National Quality Evaluation of Medication Use for Pediatric Otitis Media With Effusion

OBJECTIVES

Otitis media with effusion (OME) is characterized by the presence of fluid in the middle ear without the presentation of signs or symptoms of acute ear infection. The point prevalence of OME reaches as high as 60% in children younger than 2 years of age. We used the National Health Insurance Research Database (NHIRD) to investigate the use of medication in children with OME before receiving ventilation tube insertion (VTI).

METHODS

Data of pediatric patients (age ≤ 12 years) who had OME and received VTI from January 1, 2011, to December 30, 2012, were retrieved from the Taiwan NHIRD. We surveyed the use of 4 medications to understand whether health care providers achieved the standards of medication use recommended by clinical practice guidelines.

RESULTS

This study examines the factors affecting the use of medication for pediatric OME. Overall, according to the study's operational definitions, the use of systemic antibiotics was most common (59.9%), followed by systemic antihistamines (23.4%), systemic steroids (8.8%), and intranasal steroids (9.6%). Systemic antibiotics use was associated with 12 factors. Ten of the 12 factors increase the use of systemic antibiotics, including namely age (age > 2 years), comorbidities, teaching hospital, and community hospital. In contrast, namely catastrophic illness and watchful waiting are the 2 factors that decrease systemic antibiotics use. For the use of systemic antihistamines, systemic steroids, and intranasal steroids were related to 6, 5, and 2 factors, respectively.

CONCLUSIONS

The rate of drug use differs from the rate of use recommended by commonly used clinical practice guidelines. We found that the higher the number of factors that influenced the patients' drug use, the higher the rate of drug use. According to these results, drafting a treatment guideline for OME patients in accordance with current clinical practices in Taiwan is highly recommended.

Association between Exposure to Particulate Matter Air Pollution during Early Childhood and Risk of Attention-Deficit/Hyperactivity Disorder in Taiwan

(1) Background: Recently, a growing number of studies have provided evidence to suggest a strong correlation between air pollution exposure and attention-deficit/hyperactivity disorder (ADHD). In this study, we assessed the relationship between early-life exposure to particulate matter (PM)₁₀, PM₂.₅, and ADHD; (2) Methods: The National Health Insurance Research Database (NHIRD) contains the medical records, drug information, inspection data, etc., of the people of Taiwan, and, thus, could serve as an important research resource. Air pollution data were based on daily data from the Environmental Protection Administration Executive Yuan, R.O.C. (Taiwan). These included particulate matter (PM_2.5 and PM₁₀). The two databases were merged according to the living area of the insured and the location of the air quality monitoring station; (3) Results: The highest levels of air pollutants, including PM_2.5 (adjusted hazard ratio (aHR) = 1.79; 95% confidence interval (CI) = 1.58-2.02) and PM₁₀ (aHR = 1.53; 95% CI = 1.37-1.70), had a significantly higher risk of ADHD; (4) Conclusions: As such, measures for air quality control that meet the WHO air quality guidelines should be strictly and uniformly implemented by Taiwanese government authorities.

Kawasaki disease in childhood and psychiatric disorders: A population-based case-control prospective study in Taiwan

BACKGROUND

Kawasaki disease (KD) is a common childhood acute inflammatory disease and potentially triggers a chronic inflammation. Although some researches have investigated neurodevelopmental consequences following KD, the findings have been inconsistent. This is the first population-based study targeted on KD and common psychiatric disorders.

OBJECTIVES

We aimed to investigate the association between KD and psychiatric disorders and hypothesized that standard anti-inflammatory treatment by intravenous immunoglobulin (IVIG) may protect against development of psychiatric disorders.

METHOD

We retrieved data from Taiwan's National Health Insurance Research database (NHIRD). Patients (n = 282,513) with psychiatric disorders (the case group) during 1997-2013 were included, and the control group was matched with age, sex, income and urbanization (1:1). We calculated the prevalence of KD in both groups and estimated odd ratios (ORs) and 95% confidence intervals (CIs) in the subgroup analyses for KD in conditions of age, severity, and common psychiatric comorbidity.

RESULTS

Numbers of patients with KD were 460 in the cases and 380 in the controls (p = .006), and the crude OR of KD was 1.21 times greater (95% CI = 1.06-1.39, p = .006) in the case than the control groups. KD patients without IVIG treatment (n = 126) were higher in the cases than those in the controls (n = 54), with the OR of 2.33 (95% CI = 1.70-3.21, p < .0001). Subgroup analyses showed that KD survivors were at significant risk for autism spectrum disorders (ASD) (OR = 2.15, 95% CI = 1.27-3.65; p = .005) and attention deficit and hyperactivity disorders (ADHD) (OR = 1.19, 95% CI = 1.02-1.39; p = 0.03), and a trend of increased risk for anxiety disorders (OR = 1.36, 95%CI = 0.99-1.86; p = 0.05).

CONCLUSIONS

Patients with KD were more likely to have comorbid psychiatric disorders, including ASD and ADHD. Moreover, anti-inflammatory treatment with IVIG may have potential prophylactic effects against the development of psychiatric disorders.

Metformin Use Was Associated With Reduced Risk of Incidental Sjögren's Syndrome in Patients With Type 2 Diabetes: A Population-Based Cohort Study

Purpose: Previous studies have shown that metformin exhibits an anti-inflammatory effect and may decrease the risk of incidental diabetes. But the effect of metformin on incidental Sjögren's syndrome is unknown. The aim of the study was to examine the association between metformin exposure and Sjögren's syndrome in diabetic patients. Methods: The dataset in this retrospective cohort study was obtained from the National Health Insurance Research Database (2000-2013) in Taiwan. In total, 15,098 type 2 diabetic patients under metformin treatment and an equivalent number without metformin treatment matched for comparison were included. The primary endpoint was the incidence of Sjogren's syndrome. Univariate and multivariate Cox proportional hazards models were used for data analysis. A subgroup analysis and sensitivity test were also performed. Results: The incidence rate of Sjögren's syndrome in non-metformin controls was 40.83 per 100,000 person-years and 16.82 per 100,000 person-years in metformin users. The adjusted hazard ratio (aHR) in diabetic patients under metformin treatment was 0.46 (95% CI, 0.23 to 0.92). In subgroup analysis, men had a lower risk of developing Sjögren's syndrome than women [aHR = 0.15, 95% CI = (0.05, 0.41)]. After prescribing metformin to type 2 diabetic patients aged 60 years or more, those patients had a lower risk of developing Sjögren's syndrome [aHR = 0.34, 95% CI = (0.12, 0.96)]. Conclusion: In this large population-based cohort study, metformin exposure was associated with a reduced risk of developing Sjögren's syndrome in type 2 diabetic patients.

Pediatric Nuclear Medicine Examinations and Subsequent Risk of Neoplasm: A Nationwide Population-Based Cohort Study

Objective: To evaluate the association between radiation exposure from repeated nuclear medicine (NM) examinations and the subsequent risk of neoplasm in pediatric patients.

Methods: From 2000 to 2017, participants under 18 years of age who underwent NM scanning were identified using the Health and Welfare Data Science Center (HWDC) dataset, which was extracted from the Taiwan National Health Insurance Research Database (NHIRD). Both the exposed cohort and unexposed subjects were followed up with until the presence of any malignancy arose, including malignant brain, lymphoid and hematopoietic tumors and benign brain or other central nervous tumors.

Results: There were 35,292 patients in the exposed cohort and 141,152 matched subjects in the non-exposed group. The exposed cohort had an overall higher IR (IR: incidence rate, per 100,000 person-years) of any malignancy and benign central nervous tumor than the non-exposed group [IR, 16.9 vs. 1.54; adjusted hazard ratio (HR), 10.9; 95% CI, 6.53–18.2]. Further stratifying the number of NM examinations into 1-2, 3-4, and 5 or more times revealed that the IR of pediatric neoplasms increased gradually with the increased frequency of NM examinations (IR, 11.5; adjusted HR, 7.5; 95% CI, 4.29–13.1; IR, 25.8; adjusted HR, 15.9; 95% CI, 7.00–36.1; IR, 93.8; adjusted HR, 56.4; 95% CI, 28.8–110.3).

Conclusion: NM examination is significantly associated with a higher risk of pediatric neoplasms, according to our population-based data. Thorough radiation protection and dose reduction in pediatric NM procedures should be an issue of concern.

All-cause mortality, cardiovascular mortality, major cardiovascular events and hypoglycaemia of patients with diabetes onset at an older age: results from the 10-year nationwide cohort study

BACKGROUND

Longitudinal adverse outcomes are unclear among adults with diabetes according to the age of onset.

OBJECTIVE

To investigate the longitudinal diabetes-related outcomes in adults with new-onset diabetes stratified by age.

DESIGN

Retrospective cohort study.

SETTING

Taiwan National Health Insurance Research Database claims data from 2000 to 2015.

SUBJECTS

In total, 115,751 participants aged ≥40 years with new-onset diabetes in 2003 were recruited and stratified by the ages 40-64 (64.3%), 65-74 (21.2%), 75-84 (11.8%) and ≥85 years (2.7%) at the time of diagnosis.

