Transcript Source: REITOX National Focal Points. TDI detailed tables EMCDDA..

1
Data quality and feedback
from the assessment on
2000 and 2001
TDI data
EMCDDA, TDI meeting, 23/24 June 2003
Luis Royuela.
Data quality and feedback from the
assessment on 2000 and 2001TDI data
- Introduction
- Years covered and reporting period
- Coverage: Geographical and Agencies
- Availability of TDI items and comparability with TDI
standards.
- Completeness of each tables.
- Double-counting.
- Internal consistency.
- Other problems.
- Modification of TDI tables.
- Conclusion.
- Recommendations
2
Introduction
3
Objective of indicator
To provide comparable, reliable and anonymous
information on the number and characteristics of people
treated for their drug use in the Member States.
Purpose of information:
Epidemiological
- Description of the profile of drug
users within Europe.
- Identification of patterns of use.
- Changes in patterns of use.
Management
- Which is the population we are
interested in and why?
- What are the obstacles in getting
this information?
ACTIVITIES CARRIED OUT SO FAR.
• REITOX sub-task 3.1
• REITOX sub-task 3.2
• REITOX sub-task 6.1
• Procedures to avoid double
counting in drug treatment
reporting systems
• Feasibility study on the
implementation of the proposals
given in the final reports of
REITOX sub-tasks on improving
the quality and comparability of
treatment reporting systems
• TDI protocol December 2000.
Introduction
4
Implementation of data quality in the EU member states
and analysis of aggregated data in order to define a
“basic profile of clients in treatment”.
General Objective:
Analysis of TDI data available in the EU member states
in order to define a basic “profiles” of clients in
treatment.
Specific Objectives:
I.
To make data ready for analysis, through a detailed data
check
II. Assess quality of TDI data
III. To identify methodological problems, in the data collected,
which could affect the quality of the analysis outputs.
IV. To transform data for analysis.
V. To define analysis outputs.
VI. Final report
5
Years covered and reporting period
Reference period
Belgium
12 months:January_December
Denmark
12 months:January_December
Germany
12 months:January_December
Greece
12 months:January_December
Spain
12 months:January_December
France
One month (November)
Ireland
12 months:January_December
Italy
12 months:January_December
Luxemburg
12 months:January_December
Netherlands
12 months:January_December
Austria
12 months:January_December
Portugal
12 months:January_December
Finland
12 months:January_December
Sweden
12 months:January_December
United Kingdom
2000
Six months
1 October 2000- 31 March 2001;
1 April 2001 - 30 September 2001
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
2001
Geographical coverage
Outpatient data available in TDI Tables
2000
2001
Not available
Available
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
6
Agency and clients coverage
OUTPATIENTS
TREATMENT
CENTRES
2000
Finland
Germany
Greece
Ireland
Italy
Luxembourg
Netherlands
Spain
2001
Denmark
Finland
Germany
Greece
Netherlands
Spain
Sweden
Nº of unit
reporting
(1)
53
401
8
96
92
6
110
446
-
Nº of unit in
the coutry
(2)
% of unit
covered
(3)
% cases
covered
(4)
-
-
-
944
17
129
515
6
116
495
-
73
368
15
166
459
53
7
42%
47%
74%
18%
100%
95%
90%
-
180
1049
18
175
517
400
40%
35%
83%
95%
90%
13%
13%
70%
95%
95%
16%
(1) N of units covered: number of outpatient treatment centres reporting to the monitoring system
(2) N of units in the country: Number of outpatient treatment centres that exist in the country.
(3) % of centres covered: estimated proportion of reporting outpatient treatment centres, with regard to the overall number
of the outpatient treatment centre in the country covered by the monitoring system.
(4) % of cases covered: proportion of cases/episodes reported to the monitoring system with regard to all cases treated
nationally.
Agency coverage
8
Estimated proportion of reporting treatment centres
(with regard to the overall number of the respective
treatment centre type in the country) covered by the
monitoring system.
OUTPATIENTS
Not Known
10% - 30%
31% - 70%
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
> 70%
Geographical coverage
Inpatient data available in TDI Tables
2000
2001
Not available
Available
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
9
Geographical coverage
10
Low threshold agencies data available in TDI Tables
2000
2001
Not available
Available
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
Geographical coverage
G.P. data available in TDI Tables
2000
2001
Not available
Available
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
11
Geographical coverage
Treatment units in prison data available in TDI Tables
2000
2001
Not available
Available
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
12
Comparability with TDI items
Belgium
France
Integrating the system in one national system.
Not all items covered.
Not all items covered.
