Quality of Life, Healthcare Use and Cost of Practice From a Nationally-Representative Australian Survey to Inform Future Economic Evaluations of Contemplative Practices

All adults aged 18 or over, residing in Australia who were able to read, write and understand English were eligible for this study. Participants were excluded at the screening stage if they either did not answer the screening questions or if the quota for their demographic characteristics was full. Data were collected until the quotas were completed. The data were nationally representative for age, gender, ethnicity (Australian, Australian Indigenous, European, Asian, Other), region (Australian state), income category ($0-$26 k, $26 k-$52 k, $52 k-$91 k, $91 k-$156 k, $156 k-$208 k, $208 k +). Quotas were not interlocking.

Data were drawn from an Australian cross-sectional, nationally-representative survey conducted by our group to obtain population estimates for meditation and other contemplative practices, and to describe the sociodemographic qualities, health behaviours and utilisation, faith and spirituality, and mental health/wellbeing of practitioners and non-practitioners. Data were self-reported. The survey, included consent and screening items, sociodemographic questions, health characteristics, faith/spirituality, meditation/contemplative practice use, psychological distress (Kessler 6; Kessler et al., 2003), and health and wellbeing-related quality-of-life (EQ-HWB-9; Brazier et al., 2022). Data were collected through the Qualtrics platform from October 2023 to February 2024. The survey was preregistered (https://osf.io/etkh4) and pre-registered aims have been published previously (current aims were not included in the pre-registration); further information on the methods can be found in our other studies (Davies et al., 2025).

Demographic and baseline variables in this current study include gender, age, education level, household income, employment status, household composition, whether the participant had a disability, identified as LGBTQIA + , and Australian Indigenous status.

A recently developed quality-of-life instrument that may be suitable for capturing these outcomes is the EuroQol Health and Wellbeing (EQ-HWB) instrument (Brazier et al., 2022). Designed to be more sensitive than earlier preference-based instruments (such as EQ-5D, Devlin et al., 2022 or SF-6D, Brazier et al., 2002), the EQ-HWB is suitable for use in health, social care, and carer-related populations. Its psychometric performance has been validated in diverse Australian samples, including caregivers of children with health conditions, caregivers in families experiencing adversity, and in the general adult population (Bailey et al., 2024, 2025; Lee et al., 2024). Given its sensitivity to a broad range of well-being outcomes, the EQ-HWB may be particularly useful for assessing the impact of contemplative practices on mental health.

The EQ-HWB-9 is the short form of the EQ Health and Wellbeing instrument with 9 items (the long form has 25 items), including difficulties getting around inside and outside (mobility), difficulties doing day-to-day activities (activities), feeling exhausted (exhaustion), feeling lonely (loneliness), having trouble concentrating or thinking clearly (cognition), feeling anxious (anxiety), feeling sad or depressed (sad/depressed), feeling like one has no control over day-to-day life (control) and how much pain was experienced (pain) over the last seven days (EuroQol Research Group, 2022). Pilot UK preference-weights (index-scores) are now available for the EQ-HWB-9 (Mukuria et al., 2023) for use in economic evaluation. The instrument is still in experimental phase, prior to general release, and at the time of publication is only available to researchers for the purpose of validation (EuroQol Research Foundation, 2025).

The Meditation group was defined as persons having meditation as their main contemplative practice over the previous 12 months. Contemplative practices included: Meditation, Yoga, relaxation techniques, guided imagery, Tai Chi, Qi Gong, breathing techniques, visualisation, and other contemplative practices over the previous 12 months. The non-practitioner group included participants who were not represented in the first two groups.

Response options for all three questions were: (1) No, (2) No, I wanted to, but wasn't able, (3) Yes, I went once or twice, (4) Yes, I went several times, and (5) Yes, I went often. Unmet need was defined as participants who selected option (2), that they were not able to access the service, compared to all other options. For the regression analysis, responses 3–5 were coded as ‘Used mental health service’.

To investigate costs for meditation and other complementary practices, we asked two questions: “In the past 12 months, have you spent any money on the following to support your main practice? (select all that apply)”, with response options: (1) Classes, courses, retreats or workshops, (2) Smartphone apps or subscriptions, (3) Resources (e.g., mats or cushions, props, books, clothes), (4) Donations to practice organisations or communities, (5) Health or wellness related travel, (6) Something else, (7) None of the above. Participants could select multiple categories. “In the past 12 months, about how much have you spent overall on your main practice?”, with response options: (1) Nothing ($0), (2) $1–50, (3) $50–100, (4) $100–250, (5) $250–500, (6) $500–1,000, (7) $1,000–2,000, (8) $2,000–5,000, (9) $5,000–10,000, (10) $10,000–20,000, (11) $20,000–50,000, (12) $50,000 + .

