. 2024 Oct 8;3:46. doi: 10.1038/s44184-024-00092-9

Table 4.

STROBE guidelines

	Item No.	Recommendation	Page No.	Relevant text from manuscript
Title and abstract	1	(a) Indicate the study’s design with a commonly used term in the title or the abstract	1, 5	“associations” in the title and abstract. “linear mixed-effects regression models“ in the abstract.
Title and abstract	1	(b) Provide in the abstract an informative and balanced summary of what was done and what was found	5	What was done: “RC fidelity data were collected from 169 of all 221 RCs currently operating, spanning 28 WEIRD and non-WEIRD countries. Hofstede’s cultural dimension scores were entered as predictors in linear mixed-effects regression models, controlling for GDP percentage spent on healthcare and Gini coefficient.” What was found: “Higher Individualism (β = 0.06) and Indulgence (β = 0.05), as well as lower Uncertainty Avoidance (β = −0.04) were associated with higher RC fidelity with Long-Term Orientation being a borderline negative predictor of the fidelity (β = −0.03).”
Introduction
Background/rationale	2	Explain the scientific background and rationale for the investigation being reported	6–8	The scientific background: Despite RCs located in many countries including both WEIRD countries and non-WEIRD countries, evidence base for how RCs can help mental health recovery is oriented to WEIRD countries only. There are six reviews about RCs, and all included studies apart from one international study were from WEIRD countries (185 included studies in total). Rationale: Cross-cultural understanding of RCs remains to be developed. RCs are in operation in non-WEIRD countries including low- and middle-income countries.
Objectives	3	State specific objectives, including any prespecified hypotheses	8	“This study aimed to explore the relationships between cultural characteristics and fidelity in all currently-operating RCs internationally. RC inclusion criteria were targeting to support personal recovery, and prioritising co-production and adult learning. Our research questions were; 1. Are there associations between cultural characteristics and the operational indicators of RC fidelity?, and 2. If there are, which cultural characteristics are associated with the operational indicators of RC fidelity?” Addressing these research questions is intended to identify cultural impact on RC fidelity. Because RCs are in operation in many countries, identifying the cultural impact of RC operation will be useful to cross-cultural understanding of RC operation, and by extension, will have relevance to other recovery-oriented global innovations, such as mental health peer support work³⁷. We recruited RCs that were currently in operation from 28 countries across different cultures (Table 2 for the participating WEIRD and non-WEIRD countries), and evaluated whether differences in cultural characteristics could predict variance in their fidelity scores. Mixed-effects linear regression models with a country-level random intercept were used to allow us to identify associations between the cultural characteristics and RC fidelity while accounting for variability between countries. No hypotheses were predefined due to the exploratory and inductive nature of the research³⁸.
Methods
Study design	4	Present key elements of study design early in the paper	1, 5, 8 and 12	Noted above, in 1(a) and 3. Additionally, “We conducted a cross-sectional, observational survey in two rounds: first of all RCs in England (“England survey”)³, then of RCs in all other countries (“international survey”)². Approval was obtained from King’s College London Research Ethics Psychiatry Nursing and Midwifery Subcommittee on 09/02/22 (MRA-21/22-28685). All participants provided written informed consent prior to completing the survey. The study was conducted as part of the RECOLLECT programme¹⁶. RECOLLECT is a five-year (2020-2025) National Institute for Health and Care Research (NIHR)-funded research programme exploring the effectiveness and cost-effectiveness of RCs¹⁶.”
Setting	5	Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data collection	12	“We included all RCs whose managers completed the RECOLLECT Fidelity Measure between August and October 2021 for the England survey, and between February and October 2022 for the international survey.
Participants	6	(a) Cross-sectional study—Give the eligibility criteria, and the sources and methods of selection of participants	13 and Supplementary Information 2	“Because not all RCs named themselves a “Recovery College” (e.g., “Recovery Academy”, “Recovery School”), we included any currently active services that met three criteria, informed by the key RC components⁴. The criteria were: (a) targeting to support personal recovery; (b) prioritising co-production and (c) adult learning, and were confirmed by the service managers. Full details are reported elsewhere², and presented in Supplementary Information 2.”
Variables	7	Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if applicable	14	Outcome variable, predictor variables, and confounder variables are defined.
