Exercise for people with severe mental illness
Systematic review
|Published
We evaluated the effect of exercise for people with severe mental illness, and included two systematic reviews. One of these reviews included eight primary studies with a total of 375 participants and the second included ten primary studies with a total of 383 participants.
Key message
People with severe mental illness will in average die about 10 to 20 years earlier than the general population. The excess mortality is mostly associated with socio-economic factors, use of medication and unhealty lifestyle.
We evaluated the effect of exercise for people with severe mental illness, and included two systematic reviews. One of these reviews included eight primary studies with a total of 375 participants and the second included ten primary studies with a total of 383 participants.
For people with severe mental illnesses such as schizofrenia, schizoaffective disorder, psychosis and bipolar disorder:
- Exercise may lead to little or no change in body mass index or weight.
- The effect of exercise on physical fitness or symptoms of depression and anxiety is uncertain.
For people with schizophrenia:
- Exercise is probably associated with a small improvement in global cognition and working memory.
- Exercise may lead to a moderate improvement of attention and social cognition.
- Exercise may lead to little or no difference on processing speed.
- The effect of exercise on verbal and visual learning and memory, and reasoning and problem solving is uncertain.
Summary
Background
People with severe mental illness have, in average, between 10 and 20 years shorter lifespan than the general population. Risk factors for premature death are linked to socioeconomic status, the use of medications and unhealthy lifestyle such as smoking, unhealthy diet and physical inactivity. These types of unhealthy lifestyle seems to be particuraly common in people with serious mental illness.
Studies evaluating the association between physical activity and premature death in the general population have found that physical activity reduces the risk of early death. The Word Health Organization recommends a minimum of 150 minutes with physical activity of moderate intensity during each week for adults 18 years or older. For further health benefits, it is recommended to increase intensity or the duration of physical activity or exercise. Exercise is defined as planned and structured regular physical activity, aimed to improve or maintain the level of physical fitness.
The Norwegian Directorate of Health has developed guidelines aimed at local health workers and for people with severe mental illness. These guidelines recommend that physical activity should be part of an overall treatment plan for people with mental illness.
Our aim was to develop an overview of systematic reviews on the effect of exercise for people with severe mental illness.
Method
We conducted a systematic literature search for systematic reviews published since 2012 in the following databases: MEDLINE, PsychINFO, Cochrane Library, HTA, DARE and Epistemonikos. The search was performed in March 2018.
Two researchers (VU and HHH) independently screened references and identified relevant systematic reviews using a form with the inclusion criteria. We assessed the quality of the systematic reviews by the use of our checklist and summarized the results from each systematic review that was included separately. We assessed our confidence in the results for each outcome following the GRADE approach.
Results
Our systematic literature search resulted in 3696 references. 3628 of these were excluded, based on title and abstract. After removing duplicate publications, we assessed 61 references in full text. 58 of these were excluded for reasons that the populations did not have serious mental illness, the intervention did not qualify as exercise, lack of estimates and confidence intervals, or that the systematic review was not of high methodological quality.
We included two systematic reviews of high methodological quality. One of these reviews evaluated the effect of moderate intensity exercise in people with severe mental illness on the outcomes related to physical and mental health. This systematic review included eight primary studies with a total of 375 particiants with schizophrenia, bipolar disorder or psychosis. Evidence with low confidence shows that exercise is associated with limited effects on body mass index (SMD -0.24 [95% CI -0.56 – 0.08]) or weight (SMD 0.13 [95% CI -0.32 – 0.58]). The effect of exercise on physical fitness or symptoms of anxiety or depression is uncertain.
The second systematic review evaluated the effects of aerobic exercise on cognitive functioning in people with schizophrenia. This review included ten primary studies with a total of 383 participants, and suggested that exercise is problably associated with a small improvement in global cognition (Hedge’s g 0,33 [95% CI 0,13 – 0,53]) and working memory (Hedge’s g 0,39 ([95% CI 0,05 – 0,73]). We have moderate confidence in these results. Exercise may lead to a moderate improvement of attention (Hedge’s g 0,66 [95% CI 0,2 – 1,12]) and social cognition (Hedge’s g 0,71 [95% CI 0,27 – 1,15]), but may lead to little or no difference on processing speed. We are uncertain of the effect of exercise on verbal and visual learning and memory, and reasoning and
problem solving because of very low confidence in the results for these outcomes.
None of the systematic reviews we included had results on longterm effects of exercise or direct comparisons of the effect of different forms of exercise.
Discussion
One of the included systematic reviews summarized the effect of physical exercise on physical and mental health outcomes for people with severe mental disorders. The search for primary studies was conducted in 2013. In our search for literature, we found similar reviews with more recent search dates, but these were excluded because they did not meet the requirements for high methodical quality. We have therefore limited knowledge about primary studies published after 2013, and whether recent studies may contribute more solid evidence about the relationship between exercise and mental and physical health outcomes.
Overall, we downgraded our confidence in the results because of high risk of bias in most of the included primary studies in the reviews. We also downgraded our confidence for most of the outcomes because the studies included for the oucomes were few and had few participants.
We have moderate confidence in the results for two of 12 outcomes (global cognition and working memory), implying that the effect estimate is likely to be close to the true effect. We have low confidence in the results for five outcomes (body mass index, weight, attention, social cognition and processing speed). This means that it is a chance that the presented effect estimates can be considerable different from the true effect. For the last five outcomes (physical fitness, symptoms of anxiety and depression, verbal learning and memory, visual learning and memory, and reasoning and problem solving), we have very little confidence in the results implying that it is difficult to draw conclusion about the size of the true effect.
With few exceptions, the knowledge base on the effect of physical training for people with severe mental disorders is very limited. It is important to point out that the fact that we have limited confidence in the results does not mean that physical exercise is uneffective.
Conclusion
For people with severe mental illness, such as schizophrenia, bipolar disorder and psychosis, exercise may lead to little or no difference in body mass index or weight, and we are uncertain about the effect on physical fitness and symptoms of anxiety and depression.
For people with schizophrenia, exercise probably leads to a small improvement in global cognition and working memory and a moderate improvement of attention and social cognition. We are uncertain of the effect of exercise on verbal and visual learning and reasoning.
There is a need for more primary studies of high quality/low risk of bias on the effect of exercise for this population. Studies should also aim to include outcomes for long-term follow up.
