Illustration of Hazards From Failing to Monitor a New Social Programme Effectively

This analysis presents the prisons situation as an experiment.

The managers in prisons might choose to view the situation as resulting from prisoner violence, causing withdrawal of psychologist’s services, due to inability to safely protect the psychologists from violent prisoners.

Both viewpoints have some degree of validity, but if the managers were being prosecuted for the manslaughters, their incompetent interpretation shouldn’t protect them from proper legal accountability. (Officer, yes I am allowed to go faster than the speed limit, on nice days when I am in the mood to go fast.)

This example is included to show why it is important to monitor the performance of new programmes, to see that they are working as intended.
If you don’t know what the outcome will be, then you are doing an experiment.
If you don’t monitor the progress of the experiment, then you are unable to stop the experiment, if there are too many adverse outcomes.

If it can be shown that you could have detected outcomes by competent monitoring, then you take full responsibility for the preventable adverse outcomes.

Suicide Prevention in NZ Prisons 1969 to 1994 (25 years) NZ Department of Justice
page 11 Implies 51 suicides in 10 years
page 18 – 1 suicide in 1993 lowest figure since 1981
and 10 in 1994 highest ever shows high degree of fluctuation in annual statistics
page 21 76 suicides from 1969 to 1994
page 68 Figure 8 suicides 1969 to 1994
selected prisons only
Auckland West Prison 1
Auckland East Prison 18
Waikeria Prison 11
Mount Eden Prison 21
Total from 1969 to 1994 51

Page 23 paragraph 77
Experiment to measure cost effectiveness of psychiatric help to prisoners.
Criteria tightened under which Auckland prisoners allowed access to psychiatric help
Experimental design ABA
A – prisoners given access to psychiatric treatment 1968 to 1983,
B – access restricted from 1984 to 1987 and
A – then given access again from 1988 to 1994
Ethical approval of experiment was not shown in any of the Justice Reports.
Prisoner consent was not shown or indicated in any of the Justice Reports.

duration suicides suicide/year
Experimental Plan:
start end years suicides in period suicides per year
A 1968 1983 16 2 0.125
B 1984 1987 4 13 3.250
A 1988 1994 7 6 0.857

Base A suicide rate average 1968 to 1983 plus 1988 to 1994 – suicide rate per year
23 8 0.348
B Experimental treatment – withheld access to psychiatric treatment – suicide rate
4 13 3.250

Expected number of suicides through test period, if treatment not withheld. 1.39
Excess deaths due to withholding access to psychiatric treatment 11.61
Increase factor in suicide rate as a result of withholding treatment 9.34
ie 834% increase in suicides
Expected number of suicides, if access to treatment restored after 6 months 1.63
Expected number of suicides, if access to treatment restored after 1 years 3.25
Expected number of suicides, if access to treatment restored after 2 years 6.50
Looking at the numbers above, it would not really be possible, at 6 months to conclude that the number of suicides was significantly higher, due to withholding access to psychiatric treatment. The number of suicides might just be a cluster of suicides. However, after 1 year and 3 excess suicides, it starts to be clear that providing psychiatric treatment is important for prisoner wellbeing in death and life terms, let alone humanitarian issues.

This study generated statistically significant experimental data, that could be used to justify spending on psychiatrist’s salaries and secure treatment facilities for making treatment available to dangerous incarcerated prisoners.

As an example of the costs of delaying monitoring of an experiment on live human beings, this experiment could have been stopped after 2 years and reduced the roadkill element from 11.61 additional suicides down to about 6.50
Thus the delayed management reaction resulted in the unnecessary crushing of about 6.5 prisoners lives.
This delayed management reaction will have saved about 4 man-years of psychologist’s employment plus about 35 prisoner years of incarceration costs.
cost per year man-years total saving
psychiatrist’s employment 250000 x4 $1,000,000
prisoner incarceration cost 90000 35 3150000
cost of preparing ethical approval application 10000
experimental management saving, by not monitoring progress 20000
4180000
cost saving per life sacrificed by delayed management decision 836000.00
The net saving is about $4.2 million in wages at a cost of 5 lives ie about $836,000 per life unnecessarily lost.
2008 salary figures used

By exercising management discretion, about $4.2 million was saved, at the cost of about 5 lives. In analysis of road safety projects, they are authorised if they can save lives at less than about $2 million capital cost (2008, 2015 about $4 million).

References:
Suicide Prevention in NZ Prisons 1969 to 1994 (25 years) NZ Department of Justice
Explaining patterns of Suicide 2005 page 86 refers to prison suicide experimental deaths
________________________________________________________________________________________________________________________________________________