obtain average total costs and QALYs per patient mim: mean total_costs total_QALYs, over(numalloc) //convert data into Stata mi estimate format mi import ice, clearįigure 1 Prediction equations generated from the ice command VariableĮq5d_B numalloc bmi age gender costs_year1 */ use MI_aggregated, clear //open multiple imputed dataset //generate total QALYs and total costs gen total_QALYs = QALY_y1 + QALY_y2 + QALY_圓 + QALY_y4 + QALY_y5 gen total_costs = costs_year1 + costs_year2_D + costs_year3_D + costs_year4_D + costs_year5_D drop if _mj=0 //this is the original dataset with missing data ‘seed’ sets a random number seed, which is useful to improve consistency across imputations. ‘genmiss’ generates an indicator of missingness =1 if observation was originally missing or =0 otherwise.
The multiple imputation generates 60 (m=60) datasets using predictive mean matching (‘match’) and separately by treatment allocation (‘by(numalloc)’). Multiple imputation with chained equations using ice ice eq5d_B QALY_y1 QALY_y2 QALY_圓 QALY_y4 QALY_y5 /* */ costs_year1 costs_year2_D costs_year3_D costs_year4_D costs_year5_D /* */ numalloc bmi age gender, saving(MI_aggregated, replace) /* */ m(60) match genmiss(ind_miss) by(numalloc) seed(10) /* This command runs multiple imputation with chained equations using the models shown in Figure 1 below and saves the multiple imputed dataset in ‘MI_aggregated’. Keep studyno eq5d_B QALY_y1 QALY_y2 QALY_圓 QALY_y4 QALY_y5 costs_year1 costs_year2_D /* */ costs_year3_D costs_year4_D costs_year5_D numalloc bmi age genderģ. Prepare data for multiple imputation //calculate discounted costs at each year gen costs_year1 = c_allsurgery + c_drug_y1 gen costs_year2_D = (c_drug_y2 + c_reGP_y2 + gen costs_year3_D = (c_drug_圓 + c_reGP_圓 + gen costs_year4_D = (c_drug_y4 + c_reGP_y4 + gen costs_year5_D = (c_drug_y5 + c_reGP_y5 + The ‘mim’ package should also appear on this list.Ģ.
#Web stata install
Install ice findit ice//choose the first option on the list.
Note that data were modified to simplify the illustration of the different methods to handle missing data.
*Total QALYs and total costs over 5 years refer to the sum of QALYs and costs discounted at a 3.5% annual rate over the individuals with complete data for the relevant variables (EQ-5D for QALYs and cost components for costs). Outcomes for cost-effectiveness total_QALYs Outcome variables for health-related quality of life eq_3mĬosts of surgery in those who had surgery Table 1 Description of variables in REFLUX Variable
0 kommentar(er)
