The health impact of residential retreats: a systematic review

Background Unhealthy lifestyles are a major factor in the development and exacerbation of many chronic diseases. Improving lifestyles though immersive residential experiences that promote healthy behaviours is a focus of the health retreat industry. This systematic review aims to identify and explore published studies on the health, wellbeing and economic impact of retreat experiences. Methods MEDLINE, CINAHL and PsychINFO databases were searched for residential retreat studies in English published prior to February 2017. Studies were included if they were written in English, involved an intervention program in a residential setting of one or more nights, and included before-and-after data related to the health of participants. Studies that did not meet the above criteria or contained only descriptive data from interviews or case studies were excluded. Results A total of 23 studies including eight randomised controlled trials, six non-randomised controlled trials and nine longitudinal cohort studies met the inclusion criteria. These studies included a total of 2592 participants from diverse geographical and demographic populations and a great heterogeneity of outcome measures, with seven studies examining objective outcomes such as blood pressure or biological makers of disease, and 16 studies examining subjective outcomes that mostly involved self-reported questionnaires on psychological and spiritual measures. All studies reported post-retreat health benefits ranging from immediately after to five-years post-retreat. Study populations varied widely and most studies had small sample sizes, poorly described methodology and little follow-up data, and no studies reported on health economic outcomes or adverse effects, making it difficult to make definite conclusions about specific conditions, safety or return on investment. Conclusions Health retreat experiences appear to have health benefits that include benefits for people with chronic diseases such as multiple sclerosis, various cancers, HIV/AIDS, heart conditions and mental health. Future research with larger numbers of subjects and longer follow-up periods are needed to investigate the health impact of different retreat experiences and the clinical populations most likely to benefit. Further studies are also needed to determine the economic benefits of retreat experiences for individuals, as well as for businesses, health insurers and policy makers.


Background
Lifestyle-related chronic diseases such as obesity, diabetes and lung disease are a global issue, which the World Health Organisation estimate account for 60% of all deaths [1]. These diseases are characterised by modifiable risk factors such as physical inactivity, unhealthy diet such as diets high in salt, sugar, fat, alcohol and tobacco, and exposure to environmental toxicants [1]. Unhealthy lifestyles are a major factor in the development of chronic disease and are directly addressed by the health retreat industry, which promises to deliver enhanced health and the reversal of chronic disease and age-related conditions by engaging people directly in healthy lifestyle behaviours and experiences [2][3][4].
Health retreats have emerged from a history of travel to foreign destinations such as spas, hot springs, sacred sites, and pilgrimage locations that have been used as places of rest and rejuvenation for countless generations [3,5]. Such locations have given rise to a booming wellness tourism industry that is estimated to have generated US$563.2 billion in revenues in 2015, with growth projections that are nearly 50% faster than for overall global tourism [6,7]. A retreat may be defined as "a purposebuilt centre which accommodates its guests for the purpose of learning/improving a body-mind activity (e.g., yoga, pilates) and/or learning-receiving complementary therapies or treatments whilst there" [2]. Retreats cover a broad spectrum of facilities ranging from low-cost ashrams in India [2,3,8] that focus on a spiritual-based lifestyle, to luxury lifestyle resorts [8], to residential centers that focus on chronic disease. Retreat guests range from people who want to improve their general health and learn positive lifestyle practices, to those facing lifethreatening illnesses, and others who seek greater spiritual awareness or body-mind-spirit rejuvenation [9].
Despite the growing popularity of retreats and the growth of wellness tourism, the health impacts (either positive or negative) of attending retreats are uncertain and it is unclear if retreats offer any economic return for individuals or other stakeholders such as businesses, health insurance companies or governments. This paper aims to systematically review published studies that report on the health, wellbeing and economic impact of retreat experiences according to the PICOS approach (Participants, Intervention, Comparators, Outcomes and Study design) [10], and thereby explore the impact of these experiences on retreat guests.

