The present study evaluated, in a real clinical practice context, quality of life and sleep satisfaction in patients under supervised use of Cannabis oil rich in THC and CBD, using validated instruments and patient-reported outcomes relevant for capturing functional and subjective impact in heterogeneous chronic conditions. In the present study, the predominance of high scores for both outcomes suggests a globally favorable self-reported perception of treatment. Consistent with prior evidence, observational studies have described favorable patient-reported outcomes in medical Cannabis cohort studies (Arkell et al., 2023; Habib et al., 2021; Aviram et al., 2021; Cahill et al., 2021). Together, these findings reinforce the importance of evaluating patient-perceived benefits, especially in complex interventions where the clinical effect may involve multiple domains (pain, sleep, mood, functioning, and wellbeing) and not just an isolated biomedical marker (Zou and Kumar, 2018; Lu and Mackie, 2016; Cristino et al., 2020; AminiLari et al., 2022; Suraev et al., 2020; Skevington et al., 2004).

Regarding the age–quality of life axis, although comparisons between age groups did not reach conventional statistical significance in bivariate analyses in the present study, the direction of effects observed in the ordinal regression model suggests that adults had a lower probability of high quality of life levels compared to more advanced elderly, which requires clinical and contextual interpretation. In moderate observational samples, the absence of significance does not necessarily exclude clinical relevance, and it is more appropriate to jointly interpret the pattern of effect, plausibility, and coherence with the literature, especially when the outcome is self-reported and multidimensional (Skevington et al., 2004; Devlin and Brooks, 2017). Synthesis studies on pain and sleep with cannabinoids also highlight important heterogeneity of response according to population and study design, which may influence subgroup comparisons by age (Häuser et al., 2018; AminiLari et al., 2022; Suraev et al., 2020; Skevington et al., 2004). Moreover, expectations and adaptation to illness vary throughout the life course and may influence how therapeutic gains are translated into global assessments of wellbeing and satisfaction (Häuser et al., 2018; Skevington et al., 2004; Winiger et al., 2021). Notably, age has been identified as a moderator of the relationship between CBD concentration and sleep outcomes, with older adults showing a stronger association between higher CBD content and improved sleep efficiency and duration (Winiger et al., 2021).

At the extremes of age, especially in pediatric and adolescent populations with neurological and neurodevelopmental conditions, interpretation must be even more cautious due to diagnostic heterogeneity and different clinical trajectories. Proof-of-concept randomized trials in autism spectrum disorder suggest possible benefit in some clinical and sleep/behavior-associated outcomes, although methodological limitations and the need for standardization of formulations and outcomes persist (Aran et al., 2021). Additionally, reviews on cannabinoids and sleep describe variability of response according to clinical profile, cannabinoid metabolite used, and context of use, reinforcing the need to avoid simplistic extrapolations between populations and products (Suraev et al., 2020; Gates et al., 2014). Thus, the findings of the present study, while encouraging, are best interpreted as a signal of perceived effectiveness in clinical practice, and not as definitive evidence of comparative efficacy by age group.

A distinctive aspect of the present study is the use of a full-spectrum (chemotype I) oil with THC predominance, in contrast to part of the literature that favors CBD-rich formulations or standardized preparations. Systematic reviews and evidence overviews describe that effects and tolerability vary substantially according to formulation, dose, THC:CBD ratio, and patient selection, which may explain discrepant results and hinder direct comparability (Pratt et al., 2019; Häuser et al., 2018; AminiLari et al., 2022). In this context, the association observed in the present study between the 9% concentration and a higher probability of high quality of life levels should be treated with caution, since cross-sectional designs do not allow inferring causality or dose-response relationships, and concentrations may function as indirect markers of therapeutic phase, titration, clinical indication, or severity profile (AminiLari et al., 2022; Suraev et al., 2020; Skevington et al., 2004; Arkell et al., 2023). In other words, the concentration may reflect a “balance point” achieved by clinical adjustment over follow-up, rather than representing an isolated and linear pharmacological effect of the cannabinoid content.

