Scholarlink

Aims/hypothesis <p>Chronic kidney disease (CKD) represents a major and costly comorbidity in type 2 diabetes management. Identifying individuals with high healthcare costs due to CKD will support decision-making for early intervention. We used latent class analysis (LCA) to classify Chinese individuals with type 2 diabetes and incident CKD based on their demographic and clinical profiles.</p> Methods <p>For this study, 2886 individuals with type 2 diabetes and incident CKD and complete data for 42 attributes were selected from the prospective Hong Kong Diabetes Register cohort (2007–2019). We used LCA to select 14 variables to classify participants, followed by a hierarchical generalised linear mixed model to evaluate longitudinal healthcare costs among class memberships.</p> Results <p>During 109,784 person-years of follow-up, the incidence of CKD was 26.29 per 1000 person-years with a per-patient-per-year (PPPY) cost of US$4395&#xa0;±&#xa0;11,947 (mean&#xa0;±&#xa0;standard deviation). The four distinct classes used in the LCA based on baseline profiles were as follows: Class 1 (18.3%; PPPY: US$6087&#xa0;±&#xa0;15,519), namely those who were young at onset (44.4&#xa0;±&#xa0;10.3 years), had moderate comorbidities (25.6% had a moderate or high score on the Elixhauser Comorbidity Index [ECI]) and used multiple medications (90.2% used at least three medications); Class 2 (21.2%; PPPY: US$3822&#xa0;±&#xa0;9816), namely those who had old-age onset (66.9&#xa0;±&#xa0;6.9 years), had moderate comorbidities (27.8% had a moderate or high ECI score) and used multiple medications (70.7% used at least three medications); Class 3 (33.9%; PPPY: US$4260&#xa0;±&#xa0;11,725), namely those who were middle-aged at onset (54.2&#xa0;±&#xa0;10.0 years), had few comorbidities (14.0% had a moderate or high ECI score) and used few medications (15.6% used at least three medications); and Class 4 (26.5%; PPPY: US$3923&#xa0;±&#xa0;10,957), namely those who were middle-aged at onset (54.1&#xa0;±&#xa0;7.6 years), had moderate comorbidities (25.3% had a moderate or high ECI score) and used multiple medications (98.9% used at least three medications). Class 1 (young onset) and Class 3 (middle-aged onset) incurred the highest cost during the year of CKD onset, with those in Class 1 having more comorbidities than those in Class 3 at baseline. Multiple healthcare services contributed to the high healthcare costs in Class 1, with costs in Class 3 attributed mainly to post-CKD outpatient and psychiatric care.</p> Conclusions/interpretation <p>Those with young-onset type 2 diabetes incurred the highest cost during the year of CKD onset. Individuals with middle-aged onset type 2 diabetes with fewer comorbidities and less intensified treatment at baseline&#xa0;also had subsequent increased healthcare costs.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Differential healthcare costs in individuals with type 2 diabetes and incident chronic kidney disease in Hong Kong: a latent class trajectory analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月18日

查看全文
  • Yanrong Du,
  • Minglu Zhang,
  • Abby Q. Y. Li,
  • Eric S. H. Lau,
  • Hongjiang Wu,
  • Alice P. S. Kong,
  • Andrea O. Y. Luk,
  • Ronald C. W. Ma,
  • Chun-Kwan O,
  • Lee-Ling Lim,
  • Jenny Y. Z. Zhang,
  • Wai Kit Ming,
  • Weijian Ke,
  • Yanbing Li,
  • Juliana C. N. Chan,
  • Juliana N. M. Lui

摘要

Aims/hypothesis

Chronic kidney disease (CKD) represents a major and costly comorbidity in type 2 diabetes management. Identifying individuals with high healthcare costs due to CKD will support decision-making for early intervention. We used latent class analysis (LCA) to classify Chinese individuals with type 2 diabetes and incident CKD based on their demographic and clinical profiles.

Methods

For this study, 2886 individuals with type 2 diabetes and incident CKD and complete data for 42 attributes were selected from the prospective Hong Kong Diabetes Register cohort (2007–2019). We used LCA to select 14 variables to classify participants, followed by a hierarchical generalised linear mixed model to evaluate longitudinal healthcare costs among class memberships.

Results

During 109,784 person-years of follow-up, the incidence of CKD was 26.29 per 1000 person-years with a per-patient-per-year (PPPY) cost of US$4395 ± 11,947 (mean ± standard deviation). The four distinct classes used in the LCA based on baseline profiles were as follows: Class 1 (18.3%; PPPY: US$6087 ± 15,519), namely those who were young at onset (44.4 ± 10.3 years), had moderate comorbidities (25.6% had a moderate or high score on the Elixhauser Comorbidity Index [ECI]) and used multiple medications (90.2% used at least three medications); Class 2 (21.2%; PPPY: US$3822 ± 9816), namely those who had old-age onset (66.9 ± 6.9 years), had moderate comorbidities (27.8% had a moderate or high ECI score) and used multiple medications (70.7% used at least three medications); Class 3 (33.9%; PPPY: US$4260 ± 11,725), namely those who were middle-aged at onset (54.2 ± 10.0 years), had few comorbidities (14.0% had a moderate or high ECI score) and used few medications (15.6% used at least three medications); and Class 4 (26.5%; PPPY: US$3923 ± 10,957), namely those who were middle-aged at onset (54.1 ± 7.6 years), had moderate comorbidities (25.3% had a moderate or high ECI score) and used multiple medications (98.9% used at least three medications). Class 1 (young onset) and Class 3 (middle-aged onset) incurred the highest cost during the year of CKD onset, with those in Class 1 having more comorbidities than those in Class 3 at baseline. Multiple healthcare services contributed to the high healthcare costs in Class 1, with costs in Class 3 attributed mainly to post-CKD outpatient and psychiatric care.

Conclusions/interpretation

Those with young-onset type 2 diabetes incurred the highest cost during the year of CKD onset. Individuals with middle-aged onset type 2 diabetes with fewer comorbidities and less intensified treatment at baseline also had subsequent increased healthcare costs.

Graphical Abstract