23 - 32: Counting All the Costs: The Economic Impact of Comorbidity

This chapter provides an overview on what is known about the economic impacts of comorbid physical and mental disorders. As later chapters discuss, illnesses such as psychoses, bipolar disorder and major depression can increase the risk of many physical health problems, including obesity, diabetes, cardiovascular disease, chronic obstructive pulmonary disorder and some infectious diseases. Poor physical health, for instance related to cancer or musculoskeletal health problems, may also increase the risk of developing mental health problems such as depression and anxiety-related disorders. As a result, there is likely to be a considerable overlap in those living with long-term physical and mental health problems. In England, for example, it has been estimated that up to 46% of all those with mental health problems also have some chronic physical health problem, while 30% of those with long-term physical conditions also have a mental health problem [1].

All of these comorbidities will have economic consequences that impact right across society. Some of these costs may potentially be avoidable. Thus, there is a growing interest from policy makers, particularly in countries where health and social care spending is under great strain, about the potential importance of actions that can effectively prevent and/or reduce the impacts of comorbidity.

This chapter therefore begins by briefly describing the concept of economic cost, before going on to look at different examples of the costs of comorbidity. It ends by examining how better information on the costs of comorbidity can be used to inform economic arguments for investment in actions to prevent and alleviate some of these morbidities, and how this evidence base may be strengthened further.

Before going further, it is worth briefly explaining what economists mean by cost, as this is about much more than simple monetary cost. In fact economists often talk about three different components of cost. There may be increased direct costs to the healthcare systems associated with the management of multiple health problems, for instance if mental disorders exacerbate the risk of adverse events and complications in chronic physical health problems or if recovery times are prolonged. This could include the salary costs of healthcare staff, the cost of medicines and the use of diagnostic procedures. In England it has been estimated that as much as GBP 1 in every GBP 8 spent on chronic long-term conditions is due to adverse outcomes arising from poor mental health [1]. There may also be direct costs for the provision of services that fall on other sectors such as social care services. In some cases there might be additional costs falling on other sectors, such as for home modifications due to physical disabilities.

There are also ‘indirect' costs, which focus on the lost opportunity to contribute to economic productivity, such as when individuals are absent from the labour market due to poor health or premature death. Productivity costs for many mental disorders already account for more than 60% of all costs because of the low rate of participation in employment [2]. A major reason for these costs is the much higher rate of mortality due to poor physical health. For instance, one study of men and women with severe mental disorders in Denmark, Finland and Sweden reported that they lived between 20 and 15 years less than the general population [3].

Other forms of productivity loss also occur. Comorbid health problems may reduce participation in school or university, potentially impacting on career possibilities. In fact, the long-term adverse costs to the economy due to children with mental health problems not obtaining employment in adulthood has been one key reason for substantial policy interest in measures to help support the health and well-being of children from a very young age [4]. Family members may also give up some of their time from employment or other activities because of the need to provide care and support to a loved one.

The third cost category is known as ‘intangible' because it refers to impacts that are often difficult to quantify and value. Examples include the stigma associated with mental illness, communicable disease or physical disabilities, as well as the grief experienced by families as a result of an unexpected death.

Remarkably, health economists have not focused much of their energies on assessing the economic impact of comorbidity in any area of health, let alone looking at the issue of mental and physical comorbidity. This may be due to the difficulties in attributing healthcare costs to any specific comorbidity, as well as the separation of the way in which mental and physical health services are organised in many countries. There is, however, an increasing, albeit still small number of studies, which have looked at the economic impacts of comorbid physical and mental health problems. The majority of these examples are from a US context and thought must be given on how they translate to other contexts. Nonetheless, they illustrate potentially substantial costs to healthcare systems and society as a whole that might be avoided through early identification of potential risk factors and early intervention to mitigate the effects of comorbidity. The next sections provide an overview of some of these economic analyses, looking at economic impacts within and beyond healthcare systems.

Comorbidities between physical and mental health problems provide major challenges to healthcare systems; they can worsen health outcomes, prolong recovery time and thus exacerbate costs to healthcare systems. Some, as shown in table 1, look at the additional excess costs to the healthcare system of a comorbid physical health problem compared to having a mental health problem alone, while others, as illustrated in table 2, focus on the additional excess costs of a comorbid mental health problem to having a physical health problem alone. Much of this evidence is from the USA, but it provides valuable insights on the extra costs of comorbidities in other countries, suggesting that there is the potential to avoid substantial costs to healthcare systems through early identification and intervention.

