Participants Characteristics
We collected 215 valid questionnaires. The study included 106 participants in the relapse group and 109 participants in the non-relapse group. Approximately 25.1% of participants with relapse were women, while 31.6% of patients were men without relapse. Among the patients with relapse, the majority (20.9%) were between the ages– 30-44, 19.1% lived in urban areas, 26.5% were unmarried, and 26.5% were unemployed. In addition, 42.3% of patients had no family history, 28.4% had a disease duration of more than 5 years, and 27.0% had a medication adherence score of more than 6 points. The average dose of relapse was 222.41 (SD=215.93), whereas the mean dose of non-relapse was 323.62 (SD=437.84). The demographic and clinical characteristics of the participants are shown in Table1. The table shows a significant difference among gender (P=0.049), medication adherence (P=0.047), and drug dosage (P=0.033). There were no significant differences among other general characteristics.
Correlations of stressful life events, relapse, and social support
Table 2 shows the relationships between stressful life events, relapse, and the three dimensions of social support: subjective support, objective support, and availability. As shown in the table, stressful life events were positively associated with relapse (r=0.560, P < 0.01), while subjective support (r=-0.143, P < 0.05), objective support (r=-0.151, P < 0.05), availability (r=-0.128, P < 0.05), and total social support (r=-0.182, P < 0.05) were all negatively related to relapse. Additionally, the score for subjective support (M=19.69, SD=5.09) was higher than that for objective support (M=9.20, SD=3.62) and availability (M=7.69, SD=2.23).
Effects of demographics, stressful life events, and social support on relapse
Binary logistic regression analysis was used to explore the different variables
associated with relapse (Table 3). In the below table, stressful life events ( β=0.012,P<0.01, OR=1.012, 95% CI [1.008-1.015]), subjective support (β=-0.058,P=0.034, OR=0.943, 95% CI [0.894-0.996]) and objective support (β=-0.079, P=0.045, OR=0.924, 95% CI [0.856-0.998]) had a significant influence on relapse. The results further revealed that gender (β=-0.547, P=0.049, OR=0.578, 95% CI [0.335-0.998]), medication adherence (β=-0.559, P=0.048, OR=0.572, 95% CI [0.328-0.995]), and drug dose (β=-0.001, P=0.047, OR=0.999, 95% CI [0.998-1.000]) were all significantly associated with schizophrenia relapse.
Moderating effects of subjective support on the relationship between stressful life events and relapse
We chose Model 1 to examine the moderating effects of subjective support in the SPSS PROCESS 3.4. The analysis outputs are listed in Table 4. From the table, it is evident that the lower limit of confidential interval (LLCI) and the upper limit of confidential interval (ULCI) did not include “zero,” which suggested that subjective support exerted a significant moderating effect on the relationship between stressful life events and relapse. The conditional effect of stressful life events on relapse was significant in subjective support values from M-1SD (-5.0910), M (0.0000) to M+1SD (5.0910) (Table 5). However, if subjective support was low, the effect of stressful life events on relapse was stronger than that of high level of subjective support.
Figure 1 shows the output model of the moderating role of subjective social support. It was evidence that after including subjective support, the relationship between stressful life events and relapse became more significant (β=-0.001), suggesting that social support was negatively related to relapse. To examine how different levels of subjective support moderated these relationships, we drew a simple slope analysis diagram (Figure 2). According to Figure 2, it can be concluded that when the value of stressful life events was low, patients with high subjective support had fewer relapses than those with low social support.
Table1 Demographic characteristics of the participants
Relapse (n=106) Non-relapse (n=109)
M±SD/n(%) M±SD/n(%) X2 /t P
|
Age 1.356 0.508
18-29 44(20.5) 42(19.5)
30-44 45(20.9) 54(25.1)
45-60 17(7.9) 13(6.0)
Gender 3.871 0.049
Men 52 (24.2) 68(31.6)
Women 54(25.1) 41(19.1)
Place of residence 3.120 0.210
City 41(19.1) 45(20.9)
Town 31(14.4) 21(9.8)
Rural 34(15.8) 43(20.0)
Marital status 0.084 0.959
Unmarried 57 (26.5) 59(27.4)
Married 30(14.0) 32(14.9)
Divorce 19(8.8) 18(8.4)
Education level 2.019 0.569
Primary school 4(1.9) 4(1.9)
Middle school 31(14.4) 31(14.4)
Senior school 31(14.4) 41(19.1)