METHODS

Time-varying multivariate Cox proportional hazards model adjusted for covariates was used to examine the associations between the ages of the patients at diabetes onset and the outcomes of interest [all-cause mortality, cardiovascular (CV) mortality, major cardiovascular events (MACE) and hypoglycaemia] during a 10-year follow-up period.

RESULTS

The results showed that compared with those patients aged 40-64 at diagnosis, patients with older-onset diabetes had significantly higher comorbidities (P < 0.01) and a higher diabetes severity (P < 0.01). Patients with older-onset diabetes had a higher risk of all-cause mortality [adjusted hazard ratio (aHR) 2.28, 4.48 and 10.07 in 65-74, 75-84 and ≥85 years old, respectively], CV mortality (aHR = 2.82, 6.06 and 15.91), MACE (aHR = 2.19, 3.01 and 4.15) and hypoglycaemia (aHR = 2.41, 3.59 and 4.62) than patients aged 40-64 during a 10-year follow-up period.

CONCLUSIONS

Patients with diabetes onset at an older age was associated with increased risks of all-cause mortality, CV mortality, MACE and hypoglycaemia after adjusting for the severity of diabetes and anti-diabetic treatment.

Previous unilateral inguinal hernia repair increase risk of new developed inguinal hernia: a nationwide Longitudinal Cohort Study in Asian male adult patients

INTRODUCTION

Our study aims to identify that patients who received hernia repair previously did have higher risk of occurrence of newly developed inguinal hernia, named as a contralateral inguinal hernia (CIH), than patients who never received inguinal hernia surgery before.

MATERIALS AND METHODS

We collected data from the National Health Insurance Research Database (NHIRD) of Taiwan retrospectively. In the study cohort, 64,089 Asian male adults who underwent primary unilateral inguinal hernia repair during 2003-2008 were included using ICD-9 diagnostic and surgical codes. Another 64,089 male adults without hernia repair history were included as control group via propensity score match.

RESULTS

The median follow-up period is 93.53 months. After multivariate analysis, the risk of newly developed inguinal hernia in unilateral inguinal hernia (UIH) repair cohort was significantly higher (adjusted HR 6.364, 95% CI 6.012-6.737, P < 0.001) than the control group. In subgroup analysis, patients without mesh repair (adjusted HR 6.706, P < 0.001) and patients with mesh repair (adjusted HR 5.559, P < 0.001) both showed higher risk of developing newly developed inguinal hernia which needs repair.

CONCLUSIONS

Asian men with UIH repair history had a higher risk of developing new inguinal hernia at the contralateral site, namely CIH, than the general population. The surgeon should inform the possibility of CIH after initial herniorrhaphy, therefore, monitoring the occurrence of CIH is necessary.

Acupuncture Treatment Reduces Incidence of Parkinson's Disease in Patients With Depression: A Population-Based Retrospective Cohort Study in Taiwan

Depression is a risk factor for subsequent Parkinson's disease (PD). Some patients with depression undergo acupuncture treatment because of other diseases in Taiwan. Therefore, the present study used data from Taiwan's National Health Insurance Research Database (NHIRD) to investigate the incidence of PD in patients having depression with and without acupuncture treatment. We conducted a retrospective study of a matched cohort of 48,981 patients with newly diagnosed depression between 2000 and 2012 who were selected from the NHIRD. The 1:1 propensity score method was utilized to match an equal number of patients (N = 9,189) in the acupuncture and non-acupuncture cohorts. We employed Cox proportional hazard models to evaluate the risk of PD. The cumulative incidence of PD in both cohorts was estimated using the Kaplan–Meier method, and the difference was examined through a log-rank test. Patients with depression who received acupuncture treatment demonstrated a lower risk of PD [adjusted hazard ratio (aHR) = 0.39, 95% confidence interval = 0.31–0.49] than those who did not undergo acupuncture treatment, after adjusting for age, sex, insurance amount, geographic region, urbanization levels, comorbidities, and drugs. The cumulative incidence of PD was significantly lower in the acupuncture cohort than in the non-acupuncture cohort (log-rank test, p < 0.001). The database did not indicate the severity of depression and acupoints. The results suggest that acupuncture treatment significantly reduced the development of PD in patients with depression; however, a future study should be conducted to provide more objective evidence.

Risk of Intracranial Hemorrhage With Concomitant Use of Antidepressants and Nonsteroidal Anti-inflammatory Drugs: A Nested Case-Control Study

BACKGROUND

Whereas previous studies found that concomitant antidepressant and nonsteroidal anti-inflammatory drug (NSAIDs) use may increase the risk of gastrointestinal bleeding, either drug alone increases the risk of intracranial hemorrhage (ICH).

OBJECTIVE

To assess the risk for ICH in patients on concomitant treatment with antidepressants and NSAIDs.

METHODS

This was a nested case-control study using national insurance claims data in Taiwan between 2005 and 2013. Drug exposure was measured and compared during 3 time windows: 1 to 30, 31 to 60, and 61 to 90 days before the index date, which is the date of the ICH event. Both traditional and newer-generation antidepressants were considered in this study.

RESULTS

Patients exposed to both antidepressants and NSAIDs 1 to 30 days before the index date presented a 50% increased odds of developing ICH (OR: 1.53; 95% CI: 1.31-1.80) compared with patients receiving antidepressants alone. Specifically, the concomitant use of nonselective NSAIDs and antidepressants increased these odds compared with antidepressants alone (OR: 1.56; 95% CI: 1.31-1.84), but using a selective cyclooxygenase-2 inhibitor with antidepressant did not alter ICH risk. Regarding antidepressant class, newer-generation antidepressants generally increase the odds of developing ICH by 60% when used concomitantly with NSAIDs.

CONCLUSION AND RELEVANCE

Our results suggested that the concomitant use of antidepressants and NSAIDs was associated with an increased odds of developing ICH. NSAIDs, especially nonselective NSAIDs, and serotonergic antidepressants played an important role in this risk. Given the prevalent use of these 2 classes of drugs, this potential drug interaction deserves more attention.

Risk of Acquired Cholesteatoma and External Auditory Canal Stenosis in Traumatic Brain Injury: A Nationwide Population-Based Cohort Study

The aim of study is to investigate the risk of developing acquired cholesteatoma and external auditory canal (EAC) stenosis after traumatic brain injury (TBI) from the Taiwan National Health Insurance Research Database (NHIRD). Each subject was individually traced from their index date to identify those who received a diagnosis of acquired cholesteatoma and EAC stenosis. Cox regression analyses were applied to determine the risk of TBI-related acquired cholesteatoma and EAC stenosis. The follow-up data collected over 10 years were obtained from the TBI and comparison cohorts, of 455,834 and 911,668 patients, respectively. Multivariate analysis demonstrated that TBI significantly increased the risk of cholesteatoma (adjusted hazard ratio (HR), 1.777; 95% confidence interval (CI), 1.494-2.114, p < 0.001) and EAC stenosis (adjusted (HR), 3.549; 95% (CI), 2.713-4.644, p < 0.001). In our subgroup injury analysis, falls had the highest associated risk (4.308 times), followed by traffic injuries (66.73%; 3.718 times that of the control group). Otolaryngologists should not neglect the clinical importance and carefully investigate the possibility of subsequent cholesteatoma and EAC stenosis, which leads to hearing impairment in patients with TBI. Our research also shows the important role in preventing TBI, especially as a result of traffic injuries and falls.

Comparison of Elixhauser and Charlson Methods for Discriminative Performance in Mortality Risk in Patients with Schizophrenic Disorders

Although Charlson Comorbidity Index scores (CCIS) and Elixhauser comorbidity index scores (ECIS) have been used to assess comorbidity in patients with schizophrenia, only CCIS, not ECIS, have been used to predict mortality in this population. This nationwide retrospective study investigated discriminative performance of mortality of these two scales in patients with schizophrenia. Exploiting Taiwan's National Health Insurance Research Database (NHRID), we identified patients diagnosed with schizophrenia discharged from hospitals between Jan 1, 1996 and Dec 31, 2007. They were followed up for subsequent death. Comorbidities presented one year prior to hospital admissions were identified and adapted to the CCIS and ECIS. Discriminatory ability was evaluated using the adjusted hazard ratio and Akaike information criterion (AIC) and Harrell's C-statistic. We identified 58,771 discharged patients with schizophrenic disorders and followed them for a mean of 10.4 years, 16.6% of whom had died. Both ECIS and CCIS were significantly associated with mortality, but ECIS had superior discriminatory ability by a lower AIC and higher Harrell's C-statistic (201231 vs. 201400; 0.856 vs. 0.854, respectively). ECIS had better discriminative performance in mortality risk than CCIS in patients with schizophrenic disorders. Its use may be encouraged for risk adjustment in this population.

