Evaluating Linkage Quality of Population-Based Administrative Data for Health Service Research.

BACKGROUND: To overcome the limitations of relying on data from a single institution, many researchers have studied data linkage methodologies. Data linkage includes errors owing to legal issues surrounding personal information and technical issues related to data processing. Linkage errors affect selection bias, and external and internal validity. Therefore, quality verification for each connection method with adherence to personal information protection is an important issue. This study evaluated the linkage quality of linked data and analyzed the potential bias resulting from linkage errors. METHODS: This study analyzed claims data submitted to the Health Insurance Review and Assessment Service (HIRA DATA). The linkage errors of the two deterministic linkage methods were evaluated based on the use of the match key. The first deterministic linkage uses a unique identification number, and the second deterministic linkage uses the name, gender, and date of birth as a set of partial identifiers. The linkage error included in this deterministic linkage method was compared with the absolute standardized difference (ASD) of Cohen's according to the baseline characteristics, and the linkage quality was evaluated through the following indicators: linked rate, false match rate, missed match rate, positive predictive value, sensitivity, specificity, and F1-score. RESULTS: For the deterministic linkage method that used the name, gender, and date of birth as a set of partial identifiers, the true match rate was 83.5 and the missed match rate was 16.5. Although there was bias in some characteristics of the data, most of the ASD values were less than 0.1, with no case greater than 0.5. Therefore, it is difficult to determine whether linked data constructed with deterministic linkages have substantial differences. CONCLUSION: This study confirms the possibility of building health and medical data at the national level as the first data linkage quality verification study using big data from the HIRA. Analyzing the quality of linkages is crucial for comprehending linkage errors and generating reliable analytical outcomes. Linkers should increase the reliability of linked data by providing linkage error-related information to researchers. The results of this study will serve as reference data to increase the reliability of multicenter data linkage studies.

Epidemiology of idiopathic sudden sensorineural hearing loss in the era of big data.

OBJECTIVES: Idiopathic sudden sensorineural hearing loss (ISSNHL) is a medical emergency, and delayed treatment can have permanent sequelae. However, the etiology of ISSNHL is diverse and unclear; thus, it is idiopathic. To develop an insight into this condition, patients with ISSNHL must be clearly identified. We propose an operational definition for the unambiguous identification of ISSNHL patients. Patients are identified through suggested definitions, and prevalence and general information are investigated. METHODS: A retrospective study of patients with ISSNHL was performed using the Health Insurance and Review Assessment-National Patient Sample from 2009 to 2016. To present a new operational definition, a systematic review was conducted for studies on ISSNHL from January 2007 to June 2021. After constructing several operant definitions using the conditions that can specify patients with ISSNHL in big data, we compared each definition to propose an operational definition. RESULTS: The important conditions required to classify patients with ISSNHL using big data were the International Classification of Diseases (ICD)-10 code, number of pure tone audiometry (PTA) tests, and whether steroids were prescribed. Among them, those who had undergone PTA tests more than twice could be clearly identified as patients with ISSNHL. CONCLUSION: As the use of big data becomes smoother, research using national medical data is being conducted; however, the results of the studies may vary depending on how a patient with ISSNHL is classified. Clear identification of patients with ISSNHL will be beneficial for better management of this condition.

Association between History of Pregnancy and Liver Fibrosis Using Fibrosis-4 Index in Korean Postmenopausal Women: A Nationwide Population-Based Study.

BACKGROUND: The association between a history of pregnancy and liver fibrosis remains unclear. Herein, we investigated the association between reproductive factors, including a history of pregnancy and liver fibrosis, in postmenopausal Korean women. METHODS: This study used nationally representative, population-based data collected from the Korea National Health and Nutrition Examination Survey 2008-2017. Of 14,624 women with natural menopause, 11,085 with no previous history of any type of cancer, hepatitis, or chronic heavy alcohol consumption were enrolled. We investigated the reproductive factors, including a history of pregnancy, total reproductive years, age at menarche and menopause, and oral contraceptive use. Liver fibrosis was defined as a Fibrosis-4 index score ≥2.67 kg/m2. RESULTS: Of the study participants, 372 (3.3%) had advanced liver fibrosis. Multivariable logistic regression analysis showed that women with a history of more than one pregnancy were associated with a lower risk of liver fibrosis compared to women who had never been pregnant, after adjusting for potential confounders (adjusted odds ratio, 0.30; 95% confidence interval, 0.15-0.59). The risk of liver fibrosis did not increase significantly with an increasing number of pregnancies (P for trend=0.135). Other reproductive factors, including total reproductive years, age at menopause and menarche, and oral contraceptive use, were not significantly associated with liver fibrosis. CONCLUSION: Postmenopausal women who had experienced one or more pregnancies had a reduced risk of liver fibrosis. Our findings reveal a potential protective role of pregnancy against liver fibrosis.

Relationship between Nutrient Intake and Hearing Loss According to the Income Level of Working-Aged Adults: A Korean National Health and Nutrition Survey.

