Literature was retrieved from 11 databases (CNKI, Wan fang Database, VIP Database, CBM Database, PubMed, EMBASE, Cochrane Library, Web of Science, APA PsycINFO, CINAHL, and MEDLINE), with the search period ranging from the inception date to 1 August 2024. In addition, we utilized internet search functions, such as www.google.com. We employed the following MeSH terms and keywords: (schizophrenia or dementia praecox or schizophrenic disorder) and (positive psychology or psychology or positive or positive psychology theory) and (randomized controlled trials) and (China or Chinese or Mainland China). Furthermore, we manually checked the reference lists and related commentaries to identify and retrieve other relevant studies. The search strategies for each database are detailed in Supplementary Appendix S1 .

We followed the PICOS criteria outlined in the “Cochrane Handbook for Systematic Reviews of Interventions” to determine the scope of studies to be included (29). (1) Population: Including Chinese patients aged 18 years and above, diagnosed with schizophrenia, with no gender restrictions. Patients with a history of alcohol or drug dependence or intellectual disabilities are excluded. All participants are from Mainland China, while Hong Kong, Taiwan, and Macau are not included in the scope of this study due to their long-term influence from foreign cultures. (2) Intervention: The experimental group receives treatment based on positive psychology theory, focusing on establishing a positive mindset and cultivating positive psychological qualities in patients. The intervention style can be implemented through face-to-face interviews, relevant lectures, or other means without considering the frequency and duration of the intervention. (3) Comparison: The control group receives no treatment, minimal intervention (such as education, pamphlets), and routine care, without involving positive psychological interventions. (4) Outcome: Wellbeing, mental health, positive symptoms of schizophrenia, negative symptoms of schizophrenia, social function, self-esteem, social adaptability, and cognitive functions. (5) Study design: Inclusion criteria: Only randomized controlled trials (RCTs) were included; patients diagnosed with schizophrenia. Exclusion criteria: Non-RCT research; incorrect population, irrelevant to the results, and failure to report data used for analysis.

In accordance with the research requirements, we have established a standardized data extraction scheme. Two reviewers (YH and YJH) independently extracted and entered various data in a double-blind manner, and a third reviewer (JHJ) conducted cross-checking. Any discrepancies encountered during this step were resolved through discussion with another reviewer (QHL) to make the final decision. The extracted data include the following: the first author of the literature, publication year, sample size of the experimental and control groups, participants’ age, content of the intervention, intervention plan (duration, frequency, and cycle), outcome indicators, and other relevant data. We determine the effect size by assessing the differences in means between two independent samples. In cases where the standard deviation (SD) is not provided, we estimate it using the standard error (SE), confidence interval (CI), t-value, or p-value. We make efforts to procure missing data by reaching out to the authors via email (30). Should the authors fail to include the data in the study but offer charts that contain the relevant information, we utilize GetData Digitizer version 2.20 software to extract the required data from these charts. When both unadjusted and adjusted data are presented in a paper, we prioritize the use of adjusted data for our research analysis.

According to the guidelines of evidence-based medicine, the risk of bias assessment tool from the Cochrane Systematic Reviews is used (31), where two reviewers (YH and YJH) evaluate the quality of included studies based on seven criteria: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. Any discrepancies are resolved through discussion with a third reviewer (JHJ). In the statistical process, the quality assessment is categorized as follows: studies with six or more criteria are considered to have a low risk of bias; those with three to four criteria have a moderate risk of bias; and those with fewer than three criteria have a high risk of bias.

The assessment of the certainty of evidence is conducted by two reviewers (YH and YJH) based on the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach, which evaluates the risk of bias, consistency, indirectness, imprecision, and publication bias. The certainty of the evidence is categorized into four levels (high, moderate, low, and very low) (32, 33).