Increased Risk of Sensorineural Hearing Loss as a Result of Exposure to Air Pollution

Whether exposure to air pollution is associated with developing sensorineural hearing loss (SHL) remains controversial. Using data from the National Health Insurance Research Database, we recruited a total of 75,767 subjects aged older than 20 years with no history of SHL from 1998 to 2010, and they were followed up until SHL was observed, they withdrew from the National Health Insurance program, or the study ended. The subjects were evenly exposed to low-level, mid-level, and high-level carbon monoxide (CO) and nitrogen dioxide (NO₂). The incidence rate ratio of SHL for patients exposed to high-level CO was 1.24 (95% confidence interval (CI) = 1.14–1.36). The NO₂ pollutants increased the incidence rate ratios of SHL in mid-level NO₂ and high-level NO₂ exposures by 1.10 (95% CI = 1.10–1.32) and 1.36 (95% CI = 1.24–1.49) times, respectively. The adjusted hazard ratio (adj. HR) of SHL in patients exposed to high-level CO was 1.45 (95% CI = 1.31–1.59), relative to that of patients exposed to low-level CO. Compared to patients exposed to low-level NO₂, patients exposed to mid-level NO₂ (adj. HR = 1.40, 95% CI = 1.27–1.54) and high-level NO₂ (adj. HR = 1.63, 95% CI = 1.48–1.81) had a higher risk of developing SHL. The increased risk of SHL following the increased concentrations of air pollutants (CO and NO₂) was statistically significant in this study. In conclusion, the subjects’ exposure to air pollution exhibited a significantly higher risk of developing SHL in Taiwan.

Time-frame comparison of hystero-preservation in the surgical treatment of uterine prolapse: a population-based nation-wide follow-up descriptive study, 2006-2013 versus 1997-2005

INTRODUCTION AND HYPOTHESIS

The purpose of our study was to describe the surgical trend and time-frame comparison between 1997-2005 (1^stperiod) and 2006-2013 (2nd period) of hystero-preservation for uterine prolapse, based upon a nationwide population-based National Health Insurance (NHI) claim data in Taiwan.

METHODS

Women, who underwent primary surgeries for uterine prolapse, either uterine suspension with hystero-preservation or hysterectomy during 1997-2013, were identified from National Health Insurance Research Database (NHIRD).We analyzed the related variables including surgical types (hystero-preservation or hysterectomy), patient age and concomitant anti-incontinence surgery, surgeon age, gender, and service volume; and hospital accreditation level, and service volume.

RESULTS

We identified a total of 46,968 inpatients, 6629 (14.11%) hystero-preservation group, and 40,339 (85.89%) hysterectomy group. Hystero-preservation significantly increased during the 17 year follow-up study period (1997 to 2013) (P value =0.0147). The overall surgeries for uterine prolapse increased among patients ≥70 years, with concomitant anti-incontinence surgery, surgeons ≥45 years, high volume surgeons, and hospitals. Multiple logistic regression revealed hysterectomy was less used in 2nd period (OR 0.45, 95%, confidence interval (CI) 0.43-0.48). Hysterectomy was more often used in patient aged ≥50 years, surgeon aged ≥45 years, and low volume hospitals. In case of concomitant anti-incontinence surgery (OR 0.48, CI 0.45-0.52), high volume surgeons (more than 30 surgeries) and hospitals (more than 73 surgeries) hystero-preservation was more often used.

CONCLUSIONS

Time-frames, younger patients, concomitant anti-incontinence surgery, younger surgeons, and high volume surgeons and hospitals increase hystero-preservation for uterine prolapse. Surgeons and hospitals should be ready to respond to the wishes of female patients who want to preserve the uterus.

A fitting machine learning prediction model for short-term mortality following percutaneous catheterization intervention: a nationwide population-based study

Background

A suitable multivariate predictor for predicting mortality following percutaneous coronary intervention (PCI) remains undetermined. We used a nationwide database to construct mortality prediction models to find the appropriate model.

Methods

Data were analyzed from the Taiwan National Health Insurance Research Database (NHIRD) covering the period from 2004 to 2013. The study cohort was composed of 3,421 patients with acute myocardial infarction (AMI) diagnosis undergoing PCI. The dataset of enrolled patients was used to construct multivariate prediction models. Of these, 3,079 and 342 patients were included in the training and test groups, respectively. Each patient had 22 input features and 2 output features that represented mortality. This study implemented an artificial neural network model (ANN), a decision tree (DT), a linear discriminant analysis classifier (LDA), a logistic regression model (LR), a naïve Bayes classifier (NB), and a support vector machine (SVM) to predict post-PCI patient mortality.

Results

The DT model was found to be the most suitable in terms of performance and real-world applicability. The DT model achieved an area under receiving operating characteristic of 0.895 (95% confidence interval: 0.865-0.925), F1 of 0.969, precision of 0.971, and recall of 0.974.

Conclusions

The DT model constructed using data from the NHIRD exhibited effective 30-day mortality prediction for patients with AMI following PCI.

Obstructive sleep apnea among HIV-infected men in the highly active antiretroviral therapy era: a nation-wide longitudinal cohort study in Taiwan, 2000-2011

OBJECTIVES

This study compared the incidence of obstructive sleep apnea (OSA) between HIV-infected and matched non-infected persons and aimed to determine the effect of highly active antiretroviral therapy (HAART) on the incidence of OSA among HIV-infected men.

METHODS

A population-based cohort design was conducted using the National Health Insurance Research Database (NHIRD), which contained 13,552 HIV-infected persons and 134,361 matched age-, sex-, and index year-controls. The incidence density rates (IRR) were calculated to estimate the OSA incidence both in the HIV group and matched controls. Cox proportional hazard models were applied to assess the HAART effect on the incidence of OSA among HIV-infected men.

RESULTS

The IRR of OSA was lower in the HIV group than in the matched controls (1.91 vs. 2.26 per 1000 person-years, respectively). The associated factors for OSA in HIV-infected men were higher monthly salary (adjusted hazard ratio [aHR], 1.97, p < 0.01), obesity (aHR, 3.99, p < 0.01), and chronic obstructive pulmonary disease (aHR, 2.02, p < 0.01). Continued HAART at least a half year was associated with less diagnosed OSA (AHR, 0.70, p = 0.03) in HIV-infected men.

CONCLUSIONS

HIV-infected persons had a lower rate of OSA diagnosis compared to the matched controls. Healthcare providers are needed to increase awareness in HIV-infected men with sleep disorders, especially for those with low social economic status, comorbidities, and non-continuous HAART use.

Artificial Intelligence Prediction Model for the Cost and Mortality of Renal Replacement Therapy in Aged and Super-Aged Populations in Taiwan

BACKGROUND

Prognosis of the aged population requiring maintenance dialysis has been reportedly poor. We aimed to develop prediction models for one-year cost and one-year mortality in aged individuals requiring dialysis to assist decision-making for deciding whether aged people should receive dialysis or not.

METHODS

We used data from the National Health Insurance Research Database (NHIRD). We identified patients first enrolled in the NHIRD from 2000-2011 for end-stage renal disease (ESRD) who underwent regular dialysis. A total of 48,153 Patients with ESRD aged ≥65 years with complete age and sex information were included in the ESRD cohort. The total medical cost per patient (measured in US dollars) within one year after ESRD diagnosis was our study's main outcome variable. We were also concerned with mortality as another outcome. In this study, we compared the performance of the random forest prediction model and of the artificial neural network prediction model for predicting patient cost and mortality.

RESULTS

In the cost regression model, the random forest model outperforms the artificial neural network according to the mean squared error and mean absolute error. In the mortality classification model, the receiver operating characteristic (ROC) curves of both models were significantly better than the null hypothesis area of 0.5, and random forest model outperformed the artificial neural network. Random forest model outperforms the artificial neural network models achieved similar performance in the test set across all data.

CONCLUSIONS

Applying artificial intelligence modeling could help to provide reliable information about one-year outcomes following dialysis in the aged and super-aged populations; those with cancer, alcohol-related disease, stroke, chronic obstructive pulmonary disease (COPD), previous hip fracture, osteoporosis, dementia, and previous respiratory failure had higher medical costs and a high mortality rate.

Acupuncture treatment is associated with a decreased risk of developing stroke in patients with depression: A propensity-score matched cohort study

BACKGROUND

Depression had a significantly increased risk of stroke. Some patients with depression seek acupuncture treatment. The aim of this study is to investigate the association between acupuncture treatment and risk of developing stroke in patients with depression.

METHODS

Patients with newly diagnosed depression above 18 years old between 1997 and 2010 were selected from the Taiwanese National Health Insurance Research Database, and followed up until the end of 2013. The 1:1 propensity score method was used to match an equal number of patients (N = 13,823) in the acupuncture and non-acupuncture cohorts, based on characteristics including sex, age, baseline comorbidity and drug used. The outcome measurement was the comparison of stroke incidence in the two cohorts. We used the Cox regression model and the Kaplan-Meier method to estimate the risk of developing stroke.

RESULTS

Compared with the non-acupuncture cohort, patients who received acupuncture treatment had a lower risk of stroke (adjusted hazard ratio (HR) = 0.49, 95% confidential interval = 0.45-0.52) after adjustment for age, sex, comorbidities and drugs used. Regarding the two types of stroke, acupuncture users also had lower risk of hemorrhagic or ischemic stroke (adjusted HR = 0.37 and 0.49, respectively). The cumulative incidence of stroke in the acupuncture cohort was significantly lower than that of the non-acupuncture cohort (log-rank test, p < 0.001). At the 5th year of follow-up, 7.22% of acupuncture users had stroke onset compared with 14% in the non-acupuncture cohort.