Search strategy and data sources
A literature search was conducted in February 2017 using the electronic databases MEDLINE, CINAHL, and PsychINFO using search terms as appropriate for each of the databases. The search used combinations of keywords and phrases (retreat, health, wellness, wellbeing, resident) to conduct the systematic review. Truncation of keywords was used where variations of these words may alter search results. In addition reference lists of all relevant studies were manually searched.

Inclusion/ exclusion criteria
Studies were included if they were published prior to February 2017, written in English, and contained before and after data related to health and wellbeing of retreat guests. As there is no strict definition of 'retreat' , we included all studies that had at least one health-related outcome and the intervention involved a residential setting of one or more nights. Studies that did not meet the above criteria or contained only descriptive data from interviews or case studies were excluded. We did not exclude studies based on the purpose of the retreat.

Data extraction
Each potentially eligible study identified in the literature search was independently screened according to the study inclusion criteria and then independently reviewed. Detailed summary tables of the studies were prepared according to the PICOS approach [10]. Participants included both healthy people and people with specific diseases who attended the relevant retreat program; the interventions were all residential retreat programs that involved one or more nights stay excluding hospital stays; comparisons were made between post-retreat and pre-retreat measures; and outcomes included any physiological, psychological, or other clinically relevant outcomes. Data from all included studies were extracted by two independent authors and presented in Tables 1, 2 and 3 along with p values when p was less than 0.05.

Risk of bias assessment
Two review authors independently assessed the risk of bias of each included Randomised Controlled Trial study using the Cochrane Risk of Bias tool including key criteria such as random sequence generation, allocation concealment, blinding of participants, blinding of personnel and outcomes, incomplete outcome data, selective outcome reporting, and other sources of bias in accordance with methods recommended by The Cochrane Collaboration [11]. The following judgements were used; low risk, high risk or unclear (either lack of information or uncertainty over poteintial for bias). Non-Randomised Controlled Studies and Longitudinal Cohort Studies were assessed using the Risk of Bias in Non-Randomised Studies-of Interventions (ROBINS-I) tool. Key criteria included confounding, participant selection, intervention classification, deviations from intended interventions, missing data, outcomes measurement and reported results. The following judgements were used; low risk, moderate risk, serious risk, critical risk or no information. Authors resolved disagreements by consensus, and a third author was available for consultation to resolve any discrepancies if necessary. Risk of bias assessments are summarised in Tables 4, 5 and 6.

Results
There were 23 studies (reported in 28 articles) included in this systematic review, published over a 22-year period from 1995 to 2017 and involving 2592 participants. Of the 23 studies included, eight were randomised controlled trials (RCTs) including, one quasi-randomised trial and one randomised multi-centre trial; six non-randomised controlled trails and nine longitudinal cohort studies. A study flow chart is provided in Fig. 1. The results from the RCTs are presented in Table 1, results from the non-randomised controlled trials are presented in Table 2 and results from the longitudinal cohort studies are presented in Table 3.      Changes in spirituality were positively associated with increased well-being, meaning in life, and confidence in handling problems, and with decreased tendency to become angry. No sig.

Interventions
The retreat length of stays ranged from two and a half days [17,35] to 15 days [34] with a duration of five to seven days being the most common [12-16, 21-23, 26, 28, 31-33, 36]. Four retreats included a follow-up intervention period [17,23,27,30,34] that ranged in frequency, duration and mode of delivery from a couple of four-hour sessions over 12 weeks [17] to weekly followup via telephone over a three-month period [27].
Retreat programs ranged from a focus on religion and spirituality to lifestyle, health and wellbeing. Spiritualityfocused retreats involved different spiritual/religious denominations and practices including mindfulness meditation [13,17,18,34], Vipassana meditation [16,19,20,25], Ayurveda [14,15] and Ignatian/Jesuit spirituality [21]. These retreats included activities such as prayer, mass, chanting, observing silence and other techniques such as breathing and mindfulness. Health and wellness-focused  retreats included activities such as exercise, yoga, body treatments, medical consultations, counselling, support groups and discussion [12,22,27,31,33]. Both spirituallyfocused and health-focused retreats commonly included meditation as an activity, sometimes optional, along with a focus on either a prescribed diet such as organic diet, [12] vegetarian diet [15,35] or low-fat diet, [24,27,28,31,33] or dietary education such as mindful eating [16,18] or nutritional counselling. [35] In the four studies with a follow-up intervention, activities included review and practice of techniques taught at the retreat such as mindfulness, [17,34] the continuance of practices and processes that began at the retreat such as lifestyle changes and cleansing, [23] or telehealth support from a study nurse [27].

