3.1.1 Characteristics of NMOSD patients
A total of 139 patients were investigated, with 128 patients completing the questionnaire. Of these, 87.5% (112) of the patients were female, the age of the patients was 36.7 ± 11.4 years old, the median course of disease was 4 years, the median number of relapses was 2, and 54% (69) of the patients had a college degree or above. The PSS was 41.14 ± 3.83, and HRHS was 100.68 ± 6.59.
3.1.2 Correlation analysis of health-related hardiness, disability, and perceived stress
Results of a correlational analysis are showed in Table 1. EDSS was negatively correlated with PSS (r = − 0.762, P < 0.01) and moderately negatively correlated with HRHS (r = 0.513, P < 0.01), while HRHS was moderately negatively correlated with PSS (r = − 0.505, P < 0.01).
Table 1
Correlation analysis of HRHS, EDSS, and PSS
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
1 Duration of disease | 1 | | | | | | |
2 Number of relapses | 0.635** | 1 | | | | | |
3 Age | 0.085 | −0.037 | 1 | | | | |
4 Education | 0.132 | 0.129 | −0.458** | 1 | | | |
5 EDSS | 0.098 | 0.037 | 0.146 | −0.002 | 1 | | |
6 PSS | 0.176* | 0.12 | 0.101 | −0.01 | 0.762** | 1 | |
7 HRHS | 0.096 | 0.092 | −0.058 | −0.013 | −0.513** | −0.505** | 1 |
PSS: Perceived stress scale, HRHS: health-related hardiness scale, EDSS: Expended disability status scale |
**: P < 0.01 |
*: P < 0.05 |
Insert Table 1. Correlation analysis of HRHS, EDSS, and PSS
3.1.3 Examination of the mediating effect of health-related hardiness between disability and perceived stress
In this study, according to the mediating effect calculation method (24), first, hierarchical regression analysis was performed, with the HRHS score as the dependent variable, and disease course, relapse frequency, age, and education level as control variables to establish model 1. Then, the EDSS score was put into the equation to develop model 2 (Table 2). The results showed that model 1 was not significant (P > 0.05), model 2 was statistically significant after adding EDSS score to model 1 (adjusted R2 = 0.258, ΔF = 9.827, P < 0.001). The result indicated that EDSS had a significant impact on HRHS, and it had a negative predictive effect (β = −0.526, P < 0.001)
Table 2
Hierarchical regression analysis with HRHS as dependent variable
| Model 1 | Model 2 |
| Standardized Coefficients β | Standardized Coefficients β |
Duration of disease | 0.087 | 0.134 |
Number of relapses | 0.043 | 0.032 |
Age | −0.098 | −0.010 |
Education | −0.075 | −0.040 |
EDSS | | −0.526** |
Adjust R2 | −0.013 | 0.258 |
ΔF | 0.595 | 9.827** |
PSS: Perceived stress scale, HRHS: health-related hardiness scale, EDSS: Expended disability status scale **: P < 0,01 |
Insert Table 2. Hierarchical regression analysis with HRHS score as dependent variable
Second, hierarchical regression analysis was conducted with PSS as the dependent variable. Model 1 was established with the course of the disease, number of relapses, age, and education level as control variables. Then, EDSS score was put into the equation to establish model 2, and the HRHS was put into the equation to establish Model 3(Table 3). The results showed that model 1 was not significant (P > 0.05), model 2 was statistically significant after adding EDSS scores to model 1 (adjusted R2 = 0.577, ΔF = 35.698, P < 0.001). The result indicated that EDSS has a significant effect on PSS and is positively predictive (β = 0.757, P < 0.001). Model 3 had statistically significant after adding the HRHS score to model 2 (adjusted R2 = 0.599, ΔF = 32.594, P < 0.001). It showed that HRHS was negatively predictive of PSS (β = − 0.183, P < 0.001).
Table 3
Hierarchical regression analysis with PSS as dependent variable
| Model 1 | Model 2 | Model 3 |
| Standardized Coefficients β | Standardized Coefficients β | Standardized Coefficients β |
Duration of disease | 0.150 | 0.083 | 0.108 |
Number of relapses | 0.027 | 0.043 | 0.048 |
Age | 0.094 | −0.033 | −0.035 |
Education | 0.009 | −0.041 | −0.048 |
EDSS | | 0.757** | 0.661** |
HRHS | | | −0.183** |
Adjust R2 | 0.008 | 0.577 | 0.599 |
ΔF | 1.250 | 35.698** | 32.594** |
PSS: Perceived stress scale, HRHS: health-related hardiness scale, EDSS: Expended disability status scale |
**: P < 0,01 |
Insert Table 3. Hierarchical regression analysis with PSS as dependent variable
The regression coefficient was used to test the effect of the mediator variable, and three equations were established (24) (Table 4). Y = cX + e1, M = aX + e2, Y = c’X + bM + e3. Substituting the corresponding regression coefficients into the equation provided the following: 1) setting EDSS as the independent variable and PSS as the dependent variable, the standardized regression coefficient was c = 0.783 (P < 0001); 2) setting EDSS as the independent variable and HRHS as the dependent variable, the standardized regression coefficient was α = −0.526 (P < 0.001); and 3) setting HRHS as the independent variable and PSS as the dependent variable, the standardized regression coefficient was b = − 0.183 (P < 0.001). After adding HRHS (M), The regression coefficient of the direct effect of (EDSS) on the PSS was c' = 0.661, and the effect coefficient was statistically significant in sequential tests (P < 0.001). The calculation method of the mediation effect value is ab/c, and it is calculated (− 0.526) × (− 0.183) / 0.757 × 100% = 12.71%; therefore, the influence of HRHS between EDSS and PSS accounted for 12.7% of the total effect. Health-related hardiness played a partial mediating effect between the disability and perceived stress in NMOSD patients (Fig. 2).
Table 4
The mediating effect test of HRHS between EDSS and PSS
| Dependent variable | Independent variable | Regression equation | Regression coefficient | t | P |
1 | PSS (Y) | EDSS (X) | Y = 0.757X + e1 | c 0.757 | 12.913 | < 0.001 |
2 | HRHS (M) | EDSS (X) | M = − 0.526X + e2 | a -0.526 | −6.774 | < 0.001 |
3 | PSS (Y) | HRHS (M) | Y = 0.661X + − 0.182M + e3 | b -0.183 | −2,743 | 0.007 |
| | EDSS (X) | | c' 0.661 | 9.864 | < 0.001 |
PSS: Perceived stress scale, HRHS: health-related hardiness scale, EDSS: Expended disability status scale |
Insert Table 4. The mediating effect test of HRHS between EDSS and PSS
Insert Fig. 2. Mediating effects of HRHS between EDSS and PSS