LIMITATIONS

The database provided information of antidepressants but not the severity of depression. Details of the life style and selection of acupoints were also not provided.

CONCLUSIONS

The present study revealed that acupuncture reduced the risk of developing stroke in depression patients in Taiwan. Further clinical studies investigating the mechanisms and efficacy of acupuncture are warranted.

Risk of attempted suicide among adolescents and young adults with traumatic brain injury: A nationwide longitudinal study

BACKGROUND

Traumatic brain injury (TBI) and suicidal behavior lead to serious morbidity and premature mortality. TBI in adulthood is associated with a higher incidence of suicide, but the risk in adolescents and young adults is not clear.

METHODS

Longitudinal follow-up data were extracted from a National Health Insurance Research Database. Adolescents and young adults (12-29 years old) with and without TBI (1:4) were included, and the incidences of following attempted suicide were analyzed. The association of TBI severity, repeated TBI, and comorbid psychiatric disorders with attempted suicide were also investigated.

RESULTS

Overall, 31,599 and 126,396 subjects were enrolled in the TBI and control cohorts, respectively. The overall incidence of attempted suicide was significantly higher in the TBI cohort than in the control cohort (4.6% versus 1.0%, P < 0.001). The age at first suicide attempt was also lower in the TBI cohort (25.71 ± 5.25 versus 28.99 ± 5.26 years, P < 0.001). After adjusting for confounding factors, severe TBI, repeated TBI, female, younger age at TBI, and comorbid psychiatric conditions (unipolar depression, bipolar disorder, alcohol and substance use disorders) were associated with higher risks of attempted suicide. Long-term cumulative risks of attempted suicide were significantly elevated in the TBI cohort (P < 0.001).

LIMITATION

Access to individual's detailed medical record was not available.

CONCLUSION

TBI is associated with an elevated risk of attempted suicide in adolescents and young adults. TBI severity, repetitive injury, female, younger age at injury, and certain psychiatric comorbidities are independent risk factors.

Utilization of Chinese herbal medicine and its association with the risk of fracture in patients with Parkinson's disease in Taiwan

ETHNOPHARMACOLOGICAL RELEVANCE

Traditional Chinese medicine (TCM) has been used for over two thousand years to treat motor impairments corresponding to the clinical manifestations of Parkinson's disease (PD).

AIM OF THE STUDY

This study aimed to investigate the prescription of Chinese herbal medicine (CHM) for the management of PD and further determine whether CHM can improve motor function and decrease the risk of incident fracture.

MATERIALS AND METHODS

Patients older than 40 years newly diagnosed with PD between January 1997 and December 2010 were selected from the National Health Insurance Research Database (NHIRD) and followed up until the end of 2013. We used 1:1 frequency matching by age, sex, index year, and initial diagnostic year to compare the TCM users and non-TCM users. We used a Cox regression model and the Kaplan-Meier method to estimate the risk of developing fracture among the TCM and non-TCM users.

RESULTS

In total, 7197 patients older than 40 years were newly diagnosed with PD between 1997 and 2010 in Taiwan. Among these patients, 3456 were TCM users, and 3730 were non-TCM users. We compared 2007 PD patients with comparable demographic characteristics and comorbidity profiles between the two cohorts. During the follow-up period, compared with the non-TCM cohort, fewer patients in the TCM cohort had incident fractures (adjusted hazard ratio: 0.5, 95% CI: 0.44-0.56). The cumulative incidence of fracture was lower in the TCM cohort (log-rank test, p < 0.0001). Shi-Chang-Pu (Acorus gramineus Aiton), Yuan-Zhi (Polygala tenuifolia Willd), Bei-Mu (Fritillaria cirrhosa D. Don), Hai-Piao-Xiao (Sepiella maindronide Rochebrune; Sepia esculenta Hoyle), and Tian-Ma (Gastrodia elata Blume) constituted the core Chinese herbal medicine prescriptions used to treat PD patients.

CONCLUSIONS

The present study identified the core prescription pattern for the management of PD in Taiwan. Complementary CHM therapy was associated with a reduced risk of fracture in PD patients.

Association of Depression and Cervical Spondylosis: A Nationwide Retrospective Propensity Score-Matched Cohort Study

Objective: Depression is a psychiatric disorder associated with poorer health outcomes. Inappropriate mechanical stress and aging are factors associated with developing cervical spondylosis. The connection between cervical spondylosis and depression is not developed. Methods: From the health insurance claims data of Taiwan, we identified 34,166 persons newly diagnosed with depression in 2000⁻2010 and 34,166 persons without the disorder frequency matched by sex, age and diagnosis year. Both cohorts were followed up to the end of 2013 to estimate incident cervical spondylosis. We further examined the risk of cervical spondylosis in depressed people taking antidepressants. Results: The incidence of cervical spondylosis was 1.8-fold greater in the depression cohort than in comparison cohort (9.46 vs. 5.36 per 1000 person-years), with an adjusted hazard ratio (aHR) of 1.79 (95% confidence interval (CI) = 1.66⁻1.92). The incidence of cervical spondylosis increased in patients who had taken medications of serotonin-specific reuptake inhibitors (SSRIs) or of non-SSRIs than in those without these medicines (9.13 or 11.5 vs. 6.54 per 1000 person-years, respectively). Conclusions: Patients with depression are at an increased risk of developing cervical spondylosis. Additional efforts in reducing the risk of cervical spondylosis might be required in depressed individuals undergoing anti-depressive therapy.

Insomnia as an independent predictor of suicide attempts: a nationwide population-based retrospective cohort study

BACKGROUND

Numerous studies have verified that insomnia is associated with suicidal ideation, suicide attempts, and death by suicide. Limited population-based cohort studies have been conducted to examine the association. The present study aimed to analyze whether insomnia increases the risk of suicide attempts and verify the effects of insomnia on suicide risk.

METHODS

This study is a cohort study using 2000-2013 hospitalization data from the National Health Insurance Research Database (NHIRD) to track the rate of suicide attempts among insomnia patients aged 15 years or older. In addition, a 1:2 pairing based on sex, age, and date of hospitalization was conducted to identify the reference cohort (patients without insomnia). Cox proportional hazard model was used to assess the effects of insomnia on suicide risk.

RESULTS

The total number of hospitalized patients aged 15 years or older was 479,967 between 2000 and 2013 (159,989 patients with insomnia and 319,978 patients without insomnia). After adjusting for confounders, suicide risk in insomnia patients was 3.533-fold that of patients without insomnia (adjusted hazard ratio [HR] = 3.533, 95% confidence interval [CI] = 3.059-4.080, P < 0.001). Suicide risk in low-income patients was 1.434-fold (adjusted HR = 1.434, 95% CI = 1.184-1.736, P < 0.001) that of non-low-income patients. Suicide risk in patients with drug dependence and with mental disorders was 1.592-fold (adjusted HR = 1.592, 95% CI = 1.220-2.077, P < 0.001) and 4.483-fold (adjusted HR = 4.483, 95% CI = 3.934-5.109, P < 0.001) that of patients without drug dependence and without mental disorders, respectively. In the female population, suicide risk in insomnia patients was 4.186-fold (adjusted HR = 4.186, 95% CI = 3.429-5.111, P < 0.001) that of patients without insomnia. Among patients aged 25-44 years, suicide risk in insomnia patients was 5.546-fold (adjusted HR = 5.546, 95% CI = 4.236-7.262, P < 0.001) that of patients without insomnia. Furthermore, the suicide risk of insomnia patients with mental disorders was 18.322-fold that of patients without insomnia and mental disorders (P < 0.001).

CONCLUSION

Insomnia, low income, drug dependence, and mental disorders are independent risk factors for suicide attempts. Female patients and those aged 25-44 years are at high risk of suicide due to insomnia. Insomnia, mental disorders, and low income exhibit a synergistic effect on suicide attempts. Clinicians should pay attention to mental status and income level of insomnia patients.

A comparison of hospital and community stay in patients who underwent compulsory admission before and after the 2007 Amendment to the Mental Health Act in Taiwan

OBJECTIVE

The main purpose of this study was to assess the empirical findings of compulsory admission for psychiatric disorders before and after the 2007 amendment to the Mental Health Act in Taiwan.

METHODS

A matched case-control study design was applied. Participants were selected using the National Health Insurance Research Database (NHIRD) in Taiwan. The control and case data were collected in 2006 and 2011, and the number of compulsory admission cases was recorded with a case-control ratio of 1:4, along with information on age (±3 years) and gender. In 2006, the number of patients recruited was 9265, including 1853 compulsorily admitted patients and 7412 voluntarily admitted patients. In 2011, the number of patients recruited was 4505, including 901 compulsorily admitted patients and 3604 voluntarily admitted patients.

