actual performance and are able to anticipate to this.
Table 13. Results of the data analysis and the compliance of agreements for TAT4 for TI1 and TI2.
TAT4
(report request – report sent to police) N Median (hours) Compliance of agreements P value Effect size (r) TI1 273 2:34:01 71,1% 0,000 -0,280 TI2 581 1:47:28 93,8%
The initial aim of the NFiDENT process was to accelerate the conventional process. The two product names for NFiDENT and the conventional process are respectively VDM017 and VDM210. The data of TAT4 for VDM017
and VDM210 were compared to each other. Table 14 shows that the TAT4 of VDM017 was significantly lower
compared to VDM210 with a large effect size. This shows that VDM017 is a large improvement compared to VDM210, which was one of the goals of NFiDENT. For this part the NFiDENT process definitely succeeded. The ratio between the TATs of VDM017:VDM210 is equal to 1:0,01. Which means that the time spent on one VDM210 request is the same as on 98 NFiDENT requests. This is a great performance and confirms that the NFiDENT process accelerated the process for TAT4.
One should keep in mind that this data only shows the process at the NFI and that for VDM017 no practical investigation is conducted anymore. The practical investigation has shifted to the forensic unit of the police. One should also keep in mind that not all units are working with the NFiDENT process yet and that more VDM210 requests will be replaced with VDM017 in the future. For the continuity of the process it is important to maintain the current TAT even when more request start coming in.
Table 14. Results of the data analysis of the TAT4 interval for the two different products in 2018 (01.01.2018 – 23.05.2018).
Interval Product N Median
(hours) Median (days) P value Effect size (r) TAT4
(report request – report sent to police)
VDM210 939 189:41:00 7,90
0,000 -0,866 VDM017 741 1:57:38 0,08
18
Comparison between unitsThe stream maps of Appendix B show some practical differences between the units. It appeared that not all units were using the digital assignment system yet and not all units worked fully digital. Not all had a suitable laboratory setting to perform the analyses consecutively. Additionally, most units did not have sufficient personnel members to perform the NFiDENT process every day.
The main reoccurring aspects together with the main results of the quantitative analysis are summarized in Table 15. This provides an overview for easy comparison of the practical situation and the performance of the process. Here only the TTAT and TAT3 are shown to provide a general view of the total process and of the specific NFiDENT
step in the process.
For some units data is missing. This would be relevant to determine in the future to provide a more complete overview of the process and be able to provide more specific recommendations for optimization.
Table 15. Summary of results per unit with the associated performances according to the following scale: --- ‘very bad’, -- ‘bad’, - ‘not ideal’, + ‘okay’, ++ ‘good’, +++ ‘very good’.
Unit Number of personnel Number of analysis days Laboratory setting Delivery of cases TTAT
New process step introduced by
NFiDENT
Den Haag 3 members
no back up - Almost every day +++ 2 laboratories no consecutive process
-- Constant delivery +++ Median of 2 months ---TAT3 Median of 0 days +++ Noord-Holland 4 members no back up - 2-3 times a week - 1 suitable laboratory for a consecutive process +++ Not constant but often once a week
-- N.A. N.A. TAT3
N.A. N.A. Noord-Nederland 6 members Sufficient ++ 1-2 times a week -- 2 laboratories no consecutive process -- Standard day
once a week - 25+ days
--TAT3 Median of 0 days ++ Oost-Brabant 5 members no back up -Once a week -- 2 laboratories no consecutive process -- Constant
delivery - N.A. N.A.
TAT3 N.A. N.A. NFI 4 members not sufficient for the future
+ N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A.
TAT4
Median 2 hours, 93% compliance
19
The TAT3 interval was also compared between the unit Den Haag and the unit Noord-Nederland, to determine ifthere was a difference in performance. This was the only interval that was available for 2018 to compare between units.
Table 16 shows that the TAT3 interval at the unit Den Haag was not significantly different from theTAT3 interval at
the unit Noord-Nederland. This indicates no differences for this process step between these two units. The comparison of the stream maps corresponded to this, as no major differences were observed between these two units.