All analyses were conducted using Review Manager (version 5.3.5; The Cochrane Collaboration, Copenhagen, Denmark) and Stata software (version 14; Stata Corp, Texas, USA). Researchers combined data by summing the mean values and calculating the standard deviations to derive the post-intervention outcomes. For continuous data, we calculated the mean difference (MD) and standardized mean difference (SMD) along with their corresponding 95% CIs, depending on whether the outcomes were measured by the same tool. Researchers will perform statistical tests for heterogeneity (Cochran’s Q and I ²), where the I ² statistic is used to assess the degree of heterogeneity. If I² ≤ 50% and P > 0.1, a fixed-effects model will be used to pool the data; conversely, if I² > 50% and p < 0.1, a random-effects model will be used (34). Sensitivity analysis by excluding trials one by one and subgroup analysis will be conducted to explore the sources of heterogeneity. We have defined short-term treatment (2–4 weeks), medium-term treatment (6–8 weeks), and long-term treatment (10–12 weeks) groups to explore the effect of intervention duration on outcomes. Finally, we will use the Egger test to assess publication bias and draw a funnel plot using Stata 14.0 software (35).

We conducted a comprehensive search of the database and identified 9602 studies that potentially met the criteria. After reviewing for duplicate articles, 8664 articles remained, and then 2354 articles were excluded after carefully reading the titles and abstracts. Based on the inclusion and exclusion criteria, we selected 54 articles for this systematic review and meta-analysis by reading the full text ( Figure 1 ).

This meta-analysis includes 54 journal articles published from 2013 to 2024. These studies involve a total of 4,108 research subjects. The average age of the subjects in the experimental group is 40.7 ± 5.78 years, while that of the subjects in the control group is 40.48 ± 5.74 years. The intervention duration ranges from 2 to 24 weeks. All 54 RCTs compared positive psychology with TAU (such as routine care and education). Details are shown in Table 1.

A total of 21 articles used the Health Education Impact Questionnaire (HEIQ) to assess wellbeing (36–56), 24 articles used the Self-rated Health Measurement Scale (SRHMS) to assess mental health (37, 39, 40, 42, 43, 46, 49, 51–67), 10 articles used the Positive and Negative Syndrome Scale (PANSS) to assess positive and negative symptoms of schizophrenia (42, 62, 68–75), 16 articles used the Social Disability Screening Schedule (SDSS) scale to assess social function (38, 42, 43, 45, 47, 48, 62, 67, 76–83), 7 articles used the Self-Esteem Scale (SES) scale to assess self-esteem (38, 48, 67, 77, 80, 82, 84), 4 articles used the Social Adaptability Scale (SAFE) scale to assess social adaptability (79, 81, 85, 86), and 4 articles used the Montreal Cognitive Assessment (Mo-CA) scale to assess cognitive function (86–89).

The quality of the included literature was assessed; 7 articles were classified as having a low risk of bias, 1 article scored 7 points, and the remaining 47 articles were all categorized as having a moderate risk of bias ( Figure 2 ). A graph of the proportion of bias risk in the included studies is presented ( Figure 3 ). The most significant risks of the included literature are the lack of randomization details and insufficient blinding methods. Specifically, 96.3% of the literature lack randomization details, and 87.1% have insufficient blinding methods.

A total of 23 articles reported on the impact of positive psychology theory on the mental health of patients with schizophrenia, including a total of 2,090 patients, with 1,047 in the experimental group and 1,043 in the control group. The results show that intervention with positive psychology theory can significantly improve the mental health of patients with schizophrenia, and it is markedly superior to the control group (MD = 43.50, 95% CI = 40.11 to 46.89, p < 0.001, I ² = 0%) ( Figure 4 ). Eleven studies reported the impact of short-term positive psychology theory intervention on the mental health of patients with schizophrenia, showing a significant increase in mental health (MD = 30.62, 95% CI = 24.59 to 36.64, p < 0.001). Twelve studies reported the impact of medium-term positive psychology theory intervention on the mental health of patients with schizophrenia, showing a significant increase in mental health (MD = 42.81, 95% CI = 37.59 to 48.03, p < 0.001). Eighteen studies reported the impact of long-term positive psychology theory intervention on the mental health of patients with schizophrenia, showing a significant increase in mental health (MD = 42.48, 95% CI = 37.34 to 47.63, p < 0.001). The differences among the three groups were statistically significant (p = 0.004), indicating that long-term and medium-term interventions with positive psychology theory are significantly better than short-term interventions for the mental health. However, there was no statistical significance between the long-term and medium-term treatment groups (p = 0.93 > 0.05) ( Figure 5 ). According to the GRADE evidence grading assessment, the certainty level of mental health was rated as low quality, with downgrades due to the presence of bias risk and inconsistency. Details are found in Supplementary Material S2 .

