Impact of the COVID-19 pandemic on mortality and loss to follow-up among patients with dementia receiving anti-dementia medications – Scientific Reports

Data source

This study was conducted using anonymous customized research data extracted from the NHIS Database between February 1, 2017, and January 31, 2021. This database is primarily based on the Korean NHIS, a single government insurer that covers approximately 97% of the Korean population and supports Korean hospitals and nursing facilities. The customized database is representative of the transmission data provided by anonymous health insurance and long-term care insurance data²⁰. The database provides healthcare utilization information for both in- and outpatients, including patient demographics, diagnoses, comorbidities, and prescribed medications. The Korean Classification of Disease (KCD), 5th–7th editions, and a modification of the International Classification of Disease and Related Health Problems, 10th revision, were used to code the diagnoses. The NHIS coding system was used to collect demographic data (including age, sex, and income) and accompanying diagnostic codes, including diabetes (E10–14), chronic obstructive pulmonary disease (J44), chronic kidney disease (N18), dyslipidemia (E78), stroke (I60–64), hypertension (I10–15), and depression (F32, F33, and F34.1). The type of anti-dementia drug (donepezil, galantamine, rivastigmine, or memantine) was also extracted, and the pharmaceutical prescription codes for the anti-dementia drugs are described in Supplementary Table 1.

Ethical approval

This study was approved by the Institutional Review Board of Hanyang University Guri Hospital (2022-04-040) and registered with the Clinical Research Information Service (CRIS) under registration number KCT0008217. All personal information in the NHIS database was de-identified, and the requirement for informed consent was waived by the ethics committee of Hanyang University Guri Hospital. Furthermore, all methods were performed in accordance with the relevant guideline (STROBE checklist) and regulations.

Study population

All individuals in the customized research database visited a medical institution with a recorded dementia-related diagnostic code and anti-dementia medication from February 2018 to January 2020 (Fig. 1). Dementia was identified in the claims data based on KCD-5, 6, or 7 codes. Patients with dementia were defined as those with a history of outpatient visits or admissions based on dementia-related diagnostic codes and anti-dementia drug use. Dementia-related diagnostic codes were F00 (Dementia in Alzheimer’s disease), F01 (Vascular dementia), F02 (Dementia in other diseases classified elsewhere), F03 (Unspecified dementia), G30 (Alzheimer’s disease), G31.00 (Behavioral variant frontotemporal dementia), G31.01 (Semantic variant primary progressive aphasia), G31.02 (Nonfluent primary progressive aphasia), G31.03 (Logopenic primary progressive aphasia), G31.04 (Primary progressive aphasia), and G31.82 (Dementia with Lewy bodies). We established a washout period commencing in 2002. We excluded all patients who had documented visits to healthcare facilities with a dementia-related diagnostic code before February 2018.

Patients with dementia receiving anti-dementia medications were divided into two groups: (1) newly diagnosed with dementia between February 2018 and January 2019, and (2) newly diagnosed with dementia between February 2019 and January 2020 (Fig. 1). The records were followed for one year until January 2020 to January 2021. No medical records with dementia-related diagnostic codes during this period were considered as “follow-up loss,” while death was considered as “mortality.” Mortality data were obtained from the death dates recorded within the NHIS database.

Charlson Comorbidity Index (CCI)

We used the ICD-10 version of the CCI, which includes 17 diagnostic categories of acute myocardial infarction, congestive heart failure, peripheral vascular disease, cerebral vascular accident, dementia, pulmonary disease, connective tissue disorder, peptic ulcer, liver disease, diabetes mellitus, diabetes mellitus with complications, paraplegia, renal disease, cancer, metastatic cancer, severe liver disease, and HIV²¹. As all participants had dementia, the other 16 diagnostic categories were weighted and calculated as the CCI. The weighted values and corresponding ICD-10 codes are listed in Supplementary Table 2.

Statistical analysis

The annual numbers of patients with dementia in 2018 (from February 2018 to January 2019) and 2019 (from February 2019 to January 2020) were identified and followed for one year to determine mortality and follow-up loss (Fig. 1). Descriptive analyses were conducted to verify the demographic features of the two cohorts. All patients with dementia receiving anti-dementia medications were divided according to age (5 groups: < 50, 50–59, 60–69, 70–79, and ≥ 80), sex, and income (quintiles) and compared. Categorical variables are expressed as percentages or frequencies. Logistic regression analyses were performed to assess the relationships between various parameters (including year of diagnosis, age, sex, CCI score, residence, classification of the medical department, and income level) and follow-up loss or mortality. Age, sex, CCI score, residence, comorbidities, classification of medical institutions, classification of diagnosed medical departments, and income level were adjusted for as potential confounders. All statistical analyses were performed using SAS system version 9.4 (SAS Institute Inc., Cary, NC, USA) and Python 3.9.11 with the SciPy library, and p < 0.05 was considered statistically significant.