Outcomes of residential and day care placements
for people with drug and alcohol problems

the 2000 evaluation for Hammersmith & Fulham Social Services 1.10.98 to 30.9.99

Christo Research Systems
May 2000
Christo Research Systems
Director: George Christo DipHE, BSc, PhD, PsychD.
E-mail: DrGeorgeChristo@tiscali.co.uk
Website: http://users.breathemail.net/drgeorgechristo/
Research funded by the London Borough of Hammersmith & Fulham Drug Action Team.
Commissioned by the Drug & Alcohol Team, Hammersmith & Fulham Social Services,
Marzell House, 122-126 North End Road, London W14 9PP (020-7385-7971 ext.5477).
Manager: John Gordon-Smith.
Participating care managers: Ed Lowe, Karen Hodgson & Emma Brooks.

 

History & current aims

The 1998 Evaluation:

In 1997 Hammersmith and Fulham Social Services, and Ealing Hammersmith and Hounslow Health Authority, commissioned the Centre for Research on Drugs and health Behaviour to produce an evaluation of substance misuse treatments purchased by their care managers. Dr Christo, having relevant clinical experience and a PhD in substance misuse treatment outcome research, was hired by the Centre to do the job. The usual outcome questionnaires (e.g., The Opiate Treatment Index, the Addiction Severity Index, & the Maudsley Addiction Profile) took a while to complete, required the presence of the client in question, and could not combine sub sections to produce a single total score for easy analysis. It became quickly evident that the H&F care managers were not going to accept the extra workload necessary for completion of these questionnaires. Something simpler was required to ensure their co-operation.

So the usual outcome domains were put together in a brief inventory. The literature indicated that treatment compliance and engagement with ongoing support were also relevant to outcome and should be included. The prototype of the Christo Inventory for Substance-misuse Services (CISS) was thus developed to elicit care managers’ impressions of their clients in a quick, standardised and reliable way. CISS forms were completed by care managers at the first assessment. They were then completed again at two-month and six-month follow-ups. For those clients who were eventually placed in treatment, placement names and dates were also recorded. The study was successfully completed and produced many useful findings that helped to shape the future of the purchasing service. Subsequent interest from other services led to further development of the CISS. A validation study has been published in a peer reviewed journal and the scale satisfied all the usual reliability and validity criteria (Christo, Spurrell & Alcorn, 2000). The CISS is now used routinely in Hammersmith & Fulham Social Services, as well as by other purchasers and providers of drug and alcohol services, abstinence oriented treatments, and substitute prescribing services throughout the UK.

 

Current aims:

 

Summary of findings

 

Overview

Tracking the sample population:

370 referrals were received between 1.10.98 to 30.9.99 of which:

 

Flow chart for clients referred to the service: (all percentages below pertain to 370 total referrals)

370 total referrals

------------26%------------

98 did Not Attend

 

|

 

 

 

272 attended first

 

61 referred elsewhere

 

assessment

------------29%------------

32 made own arrangements

 

|

 

15 had custodial sentence

 

|

 

 

 

|

 

23 had no further contact

 

164 entered

 

5 were referred elsewhere

 

in this study

-------------9%------------

3 died

 

|

 

3 didn't meet funding criteria

 

|

 

 

 

130 were allocated

 

 

 

to a care manager

 

 

 

& 105 placements

-------------8%------------

22 didn't take placement

 

had funding agreed

 

8 failed detoxification

 

|

 

 

 

|

 

30 completed

20

100 were placed

 

27 self-discharged

3

in first treatment

------------23%------------

12 used & discharged (1 died)

0

|

 

7 ejected

1

|

 

5 no record

0

15 went on to a

 

4 still in

4

second placement

 

 

11% good

|

 

 

outcomes

11 completers

 

8 completed

7

2 self-dischargers

 

3 still in

3

2 discharged users

-------------4%------------

2 used & discharged

1

 

 

2 self-discharged

1

 

 

Outcome measures & methods

How outcome was measured:

The relevant aspects of the treatment process are treatment site, duration and client dysfunction. These were measured by the Christo Inventory for Substance-misuse Services (CISS) which is a standardised, validated tool now commonly used throughout the UK for treatment evaluation (Christo, 1999a, 1999b, 1999c, 2000a, 2000b).