RESULTS

The data collected for the patients who underwent compulsory admission before and after the amended Mental Health Act included gender, diagnosis, Charlson Comorbidity Index Score (CCIS), length of stay in an acute hospital ward (days), hospital accreditation level, ownership, teaching hospital status, psychiatrist gender and age, and hospital location. Although the number of compulsory admission cases (1853 vs. 901) markedly decreased and the length of stay in an acute hospital ward (30.7 ± 25.0 days vs. 39.0 ± 22.6 days) increased from 2006 to 2011, the readmission rate was reduced from 52.6% in 2006 to 42.5% in 2011.

CONCLUSIONS

The average lengths of hospital stay and community survival time were greater for compulsorily admitted patients than those for voluntarily admitted patients. This result might be attributed to a number of changes implemented since the 2007 amendment of the Mental Health Act, including a strict review process for compulsory admissions and a new discharge planning process, which require further research for approval.

Risk of colorectal cancer in patients with periodontal disease severity: a nationwide, population-based cohort study

BACKGROUND

Periodontal disease (PD) and colorectal cancer (CRC) were associated with chronic inflammation. This retrospective cohort study examined the association between PD severity and CRC in a large-scale, population-based Chinese cohort.

METHODS

A total of approximately 106,487 individuals with newly diagnosed PD and 106,487 age-matched and sex-matched patients without PD from 2000 to 2002 were identified from Taiwan's National Health Insurance Research Database (NHIRD).

RESULTS

The Kaplan-Meier analysis revealed that the cumulative incidence of CRC was significantly higher in patients with PD than in those without PD (log-rank test, P < 0.001). After adjustment for age, sex, and comorbidities, patients with PD were associated with a significantly higher risk of CRC compared with those without PD (adjusted HR = 1.64, 95% CI = 1.50-1.80). Further, the risk of CRC appeared to increase with increasing frequency of PD medical visits [adjusted HR (95% CI) was 1.78 (1.58-2.02) and 1.53 (1.35-1.74) for annual visits > 10 and < 4, respectively].

CONCLUSION

Based on our study, PD severity was associated with an increase in the risk of CRC. Further mechanistic research is needed.

Risk factors for contralateral inguinal hernia repair after unilateral inguinal hernia repair in male adult patients: analysis from a nationwide population based cohort study

Background

To identify the rate of and risk factors for contralateral inguinal hernia (CIH) after unilateral inguinal hernia repair in adult male patients.

Methods

This retrospective cohort study identified from the Taiwan National Health Insurance Research Database (NHIRD). Information on all adult patients who underwent primary unilateral inguinal hernia repair without any other operation was collected using ICD-9 diagnostic and procedure codes. The exclusion criteria were laparoscopic hernia repair, non-primary repair, complicated hernia, other combined procedures, female and undetermined gender.

Results

A total of 170,492 adult male patients were included, with a median follow-up of 87 months. The overall CIH rate was 10.5%, with a median time of 48 months to a subsequent hernia operation. The 1-year, 2-year, 3-year and 5-year-recurrent rate was 2.6, 3, 4.3, and 6.7% respectively. Further, 3.7% patients who underwent CIH repair had a complicated inguinal hernia. Multivariate analysis demonstrated that age > 45 y, direct hernia, cirrhosis (HR = 1.564), severe liver disease (HR = 1.663), prostate disease (HR = 1.178), congestive heart failure (HR = 1.138), and history of malignancy (HR = 1.116) had a significantly higher risk of CIH repair.

Conclusions

Among adult male patients undergoing long-term follow-up, we identified several significant risk factors for CIH repair. If these risk factors are presented, the surgeon should inform the following risk of CIH repair to patients so that it can be repaired as soon as possible.

The Association Between the Use of Zolpidem and the Risk of Alzheimer's Disease Among Older People

OBJECTIVES

To evaluate the association between zolpidem use and the risk of Alzheimer's disease among older people.

DESIGN

A retrospective cohort study using data from 2001 to 2011 from the National Health Insurance Research Database.

SETTING

Taiwan.

PARTICIPANTS

A total of 6,922 patients aged 65 years or older enrolled from January 2002 to December 2004 (the enrollment period).

INTERVENTION (EXPOSURE)

Zolpidem users were identified as patients who used zolpidem during the enrollment period. The index date was the date of the first zolpidem prescription. Dosage of zolpidem use was defined using cumulative defined daily dose (cDDD) based on the cumulative dosage that patients took within one year after the index date (grouped as: less than 28, 28-90, 91-180, and more than 180 cDDD).

MEASUREMENTS

The occurrence of Alzheimer's disease was defined as the time period from the end of one year after the index date to the date of the Alzheimer's disease diagnosis. The propensity score was used to adjust the measured confounders of Alzheimer's disease. Cox proportional hazards models were used to evaluate the association between zolpidem use and the incidence of Alzheimer's disease.

RESULTS

Zolpidem users with a high cumulative dose (>180 cDDD) in the first year after initiation had a significantly greater risk of Alzheimer's disease than non-zolpidem users (HR = 2.97, 95% CI = 1.61-5.49) and low cumulative dose (<28 cDDD) users (HR = 4.18, 95% CI = 1.77-9.86).

CONCLUSION

We found the use of a high cumulative dose of zolpidem was associated with an increased risk of Alzheimer's disease among older people living in Taiwan. It is advised to use caution when considering long-term use of zolpidem in older patients.

Sleep disturbances at the time of a new diagnosis: a comparative study of human immunodeficiency virus patients, cancer patients, and general population controls

OBJECTIVES

Sleep disturbances are a prevalent and troubling symptom of patients with highly stressful illnesses, such as human immunodeficiency virus (HIV) and cancer. The aim of this study was to compare the prevalence and incidence of sleep disturbances among persons with HIV, those with cancer, and the general population of Taiwan.

METHODS

A matched cohort study design was used to compare the risk of sleep disturbances among three groups using reimbursement claims recorded in Taiwan's National Health Insurance Research Database (NHIRD). A total of 14,531 HIV-infected persons were compared with 1493 cancer patients and 1373 general population controls matched by gender and age. Cox proportional hazard regression models were used to test the hazard risk of sleep disturbances among the groups.

RESULTS

The mean durations between the date of the initial HIV/cancer diagnosis and onset of sleep disturbances of HIV-infected persons, cancer patients, and controls were 1.7, 2.3, and 1.8 years, respectively. The risk of developing sleep disturbances was significantly higher in HIV-infected persons (adjusted hazard ratio [AHR] = 3.74, p < 0.001) and cancer patients (AHR = 2.72, p < 0.001) than in controls. HIV-infected persons had a 20% higher risk of sleep disturbances than cancer patients (AHR = 1.20, p < 0.001).

CONCLUSIONS

HIV-infected persons exhibited a higher risk of developing sleep disturbances than cancer patients and general population controls. With efficacious treatments for sleep disturbances, we should focus on training and research programs for health care providers to intervene and treat earlier for the present and future health of cancer patients and HIV-infected persons.

Cost-effectiveness of granulocyte colony-stimulating factor prophylaxis in chemotherapy-induced febrile neutropenia among breast cancer and Non-Hodgkin's lymphoma patients under Taiwan's national health insurance system

RATIONALE, AIM AND OBJECTIVE

The beneficial effects of granulocyte colony-stimulating factor (G-CSF) prophylaxis on reducing the risk of chemotherapy-induced febrile neutropenia (CIFN) were well documented throughout the literature. However, existing data regarding its cost-effectiveness were conflicting. We estimated the cost-effectiveness of G-CSF prophylaxis in CIFN under Taiwan's National Health Insurance (NHI) system.

METHODS

Data on clinical outcomes and direct medical costs were derived for 5179 newly diagnosed breast cancer and 629 non-Hodgkin's lymphoma (NHL) patients from the NHI claims database. Patients were further categorized into three subgroups as "primary-", "secondary-" and "no -" prophylaxis based on their patterns of G-CSF use. Generalized estimating equations were applied to estimate the impact of G-CSF use on the incidence of CIFN. The incremental cost-effectiveness ratios of primary and secondary prophylactic G-CSF use were calculated and sensitivity analyses were performed.

RESULTS

Primary prophylaxis of G-CSF decreased the incidence of CIFN by 27% and 83%, while secondary prophylaxis by 34% and 22% in breast cancer and NHL patients, respectively. Compared with those with no prophylaxis, the incremental cost per CIFN reduced in primary prophylaxis is $931 and $52 among patients with breast cancer and NHL, respectively. In contrast, secondary prophylaxis is dominated by no prophylaxis and primary prophylaxis in both cancer patients.

CONCLUSION

Primary but not secondary prophylactic use of G-CSF was cost-effective in CIFN in breast cancer and NHL patients under Taiwan's NHI system.

Association between use of short-acting benzodiazepines and migraine occurrence: a nationwide population-based case-control study

AIM

To evaluate the association between using benzodiazepines (BZDs) with short- or long-acting durations and migraine occurrence.

METHODS

The migraine group comprised 9616 subjects older than 20 years and newly diagnosed with migraine between 2005 and 2011, and the comparison group comprised 38,464 subjects without migraine. The BZDs used in the subjects were dichotomously defined as short-acting (half-life ≤24 h) and long-acting substances. A logistic regression model was used to calculate the odds ratio (OR) of migraine associated with BZD exposure and other diseases.

