Table 2 Model inputs and data sources

CAZ-AVI + metronidazoleCeftolozane/Tazobactam + metronidazoleMeropenemColistin + tigecycline + high-dose Meropenem
Probability of clinical cure a91.7%b94.1%c92.5%b75.0%d
Probability of AEe4.9%e8.1%c3.8%e14.8%f
Probability of recurrence0.0%b0.0%g0.6%b0.0%g
Treatment duration9.5 daysh9.0 daysh9.5 daysh9.5 daysh
Probability of in-hospital deathi
Appropriate empiric treatment: 4.80%
Inappropriate empiric therapy: 10.70%
Resistant to empiric therapy: 12.84%j
Utility (quality of life)
With clinical response: 0.92k
Without clinical response: 0.61l
Hospital length of staym
With clinical response: 11.71 days
Without clinical response: 24.13 days
Proportion of hospitalisation days in ICU b
With clinical cure: 26.92%
With clinical failure: 11.45%
Daily drug costs,n (average daily dose)€ 300.77 (CAZ-AVI 7500 mg; metronidazole 1500 mg)€ 248.97 (ceftolozane/tazobactam 1500 mg; metronidazole 1500 mg)€ 55.32 (3000 mg)€ 218.55 (colistin [IV] 5 mg; tigecycline 100 mg; meropenem 6000 mg)
Hospital cost per day General ward: € 697.23o; ICU € 1383.00p
Cost of SAEo € 3027
Cost of recurrenceo € 6787
1. AE = adverse event; BNF=British National Formulary; CAZ-AVI = ceftazidime-avibactam; ICU = intensive care unit; IV = intravenous; SAE = serious adverse event.
2. aProbability of clinical cure of patients without resistance
3. bRECLAIM clinical study data [1]
4. cSolomkin et al. 2015 [35]
5. dExpert opinion
6. eAEs considered in the model included only serious AEs, as these have relevant cost impact and can result in treatment discontinuation or treatment switch. Probability of SAE (up to EOT) was based on RECLAIM clinical study data
7. fPooled data from multiple sources: Chen et al. 2010 [21], Fomin et al. 2005 [22], Oliva et al. 2005 [23], Qvist et al. 2012 [24], and Towfigh et al. [25]
8. gAssumption (due to lack of data)
9. hEuropean Medicines Agency (EMA) product labels [26, 27]
10. iSturkenboom et al. 2005 [28]
11. jAssumed to be 20% higher than mortality among patients with inappropriate empiric therapy (without resistance)
12. kSong et al. 2012 [29]
13. lDelate et al. 2001 [30], assuming similar utility for patients with different infections. Deterministic sensitivity analysis showed small impact of utility of cIAI (i.e., utility applied while patients have not been cured) on the results
14. mPayer Analysis data
15. nAIFA, Agenzia Italiana del Farmaco. 2014 (except for cost of colistin which was taken from BNF, converted to Euros using an exchange rate of £1 = €1.36) [31]
16. oItalian hospital diagnosis-related groups (DRGs 2013 and 2015) [32, 33]
17. pTan et al. 2012 [34]