Physical Activity Levels and Psychosis: A Mediation Analysis of Factors Influencing Physical Activity Target Achievement Among 204 186 People Across 46 Low- and Middle-Income Countries

Abstract

Physical activity (PA) can help reduce cardiovascular disease and premature mortality in people with psychosis. However, there is a paucity of representative data on PA in people with psychosis, especially from low- and middle-income countries (LMICs). Moreover, data on subclinical psychosis and PA is absent. This study explored whether complying with PA recommendations of 150 minutes of moderate-vigorous PA per week is related to: (1) psychotic symptoms without a psychosis diagnosis (subclinical psychosis); and (2) clinical psychosis (psychosis diagnosis). A total of 204 186 participants aged 18–64 years from 46 LMICs recruited via the World Health Survey were subdivided into those with (1) no psychosis diagnosis and no psychotic symptoms in the past 12 months (controls); (2) subclinical psychosis; and (3) psychosis diagnosis. People with a psychosis diagnosis had significantly higher odds for low PA in the overall sample (OR = 1.36; 95% CI = 1.04–1.78; P = .024) and among males (OR = 2.29; 95% CI = 1.57–3.34; P < .0001) but not females (OR = 0.93; 95% CI = 0.67–1.30; P = .6712). No difference was found among those with subclinical psychosis vs controls. Mediation analyses demonstrated that mobility difficulties explained the largest amount of low PA among males (18.5%) followed by self-care difficulties (16.3%), depression (16.1%), cognition (11.8%), pain and discomfort (11.4%), interpersonal activities (8.6%), sleep and energy (7.2%), and vision (3.0%). The results from the largest dataset on PA and psychosis and first in LMICs, found that psychosis diagnosis (especially among males) but not subclinical psychosis, is associated with physical inactivity. Population level interventions seeking to increase PA among people with psychosis may help improve health outcomes.

Introduction

The burgeoning evidence base has established that people with a confirmed psychotic illness such as schizophrenia are at increased risk of cardio-metabolic disease.^1,2 Cardiovascular and metabolic disease are the leading causes of the 15-year mortality gap reported vs the general population.^3–5 While physical activity (PA) is effective for the prevention and management of diabetes and cardiovascular disease in the general population,⁶ people with psychotic disorders are among the most inactive clinical populations, spending on average almost 13 hours a day engaging in sedentary behavior.⁷ Moreover, a recent meta-analysis established that people with first-episode and established psychosis engage in very low levels of moderate to vigorous intensity PA.⁸ However, PA interventions can improve the physical health and mental health of people with psychosis^9,10 and this information is especially important for low-income settings as they struggle to plan and provide mental health services.

Understanding what factors influence PA in those with psychosis is of high importance. To date, a number of correlates have been identified including negative symptoms, depression, social isolation, and increasing age.¹¹ However, very few representative cohort studies, and always limited to a single country,¹² have investigated PA correlates.

Some other pertinent gaps exist within the literature. First, there is increasing recognition that psychosis sits on a spectrum, ranging from psychotic-like experiences (PLE) or subclinical psychosis, through to an established diagnosis of psychosis.¹³ Moreover, literature suggests that people with PLE/psychotic symptoms (particularly those from the developing world) have poorer physical health (eg, increased diabetes and cardiovascular disease risk^14,15) compared to the general population. However, to our knowledge, no study has investigated PA levels in a representative sample of people with subclinical psychosis. Second, there is a paucity of representative data investigating PA levels in people with psychotic disorders, and an absence of data from low- and middle-income countries (LMICs). This is an important research gap given that PLE are common in LMICs^16,17 while physical health comorbidities such as diabetes, which may benefit from PA, are also highly prevalent.¹⁴ Furthermore, associations may differ in LMICs given the higher likelihood of suboptimal treatment for psychotic illnesses¹⁸ and physical health conditions, among other contextual differences. Therefore, there is an imperative to achieve a global understanding of PA levels to inform the design and delivery of interventions and identify priority areas. Finally, while PA correlates have been identified in a paucity of nationally representative samples in people with established psychosis,¹² mediation analysis of potentially important mediators (eg, mobility difficulties, pain, depression), will provide a greater understanding of the impact of such common comorbidities in those with psychosis.

