Improvement plan for health care benefits of asthma (National Health Insurance, Taiwan)
In September 2004, National Taiwan University Hospital joined this program. The program introduces incentives to improve the quality of asthma care, encourages hospitals to engage in patient-centered complete asthma care, improves compliance with guidelines, strengthens follow-up visits and health-education services for asthmatic patients, and ensures the provision of complete and continuous care.
Inclusion criteria
The enrolled patients were diagnosed with asthma (ICD-9-CM 493 or ICD-10-CM J44-J45) within 90 days before enrollment by the same physician and had completed at least two outpatient visits to that physician. The patients were willing to comply with asthma care regulations and with follow-up contacts by our asthma case managers.
End of enrollment
- Completion of self-care and self-evaluation (disease resolved).
- Refusal by the patient of further contact by asthma case managers and/or follow-up by asthma specialists.
- Loss of contact and/or follow-up for > 90 days.
- Change in place of residence.
- Death.
- Noncompliance with the program regulations for > 1 year.
- Others (e.g., transfer to other care provider, change in diagnosis).
These patients continued to receive medications and treatments for their asthma as indicated by their asthma specialist and guided by the Global Initiative for Asthma (GINA) (20). These patients were required to receive care and follow-up from their physician for at least 3 months. The data were collected at baseline (total serum IgE and blood eosinophil percentage) and clinical data were collected by asthma case managers during interviews and follow-up visits.
These enrolled patients were provided comprehensive information on asthma care, including the pathophysiology of asthma, how to diagnose asthma, asthma severity, environmental allergens and control, management of acute attacks, long-term control of asthma, asthma relievers and controllers, and the importance of monitoring PEF. We instructed the patients in the correct method of inhalation, the timing of medications, and measurement of the peak PEF at home.
The following parameters were recorded at each outpatient visit: (1) tracking/visit date, (2) frequency of daytime symptoms, (3) frequency of nighttime symptoms, (4) predicted PEF (best %) (5) PEF variability (%), (6) asthma severity assessment, (7) ACT/cACT score (21), and (8) asthma control. The asthma severity classification, asthma control classification, and PEF variability and predicted percentage were defined according to the Guidelines for the Diagnosis and Management of Asthma (EPR-3)(22). Children with uncontrolled asthma according to the GINA criteria are likely to be clinically different from those included in studies using a c-ACT or ACT score of < 20 as a criterion to define uncontrolled asthma (23). Therefore, we used both the cACT/ACT score and the GINA criteria to evaluate the level of asthma control in children.
Patients
From September 1, 2004 to December 31, 2018 we registered 1115 eligible children. Among them, 383 children with information on total serum IgE and peripheral blood eosinophil percentage (independent variables) were enrolled. Another 313 children with only peripheral blood eosinophil percentage (independent variables) were also enrolled. PEF variability, predicted PEF (best percentage), asthma severity, ACT score, and asthma control during the follow-up of these 692 patients were collected as phenotype outcomes (dependent variables). Asthma and rhinitis were optimally managed by study physicians according to the applicable guidelines.
Descriptive statistics were calculated for demographic and baseline characteristics. IgE values in excess of 5000 IU/mL were assigned a value of 5000 IU/mL. The geometric mean (GM) and the arithmetic mean (AM) were calculated to approximate the normal distribution for statistical inference and modeling. Patients were classified as allergic if their total serum IgE level was ³ 150 IU/mL and as nonallergic if their total serum IgE level was < 150 IU/mL (7).
Sensitization
Thirty six allergen-specific IgE (sIgEs) were grouped into the following seven categories: dust mites (house dust mite, farinae mite, pterony mite), danders/feathers (chicken feather, cat, dog), molds (Alternaria, Aspergillus, Cladosporium, Penicillium), grasses/trees (Bermuda grass, Willow black, Eucalyptus, Cedar Japan, mulberry mix, pigweed mix, ragweed mix, Timothy grass), foods (avocado, pork, beef, milk, cheddar cheese, shrimp, crab, clam, cod, tuna, peanut, soybean, wheat, brewer yeast, egg yolk, egg white), cockroach mix, and latex.
Peak expiratory flow (PEF)
We measured PEF variability and predicted percentage according to the Guidelines for the Diagnosis and Management of Asthma (EPR-3)(22). Individual peak flow measurements are highly variable and the PEF variability has greater predictive power for future exacerbations than individual PEF measurements (24).