The phytochemical composition is particularly relevant for interpreting the results observed with complex extracts such as the one used in this study. The “entourage effect” hypothesis proposes synergies between phytocannabinoids and terpenoids that could modulate efficacy, tolerability, and subjective experience, and is frequently invoked to explain differences between full-spectrum extracts and isolated cannabinoids (Russo, 2011). Although it remains a controversial and difficult-to-conclusively demonstrate topic in humans, this framework helps contextualize why different products may produce different response and adverse event profiles, especially in patient-centered outcomes such as sleep and quality of life (Arkell et al., 2023; Russo, 2011). Furthermore, studies and syntheses focused on prolonged use suggest that changes in composition and titration patterns over time may accompany variations in quality of life and analgesic consumption, reinforcing the need for product characterization in real clinical studies (Cahill et al., 2021; Bialas et al., 2022).

A key finding was the relationship between sleep and wellbeing as an integrating axis, evidenced by the positive correlation between psychological domains of the WHOQOL-BREF and sleep satisfaction. Consistent with prior evidence, systematic reviews and scoping reviews on cannabinoids and sleep indicate that subjective improvements are frequent but inconsistent, varying substantially according to population, metabolite, dose, ratio of cannabinoids, and route of administration (AminiLari et al., 2022; Babson et al., 2017; Amaral et al., 2023; Suraev et al., 2020; Gates et al., 2014). The PSQI, due to its wide use, allows interpretation of findings in the context of an extensive literature relating sleep to functioning, mental health, and global perception of quality of life (Buysse et al., 1989). Observational studies in outpatient settings (including rheumatology and pain clinics) also describe concomitant self-reported improvement in sleep and symptoms among medical Cannabis users (Habib et al., 2021; Aviram et al., 2021). Thus, the results are consistent with the hypothesis that cannabinoid interventions may influence wellbeing through direct and indirect pathways (sleep, relaxation, anxiety reduction, and symptom modulation), beyond the primary control of a target symptom (Zou and Kumar, 2018; Lu and Mackie, 2016; AminiLari et al., 2022; Babson et al., 2017; Amaral et al., 2023; Suraev et al., 2020; Gates et al., 2014).

One of the most relevant findings of the present study was the observed association between longer treatment duration (>3 years) and higher self-reported sleep satisfaction (Table 7). Regarding the broader literature on long-term cannabinoid use, evidence is heterogeneous and often limited by low-certainty studies, with important variation in outcomes, tolerability, and discontinuation rates (Bialas et al., 2022). Guidelines and synthesis reports highlight potential risks of THC, including cognitive and psychiatric effects in certain contexts, especially with chronic high-dose use, early onset, or psychiatric vulnerability, making it essential to avoid overly optimistic interpretations (Bialas et al., 2022; Busse et al., 2021; Nugent et al., 2017). Evidence on the effects of Cannabis for chronic pain, a common indication in the present sample, remains limited in strength, with low-certainty evidence suggesting modest analgesic benefit in neuropathic, musculoskeletal, and orofacial pain, and insufficient evidence for other pain types (Bort et al., 2024; Nugent et al., 2017). A comprehensive narrative review including 74 studies and 12,562 patients found that cannabinoids were most effective for chronic secondary musculoskeletal pain, neuropathic pain, and orofacial pain, with treatment initiation recommended at low doses and gradual titration (Bort et al., 2024). At the same time, observational evidence suggests that, under gradual titration and continuous follow-up, some patients maintain subjective benefits over time, with possible stabilization of symptoms and improvement in quality of life (Cahill et al., 2021). It is also fundamental to distinguish contexts of use (supervised medicinal vs. recreational), patient profile, and pharmaceutical form, as these elements may explain discrepancies between clinical findings and studies based on recreational patterns (Bialas et al., 2022; Busse et al., 2021).

Two biases must be explicitly considered to interpret the association observed in this sample between prolonged use and better outcomes. First, since this is a cross-sectional study, there is a risk of reverse causality: patients with better sleep/quality of life may be more likely to maintain treatment for longer periods, while those with poorer response tend to discontinue use earlier (AminiLari et al., 2022; Suraev et al., 2020; Skevington et al., 2004; Bialas et al., 2022). Second, the presence of survivorship/adherence bias is plausible, as the “>3 years” group tends to consist of individuals who tolerated the treatment better, perceived benefit, or had better access to follow-up, which may inflate estimates of perceived effectiveness when compared to short-duration groups (Bialas et al., 2022). Therefore, the finding should not be interpreted as proof of superiority of prolonged use per se, but as a signal of perceived effectiveness under specific clinical conditions and with possible selection over time (AminiLari et al., 2022; Suraev et al., 2020; Skevington et al., 2004; Bialas et al., 2022; Davis et al., 2024). Longitudinal evidence suggests a complex and bidirectional interplay between Cannabis use and insomnia, wherein Cannabis may offer short-term perceived relief while simultaneously contributing to worsening sleep outcomes over time, particularly with increased frequency of use (Davis et al., 2024).