Given these aforementioned gaps within the literature, the aims of the current study were to: (1) investigate the proportion of people with psychotic symptoms without a formal diagnosis (subclinical psychosis) or an established psychosis who met the WHO recommended 150 minutes of moderate to vigorous intensity of PA per week¹⁹ from 46 LMICs; and (2) to investigate the factors that may mediate the relation between psychosis and physical (in)activity.

Methods

The Dataset

The current study uses data from the World Health Survey (WHS), which is a cross-sectional survey carried out in 70 countries between 2002 and 2004. A single stage random sampling approach was employed in 10 countries, while a stratified multi-stage random cluster sampling method was used in 60 countries. Full details of the survey are publically available (http://www.who.int/healthinfo/survey/en/). Briefly, individuals with a registered home address aged ≥18 years were eligible to participate. Kish tables were used to ensure that all household members had an equal chance of being selected. A standardized questionnaire to collect data for the WHS was developed and a consistent translation procedure was employed to ensure cross-country comparability. Face to face, and telephone interviews were conducted to acquire data by trained interviewers. The response rate was high across all countries (98.5%).¹⁵ To adjust for non-response, sampling weights were generated using the population distribution as reported by the United Nations Statistical Division. Local ethical approval and governance was approved at each country site and informed consent was obtained from all participants.

Data were available for 69 countries. These data were nationally representative for all countries with the exception of China, Comoros, the Republic of Congo, Ivory Coast, India, and Russia. Countries with no sampling information (10 countries—Austria, Belgium, Denmark, Germany, Greece, Guatemala, Italy, Netherlands, Slovenia, United Kingdom) were excluded. Of the remaining 59, 11 countries (Finland, France, Ireland, Israel, Latvia, Luxembourg, Morocco, Norway, Portugal, Sweden, Turkey) were excluded as information on PA and/or psychosis was not collected. Furthermore, we excluded the only 2 remaining high-income countries as the focus of this study was on LMICs. Based on the World Bank classification in 2003, the remaining 46 countries consisted of 25 middle-, and 21 low-income countries.

Variables

PA (Outcome Variable).

In order to classify if participants completed the recommended PA levels of 150 minutes of moderate to vigorous PA per week,¹⁹ we used the vital sign method which has been used extensively in the general population²⁰and more recently in schizophrenia.²¹ Items from the International Physical Activity Questionnaire (IPAQ)²² were used to categorize PA. Specifically, participants were asked how many days over the past week, on average they engaged in moderate and vigorous PA. Secondly, participants were asked for how many minutes on average, they engage in PA at a moderate and vigorous level. The total amount of moderate to vigorous PA over the last week was calculated and those scoring ≥150 minutes were classified as meeting the recommended guidelines (coded 0), and those scoring <150 minutes (low PA) were classified as not meeting the recommended guidelines (coded 1).

Psychosis Diagnosis and Psychotic Symptoms (Exposure Variable).

Participants were asked whether they had ever been diagnosed of having schizophrenia or psychosis. All participants, regardless of a psychosis diagnosis, were asked questions on positive psychotic symptoms which came from the CIDI 3.0.²³ This psychosis module has been reported to accord highly with clinician ratings.²⁴ The hallucinations question excluded conditions associated with sleep-related states or substance use. Specifically, respondents were asked the following questions with answer options “yes” or “no”:

During the last 12 months, have you experienced:

1. “A feeling something strange and unexplainable was going on that other people would find hard to believe?” (delusional mood)

2. “A feeling that people were too interested in you or there was a plot to harm you?” (delusions of reference and persecution)

3. “A feeling that your thoughts were being directly interfered or controlled by another person, or your mind was being taken over by strange forces?” (delusions of control)

4. “An experience of seeing visions or hearing voices that others could not see or hear when you were not half asleep, dreaming or under the influence of alcohol or drugs?” (hallucinations)

Individuals who endorsed at least 1 of the 4 above-mentioned psychotic symptoms were considered to have psychotic symptoms. Based on information on psychosis diagnosis and psychotic symptoms, a 3-category psychosis variable was constructed: (1) no psychosis diagnosis and no psychotic symptoms; (2) at least 1 psychotic symptom but no psychosis diagnosis; and (3) psychosis diagnosis.

Health Status and Depression.

