Table 3.
Distribution of putative Etiological Variables for level of outcome (n = 47,004)
|
PEs level1 N = 13,978 (29·7%) |
PEs level2 N = 23,567 (50·1%) |
PEs level3 N = 6402 (13·6%) |
PEs level4 N = 2118 (4·5%) |
PEs level5 N = 939 (2·0%) |
Deviation From Linearity | ||||||||
|
Mean (SD) |
β |
Mean (SD) |
β (95%CI) |
Mean (SD) |
β (95%CI) |
Mean (SD) |
β (95%CI) |
Mean (SD) |
β (95%CI) |
βa | χ2 | P-value | |
| Demography | |||||||||||||
|
Age 18·19 (SD 0·91) |
18·23 (0·95) |
0·00 (Ref) |
18·18 (0·90) |
−0·05***$ (− 0·06−− 0·03) |
18·16 (0·88) |
− 0·06***$ (− 0·09−− 0·04) |
18·16 (0·87) |
−0·07**$ (− 0·11–0·02) |
18·17 (0·83) |
−0·06 (− 0·12–0·00) |
0·01 | 4·30 | 0·03 |
|
0·00 (Ref) |
−0·07 ***$ (− 0·09−− 0·05) |
−0·12***$ (− 0·15−− 0·09) |
-0·15***$ (− 0·19--0·10) |
-0·18***$ (− 0·24--0·11) |
0·00 | 0·01 | 0·93 | ||||||
| N (%) | OR | N (%) | OR (95%CI) | N (%) | OR (95%CI) | N (%) | OR (95%CI) | N (%) | OR (95%CI) | ||||
|
Male n = 23,444 (49·9%) |
7353 (52·6) |
1·00 (Ref) |
11,313 (48·0) |
0·83***$ (0·80–0·87) |
3213 (50·2) |
0·91**$ (0·86–0·96) |
1096 (51·7) |
0·97 (0·88–1·06) |
469 (49·9) |
0·90 (0·79–1·03) |
0·06 | 42·18 | < 0·001$ |
|
1·00 (Ref) |
0·91***$ (0·91–0·95) |
1·12***$ (1·05–1·20) |
1·32***$ (1·19–1·47) |
1·43***$ (1·23–1·66) |
0·10 | 98·30 | < 0·001$ | ||||||
|
Ethnic Minority n = 4777 (10·1%) |
1358 (9·7) |
1·00 (Ref) |
2375 (10·1) |
1·05 (0·98–1·13) |
693 (10·8) |
1·15** (1·04–1·26) |
248 (11·7) |
1·26**$ (1·09–1·45) |
103 (11·0) |
1·16 (0·94–1·44) |
0·02 | 1·69 | 0·19 |
|
1·00 (Ref) |
0·99 (0·92–1·07) |
0·99 (0·89–1·11) |
1·01 (0·86–1·19) |
0·84 (0·66–1·07) |
0·00 | 0·01 | 0·91 | ||||||
|
Low Family Income n = 6986 (14·9%) |
2047 (14·6) |
1·00 (Ref) |
3446 (14·6) |
1·02 (0·96–1·08) |
970 (15·1) |
1·06 (0·98–1·16) |
339 (16·0) |
1·14* (1·00–1·29) |
184 (19·6) |
1·46***$ (1·23–1·73) |
0·04 | 8·72 | 0·003 |
|
1·00 (Ref) |
0·95 (0·89–1·01) |
0·89* (0·81–0·98) |
0·87 (0·75–1·00) |
0·98 (0·80–1·19) |
0·01 | 0·36 | 0·55 | ||||||
|
Urban Birth n = 6791 (14·4%) |
2292 (16·4) |
1·00 (Ref) |
3221 (13·7) |
0·80***$ (0·75–0·85) |
835 (13·0) |
0·76***$ (0·69–0·82) |
318 (15·0) |
0·89 (0·79–1·01) |
125 (13·3) |
0·78* (0·64–0·94) |
0·06 | 16·68 | < 0·001$ |
|
1·00 (Ref) |
0·87***$ (0·81–0·92) |
0·92 (0·84–1·01) |
1·19* (1·03–1·38) |
1·19 (0·96–1·48) |
0·09 | 37·27 | < 0·001$ | ||||||
|
PEs level1 N = 13,978 (29·7%) |
PEs level2 N = 23,567 (50·1%) |
PEs level3 N = 6402 (13·6%) |
PEs level4 N = 2118 (4·5%) |
PEs level5 N = 939 (2·0%) |
Deviation From Linearity | ||||||||
| N (%) | OR | N (%) | OR (95%CI) | N (%) | OR (95%CI) | N (%) | OR (95%CI) | N (%) | OR (95%CI) | βa | χ2 | P-value | |
| Family History | |||||||||||||
|
Family History (psychosis) n = 338 (0·7%) |
67 (0·5) |
1·00 (Ref) |
159 (0·7) |
1·40* (1·05–1·86) |
69 (1·1) |
2·25***$ (1·60–3·15) |
27 (1·3) |
2·67***$ (1·70–4·18) |
16 (1·7) |
3·58***$ (2·06–6·19) |
0·05 | 1·10 | 0·29 |
|
1·00 (Ref) |
1·26 (0·94–1·69) |
1·74** (1·20–2·54) |
1·82* (1·09–3·05) |
2·02* (1·03–3·94) |
0·10 | 0·03 | 0·86 | ||||||
|
Family History (non-psychosis) n = 440 (0·9%) |
79 (0·6) |
1·00 (Ref) |
214 (0·9) |
1·58***$ (1·22–2·05) |
87 (1·4) |
2·39***$ (1·76–3·25) |