Several studies focus on schizophrenia and physical health problems. Analysis of healthcare costs for more than 1,400 individuals with schizophrenia who participated in the 18-month CATIE trial (Clinical Antipsychotic Trials of Intervention Effectiveness) in the USA reported statistically significant 25% higher costs for those who were obese [5]. Data from the US Medical Expenditure Panel Survey of more than 571,000 individuals in 2001 and 2002 found annual healthcare costs for people living with schizophrenia alone of USD 5,990, compared with USD 11,611, 10,803, 12,292 and 10,415 for those with comorbid diabetes, dyslipidaemia, hypertension or heart disease, respectively [6].

The healthcare costs of more than 31,000 older people with and without schizophrenia were analysed for the 10-year period from 1998 to 2008. Mean healthcare costs were significantly higher in the schizophrenia group; a key driver of greater healthcare costs was the significantly higher rate of physical health problems, including congestive heart failure (45.1 vs. 38.8%), chronic obstructive pulmonary disease (52.7 vs. 41.4%), hypothyroidism (36.7 vs. 26.7%) and dementia (64.5 vs. 32.1%) [7].

Turning to bipolar disorder, in the USA the medical records (spanning 1 year) of more than 28,000 people with bipolar disorder were compared with matched controls without any mental health problems [8]. Those with bipolar disorder had a significantly higher prevalence of metabolic comorbidities than the general population (37 vs. 30%). Annual healthcare costs for metabolic conditions were twice those of controls (USD 531 vs. 233). The bipolar cohort also had significantly higher overall medical service and prescription drug costs than those of the control cohort (USD 12,764 vs. 3,140). Prescription medication costs for metabolic conditions were also higher, with bipolar cohort per-patient costs of USD 571 versus 301 for the control cohort. Analysis of data on 67,000 members of a health insurance fund in seven US states also suggests that 67% of total healthcare costs of bipolar disorder were related to the treatment of comorbid physical health conditions [9].

Comorbid depression or anxiety disorders and physical health problems have also been associated with higher levels of cost to healthcare systems. One US study reported that the costs of 11 chronic health problems are significantly greater when an individual has comorbid depression. Costs related to diabetes, coronary artery disease and congestive heart failure were approximately twice the costs of individuals without depression [10].

Two reviews, one with 27 [11] and the other with 41 largely US-set studies [12], looked at the impact on healthcare resource utilisation of comorbid diabetes and depression. Both reviews consistently showed increased healthcare resource use to manage diabetes in people with depression. For example, in one study of more than 400,000 adults with diabetes in the USA, the costs of depression increased mean annual healthcare costs from USD 11,000 to 19,000 [13], while in Australia, health service use by people with comorbid diabetes and depression was 49% higher compared to those with diabetes alone [14]. In another US study, the healthcare costs of managing diabetes over a 6-month period were found to be between 50 and 75% higher in people with major depression than in people with diabetes alone [15]. Furthermore, this study observed a significant difference in the costs of managing one or more complications of diabetes in people with major depression compared to those with sub-clinical thresholds of depression. Other studies also point to higher costs of managing complications. In the USA, the costs of managing complications of diabetes, such as diabetic neuropathy, in people with comorbid depression have also been shown to be significantly greater than in those without depression [16].

Another US study looked at the healthcare costs of more than 14,000 people with depression, congestive heart failure, or both [17]. People who also had depression had significantly higher total annual healthcare costs than those without: USD 20,046 versus 11,956. Costs increased with the severity of comorbidity, but mental healthcare costs accounted for less than 1% of total healthcare costs.

Outside of the USA, significantly increased costs to healthcare systems have also been observed when comorbidity involved depression. English data from more than 86,000 patients in the General Practice Research Database were used to assess whether comorbidity increases the costs of managing patients in primary care [18]. The study found that 20% of all patients had more than one chronic health problem and that all instances of comorbidity increased the costs of primary healthcare compared to the costs of managing these conditions separately. Depression was found to be the most important cost-increasing condition, significantly increasing costs in adult patients of all ages when comorbid with a wide range of conditions. For instance, the costs of managing comorbid depression and asthma or diabetes were greater for patients of all ages, whilst depression and comorbid cancer were associated with significantly higher costs in people aged 40-59 years. Depression and comorbid obesity, heart failure or epilepsy were associated with significantly higher primary care costs in patients aged over 60 years. Increases in cost ranged from GBP 269 for people aged 40-59 years with depression and cancer to GBP 2,817 in people with comorbid depression and obesity aged 40-59 years. In younger adults, mean additional costs related to asthma and comorbid depression were GBP 1,257, while comorbid diabetes and depression increased costs by GBP 2,133 in people aged over 60 years.

