The Scottish Government

« Previous | Contents | Next »

Listen

ANNEX 2 THE ECONOMIC COSTS OF SUICIDE IN SCOTLAND IN 2004

Introduction

The profound impact of suicide both on the individuals themselves and their immediate family and friends are starkly clear. Avoiding such tragedies in itself is a strong justification for action. Policy makers are however faced with many competing claims as to how they should prioritise between different needs and wants. Having an understanding of the socio-economic impact of suicide can help inform this process. It should be stressed, however, that this is of limited use unless interventions of proven effectiveness, in either preventing suicides, or in alleviating some of the post event consequences, faced by families and friends are available.

The costs of suicide fall on everyone in society and can be substantial. Most obviously there are direct costs arising from demands placed on the emergency services, potential life saving interventions to be delivered within the health care system, investigations to be carried out by the police and coroner, and of course costs associated with funerals. For those individuals who survive suicide attempts, lengthy physical and psychological rehabilitation may follow.

There are also what economists call indirect costs. As a result of premature death, individuals lose the opportunity to contribute productively to the national economy, whether this be through paid work, voluntary activities, or family responsibilities such as looking after one's children or parents. The most fundamental impact of all, of course, is the loss of the opportunity to experience all that life holds as a result of suicide. The pain and grief that suicide can have on immediate family members and friends can be immense and long lasting. These very personal impacts are known by economists as ' intangible costs' because they are often hidden and difficult to value.

As part of this study, an estimate has been made of the costs of suicide in Scotland, informed by a literature review looking at previous studies worldwide. In addition to estimating the costs of suicide, we have also estimated the potential economic gains that might be realised should the rate of suicide by reduced as the result of an effective national suicide prevention strategy.

International estimates of the cost of suicide

Three bibliographic databases, Medline, Psychlit and Econlit were used to search for 'cost of illness' studies relating to suicide. In addition a Google search was conducted to identify governmental publications and grey literature. References of relevant publications were also searched to identify additional studies. There were no date or language restrictions on the search. Cost here have been converted to £ sterling and use 2005 prices.

This review indicated that, despite these profound human and socio-economic costs, surprisingly few international studies appear to have estimated the total population wide costs of suicide, although a number of studies have sought to put a value on suicides associated with any one specific mental disorder, as for instance with depression in one recent study in England (Thomas & Morris 2003). Of those studies of most relevance to a Scottish context, in the Canadian province of New Brunswick average direct and indirect costs of each suicide in 1996 were estimated to be £443, 076 ( CAN$1,019, 210) (Clayton & Barceló 2000). In New Zealand in 2002 the estimate including intangible costs was £1,158,768 ( NZ$ 3,094, 243) per suicide. (O'Dea & Tucker 2005). In Ireland, using a similar approach, costs in 2002 were estimated to £1,402, 438 (€1,982,667) per suicide (Kennelly et al. 2005).

Estimating the costs of suicide in Scotland

Methods

We have undertaken an incidence based costing study - that is, we have estimated the total lost lifetime costs for all suicides in Scotland that occurred in one year, 2004. Data on suicides and deaths of undetermined cause, broken down by age and gender were obtained from the General Register Office for Scotland. We have included direct, indirect and intangible costs of suicide in our analysis. All data are reported in 2005 prices. It should be noted that this analysis can be viewed as a conservative estimate of costs, as we have not included the costs of non fatal deliberate self-harm events.

Indirect costs

Indirect costs can be valued in a number of different ways - the principle method being to use what is known to economists as the 'human capital' approach. Human capital models have been applied, for example, to measure the cost of suicide in Ireland (Kennelly et al 2005) , the United States (Kashner et al. 2000), New Brunswick (Clayton and Barceló, 2000) This assumes that the lost opportunity for individuals to contribute to the national economy (lost productivity) is equivalent to the lost gross lifetime earnings of an individual.