50 (2·3) |
4·22***$ (2·95–6·03) |
10 (1·1) |
1·87 (0·97–3·62) |
-0·01 | 0·01 | 0·91 |
|
1·00 (Ref) |
1·42** (1·09–1·85) |
1·82***$ (1·29–2·55) |
2·79***$ (1·83–4·25) |
1·00 (0·48–2·11) |
-0·05 | 1·06 | 0·30 | ||||||
| Adverse Childhood Experiences | |||||||||||||
|
Loss of Parent (Divorce) n = 1102 (2·3%) |
191 (1·3) |
1·00 (Ref) |
546 (2·3) |
1·71***$ (1·45–2·02) |
219 (3·4) |
2·57***$ (2·11–3·12) |
85 (4·0) |
3·03***$ (2·34–3·94) |
61 (6·0) |
5·03***$ (3·74–6·77) |
0·03 | 0·93 | 0·34 |
|
1·00 (Ref) |
1·61***$ (1·36–1·91) |
2·21***$ (1·78–2·74) |
2·42***$ (1·80–3·25) |
3·59***$ (2·50–5·16) |
0·00 | 0·01 | 0·92 | ||||||
|
Loss of Parent (Death) n = 632 (1·3%) |
145 (1·0) |
1·00 (Ref) |
304 (1·3) |
1·25* (1·02–1·53) |
114 (1·8) |
1·74***$ (1·36–2·23) |
48 (2·2) |
2·23***$ (1·60–3·10) |
21 (2·2) |
2·19***$ (1·38–3·49) |
0·06 | 2·41 | 0·12 |
|
1·00 (Ref) |
1·16 (0·94–1·42) |
1·44** (1·09–1·90) |
1·68** (1·15–2·45) |
1·43 (0·83–2·46) |
0·03 | 0·55 | 0·46 | ||||||
|
Physical Abuse n = 14,785 (31·5%) |
3829 (27·4) |
1·00 (Ref) |
6773 (28·7) |
1·09***$ (1·04–1·14) |
2625 (41·0) |
1·88***$ (1·76–2·00) |
1019 (48·1) |
2·49***$ (2·27–2·74) |
539 (57·4) |
3·66***$ (3·20–4·19) |
0·17 | 253·88 | < 0·001$ |
|
1·00 (Ref) |
1·01 (0·96–1·06) |
1·55***$ (1·44–1·66) |
1·87***$ (1·68–2·08) |
2·40***$ (2·06–2·80) |
0·14 | 171·46 | < 0·001$ | ||||||
|
Sexual Abuse n = 1360 (2·9%) |
311 (2·2) |
1·00 (Ref) |
666 (2·8) |
1·25**$ (1·09–1·43) |
208 (3·2) |
1·46***$ (1·22–1·75) |
102 (4·8) |
2·22***$ (1·77–2·79) |
73 (7·8) |
3·67***$ (2·82–4·78) |
0·13 | 25·67 | < 0·001$ |
|
1·00 (Ref) |
1·12 (0·97–1·29) |
1·11 (0·91–1·35) |
1·46** (1·13–1·90) |
1·97***$ (1·42–2·74) |
0·08 | 9·50 | 0·002$ | ||||||
|
Neglect n = 16,056 (34·2%) |
3511 (25·1) |
1·00 (Ref) |
7707 (32·7) |
1·46***$ (1·39–1·53) |
3042 (45·7) |
2·72***$ (2·55–2·89) |
1198 (56·6) |
3·90***$ (3·55–4·29) |
598 (63·9) |
5·27***$ (4·59–6·05) |
0·12 | 128·09 | < 0·001$ |
|
1·00 (Ref) |
1·32***$ (1·25–1·38) |
2·11***$ (1·97–2·27) |
2·69***$ (2·42–2·99) |
3·06***$ (2·62–3·57) |
0·09 | 64·77 | < 0·001$ | ||||||
|
> =2 maltreatment n = 10,567 (22.5%) |
2991 (21.4) |
1.00 (Ref) |
4531 (19.2) |
0.88***$ (0.84–0.93) |
1811 (28.3) |
1.46***$ (1.36–1.56) |
790 (37.3) |
2.20***$ (1.99–2.42) |
444 (47.3) |
3.33***$ (2.91–3.81) |
0.24 | 434 | < 0·001$ |
|
1.00 (Ref) |
0.82***$ (0.77–0.86) |
1.20***$ (1.11–1.29) |
1.64***$ (1.47–1.83) |
2.17***$ (1.85–2.53) |
0.21 | 315.23 | < 0·001$ | ||||||
No adjustment for demographic variables· Adjusted for age and sex in the first row for the rest variables· Second row for each variable additionally adjusted for PHQ-9 score
*P < 0·05, **p < 0·01, ***p < 0·001
aregression coefficient for the squared term of PEs scores· A negative squared term suggests a qualitatively stronger association at the lower end of the PEs spectrum, and a positive squared term suggests a qualitatively stronger association at the higher end of the PEs spectrum. Significant deviation from linearity indicates a positive linear effect
$: significance after correction for regressions of 21 variables using Bonferroni’s correction and the new threshold is 0.05/21 = 0.0024