Participants’ health status was evaluated with 14 health-related questions pertaining to 7 health domains including (1) mobility; (2) self-care; (3) pain and discomfort; (4) cognition; (5) interpersonal activities; (6) vision; (7) sleep and energy. Each of the 7 domains corresponds to those in common health related quality of life outcome measures such as the Short Form-12 (SF-12),²⁵ the Health Utilities Index Mark 3 (HUI),²⁶ and the EUROQOL 5D.²⁷ Moreover, the domains have been used as indicators of functional health status in prior studies utilizing the WHS dataset.^16,28,29 Each domain consists of 2 questions that assessed health function in the past 30 days. The actual questions can be found in supplementary table 1. Each item was scored on a 5-point scale ranging from “none” to “extreme/cannot do.” For each separate domain, we used factor analysis with polychoric correlations to obtain a factor score which was later converted to scores ranging from 0 to 100^16,29 with higher values representing worse health function.

In order to determine the presence of depression, the DSM-IV algorithm was used and was based on the duration and persistence of depressive symptoms in the previous 12 months.³⁰

Other Variables.

Information was also examined on age (18–29, 30–39, 40–49, 50–59, ≥60 y), sex, wealth, education, and setting (urban or rural). Principal component analysis based on 15 to 20 assets was conducted to establish country-wise wealth quintiles. Education was categorized as: no formal education, primary education, secondary or high school completed, or tertiary (university) education completed.

Statistical Analysis.

The statistical analysis was done with Stata 14.1 (Stata Corp LP). We restricted all analyses to those aged 18 to 64 years, since the PA recommendations are based on adults within this age range.¹⁹ The difference in sample characteristics by psychosis category or PA level was examined with Chi-squared tests.

Multivariable logistic regression analysis was conducted to assess the association between psychosis and low PA. The 3-category psychosis variable was the exposure variable and low PA the outcome variable. Sex-stratified analyses were also conducted. The models were adjusted for age, education, wealth, setting, and country. In addition, the model including the overall sample was also adjusted for sex. In line with previous publications using the WHS, we adjusted for country by including dummy variables for each country.³¹

Next, we conducted mediation analysis to understand the extent to which various factors (depression, mobility, self-care, pain and discomfort, cognition, interpersonal activities, vision, sleep and energy) may explain the association between psychosis and low PA. The mediation analysis was done using the Karlson Holm Breen-command in Stata.³² Each potential mediator was included in the model separately. The selection of the potential mediators was based on past literature.^11,12,33 This analysis used the psychosis diagnosis (yes/no) variable as the exposure variable and was only conducted among males based on the results of the abovementioned multivariable logistic regression analysis.

The sample weighting and the complex study design were taken into account in all analyses. Results from the logistic regression models are presented as ORs with 95% CIs. The level of statistical significance was set at P < .05.

Results

The final sample size was 204 186. The mean (SD) age was 35.2 (12.5) years and 50% were men. Psychotic symptoms without a psychosis diagnosis were present in 13.8% (unweighted n = 22 987, mean age 35.4 y, 47.2% male), while the prevalence of psychosis diagnosis was 1.1% (unweighted n = 2107, mean age 37.8 y, 44.1% male; see table 1).

Table 1.

Characteristics of the Study Sample (Overall and by Psychosis Category)

				Controls		Subclinical Psychosis		Psychosis Diagnosis
				Symptom (−) Diagnosis (−)		Symptom (+) Diagnosis (−)		Diagnosis (+)
Characteristic	Category	Unweighted N	Overall (%)	Unweighted N	%	Unweighted N	%	Unweighted N	%	P-valuea
Age (y)	18–29	62 423	40.3	54 320	40.6	7947	40.3	608	28.1	<.0001
	30–39	50 643	23.7	43 621	23.6	6244	23.6	579	28.6
	40–49	37 782	18.7	32 647	18.6	4536	17.3	457	22.1
	50–59	24 364	12.7	21 098	12.5	3050	13.6	336	14.9
	≥60	9600	4.7	8332	4.7	1210	5.1	127	6.4
Sex	Male	82 572	50.0	72 220	50.2	9673	47.2	881	44.1	.0003
	Female	102 205	50.0	87 740	49.8	13 306	52.8	1226	55.9
Education	No formal	34 327	23.7	31 532	24.6	6832	25.1	708	37.1	<.0001
	≤Primary	60 549	31.6	54 293	30.8	8325	36.5	781	35.0
	Secondary completed	74 839	35.2	59 502	34.4	6541	31.0	506	21.8
	Tertiary completed	14 946	9.6	14 517	10.2	1265	7.4	107	6.1
Wealth	Poorest	39 455	19.5	34 431	19.4	5099	20.6	536	26.1	<.0001
	Poorer	36 733	19.8	31 284	19.5	4546	21.2	408	22.7
	Middle	34 628	20.0	29 234	19.9	4127	19.8	374	20.0
	Richer	33 277	20.5	27 986	20.4	3883	19.9	313	17.1
	Richest	31 356	20.2	26 738	20.8	3457	18.6	277	14.1
Setting	Rural	93 084	56.4	84 056	57.8	12 934	56.0	1243	64.0	.0183
	Urban	91 069	43.6	75 195	42.2	9915	44.0	851	36.0