Twenty-one articles reported on the impact of positive psychology theory on the wellbeing of patients with schizophrenia, including a total of 1,650 patients, with 827 in the experimental group and 823 in the control group. The results show that the intervention of positive psychology theory can significantly improve the wellbeing of patients with schizophrenia, and it is obviously better than the control group (MD = 0.61, 95% CI = 0.56 to 0.66, p < 0.001, I ² = 0%) ( Figure 6 ). Twelve studies reported the impact of short-term positive psychology theory intervention on the wellbeing of patients with schizophrenia, showing a significant increase in scores (MD = 0.34, 95% CI = 0.27 to 0.41, p < 0.001). Twelve studies reported the impact of medium-term positive psychology theory intervention on the wellbeing of patients with schizophrenia, showing a significant increase in scores (MD = 0.61, 95% CI = 0.53 to 0.68, p < 0.001). Twelve studies reported the impact of long-term positive psychology theory intervention on the wellbeing of patients with schizophrenia, showing a significant increase in scores (MD = 0.63, 95% CI = 0.57 to 0.69, p < 0.001). The differences among the three groups were statistically significant (p < 0.001), indicating that long-term and medium-term interventions with positive psychology theory are significantly better than short-term intervention for the wellbeing of patients with schizophrenia. However, there was no statistical significance between the long-term and medium-term treatment groups (p = 0.78 > 0.05) ( Figure 7 ). According to the GRADE evidence grading assessment, the certainty level of wellbeing was rated as moderate quality, with downgrades due to the presence of bias risk. Details are found in Supplementary Material S2 .

Ten research studies have reported the impact of positive psychology theory on the recovery of positive symptoms in schizophrenia. The results show that positive psychology theory can significantly improve the positive symptoms of schizophrenia and is superior to TAU (SMD = −2.68, 95% CI = −3.53 to −1.84, p < 0.001, I ² = 95%) ( Figure 8 ). The results of sensitivity analysis showed values of SMDs ranging from −2.68 (95% CI = −3.53 to −1.84) to −3.19 (95% CI = −3.70 to −2.98), with I² ranging from 95% to 77% ( Supplementary Material S4 ). The results of the sensitivity analysis were stable. One study reported the impact of short-term positive psychology theory intervention on the positive symptoms of schizophrenia, with results showing no significant effect (SMD = −0.15, 95% CI = −0.62 to 0.32, p > 0.05). Two studies reported the impact of medium-term positive psychology theory intervention on the positive symptoms of schizophrenia, showing a significant improvement in the positive symptoms of schizophrenia (SMD = −2.97, 95% CI = −3.42 to −2.53, p < 0.001). Six studies reported the impact of long-term positive psychology theory intervention on the positive symptoms of schizophrenia, showing a significant improvement in the positive symptoms of schizophrenia (SMD = −2.77, 95% CI = −4.27 to −1.27, p < 0.001). The differences among the three groups were statistically significant (p < 0.001), indicating that long-term and medium-term interventions with positive psychology theory are significantly better than short-term interventions for the positive symptoms of schizophrenia. However, there was no statistical significance between the long-term and medium-term treatment groups (p = 0.80 > 0.05) ( Figure 9 ). According to the GRADE evidence grading assessment, the certainty level of the positive symptom was rated as low quality, with the downgrade due to the presence of bias risk and publication bias. Details are found in Supplementary Material S2 .

