Table 5.

Summary of findings for outcome valuation (KQ2)

Number of included studies; sample size	Disutility (95% CI) 0 (no decrease in HRQoL) to 1 (HRQoL equal to death)	GRADEꝉ	What does the evidence say?
Disutility from hip fracture
16 studies; n = 7409	EQ-5D, at time of fracture (< 3 months) 0.53 (0.44 to 0.62)	⊕ ⊕ ⊕ ⊝  Moderatea	The disutility of a hip fracture is probably 0.53 immediately after injury
27 studies; n = 9399	EQ-5D, 12-month post-fracture 0.16 (0.12 to 0.20)	⊕ ⊕ ⊕ ⊕  High	The disutility of a hip fracture is 0.16 at 12 months after injury
Disutility from non-hip* fracture
4 studies; n = 1792	EQ-5D, at time of fracture (< 3 months) 0.57 (0.43 to 0.71)	⊕ ⊕ ⊝ ⊝  Lowab	The disutility of a non-hip fracture may be 0.57 immediately after injury
4 studies; n = 1792	EQ-5D, 12-month post-fracture 0.19 (0.10 to 0.28)	⊕ ⊕ ⊕ ⊝  Moderateb	The disutility of a non-hip fracture is probably 0.19 at 12 months after injury
Disutility from any injurious fall
0 study	No evidence	Not applicable	The disutility of an injurious fall is uncertain
Disutility from a fall (within last 12 months)
6 studies; n = 4653	EQ-5D 0.09 (− 0.04 to 0.22)	⊕ ⊕ ⊝ ⊝  LowA	The disutility after a fall may be 0.09
Disutility from functional impairment (impairment in at least one ADL**)
1 study; n = 123	HUI Mark II 0.12 (0.05 to 0.19)	⊕ ⊕ ⊝ ⊝  Lowcd	The disutility from impairment in one or more ADLs may be 0.12
Disutility of LTC admission (compared to full health)
1 study; n = 194	TTO Median (IQR): 1 (1, 1) “80% of participants said they would rather be dead"	⊕ ⊕ ⊝ ⊝  Lowcd	The disutility from a long-term care admission (compared to full health) may be 1
Relative importance across health states
EQ-5D unless otherwise specified Disutility, LTC admission: 1 (TTO) Disutility, non-hip fracture (< 3 mos): 0.57 Disutility, non-hip fracture (12 mos): 0.19 Disutility, hip fracture (< 3 mos): 0.53 Disutility, hip fracture (12 mos): 0.16 Disutility, ADL impairment: 0.12 (HUI Mark II) Disutility, fall: 0.09 Disutility, injurious fall: unknown Also see below rows for findings from other preference-based studies, used for comparison		⊕ ⊕ ⊝ ⊝  Lowe,f	LTC admission may be more important than all other outcomes
		⊕ ⊕ ⊕ ⊝  Moderatea	Fracture (hip or non-hip) is probably more important than falls and functional impairment
		⊕ ⊕ ⊝ ⊝  Lowe,f	Functional impairment may be somewhat more important than a fall
Findings of relative importance between health states
50% decrease in fracture risk:50% decrease in fall risk: ratio, SMD of coefficients: 2.43 (Milette 2013, Franco 2015; DCE) 50% improvement in daily functioning:50% decrease in fall risk: ratio, SMD of coefficients: 2.11 (Franco 2015; DCE) Ability to manage domestic activities:HRQoL: ratio, relative importance score: 1.58 (Hilingsman 2020; CA) Ability to manage domestic activities:fall frequency: ratio, relative importance score: 1.20 (Hilingsman 2020; CA) LTC admission:falls risk: ratio, relative importance score: 1.18 (Robinson 2015; CA) 50% decrease in fracture risk:50% improvement in daily functioning (Milette 2013, Franco 2015; DCE): ratio, SMD of coefficients: 1.14
Data from other utility instruments, disutilities
Hip fracture; disutility at time closest to injury Average of HUI Mark II, SF-6D, SG, & FT 1 study; n = 80 0.25 (0.20 to 0.30) Hip fracture; disutility at time closest to 12 mos TTO 2 studies; n = 471 0.57 (0.32 to 0.82) Average of HUI Mark II, SF-6D, SG, & FT 1 study; n = 80 0.12 (0.07 to 0.17) Falls: disutility any time after event TTO 1 study; n = 203 0.33 (0.26 to 0.40)

Abbreviations: ADL activity of daily living, CA conjoint analysis, DCE discrete choice experiment, FT feelings thermometer, SG standard gamble, TTO time trade-off. *Non-hip fracture refers to any fracture other than hip fracture; studies contributing to this health state reported on distal forearm, humerus, and clinical vertebral fractures (weighted average used for analysis). **This study measured utility using HUI Mark II in individuals with and without impairment in ≥ 1 ADL; participants were not asked how long they had been ADL impaired. Explanation of GRADE ratings: Lowercase superscript indicates rating down once, and an uppercase superscript indicates rating down twice. A, inconsistency in estimates across multiple studies. B, > 50% of sample size is from studies rated at high risk of bias. C, concerns about indirectness due to mean age > 80 years in the single study. D, concerns about lack of consistency due to evidence from single study. E, concerns about indirectness due to outcome measure other than EQ-5D. F, concerns about inconsistency due to lack of agreement between utility ranking and relative importance studies