4
Another question is at what stage during the course of the disease
treatment is most effective. Despite the fact that mild or moderate mental
illness is far more common than severe mental disorders, accounting for as
much as three quarters of mental disorders (OECD, 2012), the bulk of the
research concerns the latter group. According to OECD (2012), evidence
suggests that the effectiveness of drug treatments for mental illness
increases with illness severity, whereas psychotherapy may be more
effective to treat milder mental disorders. It is also recognized that
symptom improvements do not necessarily translate into improved
employment outcomes. Mild or moderate mental disorders as well as pain-
related diseases may eventually turn into severe disorders if no treatment
is provided. In terms of employment outcomes, individuals may also benefit
more from treatment at an earlier stage of the sickness episode when the
attachment to the workplace is still strong. Johansson et al. (2011) find
that vocational rehabilitation is most effective if provided at the workplace.
The contribution of the paper is to study the effects of cognitive behavioral
therapy (CBT) for mental illness and multidisciplinary treatment (MDT) for
chronic pain not only on health related outcomes such as health care
consumption and drug prescriptions, but also on outcomes related to the
return to work, namely sick leave. The analysis also sheds light on
treatment effectiveness at different stages of the sickness episode. To
identify these effects, we exploit a government initiative (
the medical
rehabilitation guarantee
) in Sweden. From a public economics perspective,
it is crucial to investigate whether different types of initiatives aimed at
improving labor market outcomes of individuals at risk of becoming ill are
worthwhile to pursue. Since the two types of treatment are targeted at
different diagnosis groups, we do not per se make a comparison between
treatments. However, since the two types of treatment concern the two
main causes of work absence and the medical rehabilitation guarantee
affected both in a similar manner it is interesting to consider both
treatments in the same paper.
We focus on the treatments in Skåne, a region in the south of Sweden. The
reason for restricting the analysis to the Skåne region is the availability of
detailed individual health data together with information on certified CBT
and MDT clinics. The introduction of the medical rehabilitation guarantee
can be shown to cause an asymmetric and gradual expansion of these
treatments in the Skåne region. This implies that individuals with similar
potential to benefit from treatment to varying degrees were exposed to
treatment across time and residence, which means that individuals with the
same health status differ in their probability of being treated. In addition of
explaining why there is a common support in the selection of observables
estimator used in the evaluation the gradual and asymmetric expansion is
used as an instrument to test the validity of the maintained condition
independence assumption (cf. de Luna & Johansson, 2014). Estimation is
performed using propensity score matching, and we analyze the effects for
up to two years after treatment. To study the benefits of CBT and MDT at
different stages of the sickness episode, the analysis is performed
separately for individuals receiving treatments before entering sick leave
and when on sick leave. We also provide rough calculations of the public
finance implications up to two years after the initiation of treatment.