Note: Percentages are based on weighted sample. Symptoms refer to having at least one of: delusional mood, delusions of reference and persecution, delusions of control, and hallucinations in the past 12 mo. Diagnosis refers to self-reported lifetime diagnosis of schizophrenia/psychosis.

^a P-values are based on Chi-squared tests.

Across the entire population, the prevalence of people achieving less than 150 minutes of moderate-vigorous PA was 26.9%. Overall, the prevalence of low PA was 24.3% in people with psychotic symptoms and no diagnosis and 33.0% in people with a diagnosis of psychosis. The prevalence of low PA in people without any psychosis was 27.0% (figure 1). Compared to those without a psychosis diagnosis, patients with a diagnosis were more likely to be physically inactive in the overall sample and in males while this was not observed in females. Across the entire dataset, older age, female sex, higher levels of education and wealth, and urban setting were significantly associated with physical inactivity (P < .0001) (table 2).

Fig. 1.

Prevalence of low physical activity by psychosis category (overall and by sex). Bars denote 95% CIs. Estimates are based on weighted sample. Symptoms refer to having at least one of: delusional mood, delusions of reference and persecution, delusions of control, and hallucinations in the past 12 mo. Diagnosis refers to self-reported lifetime diagnosis of schizophrenia/psychosis. The total amount of moderate to vigorous physical activity over the last week was calculated and those scoring <150min were considered to have low physical activity.

Table 2.

Characteristics of the Study Sample by Low Physical Activity

		Low Physical Activity
		No		Yes
Characteristic	Category	Unweighted N	%	Unweighted N	%
Age (y)	18–29	45 715	41.3	16 708	37.7
	30–39	37 385	24.5	13 258	21.6
	40–49	27 509	18.8	10 273	18.3
	50–59	16 662	11.7	7702	15.2
	≥60	5785	3.7	3815	7.2
Sex	Male	61 560	52.5	21 012	43.3
	Female	71 477	47.5	30 728	56.7
Education	No formal	24 957	23.1	9370	25.3
	≤Primary	45 073	32.8	15 476	28.2
	Secondary completed	53 234	35.1	21 605	35.4
	Tertiary completed	9696	9.0	5250	11.1
Wealth	Poorest	29 595	20.4	9860	17.1
	Poorer	27 147	20.3	9586	18.3
	Middle	25 000	20.5	9628	19.0
	Richer	23 723	20.4	9554	20.8
	Richest	20 932	18.5	10 424	24.8
Setting	Rural	72 600	60.5	20 484	45.1
	Urban	60 000	39.5	31 069	54.9

Note: Percentages are based on weighted sample. The difference between low physical activity (no vs yes) was statistically significant for all characteristics (P < .0001) (Chi-squared test). The total amount of moderate to vigorous physical activity over the last week was calculated and those scoring <150min were considered to have low physical activity.

The association between psychosis and low PA estimated by multivariable logistic regression is shown in table 3. Psychotic symptoms without a diagnosis were not associated with low PA neither in the overall sample or the sex-wise samples. However, adults with a psychosis diagnosis had significantly higher odds for low PA in the overall sample (OR = 1.36; 95% CI = 1.04–1.78; P = .024) and among males (OR = 2.29; 95% CI = 1.57–3.34; P < .0001) while this was not the case for females (OR = 0.93; 95% CI = 0.67–1.30; P = .6712).

Table 3.