In the case of sometimes very premature mortality from suicide this can mean many years of potential participation in the labour market foregone. Of course future employment patterns may be very different from those seen today as economies change, however we have made the assumption that employment patterns seen today will continue over a period of as much as 50 years. This is potentially a significant limitation - moreover the demographic profile of the population is changing and it is likely that the retirement age of those entering the labour market in the UK today will be higher than 65. This might also mean that future productivity losses would be higher than those mentioned here, for instance in the state retirement age were to raise to say 67 or 68.

Ideally, in estimating these productivity costs we would use actual wages foregone by those lost to suicide. In the absence of such data we have used the average gross wage rate, adjusted for gender and age group. We have followed the convention adopted by Kennelly et al and made an adjustment to labour force participation. In our baseline analysis we have assumed that this would be 2.25% than in the general population. We recognise that this adjustment is controversial and conservative as it assumes that a significant proportion of individuals who complete suicide are more likely to have mental health problems. We have also estimated the costs of suicide without making this adjustment and this is also reported in the section on sensitivity analysis.

The impact of mental health problems in childhood or early adulthood can also have an impact on success in school, higher education and longer term career prospects which again may limit individuals potential wage rates. It can be argued that the productivity levels of this group will be lower than that for the population as a whole due to increased levels of absenteeism, presenteeism (reduced performance when at work) and early retirement from the labour force (Almond & Healey 2003; Dewa et al. 2004). We have not however made an adjustment for this lower potential rate of productivity in our analysis. In some instances this impact may in any event be temporary: for instance around two thirds of individuals will recover from depressive disorders.

In addition to lost working time there are also productivity losses for the retired , the unemployed, those engaged in voluntary activities and those with family/home responsibilities. All of these individuals contribute to the national economy. We have used UK survey data on time spent on these activities to estimate the potential loss to society. (Short 2001). Life tables have been used to provide an estimate of the average life expectancy of an individuals by different ages and genders, so as to capture the costs beyond 65 (Government Actuary's Department 2005).

One limitation of our analysis is that we have not included any productivity losses experienced by the family and friends of those who take their own lives. In general these might be expected to be relatively short-term in nature, but such events can trigger potential mental health problems in individuals. Moreover it may be the case that the partner of someone who commits suicide may have to give up work to care for children, for instance. Ideally such costs should be estimated - at present we are unaware of any study that has incorporated these costs. Another limitation of our analysis which potentially might overvalue costs is that we have not adjusted our estimate of productivity losses to take account of the probability of death from all other causes below the age of 65.

Is it necessary to estimate long-term productivity losses? The friction cost debate

Some economists would argue that the calculation of long-term productivity costs are not necessary. They argue that no economy is in a position of full employment and if individuals fall out of the workforce, in this case because of suicide, they will be replaced by someone from the existing pool of spare labour. They also further argue that human inputs might also be substituted by additional capital investment in new equipment. They suggest that only the costs of one year's lost wages plus the costs of recruiting a replacement worker or investing in capital should be included in any economic analysis as in effect this is a temporary or 'friction' cost (Koopmanschap et al. 1995; Goeree et al. 1999). We consider the impact of this approach on overall estimates of the cost of suicide in sensitivity analysis.

Direct costs

We have insufficient data at this stage to estimate the total direct costs of suicide in Scotland (although some data has been collected on this). We have adopted the convention used by Kennelly at al in their analysis of the costs of suicide in Ireland and also used estimates reported in the international literature to provide a proxy value for these costs. In their analysis Kennelly et al followed the assumption set out by Clayton and Barceló, based on work undertaken in New Brunswick, that the direct costs associated with suicide mortality were equivalent 0.67 of one per cent of the indirect costs of suicide. In New Zealand a detailed attempt was made to estimate the costs of suicide including police and fire service attendance, funeral directors and funeral expenses victim support as well as the cost of post mortems and inquests. Direct costs in their analysis were considerably higher at 2.33 per cent of indirect costs; we have used this estimate in our base line analysis, and consider the impact of using the New Brunswick valuation method in sensitivity analysis.

