People with severe mental illness (SMI) have poor physical health and experience premature mortality of approximately 15–20 years compared to the general population (1); up to 85% of this prematurely mortality has been attributed to physical illnesses such as cardiometabolic disease (CMD), and cancer (2). Modifiable risk factors such as physical inactivity, poor nutrition, and low cardiorespiratory fitness play a significant contribution to this increased risk (3, 4). While the poor physical health of people with SMI has gained increasing international recognition (5), there is evidence that physical health disparities persist and the mortality gap for people with SMI may actually be widening (1). Addressing modifiable risk factors and improving the physical health of people with SMI should be a primary concern in a mental health setting (5).

Several meta-analyses of lifestyle (diet and exercise) interventions for people with SMI have demonstrated improvements in metabolic risk factors, such as waist circumference, body mass index (BMI), fasting blood glucose, triglycerides, and blood pressure (6–8). A recent meta-review found lifestyle interventions were comparable to pharmacotherapy for cardiometabolic risk reduction (9). However, the majority of lifestyle interventions have been conducted in outpatient community settings, or with early psychosis populations (10).

One setting in which few lifestyle interventions have been conducted are residential rehabilitation and supported housing units for people with SMI. People accessing these services typically experience significant burden from negative and cognitive symptoms, and have lower functioning than people living in community settings (11, 12). In Australia, Community Care Units (CCU’s) are a type of residential rehabilitation which provide independent housing in the community, with access to 24 h on-site multidisciplinary mental health staff, but do not provide food or meals.

Previous lifestyle studies conducted within supported housing, residential rehabilitation, or long stay inpatient units have broadly employed two different methods; either via evaluating the effect of the addition of combined structured nutrition and exercise interventions provided directly to residents with SMI (13, 14), or by more broadly focusing on attempts to utilize existing resources to improve the environment and culture of facilities, via policy change and by upskilling mental health staff to promote healthy behaviors for residents of the facilities (15–17). Both types of studies have found various improvements in components of the metabolic syndrome [weight, BMI, waist circumference, blood pressure, high density lipoprotein (HDL) cholesterol], (13–17) and objective physical activity (15). In an Australian context, to date, no combined diet and exercise interventions have been formally evaluated within a residential rehabilitation setting.

Our team conducted a small (n = 10) exercise physiology student-led pilot supervised exercise intervention (three sessions per week of supervised exercise) in 2016 in a single residential rehabilitation unit, Coorparoo CCU (18). We found the intervention to be feasible and improved functional exercise capacity and total exercise volume/week. However, no changes in metabolic risk factors were observed. We postulated that this may be due to the poor diet quality of the participants (19), and the addition of a dietary intervention may be beneficial.

In this follow-up study, we expanded on the pilot study by including (i) the dietary intervention (six individual and group sessions), and (ii) two additional CCU sites. Research from large scale clinical trials are often significantly resourced compared to usual clinical settings, with interventions typically led by external research teams, which limits generalizability to usual care settings (20, 21). There are increasing calls for the research agenda of lifestyle interventions to include studies that report outcomes from pragmatic and real-world settings (22, 23). Hence, in the current study, we implemented and evaluated the feasibility of a 10-week, exercise physiology and dietetics student-led, pragmatic lifestyle intervention integrated within the usual mental health care setting of three CCU’s.

Based on the results of our pilot exercise study, we hypothesize that the combined dietary and exercise lifestyle intervention will be feasible and acceptable increase fitness (measured via functional exercise capacity) and the amount of exercise of participants. Further, we hypothesize the addition of a dietary component to the intervention will decrease metabolic risk factors of participants with SMI located within CCU’s.

Implementation of a brief student-led lifestyle (combined dietary and exercise) intervention integrated into usual mental health care within three residential rehabilitation settings was feasible, safe, and acceptable with no significant adverse events resulting in drop out. We found increases in functional exercise capacity, total exercise volume, and a decrease in negative and total psychiatric symptoms, confirming and extending our 2018 pilot exercise findings across two additional rehabilitation settings. However, addition of the dietary intervention did not improve markers of metabolic health or anthropometric measures.

The recruitment rate of the combined diet and exercise lifestyle intervention (68.0%) was similar to previously reported lifestyle interventions conducted within both outpatient (30, 43) and residential rehabilitation settings (13, 14), (63–70%), indicating the large majority of people accessing rehabilitation facilities may be interested in participating in lifestyle interventions. This is encouraging given that rehabilitation facilities typically have a holistic focus and are often resourced with shared cooking facilities making these settings opportune for implementation of lifestyle interventions (43).

Our retention rates (77%) were high and comparable with other lifestyle intervention studies in SMI populations (73–84%) (29, 30, 43, 44). Features of our program that may have promoted participant retention included support by mental health staff to participate, provision of the intervention on site, and integration into usual mental health care (26).

