Table 2 Feature descriptions

Baseline features

69

 Date features

The year, month, and day of the week of admission

N

3

 Gender

Male or Female

D

2

 Age

Age of the patient

N

1

 Hospital affiliation

The affiliation of the hospital

N

1

 Admission status

1. Danger 2. Urgent 3. General

N

1

 Patient's and Hospital's address code

The smaller the value, the closer to the city center

N

2

 Address flag

Whether the patient's address code is equal to the hospital's address code

N

1

 Hospital levels

Measuring hospital quality

N

2

 Number of diseases

Number of diseases at the PoA

N

1

 Hospital admission source

1. Emergency treatment 2. Outpatient service 3. Transferred from Other medical institutions 4. Others

D

4

 Ethnic group

Han or minority

D

2

Job

The occupation of the patient

D

13

 Marital status

1. Spinsterhood 2. married 3. Divorce 4. Missing

D

4

 Elixhauser comorbidity index [36]

Including AIDS HIV, alcohol abuse, blood loss anemia, and so on

D

31

 Elixhauser comorbidity score [37]

A mapping score to represent one's health condition

N

1

Historical features

8

 Descriptive statistics of historical LOS

Extract the counts, mean, standard deviation, median, min, and a max of these LOS

N

6

 Last discharge interval

The days between the last discharge date and the date of current admission

N

1

 Last LOS

The LOS of the last hospital admission

N

1

MN features

657

 Eigenvector centrality features

For each chronic disease in the MN, extracting its eigenvector centrality value as features

N

653

 Disease risk features

Extract the counts, maximum, mean, and sum of disease risk scores

N

4

PSN features

5

 Descriptive statistics of neighbor's LOS

Extract the mean, standard deviation, median, min, and a max of these LOS

N

5