The CISS is a single page outcome evaluation tool completed by drug / alcohol service workers either from direct client interviews or from personal experience of their client supplemented by existing assessment notes. Its purpose is to elicit workers’ impressions of their clients in a quick, quantitative, standardised and reliable way. The 0 to 20 scale consists of 10 items reflecting clients’ problems with:

These outcome areas are scored on a three point scale of problem severity (0 = none, 1 = moderate, 2 = severe), each point is illustrated with relevant examples for guidance (see Appendix). Thus, a CISS score of 0 would indicate no problems and a score of 20 would indicate severe problems in all outcome areas.

The original Hammersmith & Fulham evaluation (Christo, 1998) was used as a trial run to test the CISS in terms of its usefulness, sensitivity to change and ease of completion. It was very successful and the CISS is now incorporated as a regular part of intake and follow-up procedures.

 

Evaluation procedure:

Baseline CISS forms were completed at the first assessment (with the Joint Assessment Form). They were then completed again at approximately two months into treatment and at six months. For those clients who were eventually placed in treatment, placement names and dates were also recorded. Each care manager kept a ‘research’ ring binder file where completed CISS forms were placed in order of date of baseline assessment. Follow-up CISS forms were simply stapled to the baseline CISS form as and when they were completed. A table of client names and interview dates was kept at the front of each file to serve as a reminder of when follow-up CISS forms were due.

The research files were started on 1.10.98 and closed to further baseline forms on 30.9.99, follow-ups continued to come in for the next six months. Once all available follow-ups were collected, the three ringbinders were delivered to Christo Research Systems where the forms were coded onto a statistical research database. The data were then cleaned, errors in form completion were identified, and a few faxes and telephone calls were sufficient to obtain the necessary information to rectify the errors.

The data were then analysed and all the statistics in this report were derived from the information contained in the CISS forms. This method has been able to produce high quality research with minimal effort from the service. There was no need for the service to maintain or analyse a complex research database, they simply sent off their year’s CISS forms just as one might send end of year accounts to an accountant.

 

Good / poor outcome:

For ease of interpretation, some of this report uses a dichotomous criterion of good / poor outcome.

Poor outcome:

Good outcome:

 

Statistical information:
 

Findings

Client details:

This evaluation concentrates on outcomes for 164 clients (112 males, 52 females) who were mostly allocated to three care managers between 1.10.98 and 30.9.99.

Table 1 Breakdown of client details by care manager

Care

number

%age

drug

clients

good

mean

mean

Manager

of clients

male

users

placed

outcome

age

CISS

 

 

 

 

 

 

 

 

Emma

62

71%

47%

46%

38%

36.0

10.9

Karen

40

65%

40%

70%

39%

39.6

10.9

Ed

58

69%

38%

67%

38%

39.7

10.2

Other

4

 

 

 

 

 

 

Total

164

68%

42%

61%

40%

38.5

10.6

Table 1 above indicates little difference between caseloads of the three care managers. Karen recorded fewer clients than the others, and Emma placed fewer clients than the others. The latter finding may have been an artefact due to the greater number of clients recorded by Emma, the others may have been less likely to record clients that dropped out before being placed. Among clients who were eventually placed in treatment, the rates of good outcome (about 40%) were similar for all three care managers.

Figure 1 Age: Clients’ mean age was 38.5 years (sd = 10.0, range = 18 - 69).

 

Figure 2 Baseline CISS score distribution

The mean baseline CISS score was 10.6 (sd = 2.9, range 3 - 18) and there was no difference in CISS scores between males and females. This score is indicative of a high average level of dysfunction and indicates that these clients are generally more dysfunctional than drug users engaged in a methadone service (see CISS comparison scores in Appendix).

 

Drugs of choice:

Among all 164 recorded clients, the drugs of choice were as follows:

Opiates, cocaine, tranquillisers and antidepressants were collapsed into a single category (Drugs 41%) for the purposes of subsequent analyses.

Table 2 Breakdown of client details: alcohol versus drug users

Drug

number

%age

clients

good

mean

mean

type

of clients

male

placed

outcome

age

CISS

 

 

 

 

 

 

 

Alcohol

96

74%

58%

45%

42.1

10.1

Drugs

68

60%

65%

34%

33.5

11.4

Table 2 above indicates that more alcohol users were recorded than drug users. Although there was a slight trend for a greater proportion of males among the alcohol users, this failed to reach statistical significance. Similarly there were no significant differences between drug and alcohol users in terms of the proportion getting into treatment, or treatment outcome once having got there. However, alcohol users were on average about ten years older than drug users (t [158] = -6.2, p <.001), and they scored about one CISS point less than the drug users (t [161] = 2.7, p =.008). Among the individual CISS items, drug users scored higher than alcohol users for ‘criminality’ (U [161] = 1935, p < .001) and ‘sexual / injecting risk behaviour’ (U [161] = 1540, p < .001). These differences are as would be expected from current literature (Christo et al., 2000).