Table 16. Results of interval comparison between units.
Units Year Interval P value
Noord-Nederland &
Den Haag 2018 TAT3 0,847
To determine best practices for specific intervals of the process more data is needed to compare the performances of the units. This would be especially interesting for the other intervals, as more differences were observed in the practical execution of the process. It is logical to not find many difference between the TAT3 intervals at the different
units, as this part of the process was taught by the NFI in the same manner to all units.
Limitations of this study
There are some limitations of this study concerning the data analysis.
In the data analysis of the forensic units there are some factors that could have influenced the data. The first factor is the manual registration. The data is manually registered by the police personnel, which makes it more prone to mistakes. Additionally, the interpretation of a time step could have been different per person. For the data of Noord-Nederland it is also known that not all analyses were registered. Therefore, there is no way of knowing if the data was actually random or not, neither if it was representative. During the data analysis this was assumed, otherwise no conclusions based on this data could have been drawn. Absence due to illness, vacation or other reasons could have influenced the data as well. It would have been ideal to set up registration form which could have been used to register all time steps mentioned in Table 1 at all units, to obtain the same data from all units and have a more complete dataset. This would have provided a more regulated study of the performance.
In the comparison between the different units the Mann-Whitney U test was applied [19]. One of the assumptions of this test is that the distribution functions are of equal shape and variance [22]. This was not entirely the case in this comparison, therefore the assumptions were not fully met. However, if one would want to perform the analysis completely correctly, a specific test model should be created. Due to time limitations this was not possible and this seemed unnecessary as conclusions could be drawn from the descriptive statistics as well and this test was the best fit available.
The Wilcoxon test compares paired samples [19]. When analyzing the data over different time intervals the data is not actually paired, as it is not the same case that is analyzed in a different time. However, it was assumed that all NFiDENT cases are of the same type which made the Wilcoxon test the most suitable test in this case. The data was randomly paired in the analysis to resemble paired data.
The available data for 2018 was less than for 2017, due to the fact that it was not even halfway through the year at the time of this study. This makes the analysis of the 2018 data not representative for the whole year. However, the purpose of the analysis was to determine the current performance which is based on the data available at that time. It would have been more ideal to analyze the data of two full years. However, the process during NFiDENT cases is not influenced by the time of the year and it is therefore assumed to be constant over a year. Therefore, the data of one year could be compared to the data of a part of a year.
For this study, only quantitative data was available for the unit Den Haag and Noord-Nederland. This data cannot be assumed to be representative for all units. If the data was available from all units more comparison between units could be performed. This would be interesting to determine best practices in practical execution based on the performance of the TAT. Currently this could only be performed for TAT3 between Den Haag and Noord-Nederland,
which did not differ significantly. This comparison would also be more interesting for the other intervals, because the TAT3 interval would not differ much between units as this is taught in the same manner by the NFI.
20
RecommendationsIn this section recommendations are discussed based on the results described under ‘Results & discussion’. More detailed and specified recommendations are discussed in Appendix A – 6.
- PPS
The PPS should be kept informed of the performance of the NFiDENT process. In this way they are able to anticipate properly and more involved in the optimization of the process.
When the performance of the preceding processes is improved the processes at the PPS could be further investigated to determine the performance of the process in these steps as well.
The different units should also communicate their methods with each other to learn from each other. The unit Noord-Nederland is for instance a unique example of cooperation between the police and the PPS. Other units could learn from their way of working, which is very efficient for NFiDENT.
- Forensic units of the police
As shown in Table 15 there are some reoccurring issues at the forensic units of the police. This leads to the recommendation to make sure enough personnel are present to be able to perform NFiDENT according to the case flow. The specific number of personnel would depend on the availability of the personnel in workdays and the amount of illicit drug cases. It is also recommended that the number of personnel should be sufficient to provide back up in cases of absence.
Another recommendation involves the delivery of cases, which is not constant nor regular at every unit. It is recommended to adjust the delivery moments to the NFiDENT process in cooperation with the ‘ketenbeslaghuis’ (KBH) which is responsible for the deliveries.