Other examples of additional costs from around the world can be identified. In Singapore, adults attending a specialist diabetes centre over a 1-year period who also had depression had a 30% chance of hospitalisation compared to 10% in the diabetes-alone group. They were four times more likely to be hospitalised for non-psychiatric conditions and three times more likely to be hospitalised for complications of diabetes [19].

In Hungary, a survey of more than 12,000 people looking at their use of health services over a 12-month period found that those with comorbid diabetes and depression had a 2.6 times greater risk of a lengthy period of hospitalisation and had almost double the risk of multiple hospital admissions compared to people with diabetes alone [20]. Another study looked at the impacts of physical comorbidity on healthcare costs for 65,000 people receiving primary care in Spain in 2004. Individuals with a depressive disorder had a significantly greater number of comorbid conditions or risk factors, including obesity, dyslipidaemia and smoking per year compared to other primary care service users (7.4 conditions vs. 4.3). Overall the annual costs of care were EUR 1,084 and 684 per patient in the comorbid and control populations, respectively [21]. In the UK, the economic impacts of smoking in people with mental disorders have been estimated to cost primary and secondary care service providers GBP 720 million per annum in treating smoking-related disease. The study also estimated that about a third of all cigarettes smoked in England are smoked by people with a mental disorder. This could mean that there are 2.6 million avoidable hospital admissions, 3.1 million avoidable primary care consultations and 18.8 million prescriptions that can be avoided each year [22].

In Germany, analysis at one teaching hospital over a 2-year period again reported that the average total costs of hospitalisation for people with cardiovascular disease alone compared to those with psychiatric comorbidity (largely depression and anxiety disorders) differed significantly (EUR 5,142 vs. 7,663). The average length of stay for patients with comorbidity was 13.2 days compared to 8.9 for patients with cardiovascular disease only [23]. Furthermore, this paper highlighted that the funding system in that hospital did not fully cover the costs of comorbidity, which means that patients might not receive appropriate levels of care.

There is also some limited information looking at the association between healthcare costs and comorbid asthma and mental disorders. One systematic review found 20 studies, largely focused on depression or anxiety disorders and asthma. It reported increased rates of hospitalisation, emergency department visits and visits to primary care practitioners in people with asthma and a mental disorder [24]. In the USA, a telephone survey of adolescents (aged 11-17 years) with asthma found that those assessed to have depressive disorders as well had on average 51% higher healthcare costs [25]. Most of these additional costs were related to non-asthma and non-mental health-related healthcare costs.

There appear to be fewer estimates of the impacts of comorbidity beyond the healthcare system. Published estimates concentrate on indirect costs - mainly productivity losses from employment due to absenteeism, with much less discussion of poor performance at work (presenteeism); there does not appear to be much information on the intangible costs of comorbidity. One review looked at the impact on participation in employment of people with coronary artery disease and mental disorders [26]. Only 13 studies were identified, 10 of which focused on comorbid depression. The review concluded that people with comorbid depression had a reduced likelihood of returning to work following the onset of illness (odds ratio 0.37) compared to people with coronary artery disease alone. A systematic review on comorbid diabetes and depression identified 11 studies that looked at the impacts on productivity, but only two of these studies assigned a monetary value to productivity losses and none included losses from premature mortality or informal care [12].

Data from a cross-sectional survey of 78,000 workers in Australia [27] show higher relative risks of both absenteeism and poor functioning while at work in individuals with comorbid psychological distress and physical health problems compared to those with physical health problems only. For instance, compared to people with no health problems, the risk of absenteeism from work was 33% higher for people who were experiencing psychological distress alongside obesity, and 27% higher for those who had high levels of cholesterol. Rates of presenteeism were between 2.5 and 5 times greater in populations with comorbid asthma, obesity, arthritis, diabetes and high cholesterol compared with the reference population.

Data from the 2007 Australian National Survey of Mental Health and Wellbeing (n = 8,841) have also been used to compare work functioning and absenteeism rates in people with depression, cardiovascular disease, or both conditions, with a disease-free population [28]. As table 3 shows, compared to a population with neither condition, and adjusted for various social and demographic characteristics, the odds ratio for people with comorbid depression and cardiovascular disease participating in work was significantly lower at just 0.4. This compared with odds ratios for work participation of 0.8 for depression or cardiovascular disease only. The comorbid group were also 10 times more likely to experience poor work functioning compared to the healthy workforce, 9 times greater than for people with cardiovascular disease alone and 2.5 times greater than for people with depression alone. Rates of absenteeism were also significantly higher for the comorbid group.