 

Pre-treatment dropout:

As seen in the overview on page 3, only 100 (37%) of the 272 interviewed referrals eventually made it to a funded placement. For those who did not meet the eligibility criteria for funding such service provision (see H&F information leaflet "Drug & alcohol team"), the team endeavoured to arrange alternative provision or help. Some of the detailed reasons for pre-treatment dropout were as follows.

Referred elsewhere
No further action
Failed detoxification

 

Predictors of pre-treatment dropout:

CISS forms were available for 64 clients who did not make it to a funded placement, and these were compared with the 100 cases that did get into a placement. No significant differences were found.

Table 3 pre-treatment dropout versus treatment entry

Treatment

number

%age

%age

mean

mean

status

of clients

male

drug users

age

CISS

 

 

 

 

 

 

Entered

100

65%

44%

38.3

10.7

Dropout

64

73%

37%

38.7

10.5

Table 3 above indicates that there was a slight tendency for proportionately more females and drug users to enter treatment placements but these differences are non-significant. There were no differences on age, or intake CISS total scores, or CISS item scores.

 

Delays in treatment entry:

Two clients had excessive delays between first assessment and treatment entry. One man (286 days) had to complete a prison sentence, and one woman (267 days) had to wait until the resolution of some family issues (neither had good outcome). These clients were excluded from subsequent analyses to avoid biasing the results.

Figure 3 Details of delays in treatment entry

Fig 3 illustrates that among the remaining 98 placed clients, the majority were placed within a month of assessment. The average delay between first assessment and treatment entry was 25.8 days (n = 98, sd = 24.0, range = -12 - 119). This is a reduction on the average placement delay time from the 1998 evaluation (m = 42 days, n = 45, sd = 42, range = -14 - 156).

There was no relationship between age or gender and treatment delay. However, the average placement delay (34.0 days) for drug users was longer than that (19.5 days) among alcohol users (t[64.7] = -2.9, p = .005). ‘Sexual / injecting risk’ was the only baseline CISS item associated to placement delay (Spearman's rho = .23, n = 98, p = .02). This is similar to the finding of the 1998 evaluation and indicates that drug users, perhaps injectors in particular, are likely to experience greater delays before entering treatment.

The average placement delay (26.8 days) for those who eventually had good outcomes, was no different from that (25.2 days) among those with poor outcomes (t[96] = -.3, p = .8). (Analysis with non-parametric tests to allow for the skewed delay distribution made no difference to this finding.) So among placed individuals, those who waited longer to get into treatment fared no better or worse than the faster admissions. Nor was there any relationship between placement delay and time eventually spent in placement. This is similar to the finding of the H&F 1998 evaluation, and both evaluations disagree with the European Association for the Treatment of Addiction’s assertion that longer waiting times are associated with poorer outcomes (Addiction Today, Jan / Feb 2000, p34). However, it remains likely that shorter waiting times are associated with an increased uptake of treatment placements. This research cannot test the latter point, but with limited funds available for more placements this may be an irrelevant issue in any case.

 

Follow-up periods:

In practice, precise follow-up times were hard to adhere to. However, the spread of follow-up periods is acceptable for a study of this type and the average periods were close to two months and six months as in the previous 1998 evaluation.

 

Missing CISS Follow-ups:

Of the 100 placed clients, a large number were lost to follow-up. Either because they withdrew themselves from treatment against staff advice or were prematurely discharged. It would not have been cost effective to follow-up such clients, and unless there is information to the contrary it is assumed they will have made no change or improvement in their status:

 

Process of change (Figures 4 to 6):

This section concentrates on 47 clients for whom CISS forms were available at baseline and both follow-ups. The adjacent figure illustrates a score distribution indicating substantial dysfunction not dissimilar to that illustrated in Fig.2. No clients achieved the minimum score of 0 or the maximum score of 20. The absence of a ‘floor’ or ‘ceiling’ effect indicates that the CISS is well suited to its target population.