Ten research studies have reported the impact of positive psychology theory on the recovery of negative symptoms in schizophrenia. The results show that positive psychology theory can significantly improve the negative symptoms of schizophrenia and is superior to TAU (SMD = −2.63, 95% CI = −3.40 to −1.87, p < 0.001, I ² = 94%) ( Figure 10 ). The results of sensitivity analysis showed values of SMDs ranging from −2.63 (95% CI = −3.40 to −1.87) to −3.06 (95% CI = −3.56 to −2.56), with I² ranging from 94% to 78% ( Supplementary Material S5 ). The results of the sensitivity analysis were stable. One study reported the impact of short-term positive psychology theory intervention on the negative symptoms of schizophrenia, with results showing no significant effect (SMD = −0.15, 95% CI = −0.62 to 0.32). Two studies reported the impact of medium-term positive psychology theory intervention on the negative symptoms of schizophrenia, showing a significant improvement in the negative symptoms of schizophrenia (SMD = −3.64, 95% CI = −4.37 to −2.92, p < 0.001). Five studies reported the impact of long-term positive psychology theory intervention on the negative symptoms of schizophrenia, showing a significant improvement in the negative symptoms of schizophrenia (SMD = −2.63, 95% CI = −4.04 to −1.22, p < 0.003). The differences among the three groups were statistically significant (p < 0.001), indicating that long-term and medium-term interventions with positive psychology theory are significantly better than short-term interventions for the negative symptoms of schizophrenia. However, there was no statistical significance between the long-term and medium-term treatment groups (p = 0.21 > 0.05) ( Figure 11 ). According to the GRADE evidence grading assessment, the certainty level of the negative symptom was rated as low quality, with the downgrade due to the presence of bias risk and publication bias. Details are found in Supplementary Material S2 .

Seven research studies have reported on the improvement of schizophrenia symptoms through positive psychology theory. The results show that positive psychology theory can significantly reduce PANSS-Total, and is superior to TAU (SMD = −1.99, 95% CI = −2.20 to −1.78, p < 0.001, I² = 42%) ( Figure 12 ). According to the GRADE evidence grading assessment, the certainty level of PANSS-Total was rated as low quality, with the downgrade due to the presence of bias and imprecision risks. Details are found in Supplementary Material S2 .

Sixteen research studies have reported on the impact of positive psychology theory on the recovery of social function in patients with schizophrenia. The results indicate that positive psychology theory can significantly improve the social function of patients with schizophrenia and is superior to TAU (MD = −2.68, 95% CI = −3.26 to −2.10, p < 0.001, I ² = 94%) ( Figure 13 ). The results of sensitivity analysis showed values of SMDs ranging from −2.68 (95% CI = −3.26 to −2.10) to −2.63 (95% CI = −3.19 to −2.08), with I² ranging from 94% to 92% ( Supplementary Material S6 ). The results of the sensitivity analysis were stable. Two studies reported the impact of short-term positive psychology theory intervention on social function, showing a significant improvement in the social function of patients with schizophrenia (SMD = −0.95, 95% CI = −1.53 to 0.36, p = 0.002). Six studies reported the impact of medium-term positive psychology theory intervention on social functioning, showing a significant improvement in the social function of patients with schizophrenia (SMD = −3.09, 95% CI = −4.32 to −1.85, p < 0.001). Six studies reported the impact of long-term positive psychology theory intervention on social function, showing a significant improvement in the social function of patients with schizophrenia (SMD = −2.15, 95% CI = −2.80 to −1.51, p < 0.001). The differences among the three groups were statistically significant (p < 0.001), indicating that long-term and medium-term interventions with positive psychology theory are significantly better than short-term intervention for the social function of patients with schizophrenia. However, there was no statistical significance between the long-term and medium-term treatment groups (p = 0.19 > 0.05) ( Figure 14 ). According to the GRADE evidence grading assessment, the certainty level of the social function was rated as low quality, with the downgrade due to the presence of bias risk, inconsistency, and publication bias. Details can be found in Supplementary Material S2 .

