Background
Methods
Data sources and study selection
Data extraction and synthesis
Category | Data field |
---|---|
Reporting and methodological quality checklist | The checklist designed for falls prevention economic evaluations by a panel of falls prevention experts [32] was adapted to specifically suit decision models. There were 32 items, each scored 0 (recommendation not followed), 0.5 (partially followed), and 1 (fully followed), giving maximum score of 32. See Table A2 in Supplementary Materials for adapted version. |
(A) Model and evaluation overview | 1. Bibliography: author(s); publication year 2. Setting and aim: country; region; decision-maker; evaluation aim 3. Target population demographics and comorbidities (e.g., residence,a age, sex, socioeconomic status, health conditions unrelated to falls risk) 4. Type of analysis: e.g., CEA; CUA; CBA; ROIb 5. Perspective (e.g., public sector, societal) 6. Cost-effectiveness threshold: monetary amount and type (e.g., health opportunity cost in healthcare system, willingness to pay as consumer) 7. Model type (e.g., decision tree, Markov) 8. Model time horizon 9. Discount rates (if time horizon longer than 1 year) 10. Model cycle length (if any) |
(B) Falls epidemiology features | 1. Characterising baseline falls risk of target population 2. Characterising multiple falls per year (recurrent falls) 3. Risk factors for falls 4. Health consequences of falls: fall/injury type; long-term health consequences (e.g., institutionalisation, excess mortality risk) 5. Health utility data: fall-related loss; comorbidity status 6. Economic consequences of falls: care resource types; unit costs; all-cause/comorbidity care costsc |
(C) Falls prevention intervention features | 1. Intervention characteristics: type;d comparator(s); component; access pathwaye 2. Falls risk screening methodf 3. Intervention resource use and costs: auxiliary implementation resources (e.g., marketing to improve uptake); therapeutic resources (e.g., staff labour). 4. Intervention efficacy: metric;g fall type;h effectiveness periodi 5. Wider health effects of interventions beyond falls preventionj |
(D) Evaluation methods and results | 1. Model validity: structural/face;k internal; external; crossl 2. Assessing parameter uncertainty: DSA; PSA 3. Scenario analyses: to assess impact of structural assumptions on outcomes. 4. Aggregate health and cost outcomes (e.g., total intervention cost, total QALY gain, total number of falls prevented) 5. Cost-per-unit ratios (e.g., incremental cost per QALY gain) 6. Wider decisional outcomes (e.g., reduction in social inequities of health) 7. Currency: original type/year; conversion to same currency for comparison 8. Discussion by evaluation authors: generalisability; policy implementation; model strengths and limitations |
(E) Key methodological challenges for public health economic model | 1. Capturing non-health outcomes and societal intervention costs 2. Considering heterogeneity and dynamic complexity: e.g., long-term progression of falls risk factors/profile 3. Considering theories of human behaviour and implementation: e.g., implementation quality (i.e., uptake and adherence rates) 4. Considering social determinants of health and conducting equity analyses |
Model overview and checklist scores for reporting and methodological quality
Narrative synthesis of methodological features and methodological recommendations
Developing commissioning recommendations by this systematic review
Results
Search results
Overview of included decision models
# | Reference | Setting | Target population | Type of analysis | Perspective | Intervention type | Comparator | Model type | Time horizon |
---|---|---|---|---|---|---|---|---|---|
1 | Agartioglu (2020) [50] | Turkey, Izmir | CD adults aged 65+ | CEA | Public sector | HAM | UC | DT | 1 year |
2 | Albert (2016) [51] | US, Pennsylvania | CD adults aged 50+ (mean age 75.5) | CUA | Public sector | MF int. | UC | DT | 1 year |
3 | Alhambra-Borras (2019) [52] | Spain, Valencia, hospital level | CD adults aged 65+ at high falls risk or frail with no severe physical or cognitive limitation | CUA | Public sector | Exercise | UC | Markov cohorta | Lifetime |
4 | Beard (2006) [53] | Australia, NSW | CD adults aged 60+ | CBA; ROI | Public sector; Societal | MC (intersectoral) int.b | UC | Binary decisionc | 5 years |
