Psychiatric outpatient care is designed to offer comprehensive, continuous treatment and support for individuals to cope with mental health conditions to enabling them to sustain their everyday lives within their community (1). A primary goal of the provided care is to improve patients’ quality of life (QoL) (2–6), which encompasses their perceptions of life circumstances within their cultural context, considering their aspirations, expectations, and concerns (7). However, beyond psychiatric clinical symptoms (3, 8–11), various sociodemographic, social, occupational, and financial factors may contribute to patients’ QoL. For instance, studies have indicated that older age (12, 13), female gender (12, 14), relationship problems (11, 12), lower education levels (10, 13, 15), and unemployment (11, 13, 16–18) are associated with lower QoL in patients with mental illness.

However, the multitude of variables linked to QoL makes the research in this area fragmented and challenging to comprehend (19). While attempts have been made to propose models demonstrating the relationships between various variables and QoL (20–22), explicit models that effectively guide the understanding of how mental health service provision impacts QoL are still lacking. The relationship between mental health service provision and patients’ perceived QoL is intricate and may not be adequately explained solely through simple correlations. Employing more complex modelling techniques, such as mediation analysis (23, 24), can help elucidate the complex relationships, where an intermediate (mediating) variable or factor may clarify the connection between the quality of mental health services and patients’ perceived QoL.

Upon reviewing existing literature, it becomes evident that while numerous factors influence QoL, only a selected few have been thoroughly investigated as potential mediators impacting QoL. For instance, studies have delved into variables such as recovery (21, 25) and symptom severity (22, 25–27) but have yet to explore many others in depth. Within the scope of recovery, notable components encompass empowerment, agency, and hope (21, 25, 28). This emphasis on recovery has spurred a growing body of research over the past three decades, prompting mental health systems globally to advocate for recovery-oriented services. These services encourage active involvement and choice for individuals seeking treatment and support (29).

While the concept of recovery from mental illness traditionally involved symptom absence and restoration to a pre-illness state, the contemporary viewpoint extends beyond this notion. Modern perspectives emphasize personal growth and development, surpassing the adverse effects of mental illness to establish a fulfilling and meaningful life (30). Therefore, recovery encompasses more than mere symptom remission or a return to prior functioning levels. Interestingly, a meta-analysis of 20 articles highlights that individuals with schizophrenia can experience personal recovery despite persisting symptoms of psychosis (31). Consequently, recovery and symptom severity may function as partially independent markers of mental health status.

Building upon the findings of prior research, a conceptual model (Figure 1) has been proposed. According to this model, variables within mental health services can directly influence QoL or operate via mental health status variables that directly impact QoL. Additionally, sociodemographic variables directly influence QoL alongside patients’ clinical diagnoses.

Twenty significant relationships were found after the iterative testing procedure (Table 2), and we reached a final mediation model (Table 3) that received excellent goodness of fit for all evaluated goodness-of-fit indices. The non-significant chi-square shows that the model did not deviate significantly from observed data. The R² indicates that the model explained 55% of the variance in quality of life, which was deemed adequate. Adjustment of the model with sociodemographic and diagnosis variables had a small and non-significant effect on the model fit (Δχ² = 17.04 with a Δdf = 21 giving a p = 0.71, a ΔRMSEA = 0.011, and a ΔCFI = 0.001), demonstrating that the associations between the variables in mental health service provision and QoL were not affected by the observed socioeconomic or diagnosis variables.

The final model indicated that the relationship between the variables in mental health service provision and the patient’s perceived QoL is relatively complex. As shown in Figure 2, QoL was directly associated with six variables, whereas only one, participation information, was a service provision variable. Most service provision variables were mediated by the mental health status variables (symptom severity and recovery) mediating QoL. Among the quality-of-care variables, all but one were mediated by the discharge variable. Moreover, the patient-staff relationship, measured by the VSI, had an indirect association with QoL and a significant direct association with recovery. One treatment variable, psychotherapy, had a significant association with QoL, i.e., a mediated association via symptom severity and recovery, resulting in a total negative association with QoL.

Although sociodemographic and diagnosis variables had no significant effects on the model fit, some relationships with QoL were significant. Living with partner had a significant and direct positive association with QoL. For clinical diagnosis, bipolar disorder, schizophrenia, and anxiety had a significant positive association with QoL. Notably, the relationship between patients’ anxiety and QoL was somewhat complex. Patients with anxiety were more likely than those with other diagnoses to receive psychotherapy. Those receiving psychotherapy (whether or not with anxiety) perceived slightly more severe symptoms and less QoL than those receiving other psychiatric treatments, resulting in a negative association between anxiety and QoL. In contrast, patients diagnosed with schizophrenia and bipolar disorders reported better QoL than patients with other diagnoses.

This study examined the relationship between perceived mental health service provision and QoL among patients in psychiatric outpatient care. The final model achieved excellent goodness of fit, showing that variables in mental health service provision mainly had an indirect effect on patients’ perceived QoL and that this effect was mediated by symptom severity and recovery.

This result is consistent with research indicating that service provision variables have a minor direct association with perceived QoL, suggesting that other factors beyond mental health care service are more impactful (5, 31, 40). Moreover, the results correspond to studies using complex mediator models. For instance, Fleury and colleagues (21) found that sociodemographic variables (gender and living) and clinical diagnosis variables (mood disorders and substance use disorders) were directly related to patients’ perceived QoL but that only two of the service provision variables (service continuity and adequacy of help) were directly associated with QoL. Similarly, the present study shows that recovery serves as an important mediator but that symptom severity, which was not assessed by Fleury and colleagues (21), also acts as a mediator and has a similar magnitude of association with QoL, suggesting that both recovery and symptom severity have a vital role for patients’ perceived QoL. However, in this study the association of symptom severity with QoL is primarily mediated by recovery, which correspond with findings by Saavedra and colleagues (24), suggesting that the symptom severity and QoL association may depend on the presence of recovery, which aligns with previous research (e.g., (30)). The present results thus suggest that symptom relief is unnecessary for recovery, but in relation to QoL, the absence of symptom relief will reduce recovery, resulting in a negative association with QoL. Hence, the model indicates that recovery would have less impact on patients’ perceived QoL if there was no symptom relief.