Intangible costs of suicide

In addition to productivity losses and direct costs of suicide there are also intangible human costs. These can include grief experienced by relatives as well as the value individuals would generally place on years of life lost, with all their opportunity to enjoy life experiences. (Some have argued that the individual who completes suicide places a greater value on death than on life but we have decided not to follow this approach in our analysis). Economists have adopted different approaches to trying to place a value on these type of intangible costs - in some instances, individuals may de facto reveal their preferences - for instance the premium on wages to work in dangerous professions such as in the nuclear industry or in mining has sometimes been used to estimate the value of life, whilst other have used the level of compensation awarded by juries to individuals in the case of compensation claims for death and disabling accidents.

Another way in which economists have sought to estimate such values is through the use of elicitation techniques to measure the willingness to pay of individuals to fund interventions to reduce suicides. No such studies have as yet been conducted in respect of suicide, although work is currently underway in Ireland. A good proxy for this however may be the use of willingness to pay studies conducted in respect of road accidents. In our analysis we have made use of data from the UK Department of Transport's 2004 estimate that the public have placed on the value of the prevention of road fatalities (Department of Transport 2004). This provides specific data on the intangible costs of death including pain, grief and suffering and lost life experiences for both the potential victim and their relatives. This data source has previously been used in estimates of the cost of suicide in England and as it separates productivity losses from deaths from the human costs of death, it can avoids any double counting of these costs.

Results

Indirect costs

Waged time

Data on deaths from intentional injuries and undetermined deaths in 2004 were categorised by age and gender. As the distribution of deaths within age cohorts was not known, it was assumed that on average death would occur midway through each age category, with the exception of those aged 0-14. In this latter case, as there were only two deaths in 2004, it was assumed for simplicities sake, that these occurred at the age of 14 and costs of lost productivity would thus cover the whole of the first age category from 16 to 24. Productivity losses in this case would not begin until 2006 when these children would have reached the age of 16.

It was assumed that individuals would have earned average UK wide gross wage rates for each age and gender category. These wage rates taken account of the differences between those who work full and part time (Office for National Statistics 2004). However they do not take into account the income of the self employed and thus this is not factored into our analysis. The likelihood of being in work was determined using data on participation rates in the labour market by age and gender in Scotland (Office for National Statistics (Scotland) 2004). Average wages have been adjusted to take account of real wage growth over time and in our model this is set at a rate of 2% as the Bank of England in economic forecasts have observed that since 2001 this has remained below 3% and often well below 3%.To reflect the concept of positive time preference for money, that is a pound today is generally preferred to a pound at some time in the future, all future earnings have been discounted using the Treasury Green Book's recommended discount rate of 3.5% (H M Treasury 2003). Differences to the discount rate have been considered in sensitivity analysis.

To calculate lifetime lost earnings from premature death it was assumed that if individuals had proceeded through the life cycle they would have had the opportunity to earn the average gross wage in progressive age gender categories. For instance an individual who died at the age of 23 in 2004 would at the age of 30 in 2011 be projected to earn the average wage for men in the 25 - 34 age bracket and similarly at the age of 37 in 2018 they would have had the opportunity to earn the average wage rate in the 35-44 age group.

Table A6 presents the results broken down by age and gender group; overall lost lifetime output for men and women in Scotland due to premature death from suicide were £201 and £31 million respectively. As the majority of suicides occur in people between the ages of 35 and 44 this age group account for 35% of all lost productivity costs of suicide.

