Table 2.

Causal effects of composite cognitive measure (COGN_t), PANSS symptoms – sum (log(PANSSsum_t)) and negative (log(PANSSneg_t)) in separate models, on number of workinghours per week (WH_t), by g-estimation in JUMP. Parameters refer to SNMM = (ψ₀ + ψ₁f(t))a_t (2), where a_t is COGN_t, log(PANSSsum_t), or log(PANSSneg_t), and f(t) is the function in eq. (3) with λ = 0.8

	COGNt			log (PANSSsumt)			log (PANSSnegt)
	Estimate	95% CI	p-value	Estimate	95% CI	p-value	Estimate	95% CI	p-value
ψ0	0.224	0.067, 0.381	0.005	1.55	−4.25, 7.35	0.6	3.59	−0.383, 7.564	0.077
ψ1	.	.	.	−0.31	− 0.57, − 0.052	0.018	−0.288	− 0.476, − 0.099	0.0029