University 40(18.6) 33(15.3)
Occupation 2.803 0.246
None 57(26.5) 55(25.6)
|
Manual labor 27(12.6) 38(17.7)
Mental labor 22(10.2) 16(7.4)
MHIPC 2.232 0.328
<1000 19(8.8) 25(11.6)
1000–3000 45(20.9) 51(23.7)
>3000 42(19.5) 33(15.3)
Family history 0.007 0.934
Yes 15(7.0) 15(7.0)
No 91(42.3) 94(43.7)
Payment method 2.846 0.241
Self-paying 57(26.5) 60(27.9)
City medical 27(12.6) 35(16.3)
insurance
Rural medical 22(10.2) 14(6.5)
insurance
Duration of disease 0.137 0.712
≤5years 45(20.9) 49(22.8)
>5years 61(28.4) 60(27.9)
Medications 6.415 0.093
None 27(12.6) 19(8.9)
Single 30(14.0) 41(19.1)
Multiple 49(22.8) 49(22.8)
Medication adherence 3.934 0.047
<6points 48(22.3) 35(16.3)
≥6points 58(27.0) 74(34.4)
Drug dose 222.41±215.93 323.62±437.84 -2.140 0.033
Psychiatric 44.76±11.13 44.54±9.19 0.160 0.873
symptoms
M, Mean value; SD, Standard deviation; Bold values indicates P<0.05; MHIPC, Monthly household income per capita;Drug dose, Chlorpromazine equivalent
|
Table2 Inter-correlations of stressful life events, relapse and social support
Variables
|
M±SD
|
1
|
2
|
3
|
4
|
5
|
6
|
1.Stressful life events
|
173.32±130.20
|
1
|
|
|
|
|
|
2.Subjective support
|
19.69±5.09
|
-0.143*
|
1
|
|
|
|
|
3.Objective support
|
9.20±3.62
|
-0.151*
|
0.436**
|
1
|
|
|
|
4.Availability
|
7.69±2.23
|
-0.128
|
0.392*
|
0.257*
|
1
|
|
|
5.Social support
|
36.53±8.70
|
-0.182*
|
0.892**
|
0.718**
|
0.592
|
1
|
|
6.SCH
|
0.49±0.50
|
0.560**
|
-0.143*
|
-0.151*
|
-0.128*
|
-0.182*
|
1
|
SCH, schizophrenia relapse
*P<0.05, ** P<0.01
Table 3 Binary logistic regression results of the relationships between variables and relapse
Variables
|
β
|
SE
|
Wald
|
P
|
OR
|
OR (95% CI)
|
Stressful life events
|
0.012
|
0.002
|
42.903
|
P<0.01
|
1.012
|
1.008 ~ 1.015
|
Subjective support
|
-0.058
|
0.028
|
4.476
|
0.034
|
0.943
|
0.894~0.996
|
Objective support
|
-0.079
|
0.039
|
4.035
|
0.045
|
0.924
|
0.856~0.998
|
Availability
|
-0.110
|
0.062
|
3.107
|
0.078
|
0.896
|
0.793~1.012
|
Social support
|
-0.040
|
0.016
|
6.033
|
0.014
|
0.960
|
0.930~0.992
|
Age
|
-0.009
|
0.035
|
0.064
|
0.800
|
0.991
|
0.926~1.061
|
Gender
|
-0.547
|
0.278
|
3.870
|
0.049
|
0.578
|
0.335 ~ 0.998
|
Residence
|
-0.063
|
0.157
|
0.162
|
0.688
|
0.939
|
0.690 ~ 1.277
|
Marital Status
|
0.031
|
0.18
|
0.029
|
0.865
|
1.031
|
0.725 ~ 1.466
|
Duration of disease
|
0.102
|
0.275
|
0.137
|
0.712
|
1.107
|
0.646~1.898
|
Education
|
0.083
|
0.155
|
0.288
|
0.592
|
1.087
|
0.802 ~ 1.473
|
Occupation
|
0.337
|
0.203
|
2.749
|
0.097
|
1.401
|
0.940 ~ 2.086
|
MHIPC
|
0.271
|
0.189
|
2.067
|
0.151
|
1.312
|
0.906 ~ 1.899
|
Family history
|
0.163
|
0.217
|
0.565
|
0.452
|
1.177
|
0.769 ~ 1.803
|
Payment method
|
0.162
|
0.181
|
0.793
|
0.373
|
1.175
|
0.824 ~ 1.677
|
Medication
|
-0.111
|
0.175
|
0.402
|
0.526
|
0.895
|
0.636 ~ 1.261
|
Medication
adherence
|
-0.559
|
0.283
|
3.905
|
0.048
|
0.572
|
0.328~0.995
|
Drug dose
|
-0.001
|
0.001
|
3.953
|
0.047
|
0.999
|
0.998~1.000
|
Psychiatric
symptoms
|
0.002
|
0.013
|
0.026
|
0.872
|
1.002 0.976~1.029
|
β, Standardized regression coefficient; SE, Standard error; Bold values indicates P<0.05; OR, Odds ratio; MHIPC: Monthly household income per capita; Drug dose, Chlorpromazine equivalent;
Table4 Subjective social support on the relationship between stressful life events and relapse
Model summary
|
Coeff
|
SE
|
Z
|
P
|
LLCI ULCI
|
constant
Gender
Duration of disease
Drug dose
Medication adherence
Psychiatric symptoms
X
M
Int-1
|
0.3502
-0.8230
-0.0240
-0.0017
-0.2412
-0.0073
0.0160
-0.1780
-0.0010
|
0.2032
0.3710
0.1270
0.0009
0.3557
0.0174
0.0020
0.0381
0.0004
|
1.7238
-2.2200
-0.1930
-1.9744
-0.6782
-0.4170
6.7250
-4.0300
-2.7940
|
0.0847
0.0260
0.8470
0.0483
0.4977
0.6767
0.0000
0.0001
0.0050
|
-0.0480 0.7484
-1.5500 -0.0960
-0.2720 0.2240
-0.0034 0.0001
-0.9383 0.4559
-0.0414 0.0269
0.0110 0.0197
-0.2590 -0.0970
-0.0020 -0.0003
|
Drug dose, Chlorpromazine equivalent; X, Stressful life events; M, Subjective social support; Int-1, Stressful life events X Subjective social support;Coeff, Coefficient; LLCI, Lower limit of confidential interval;ULCI, Upper limit of confidential interval
Table5 Conditional effect of stressful life events on relapse at values of the moderator
Value
|
Effect
|
SE
|
Z
|
P
|
LLCI
|
ULCI
|
-5.0910
|
0.0220
|
0.0035
|
5.8620
|
0.0000
|
0.0136
|
0.0290
|
0.0000
|
0.0152
|
0.0022
|
6.7250
|
0.0000
|
0.0109
|
0.0196
|
5.0910
|
0.0100
|
0.0021
|
4.6981
|
0.0000
|
0.0058
|
0.0150
|
SE: Standard error; P- values < 0.05 were considered significant (in bold); LLCI: Lower limit of confidential interval; ULCI: Upper limit of confidential interval.