Table A6 Value of lost market output due to suicide in Scotland in 2004 by age and gender (2005 prices)

Non waged outputs

Table A7 provides an estimate of the costs of non waged productivity losses as a result of suicide. The relative similarity in costs reflects both the greater contribution of women to these activities and their longer life expectancy. Time spent on daily unpaid household activities were taken from a time survey undertaken as part of the UK Household Satellite Accounts (Short 2001). This survey, using data from 1999, estimated that on average men spend 90 minutes a day on household activities and women 196 minutes. This covered the task of cleaning and maintaining one's house, cooking, washing, caring for children and adults, charitable work and attendance and religious, political and other meetings. Time spent gardening and looking after pets were not included in the analysis. We have followed the assumption of Kennelly et al that household production would drop by 25% at age 65 for both men and women. Household production time was valued using the three minimum wage rates set in 2004 of £3 for 16-17 year olds, £4.10 for 18-21 and £4.85 for those aged 22 or over. The value of household production was assumed to grow at the same rate of 2% per annum as the predicted increase in real wages. Again we have discounted future household production at a rate of 3.5% per annum.

Table A7 Value of lost non-waged output due to suicide in Scotland 2004 by age and gender (2005 prices)

Direct costs

As shown in table A8, if direct costs as a proportion of indirect costs in Scotland are similar to those found in New Zealand, it is estimated that the direct costs for suicides of men and women respectively were approximately £5.66 and £1.44 million respectively. Using the more conservative estimate advocated by Clayton and Barceló these costs would be £1.62 million and £0.41 million.

Estimating the human costs of suicide

In 2004 using figures from the UK Department of Transport the intangible human costs of suicide were estimated to be £927, 035 per fatality. Caution must however be exercised in making use of this estimate, as we do not know how comparable the risk level used for road accidents is with suicide. Moreover we might hypothesise that in addition to the pain and grief that relatives may endure regardless of whether a fatality occurs through suicide or road accident, there may also be additional costs associated with dealing with the stigma, shame and guilt of suicide.

Overall costs of suicide in Scotland

Table A8 provides a breakdown of the total costs of suicide using our baseline assumptions. In total the lifetime costs of suicides in 2004 are estimated to be almost £1.08 billion of which 75% of costs would be due to suicides by men. This represents an average cost of £1.29 million per completed suicide and are comparable to those reported in New Zealand and Ireland. By far the largest single component of the total costs of suicide (more than 70%) are the intangible human costs experienced by families; indirect lost productivity costs account for 21% of the total costs.

Again we should emphasise that this estimate is conservative as we do not include the costs of non-fatal deliberate self-harm events. Their costs can be substantial and in Ireland were recently estimated to be almost £22 million (â'¬31 million). (Kennelly et al. 2005) These include both the direct health and other costs from dealing with deliberate self-harm events plus the loss of both waged and non waged contributions to the economy.

Table A8 Total costs of suicide in Scotland in 2004 (2005 prices)

What would the potential socio-economic benefits of a reduction in the rate of suicide?

Using our baseline assumptions if a 20% reduction in the overall level of suicide (based on 2004 suicide rates) were to be achieved for any one year then the lifetime costs averted for all suicides avoided during that year those would amount to some £216 million. Even just a 1% reduction in the suicide rate would reduce overall costs by £10.8 million over the lost lifetimes of these individuals.

Given the modest cost of investing in Choose Life compared to the potential lives that may be saved and costs averted, if effective interventions that reduce the rate of suicide at reasonable cost can be identified, investing in Choose Life is likely to be cost saving that is leading to better health outcomes and also reducing overall cost.

Of course as we have indicated the total costs of suicide vary by gender and age of the population - clearly if most significant progress was made in reducing suicides among younger populations then cost savings would be greater. However from an ethical perspective it may be deemed inappropriate to target suicide interventions simply on the basis of lifetime costs that might be avoided. It might also be argued that some of these costs are not in fact averted, but merely delayed - for instance some of the costs associated with dealing with death; moreover some costs might even be greater - for instance treating the complications of poor health in older age. Again in conventional economic analysis when looking at preventive interventions it is usually deemed unethical to argue against preventive interventions on the basis that individuals might potentially face later in life.