Comparators
Of the eight controlled trials reviewed, five included vacation groups [13-15, 18, 26], who visited the same  retreat purely for relaxation purposes without participation in organised retreat activities. One of these studies included an additional comparator group to compare results between novice and experienced meditators [13].
One study for HIV/AIDS patients had a group who continued to receive their standard care [34], and two studies; one for Human Service Professionals and another for non-metastatic breast cancer patients, had a group who received no intervention [17,30].
Of the six non-randomised controlled trials reviewed, five studies included comparator groups who received no intervention [16,19,20,25,29] and one study included a group who received standard care [27]. Only two of the nine longitudinal cohort studies included a comparator group, with one study comparing results from novice to experienced meditators [22] and another comparing results from healthy heart patients to their partners [35].

Summary of objective and subjective outcomes
All studies reported statistically significant improvements in at least one measured outcome with only one study of 44 human service professionals undertaking a two-and-a-half day Mindfulness with Metta Training (MMTP) Retreat, reporting no significant differences in self-report measures of resilience, mindfulness and selfcompassion immediately after the retreat experience, despite significant improvements for mindfulness and self-compassion at one and four months and for resilience at four months post-retreat [17]. A further study of 47 patients with HIV/AIDS who participated in a 15-day ' Art-of-Living with HIV' retreat reported significant positive changes in wellbeing immediately after the retreat, that were not evident after 6 and 12 weeks [34].

Objective/ quantitative outcomes
All seven studies investigating objective outcomes reported statistically significant improvements immediately after the intervention. Three of these studies reported significant improvements in anthropometric measures such as weight, abdominal girth and blood pressure [12,23,27,41] and reductions in blood lipids [41]. Statistically significant results were also reported for decreases in dopamine transporter binding in the basal ganglia and serotonin transporter binding in the midbrain [22]; changes in resting state functional connectivity (rsFC) in the right amygdala-subgenual anterior cingulate cortex (sgACC) [18]; reductions in 12 phosphatidylcholines and an additional 57 metabolites such as amino acids, biogenic amines, acylcarnitines, glycerophospholipids and sphingolipids [15]; gene expression changes associated with improved regulation of stress response, immune function and amyloid beta (Aβ) metabolism [13]; and electroencephalogram (EEG) changes [22]. Ornish et al. [27] further documented increases in relative telomere length after five years that was associated with the degree of adherence to lifestyle changes in ten of 35 men with biopsy-proven prostate cancer [27,41].

Subjective/ qualitative outcomes
Fifteen of the 16 studies investigating subjective or surveybased outcomes reported statistically-significant improvements immediately post-retreat including significant improvements in quality of life, perceived physical health and health symptoms, as well as a variety of psychological and spiritual measures [12-14, 16, 17, 19-21, 23, 25, 28, 29, 31, 34, 36]. Two studies reported improvements in overall health-related quality of life [28,31] and four studies improvements in perceived physical health [20,21,25,31]. Cohen et al. [12] reported improvements in both subjective and outcome measures including cognitive function and Conboy et al. [23] reported improvements in positive health behaviours and self-efficacy.
Eight of the nine studies measuring psychological wellbeing reported statistically significant improvements in a variety of indicators including depression, anxiety, tension, stress, fatigue, mindful awareness and vitality [12,13,20,21,25,28,29,31,36]. Khurana and Dhar [19] reported improvements in subjective wellbeing and criminal propensity, however this improvement was only seen in male inmates of the intervention group, and not in female inmates or the control group that did not receive the intervention. All six studies measuring spiritual wellbeing reported significant improvements in various religious and spiritual measures [14,16,21,24,28,35]. Vella and Budd [28] reported improvements in overall spiritual wellbeing and Mills et al. [14] reported a significant increase in spirituality and gratitude in the intervention group that participated in a six-day Panchakarma Ayurvedic program, compared with no change in the control group that were on vacation at the same resort. Newberg et al. [21] reported significant changes such as more intense religious and spiritual beliefs, feeling more religious and more spiritual, and an increase in feelings of self-transcendence in 14 participants of a Christian faith.
Two studies [24,35] investigating the relationship between spirituality and health measures, found that measures of spirituality increased after a retreat along with increased well-being, sense of meaning and purpose in life, confidence in handling problems and a decreased tendency to become angry. Similarly, Emavardhana and Tori [16] found that heightened belief in Buddhist precepts was associated with positive change in self-concept and less self-criticism and increased Buddhist religiosity was correlated with reductions in the defences of displacement, projection and regression [16].