Sequencing of miRNAs with RNA-seq
Peripheral blood samples were obtained from four nonallergic and five allergic asthmatic patients. Total RNA was extracted from peripheral blood white blood cells using the miRNeasy Extraction Kit (Qiagen) according to the manufacturer’s protocol. Sequencing was performed using high-quality RNA with RNA integrity number (RIN) ≥ 7. A total of 1.2 µg of total RNA per sample was used as input material for the small RNA library. Sequencing libraries were generated using the TruSeq Small RNA Library Prep Kit (Illumina) following the manufacturer’s recommendations and index codes were added to attribute sequences to the samples. Briefly, 3'- and 5'-adapters were specifically ligated to the 3'- and 5'-ends of small RNAs. Next, first-strand cDNA was synthesized using SuperScript II Reverse Transcriptase. PCR amplification was performed using 2´ PCR Master Mix and the PCR products were resolved in a BluePippin 3% agarose gel. DNA fragments of 120–160 bp were recovered and dissolved in 15 μL of double-distilled water. Library quality was assessed using the Agilent Bioanalyzer 2100 system and DNA High-Sensitivity Chips. The libraries were sequenced on the Illumina NextSeq 500 platform and 75-bp single-end reads were generated. MiRNA expression levels were expressed as RPM (reads per million). The heat map were generated by uploading the differential expressed miRNAs data to ClustVis (https://biit.cs.ut.ee/clustvis/).
Ingenuity pathway analysis
miRNA expression profiles were analyzed by ingenuity pathway analysis (IPA), using the features core analysis and pathway explore. Finally, the direct interaction pathway and the node and target molecules were overlaid with the differentially expressed miRNAs between allergic and nonallergic asthma to plot the network diagram.
Statistical analysis
Generalized estimating equations (GEEs) are important in the analysis of correlated data (25). These data sets can be generated in longitudinal studies, in which patients are measured at different points in time, or from clustering, in which measurements are taken of patients who share a common characteristic. We performed GEE analysis using the GENMOD and GEE procedures in SAS software. Missing data are common in longitudinal studies and can be caused by dropouts or skipped visits. Both procedures implement the standard GEE approach for longitudinal data; this approach is appropriate for complete data or when data are missing completely at random.
We first analyzed the association between independent variables and dependent variables (phenotype outcome). The independent variables were total serum IgE level (original value divided by 100), peripheral blood eosinophil percentage, and inhaled corticosteroid (ICS) dosage (low, medium, and high). Low, medium, and high daily doses of inhaled corticosteroids (ICS) of estimated comparability were defined by GINA (20). The dependent variables were predicted PEF, PEF variability, ACT score, asthma severity, and asthma control. All outcome variables were measured at different occasions (e.g., time points/visits), so time was considered a covariate of the independent variables (total serum IgE level and blood eosinophil percentage) to determine the time effect. That is, the numbers of time points/visits were conceptually equivalent to disease duration. The interaction terms were total serum IgE * time and blood eosinophil * time. We took the time factor into consideration and analyzed the association between independent variables and dependent variables under the interaction between time and the independent variables.
Odds ratio and beta coefficient
The odds ratio (OR) quantifies the strength of the association between an outcome (dependent variable) and a change (independent variable). An OR of < 1 indicates lower odds that an outcome is attributed to a of change; and an OR of > 1 indicates higher odds that an outcome is attributed to a change.
The beta coefficient is the degree of change in the dependent variable for each unit of change in the independent variable. The beta coefficient can be negative or positive. If the beta coefficient is not statistically significant, the independent variable is not significantly predictive of the dependent variable (outcome). If the beta coefficient is positive, for each unit of increase in the independent variable, the outcome will increase by the beta-coefficient value. If the beta coefficient is negative, for each unit increase in the independent variable, the outcome variable will decrease by the beta-coefficient value.
Longitudinal outcome pattern analysis
To evaluate the longitudinal trends of childhood asthma, we divided 383 patients into IgE < 150 IU/mL (n = 118), IgE 150–550 IU/mL (n = 133), and IgE > 550 IU/mL (n = 132) groups. Among these three group of patients, we selected patients with a follow-up duration of more than 5 years for pattern analysis. Locally weighted scatterplot smoothing (LOWESS) in Prism version 8 (GraphPad Software) enabled fitting of a curve without selecting a model (26). We plotted follow-up duration against asthma severity changing with time for the patients received followed-up for more than 5 years.