Age-period-cohort analysis of asthma prevalence among school children.

Prevalence of age-dependent diseases such as asthma is confounded not only by aging effects but also by cohort and period effects. Age-period-cohort (APC) analysis is commonly performed to isolate the effects of these three factors from two-way tables of prevalence by age and birth cohort. However, APC analysis suffers from technical difficulties such as non-identifiability problems. We isolated the effects of these three factors in a step-by-step manner by analyzing Japan's school health data collected from 1984 to 2004 focusing on asthma prevalence among school children aged 6-17 years consisting of 30 birth cohorts (entering classes). We verified the accuracy of our method showing high agreement of the observed age-, period- and cohort-specific data and the data predicted by our method. The aging effects were found to follow cubic equations whose multinomial coefficients were determined by an optimization technique. The obtained aging effect curves of age-specific asthma prevalence showed that boys reach the peak prevalence at 13 and girls at 14, declining markedly afterward. The cohort effects, defined as the arithmetic asthma prevalence means for ages 6-17 years, showed consistent upward trends for the 30 birth cohorts born in 1968-97 for both sexes. The period effects showed a consistent decline since 1984 but abruptly increased in 1999 and then declined again. We were not able to identify the exact cause of the increase in 1999, therefore, this should be examined in the future studies. Because the cohort effects show no sign of leveling off yet, asthma prevalence will likely increase in the foreseeable future.

Estimation of disease-specific costs in health insurance claims: a comparison of three methods.

OBJECTIVE: To compare the accuracy and validity of three different methods (Proportional Disease Magnitude method [PDM] with two different magnitude estimations: arithmetic means with correction by the authors; Proportional Allotment Estimator [PAE] by Tango; Maximum Likelihood Estimator [MLE] also by Tango) for estimating disease-specific costs in health insurance claims. METHODS: Application of the three methods to a computer-generated simulation dataset whose disease-specific costs were known and to actual outpatient claims whose disease-specific costs were unknown. OUTCOME MEASURES: For simulation data, the accuracy was assessed by correlation between known disease-specific costs and estimated disease-specific costs by the three methods. For actual claims, concurrent validity was assessed by inter-method correlations between pairs of the two methods. RESULTS: All three methods showed good agreement and accuracy with the simulation data but marked disagreement when they applied to actual claims. MLE yielded an aggregate total of disease-specific costs exceeding the actual total by 21.3% and showed negative disease-specific costs in 18 out of 154 categories. Inter-method correlations showed that PDM with PAE and MLE correlated most strongly (R2 = 0.9022) while the least correlation was observed for PDM with arithmetic means and MLE (R2 = 0.6861). CONCLUSION: MLE is not usable for claims analysis but PDM yielded good estimates with two different methods of magnitude estimation using actual claims.

[Estimation of disease-specific costs in a dataset of health insurance claims and its validation using simulation data].

PURPOSES: To estimate disease-specific costs in a dataset of health insurance claims with multiple diagnoses with known aggregate cost per claim and unknown disease-specific cost of each diagnosis using PDM (Proportional Disease Magnitude) method, validate its accuracy using simulation data with Monte Carlo method and improve its accuracy by developing an adjustment formula. METHODS: Developed simulation data with pre-assigned disease-specific costs, applied PDM method using arithmetic means of per-diem-per-disease cost as magnitude, validated its accuracy by observing the correlation between estimates by PDM method and known disease-specific costs and formulated an adjustment formula to improve accuracy. The reproducibility of the findings was assessed using Monte Carlo method by repeating the same procedures. RESULTS: The observed arithmetic means of per-diem-per-disease cost did not match well with actual values resulting in unsatisfactory accuracy. However, when the observed means were adjusted with a formula in which the observed mean is multiplied by (observed mean/overall mean) in the power of 2, PDM method yielded an accurate estimate of disease-specific cost. The accuracy was reproduced by Monte Carlo method with 0.9 or above R square value and slope of regression line in 76, 56 out of 100 iterations respectively. CONCLUSIONS: PDM method proved to be an objective, reproducible and accurate method for estimation of disease-specific costs of health insurance claims.