Skip to content

Datasets

The NHIS offers several research databases that differ in population, sampling, and variable names. Each has its own page with the same sections, namely sample design, versions, tables, linkage, and known issues.

  1. Sample Database

  2. Customized Database

The documents each page follows, and the conflicts found between them, are listed on the sources and versions page. The NHISS portal also lists an educational sample research DB (교육용 표본연구DB), which is not documented here.

Comparison

The table compares the four ready-made samples on the features that most often decide which one fits a study.

National Sample Cohort Health Screening Cohort Senior Cohort Low-Risk Sample Dataset
한글 표본코호트DB 건강검진코호트DB 노인코호트DB 저위험 표본데이터셋
Current release 2.2 Manual version 2.3 2.0 Single release
Population Enrolled throughout 2006, all ages Aged 40 to 79 and screened in 2002 or 2003 Aged 60 to 80 and enrolled in 2008 Used medical care in 2014 to 2023
Sample 1,000,000, about 2 percent 514,866, 10 percent 511,953, 8 percent 1,000,000
Sampling Stratified Simple random Stratified Random
Refreshment Newborns every year None 60-year-olds every year None
Years 2002-2023 2002-2023 2002-2023 2014-2023
Age Birth year Age in years Age in years Birth year
Residence District (시군구) Province (시도) Province (시도) Province (시도)
Claims Full claims Full claims Full claims Annual days by diagnosis group
Health screening General General, life-transition, and oral General Selected results
Long-term care 2008-2023 No 2008-2023 No
Person identifier RN_INDI PERSON_ID PERSON_ID RN_INDI

Variable name crosswalk

The National Sample Cohort renamed most variables when it was rebuilt in 2017, while the Health Screening and Senior cohorts kept the older names. The crosswalk maps the main variables, using the Korean labels of the manuals.

Concept 한글 National Sample Cohort Health Screening Cohort Senior Cohort
Reference year 기준년도 STD_YYYY STND_Y STND_Y
Person identifier 개인고유번호, 개인일련번호 RN_INDI PERSON_ID PERSON_ID
Claim identifier 청구고유번호, 청구일련번호 RN_KEY KEY_SEQ KEY_SEQ
Provider identifier 요양기관고유번호, 요양기관식별대체번호 RN_INST YKIHO_ID YKIHO_ID
Age 출생연도, 연령 BTH_YYYY (birth year) AGE AGE
Date of death 사망연월, 사망일자 DTH_YYYYMM (year and month) DTH_MDY DTH_MDY
Cause of death 사망원인 COD1, COD2 DTH_CODE1, DTH_CODE2 DTH_CODE1, DTH_CODE2
Residence 시군구, 시도코드 SGG (district) SIDO (province) SIDO (province)
Insurance type 가입자격 구분, 가입자구분 GAIBJA_TYPE IPSN_TYPE_CD IPSN_TYPE_CD
Premium decile 보험료 10분위, 보험료분위 CTRB_Q10 CTRB_PT_TYPE_CD CTRB_PT_TYPE_CD
Disability severity 장애중증도 DSB_SVRT_CD DFAB_GRD_CD DFAB_GRD_CD
Disability type 장애유형 DSB_TYPE_CD DFAB_PTN_CD DFAB_PTN_CD
Start date of care 요양개시일자 MDCARE_STRT_DT RECU_FR_DT RECU_FR_DT
Medical specialty 진료과목코드 MCARE_SUBJ_CD DSBJT_CD DSBJT_CD
Primary diagnosis 주상병 SICK_SYM1 MAIN_SICK MAIN_SICK
Secondary diagnosis 부상병 SICK_SYM2 SUB_SICK SUB_SICK
Diagnosis code in the diagnosis table 상병기호 MCEX_SICK_SYM SICK_SYM SICK_SYM
Diagnosis type in the diagnosis table 상병분류구분 SICK_CLSF_TYPE SICK_DIV_TYPE_CD SICK_DIV_TYPE_CD
Days of care 요양일수 MDCARE_DD_CNT RECN RECN
Days of visits or inpatient days 입내원일수 VSHSP_DD_CNT VSCN VSCN
Total covered cost after review 심결요양급여비용총액 ED_RC_TOT_AMT EDEC_TRAMT EDEC_TRAMT
Patient cost sharing after review 심결본인부담금 EDC_SBA EDEC_SBRDN_AMT EDEC_SBRDN_AMT
Insurer payment after review 심결보험자부담금 EDC_INSUR_BRDN_AMT EDEC_JBRDN_AMT EDEC_JBRDN_AMT

A shared concept does not guarantee a shared coding. For example, SGG and SIDO differ in geographic detail, and the disability and premium variables were regrouped at different times in different databases.

Page template

A new dataset page should use the following sections, leaving out a section when the information is not yet known.

  • Overview. What population the database covers and what it is typically used for.
  • Sample design. How the sample was drawn, including the base year, strata, and sampling fraction.
  • Versions. Each release, its years of coverage, and what changed between releases.
  • Tables. The tables in the database, with Korean and English names and their unit of observation.
  • Linkage. How tables link to one another, with the Korean and English names of each identifier.
  • Known issues. Pitfalls specific to this database.