It is also recommended that all units make sure that the laboratory setting is adjusted to the NFiDENT process, meaning that T2 and T3 can be performed consecutively.
Another recommendation is to have close contact with the PPS to be able to specify deadlines per case, to be able to deliver reports more effectively as not all cases require the same speed of the process. The unit Noord-Holland provides a good example for this.
The unit Noord-Holland also provides a good example for finishing the case file, which is performed by a backoffice. For the other units it is recommended to work in this manner as well.
For specific measures per unit, the results of this study need to be discussed. Therefore, it is recommended for all units to have the relevant personnel present at a workshop for the optimization of the NFiDENT process. This workshop is scheduled to determine specific plans for optimization per unit in cooperation with the police, NFI and PPS.
- NFI
For the NFI it is recommended to, at least, maintain the current performance of TAT4 and the compliance
to the agreements. Additionally, it is recommended to keep sharing the performance of compliance to the agreements to create constant awareness.
It is recommended to set up a plan for future reporting. Many more requests will be coming in and a proper preparation is needed to at least maintain the current performance.
Another recommendation is a stricter plan of requirements before the forensic units start with the NFiDENT process. A detailed example of this plan is explained in Appendix A – 6.
21
ConclusionsThis study has shown that the new steps at the forensic units of the police, introduced by the NFiDENT process, perform very well. However, the accompanying processes are not well adapted to the new process yet. This causes the process to not reach its goal yet, which is to accelerate the total process for illicit drug cases. To achieve this, a faster process at the police and the NFI is needed. If the desired performance is proven to the PPS, they will be able to anticipate more in the process. This anticipation will eventually lead to a faster total process.
To answer the research question: the current executed NFiDENT process can be optimized by improving the connection of the more general processes to the NFiDENT part of the process. This improvement involves better personnel planning, more regular delivery of goods and a consecutive laboratory process. But also, more communication between the three fields to connect their processes and share experiences to create awareness. More detailed explanations and recommendations are provided in the Results & Discussion and Recommendations section and in Appendix A.
Coming back to the hypotheses of the sub questions;
- The first hypothesis is rejected as it was shown that not all reports were delivered within one day nor according to the agreements. The hypothesis of the process being significantly faster than the conventional process was accepted for the process at the NFI. The other process steps could not be compared to the conventional process.
- The second hypothesis was accepted for some steps at some units, as this differed per step and per unit. It was accepted for the TAT3, TAT5 intervals at the unit Den Haag, the TAT5 at the unit Noord-Nederland and
the TAT4 interval at the NFI. For these intervals the TAT improved, largely due to habituation as no other
aspects of the process changed. For the other analyzed intervals this hypothesis was rejected. Conclusions per unit:
- Den Haag
It is clear that TAT1 is the bottleneck in the process at this unit. The process could therefore be optimized
by eliminating the backlog.
- Noord-Holland
At this unit no clear bottleneck could be determined. There are some aspects that could contribute to optimizing the process such as the delivery from the KBH. These aspects are explained under Recommendations.
- Noord-Nederland
At this unit the TAT2 interval is the bottleneck determined using the available data. The process could
therefore be optimized by a suitable laboratory setting.
- Oost-Brabant
Not enough data was available to conclude on any bottleneck in the process. Therefore, no specific optimization could be suggested. However, the processes of T2 and T3 have shown obstruct the process
by the separated laboratory at other units. Therefore, the process at this unit could be optimized by a consecutive execution of T2 and T3.
- NFI
For the NFI the TAT could be further optimized by a realistic insight into the compliance of agreements. This together with a more planned schedule of checking the request will optimize the performance of the NFI in the NFiDENT process.
22
AcknowledgementsI want to thank Annette van Esch, Annette Sprong, Ard Schoep, Bert Corbijn, Caecilia van Venrooij, Carolien Diever, Caroline Krol, Daniël Takken, Jeannette Bronsgeest, Jeannette Mulder, John Mertens, Leonie Potijk, Mirjam Warnaar, Peter Strookappe, Rik Walinga, Rob van der Veer and Titia Kleffens for making time for me to interview them and providing me with the necessary information.