5 | Boyd (2020) [54] | New Zealand | Adults aged 65+ | CUA | Public sector | Cataract surgery (expedited, routine) | NR | Markov cohort | Lifetime |
6 | Carande-Kulis (2015) [55] | US, private insurers | CD adults aged 65+ | ROI | US health insurance payer | Exercise (2 forms); MC int. (Stepping On) | NR | Binary decision | 1 year |
7 | CSP (2016) [56] | UK, varying regions | CD adults aged 65+ | ROI | Public sector | FRS + Exercise (physiotherapy) | NR | DT | 1 year |
8 | Church (2011) [57] | Australia, NSW | CD adults aged 65+ (separate model for residential care) | CEA; CUA | Public sector | Exercise (3 forms); MC int.; MF int.; MRA; Exp. cataract surgery; Med. modification; Cardiac pacing | NR | Markov cohort | 10 years |
9 | Church (2012) [58] | Australia, NSW | CD adults aged 65+ | CEA; CUA | Public sector | Exercise (4 forms); MC int.; MF int. (2 forms); MRA; HAM; Exp. cataract surgery; Cardiac pacing; Med. modification | NR; Cross-comparison | Markov cohort | Lifetime |
10 | Comans (2009) [59] | Australia, Brisbane | CD adults aged 65+, falls history in past 6 months or gait/functional decline and cognitively intact | ROI | Societal | MF int. (2 forms) | NR | Binary decision | 1 year |
11 | Day (2009) [60] | Australia, varying regions | CD adults aged 50+ (age and characteristics differ by intervention type)d | CEA | Public sector; Societal | Exercise (2 forms); HAM; MF int.; Med. modification; Cardiac pacing | NR | DT | 1 year |
12 | Day (2010) [61] | Australia | CD adults aged 70+ | CEA | Public sector; Societal | Exercise (Tai Chi) | NR | DT | 1 year |
13 | Deverall (2018) [62] | New Zealand | CD adults aged 65+ | CUA | Public sector; Societal | Exercise (3 forms) | NR | Markov cohort | 25 years |
14 | Eldridge (2005) [63] | UK, primary care trust | Adults aged 65+ in community or nursing home | CUA | Public sector | FRS + MF int. or Exercise | UC | DT + Markov cohort | Lifetime |
15 | Farag (2015) [64] | Australia | CD adults aged 65+ without falls history | CUA | Public sector | Non-specific intervention | NR | Markov cohort | Lifetime |
16 | Franklin (2019) [65] | UK, city level | CD adults aged 65+ | CUA | Public sector (2 types) | FRS + Exercise (3 forms) or HAM | NR; Cross-comparison | DT + Markov cohort | 2 years |
17 | Frick (2010) [66] | US | CD adults aged 65+ | CUA | US healthcare payere | Exercise (2 forms); HAM; MF int. (2 forms); Vit. D; Med. modification | Cross-comparison | Binary decision | 1 yearf |
18 | Hektoen (2009) [67] | Norway | CD women aged 80+ | CEA | Societal | Exercise | NR | Binary decision | 1 year |
19 | Hiligsmann (2014) [68] | Belgium | Adults aged 60+ with osteoporosis | CUA | Societal | Vit. D and calcium | NR | Markov patienta | Lifetime |
20 | Hirst (2016) [69] | UK | Women aged 75+ on chronic pain medication | CUA | Public sector | Med. modification (Transdermal Buprenorphine) | Tramadol | DTg | 1 year |
21 | Honkanen (2006) [70] | US, Medicare/aid | Adults aged 65+ living in community at baseline | CUA; ROI | Societal | Hip protectors | NR | Markov cohort | Lifetime |
22 | Howland (2015) [71] | US, Massachusetts | CD adults aged 65+ admitted to A&E due to fall | ROI | US healthcare payere | MC int. (MoB/VLL) | NR | Binary decision | 1 year |
23 | Ippoliti (2018) [72] | Italy, Piedmont | CD adults aged 65+ living in mountainous areas | ROI | Public sector | MF int. | NR | Binary decision | 3 years |
24 | Johansson (2008) [73] | Sweden, Stockholm | CD adults aged 65+ | CUA | Societal | MC (intersectoral) int.h | UC | Markov cohort | Lifetime |
25 | Lee (2013) [74] | US, Medicare/aid | CD adults aged 65–80 without falls history | CBA | Public sector | Vit. D (targeted, universal) | NR | DT + Markov cohort | 3 years |
26 | Ling (2008) [75] | US, Hawaii | CD adults aged 65+ with falls history or other risk factors | ROI | US healthcare payere | HAM | NR | Binary decision | 1 year |
27 | McLean (2015) [76] | Australia, Melbourne | CD adults aged 70+ | CEA; CUA | Public sector | Exercise | UC | DT | 18 months |
28 | Miller (2011) [77] | US, Texas | CD adults aged 50+ at high falls risk | ROI | US healthcare;e Societal | MC int. (MoB/VLL) | NR | Binary decision | 2 years |