Seven research studies have reported on the impact of positive psychology theory on the self-esteem of patients with schizophrenia. The results show that positive psychology theory can significantly improve the self-esteem of patients with schizophrenia and is superior to TAU (MD = 7.98, 95% CI = −7.53 to 8.42, p < 0.001, I² = 0%) ( Figure 15 ). According to the GRADE evidence grading assessment, the certainty level of the self-esteem was rated as low quality, with the downgrade due to the presence of bias and imprecision risks. Details are found in Supplementary Material S2 .

Four research studies have reported on the impact of positive psychology theory on the social adaptability of patients with schizophrenia. The results show that positive psychology theory can significantly improve the social adaptability of patients with schizophrenia and is superior to TAU (MD = −8.72, 95% CI = −9.16 to −8.27, p < 0.001, I² = 0%) ( Figure 16 ). According to the GRADE evidence grading assessment, the certainty level of the social adaptability was rated as low quality, with the downgrade due to the presence of bias and imprecision risks. Details are found in Supplementary Material S2 .

Four research studies have reported on the impact of positive psychology theory on the cognitive function of patients with schizophrenia. The results show that positive psychology theory can significantly improve the cognitive function of patients with schizophrenia and is superior to TAU (MD = 2.38, 95% CI = 1.97 to 2.78, p < 0.001, I ² = 38%) ( Figure 17 ). According to the GRADE evidence grading assessment, the certainty level of the cognitive function was rated as low quality, with the downgrade due to the presence of bias risk and imprecision. Details are found in Supplementary Material S2 .

This study included 54 articles and conducted a publication bias test on various outcome indicators (90). The funnel plots of mental health and wellbeing were symmetrical ( Supplementary Material S10 ), and the Egger’s test was conducted with p > 0.05. The funnel plots of positive symptoms, negative symptoms, and social function showed asymmetry, and the Egger’s test was conducted with p < 0.05 ( Supplementary Material S10 ). The results of the trim-and-fill method were consistent with the original results (35). Details can be found in Supplementary Materials S7 , S8 , and S9 .

In this meta-analysis, we have identified certain limitations in the study design and implementation that may have impacted our conclusions. Firstly, the large time span included in the study, coupled with regional differences, may have led to modifications of the scale. As a result, when using the PANSS for assessment, the heterogeneity of the measurement results is relatively high. In addition, when conducting subgroup analyses on wellbeing, mental health, positive symptoms, negative symptoms, and social functioning, we found no significant differences between medium-term and long-term treatments. This may be due to the different number of studies included in each group and the relatively small sample sizes. We also noted that most of the studies included did not provide detailed descriptions of their randomization methods, and blinding techniques were not employed, which may have affected the objectivity and accuracy of the results. Furthermore, the necessity for personalized treatment to adapt to the patients’ evolving needs and therapeutic responses may have led to inconsistencies in treatment outcomes, thereby affecting the stability of the study. The implementation of long-term treatment also faces challenges such as resource constraints, low patient engagement, and questions of treatment sustainability.

Regarding the issue of publication bias. For the indicators of social functioning as well as positive and negative symptoms, the detected publication bias poses a challenge to the robustness and generalizability of our research results. The asymmetry presented by the funnel plot and the results of Egger’s test strongly suggest that the existing literature may not cover all the studies conducted in this field. It is highly likely that studies with non-significant or negative results have not been published, while those with desirable outcomes are more likely to enter the public eye. This selective dissemination of research findings can lead to a biased overall situation, thereby causing an overestimation or underestimation of the true effect sizes related to improvements in social functioning and changes in symptoms.