Search for


TEXT SIZE
search for




 

장기요양 필요 발생의 고위험 대상자 발굴을 위한 예측모형 개발
Development of prediction model identifying high-risk older persons in need of long-term care
Korean J Appl Stat 2022;35(4):457-468
Published online August 31, 2022
© 2022 The Korean Statistical Society.

송미경a, 박영우a, 한은정1,a
Mi Kyung Songa, Yeongwoo Parka, Eun-Jeong Han1,a

a국민건강보험공단 건강보험연구원

aHealth Insurance Research Institute, National Health Insurance Service
Health Insurance Research Institute, National Health Insurance Service, 2 segye-ro, Wonju-si, Gangwon-do, Korea, Republic of. E-mail: 9739han@nhis.or.kr
This paper was reconstructed from “An exploratory study on risk factors for long-term care needs using the National Health Insurance Service database (Song et al., 2021)”
Abstract
고령인구가 증가함에 따라 국가차원에서 노인의 건강노화 실현을 위한 장기요양 필요 발생의 예방 방안을마련하는 것은 매우 중요하며, 정책적 효과를 극대화하기 위해서는 적절한 대상자의 선정이 선행되어야 한다. 이에 본 연구는 국민건강보험공단의 국민건강정보를 활용하여, 장기요양 필요를 야기하는 기능장애 발생가능성이 높은 대상자를 발굴하기 위한 예측모형을 개발하고자 한다. 본 연구는 연구대상자의 과거 수집된 자료를 활용하는 후향적 연구로, 본 연구의 연구대상자는 만 65세 이상 의료보장등록인구이다(총 7,724,101명).예측모형 개발을 위해 고유 방법인 로지스틱 회귀모형, 머신러닝 방법인 의사결정나무와 랜덤포레스트, 딥러닝 방법인 다층퍼셉트론 신경망을 분석하였다. 체계적 분석절차를 통해 각 분석방법별 모형을 적합하였고,내적 타당성 및 외적 타당성 평가 결과를 기반으로 최종 예측모형을 랜덤포레스트로 선정하였다. 랜덤포레스트는 모집단에서의 4.50%밖에 되지 않는 장기요양 필요 대상자의 약 90%를 장기요양 필요 발생 고위험대상자로 예측할 수 있다. 본 연구의 예측모형 및 고위험군 기준은 노인의 욕구 중심에서 예방 서비스가 필요한 대상자를 선제적으로 발굴하는데 기여할 것으로 기대된다.
In aged society, it is important to prevent older people from being disability needing long-term care. The purpose of this study is to develop a prediction model to discover high-risk groups who are likely to be beneficiaries of Long-Term Care Insurance. This study is a retrospective study using database of National Health Insurance Service (NHIS) collected in the past of the study subjects. The study subjects are 7,724,101, the population over 65 years of age registered for medical insurance. To develop the prediction model, we used logistic regression, decision tree, random forest, and multi-layer perceptron neural network. Finally, random forest was selected as the prediction model based on the performances of models obtained through internal and external validation. Random forest could predict about 90% of the older people in need of long-term care using DB without any information from the assessment of eligibility for long-term care. The findings might be useful in evidencebased health management for prevention services and can contribute to preemptively discovering those who need preventive services in older people.
주요어 : 장기요양, 머신러닝, 예측모형, 예방
Keywords : long-term care, machine learning, prediction model, prevention
 
References
  1. Chung KH (2015). Characteristic changes and living status of older Koreans, Health and Welfare Policy Forum.
  2. Freedman VA (2009). Adopting the ICF language for studying late-life disability: A field of dreams?, Journals of Gerontoloty: Medical Sciences, 64, 1172-1174.
    Pubmed KoreaMed CrossRef
  3. Han EJ, Kwak MJ, and Kan IO (2011). A determining system for the category of need in long-term care insurance system using decision treem model, The Korean Journal of Applied Statistics, 24, 145-159.
    CrossRef
  4. Han E, Kwon J, Song MK, Lee Y, Jang H, Kim M, Lee HS, and Kim JH (2018). Establishing a Prospective Cohort Study of Oder People with Long-Term Care Needs II, National Health Insurance Service, Wonju.
  5. Han ST, Kang HC, Choi BS, and Lee SK (2011). A study on the judgement rating for level of need for longterm care insurance using a decision tree, Communications for Statistical Applications and Methods, 18, 137-146.
    CrossRef
  6. Kim Y (2020). A Study on the Role Allocation of Private Insurance for Enhancing the Sustainability of Long-Term Care Insurance, Korea Insurance Research Institute, Seoul.
  7. Lim JG (2008). A Study on the projection of long-term care for the disabled in Korea, Health and Social Welfare Review, 25 185-208.
    CrossRef
  8. Lionis C and Midl¨ov P (2017). Prevention in the elderly: A necessary priority for general practitioners, European Journal of General Practice, 23, 202-207.
    Pubmed KoreaMed CrossRef
  9. Nagi SZ (1991). Disability concept revisited: implications for prevention, Disability in America. Toward a national agenda for prevention, 309-327.
  10. National Council on Disability (2008). National Disability Policy: Progress Report, Chapter 5: Long-term Services and Supports, National Council on Disability, Washington, DC.
  11. OECD (2021). OECD health statistics 2021. https://www.oecd.org/els/health-systems/health-data.htm.
  12. Rowe JW and Kahn RL (1987). Human aging: usual and successful. Science, 237, 143-149.
    Pubmed CrossRef
  13. Seo DM and Jang BW (2005). A study on estimating the cost of long-term care for the elderly by social insurance model in Korea, Korean Social Security Studies, 21, 161-198.
  14. Statistics Korea (2021). 2021 Elderly statistics.
  15. Statistics Korea (2021). Life table. https://www.index.go.kr/unify/idx-info.do?idxCd=4234
  16. Thomas MG, Heather GA, and Theodore RH (2004). Zhenchao G. Hospitalization, Restricted Activity, and the Development of Disability among Older Persons, JAMA, 292, 2115-2124.
    Pubmed CrossRef
  17. Verbrugge LM and Jette AM (1994). The disablement process. Social Science & Medicine, 38, 1-14.
    CrossRef
  18. WHO(1991). Disability prevention and rehabilitation. Executive board. Eighty-ninth session: provisional agenda item 8. World Health Organization.
  19. WHO (2015). World Report On Ageing and Health. World Health Organization. World Health Organization.
  20. WHO (2017). Global strategy and action plan on ageing and health. Geneva. World Health Organization.World Health Organization.
  21. Won JW (2014). Health promotion and disease prevention in the older adults, J Korean Med Assoc, 57, 756-762.
    CrossRef
  22. Yun E and Jo Y (2014). Factors for environment affecting the social activity disability of the elderly, Korean Journal of Care Management, 12, 21-50.
December 2024, 37 (6)
Full Text(PDF) Free