The Association Between Psychosis and Low Physical Activity Estimated by Multivariable Logistic Regression (Overall and by Sex)

	Overall		Male		Female
Characteristic	OR	95% CI	OR	95% CI	OR	95% CI
Psychosisa
Symptoms (−) Diagnosis (−)	1.00		1.00		1.00
Symptoms (+) Diagnosis (−)	0.97	[0.88, 1.07]	0.96	[0.83, 1.11]	1.01	[0.90, 1.13]
Diagnosis (+)	1.36*	[1.04, 1.78]	2.29**	[1.57, 3.34]	0.93	[0.67, 1.30]
Sex
Male	1.00
Female	1.51**	[1.42, 1.61]
Age (y)
18–29	1.00		1.00		1.00
30–39	0.94	[0.87, 1.01]	1.11	[1.00, 1.24]	0.81**	[0.74, 0.89]
40–49	1.10*	[1.02, 1.18]	1.26**	[1.14, 1.40]	0.99	[0.89, 1.09]
50–59	1.48**	[1.37, 1.60]	1.67**	[1.48, 1.89]	1.36**	[1.22, 1.52]
≥60	2.39**	[2.11, 2.70]	2.70**	[2.26, 3.22]	2.24**	[1.89, 2.66]
Education
No formal	1.00		1.00		1.00
≤Primary	0.97	[0.90, 1.06]	1.17*	[1.03, 1.33]	0.94	[0.84, 1.05]
Secondary completed	1.10	[0.99, 1.21]	1.31**	[1.14, 1.52]	1.07	[0.93, 1.24]
Tertiary completed	1.28**	[1.11, 1.48]	1.63**	[1.33, 2.00]	1.17	[0.96, 1.42]
Wealth
Poorest	1.00		1.00		1.00
Poorer	1.05	[0.97, 1.13]	1.13*	[1.00, 1.27]	0.98	[0.88, 1.09]
Middle	1.04	[0.94, 1.15]	1.10	[0.94, 1.27]	0.99	[0.88, 1.12]
Richer	1.11*	[1.01, 1.22]	1.29**	[1.13, 1.47]	0.95	[0.83, 1.08]
Richest	1.33**	[1.20, 1.48]	1.47**	[1.27,1.71]	1.21*	[1.04, 1.40]
Urban setting	1.61**	[1.46, 1.77]	1.65**	[1.48, 1.84]	1.59**	[1.39, 1.82]

Note: Models are adjusted for all covariates in the respective columns and country. The total amount of moderate to vigorous physical activity over the last week was calculated and those scoring <150min were considered to have low physical activity.

^aSymptoms refer to having at least one of: delusional mood, delusions of reference and persecution, delusions of control, and hallucinations in the past 12 mo. Diagnosis refers to self-reported lifetime diagnosis of schizophrenia/psychosis.

*P < .05; **P < .001.

In order to understand the degree to which the association between a psychosis diagnosis (Y/N) and low PA can be explained by other factors, we conducted mediation analyses only among males while adjusting for age, education, wealth, setting, and country. For this analysis, a dichotomized psychosis variable was used where the “no psychosis diagnosis” category also included those with psychotic symptoms as the magnitude of the association with PA for this group was similar regardless of the presence of psychotic symptoms. While the total effect was also significant in the overall sample, we did not conduct this analysis using the overall sample as it was likely that the significant results are mainly driven by the strong association observed in males. The direct and indirect effects were both significant in all models and are summarized in table 4. While a significant mediational effect was observed for all the potential mediators tested, mobility explained the largest percentage of the association (18.5%) followed by self-care (16.3%), depression (16.1%), cognition (11.8%), pain and discomfort (11.4%), interpersonal activities (8.6%), sleep and energy (7.2%), and vision (3.0%).

Table 4.

Health Status Outcomes as Mediators in the Association Between Psychosis Diagnosis and Low Physical Activity Among Males

Mediator	Effect	OR (95% CI)	P-value	% Mediated
Depression	Total	2.27 (1.54–3.36)	<.0001	16.1
	Direct	1.99 (1.33–2.97)	.0007
	Indirect	1.14 (1.07–1.22)	.0002
Mobility	Total	2.33 (1.60–3.37)	<.0001	18.5
	Direct	1.99 (1.37–2.88)	.0003
	Indirect	1.17 (1.12–1.23)	<.0001
Self-care	Total	2.31 (1.63–3.29)	<.0001	16.3
	Direct	2.02 (1.42–2.86)	.0001
	Indirect	1.15 (1.08–1.21)	<.0001
Pain and discomfort	Total	2.44 (1.69–3.52)	<.0001	11.4
	Direct	2.20 (1.53–3.18)	<.0001
	Indirect	1.11 (1.06–1.15)	<.0001
Cognition	Total	2.32 (1.61–3.35)	<.0001	11.8
	Direct	2.10 (1.47–3.02)	.0001
	Indirect	1.10 (1.06–1.15)	<.0001
Interpersonal activities	Total	2.31 (1.60–3.33)	<.0001	8.6
	Direct	2.15 (1.49–3.09)	<.0001
	Indirect	1.07 (1.03–1.12)	.0003
Vision	Total	2.33 (1.60–3.39)	<.0001	3.0
	Direct	2.27 (1.56–3.30)	<.0001
	Indirect	1.03 (1.01–1.05)	.0113
Sleep and energy	Total	2.30 (1.59–3.33)	<.0001	7.2
	Direct	2.17 (1.49–3.14)	<.0001
	Indirect	1.06 (1.03–1.10)	.0005

