Costs and expected gain in lifetime health from intensive care versus general ward care of 30,712 individual patients: a distribution-weighted cost-effectiveness analysis

Frode Lindemark; Øystein A. Haaland; Reidar Kvåle; Hans Flaatten; Ole F. Norheim; Kjell A. Johansson

doi:10.1186/s13054-017-1792-0

Table 1 Key model assumptions and parameters

From: Costs and expected gain in lifetime health from intensive care versus general ward care of 30,712 individual patients: a distribution-weighted cost-effectiveness analysis

Key model assumptions		Description			Sources
Patients		General adult ICU population (n = 40,916), Norway 2008–2010. One or more organ failures, case mix, >90% of annual ICU admissions in the country			Norwegian Intensive Care Registry
Intervention		ICU admission
Comparator		Hypothetical ICU refusal and general ward treatment of the same patients
Setting		38 mixed medical-surgical ICUs out of a total of 42 ICUs in publicly funded hospitals operated by local health trusts under four state-owned regional health authorities.
Health outcome		Quality-adjusted life years (QALYs)
Resource use		Hospital costs were summed in Norwegian kroner (NOK), converted to NOK in 2016 using the CPI and to Euro (€) using OECD PPP: 1 USD = 9.4 NOK = 0.75 Euros			Statistics Norway, [51]
Severity of disease		Expected lifetime health if rejected to the ICU. Measured as number of lifetime QALYs.			[13]
Distribution weights		Higher weights assigned to health gains to patients with more severe conditions; results in more favourable cost-effectiveness estimates.			Additional file 2, [13, 62]
Perspective		Hospital, health care provider			[50]
Time horizon		Lifetime			[50]
Discounting		4% annually of health benefits and incremental health care costs post-hospital discharge.			[50]
Heterogeneity		Analysis of total study population, subgroups by type of admission and individual admissions.
General cost-effectiveness threshold range		275,000–800,000 NOK per QALY ≈22,000–64,000 PPP-adjusted € per QALY.			[56, 57]
Model parameter		Parameter values			Sources
.		Base case	Range	Distribution	.
Age		Individual age	Individual age	Observed	ICU cohort 2008–2010
Transition probabilities:
Effect of ICU admission on hospital mortality
Predicted hospital mortality if ICU patient		Individual SAPS II, calibrated model		Observed	ICU cohort 2008–2010, [35]
Predicted hospital mortality if general ward		Individual SAPS II, modified model (β1 = 0.09)	β1 = 0.0737 to 0.14	Scaled beta^a	Fig. 1, Additional files 1 and 2
Annual mortality after hospital discharge		Age-specific *3–1 over 3 years	1 to 5–1 over 10 years	Uniform	Life table Norway 2011, [34], Additional file 2
Length of hospital stay, mean (median) days
If ICU, weighted sum of
Died ICU	LOS ICU	4.8 (1.7)	0.6–5.0 (IQR)	Observed	ICU cohort 2008–2010
	LOS ward post ICU	0	NA	NA
Died ward	LOS ICU	5.4 (2.6)	1.3–5.9 (IQR)	Observed	ICU cohort 2008–2010
	LOS ward post ICU	2.5	NA	Gamma (scale = 0.5, shape = 5)
Survived hospital	LOS ICU	4.1 (2.0)	1.1–4.1 (IQR)	Observed	ICU cohort 2008–2010
	LOS ward post ICU	10	NA	Gamma (scale = 2, shape = 5)
If general ward, weighted sum of
Died hospital	LOS ward	LOS if ICU and dead *0.9	NA	Gamma
Survived hospital	LOS ward	LOS if ICU and hospital survivor *1.1	NA	Gamma
Cost units
Cost of ICU bed-day		€3980	€2390 to €5570	Scaled beta^b	Additional file 2, [18]
Cost of general ward bed-day		€640	€320 to €950	Scaled beta^b	Additional file 2
Annual health care cost survivors, year 1		€6400	€80 to €12,700	Scaled beta^b	[47, 48]
Years 2–5		Year 1 0.6, 0.5, 0.4, 0.3
Health-related quality of life weights		*0.8	*0.6 to 0.9	Scaled beta^c	Additional files 1 and 2, [19, 38,39,40,41,42]
Age-matched reference value reduced 20% for 5 years after discharge, equally for both treatment options	18–19: 0.90	0.72	0.54 to 0.81
	20–29: 0.89	0.71	0.53 to 0.80
	30–39: 0.88	0.70	0.53 to 0.79
	40–49: 0.86	0.69	0.52 to 0.77
	50–59: 0.84	0.67	0.50 to 0.76
	60–69: 0.81	0.65	0.49 to 0.73
	70–79: 0.79	0.63	0.47 to 0.71
	80+: 0.73	0.59	0.44 to 0.66
Distribution weights based on lifetime QALYs		Linear	NA	NA	Additional file 2, [13, 62]

Scaled beta: min + (max – min) × beta(alpha,beta)
^abeta(alpha = 2,beta = 3.5)
^bbeta(alpha = 2,beta = 2)
^cbeta(alpha = 4,beta = 2
CPI consumer price index, ICU intensive care unit, IQR interquartile range, LOS length of stay, NA not applicable, NOK Norwegian kroner, PPP purchasing power parity, SAPS Simplified Acute Physiology Score, USD United States dollar

Back to article page