Factor analysis of the relationship between PANSS score and family burden of patients with schizophrenia

Abstract Objective This study aims to investigate the burden of family caregivers of patients with schizophrenia, and its influencing factors Methods A total of 105 patients with schizophrenia and their caregivers were investigated using the positive and negative symptom scale (PANSS) and family burden scale of disease (FBS) Results There was a strong correlation between the patient's recovery and family burden, especially between positive and negative symptoms and family financial burden, family daily activities, family recreational activities, and family relationship Conclusion There is a strong correlation between the patient's recovery and family burden, and this is especially correlated to family economic burden, family daily activities, family recreational activities, and family relationship. Medical staff should pay attention to the psychological characteristics of patients and fully understand and avoid the adverse effects of family burden on the rehabilitation of patients.

consists of four sub-scales: positive scale, negative scale, general mental scale, and compound scale (Li et al., 2021). This scale is divided into seven grades of 1-7, in which 1 refers to asymptomatic and 7 refers to severe symptoms (Kay et al., 1987).
The family burden scale of disease (FBS) is a general questionnaire to comprehensively and systematically understand the family burden of patients (Hirsch et al., 2016). There are 24 items in six dimensions in the FBS, which include family economic burden (six items), family daily activities (fix items), family entertainment activities (four items), family relationship (five items), physical health of family members (two items), and mental health of family members (two items) (Frias et al., 2020).
The scale adopts three grades of 0-2, in which no burden has a rating of 0 and severe burden has a rating of 2 (Frias et al., 2020). Since the number of items in each dimension is different, and in order to standardize the score, the score of each dimension was divided by the number of items to obtain the standardized average score and total average score . Then, the average score was divided into three grades, with a score of 0 indicating no burden and a score of 1 or 2 indicating a positive result (Stansfeld et al., 2017).
The hypothesis of this study is that there is a correlation between the patient's recovery and family burden. In order to prove it, the PANSS and FBS were used to survey 105 schizophrenic patients and their caregivers, and the factor analysis method of the multivariate statistics was used to construct a model for comprehensive evaluation.
Finally, the burdens of the disease families were classified and ranked in order to determine the main problems of the burden of patient families.

Inclusion criteria for subjects
The subjects were schizophrenic patients in the rehabilitation period.

Statistical data processing
The data were statistically analyzed using the statistical software SPSS 17.0 and using t-test or factor analysis.

General situations of the PANSS and FBS scores
The absolute value of the correlation coefficient between the total PANSS and FBS scores was 0.604 (p < .01). This indicates that these two are moderately correlated.
In Table 1, the results in the correlation coefficient table reveal that the absolute value of the correlation coefficient between the total PANSS and FBS scores was 0.619 (p < .01). This indicates that these two are strongly correlated. The absolute value of the correlation coefficient between the total scores of the PANSS and family daily activities was 0.586 (p < .01). This indicates that these two are moderately correlated. The absolute value of the correlation coefficient between the total scores of the PANSS and family entertainment activities was 0.566 (p < .01). This indicates that these two are moderately correlated. The absolute value of the correlation coefficient between the total scores of the PANSS and family relationship was 0.458 (p < .01).
This indicates that these two are moderately correlated. The absolute value of the correlation coefficient between the total scores of the PANSS and the physical health of family members was 0.149 (p > .05).
This indicates that these two are weakly correlated. The absolute value of the correlation coefficient between the total scores of the PANSS and the mental health of family members was 0.227 (p < .05). This indicates that these two are weakly correlated.
The above analysis reveals that the total PANSS and FBS scores were moderately positively correlated. This indicates that when the number of positive and negative symptoms increases, the burden of the patient's family would also increase.
In Table 2, the results of the questionnaire survey revealed that the family burden of the patient's family members are as follows:

Factor analysis of the FBS
Six indexes were selected for the factor analysis, and the specific indexes were as follows: economic burden X1, family daily activities X2, family entertainment activities X3, family relationship X4, family member physical health X5, and family member mental health X6.
From the correlation coefficient matrix above, it is concluded that most of the variables were greater than 0.3. This indicates that the selected data is suitable for the factor analysis.
The results exported by the SPSS software are shown in Table 3

Extraction of common factors and determination of the cumulative variance contribution rate
The common factors were extracted by principal component analysis.
In general, the extracted common factor eigenvalue is greater than 1.
However, sometimes, in order to explain the index variables more accurately, factors with eigenvalues of less than 1 can also be extracted. In the present study, some common factors with eigenvalues greater than 0.8 were extracted. In order to verify the degree of information loss of the original variables, the commonality of variables was calculated: That is, after being extracted from the two common factors, the commonalities of variables were almost above 80% (the detailed results are presented in Table 1). This indicates that the extracted common factors already contain most of the information of the original variables, and lost information was relatively less. Therefore, the effect of the factor analysis was relatively ideal.

F I G U R E 1 Gravel map
Eigenvalues can also be considered as variance contributions: That is, The cumulative variance contribution rate of the three common factors extracted was 63.718% + 16.285% = 80.003%. This indicates that it is feasible to extract two common factors. On the basis of the gravel map in Figure 1, it can be observed that through the third factor, the eigenvalue began to stabilize. Therefore, it would be most appropriate to extract two common factors.

Establishment of the factor model
In order to make the meaning of the common factor easier to explain, factor rotation was required to be conducted. In the present study, the initial factor load matrix was rotated using the orthogonal rotation method, in which the initial load matrix was multiplied to the orthogonal matrix on the right. The initial factor load matrix was rotated using the SPSS software. According to the factor score coefficient matrix, the expression of the factor score can be obtained as follows: F j = j1 X 1 + j2 X 2 + ⋯ + jp X p ,Xj = 1, 2, … , m.
Finally, the comprehensive score was calculated, with the proportions of the variance contribution rate of each factor in the cumulative contribution rate taken as the weight for the weighted summary. Afterwards, the comprehensive score of each area was calculated. That is, Then, the scores of each factor and comprehensive scores were ranked. The results are presented in Table 4.

DISCUSSION
The present study result confirmed our hypothesis that the rehabili- For instance, the 7th patient ranked first on the common factor F1, 86th on F2, and 2nd on the comprehensive score F. This indicates that the family of the 7th patient had heavy burdens in terms of economic concern, daily activities, entertainment activities and family relationship. However, they did not have very heavy burdens in terms of physical health and mental health. In general, the family burden of the 7th patient was heavy. The 97th patient ranked 16th on the common factor F1, 103rd on F2 and 20th on the comprehensive score F. This indicates that the family of the 97th patient had heavy burdens in terms of economic concern, daily activities, entertainment activities, and family relationship. However, they had relatively light burdens in terms of physical health and mental health. In general, the family burden of the 97th patient was relatively heavier. The 98th patient ranked 103rd on the common factor F1, 38th on F2 and 103rd on the comprehensive score F. This indicates that the family of the 98th patient had light burdens in terms of economic concern, daily activities, entertainment activities, and family relationship. However, they had relatively heavy burdens in terms of physical health and mental health.
In general, the family burden of the 98th patient was not light.
According to the results of the analysis, the problems and difficulties of family burdens of different patients can be determined, and reasonable and feasible policy recommendations can be put forward.
Therefore, it is of great significance to analyze and investigate the fam-   et al., 2014). However our limitation is that the sample size is not large enough, thus further study is needed.

CONCLUSION
There is a strong correlation between rehabilitation and family burden.
In particular, family economic burden is correlated to family daily activities, family entertainment activities, and family relationship. The medical staff should focus their attention to the psychological characteristics of these patients and fully recognize and avoid the adverse effects of family burden during the patient's rehabilitation.

FUNDING INFORMATION
The work was supported by Guizhou Qiandongnan Science and Tech- (NJD)

DATA AVAILABILITY STATEMENT
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.