At two-month follow-up, the adjacent figure shows the clients’ scores have shifted to the lower ends of the CISS range. The big reductions in dysfunction are due to the fact that nearly all clients are still in treatment at this time. Their average score of 3 is the same as the CISS comparison score for successfully treated abstinent drug users.

This figure illustrates a 'bimodal' score distribution by six months. There is a central dip as scores tend to bunch up at the top and bottom ends of the score distribution. This is indicative of ongoing reduced dysfunction for those individuals for whom treatment has been ‘successful’. Conversely, the treatment ‘failures’ return to prior high levels of problem severity as any transient treatment related gains are lost on leaving treatment.

 

Figure 7 Changes in individual CISS item scores (mean CISS item scores, scale 0 to 2)

Figure 7 illustrates average CISS item scores for the 47 clients for whom CISS forms were available at baseline and both follow-ups. It indicates baseline problems listed in order of degree of severity and is similar to the 1998 finding. The greatest problem being that of drug use, followed by problems of occupation and support, etc. Problems of criminality, compliance, sexual / injecting risk, and working relationships were noted only for a minority of the assessed clients.

On entering treatment, problems of drug use, support, and criminality fell to zero. Problems of occupation were reduced as treatment at least provided some part time structure and occupation (e.g., many treatments include 'therapeutic duties' to help build self-esteem). Residual psychological, social, health, and sexual / injecting risk problems were also greatly reduced. However, ‘working relationships’ problems were quite resistant to treatment. ‘Difficult to work with’ clients may remain just as difficult even when they have entered treatment and stopped using drugs. Other CISS based research (Christo, 2000c) has indicated that ‘working relationships’ problems are abnormally elevated among dual-diagnosis populations. The phenomenon here may thus be due to a small minority of clients with personality disorder which by definition is an enduring condition.

Drug use and support problems became apparent again at six months as a minority of the sample relapsed and actively avoided contact with supportive structures that interfere with their resumed drug / alcohol use. Problems of occupation recurred on leaving treatment. Although structured support (e.g., NA & day care) and voluntary work provided a stopgap for those who remained abstinent.

 

Predictors of good outcome:

Among the 100 placed clients, 40 had demonstrated a good outcome by the criteria described earlier in the outcome measures & methods section. Baseline information on these clients was compared to that from 60 poor outcome clients in an effort to identify factors that could predict eventual treatment outcome.

Table 4 good versus poor treatment outcome

Outcome

number

%age

%age

mean

mean

status

of clients

male

drug users

age

CISS

 

 

 

 

 

 

Good

40

60%

38%

39.4

10.1

Poor

60

68%

48%

37.6

11.2

Table 3 above indicates no significant differences other than baseline CISS scores. The average CISS score of good outcome clients was one point lower than that for poor outcome clients (t[98] = 2.1, p = .04). A post hoc examination of the 10 CISS items revealed that clients with higher baseline scores on problems of compliance (U[98] = 900, p = .01) and working relationships (U[98] = 958, p = .02) were more likely to then go on to have a poor treatment outcome.

 

Outcome and placement duration:

The average number of days in good outcome placements was 142.6 (sd = 57.0, n = 38, range = 12 - 273): The average number of days in poor outcome placements was 53.9 (sd = 48.9, n = 52, range = 1 - 208). Those clients who eventually demonstrated a poor outcome tended to spend much less time in placements. This difference in placement duration is highly significant (t [88] = -7.9, p < .001) and could not have happened by chance.

The above finding may indicate that placement funding was ‘naturally’ put to effective use, as clients who chose to resume their drug / alcohol use did not tend to remain in placements thus wasting money. Most placements discourage the use of drugs / alcohol on their premises and covert drug / alcohol use may be too stressful for clients to maintain for any significant period. Alternatively, it may simply be that placements requiring longer residency or extended care, produce better outcomes.

 

Outcome and placement type:

Table 5 below has divided the 100 placed clients into 60 clients who had poor outcomes and 40 clients who had good outcomes. Cases were then distributed across the various first placements that were encountered.

