Sas studio 5

Fumigant concentration mortality assays and time response assays were analyzed using probit regression SAS Studio^® 5.1 software [29 ]. Sorption and desorption mortality assay data were transformed with an angular transformation and tested for normality and individually analyzed using a one-way ANOVA with general linear mixed models (GLMMIX) [29 ]. Means were separated using Tukey’s HSD (honestly significant difference) post hoc test at p < 0.05.

The characteristics of the cohorts were described by frequencies for categorical variables and by median for continuous variables.
To identify variables associated with ‘dying at home’ we performed a multivariate logistic regression analysis to estimate odds ratios (OR) and 96% confidence intervals (CI). We made separate analyses for the three cohorts that were defined by the causes of death with two groups of variables included, variables on patient and variables on municipal level. Variables on the individual level included gender, age categorized in 10-year age bands from 50 to 89 years and with patients below 50 and above 90 years in separate groups, marital status (indicator of informal care), education (primary, secondary, and higher education) and the number of comorbidities (0, 1–2, 3–4, 5 and above). The variables on the municipal level included the person years of GPs and caring personnel in total, both normalized by 10,000 inhabitants.
We registered data on 15 comorbidities (see Appendix A) from up to two years before the death day. Comorbidities were generated from the registration of both primary and secondary diagnoses and from both hospital inpatient and outpatient stays as well as consultations with GPs registered in the Municipal Patient and User Register.
Data management and analyses were conducted in SAS Studio 5.1 (SAS Institute Inc., Cary, NC, USA).