Differences in reported sepsis incidence according to study design: a literature review

Table 3 Code-based studies of sepsis and severe sepsis incidence in the general population

	CDC, 1990 [57]	Angus, 2001 [14]	Martin, 2003 [63]	Flaatten, 2004 [21]	Dombrovskiy, 2005 [60]	Esper, 2006 [61]	Dombrovskiy, 2007 [59]	Shen, 2010 [56]	Wilhelms, 2010 [22]	Kumar, 2011 [62]	Lagu, 2012 [16]	Chen, 2013 [58]
Country/Region	USA	USA (7 states)	USA	Norway	New Jersey, USA	USA	USA	Taiwan	Sweden	USA	USA	Taiwan
Coding system	ICD-9	ICD-9	ICD-9	ICD-10	ICD-9	ICD-9	ICD-9	ICD-9	ICD-9/10^vi	ICD-9	ICD9	ICD9
Data source	NHDS	Constructed database	NHDS	NPR	New Jersey SID	NHDS	NIS	NHIRD	SHDR	NIS	NIS	NHIRD
Study population	NA	6,621,559	NA	700,107	7,364,550	NA	NA	201,657^iv 200,000^v	2,024,793	NA	NA	NA
Number of cases (sepsis/severe sepsis)	NA	NA/192,980	NA	6665/NA	24,765 - 30,081/8096 - 13,453	NA	NA	NA/7531^iv NA/5258^v	NA/37,990^vii NA/27,655^vii NA/12,512^vii	NA	NA	NA/40,856 - 116,749
Exclusion criteria	<1 year			Neonate sepsis	<18 years			Previous episode of severe sepsis	Neonate sepsis, previous episode of severe sepsis	<18 years	<18 years
Internal validation	No	Yes	Yes	No	No	(Yes)ⁱⁱ	No	Yes	No	No	No	No
Calendar year	1979; 1987	1995	1979; 2000	1999	1995–2002	1979; 2003	1993–2003	1997; 2006	1987; 2005	2000; 2007	2007	1997-2008
Sepsis incidence 100.000 person yrs⁻¹	74; 176	–	83; 240ⁱ	149	–	83; 275ⁱ	–	–	–	–	–	-
Severe sepsis incidence 100.000 person yrs⁻¹	–	300	–	–	135–208	–	65–135	153; 359^iii,iv 135; 217^iii,v	10; 35^vii 25; 43^vii 3; 13^vii	143; 343	1074^viii 303^viii	188 - 507

Characteristics of code based studies of sepsis and severe sepsis incidence extrapolated to the general population. i) Age-standardized to fit the population distribution in the 2000 U.S. consensus. ii) Method validated by Martin et al. iii) Age-standardized using 2000 world population reported by WHO as standard. iv) No exclusion criteria. v) Exclusion criteria as stated. vi) Discharge diagnoses were classified according to ICD-9 until the end of 1996. These were translated into ICD-10 for the methods of Angus et al. and Martin et al. vii) Using the method proposed in Angus et al., Flaaten et al. (time of incidence measure: 1997; 2005) and Martin et al., respectively. viii) Using the method proposed in Angus et al. and Dombrovskiy et al., respectively. Abbreviations: −, not calculated; SHDR Swedish hospital discharge register, NA not available, NHDS national hospital discharge survey (USA), NHIRD national health insurance research (Taiwan), NIS nationwide inpatient sample (USA), NPR Norwegian patient register; yrs, years

ISSN: 1471-2288