Skip to content

Latest commit

 

History

History
49 lines (43 loc) · 2.89 KB

README.md

File metadata and controls

49 lines (43 loc) · 2.89 KB
Raw

Prediction-Model-of-COVID-19-Exacerbation

Abbreviations of Dataset

We open two datasets master.xlsx, and vital.xlsx. We encoded datasets as .xlsx format, and note that master.xlsx dataset includes Korean language data. master.xlsx is a comprehensive EMR chart dataset for patients hospitalized for each inpatient event. vital.xlsx vital sign time series dataset of patients. This is composed of 'serial', 'bt', 'pr', 'spo2', 'sbp', 'dbp' columns. corresponding to The datasets are in ./data folder. The folder structure is as belows:

.
├── data
    └── master.xlsx
    └── vital.xlsx
└── README.md

These raw datasets are original dataset before data cleaning preprocess. Notice few columns values related to identification information of patients were transformed. Especially we random shifted the 'ins' column information and align 'out', 'occur' columns to ins column for securing private identification information of patients.

  • Total number of events: 3748
  • Total number of objects: 3743
  • Details of columns:
    • serial : randomly shifted index number of events
    • ins (str dtype): date(YYYY-MM-DD) of patients entering in hospital
    • out (str dtype): date(YYYY-MM-DD) of a patient going out hospital
    • occur (str dtype): occuring date(YYYY-MM-DD) of patients showing their symptoms
    • trans (str dtype): binary categorical value whether the patient is transferred to next hospital like tertiary general hospital under Korean Hostpital system at that event ('전원' corresponds to 'transferred', and blank corresponds to 'not transfrred', at that event)
    • trans_date (st dtype): (HH-MM) time value of being transffered
    • gender (str dtype):
    • age (int dtype)
    • abroad (str dtype): O corresponds to True wheter the patients have experience going abroad recently at that evenvt
    • high_risk dtype)
    • history_yn (str dtype): Categorical values of which '유', '무' correspond to True, False respectively
    • history (str dtype): notes of patient's disease history
    • symptom1 (str dtype): initial symptoms of a patient at the event
    • symptom (str dtype): symptoms of a patient during hospitalization
    • result (str dtype)
    • wbc, crp, bun, cratinine, tbilirubin, procalcitinin (str dtype): blood test results of a patitent (We recommend you converting columns' data dtype from str to float or int dtype.)
    • bt, pr, spo2, sbp, dbp (str dtype): vital sign values of the patients at that event (We recommend you converting columns' data dtype from str to float or int dtype.)
    • created (str dtype): time information when an vital sign event of of a patient was created. Its format is 'yyyy"-"mm"-"dd" "hh"-"mm' in vital.xlsx.

Contributors

Corresponding Author

Contact information

Jung Hwa Lee, M.D., Ph.D. Tel: +82-2-2650-5589 Fax: +82-2-2650-5958 Email: [email protected], [email protected]

License

Corresponding Author