Background

Sepsis is defined as acute organ dysfunction caused by an acute infection. There is no single diagnostic test to confirm diagnosis of sepsis, rather it is a complex syndrome. This is a problem in the clinical domain, making sepsis difficult to identify. This problem is amplified in the epidemiology of sepsis. Different groups of experts have created groups of International Calssificationof Diseases diagnosis and procedure codes to identify sepsis cases in administrative data. These data sources continue to be important epidemiological tools given their widespread use and ability to capture population-level data around the world.

In 2015, the US adopted the ICD10 coding system. In this brief report we create a longitudinal timeseries examining the incidence of adult sepsis by different ICD9 and ICD10 definitions.

Data Source

We utilize the Healthcare Utilization Project’s Nationwide Inpatient Sample (NIS), years 2012-2014 and 2016. We omitted 2015, given that the US adopted ICD10 in hte fourth quarter of this year. This makes the comparative estimation of annual incidences from seasonal quarters more complicated in a disease process that already has a seasonal variation. Furthermore, we hypothesized that the first few months of ICD10 adoption would be less reliable.

The “NIS is the largest publicly available all-payer inpatient health care database in the United States, yielding national estimates of hospital inpatient stays. Unweighted, it contains data from more than 7 million hospital stays each year from ~1,000 hospitals. Weighted, it estimates more than 35 million hospitalizations nationally”. The data from these years represents a 20% stratified sample of discharges from all available US community hospitals in states participating in HCUP.

Defintions

ICD9 Definitions

  • icd9_gbd_explicit =
  • icd9_gbd_implicit =
  • icd9_gbd_sepsis =
  • icd9_angus_explicit =
  • icd9_angus_implicit =
  • icd9_angus_sepsis =

ICD10 Definitions

  • icd10_cms =
  • icd10_gbd_explicit =
  • icd10_gbd_implicit =
  • icd10_r =

Data Table
VarName Sum StdDev year Pop Rate100K upper lower
icd9_gbd_explicit 1520500 19286.0 2012 240291024 632.8 1559072.0 1481928.0
icd9_gbd_implicit 3519636 37531.0 2012 240291024 1464.7 3594698.0 3444574.0
icd9_gbd_sepsis 4222136 44281.0 2012 240291024 1757.1 4310698.0 4133574.0
icd9_angus_explicit 614795 8328.4 2012 240291024 255.9 631451.8 598138.2
icd9_angus_implicit 2611676 28446.0 2012 240291024 1086.9 2668568.0 2554784.0
icd9_angus_sepsis 2779741 30138.0 2012 240291024 1156.8 2840017.0 2719465.0
icd9_gbd_explicit 1716384 21638.0 2013 242625484 707.4 1759660.0 1673108.0
icd9_gbd_implicit 3736723 38845.0 2013 242625484 1540.1 3814413.0 3659033.0
icd9_gbd_sepsis 4519758 46599.0 2013 242625484 1862.9 4612956.0 4426560.0
icd9_angus_explicit 683245 9229.1 2013 242625484 281.6 701703.2 664786.8
icd9_angus_implicit 2751494 29381.0 2013 242625484 1134.0 2810256.0 2692732.0
icd9_angus_sepsis 2945534 31384.0 2013 242625484 1214.0 3008302.0 2882766.0
icd9_gbd_explicit 2007905 24949.0 2014 244986302 819.6 2057803.0 1958007.0
icd9_gbd_implicit 4008747 41943.0 2014 244986302 1636.3 4092633.0 3924861.0
icd9_gbd_sepsis 4916582 51021.0 2014 244986302 2006.9 5018624.0 4814540.0
icd9_angus_explicit 799905 10517.0 2014 244986302 326.5 820939.0 778871.0
icd9_angus_implicit 2930596 31361.0 2014 244986302 1196.2 2993318.0 2867874.0
icd9_angus_sepsis 3166436 33728.0 2014 244986302 1292.5 3233892.0 3098980.0
icd10_cms 2360084 26581.0 2016 249485228 946.0 2413246.0 2306922.0
icd10_gbd_explicit 2364874 26623.0 2016 249485228 947.9 2418120.0 2311628.0
icd10_gbd_implicit 3293437 33389.0 2016 249485228 1320.1 3360215.0 3226659.0
icd10_r 919109 11498.0 2016 249485228 368.4 942105.0 896113.0