the date of discontinuation, disenrollment, switch , or end of the FTY720 study period . Patients were assigned to a persistency category based on length of treatment with ZOL . Study outcomes. Patient mortality was assessed during the postindex period. Occurrence of death was identified from facility claims based on patient discharge status for death on facility claims or a diagnosis code for death on a facility or physician claim. Service dates at >60 and >90 days after the death date were analyzed for patients with a death indicator to confirm the mortality algorithm. The time to death was calculated as the number of days between the index date and the death date. Because only one of the two health plan databases used for this study included death data, patient mortality could be determined for only a subset of the patient sample.
Fractures were identified from medical claims based on the presence of ICD 9 CM diagnosis codes or CPT procedure codes for surgery to the bone . The occurrence of SREs was also determined in the preindex period and postindex period, including Ramelteon 196597-26-9 fracture and surgery to the bone , as well as SCC or radiation to the bone . Statistical analysis. Unadjusted bivariate comparisons of baseline characteristics and outcome measures were performed, using appropriate tests based on the distribution of the measure. Because the length of follow up time varied, person time was used because it estimates the actual time atrisk that all persons contributed to the study. Incidence rates, allowing for variable follow up time, were calculated using Stata1 stptime .
Because of variability in the length of follow up periods, study outcomes were measured as risk per 100 person years. To test for trends across persistency categories, we examined whether there were buy Pimobendan significant differences in the SREs across persistency categories . This test is similar to the log rank test except that the ordinal ranking across persistency categories was considered. PAM persistence was not tracked in our study; therefore, no comparisons with ZOL were possible. Multivariate analysis adjusting for covariates was conducted using the Cox proportional hazards model. Confounding factors that were adjusted for include gender, age, preindex Charlson Quan comorbidity score, preindex evidence of oral bisphosphonate use, and preindex SREs. Analyses were conducted using version 10.
0 of Stata1 statistical software . Results Patients meeting the study selection criteria were assigned to either a ZOL cohort or a no bisphosphonate cohort . Patients who initiated treatment on PAM were not included purchase Cilostazol in the study; however, the patient population was similar in size to the ZOL cohort. A majority pericardium of patients in the sample, 64%, were assigned to the ZOL cohort, while 36% of patients were assigned to the no bisphosphonate cohort. In total, 1,655 patients were included in the study . Some differences in demographic and clinical characteristics were observed between patients in the ZOL cohort and the no bisphosphonate cohort . Patients in the ZOL cohort were on average slightly older , more often male , and had a longer average follow up period than patients in the no bisphosphonate cohort. Also, patients in the ZOL cohort were more likely to receive oral bisphosphonate therapy during earlier treatment than patients in the no bisphosphonate cohort. However, patients in the no bisphosphonate cohort had a higher average Charlson Quan comorbidity score and slightly higher.