Minor conversions needed to fit categories of the TDI items.
Not all items covered.
Some Items with multiple answers.
Not all items covered.
Minor conversions needed to fit categories of the TDI items.
Not all items covered.
Minor conversions needed to fit categories of the TDI items.
Moving to system based on TDI .
Ireland
Italy
Not all items covered.
Integrating the system in one national system.
Luxemburg
Netherlands
Austria
Portugal
Finland
Sweden
United Kingdom
No specific problems
Denmark
Germany
Greece
Spain
Not all items covered.
Implementing data collection according to TDI Protocol
Portugal is in a pilot phase
Not all items covered.
No specific problems
Not all items covered.
Categories route of administration do not match
categories.
13
TDI
Completeness of TDI tables
14
Completeness of TDI tables
15
1. Missing values: Primary drug.
Definition: The main drug is defined as the drug that
causes the client the most problems.
-
Based on problems as defined by clients
-
Based on the ICD 10 diagnoses.
Records which primary drug is not known.
Should be excluded?
Double-counting
16
TDI collects information on persons who start
treatment for their drug use at a treatment
centre during a given period of time. (one
calendar
year).
Each
person
counted only once in the period.
should
be
Double-counting.
Belgium
Denmark
Germany
Greece
Spain
France
Ireland
Italy
Luxemburg
Netherlands
Austria
Portugal
Finland
Sweden
United Kingdom
-
Level of control
Treatment centre
National
Treatment centre
National
Regional
-
Treatment centre
Very limited control
National
National
No double-counting control
Regional
National
No double-counting control
Regional
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
17
Internal inconsistency
Impossibility to know the real number of:
- Males
- Females
- Drugs
- And other variables
Increases the delay of the data analysis due to:
- Have to request for correction.
- Have to wait for the answer.
18
Internal inconsistency
19
Treatment contact details
4. Previously treated
All treatments
4. Ever previously treated
(N of cases)
1. never
2. previously treated
0. not known
Total
Missing cases:
M
3722
12003
0
15725
F
589
2473
0
3062
Data from 14. 1st drug x ever treated
All treatments
not known Total
0
4311
0 14476
0
0
0 18787
M
F
not known Total
2227
6494
9502
18223
0
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
44
18267
Internal inconsistency
All treatment
1. Opiates (total)
11 heroin
12 methadone
13 other opiates
2. Cocaine (total)
21 cocaine
22 crack
3. Stimulants (total)
31 amphetamines
32 MDMA and other derivates
33 other stimulants
4. Hypnotics and Sedatives (total)
41 barbiturates
42 benzodiazepines
43 others
5. Hallucinogens (total)
51 LSD
52 others
6. Volatile Inhalants
7. Cannabis (total)
9. Other Substances (total)
99. Not known
Total
Missing cases:
Total + missing
% of missing
20
14. 1st drug x 14.3 Primary drug 16. Route of
18. Age 1st
20.1 Ever
ever treated
x age (total)
administration
use
injected (drugs)
15573
7822
15573
15573
15573
15264
7579
15264
15264
15264
282
235
282
282
282
27
8
27
27
27
893
279
893
893
893
891
278
891
891
891
2
1
2
2
2
219
98
219
219
219
50
31
50
50
50
169
67
169
169
169
118
8
110
31
2
29
118
8
110
118
8
110
118
14
12
2
14
12
14
12
14
12
11
1230
165
516
18739
44
18783
0,234
11
1230
165
520
18743
44
18787
0,234
11
1230
165
521
18744
44
18788
0,234
11
1230
165
520
18743
44
18787
0,234
469
20
275
8996
9791
18787
52,116
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
110
Internal inconsistency
21
Treatment contact details
4. Previously treated
7. Age distribution (N of cases)
All treatments
All treatments
4. Ever previously treated
(N of cases)
1. never
2. previously treated
0. not known
Total
Missing cases:
M
F
4637
4433
284
9354
not known
1136
1217
89
2442
0
0
0
0
Total
5773
5650
373
11796
M
F
not known
4780
1190
284
10709
89
2898
1811?