Surprisingly, the discharge dimension was a mediator between most of the other service provision variables and recovery. It is not apparent why discharge has this role. However, visits to outpatient clinics are often intermittent and a regular discharge procedure does not always occur. Thus, the QPC-OP discharge dimension items are mainly about closure and the future. These aspects can be related to findings by Fleury and colleagues (21), who showed that recovery, which included elements found in patients’ experience at discharge, was a mediating factor between service continuity and QoL. Discharge may thus be a significant event and a crucial component of the treatment process. As such, discharge is a creative process that can be customised to meet each patient’s unique needs and comprehensively address them across multiple health systems in a continuous and coordinated manner (41). As mentioned above, the discharge process in outpatient care is less formal, and in this study, it was measured by questions including aspects of closure and the future. These questions encompass parts of the personal agency, an essential factor in understanding the ‘service provision-recovery-quality of life’ relationship (22).

From a clinical perspective, this finding raises questions about whether interventions in outpatient psychiatric care need to focus more on agency rather than providing more support and that it is important to include questions about closure and the future from the start of the care process. Prospects encouraging hope can be positive for perceived QoL (42). The quality-of-service provision may also increase patients’ capacity, allowing them to influence their future (43), potentially affecting their QoL. Our model aligns with the assumption that the experience of controlling and being able to master situations is an important recovery factor for developing a higher level of QoL. In this regard, our findings correspond with a holistic perspective on mental health service providers aiming to support recovery by focusing on both symptom reduction and increased well-being related to hope, self-esteem, social connectedness, and a sense of control of one’s life (35–47). However, the current outcome on the role of the discharge dimension should be taken with caution until it has been replicated in independent studies. Further research is needed to clarify the link between the quality of the service provided and the patients’ view of their future.

In addition to the discharge issues, the present results show that the patient-staff relationship, as measured by the VSI, may have an independent effect on recovery. Among the three elements of the patient-staff relationship measured by the VSI-OP, only Inviting the patient to establish a relationship had an association with QoL. This association was mediated entirely by recovery. In other words, the greater the patient-staff relationship, the greater the recovery and thus better QoL. A good patient-staff relationship has been noted as an important factor that can make a difference in recovery during an episode of mental illness, including involuntary admission (48). Using the VSI-OP, Rask and colleagues (35) showed that patients reported that Inviting the patient to establish a relationship, as well as, Showing interest in patients’ feelings, experiences, and behaviour were rated as the most frequent actions performed by staff. In another study, using VSI for supported housing, the residents rated inviting the patient to establish a relationship as the most frequently performed and the most important facet of the patient-staff interaction (49). These findings confirm Green et al. (50), who reported that patients value a staff perceived as competent, caring, trustworthy, and trusting and that these factors were important for patient recovery. This position is in line with our findings that a positive patient-staff relationship, including staff showing interest in the patients’ thoughts and experiences, is vital for patient recovery.

Concerning clinical diagnoses, schizophrenia and bipolar disorders were directly associated with QoL. The associations were positive, indicating that patients with schizophrenia and bipolar disorder perceived QoL better than patients with other mental diagnoses. In addition, schizophrenia was associated with less symptom severity, leading to a positive indirect association with QoL. This observation aligns with previous findings showing that patients with schizophrenia report higher self-reported QoL than patients with other mental disorders (51). However, our finding differed from the finding of Priebe et al. (51) that schizophrenia is associated with more severe reported symptoms.

Anxiety was not directly associated with QoL. Instead, the association between anxiety and QoL was mediated by psychotherapy and reported symptom severity, indicating that the relationship between anxiety and QoL is complex. Specifically, patients with anxiety reported receiving more psychotherapy compared to patients with other mental diagnoses and receiving psychotherapy was associated with reporting greater symptom severity, suggesting that patients with anxiety in ongoing psychotherapy also struggle with severe symptoms. This circumstance may thus explain why psychotherapy was associated with greater reported symptom severity.

Concerning the sociodemographic variables only one, living with a partner, was positively associated with QoL. Although associations between sociodemographic variables and QoL seem to depend on the study population, similar results have been demonstrated in studies on patients with schizophrenia (52).

The present study shows that patients’ perception of mental health service provision is positively associated with their perceived QoL; however, this association is mostly indirect and mediated by reduced symptom severity and increased recovery. This finding can help design future interventions to enhance service provision and thus promote patients’ QoL. Further studies are needed to capture a causal path to QoL.

The datasets presented in this article are not readily available because of ethical restrictions. Requests to access the datasets should be directed to the corresponding author.

The studies involving humans were approved by the Swedish Ethical Review Authority. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because the participants gave verbal informed consent to participate.

L-OL: Conceptualization, Formal analysis, Methodology, Writing – original draft, Writing – review & editing. PR: Conceptualization, Writing – review & editing. MR: Conceptualization, Writing – review & editing. DB: Conceptualization, Writing – review & editing. TS: Conceptualization, Writing – review & editing. KG: Conceptualization, Writing – review & editing. IR: Conceptualization, Writing – review & editing. AS: Conceptualization, Funding acquisition, Writing – review & editing.