SELECTING SOFTWARE PACKAGES FOR CONDUCTING COST-EFFECTIVENESS MODELLING IN HEALTH CARE
Author(s)
Stephens TJ, Selya-Hammer C, Gauthier A
Amaris, London, UK
Presentation Documents
OBJECTIVES: Microsoft Excel® is considered the essential software package for developing health economic models, and is one of four preferred by NICE for technology assessment submissions. As models become more complex and analyses become computationally expensive, the use of alternative software packages becomes a pertinent consideration. This study aimed to identify the different packages used for cost-effectiveness modelling in HTA submissions to NICE and the criticisms of the chosen software by Evidence Review Groups (ERG). Recommendations on choice of software for models of varying complexity across disease areas were then developed. METHODS: All cost-effectiveness models submitted to NICE by manufacturers and published between 2006 and 2016 were assessed for software used to develop the model. Data were extracted from submitted models to determine predictors of each package’s use, and likelihood of criticisms and acceptability by the ERGs. Results were then used to develop a decision algorithm to guide software choice. RESULTS: The search identified 181 submissions utilising 6 different software packages. Excel® was the most common, having been used in 90% of submitted models. Other commonly used programs included TreeAge, SAS, and SIMUL8. The principal factor identified in choosing to model in non-Excel software was for the evaluation of non-oncology drugs (43% of submissions). The number of health states, varying cycle length, or model structure (i.e. discrete event simulation) necessitated advanced software to manage these aspects. ERGs were sometimes critical of non-Excel models perceived as overcomplicated, but overall such models were well received. CONCLUSIONS: Microsoft Excel® is still the most widely used and accepted package for developing health economic models, however the use of different packages is justified under certain conditions. ERGs can be critical of non-conventional models developed in alternative software. The results of our analysis are presented in a decision algorithm to guide the choice of modelling software.
Conference/Value in Health Info
2017-11, ISPOR Europe 2017, Glasgow, Scotland
Value in Health, Vol. 20, No. 9 (October 2017)
Code
PRM132
Topic
Methodological & Statistical Research
Topic Subcategory
Modeling and simulation
Disease
Multiple Diseases