Table A1.

Multivariable zero-truncated negative binomial regression model for length of hospital stay (LOHS) estimation in pulmonary embolism survivors with pulmonary embolism severity index (PESI) items separately assessed (n = 234).

	LOHS prediction	IRR (95%CI)	Pr(>|z|)
Intercept	4.638	1.774(1.148–2.742)	0.01
Age	5.207	1.138 (1.071–1.21)	<0.001
Sex	5.864	1.296 (1.083–1.55)	0.005
History or active cancer	4.72	1.02 (0.829–1.255)	0.851
BP systolic < 100 mmHg	5.452	1.197 (0.761–1.883)	0.436
Heart rate ≥ 110 bpm	6.139	1.361 (1.075–1.724)	0.01
Respiratory rate ≥ 30 breaths/min	4.876	1.058 (0.707–1.583)	0.783
Oxygen < 90%	6.188	1.373 (1.109–1.699)	0.004
Chronic lung disease	5.107	1.114 (0.878–1.414)	0.375
Altered mental status	5.003	1.089 (0.51–2.324)	0.826
Hypothermia < 36°C	4.821	1.045 (0.612–1.782)	0.872
Heart failure	9.345	2.113 (1.165–3.832)	0.014

The analysis of the items composing the PESI score showed that age, sex, heart rate over 110 bpm, oxygen saturation <9% and heart failure were significantly associated with LOHS. Statistically significant variables p < 0.05.