Table 5 Client outcome status by first placement name

First placement

poor outcome

good outcome

 

total

name

clients

clients

 

clients

 

n

n

%age

n

 

 

 

 

 

SHARP

11

7

39%

18

Clouds House

9

5

36%

14

Accept

5

5

50%

10

Hartley House

7

2

22%

9

CORE Trust

3

2

40%

5

Canterbury Alcohol Project

2 good

2 poor

50%

4 total

Phoenix House

3

1

25%

4

St. Luke's

0

3

 

3

Ravenscourt

1

2

 

3

Agar Grove

2

1

 

3

Davies Centre

2

1

 

3

Blenheim

2

1

 

3

Oak Lodge

3

0

 

3

Alpha House

0

1

 

1

Hope House London

0

1

 

1

Hope House Luton

0

1

 

1

Cranstoun House

0

1

 

1

Francis House

0 good

1 poor

 

1 total

Allington House

0

1

 

1

Frenchay Mews

0

1

 

1

Mount Carmel

0

1

 

1

Thurston House

1

0

 

1

Base Project

1

0

 

1

Hove Family Project

1

0

 

1

Phoenix Sheffield family

1

0

 

1

Ashley House

1

0

 

1

Camberwell Alcohol Project

1 good

0 poor

 

1 total

Sefton Park

1

0

 

1

Trelawn

1

0

 

1

DAF Dartmouth Street

1

0

 

1

Maya

1

0

 

1

 

 

 

 

 

Totals

60

40

 

100

A cursory view of the top of Table 5 indicates that no treatments differed noticeably from the overall treatment success rate of 40%. However, the client numbers were too small to allow a reliable comparison of placement effectiveness and outcomes for placements with single clients should be viewed with extreme caution.

Table 6 Client outcome status by second placement name

First placement

poor outcome

good outcome

total

name

clients

clients

clients

 

n

n

n

 

 

 

 

Quinton House

1

2

3

Thurston House

0

2

2

Pelham Road Hostel

1

1

2

Hove Family Project

0

1

1

Nelson House

0

1

1

Phoenix Bexhill

0

1

1

Phoenix London

0

1

1

St. Luke's

0

1

1

Camberwell Alcohol Project

0

1

1

Hope House London

0

1

1

Accept

1

0

1

 

 

 

 

Totals

3 poor outcome

12 good outcome

15 total

Fifteen of the 100 placed clients went on to a second placement and table 6 above indicates their detailed outcomes. As expected, the overall success rate of 80% was higher than that for primary placements.

 

Outcome for day care versus residential placements:

Clients generally chose their own placements with some guidance and advice from their care managers. The details of 38 clients who entered day care placements were compared to those of 62 clients who entered residential placements. Outcome rates were the same, but those who went to residential placements had greater initial levels of dysfunction.

Table 7 day care versus residential treatment

Treatment

 

%age

%age

%age

mean

mean

mean

type

 

male

drug

good

age

treatment

intake

 

n

 

users

outcome

years

duration

CISS

 

 

 

 

 

 

 

 

Day care

38

66%

45%

39%

38.8

74.8 days

9.7

Residential

62

65%

44%

40%

38.1

79.8 days

11.3

Table 7 above indicates no significant differences other than baseline CISS scores. The average CISS score of day care clients was one and a half points lower than that for residential clients (t[98] = -3.3, p = .001). A post hoc examination of the 10 CISS items revealed that clients with higher baseline scores on social functioning (U[98] = 855, p = .01) and criminality (U[98] = 894, p = .01) were more likely to be subsequently placed in a residential setting.

 

Outcome and reasons for leaving treatment:

Clients still in treatment at 6-month follow-up or with no record of outcome were excluded from this analysis. Of the remaining 88 placed clients, 50 had left treatment prematurely.

Table 8 Reasons for leaving treatment by treatment outcome

 

Poor Outcome

Good Outcome

total

 

Clients

Clients

 

 

 

 

 

(not included) Still in treatment at follow-up

0

7

7

(not included) Lost contact - no record

5

0

5

 

 

 

 

*Ejected for poor compliance

6

1

7

*Ejected for using drugs / alcohol

13

1

14

*Left prematurely against staff advice

25

4

29

~Left on treatment completion

11

27

38

 

 

 

 

*Left prematurely (for whatever reason)

44 (88%)

6 (12%)

50 (100%)

~Left on treatment completion

11 (29%)

27 (71%)

38 (100%)

Table 8 above illustrates that poor outcome was generally associated with premature treatment termination and this association is highly significant. Among the 50 premature leavers only 12% eventually had a good outcome, whereas 71% of the 38 treatment completers had a good outcome. So it appears that the relationship between poor outcome and shorter treatment duration is due to poor client compliance rather than variations in the designed lengths of treatment programmes. This finding replicates the result of the 1998 evaluation and it corresponds with the predictive effects of compliance and working relationships problems on treatment outcome. Clients should be forewarned of this robust effect before they accept a placement.