Note: Models are adjusted for age, education, wealth, urban location, and country. Psychosis diagnosis refers to self-reported lifetime diagnosis of schizophrenia/psychosis. The total amount of moderate to vigorous physical activity over the last week was calculated and those scoring <150min were considered to have low physical activity.

Discussion

To the best of our knowledge the current data are the first to suggest on a multi-national level across LMIC that a diagnosis of psychosis is associated with physical inactivity, especially among males. Second, our data are the first to demonstrate that psychotic symptoms without a diagnosis of psychosis do not appear to be associated with physical inactivity. We hypothesized that PLE might be associated with low PA based on the finding that this group has elevated levels of diabetes and cardiovascular disease. The finding that a psychosis diagnosis but not PLE was associated with low PA, possibly indicates that this difference may partially be explained by the use of antipsychotics as those with PLE are unlikely to be taking antipsychotic medication. Third, our study investigated mediators of inactivity in people with diagnosed psychosis for the first time, suggesting that mobility limitations, difficulties with self-care, depression, reduced cognition and pain are key explanatory factors that influence PA participation in male patients. Given the fact that physical inactivity increases the risk of weight-gain and chronic physical diseases³⁴ and that PA may have a protective role in preventing adverse outcomes in people with psychosis,^9,10 there is a need to develop appropriate future interventions to increase PA in this population.

A previous systematic review of PA correlates found inconsistent evidence of a relationship between sex and PA in people with psychosis.¹¹ A recent nationally representative study also found no difference for sex and PA among people with psychosis.¹² Our data is the first adequately powered data set to demonstrate the importance of sex on PA levels in psychosis. The exact reasons why inactivity affects males more than females with psychosis is unclear. However, the higher levels of inactivity among males could be due to the typical earlier onset of illness in males, higher burden of negative symptoms, substance use disorders and worse remission and recovery rates.³⁵ These higher levels of inactivity may also partially explain the increased risk of cardiovascular disease and premature mortality observed particularly among males with psychosis.^36,37

Importantly, our data include the first mediation analyses offering an understanding of the pathway between psychosis and physical inactivity among males who were at greatest risk. Among the most significant mediators were mobility difficulties (accounting for 18.5%) and pain (11.4%), both of which are closely related to each other and have been identified as being related to reduced PA in the general population.^38,39 While previous literature has identified that pain is highly prevalent among people with psychosis⁴⁰ and those with psychotic experiences,⁴¹ the potential deleterious influence of pain beyond its negative impact on quality of life⁴² is largely unknown. Pain is known to decrease an individual’s confidence in mobilizing and is associated with falls and could therefore result in less PA in people with psychosis. Moreover, a primary symptom of diabetes is painful neuropathy, which could also account for lower PA in people with psychosis and metabolic abnormalities.⁴³ Physiotherapists are experts in assessing and developing PA interventions and may have a unique role in developing adaptive PA interventions in people with psychosis and may help to overcome many of these potential mediators.⁴⁴ Despite their vital role as indicated above, most of the mental health units in low income countries lack physiotherapists or other qualified exercise professionals and, neither are their activities performed by other cadres of staff (Vancampfort, D, Stubbs B, Probst M, Mugisha J. Physiotherapy for people with mental health problems in Sub-Saharan African countries: a systematic review. Submitted). Even in the few mental health units where such specialists exist, they are normally overwhelmed by patient load which puts doubt on the quality of services they provide.