RESULTS

The adjusted OR of migraine associated with BZD exposure was 1.73 (95% confidence interval [CI] = 1.63-1.84). Either exposure to a short-acting BZD alone or using it combining with a long-acting BZD had significant higher risks of migraine (adjusted OR = 1.69, 95% CI = 1.59-1.80; adjusted OR = 2.06, 95% CI = 1.91-2.24, respectively), whereas only long-acting BZD use was not associated with an increase of migraine. Meanwhile, sleep disorders, anxiety, and stroke were strongly associated with migraine (adjusted OR = 2.00, 1.91, and 1.57, respectively).

CONCLUSIONS

We observed a significant increase of migraine occurrence in subjects using short-acting BZDs, either alone or in combination with long-acting ones.

Association between schizophrenia and urinary incontinence: A population-based study

BACKGROUND

Urinary incontinence (UI) is defined as the involuntary loss of urine and can constitute a hygiene or social problem for affected patients. Whether the UI incidence is higher in patients with schizophrenia than in the general population remains undetermined.

METHODS

The patients were selected from the Taiwan National Health Insurance Research Database, using claims data from 2005 to 2011. We compared the risk of UI between schizophrenia and nonschizophrenia groups. Logistic regression models were employed for analyzing the risk of UI after adjustment for sex, age, and comorbidities.

RESULTS

The adjusted odds ratio for UI in the schizophrenia group was 1.78-fold higher than that in the nonschizophrenia group. Furthermore, male patients with schizophrenia were more susceptible to UI than their female counterparts, and a higher risk of UI was observed among younger patients with schizophrenia.

CONCLUSION

Compared with the general population, the risk of UI was higher among the patients with schizophrenia. Early diagnosis and intervention can increase the quality of life among patients with schizophrenia.

Acute Anterior Uveitis as a Risk Factor of Ankylosing Spondylitis-A National Population-Based Study

Introduction: In clinical settings, acute anterior uveitis (AAU) could be the first presentation of ankylosing spondylitis (AS). Based on this hypothesis, we investigate whether AAU is a risk factor in developing AS later by using National Health Insurance Research Database (NHIRD) in Taiwan. Materials and Methods: This cohort comparison study used longitudinal Taiwanese NHIRD to probe the relative risk odds of AAU for AS development, and consisted of all patients diagnosed with AAU (n = 5621) (ICD-9-CM codes 364.00). The relative risks of AS between AAU patients and controls were compared by estimating the crude hazard ratio with logistic regression. Kaplan–Meier analysis was used to calculate the cumulative incidence rates of developing AS, and a log-rank test was used to analyze the differences between the survival curves. Separate Cox proportional hazard regressions were performed to compute the AS-free rate after adjusting for possible confounding factors such as age and sex. Results: The crude hazard ratio was 2.667 for the AAU group, and the adjusted hazard ratio was 2.705 for the AAU group. The observation time of the AS-free group was shorter for AAU patients compared with the control group (1507 versus 1578 days). Moreover, in the AAU patients, the younger age onset of AAU (less than 30 years old here) would lead to an earlier diagnosis of AS later with a median of 1445.5 (742–2241) versus 1544 (819–2289) days of survival for the group of age onset of AAU greater than 30 years old. The difference is statistically significant (p < 0.05). Conclusions: AAU was a risk factor for AS. To identify AAU as an extra-articular manifestation is crucial for early diagnosis and treatment of AS and containing functional loss accordingly.

The association between comorbid anxiety disorders and the risk of stroke among patients with diabetes: An 11-year population-based retrospective cohort study

BACKGROUND

Diabetes and anxiety disorders are independent risk factors for stroke. However, it remains unclear whether the risk of stroke is higher among diabetic patients with comorbid anxiety than without comorbid anxiety. Therefore, the purpose of this study was to investigate the association between comorbid anxiety and the risk of stroke among patients with diabetes.

METHODS

This is a retrospective cohort study. We used the National Health Insurance Research Database in Taiwan to identify a diabetes cohort with a new diagnosis of an anxiety disorder but without a history of stroke. The enrollment period was 2001-2006 with up to 11 years of follow-up data. Comorbid anxiety was defined by both a clinical diagnosis of the DSM-IV (ICD-9-CM) and prescriptions for anxiolytic medications. Propensity score matching was performed to balance the selected confounders between the anxiety-exposed group and anxiety non-exposed group. Cox-propositional hazard regression models were used to evaluate the association between comorbid anxiety and the risk of stroke.

RESULTS

Among patients with diabetes (N=40,846), an estimated 5.8% (N=2374) of patients had comorbid anxiety disorders. Diabetic patients with comorbid anxiety were significantly associated with a higher risk of stroke compared to patients without comorbid anxiety (hazard ratio: 1.33, 95% confidence interval: 1.02-1.72).

LIMITATIONS

The severity of anxiety or diabetes could not be measured from the claims data. Residual confounding may still exist.

CONCLUSION

A significantly elevated risk of stroke was observed in association with comorbid anxiety among patients with diabetes. Psychiatrists should consider routine screening for anxiety disorders to prevent a stroke event among patients with diabetes.

Statin Use and the Risk for Incident Diabetes Mellitus in Patients with Acute Coronary Syndrome after Percutaneous Coronary Intervention: A Population-Based Retrospective Cohort Study in Taiwan

OBJECTIVES

The purpose of this study was to examine the association between statin use by individuals and the risk for incident diabetes mellitus in patients with acute coronary syndrome (ACS) following percutaneous coronary intervention (PCI).

METHODS

We conducted a retrospective cohort study of patients who were hospitalized for ACS between January 1, 2006, and December 31, 2010, and who had undergone PCI (n=30,665); the data were retrieved from the Taiwan National Health Insurance Research Database. A propensity score technique was used to establish a 1:1 matched cohort for statin users and non-statin users (n=9043 for each group). The risk for incident diabetes mellitus in statin users compared to non-statin users for patients with ACS after PCI was estimated by the multivariable Cox proportional hazards regression model.

RESULTS

Statin use was associated with a significant increase of 27% in the risk for new-onset diabetes mellitus (adjusted hazard ratio [HR] 1.27, 95% CI 1.14 to 1.41) compared to non-statin use in the matched cohort. The matched cohort analysis indicated that almost all individual statins were associated with a statistically significant increase in the risk for new-onset diabetes mellitus compared to those without statin use.

CONCLUSIONS

Our study indicated an association between increased risk for new-onset diabetes mellitus and statin use. Because the benefits of statins in prevention of morbidity and mortality in patients with ACS are well-established, clinical decision making should not be changed for patients with existing cardiovascular disease in whom statin therapy is recommended.

Decision to adopt medical technology under the National Health Insurance System in Taiwan: case study of new molecular targeted drugs among non-small cell lung cancer patients

RATIONALE, AIMS AND OBJECTIVES

Molecular targeted drugs (MTD), gefitinib and erlotinib, have been proven to provide clinical benefits to advanced stage non-small cell lung cancer (NSCLC) patients. Therefore, access to such medical innovations in time is critical for patients with the right indications. The aim of this study was to explore determinants of the adoption of MTD among NSCLC patients under Taiwan's National Health Insurance (NHI) system.

METHODS

Using Taiwan's Longitudinal Health Insurance Database and Cancer Registry database as the data source, we identified 1555 newly diagnosed NSCLC patients who started their cancer treatment between 2004 and 2009. Patients were categorized into 'non-MTD' and 'MTD' groups based on the cancer treatment they received. Hierarchical logistic regression was used to explore potential determinants associated with the adoption of MTD by adjusting for patient-, disease-, doctor-, hospital-level characteristics and changes of reimbursement schemes.

RESULTS

During the study period, 562 (36%) NSCLC patients were classified as the 'MTD group and 993 (64%) patients were the 'non-MTD' group. Hierarchical logistic regression model showed that the elderly were less likely to receive MTD (OR = 0.41; 95% CI: 0.28-0.60). In contrast, patients whose cancer type was adenocarcinoma (OR = 2.99; 95% CI: 2.09-4.26) were more likely to receive MTD. Those who went to private hospitals (OR = 1.75; 95% CI: 1.01-3.03) and hospitals with higher economic scale (OR = 3.00; 95% CI: 1.72-5.25) were more likely to receive MTD as well.

CONCLUSIONS

Our findings suggest that both patient- and hospital-level characteristics significantly affected the adoption of MTD among NSCLC patients.

Long-term risk of pancreatitis and diabetes after cholecystectomy in patients with cholelithiasis but no pancreatitis history: a 13-year follow-up study

BACKGROUND & AIM

Patients with biliary pancreatitis are suggested to undergo cholecystectomy to prevent the recurrence of pancreatitis. However, it remains controversial whether cholecystectomy is associated with reduced risks of pancreatitis and diabetes in patients with cholelithiasis and no history of pancreatitis.

METHODS

From Taiwan's National Health Insurance Research Database, we identified the following cohorts and analyzed the long-term risks of pancreatitis and diabetes in each cohort: 1) cholecystectomy cohort: cholelithiasis patients who had no history of pancreatitis and diabetes and underwent cholecystectomy; and 2) comparison cohort: cholelithiasis patients who had no history of pancreatitis and diabetes and did not undergo cholecystectomy.