I want to thank Jerien Koopman, Radboud Winkels, Saskia Verheij and Suzan Derksen for guidance during this project.
23
References[1] Verheij, Saskia; Koper, Carola; Schutte, Leon; Hoogland, Ellen; van der Veer, Rob; Nederpel, Ben; Hurkens, Martin; Bouwman, Marja, Annex 1 van samenwerkingsovereenkomst NFiDENT, Den Haag: Nederlands Forensisch Instituut, 2016.
[2] S. A. Cole, "A surfeit of science: The “CSI effect” and the media appropriation of the public understanding of science," Public Understanding of Science, vol. 24, no. 2, pp. 130-146, 2015.
[3] E. Sikkema and F. G. H. Kristen, "Strafbeschikking en ZSM: verschuivingen binnen de
strafrechtshandhaving," in Relaties van gezag en verantwoordelijkheid: strafrechtelijke ontwikkelingen, Den Haag, Boom Lemma, 2012, pp. 179-205.
[4] A. Kloosterman, A. Mapes, Z. Geradts, E. van Eijk, C. Koper, J. van den Berg, S. Verheij, M. van der steen and A. van Asten, "The interface between forensic science and technology: how technology could cause a paradigm shift in the role of forensic institutes in the criminal justice system," Philosophical Transactions of the Royal Society B Biological Sciences, vol. 370, no. 1674, 2015.
[5] W. King and E. Maguire, "Assessing the Performance of Systems Designed to Process Criminal Forensic Evidence," Forensic Science Policy & Management: An International Journal, pp. 159-170, 2010. [6] C. Vilela Rodrigues and J. Carlos de Toledo, "A value based method for measuring performance on
Forensic Science service," Gestão & Produção, vol. 24, no. 3, pp. 538-556, 2017.
[7] G. C. Omari, S. V. Manyele and G. Mwaluko, "Analysis of Turnaround Time during Casefile and Sample Processing in Forensic Science Laboratory," Scientific Research Publishing, vol. 10, pp. 43-73, 2018. [8] S. Jenner, Managing Benefits, 2nd ed., APMG International, 2014.
[9] H. Kleijn and F. Rorink, Verandermanagement, 3rd ed., Pearson Education Benelux, 2012.
[10] M. W. Dale and W. S. Becker, Forensic Laboratory Management, Boca Raton: CRC Press Taylor & Francis Group, 2015.
[11] C. Freitag and B. Found, "Developing tailored planning models for forensic organisations," Australian Journal of Forensic Sciences, vol. 49, no. 4, pp. 379-391, 2017.
[12] R. C. Hawkins, "Laboratory Turnaround Time," The Clinical biochemist, vol. 28, no. 4, pp. 179-194, 2007. [13] L. H. Hilborne, R. K. Oye, J. E. Mcardle, J. A. Repinski and D. O. Rodgerson, "Evaluation of Stat and
Routine Turnaround Times as a component of laboratory quality," American journal of clinical pathology, vol. 91, no. 3, pp. 331-335, 1989.
[14] B. A. Stotler and A. Kratz, "Determination of Turnaround Time in the Clinical Laboratory," American journal of clinical pathology, vol. 138, no. 5, pp. 724-729, 2012.
[15] H. Chima and V. Ramarajan, "Success with Turnaround Time," Laboratory Medicine, vol. 39, no. 9, pp. 518-520, 2008.
[16] G. R. Lowe, Y. Griffin and M. D. Hart, "Analysis of STAT Laboratory Turnaround Times Before and After Conversion of the Hospital Information System," Respiratory care, vol. 59, no. 8, pp. 1275-1280, 2014. [17] Verheij, Saskia, "NFiDENT Pilotrapport v1.1," Nederlands Forensisch Instituut, Den Haag, 2015. [18] R. D. Behn, "Why measure performance? Different purposes require different measures," Public
Administration Review, vol. 63, no. 5, pp. 586-606, 2003.