Participation rate for those completing the exercise intervention was 70%, and was broadly comparable to the 72% reported in a meta-analysis of exercise interventions for people with SMI (45). Of the 32 completers, one third (9) of participants attended 100% of the sessions and were observed to make friendships with other participants during partnering and teamwork involved with circuit sessions, which likely facilitated autonomous motivation to exercise (46). As was highlighted in previous pilot work, (25) an advantage of the student placement design was that EP students had the flexibility to offer both group and individual make-up sessions which increased participation rates. Broadening the exercise program to include less structured exercise formats and greater variety (45), i.e. more sports and use of green spaces (47), in keeping with participants’ preferences, may increase participation further.

For those who completed exercise questionnaires, there was overall acceptance and enjoyment of the exercise program. Confidence to exercise alone and when having a bad day was particularly encouraging as exercise could be included as a coping strategy for symptom exacerbations in the setting of treatment resistance, commonly encountered by residents in residential facilities (12). We found resistance training elements were popular; future exercise interventions could focus on expanding the provision of resistance training, particularly given emerging evidence for this type of training for depressive symptoms, quality of life, and symptoms of psychosis (48–50).

While the brief intervention was acceptable to participants, the most common reason for withdrawal was a lack of motivation to exercise. Additionally, for those completing the intervention we found an increase in amotivation and decrease in the value of exercise on the BREQ-2 questionnaire over ten weeks. Together, these findings confirm the challenges people with SMI can face regarding motivation to adapt and sustain a healthy lifestyle (27), which raises two points. Firstly, while participation was high over ten weeks, amotivation was already increasing which may impact longer term engagement. The vast majority of lifestyle intervention evidence has been reported from studies under 6 months (10), hence longer pragmatic interventions need to be evaluated, as these findings may have a significant impact on lasting health benefits for people with SMI. Secondly, these findings highlight the importance of addressing autonomous motivation when delivering lifestyle interventions (51). While we employed motivational strategies, students leading the lifestyle intervention were not fully qualified; and interventions delivered by qualified practitioners may improve adherence and motivation (52). In future, studies could also investigate the use of objective physical activity measurement devices (i.e. pedometers or accelerometers) and mobile phone apps to target autonomous motivation to address lifestyle behaviors (53–55). Additionally, in future, utilizing principles of co-design, by consulting service users regarding optimal development and implementation of exercise interventions, may increase autonomous motivation to engage (23, 56, 57).

There was broad level of acceptance of the dietetic program by those completing acceptability questionnaires: the majority felt that they had improved their eating habits and would continue to use healthy cooking skills independently, however this was not assessed objectively.

While no adverse events resulted in participant withdrawal, several participants experienced exacerbations of pre-existing injuries that required program modifications. This highlights the importance of exercise professionals prescribing and delivering exercise programs for people with SMI, who are known to experience higher levels of pain and physical comorbidity than the general population (26, 58, 59). Nonetheless, all participants with pre-existing injuries were able to complete a modified program, in keeping with several recent studies suggesting that exercise is safe in people with SMI (22, 51).

Consistent with our hypothesis, and pilot study findings, we found a significant increase in functional exercise capacity (26) of 57 m—these findings are in keeping with the literature regarding exercise interventions and people with SMI (26, 60). Increases of greater than 50 m (achieved by over half of those completing our program) have been associated with clinically significant improvements in walking capacity in people with cardiovascular disease (29, 61) and at improving self-perceived fitness levels in other disease populations (62). Our results were higher than several outpatient lifestyle interventions (increases of 7–34 m), (29, 43, 44), however the average BMI of participants in these studies (BMI 34–37) was higher than ours (BMI 28.5), where obesity is known to cause impairments in functional exercise capacity (63). Our functional exercise capacity increases may also have been higher than other lifestyle studies in outpatients and residential rehabilitation studies due to the greater frequency of sessions we offered (three supervised sessions per week versus one), (13, 29, 44). Engagement in three or more sessions per week has been associated with greater fitness improvements for people with SMI (60), which may have relevance when planning exercise interventions in a residential rehabilitation setting.

There were no significant anthropometric or metabolic changes following our 10-week study. Previous lifestyle interventions in both residential and outpatient rehabilitation settings have demonstrated various reductions in anthropometric (weight, BMI, waist) and metabolic syndrome components in people with SMI (13, 14, 30, 64). There may be several reasons for this finding.

Firstly, people accessing 24 h residential services can have substantial motivational and cognitive deficits (11, 65) and may require more intensive support to integrate changes into their diet. The number of diet intervention sessions we offered (six same-day cooking groups and individual sessions) may have been insufficient to facilitate behavior change. Previous research in both a supported housing and outpatient rehabilitation setting demonstrated, improvements in metabolic syndrome parameters (HDL, systolic BP, body weight) occurred after substantially more diet interventions than our study employed (35 group sessions and 21 combined individual and group sessions respectively), (13, 30).

Secondly, whereas weight loss tends to peak early after lifestyle interventions in the general population, weight loss may occur more gradually for people with SMI (14, 30, 66), hence our 10-week intervention may have been too short to detect significant change.

