Analyses of Idiopathic Pulmonary Fibrosis (IPF) Inpatients in the United States

ABSTRACT majority of patients, regardless of treatment, within 3-5 years of diagnosis. 1-3 Pulmonary fibrosis is often the final common pathway of many known causes of interstitial lung disease, such as sarcoidosis, silicosis, drug reactions, infections and collagen vascular diseases. The overall goal of the project is to identify the factors and costs associated with IPF patients in terms of mortality, length of stay and costs in different types of clinical settings across the United States.

We used data rom the Nationwide Inpatient Sample (NIS) for calendar years 2007-2012 to perform a series of analyses to identify any factors and costs associated with IPF patients, particularly those pertaining to mortality, length of stay (LOS), and costs across the various clinical settings in the US. A series of parametric and non-parametric methods were used. Parametric methods included linear regression models, correlation analysis using Pearson correlation coefficients, paired and unpaired t-tests, and one-way ANOVA. A series of non-parametric methods were also used including: Wilcoxon rank sum tests, Mann-Whitney tests, Spearman correlations, and Kruskal-Wallis tests. Descriptive analyses were employed to compute sample mean, median, standard deviation, and interquartile ranges for study variables. Finally, binary outcomes and categorical variables were examined using the following: logistic regression models, chi-square tests, contingency coefficients, and McNemar’s test.

We found a mean expenditure per IPF patient of $68,925 (standard deviation = $16,4049), with a median of $28,257. We found the distribution of total charges to be significantly different from the normal distribution (p<0.001). Further investigation revealed a maximum charge of $4,162,849, which is approximately six times as high as the 99^th percentile total charge of and 147 times as high as the median. These results indicate that the distribution of total costs are skewed heavily by a relative few patients with extremely high bills. LOS also varied wildly with the mean stay being 7.84 days with a standard deviation of 13.60 days. LOS ranged from zero days to 1158 days, which corroborates our conclusion about the distribution of total costs. The patient who stayed 1158 days was hospitalized for 858 days more than any other patient. The next highest LOS values were 300 days, 250 days, 196 days, and 160 days, which further plays to our notion of the extremely expensive few. LOS varied in a statistically significant way between white and black patients (p<0.05).

Only 11.2% of all patients in the data set died during the study period (11.2% of whites, 9.5% of blacks, 11.2% Hispanics, 12.2% of Asian/Pacific Islander, 13.0% of Native Americans). Finally, only 12% of all IPF patients passed away, while slightly more (13.8%) of all IIP patients died during the study period.