Objective: Nosocomial infections are increasing. This study focuses on soft tissue infections, S aureus infections and MRSA nosocomial infections. Trends are quantified and analyzed for hospital and patient characteristics. Methods: Data from the HCUP National Inpatient Sample (NIS) of US hospital discharges for 1998 through 2004 were used to estimate national trends in hospitalizations for three infectious conditions. Discharges where ICD9-CMs were recorded for soft tissue infections (680.xx, 681.xx, and 682.xx), S aureus infections (038.11, 041.11, and 482.41), and MRSA (V09.0) were included in the study. NIS discharge weights were used to generate nationally representative estimates of discharges for each infection and for their combinations with MRSA. Logistic regression was used to obtain odds ratios (ORs) for each infection and MRSA combination (1=yes, 0=no) and the independent variables of patient age, race, gender, income, hospital bedsize, location (urban/rural), mission (teaching/non-teaching), and expected payment source. Results: From 1998 to 2004 discharges for soft tissue infections increased 47%, S aureus infections increased 28%, MRSA increased 223%, MRSA with S aureus increased 317%, and MRSA with soft tissue infection increased 608%. Regression results for the S aureus with MRSA combination in 1998 revealed that being white, and older increased the odds of this infection (ORs=1.160 and 1.020 respectively) while being female reduced the odds (OR=0.692). Larger hospitals had a lower odds (OR=0.918) as did teaching hospitals (OR=0.681). However urban hospitals odds were higher (OR=1.384) and being a Medicare or Medicaid patient also was associated with an increased odds of the infection (ORs=1.486 and 1.743 respectively). Only patient income was not significant (p<0.001 for all other variables). Results for 2004 were similar but not the same. Conclusion: Nosocomial infections continue to increase as a percent of hospital stays, and the odds of infection are not equal across patient and hospital types.