An additional interpretive component when considering the present findings is the possibility of expectation effects in therapies with strong symbolic burden, high patient engagement, and a context of stigma. Evidence on patient experience with medical Cannabis suggests that therapeutic expectation, trust in care, and meaning attributed to treatment modulate self-reported outcomes, especially quality of life and sleep, which are highly sensitive to context and perception (Sat et al., 2015; Busse et al., 2021; National Academies of Sciences et al., 2017; Kuhat et al., 2022). Notably, Cannabis use for insomnia management has been reported across populations with depression, anxiety, and comorbid conditions, with perceived benefits varying by product form and strain category (Kuhat et al., 2022). Thus, sustained psychosocial effects, combined with clinical follow-up and individualized adjustment, may contribute to perceived improvements even when the isolated pharmacological effect is difficult to separate in observational designs (AminiLari et al., 2022; Gates et al., 2014; Bialas et al., 2022).

In the present study, the finding that absence of adverse effects was associated with a lower probability of high sleep satisfaction is counterintuitive, but plausible under interpretive models centered on effect perception. Reviews indicate that mild adverse events (e.g., drowsiness, xerostomia, dizziness) are relatively frequent and vary according to formulation and dose (Bort et al., 2024; AminiLari et al., 2022; Babson et al., 2017; Amaral et al., 2023; Suraev et al., 2020; Gates et al., 2014; Bialas et al., 2022). In interventions aimed at sleep, mild drowsiness may be perceived as a sign of useful pharmacological effect, while total absence of effects may reflect a subtherapeutic dose, tolerance, underreporting, or lower individual responsiveness, especially in extracts with variable usage (Gates et al., 2014; Bialas et al., 2022; National Academies of Sciences et al., 2017). Therefore, this result should be discussed as a signal of interaction between tolerability, expectation, and perception of benefit, and not as evidence that adverse events are desirable (Bialas et al., 2022; Busse et al., 2021; National Academies of Sciences et al., 2017).

Regarding the associations observed in the present study between racial self-declaration and outcomes, it is recommended to avoid biologizing interpretations and to privilege contextual mechanisms that influence self-reported outcomes. Studies on medical Cannabis indicate that stigma, previous experiences with the healthcare system, access barriers, and social insecurity may modulate the experience of treatment and the way benefits are reported (Sat et al., 2015; Busse et al., 2021; National Academies of Sciences et al., 2017). Since quality of life and sleep are outcomes highly sensitive to psychosocial context, differences between groups may reflect variations in therapeutic experience, trust in care, social support, and degree of perceived stigmatization, rather than intrinsic differences in biological response (Skevington et al., 2004; Devlin and Brooks, 2017; Sat et al., 2015; Busse et al., 2021; Arkell et al., 2020). This point reinforces the importance of future studies incorporating socioeconomic variables and stigma/expectation measures to elucidate underlying mechanisms (Sat et al., 2015; Busse et al., 2021; National Academies of Sciences et al., 2017).

In the present study, the absence of a significant association between gender and the outcomes is consistent with part of the literature, which describes greater consistency in differences in use patterns than in perceived efficacy, suggesting heterogeneity according to clinical indication, dose, and sociocultural context (Arkell et al., 2020; Cuttler et al., 2016). Studies also indicate that sex/gender differences may emerge in specific subdomains (e.g., anxiety, consumption patterns, and perceived effects), justifying stratified analyses in larger samples with greater statistical power (Cuttler et al., 2016). Additionally, in diverse clinical populations, gender effects may be mediated by diagnosis, comorbidities, concomitant treatments, and adherence patterns, which is frequently pointed out as a source of heterogeneity in broad reviews (Suraev et al., 2020; Bialas et al., 2022).