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
Total
0
5970
0
0
373
0
13607
0
Internal inconsistency
22
14. 1st drug x 14.3 Primary drug 16. Route of
18. Age 1st
20.1 Ever
ever treated
x age (total)
administration
use
injected (drugs)
All treatment
1. Opiates (total)
182
182
182
100
11 heroin
19
133
133
133
95
12 methadone
14
14
14
5
13 other opiates
35
35
35
2. Cocaine (total)
10
10
10
17
21 cocaine
4
10
10
10
17
22 crack
5
3. Stimulants (total)
0
3
3
3
2
31 amphetamines
2
32 MDMA and other derivates
18
3
3
3
2
33 other stimulants
4. Hypnotics and Sedatives (total)
4
6
6
6
4
41 barbiturates
42 benzodiazepines
3
3
3
3
2
43 others
3
3
3
2
5. Hallucinogens (total)
16
51 LSD
1
52 others
6. Volatile Inhalants
1
1
1
7. Cannabis (total)
26
26
26
18
9. Other Substances (total)
2
2
2
2
99. Not known
164
19
19
19
145
Total
184
249
249
248
289
Missing cases:
0
0
0
0
0
Total + missing
184
249
249
248
289
% of missing
0,000
0,000
0,000
0,000
0,000
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
Internal inconsistency
23
 Ten countries were included in the analysis.
Denmark
Germany
Greece
Spain
Ireland
Italy
Luxemburg
 Only one did not present any
problem.
 Two presented minor problems.
 Seven presented serious problems.
Netherlands
Sweden
Finland
Source: REITOX National Focal Points. TDI detailed tables EMCDDA..
Other problems
24
 It is a HARD WORK filling in all TDI tables.
o Takes many hours.
o Probability of error is high.
o The cross check of the tables, to ensure they
have the same information as the source, is a
very complex process.
 Facilitate the work of people responsible at NFPs.
o Study the modification of the tables.
o Study methods to automate the process:
1.For filling in tables. (SPSS-SAS)
2.For data checking. (Validation tables)
Modification of the tables.
 Fundamental
aim
to
go
ahead
25
with
a
modification:
o Increase and ensure the quality of data.
 How data should be collected would depend on
what we are going to do with the data.
o Describe general characteristics of clients in
drug treatment of EU member states?
o Comparative analysis?
Modification of the tables
 First option:
1. ADD a new column and row for missing values
All treatments
4. Ever previously treated
(N of cases)
1. never
2. previously treated
0. not known
9. Missing
Total
M
3722
12003
0
0
15725
F
589
2473
0
0
3062
Not known Missing Total
0
0
4311
0
0 14476
0
0
0
0
0
0
0
0 18787
2. Modify the formulas to allow ceros.
26
Modification of the tables.
27
 Second option:
1.
2.
3.
4.
5.
6.
ADD a new column and row for missing values
Drop table 4. Previously treated.
Double table 14. 1st drug x ever treated by gender.
Drop table 7. Age x gender. (Same data in 14.1,2,3)
Drop 19.1,2,3 Other drug x age.
Double 9. Living status (with whom)
10. Living status (where)
11. Nationality by first treatment
7. Drop 20. Ever injected
8. Double 20.1 Ever injected (drugs) by gender.
9. ADD a table for missing and a validation table.
Modification of the tables.
 Third option: Same structure for all the tables.
1. All cross-tabulations:
All Treatments
MEN
Drugs
Opiates
....
...
WOMEN
Drugs
Opiates
....
...
TDI Items
TDI Items
First Treatments
MEN
Drugs
Opiates
....
...
WOMEN
Drugs
Opiates
....
...
TDI Items
TDI Items
2. ADD a table for missing and a validation table.
28
Modification of the tables.
29
 Fourth option:
Create a databank on National Treatment Demand.
1. Report individual records.
(On a voluntary basis?)
2. Report TDI tables.
Data aggregation at EU level:
Databank on National Treatment Demand
30
1. Better control of the data and increase of quality.
2. No Influence on the level of data protection
- Confidentiality
(identification variables should not be reported).
3. Extraordinary increase of the power and richness
of the data analysis.
Conclusions
1. Not ALL European Member States report TDI data
annually.
2. Quite good geographical coverage of outpatient data
in EU, but there is low coverage of agencies and
clients in some Member States.
3. Internal inconsistency were found in nine of the
countries with data available.
4. There is a high probability of error when TDI tables
are filled in.
5. Now is the best moment to face the modification of
the reporting system before the new members join
it.
6. Still have a long way to go until TDI is used to
establish a standard baseline for comparisons.
31
Recommendations
32
1. Data quality and reliability should be improved.
2. Concentrate efforts on outpatients and inpatients and
improve the coverage of agencies and clients in a 1st
phase.
3. Improve the data collection of the other type of units
in the next phase.
4. Simplify spreadsheet. The best way to avoid internal
inconsistency and facilitate the filling in of the tables
is to uniform the structure of the tables.
5. Make more use of the data, analysis and
dissemination.
6. TDI should be used to establish a standard baseline
for comparisons.
7. Create a databank on National Treatment Demand.