Among the other mediators in males with psychosis, depression and cognitive impairment also appear to influence PA participation, which is line with previous correlates reported in smaller study samples.¹¹ While depression and cognition appear to mediate inactivity, both are known to be exacerbated by sedentary behavior in the general population.^45,46 However, PA is known to be an effective intervention to improve depression symptomology in those with major depression,⁴⁷ post-traumatic stress disorder⁴⁸ and psychosis.⁹ Therefore, this demonstrates the need for supportive interventions to help overcome such pertinent barriers to PA.

The development of effective PA interventions in LMICs is important for the prevention and management of all mental disorders due to the considerable burden in LMICs.⁴⁹ Despite the high burden due to mental ill health, only a very small proportion of the health budget in LMICs is allocated to the treatment and prevention of these disorders with an average of 2% in lower-middle income countries, and 0.5% in low-income countries.⁵⁰ As a consequence, mental health services are poorly resourced with treatment rates for people with mental disorders remaining low and reported treatment gaps over 90%.⁵¹ Currently, community-based rehabilitation, psychoeducation and support for families (delivered by nonspecialists) are recommended for low resource settings, with assertive community care and cognitive therapy recommended as additions in higher resourced settings with stronger service-deliver platforms.⁵² The potential role of low-cost PA interventions in the prevention and treatment of mental disorders seems to be neglected but need to be explored in mental health systems studies. Given the large numbers of people affected by psychosis, and the increased associated morbidity, investment in trained public health professionals and resources (including infrastructure) to increase PA in daily life for people with psychosis may offer excellent value, both economically and from a mental and physical health perspective.

Several strategies to initiate and facilitate an increase in PA levels through the health care systems of LMICs are possible. For example, continued medical education (CME) should be used to equip staff on the importance of PA for people with mental illness. These CME are cheaply delivered since they are part of other routine trainings in the health facilities. Health care professionals in primary care should be informed about the importance of PA for people with psychosis and should be trained in skills to deliver easy implementable interventions. Asking about and advising regarding PA among people with psychosis might help to improve PA uptake or maintenance. One approach that should be tested more in detail is the one used in this survey, ie, the Physical Activity Vital Sign (PAVS) method. The brevity of the PAVS may help promote the importance of PA and exercise assessment and prescription in lower resourced settings. A recent study²¹ in people with established schizophrenia in Belgium clearly demonstrated that those who did not adhere to the minimum PA recommendations, as formulated by this method, have higher metabolic risks.

Limitations and Strengths

The current findings should be interpreted in light of some limitations. First, the study is cross-sectional. In addition, the range of psychotic experiences was limited, and not assessed by a clinical interview. Longitudinal studies with more experienced interviewers are needed in order to analyze the natural history of these symptoms in the general population.⁵³ In addition, we did not capture data on psychotropic medication which is known to influence activity levels.⁸ PA was captured with a self-report measure and their accuracy has been questioned.^8,54 Nonetheless, the strengths of the study include the large sample size (over 200 000) and the multi-national scope, including most regions of the world, but in particular LMICs in Africa, Latin-America, Asia and Eastern Europe. Most of the research in the domain of psychosis and physical activity has been conducted in Western countries. The present study was furthermore performed with nationally representative samples of non-institutionalized persons. In addition, we present the first representative data on PA in those who have experienced psychotic episodes without a formal diagnosis. Moreover, by conducting mediation analyses, we have advanced the understanding of factors influencing PA participation, which to date, has relied almost entirely on correlational analysis from biased studies in single countries.

In conclusion, people with a psychosis diagnoses (particularly males) but not subclinical psychosis engage in lower levels of PA vs controls. Several pertinent mediators including mobility limitations, pain, depression and cognition appear to influence the psychosis and inactivity relationship. Patients with psychotic disorders in LMICs should be screened for their PA behavior and public health care providers should be trained in the identification of patients not complying with the general PA recommendations.

Supplementary Material

Supplementary material is available at http://schizophreniabulletin.oxfordjournals.org.

Funding

B.S. and F.G. receive funding from the National Institute for Health Research Collaboration for Leadership in Applied Health Research & Care Funding scheme. J.F. is funded by an MRC Doctoral Training Grant. S.R. is funded by a Society for Mental Health Research early career fellowship (Australia). A.K.’s work was supported by the Miguel Servet contract financed by the CP13/00150 and PI15/00862 projects, integrated into the National R + D + I and funded by the ISCIII—General Branch Evaluation and Promotion of Health Research—and the European Regional Development Fund (ERDF-FEDER). D.V. is funded by the Research Foundation—Flanders (FWO-Vlaanderen).