Sensitivity analysis

It should be noted that these cost estimates of suicide can vary substantially. By far the most significant decrease in costs would be observed if we adopted the 'friction' cost approach and only included the productivity costs for the first year following death in our analysis. However overall these costs only account for just over 20% of total costs. The value of intangible human costs is the key factor; the appropriateness of using the Department of Transport's estimation can only truly be judged if an elicitation exercise to place a value on these costs were to be undertaken in Scotland. While this may be an expensive undertaking much might be learnt from analysis currently underway in Ireland by Kennelly and colleagues which is anticipated to be completed in late 2006.

One important factor in sensitivity analysis can be the discount rate used to reflect the present value of present earnings. In this case the overall costs remain robust even at a high discount rate. Table A9 indicates how overall costs will change if the discount rate is varied - however even discounting at a rate of 8% total costs remain well in excess of £0.9 billion. The adjustment to labour force participation has little impact on costs. With no adjustment these costs increase slightly to £1.085 billion. If participation is adjusted downwards by 10% compared with the general population then costs would still be £1.06 billion. Again there is little difference in overall costs if the assumption about decreased productivity after the age of 65 were dropped. This would see total costs rise to £1.085 billion, while if productivity halved after age 65 costs would still be more than £1.07 billion.

Table A9 Impact of discount rates on overall total costs

Conclusions

The lifetime costs associated with completed suicides in any one year are substantial and many time in excess of the annual investment in Choose Life. From a societal perspective, taking account not only of direct costs to health and other services, but also the loss of contribution to the economy and intangible costs including grief and the loss of the future opportunity to enjoy life, even a modest level of success in reducing the rate of suicide is likely to be highly cost effective and potentially cost saving.

References

Almond, S. and Healey, A. (2003). Mental health and absence from work: new evidence from the UK Quarterly Labour Force Survey . Work, Employment and Society,17(4): 731-42.

Clayton, D. and Barceló, A. (2000). The cost of suicide mortality in New Brunswick, 1996. Chronic Diseases in Canada, 20(2): 89-95.

Department of Transport (2004). Highways Economic Note No 1. London, Department of Transport.

Dewa, C. S., Lesage, A., Goering, P. and Caveen, M. (2004). Nature and prevalence of mental illness in the workplace. HealthcarePapers,5(2): 12-25.

Goeree, R., O'Brien, B. J., Blackhouse, G., Agro, K. and Goering, P. (1999). The valuation of productivity costs due to premature mortality: a comparison of the human-capital and friction-cost methods for schizophrenia. Can J Psychiatry, 44(5): 455-63.

Government Actuary's Department (2005). Life expectancy tables for Scotland by age and gender 2002-2004. London, Government Actuary's Department.

H M Treasury (2003). Green Book. Appraisal and evaluation in central government. London: HM Treasury.

Kashner, M. T., Shoaf, T. and Rush, A. J. (2000). The economic burden of suicide in the United States in the year 2000. Trends in Evidence-Based Neuro-Psychiatry, 2(3): 44 -48.

Kennelly, B., Ennis, J. and O'Shea, E. (2005). Economic cost of suicide and deliberate self harm. Reach out. National strategy for action on suicide prevention 2005 - 2014. Dublin: Department of Health and Children.

Koopmanschap, M. A., Rutten, F. F., van Ineveld, B. M. and van Roijen, L. (1995). The friction cost method for measuring indirect costs of disease. J Health Econ,14(2): 171-89.

O'Dea, D. and Tucker, S. (2005). The Cost of Suicide to Society. Wellington, New Zealand Ministry of Health.

Office for National Statistics (2004 ). Annual survey of hours and earnings 2004. London: Office for National Statistics.

Office for National Statistics (Scotland) (2004). Economic activity by age band and gender. Scotland 1996 - 2003. Edinburgh: Office for National Statistics (Scotland).

Short, S. (2001). Time Use Data in the Household Satellite Account: October 2000 . Economic Trends, 563: 47-55.

Thomas, C. and Morris, S. (2003). Cost of depression among adults in England in 2000. British Journal of Psychiatry, 183: 514-19.

« Previous | Contents | Next »