Analyses were performed in STATA version 18 and Microsoft Excel. We calculated baseline demographics by three groups, meditators (as main practice), other contemplative practitioners and non-practitioners. Baseline characteristics were calculated for the three contemplative practices groups, using one-way ANOVAs. We applied the pilot UK preference-weights to the EQ-HWB-9 data (Mukuria, et al., 2023).

We conducted a one-way ANOVA to compare EQ-HWB-9 index scores across groups (meditation, other contemplative, no practice). Stepwise generalised linear regression models (GLMs) were conducted to investigate whether the differences observed between contemplative practice group (three groups) for EQ-HWB-9 index scores were maintained when we controlled for demographic and other variables. We used a gamma distribution and log link (as per Lee et al., 2024) for the GLM models, to address the skewed distribution of the EQ-HWB-9 utility values. In Model 1, we included contemplative practice group as the independent variable and EQ-HWB-9 index-scores as the dependent variable. In Model 2 we added basic demographics including gender(female/male), age (continuous), education level (continuous), income level (continuous), employment (not working/working full or part-time), household composition (factor; living alone, single parent, couple with children, couple without children, other) and disability (no or not stated/yes) based on differences in groups found in previous research (Davies et al., 2024). We also included Indigenous status (no/yes) and LGBTQIA + status (no- including prefer not to say/yes) to investigate these variables that had not been included in the cited papers, and we hypothesised that participants who sought mental health services may also turn to mindfulness and other contemplative practices. Variables that were not significant in the model were progressively removed from the highest p-value until only significant variables remained at the less than 0.05 level. In Model 3 we added in mental health service provision on the basis that we hypothesised that participants may consider meditation and other contemplative practices when they need mental health support, on the basis that participants seeking mental health support are regularly recommended to take up mindfulness practices (Russell & Siegmund, 2016).

People with unmet healthcare provision were defined as those who wanted to, but were not able to, access medical, mental health or complementary healthcare. The health service use variables were recoded to those with unmet need (participants who wanted to access the type of healthcare but were not able to) versus all other participants. Population estimates were calculated based on the Australian population of adults (18 +) as at June 2023 of 20,883,717 (divided by the sample size of 2069, making a multiplier of 10,094). One-way ANOVA analyses were conducted to measure mean differences across the three practice groups (meditation as a main practice, all other contemplative practices, no contemplative practice) by unmet need. EQ-HWB-9 preference-weighted scores for health service use (not needed, unmet need, used service) were calculated for general, mental and complementary health questions using one-way ANOVAs with post-hoc Sheffe tests.

For the analysis of the cost item data, we used numbers and percentages. For the cost of practice estimates, the options were in ranges. Therefore, we calculated the mid cost in dollars, as well as upper and lower cost limits (i.e. for the bin $1–50, the mid-point was $25, the lower limit was $1, and the upper limit was $50). We multiplied the number in each of the response categories by these to estimate the average plus lower and upper limits of the dollar amount spent overall and per person for the participant’s main practice.

Demographics by contemplative practice categories are displayed in Table 1. The analysis demonstrated significant differences across most demographic categories. Compared to meditators and other contemplative practitioners, non-contemplative practitioners were more likely to be male, older, less well educated, have lower incomes, more likely to be retired and less likely to be working full-time, more likely to be living with a partner, less likely to identify as LGBTQIA + , less likely to be Indigenous, and less physically active. Non-contemplative practitioners also had higher quality-of-life (higher EQ-HWB-9 index scores) and lower mental health distress scores (K6) than contemplative practitioners (meditators and other practitioners). There were significant differences in one-way ANOVA results between groups for age (meditation mean years (SD) = 41 (15), contemplative = 44 (17) and non-contemplative = 56 (17) (F(2,2062) = −118.00, p < .001)) and EQ-HWB-9 index-scores (meditation mean (SD) = 0.719 (.201), other contemplative = 0.720 (0.233) and non-contemplative = 0.804 (0.227) (F(2, 2062) = 29.51, p < 0.001)).

There were significant differences between groups on EQ-HWB-9 index-scores for meditators (0.719 (SD = 0.201)), other contemplative practitioners (0.720, SD = 0.233) and non-practitioners (0.804, SD = 0.226; F(df) = 29.51(2,2062), p < 0.001), with significant post-hoc tests between non-practitioners and the two contemplative practitioner groups (both p < 0.001) but not between the two practicing groups where the index-scores were almost identical.