 

Outcome and location at intake:

At the intake assessment, clients were recorded as being housed in various different locations (e.g., home, prison, etc. Table 9 below illustrates the outcomes for clients from each location. The fourth column of Table 9 provides the percentage of all clients from a particular location who eventually demonstrated good outcomes.

 

Table 9 Outcome by location at intake:

Location

Pre-treatment

Poor Outcome

Good Outcome

Good Outcome

total

at intake

dropouts

Clients

Clients

(%)

 

 

 

 

 

 

 

At home

46

37

19

19%

102

Detoxification clinic

10

12

10

31%

32

No fixed abode

2

2

6

60%

10

Hospital

1

5

3

33%

9

Prison

1

3

1

20%

5

Hostel

4

0

0

0%

4

Already in residential treatment

0

1

1

50%

2

 

 

 

 

 

 

Total

64

60

40

 

164

Surprisingly, those who had no fixed abode at the time of assessment demonstrated the greatest rate of good outcome (60%). Perhaps they chose to remain in treatment because they had nowhere else to go. Individuals who were already engaged with treatment in hospitals or detoxification clinics had Intermediate rates of outcome (30%). Low rates of good outcome (20%) were achieved by those initially at home or in prison, many of these individuals did not even get into treatment. All hostel residents dropped out pre-treatment.

 

How ‘poor’ is a poor outcome? Changes in CISS scores within individuals:

The assumption of the 1998 report was that ‘poor outcome’ clients would have returned to previous levels of dysfunction and all treatment related gains would have been lost. This assumption is tested below by looking at changes in clients’ CISS scores between baseline and follow-ups. For each client, their baseline CISS score was subtracted from their follow-up score, a negative number thus indicating a CISS score reduction and consequent improvement in functioning.

Table 10 Changes in CISS score between intake and 2-month follow-up

 

n

mean CISS change

(sd)

range

t

p

 

 

 

 

 

 

 

Good outcome clients

40

-7.4

(2.6)

-1 to -13

17.9

<.001

Poor outcome clients

39

-7.1

(4.9)

+5 to -16

9.1

<.001

Table 10 shows that 79 (of 100) placed clients were captured by the 2-month follow-up and half of them eventually went on to have a poor outcome. However, both groups showed similar improvements corresponding to an average drop of 7 CISS points at two months into treatment. Paired samples t-tests show this large change in CISS scores to be highly significant. So even the relapsers made dramatic improvements in functioning while still in treatment (see figure 8 for details).

Table 11 Changes in CISS score between intake and 6-month follow-up

 

n

mean CISS change

(sd)

range

t

p

 

 

 

 

 

 

 

Good outcome clients

34

-7.6

(2.1)

-4 to -11

21.5

<.001

Poor outcome clients

13

-2.1

(2.7)

+3 to -7

2.8

.02

Table 11 shows that 47 (of 100) placed clients were captured by the 6-month follow-up and 13 of them had a poor outcome. However, the poor outcome group still retained significant improvements corresponding to an average drop of 2 CISS points. Their average intake CISS score was 12.5 and their follow-up score was 10.4. This improvement of 2 CISS points is not dissimilar to that which might have been achieved had the clients been engaged in a long-term substitute prescribing programme (see CISS comparison scores in Appendix). So even those with ‘poor outcomes’ benefited from a period of respite from their drink / drug using lifestyle. However, most of the poor outcome clients were not caught by the 6-month follow-up and it is possible that the few poor outcome clients sampled above may have been among the less dysfunctional.

 

Summary of current findings

 

Summary of findings from the 1998 evaluation

 

Conclusions

Hammersmith & Fulham have again demonstrated how easy it is to produce high quality research within the limitations of a busy service setting. It was not even necessary to maintain computerised records, as CISS forms in ring binders were quite sufficient. The notion of evidence led practice is frequently discussed by other services, but research aims are often set too high to be practicable and nothing gets done. Perhaps this may be a deliberate ploy to avoid having to do any such work. It could be argued that experienced practitioners already make best use of their resources. Thus, the purpose of such research could only be to illustrate to more ignorant outsiders that these ‘experts’ know what they are doing (e.g., practice led evidence). This view may well be partially justified, as some of the findings in this study are obvious to anyone who is familiar with the field. However, some findings here are obvious only with the benefit of hindsight and others may inform better practice at a local level.