Risk of bias
All Randomised Controlled Trial Studies were found to have low, unclear or high risk of bias in one or more domains. The most high risk was reported for blinding of outcome assessment. Allocation concealment and reported data selection was not reported for the majority of studies and therefore unclear. All Non-Randomised Controlled Trial Studies and Longitudinal Cohort Studies were found to have low risk of bias for all domains except outcomes measures. All but one study [41] failed to report whether or not the outcomes assessor was aware of the participant intervention and were therefore found to have a moderate risk of bias for outcomes measurement. Given these findings, all studies are comparable to a well-performed randomised trial with regard to the majority of domains (low risk) except outcomes measurement (moderate risk) indicating the studies are sound for a non-randomised study with regard to this domain but cannot be considered comparable to a wellperformed randomized trial,

Discussion
The retreat industry is a niche sector of the wellness tourism industry that focuses on transformative experiences that aim to improve the health of participants through healthy lifestyle experiences, along with providing the skills and knowledge to help maintain healthy behaviours. The findings from the reviewed studies suggest there are many positive health benefits from retreat experiences that includes improvements in both subjective and objective measures. Most studies used a quasiexperimental design with small sample sizes, poorly described methodology with little follow-up data and reliance on self-report questionnaires to report on psychological and spiritual benefits. The results from the most rigorous studies that used randomized controlled designs were consistent with less rigorous studies and suggest that retreat experiences can produce benefits that include positive changes in metabolic and neurological pathways, loss of weight, blood pressure and abdominal girth, reduction in health symptoms and improvements in quality of life and subjective wellbeing.
In addition to facilitating general health improvements, there is evidence that retreat experiences can have a positive impact on chronic disease processes and provide benefits for some people with life threatening and/or chronic diseases. Of the four studies of retreat experiences aimed at improving quality of life for cancer patients [27,28,30,36], all showed some benefits from retreat participation, including improvements in quality of life, depression and anxiety scores, and increased telomere length, with benefits being recorded up to five years post-retreat. Similarly, benefits of retreat participation are reported for people with multiple sclerosis with improvements in quality of life along with physical and mental health being evident up to five years post-retreat [31,32]. Not all measures in the studies of lifethreatening chronic diseases improved [30,36], and as they are all small, poorly-controlled studies, more rigorous research is needed.
The finding that retreat experiences can lead to sustained and significant health improvements long after participants return home suggests that these experiences assist guests in making positive lifestyle changes and adopting healthy behaviours that lead to a variety of positive psychological, physiological, cognitive, clinical and metabolic effects. The ability to influence participants' health once they return home is dependent on many factors including the type of participants involved, the education and experiences provided during the retreat program, and the provision of follow-up activities such as online coaching, nutrition programs, or followup consultations with practitioners. Of the four studies that showed a reduced effect over time in some measures [12,17,34,35], two studies did not include a follow-up retreat component [12,35].While it is not possible to determine which parts of the retreat intervention have the greatest influence, it is likely that improvements in health are due to a combination of psychological and behavioural factors that lead to better coping mechanisms and enhanced resilience to stress, as well as metabolic factors that lead to alterations in gene expression and DNA repair mechanisms that are evident in the observed changes in the metabolome [15] and teleomere length [27,41].