Data sources/ measurement	8	For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe comparability of assessment methods if there is more than one group	13-14	“Fidelity was measured using the seven nonmodifiable components of the RECOLLECT Fidelity Measure⁴, completed by the manager of each RC.” Predictor variable: “Data for the cultural characteristics were obtained from Hofstede⁷⁰…. The data were collected using the Value Survey Module 2013” Confounder variables: “Two confounder variables were included in the fully adjusted analyses, as relevant to mental health treatment resources⁶⁵ and Hofstede’s index^35,36, as well as the significant global variation in financial statuses for RC operations². The percentage of GDP spent on health is the amount spent on healthcare relative to the economy size, calculated by the total health expenditure divided by GDP⁷². The Gini coefficient for each country indicates the income inequality within a nation, expressed from 0 (perfect equality) to 1 (maximum inequality), obtained from the World Bank⁷³.”
Bias	9	Describe any efforts to address potential sources of bias	14	“Fidelity scores were summarised as medians and interquartile ranges where possible (i.e. for countries which provided fidelity data for multiple RCs). In order to examine unadjusted and adjusted associations between each cultural characteristic (country-level) and fidelity scores (college-level), we used mixed-effects linear regression models with a country-level random intercept in order to account for variability between countries. Adjusted associations included the percentage of GDP spent on healthcare and the Gini coefficient for each country as potential confounders. Uganda were missing data for Power Distance, Individualism, Success-Drivenness, and Uncertainty Avoidance, and was therefore excluded from analyses involving these cultural predictors. Gini coefficients for Hong Kong and New Zealand were unavailable from the World Bank due to high costs (Personal communication on 28 April 2023, The World Bank, Development Economics Data Group): these two countries were omitted from adjusted mixed-effects linear regression models.”
Study size	10	Explain how the study size was arrived at	12-13	“We included all RCs whose managers completed the RECOLLECT Fidelity Measure between August and October 2021 for the England survey, and between February and October 2022 for the international survey.”Three steps were followed for both surveys: (1) Developing RC inclusion criteria, (2) Identifying and approaching potentially eligible RCs, and (3) Disseminating and collecting the survey.”
Quantitative variables	11	Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why	13-14	All quantitative variables were handled as continuous variables. This is detailed for Fidelity, Cultural Characteristics, and confounder variables.
Statistical methods	12	(a) Describe all statistical methods, including those used to control for confounding	14-15	Mixed-effects linear regression models - efforts to control for confounding are described in the Statistical analysis section.
		(b) Describe any methods used to examine subgroups and interactions	N/A	The examination of subgroups or interactions was not applicable in this study
		(c) Explain how missing data were addressed	14-15 and Table 3	Complete case analysis - countries were omitted from analyses if data were missing. Imputation was not appropriate.
		(d) Cross-sectional study—If applicable, describe analytical methods taking account of sampling strategy	N/A	A complex sampling strategy was not used in this study, so we did not need to adjust for this.
		(e) Describe any sensitivity analyses	N/A	No sensitivity analyses were performed.
Results
Participants	13	(a) Report numbers of individuals at each stage of study—e.g. numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and analysed	14 and Table 5	“The surveys were completed by 169 (76%) RC managers from 28 countries, with more than 55,000 students attending in total. A description of the sample and summaries of variables of interest are provided in Table 5.”
		(b) Give reasons for non-participation at each stage	14–15 and Table 5	Statistical analysis section and Table 5 provide sample sizes that indicate non-participation in analyses.
		(c) Consider use of a flow diagram	N/A
Descriptive data	14	(a) Give characteristics of study participants (ego demographic, clinical, social) and information on exposures and potential confounders	Table 5
		(b) Indicate number of participants with missing data for each variable of interest	Table 3	The number of colleges with missing data for certain variables are provided in Table 3 (e.g., Gini coefficient missing for New Zealand = 2 RCs).
		(c) Cohort study—Summarise follow-up time (e.g., average and total amount)	N/A
Outcome data	15	Cross-sectional study—Report numbers of outcome events or summary measures	Table 3
Main results	16	(a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (ego, 95% confidence interval). Make clear which confounders were adjusted for and why they were included	Table 3	Unadjusted and adjusted estimates and 95% CI are provided. Table 3 also reports the results of all outcomes, both significant and no significant associations. Clear list of covariates in table footnote.
		(b) Report category boundaries when continuous variables were categorized	N/A	No continuous variables were categorised.
		(c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period	N/A	We do not report relative risk.
Other analyses	17	Report other analyses done—e.g. analyses of subgroups and interactions, and sensitivity analyses	None
Discussion
Key results	18	Summarise key results with reference to study objectives	10	“In this global study, we found that higher levels of Individualism and Indulgence, and lower levels of Uncertainty Avoidance were associated with higher RC fidelity scores, with Long-Term Orientation being a borderline negative predictor.”