The GLM models of the association between contemplative practices and EQ-HWB-9 preference-weighted scores are displayed in Table 2. Results from Model 1 suggested that there were significant differences for EQ-HWB-9 preference-weighted values between non-practitioners and meditators, and non-practitioners and other contemplative practices at the univariate level (deviance = 276.95, log-likelihood = −1450.50, AIC = 1.41) as had been seen in the ANOVA results. In Model 2, we added all demographic variables that were significant at baseline (gender, education level, income level, employment, household composition, disability, Indigenous status, LGBTQIA + status and age). We progressively removed household composition, Indigenous status, and employment variables, leaving gender, education level, income level, disability, LGBTQIA + status and age to arrive at the final model. Significant differences in EQ-HWB-9 preference-weighted scores between contemplative practice groups remained in this model (deviance = 241.81, log-likelihood = −1429.73, AIC = 1.40). In Model 3 we added mental health service utilisation and EQ-HWB-9 preference-weighted scores between contemplative practice groups were no longer significantly different (deviance = 226.72, log-likelihood = −1420.76, AIC = 1.40).

Numbers, percentages and population-estimates of people in each health service use category (not needed, unmet need, service used at least once) are presented in Table 3. We found significant differences between groups for general, mental and complementary health. Likely unmet mental health needs were found in 9% of the sample. For general health and complementary health, there were significant differences in EQ-HWB-9 index-scores between having unmet need and those who received a service (Scheffe value = 0.15, p < 0.001; Scheffe value = 0.11, p < 0.001, respectively). This was not the case in mental health service use where there were significant differences between people who did not seek services with those with unmet needs and used care (Scheffe value = 0.19, p < 0.001; Scheffe value = 0.18, p < 0.001, respectively), but no significant differences in index-scores between the unmet need and used services groups (Scheffe value = 0.01, p = 0.914), as can be seen graphically in Fig. 1.

Afbeelding vergroten

Unmet medical, mental health and complementary health care needs by practice categories are presented in Table S1, and visually in Fig. 2 for three health service types (general health, mental health, complementary health). Comparisons between those with unmet needs and all other groups were significant (p < 0.001) across all three types of health service use. Unmet general health need was particularly high for meditators, who were 9.1% more likely to have unmet general health needs than those not doing any contemplative practice. The "other contemplative practices" group was 4.5% more likely to have unmet general health needs than those not doing any contemplative practice (chi-squared = 37.63 (2), p < 0.001). Meditators were 11.5% more likely to have unmet mental health needs than those not doing any contemplative practice, and the other contemplative practice group was 8.2% higher than those not doing any contemplative practice (chi-squared = 47.68 (2), p < 0.001). Complementary health service use was also significantly different between groups (chi-squared = 35.01 (2), p < 0.001).

Afbeelding vergroten

The number of participants who spent any money on any contemplative practice item was calculated. Percentages were calculated by percentage of items and percentage of participants (participants could select more than one item) and are displayed in Table 4. The categories with the most prevalently used categories were resources and smartphone apps/subscriptions, followed by donations.

This study used a nationally representative survey of meditation practice in Australia, with quotas across a range of demographic variables. We did not use interlocking quotas which may lower representativeness. A limitation was that we could not include in-depth questions around expenses specific to different types of practice costs. The data for unmet health care need was derived from a multi-option variable; it is possible that a direct question may have elicited different results. Despite these limitations, we were able to determine estimates of health service use and practice costs. We used the pilot index-scores from the UK. As preference-weights become available for a range of different countries in the future, it would be more accurate to apply Australia-specific weights. For the health service use questions, we asked participants to think back over the previous year. It may be more accurate to ask participants to estimate their health service use over a shorter time-period, such as every month; however, this was not feasible in the current survey. It is therefore possible that health service use across the three variables has been underestimated. We note that quality-of-life differences between practitioners and non-practitioners were most likely driven by practitioners seeking contemplative practices to address mental health needs, rather than as a reflection of the efficacy of the practices themselves. We recommend that future studies aiming to estimate the effectiveness of mindfulness practices use different methods such as randomised controlled trials or cohort studies.

This study provides preliminary evidence on variables for conducting future cost-effectiveness studies in contemplative practice interventions. There were large demographic differences between contemplative practice groups and mental health service use, and unmet mental health need was a strong predictor for being in a contemplative practice group. EQ-HWB-9 index scores appeared useful as an outcome measure in this population, demonstrating differences between groups as well as performing well in regression analyses. The EQ-HWB-9 may be a useful tool for economic evaluation in this context. Group differences across meditation groups point to the need to adjust data by demographics, and further research on the relationship between contemplative practice, mental health service use and participants’ quality-of life outcomes in trial-based research is indicated. We provided estimates of unmet health service use, cost estimates of practice per person, plus a breakdown on what types of expenses for contemplative practices participants incurred. We anticipate that these results will be useful in future cost-effectiveness research in contemplative studies.