About Dataset

Context

The data are simulated based on real hospital administrative data for England called Hospital Episodes Statistics. Every public (National Health Service, NHS) hospital in the country must submit records for every admission; private hospitals also submit records for any NHS patients that they treat.
Your simulated extract contains a random sample of emergency (unplanned) admissions for heart failure (ICD10 code I50). Here's a list of the fields and an explanation for some of them. Many of the fields are comorbidities coded as 0/1, where 1 indicates that the patient had it recorded.

Content :

death (0/1)

los (hospital length of stay in nights)

age (in years)

gender (1=male, 2=female)

cancer

cabg (previous heart bypass)

crt (cardiac resynchronisation device - a treatment for heart failure)

defib (defibrillator implanted)

dementia

diabetes (any type)

hypertension

ihd (ischaemic heart disease)

mental_health (any mental illness)

arrhythmias

copd (chronic obstructive lung disease)

obesity

pvd (peripheral vascular disease)

renal_disease

valvular_disease (disease of the heart valves)

metastatic_cancer

pacemaker

pneumonia

prior_appts_attended (number of outpatient appointments attended in the previous year)

prior_dnas (number of outpatient appointments missed in the previous year)

pci (percutaneous coronary intervention)

stroke (history of stroke)

senile

quintile (socio-economic status for patient's neighbourhood, from 1 (most affluent) to 5 (poorest))

ethnicgroup (see below for categories)

fu_time (follow-up time, i.e. time in days since admission to hospital)

Ethnic group has the following categories in this extract:

1=white

2=black

3=Indian subcontinent

8=not known

9=other

Inspiration :

Factors associated with hospital emergency readmission and mortality rates in patients with heart failure.