Finally, it is important to acknowledge limitations while simultaneously highlighting the contributions of the study. The cross-sectional design, the absence of baseline measures, and the potential selection/response bias limit causal inference and estimation of effect magnitude, and such limitations are widely acknowledged in evidence syntheses on medical Cannabis (AminiLari et al., 2022; Suraev et al., 2020; Bialas et al., 2022). Additionally, the ordinal logistic regression models should be interpreted as exploratory, as the sample size (n = 71) and the number of predictors with sparse categories produced parameter instability in specific subcategories, including extreme regression coefficients and non-estimable standard errors consistent with quasi-complete separation. These results are not suitable for direct clinical inference and should be confirmed in larger samples.

The low response rate (71/689; 10.3%) warrants careful consideration, though it is consistent with rates reported in email-based health surveys in real-world clinical populations (Busse et al., 2021; Davis et al., 2024). Online self-administered questionnaires are known to yield substantially lower response rates than paper-based or telephone-administered instruments (Tumyan et al., 2025; Romero et al., 2025), and response rates in this range have been documented even in well-designed randomized studies of survey delivery methods in healthcare settings (Tumyan et al., 2025). In the specific context of medical Cannabis, non-response may be further compounded by stigma-related concerns, as a substantial proportion of medical Cannabis users conceal their use from healthcare providers due to fear of disapproval or legal concerns (Kongsved et al., 2007). Nonetheless, selection bias remains a concern: respondents may systematically differ from non-respondents in terms of treatment adherence, satisfaction with outcomes, and capacity to complete a lengthy self-administered instrument combining two validated multidomain instruments, potentially overestimating perceived effectiveness and underrepresenting patients with poor outcomes or treatment discontinuation. Furthermore, data were collected via a self-administered online questionnaire, which carries inherent response biases: interpretation bias may arise from individual variability in how participants understand questionnaire items; memory bias is inherent to retrospective self-report instruments such as the PSQI; expectation bias is particularly relevant in the context of medical Cannabis, where strong prior beliefs about therapeutic benefits may inflate self-reported outcomes; and social desirability bias may lead participants to report more favorable outcomes than actually experienced. These limitations are particularly pertinent given that the primary outcomes (quality of life and sleep satisfaction) are highly subjective and context-sensitive.

The need for variable grouping may reduce granularity, but it is a defensible strategy for statistical stability in moderate and heterogeneous samples. The clinical heterogeneity of the sample, encompassing adults, children, adolescents, and patients with psychiatric, neurological, pain-related, and other conditions analyzed jointly, represents an important contextual consideration. The comparative analyses conducted were explicitly exploratory in nature, designed to identify preliminary signals and generate hypotheses rather than to establish definitive subgroup-specific conclusions. Given the sample size and the distribution of responses across clinical subgroups, drawing direct or conclusive comparisons between groups was neither statistically appropriate nor the primary objective of the study. This approach is consistent with the broader aim of characterizing real-world patterns of Cannabis oil use in a heterogeneous clinical population, reflecting the complexity of actual clinical practice. Subgroup-specific confirmatory analyses would require substantially larger and more homogeneous samples, and are strongly recommended as a priority for future longitudinal research. On the other hand, clinical heterogeneity may be interpreted as an external validity strength, as it mirrors the reality of services that follow patients with multiple indications and individualized therapeutic regimens, an aspect frequently absent in highly selected randomized clinical trials (AminiLari et al., 2022; Gates et al., 2014; Aran et al., 2021; Bialas et al., 2022). Thus, the findings contribute by describing the performance of a cannabinoid intervention with a focus on PROMs, offering hypotheses for future longitudinal investigations with baseline assessment, follow-up, and better control of confounders, in addition to more detailed product characterization, including phytochemical profile.

In summary, the results suggest that, in supervised clinical practice, patients using full-spectrum THC and CBD rich oil reported favorable perceived quality of life and sleep satisfaction. However, given the cross-sectional design and the absence of pre-treatment assessment, these findings reflect self-reported perceptions rather than evidence of treatment efficacy, and causal interpretation is not warranted. Future longitudinal studies, with pre-treatment assessment, systematic follow-up, and more detailed product characterization, are needed to clarify mechanisms, responding subgroups, and the risk-benefit balance of prolonged use, especially in products with higher THC content.