[19] D. LeBlanc, Statistics: Concepts and Applications for Science, Jones and Barlett Learning, 2004, pp. 228-233.
24
[21] J. Cohen, Statistical Power Analysis for the Behavioral Sciences, 2nd ed., Hillsdale: Lawrence EarlbaumAssociates, 1988.
[22] E. Skovlund and G. U. Fenstad, "Should we always choose a nonparametric test when comparing two apparently nonnormal distributions?," Journal of Clinical Epidemiology, vol. 54, no. 1, pp. 86-92, 2001.
25
AppendicesAppendix A – Detailed qualitative information and recommendations 1 – Den Haag
The public prosecution service
The PPS of the unit Den Haag was enthusiastic about the NFiDENT process, but also mentioned that it did not reach its full potential yet. The PPS could not anticipate ideally to the preceding processes due to a lack of continuity and security of supply. They explained that court hearings were not planned earlier if it was not sure that if the expert report would be delivered in time. For the PPS to anticipate ideally it demands a certain performance of the police and the NFI. A maximum delivery time of three days after confiscation was suggested to connect the processes efficiently. This would mean that a detained suspect could be sent home with a subpoena or the information could be used during an arraignment. In this way the processes connect better to each other’s time tables. However, it is difficult to determine a standard delivery time, as not all cases might demand such a speed of the process. The ideal situation would be that the delivery time would be discussed and specified per case. The question would be if this is possible. This depends on the structure and contact between the PPS and the forensic unit.
It was also brought to the attention that not all public prosecutors were well informed about the new process and it was not clear what could be expected of it. The public prosecutors should be well informed to be able to anticipate in the process. When the NFiDENT process is a success in a case, this should be shared with the parties involved to create awareness.
Backlog
At the forensic unit of the police it was explained that this unit had some troubles over the past year, which lead to a temporary understaffing and a backlog in general. This has also influenced the NFiDENT process. Although the NFiDENT process by itself (T3, T4 and T5) was running very well at this unit, the processes around it caused that
the NFiDENT cases had a backlog as well. The personnel estimated the backlog was at approximately two months. For the NFiDENT process to perform optimally the backlog needs to be eliminated.
Personnel
The process was in use for more than two years and they were one of the first units joining in the starting phase of the NFiDENT project. Not all personnel were working with NFiDENT for those two years, but one was there from the start. This provided a good basis as this one personnel member was very accustomed to the process. Somewhat newer personnel were still getting accustomed to the process, but understood the process well. At the time of the interview three personnel members in total were working on NFiDENT. It was planned to have two more personnel members for the NFiDENT process. However, this unit did manage to perform GC-MS analyses almost every day despite of the understaffing. The personnel members that were working on NFiDENT also had other work duties at the unit. One of the three worked in the first line as well (at the crime scene). This means that one could be called to a crime scene at any moment. This makes it more difficult to create continuity in the process as one cannot predict what other work might occur. This needs to be considered when planning the NFiDENT work, there needs to be some back up to make sure the work can be performed.
The personnel agreed that every member handled their own reports for which they performed the analysis. This means that when an export report is delivered at a time that the investigator of that case is not present, the report is just waiting to be sent to the requesting authority. This wastes time especially because it involves a simple action which can be performed by others as well. This would save time and contribute to the continuity of the process. Delivery
Illicit drugs that had to be investigated were delivered every day by the ‘ketenbeslaghuis’ (KBH, location where confiscated goods were securely stored by the police) if needed. This means that there is a constant supply of material to be investigated. In practice this meant that generally four or five cases were brought in for illicit drug analysis via NFiDENT. This will form no problem in the NFiDENT process and is ideal already.
Laboratory setting
The laboratory setting was not ideal yet, as the preliminary investigation (T2) took place in a different laboratory
than the GC-MS analysis (T3). At the time of the interview there was a renovation taking place at another side of
the building, to build a laboratory where these two steps could take place consecutively. If this was available this would result in a faster and more efficient process, as less movement and waiting time would be involved in the process.