The relationship between hearing impairment and nutrition has been extensively investigated; however, few studies have focused on this topic in working-age adults by income level. Herein, we aimed to determine the differences in hearing impairment among working-age adults by income level and identify the nutritional factors that affect hearing loss in various socioeconomic groups. Seven-hundred-and-twenty participants had hearing impairment, while 10,130 had normal hearing. After adjustment for propensity score matching, income and smoking status were identified as significant variables. By assessing the relationship between hearing impairment and nutrient intake by income level using multiple regression analyses, significant nutrients differed for each income category. Carbohydrate and vitamin C levels were significant in the low-income group; protein, fat, and vitamin B1 levels were significant in the middle-income group; and carbohydrates were significant in the high-income group. Income was significantly associated with hearing impairment in working-age adults. The proportion of individuals with hearing impairment increased as income decreased. The association between hearing impairment and nutritional intake also differed by income level. Our findings may enable the establishment of health policies for preventing hearing impairment in working-age adults by income level.

Validation of BMI genetic risk score and DNA methylation in a Korean population.

When DNA profiles obtained from biological evidence at a crime scene fail to match suspects or anyone in the database, forensic DNA phenotyping, which is the prediction of externally visible characteristics, can facilitate a traced search for an unknown suspect by limiting the search range. Therefore, age, trait, or lifestyle predictors, as well as the predictor for colorations, have been researched in the forensic field. In the present study, for the development of a prediction model for BMI or obesity, we investigated several previously reported BMI- or obesity-associated genetic and epigenetic markers that included four CpGs (cg06500161, cg00574958, cg12593793, and cg10505902 of the ABCG1, CPT1A, LMNA, and PDE4DIP genes, respectively), and eight SNPs (rs12463617, rs1558902, rs591166, rs11030104, rs11671664, rs6545814, rs16858082, and rs574367 near the TMEM18, FTO, MC4R, BDNF, GIPR/QPCTL, ADCY3/RBJ, GNPDA2, and SEC16B genes, respectively) in 700 Koreans within the BMI ranging from 16.1 to 40.6 (27.6 ± 4.5) kg/m2. Linear regression analysis showed that DNA methylation of the four CpG sites explained 10.9% total variance in BMI, and the model constructed using age information, genetic score from eight SNPs, and DNA methylation at four CpG sites could account for 17.4% of BMI variance. Using data mining techniques, i.e., decision tree (Entropy and Gini), random forest, and bagging, a total of eight models with BMI 31 or 32 as a cutoff value were also constructed based on the data obtained from 490 training samples with age and sex as a covariate. Among them, a random forest model with a cutoff value of 31 showed the best performance with 63.3% accuracy and the AUC value of 0.682 in 210 test set samples. In the present study, we could replicate the previous finding that DNA methylation contributes more to BMI than do genetic factors. In addition, although the accuracy for the prediction of BMI was not high, our study is meaningful in respect of the ability to use a small number of markers to achieve similar prediction accuracy to that obtained from a model composed of more than a thousand markers, which adds support to continued research to identify a small set of predictive markers for practical application in the forensic field.

Association between cotinine-verified smoking status and moderately increased albuminuria in the middle-aged and older population in Korea: A nationwide population-based study.

OBJECTIVES: Although several self-reported questionnaire-based studies have found an association between smoking and moderately increased albuminuria, this result remains controversial. We investigated whether moderately increased albuminuria was associated with smoking status, verified by urinary cotinine (an objective biomarker of tobacco exposure), using population-based, nationally representative data. METHODS: This study included 2059 participants aged ≥ 50 years from the 2014 Korean National Health and Nutrition Examination Survey. Individuals with a urinary cotinine level ≥ 50 ng/mL were identified as cotinine-verified smokers. Moderately increased albuminuria was defined as a urine albumin-to-creatinine ratio ranging between ≥ 30 mg/g and < 300 mg/g. Multivariable logistic regression was used to evaluate the association between cotinine-verified smoking status and moderately increased albuminuria. RESULTS: Among the study participants, 16.9% were cotinine-verified smokers, 84.8% of whom were men. After adjustment for multiple covariates, cotinine-verified smokers showed a significant positive association with moderately increased albuminuria (adjusted odds ratio: 4.37, 95% confidence interval: 1.63-11.71) compared with cotinine-verified non-smokers. The association between urinary cotinine and moderately increased albuminuria did not differ with age, sex, obesity, or comorbidities (P-value for interaction > 0.05 in all cases). CONCLUSION: This large-scale observational study showed that cotinine-verified smoking is associated with moderately increased albuminuria in the Korean middle-aged and older general population, suggesting that smoking must be strictly controlled to reduce the risk of moderately increased albuminuria.

Using GA-Ridge regression to select hydro-geological parameters influencing groundwater pollution vulnerability.

For groundwater conservation and management, it is important to accurately assess groundwater pollution vulnerability. This study proposed an integrated model using ridge regression and a genetic algorithm (GA) to effectively select the major hydro-geological parameters influencing groundwater pollution vulnerability in an aquifer. The GA-Ridge regression method determined that depth to water, net recharge, topography, and the impact of vadose zone media were the hydro-geological parameters that influenced trichloroethene pollution vulnerability in a Korean aquifer. When using these selected hydro-geological parameters, the accuracy was improved for various statistical nonlinear and artificial intelligence (AI) techniques, such as multinomial logistic regression, decision trees, artificial neural networks, and case-based reasoning. These results provide a proof of concept that the GA-Ridge regression is effective at determining influential hydro-geological parameters for the pollution vulnerability of an aquifer, and in turn, improves the AI performance in assessing groundwater pollution vulnerability.