Despite the potential for retreat experiences to benefit people with chronic and life threatening disease, the retreat industry does not routinely interact with the health care sector with few patients being referred to retreats by medical practitioners and retreat experiences generally not covered by third party payment schemes or eligible for tax deductions or incentives. The lack of integration between the healthcare and retreat sectors may be partly due to a lack of data with which to evaluate retreat experiences. Few retreats routinely collect and/or communicate data relevant to the healthcare sector, and even when formal studies such as those reviewed here are conducted, there is great heterogeneity in the range and scope of outcome measures, with few measures being comparable across studies. The retreat industry would therefore benefit from the use of a standardised dataset collected from guests on a routine basis. Such data could include a combination of psychological, cognitive, physiological, anthropometric and biochemical measures that together provide a holistic assessment of outcomes. This would allow retreat participants to evaluate and monitor the impact of their experiences and provide data to engage the medical profession and third party payers. It would also be beneficial for the industry to develop a standardised reporting system for retreat activities so that the influence of different types of retreat experiences can be assessed and results meaningfully compared across retreats and studies.
While retreat experiences appear to have positive health impacts, there is no published data on the economic impact of retreat experiences. There is however, substantial evidence that non-residential wellness programs, which share a similar focus on health promotion and lifestyle modification, provide a substantial economic return [42][43][44]. A review of 28 studies of corporate wellness programs [45] finds that the economic benefit of participation is substantially higher than the costs of providing the program. Stead [45] reports benefit-to-cost ratios averaging 3.4-1 which indicates that corporate companies receive on average US$3.40 for every US$1 invested in the respective wellness program. In addition to return on investment, employees benefit from participating in corporate wellness programs through experiencing better health, lowered disability payments and reduced health care expenditures, while companies benefit from reduced employee turnover, increased productivity [45] and reduced absenteeism and presenteeism along with intangible benefits such as being an employer of choice and attracting highly skilled employees and creating a positive corporate culture [45,46].
While the economic benefits of corporate wellness programs are becoming well established, it is unclear if similar benefits are offered by residential retreats. Future studies that include a health economic analysis are therefore needed to determine the cost-benefits of retreat experiences and the return on investment for participants, businesses, health insurers and policy makers. This may enable retreat operators to advocate for tax benefits, as well as inclusion in health insurance policies, and corporate wellness schemes. Furthermore, there is no data on the occurrence of adverse events. Future studies would benefit from including measures of adverse outcomes to confirm the safety and efficacy of retreat interventions.
Despite the consistent reporting of positive health effects from retreat interventions across multiple study designs and locations, the ability to draw definitive conclusions for any one condition or population is limited due to poor methodological rigor and substantial heterogeneity in study design, length and type of retreat program, target population, outcome measures and length of follow-up. Furthermore, while the reviewed studies included subjects from a wide range of demographic groups in multiple countries, only published English language studies were reviewed and it is uncertain if the findings can be generalized to the wider population. The use of mostly self-selected populations also introduces the possibility of selection bias, while a lack of blinding and adequate controls may introduce performance bias due to exposure to factors other than the specific intervention such as the vacation effect whereby health can improve from simply being removed from normal routines and behaviours. The lack of any reported adverse events may further indicate reporting bias with researchers not actively looking to identify adverse outcomes, or outcome measurement tools not being designed to capture adverse outcomes. Future studies, with more rigorous methodology and long-term follow-ups are now needed to determine the longevity of any effects, their mechanisms of action and the conditions most likely to respond.

Conclusion
The findings of this review suggest that retreat experiences appear to have positive health benefits that include benefits for people with chronic diseases. As the observed improvements in chronic diseases are based on a small number of patients, future research using larger numbers of subjects and longer follow-up periods is needed in order to determine the populations most likely to benefit and quantify any long-term health benefits. Future studies could also benefit from more rigorous study designs including the use of standardized outcome measures, more detailed descriptions of the retreat interventions and study population, and the inclusion of a health economics analysis in order to determine the economic benefits of retreat experiences for individuals, as well as for businesses, health insurers and policy makers.