Individual CISS items consist only of 3-point scales and it has been argued that they are not sensitive to change in individual cases. However, the CISS total scale multiplies that sensitivity by having ten such items. Similarly, significant and useful findings have been derived by analysing individual items when multiplying their power by ten or more individuals. For example, although day care and residential services produced similar outcomes, their intake differed in that residential treatments tended to get more dysfunctional individuals with social functioning and criminality problems. This illustrates the importance of considering individual intake variables when comparing outcomes.

Many findings agreed with those of the 1998 study. Thus indicating the reliability of this research design, and the value of having many small studies that are temporally and geographically relevant (as opposed to one big one). For example, both studies found that poor outcomes were closely related to the absence of ongoing structured support. They also found that the CISS total score and the ‘compliance’ item in particular were able to predict who were likely to have poor outcomes. Unreliable clients were more likely to fail treatment when they eventually got in. It may thus seem justified to ensure that clients reliably attend appointments before treatment entry in order to demonstrate their motivation for change. Any resulting delays would appear to have no impact on treatment outcome in any case. Delay times have generally reduced since the 1998 study, but it remains unclear why drug users experience greater delays before being placed. It is interesting to note that the new CISS item ‘working relationships’ also predicted treatment failure and was the only CISS item resistant to treatment change. Further research would be necessary to find out how many of such ‘difficult to work with’ clients suffer from long term personality disorder and need special treatment provision. However, while resources are limited, it may be reasonable to continue to allow non-compliant / difficult to work with clients to select themselves out during the assessment process, as they are more likely to fail treatment by leaving prematurely in any case. Nonetheless, even the treatment ‘failures’ appeared to have benefited from their experience in treatment.

In this analysis, some of the most useful and informative variables have been ‘ongoing support’, ‘compliance’ and ‘working relationship’. It is interesting to note that older outcome evaluation tools have no equivalent to these domains.

 

References

Christo, G. (1988). Outcomes of residential care placements for people with drug and alcohol problems: an evaluation of Hammersmith and Fulham Social Services. Centre for Research on Drugs and Health Behaviour, Hammersmith & Fulham Social Services, Ealing Hammersmith & Hounslow Health Authority.

Christo, G. (1999a). Outcome monitoring: service evaluation made simple. GLAAS mailing, Issue 95 January 1999, Greater London Association of Alcohol Services.

Christo, G. (1999b). Keep it simple. Drug & Alcohol Findings, June 1999, Issue 1, p27.

Christo, G. (1999c). CISS: keeping it sweet and simple. Addiction Today, Vol 11, No 61, pp14-15.

Christo, G. (2000a). Clear classification: simple service evaluation. Druglink (ISDD journal), 15(1), pp19-21.

Christo, G. (2000b). CISS: keeping it sweet and simple (part 2). Addiction Today, Vol 11, No 62, pp14-15.

Christo, G. (2000c). New assessments at the Royal Free Drug Service. Unpublished internal report, Royal free Drug Service.

Christo, G., Spurrell, S. and Alcorn, R. (2000). Validation of the Christo Inventory for Substance-misuse Services (CISS): a simple outcome evaluation tool. Drug and Alcohol Dependence. 59, pp189-197.

Christo, G. (in press). The Christo Inventory for Substance-misuse Services. In J. Maltby, C.A. Lewis, and A. Hill (Eds). A Handbook of psychological tests. Lampeter, Wales, UK: Edwin Mellen Press.

 

Appendix, CISS form

The front of the form contains:

The back of the form contains:

 

  

 Links within CISS site 

CISS home page and index.

CISS form

CISS general overview. Unedited version of an explanatory article about CISS which appeared in Addiction Today Magazine (Nov/Dec, 1999)

CISS comparison scores for harm minimisation oriented methadone prescribing outpatient services

CISS comparison scores for an outpatient alcohol service (item score comparisons with drug users)  

CISS comparison scores and cutoffs for abstinence oriented services

CISS technical information (reliability, validity, correlations with other scales)  

CISS detailed information. Validation of the Christo Inventory for Substance-misuse Services (CISS): a simple outcome evaluation tool (from Drug and Alcohol Dependence, 2000).

 

ã 1998 George Christo PhD, PsychD.