RESULTS

The cholecystectomy group and the comparison group had similar distributions of age, sex, and comorbidities, except for hyperlipidemia. The proportion of patients in the cholecystectomy group who underwent endoscopic cholangiographic procedures was higher than that in the comparison group. Cholecystectomy was associated with a reduced risk of pancreatitis (adjusted hazard ratio [HR], 0.49; 95% confidence interval [CI], 0.36-0.68). Age-specific analyses showed that pancreatitis risk was decreased in patients younger than 50 and older than 65years. Both men and women exhibited reduced risks of pancreatitis after cholecystectomy. However, cholecystectomy was not associated with changes in the risk for diabetes.

CONCLUSION

Cholecystectomy for cholelithiasis is associated with a reduced risk of pancreatitis, but not of diabetes, in patients without previous history of pancreatitis and diabetes.

Prescription patterns of Chinese herbal products for patients with sleep disorder and major depressive disorder in Taiwan

ETHNOPHARMACOLOGICAL RELEVANCE

Chinese herbal products (CHPs) are commonly prescribed for sleep disorder and major depressive disorder (MDD). The aim of this study was to investigate the prescription patterns of CHPs and Western medicine for patients with these disorders in Taiwan, and analyze the frequency of using single herbs (SHs) and herbal formulas (HFs).

MATERIALS AND METHODS

In this retrospective population-based study secondary data analysis was performed using data from Taiwan's Longitudinal Health Insurance Database (LHID) between January 2007 and December 2011. In total, 1000,000 beneficiaries from the LHID were randomly selected from the 2010 registry for beneficiaries of the National Health Insurance Research Database. Patients with sleep disorder and MDD according to the International Classification of Diseases, Ninth Revision, Clinical Modification codes 307.40 and 311, respectively.

RESULTS

Among a total of 11,030 patients with sleep disorder, 9619 used Western medicine, 1334 used CHPs, and 77 used both, Among a total of 11,571 patients with MDD, 11,389 used Western medicine, 131 used CHPs, and 51 used both. Regardless of disorder type, women were predominant The majority of the patients were aged 22-44 years, had a monthly income of NT$17,281-NT$22,800, and lived in an area with Level 1 and Level 2 urbanization. Of the patients with sleep disorder, 1411 had used CHPs and visited a clinic 5298 times on average. Of the patients with MDD, 182 had used CHPs and visited a clinic 755 times on average. The three most commonly used SHs and HFs were Ziziphi Spinosae Semen, Polygoni Multiflori Caulis, and Polygalae Radix, and Jia-Wei-Xiao-Yao-San, Suan-Zao-Ren-Tang, and Chai-Hu-Chia-Lung-Ku-Mu-Li-Tang, respectively.

CONCLUSION

Chinese herbal products including SHs and HFs are prescribed for patients with sleep disorder and MDD. However, the efficacy and safety of CHPs for sleep disorder and MDD need to be further evaluated.

The epidemiology and progression time from transient to permanent psychiatric disorders of substance-induced psychosis in Taiwan

INTRODUCTION

Substance-induced psychosis (SIP), including alcohol-induced psychotic disorder (AIPD) and substance-induced psychotic disorder (SIPD), is gradually increasing in importance in clinical practice. However, few studies have investigated the epidemiology and progression time from transient to permanent psychiatric disorders for AIPD and SIPD patients.

METHODS

We utilized the National Health Insurance Research Database (NHIRD) to investigate the incidence and prevalence of AIPD and SIPD in Taiwan and determined the timing of AIPD or SIPD followed by the development of persistent psychotic conditions.

RESULTS

The average incidence and prevalence were 1.97 and 2.94 per 100,000 person-years for AIPD, 3.09 and 5.67 per 100,000 person-years for SIPD in Taiwan. Moreover, 10.9% to 24.3% of subjects with either AIPD or SIPD had a change in diagnosis to either schizophrenia or affective disorder, and ~50% of patients had a psychotic or affective transformation in their first year after AIPD and SIPD diagnoses. The mean progression time of psychotic or affective transformation was 1.9 to 2.7 years.

CONCLUSIONS

SIP is a predictive factor for persistent psychotic and affective transformation, and a three-year follow-up may be an optimal clinical practice to prevent psychotic or affective transformation in 60% of patients.

The Prevalence of Dry Eye Syndrome's and the Likelihood to Develop Sjögren's Syndrome in Taiwan: A Population-Based Study

BACKGROUND

Dry eye syndrome (DES) is one of the key clinical features and possibly an early clinical presentation of Sjögren's syndrome (SS). We explore DES prevalence and assess the likelihood of DES patients to develop SS in Taiwan through the National Health Insurance Research Database (NHIRD).

METHODS

Through a cohort comparison study, longitudinal data from the NHIRD (2000 to 2008) in Taiwan was used to probe the prevalence of DES and the odds that DES patients would later develop SS.

RESULTS

The prevalence of DES in the present study is 4.87%. The incidence rates of developing SS were 4.8% for the DES group and 1.5% for comparison group. The median age and interquartile range of DES and comparison patients was 49.8 (10) and 48.7 (15) years old, respectively. The crude hazard ratio (with 95% confidence interval) for DES patients to develop SS was 3.13 (3.10-3.50) for the DES group, and the adjusted hazard ratio (with 95% confidence interval) was 3.64 (3.43-3.87). The observation period and interquartile range for DES and comparison patients to develop SS later were 1418 (781-2316) versus 1641 (971-2512) days respectively.

CONCLUSIONS

DES patients carried a higher risk for developing SS (hazard ratio 3.13) and presented for SS 3.88 years earlier than comparison group patients in this study.

Changes in prevalence of diabetic complications and associated healthcare costs during a 10-year follow-up period among a nationwide diabetic cohort

OBJECTIVE

Long-term health and economic consequences of diabetes mellitus are of significant importance to health policy makers to identify the most efficient interventions for disease managements. However, existing data are mainly from simulation models instead of "real-world" data. The objective of this study was to longitudinally evaluate the changes of prevalence of diabetic complications and associated healthcare costs in a nationally-representative diabetic cohort.

METHODS

We used the 2000-2011 Taiwan's Longitudinal Health Insurance Database (LHID) to conduct a population-based cohort study of 136,372 patients with type 2 diabetes. Diabetic complications of each patient were calculated annually after the cohort entry by the adapted Diabetes Complications Severity Index (aDCSI) score (sum of diabetic complication with severity levels, range 0-13) using diagnostic codes recorded in the LHID. Study subjects were further categorized into six subgroups according to their aDCSI score (0, 1, 2, 3, 4, 5+) at cohort entry. Healthcare utilizations (including outpatient and inpatient visits) as well as direct medical costs for the six subgroups were estimated annually using patient-level data from the LHID.

RESULTS

We found the severity of diabetic complications increased over time, especially for patients with aDCSI score of 2 and above at cohort entry (at 10years of follow-up: aDCSI=0 (cohort entry), 2.37; aDCSI=1, 3.59; aDCSI=2, 4.60; aDCSI=3, 5.14; aDCSI=4, 5.96). There were significant differences in healthcare utilizations and associated medical costs among patients stratified by aDCSI score (e.g. at 1year after cohort entry, mean counts of inpatient visits: 0.14 vs. 1.81 for aDCSI=0 vs.5+). Relatively high healthcare utilizations and associated medical costs in the first year of cohort entry were observed for patients with aDCSI score of 4 and above at cohort entry.

CONCLUSIONS

We provided the important empirical data for patient-level longitudinal changes in diabetic complications and associated healthcare utilization and medical costs among patients with diabetes.

The association of healthcare seeking behavior for anxiety and depression among patients with lower urinary tract symptoms: a nationwide population-based study

Lower urinary tract symptoms (LUTS) have a negative impact on the quality of life, and may relate to anxiety and depression. The objective of this study was to test the hypothesis that LUTS are associated with anxiety and depression using a nationwide population-based database in Taiwan. Data were obtained from a random population sample of about one million enrollees in the National Health Insurance program from 2001 to 2009, and consisted of 22,980 LUTS patients and 45,960 matched controls. The records of healthcare seeking for anxiety and depression were collected 2 years before and after the diagnosis of LUTS. The results showed that patients with LUTS had a significantly higher prevalence of anxiety or depression than the matched controls (11.45% vs. 5.72%). After controlling for sociodemographic variables and other major systemic diseases, the odds ratios for anxiety, depression, either anxiety or depression, and both anxiety and depression, were 2.05, 2.19, 2.14, and 2.56, respectively. There was an association between LUTS and the stress-related common mental disorders, and there seemed to be an additive effect of anxiety and depression on the association with LUTS. These findings imply a psychological role in the pathogenesis or sequelae of LUTS.

The risks of major osteoporotic fractures in patients with schizophrenia: a population-based 10-year follow-up study

OBJECTIVE

The aim of the study is to explore the incidence and the risks associated with major osteoporotic fractures, all-cause mortality with osteoporotic fractures and the effect of the psychiatric drug exposure in patients with schizophrenia during a 10-year follow-up period.

