Table 4.

Interactions between any cannabis use and predisposition for psychosis

				Difference in risk
Cannabis use at baseline	No with psychosis outcome*	No without psychosis outcome*	Risk of psychotic symptoms at follow up	Unadjusted	Adjusted†(95% CI)
No predisposition for psychosis at baseline
None	294	1642	15%	6%	5.6% (0.4 to 10.8) P=0.033
Any (≥5 times)	59	216	21%
Predisposition for psychosis at baseline‡
None	47	133	26%	25%	23.8% (7.9 to 39.7) P=0.003
Any (≥5 times)	23	22	51%

^*Numbers total 2436 because of one missing value on predisposition for psychosis at baseline.

^†Age, sex, socioeconomic status, urbanicity, childhood trauma, and predisposition for psychosis at follow up. Test for additive interaction 18.2% adjusted difference in risk (95% confidence interval 1.6 to 34.8), P=0.032 (tests whether risk difference in “predisposition” group is significantly greater than risk difference in “no predisposition” group).

^‡Total score ≥90th centile on “paranoid ideation” and “psychoticism” subscales of symptom checklist.