Data from a large Canadian survey of more than 130,000 people found (even after adjusting for socio-demographic characteristics, alcohol dependence and chronic physical illness burden), that the presence of comorbid major depressive disorders was associated with twice the likelihood of healthcare utilisation, and increased functional disability and work absence compared to the presence of a chronic physical illness without comorbid depression [29]. Significant increases in resource use and productivity losses have also been reported in a population survey of more than 12,000 adults in Hungary [20]. People with comorbid diabetes and depression were more than twice as likely to have lengthier stays in hospital (>20 days) and to have more hospital admissions. They were also more than 3 times as likely to have a prolonged absence (>10 days) from paid work and to be unemployed.

In Finland, analysis of certified sickness absence in 33,000 public sector employees reported that non-cardiovascular comorbid conditions for employees with diabetes, including depression, accounted for over 50% of excess risk of sickness absence [30]. A US study of a manufacturing company with 15,000 employees reported 13.5 sick days per annum on average due to depression and physical comorbidity (diabetes, heart disease, hypertension or back problems) compared to 6.6 and 8.8 days for those with a physical condition or depression alone. Total mean costs, including healthcare and disability, per employee with comorbidity were USD 7,906. This is conservative as the costs of poor performance (presenteeism) at work were not included [31].

Data from the UK Psychiatric Morbidity Survey also indicate that over a year individuals with diabetes and depression were 7.7 and 5.3 times more likely to take sick days and have their work and other activities impaired by poor health compared to people with diabetes alone [32]. The number of working days lost was also found to be significantly higher in people with comorbid diabetes and depression compared to diabetes alone in Singapore - 1.9 versus 1.4 working days lost over a 3-month period [19].

This chapter has illustrated that there are substantial economic costs associated with comorbidity. These costs fall both within and beyond healthcare systems. It is not, however, enough to identify these costs. Given that healthcare budget holders have to make difficult choices on how to allocate scarce resources to mental health and other services, it is critical is to identify cost-effective ways of reducing the risks or consequences of comorbidity.

Potentially, given the high additional costs of comorbidity, even modest success in reducing its prevalence may have economic payoffs. Policy makers will want to know whether early intervention and investment in actions to protect the physical health of people with mental health problems are a cost-effective use of resources, generating benefits not only to healthcare systems, but also more widely, perhaps reducing rates of absenteeism from work or the need for informal family care. They may also want to know whether a greater focus on managing the mental health of people with chronic physical health problems is a cost-effective way of improving outcomes.

There is a need to strengthen the evidence base, and in particular to look more at the cost-effectiveness of interventions outside of a US context. To date, the evidence on cost-effective approaches to manage comorbidity is modest, although it supports investment, suggesting that a number of cost-effective interventions are available. For instance, empirical work and modelling studies focusing on comorbid diabetes and depression includes several studies that have highlighted the cost-effectiveness of better integration between psychological and diabetes care [33,34,35]. Only a handful of studies have considered the economic, as well as the clinical impact, of interventions to promote and protect the physical health of people with mental disorders [36].

One way of strengthening the evidence base in the short-term is to make use of economic modelling techniques to take existing data on the effectiveness of different actions to tackle comorbidity and model plausible costs and benefits under different scenarios in different country contexts and over different time periods [37]. One recent example of a modelling approach suggests that it may be cost-effective to invest in group-based actions to promote better weight management in people with schizophrenia and diabetes [38]. In the mid to longer term it would be helpful to embed economic analysis, collecting data on resource use, costs and economic consequences, into evaluations of interventions to tackle comorbidity. This can facilitate better comparability across different interventions and help policymakers prioritise the way in which they allocate resources.

Where mental health services are largely separate from physical health services, the provision of seamless care for both physical and mental health problems becomes more difficult to achieve. Therefore, another important step may be to evaluate the cost-effectiveness of different incentives and organisational structures to encourage a more collaborative approach to the detection and management of comorbidity; there are often substantive silos in both the financing and organisation of mental health and general health services [39].

This chapter has illustrated that economics can help provide powerful arguments to support a case for tackling comorbidity. It has shown that there are substantial impacts both within and beyond health systems linked to comorbidity. For too long, much of these costs remained hidden; this is now changing and in the future more information will be available on these additional costs as well. This hopefully will provide more impetus for actions to address an issue which leads not only to some avoidable costs to health and society, but also has profound personal human consequences.