METHODS

Two nationwide cohorts were selected from the Taiwan National Health Insurance Research Database (NHIRD) consisting of 30,335 patients with schizophrenia (age ≥ 40 years) and 121,340 age- and sex-matched control participants without schizophrenia. The psychiatric proportion of days covered (PDC) is an indicator of the intensity of drug exposure in patients with schizophrenia. The incidence and risk factors of major osteoporotic fractures were calculated for both cohorts. Additionally, the patient survival rate after major osteoporotic fractures was also calculated.

RESULTS

During a 10-year follow-up period, 1677 (5.53%) schizophrenia and 4257 (3.51%) control subjects had major osteoporotic fractures (P < 0.001). The schizophrenia patients with a PDC > 0.1 showed a significantly higher incidence of major osteoporotic fractures than did the non-schizophrenia controls; however, those with a psychiatric PDC ≤ 0.1 did not. After adjustment, the psychiatric PDC was significantly and independently associated with the risk of major osteoporotic fractures except some medical morbidities but the schizophrenia diagnosis was not. In addition, among all 5934 patients with major osteoporotic fracture, the adjusted mortality hazard ratio for psychiatric PDC was 1.92 (95% CI = 1.63-2.26).

CONCLUSIONS

Patients with schizophrenia are at a higher risk for major osteoporotic fractures than the general population and also have a higher mortality rate due to major osteoporotic fractures. These findings may be caused by psychiatric drug use rather than schizophrenia, which suggests that directions can be taken in future studies.

Risk of hospitalization and healthcare cost associated with Diabetes Complication Severity Index in Taiwan's National Health Insurance Research Database

OBJECTIVE

The aim of this study is to test the validity of adapted Diabetes Complication Severity Index (aDCSI) in predicting the risk of hospitalization and healthcare cost in type 2 diabetic patients using a nationally-representative claims database.

STUDY DESIGN

Retrospective cohort study used 4years of claims data from Taiwan's National Health Insurance Research Database (NHIRD).

METHODS

Type 2 diabetic patients who had 4-years of enrollment were identified as study subjects (N=136,372). The aDCSI score (sum of diabetic complication with severity levels, range 0-13) and complication count (sum of diabetic complications, range 0-7) were generated using diagnostic codes for each patient. Poisson model and linear regression model were conducted to predict risk of hospitalization and healthcare costs associated with aDCSI score and count of diabetic complications.

RESULTS

The aDCSI score (risk ratio 1.51 to 10.32 categorically, and 1.41 linearly) and count of diabetic complications (risk ratio 1.56 to 12.20 categorically, and 1.66 linearly) were significantly positively associated with risk of hospitalization. A one-point increase in the aDCSI score was positively associated with increased healthcare costs.

CONCLUSIONS

The performance of aDCSI in predicting risk of hospitalization and healthcare cost in the nationally-representative claims database is similar to those reported in the original study. It may serve as an efficient tool for stratifying type 2 diabetic patients for disease management programs and population-based studies.

Hyperlipidemia, statin use and the risk of developing depression: a nationwide retrospective cohort study

OBJECTIVE

Depression is a highly prevalent disorder that is associated with disability. The aim of this study was to determine the relationship between depression and hyperlipidemia and whether the onset of depression is associated with administering statins to patients with hyperlipidemia.

MATERIAL AND METHODS

The data analyzed in this study were retrieved from the National Health Insurance Research Database in Taiwan. We identified newly diagnosed hyperlipidemia in 26,852 patients without a history of depression as the exposure group in the period of 2000-2002, and a comparison group comprised 107,408 patients. The differences between the exposure group and the comparison group were examined using a chi-square test to calculate categorical variables. The hazard ratio and the 95% confidence interval for depression were used in the logistic regression.

RESULTS

The hyperlipidemia patients demonstrated a high risk for depression and comorbidities, such as hypertension, diabetes and sleep disorder, which indicated synergistic effects related to a high risk of depression in hyperlipidemia patients. Hyperlipidemia patients who had received statins exhibited a lower risk of depression than did those who had not received statins.

CONCLUSION

Our results suggested that hyperlipidemia increases the risk of depression and that using statins is associated with a decreased risk of depression in patients with hyperlipidemia.

Statins and the risk of liver injury: a population-based case-control study

OBJECTIVE

This case-control study investigated the association between statin use and liver injury using Taiwan's National Health Insurance Research Database.

METHODS

Our study subjects included 4165 cases (patients who had been admitted with a primary diagnosis of liver injury between 2002 and 2009) and 16 660 age-matched, sex-matched and index date-matched controls. Multivariable conditional regression models were used to estimate the association between statin use and liver injury.

RESULTS

Users of statins were not associated with risk of liver injury (adjusted odds ratio [aOR] 1.04; 95% confidence interval [0.90-1.19]) when compared with nonusers. Nevertheless, a higher dose of statin (≥1 defined daily dose; aOR 1.55 [1.14-2.11]) and use of rosuvastatin before event of liver injury (aOR 1.38 [1.03-1.85]) were significantly associated with liver injury.

CONCLUSIONS

This population-based study extends previous evidence by exploring the potential association between statins use and risk of liver injury. Overall, we found that statin was not associated with risk of liver injury. Nevertheless, special concern should be paid to those who used statin ≥1 defined daily dose and rosuvastatin.

Risk of accidental injuries amongst Parkinson disease patients

BACKGROUND AND PURPOSE

To investigate the spectrum and risks of accidental injuries (AIs) amongst Parkinson disease (PD) patients.

METHODS

The participants comprised PD patients aged 50 years and older who were initially diagnosed between 2000 and 2009, and a comparison group of non-PD patients. The incidence rates of accidental injury types amongst PD and non-PD patients were calculated; hazard ratios were calculated and adjusted for comorbidities, using 95% confidence intervals (CIs) of developing such outcomes in PD patients.

RESULTS

In total, 4046 PD patients and 16 184 non-PD patients were followed over time. The PD patients demonstrated the following incidence rates and hazard ratios in comparison to the control cohort for accidental injuries: all injuries, 19.78 per 100 person-years (100 PYs), adjusted hazard ratio (HR) 1.30 (95% CI 1.24-1.36); head injury, 2.95 per 100 PYs, HR 1.88 (95% CI 1.64-2.15); bone fracture and dislocation, 4.61 per 100 PYs, HR 1.39 (95% CI 1.25-1.54); burns, 0.66 per 100 PYs, HR 1.01 (95% CI 0.78-1.32); injury to spinal cord, plexus and nerves, 0.15 per 100 PYs, HR 1.25 (95% CI 0.72-2.17); superficial injuries and contusions, 11.41 per 100 PYs, HR 1.20 (95% CI 1.12-1.27). The injury risk for the 69-79 years age group in PD compared with controls of the same age (HR 1.38) was significantly higher compared with that of the 50-69 age groups in PD and controls (HR 1.16).

CONCLUSIONS

Parkinson disease patients demonstrate a significantly elevated risk of developing all accidental injury types except burn injuries and injuries to spinal cord, plexus and nerves, compared with age-matched controls. The risk increases as age increases.

A nationwide population study of trazodone use in urology patients

BACKGROUND

Erectile dysfunction (ED) has been a prevalent disease worldwide. The mainstay of treatment for ED focused on oral medications such as phosphodiesterase-5 (PDE-5) inhibitors or other medical aids with variable acceptance. Trazodone is a second-generation antidepressant approved by the U.S. Food and Drug Administration in 1981. Some evidence has suggested that trazodone may be helpful in improving ED, especially for problems dealing with male arousal. Although controversial, trazodone can sometimes be used for ED as an off-label medication for urological patients. By using a nationwide health insurance database, we attempted to estimate and thereafter analyze the existence and extent of off-label use of trazodone for ED in Taiwan.

METHODS

The 1/500 randomly sampled outpatient visits dataset and the 1/1 million randomly selected dataset issued in 2000 in the National Health Insurance Research Database were used to estimate the total number of visits and the urological visits involving trazodone prescriptions. The refill rate, patient age at first prescription, and the diagnoses assigned to the prescription visits were also gathered and analyzed.

RESULTS

The prescription visits in urological clinics consisted exclusively of male patients (99.0 ± 1.3%, p < 0.001). The use of trazodone was more prevalent and regular among the elderly (p < 0.001), which happens to parallel the trend of incidence rates of ED. Starting in 1998, the prescription rate for trazodone increased rapidly, but then slowed, equalized, and topped out in 2003. One of the reasons for this marked change in prescriptions rates in Taiwan may be the increased awareness of ED as a treatable disease under the promotion and marketing of three PDE-5 inhibitors. ED-related diagnoses assigned to the prescription visits accounted for 55% of total visits, whereas anxiety/depression-associated diagnoses were the bases of less than 13%, and those attributed to insomnia accounted for 2.3% of visits.

CONCLUSION

Despite the fact that trazodone was not officially approved for patients with urological conditions, this medication has been used as an ED treatment in Taiwan at least since 1997; the prescriptions of trazodone were more prevalent among the elderly, and this trend increased, then evened out, and eventually reached its maximum point in 2003, correlating with the introduction of PDE-5 inhibitors.