Thirdly, CCU’s provide support to residents to cook and shop independently but do not provide food on site. Successful weight loss has been reported in rehabilitation settings where food was provided (14, 30) and hence caloric intake and nutrition quality could be modified more easily than in a CCU where participants had to shop for and cook their own food. Healthy meal delivery to replace meals for some CCU residents who experience ongoing barriers to cooking healthy meals regularly despite attempted engagement in rehabilitation to increase independent cooking skills (ie prominent negative and cognitive symptoms), may be an option to improve metabolic outcomes for some residents in future, which may be preferable to “take-away” and discretionary food, with the caveat that this option may increase dependency on external services and need to be weighed carefully against the priority of rehabilitation services to increase independent functioning.

Finally, our sample included participants with chronic illness (mean 8.5 years) in which more than half (62%) were taking the most metabolically unfavorable medications (clozapine and olanzapine), known to substantially increase body weight via hunger and metabolic disturbances (67, 68). Greater anthropometric reductions have been found in samples where olanzapine and clozapine prescriptions were much lower (30) and when dietary interventions have been conducted at antipsychotic initiation, highlighting the importance of early intervention (69).

Of note, our participants increased their fitness while maintaining their weight during the 10-week intervention. This may be a more realistic cardiometabolic disease risk reduction goal for people living with longer term exposure to metabolically unfavorable antipsychotics, and particularly in this population with high levels of treatment resistance (11), where switching to antipsychotics with a more favorable cardio-metabolic profile may not be possible due risk of mental state deterioration.

There was a significant reduction in negative symptoms of schizophrenia and general psychiatric symptoms following the lifestyle intervention, consistent with the literature regarding exercise interventions and people with SMI (45, 70). While we observed improvements in social functioning during exercise participation in a subset of our participants, this was a small, uncontrolled study and improvements in symptoms could have occurred due to other aspects of the rehabilitation treatments available at the CCU. We found the physical health component of quality of life improved, confirming the physical health focus of the intervention, however none of the psychological quality of life domains improved. This may relate to the length of the intervention (14, 71), that the lifestyle intervention was not specifically designed to improve psychological constructs (72), or that weight loss did not occur in our participants (a pathway to improved self-efficacy and quality of life) (73).

This brief student-led diet and exercise intervention was found to be feasible, acceptable, and safe for senior students to lead while on practicum placement and could be replicated across residential rehabilitation facilities with minimal resources. Results of this brief study led to the development of an ongoing relationship with local universities providing senior MDiet and EP students on practicum placement at each of the sites throughout the year, resulting in the provision of student-led diet and exercise service to people living in CCU’s. It is beyond the scope of this study to report cost-effectiveness data, however future studies should address this as it is a key factor in determining successful implementation (23). While senior MDiet, EP students and mental health staff were afforded mutual learning opportunities in this novel design, qualified exercise physiology and dieticians could achieve better adherence and physical and mental health outcomes (52, 69).

This was a small, unblinded, uncontrolled pilot study. Psychiatric symptom scales were most prone to observer bias, and these results should be interpreted with caution and future real-world studies in a residential setting should include a control group. Preliminary effectiveness revealed improvements in a number of mental and physical health outcomes, however when multiple analyses were corrected for, only exercise capacity and negative symptoms remained significant; this was a small feasibility study and larger studies are required to confirm our results.

Additionally, while motivational strategies were employed by students, we did not utilize a specific behavior change theory to inform intervention design, limiting discussion regarding implications for clinical practice in light of the reduction in motivation we found and also limiting potential replication of the design in real world settings. Future lifestyle intervention studies in residential settings should utilize and articulate a clear behavior change model (74). While commonly utilized, physical activity self-report measures are prone to recall bias (75), particularly in this population of people with SMI utilizing residential services who can experience significant cognitive impairment. Future studies using device-based measures, (ie accelerometry), could detect differences in sedentary behavior and physical activity more objectively, and particularly as recent research has reported accelerometry may be the most preferable measure of PA by people with SMI (76). Participants reported improvements in their diet following the intervention, however we did not measure energy consumption, diet quality or macro- and micro-nutrient intake, and this should be an important focus in future dietary interventions. Finally, the lower completion rate of the acceptability questionnaires may not have been representative of the entire sample recruited into the study.

Implementation of a brief student-led lifestyle intervention within mental health residential rehabilitation facilities for people with SMI is feasible, safe, scalable, and may improve functional exercise capacity, exercise participation, and negative symptoms. This pragmatic study may be relevant to other CCU’s and other residential rehabilitation units. People living in CCU’s may require more intensive dietetic support to positively impact metabolic risk. Longer duration of lifestyle interventions and investigation of additional interventions to support sustaining healthy lifestyles are warranted.

The datasets generated for this study are available on request to the corresponding author.

The studies involving human participants were reviewed and approved by Metro South Human Research Ethics Committeein March 2016: (HREC 16/QPAH/042). The patients/participants provided their written informed consent to participate in this study.

NK wrote the protocol in conjunction with DS. CD, NK, SP, CC, and SS collected data. NK, CD, and HF entered data. NK, HF, and DS performed statistical analyses in conjunction with TS. CD supervised the exercise intervention. NK and HF wrote the first draft of the manuscript. All authors NK, HF, CD, TS, JC, SS, DS, FD, CC, and SR revised the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.