Limitations	19	Discuss limitations of the study, taking into account sources of potential bias or imprecision. Discuss both direction and magnitude of any potential bias	11-12	“Strengths of our study include it being the first global cross-cultural study of RCs in 28 countries, informing RC development globally. Service disparity for minority cultures is a global mental health concern, associated with poor service uptake, adverse mental health outcomes, and increased costs²⁵. RCs are operating in 28 countries including low- and middle-income countries (LMICs). Our findings can inform cultural adaptation of RCs, helping to address service disparity. Additionally, the numbers of RCs in non-WEIRD countries, including LMICs (Table 4), suggest they may be in the early stages of RC implementation. Our findings could guide the initial steps towards scaling RCs in those countries. The establishment of more RCs in non-WEIRD countries would promote the inclusion of diverse cultures in RC operations worldwide. Although our study included all operational RCs globally, participation from non-WEIRD countries was limited, with only 15 RCs from four such countries represented. To better understand RC cultural adaptations in non-WEIRD contexts, a greater presence of non-WEIRD RCs is needed. Several study limitations can be identified. First, there are other cross-cultural frameworks that could have been used (e.g., tightness-looseness⁵⁷). However, data for many of the 28 countries were not available, making meaningful comparisons problematic. Relatedly, critiques of Hofstede’s definition of culture³² include overgeneralisation such as treating nations as a cultural unit⁵⁸ and under-emphasis on non-psychological cultural aspects such as socioeconomic and ecosocial factors^59,60. To effectively inform the cultural adaptation of RC, these factors must be assessed using more in-depth approaches⁶¹. For example, community-based participatory research, which involves close collaboration with local communities, stakeholders, and cultural minority groups, is recommended to identify the most appropriate cultural adaptations⁶². Common research processes in WEIRD countries, such as interviews, can make people in non-WEIRD countries feel like ‘subjects’ thus may not capture authentic responses. Culturally appropriate processes, such as casually asking around in their natural environment (e.g., Pagtatanong-tanong in the Philipines⁶³), can be more effective in eliciting genuine responses that are useful for cultural adaptations⁶⁴. Moreover, the person-centred approach to recovery that RCs emphasise, may seem contradictory to our evaluation on cultures. However, the cultural dimension scores regard collective tendencies, instead of personal factors³². Therefore, our findings inform associations between cultural characteristics and RC operation. Second, although we included two relevant confounders in fully adjusted analysis, it is possible there were unmeasured confounders that may bias our results. We used these two confounders due to their relevance to mental health treatment resources⁶⁵ and Hofstede’s index^35,36, and the significant global variation in financial statuses for RC operations². However, since no studies have directly examined the relationship between Hofstede’s index and mental health intervention fidelity, other country-level confounders, such as trust in government⁶⁶, may also be relevant. Third, RCs with missing data for outcomes of interest or confounders were excluded. The uneven distribution of RCs across countries limits the robustness of the findings. Fourth, the survey was completed by service managers, therefore may not reflect the other people’s perspectives. Following the philosophy of RCs, the fidelity assessment should be done by RC students too. This raises the deeper issue of reducing fidelity to a quantitative score (as done with the RECOLLECT Fidelity Measure in this study), which may not capture many important operating characteristics, such as psychological safety and the impact of the built environment on student and trainer wellbeing. Future research should involve student assessment after addressing ethical concerns, with rigorous and reasonable sampling methods in each country or context (e.g., how to identify people who can assess an RC comprehensively). Consideration should also be given to developing more qualitative approaches to characterising fidelity, for example using the Impacts of Recovery Innovations (IMRI) framework⁶⁷. Lastly, as cultures and practice change over time, cross-cultural understanding of RCs needs to be investigated periodically.”
Interpretation	20	Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from similar studies, and other relevant evidence	10-12	From “The results indicated that characteristics typically associated with WEIRD countries predicted RC fidelity.” To “Lastly, as cultures and practice change over time, cross-cultural understanding of RCs needs to be investigated periodically.”
Generalisability	21	Discuss the generalisability (external validity) of the study results	10	From “The results indicated that characteristics typically associated with WEIRD countries predicted RC fidelity.” To “leading to a more holistic understanding of mental health and recovery^46,47.”
Other information
Funding	22	Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based	12	“The study was conducted as part of the RECOLLECT programme¹⁶. RECOLLECT is a five-year (2020-2025) National Institute for Health and Care Research (NIHR)-funded research programme exploring the effectiveness and cost-effectiveness of RCs¹⁶.”