Table 4. Hazard ratios and 95 % confidence intervals (CIs) for cancer-specific death occurred after 90 days of surgery among patients with different psychiatric statuses, subgrouped by the history of psychiatric disorder.

Psychiatric morbidity	Hazard ratios (95% CI)*
	All patients alive at least for 90 days after surgery (n = 2,485)	Patients with history of psychiatric disorder (n = 490)	Patients without history of psychiatric disorder (n = 1,995)
Outcome: overall mortality
Overall psychiatric morbidity
No	Reference	Reference	Reference
Yes	1.97 (1.72–2.26)	1.51 (1.03–2.22)	2.08 (1.80–2.41)
Specified by time of occurrence
No	Reference	Reference	Reference
Preoperative psychiatric morbidity	0.89 (0.67–1.19)	0.99 (0.56–1.78)	0.86 (0.60–1.21)
Pre- and post- operative (continued) psychiatric morbidity	1.52 (1.24–1.86)	1.46 (0.95–2.22)	1.57 (1.24–2.00)
Postoperative psychiatric morbidity only	3.22 (2.68–3.85)	2.92 (1.63–5.23)	3.30 (2.72–3.99)
Outcome: cancer-specific mortality
Overall psychiatric morbidity
No	Reference	Reference	Reference
Yes	1.94 (1.67–2.24)	1.31 (0.86–1.97)	2.08 (1.78–2.42)
Specified by time of occurrence
No	Reference	Reference	Reference
Preoperative psychiatric morbidity	0.83 (0.53–1.30)	0.68 (0.38–1.31)	0.90 (0.65–1.17)
Pre- and post- operative (continued) psychiatric morbidity	1.43 (1.15–1.77)	1.31 (0.84–2.06)	1.46 (1.13–1.89)
Postoperative psychiatric morbidity only	3.32 (2.75–4.01)	2.86 (1.52–5.36)	3.42 (2.80–4.18)

^* Evaluated by inverse probability of treatment weights (IPTW) weighted Cox proportional hazards regression model where psychiatric morbidity (or specific type of psychiatric morbidity) was included as a time-varying variable, weighed by propensity score and adjusted for hospital volume (low, median, or high), preoperative chemotherapy (yes/no), preoperative radiotherapy (yes/no), operation type (curative/palliative), history of mental disorder (yes/no), postoperative general complication (yes/no), and postoperative surgical complication (yes/no).