29 | Mori (2017) [78] | US | CD women aged 65+ at osteoporosis risk without previous fracture | CUA | Societal | Exercise and bisphosphonate combined | Cross-comparison: single or no intervention | DT + Markov patient | Lifetime |
30 | Moriarty (2019) [79] | Ireland | CD adults aged 65, no current/previous adverse events for benzodiazepine/PPI | CUA | Public sector | Med. modification (Benzodiazepine, PPIn) | Inappropriate prescribing | DT + Markov patient | 35 years |
31 | Nshimyu-mukiza (2013) [80] | Canada | Women aged 40+ (with subgroup aged 65+) | CEA; CUA | Public sector | Fracture risk screening + Physical activity, Vit. D and calcium, and/or Osteoporosis screen & treat | NR; Cross-comparison | DT + Markov patient | Lifetime |
32 | OMAS (2008) [81] | Canada, Ontario | CD adults aged 65+ | CEA; ROI | Public sector | Exercise; HAM; Vit. D and calcium; Med. modification; gait-stabilizer | NR | Markov cohort | Lifetime |
33 | Pega (2016) [82] | New Zealand | CD adults aged 65+ | CUA | Public sector | HAM | NR | Markov cohort | Lifetime |
34 | Poole (2014) [83] | UK | Adults aged 65+ | ROI | Public sector | Vit. D | NR | Binary decision | 1 year |
35 | Poole (2015) [84] | UK | CD adults aged 60+ | CUA; ROI | Public sector | Vit. D | NR | Markov cohort | 5 years |
36 | PHE (2018) [85] | England, varying regions | CD adults aged 65+ | CUA; ROI | Public sector | Exercise (3 forms); HAM | NR | DT | 2 years |
37 | RCN (2005) [34] | England & Wales | CD adults aged 60+ | CUA | Public sector | Exercise; MF int. | NR | Markov cohort | Lifetime |
38 | Sach (2007) [86] | UK | Women aged 70+ with bilateral cataracts | CEA; CUA | Public sector; Societal | Exp. cataract surgery (first eye) | UC (routine surgery) | Binary decision | Lifetime extrapol.i |
39 | Sach (2010) [87] | UK | Women aged 70+ with second operable cataract | CUA | Public sector; Societal | Exp. cataract surgery (second eye) | UC (no surgery) | Binary decision | Lifetime extrapol.i |
40 | Smith (2016) [88] | UK, NW London | Adults aged 65+ covered by GP practice and hospital | ROI | Public sector | FRS + MF int. | Cross-comparison | Risk prediction | 1 year |
41 | Tannenbaum (2015) [89] | US, Medicare/aid | CD adults aged 65+ with insomnia | CUA | Public sector | Med. modification; CBT | NR; Cross-comparison | Markov cohort | 1 year |
42 | Turner (2020) [90] | Canada, Quebec | CD adults aged 65+ who are chronic users of sedatives for insomnia | CUA | Public sector | Med. modification | NR | DT + Markov cohort | 1 year |
43 | Velde (2008) [91] | Netherlands | CD geriatric outpatient population with falls history (mean age 78) | CEA | Public sector | Med. modification | NR | Binary decision | 1 yearf |
44 | Wilson (2017) [92] | New Zealand, Manukau | CD adults aged 65+ | CUA | Public sector | HAM | NR | Markov cohort | Lifetime |
45 | Wu (2010) [93] | US, Medicare/aid | CD Medicare beneficiaries aged 65+ with falls history | CEA; ROI | Public sector; Societal | MF int. | NR | Binary decision | 1 year |
46 | Zarca (2014) [94] | France | Adults aged 65+ without previous hip fracture | CEA; CUA | Public sector | Vit. D (targeted (2), universal) | NR; Cross-comparison | DT + Markov patient | Lifetime |
Checklist scores for methodological and reporting quality
Narrative synthesis: falls epidemiology features
Baseline falls risk
Data source | N | Study label (n = 46)a |
---|---|---|
(1) Individual-level epidemiological data | 8 | |
(2) Published epidemiological data or expert/author opinion | 25 | Agartioglu (2020) [50]; Carande-Kulis (2015) [55]; CSP (2016) [56]; Church (2011) [57]; Church (2012) [58]; Farag (2015) [64]; Franklin (2019) [65]; Frick (2010) [66]; Hiligsmann (2014) [68]; Hirst (2016) [69]; Honkanen (2006) [70]; Howland (2015) [71]; Lee (2013) [74]; Ling (2008) [75]; Miller (2011) [77]; Mori (2017) [78]; Moriarty (2019) [79]; Nshimyumukiza (2013) [80]; Poole (2014) [83]; Poole (2015) [84]; RCN (2005) [34]; Tannenbaum (2015) [89]; Turner (2020) [90]; Wu (2010) [93]; Zarca (2014) [94] |
(3) Internal intervention studyb evidence | 9 | |
(4) Risk/rate from external RCT control group | 4 |
Recurrent falls
Study label (n = 46)a | Transition entityb | Cycle length | Main fall-related event | Possible to model recurrent falls |
---|---|---|---|---|
Binary decision model | ||||
Beard (2006) [53] | Fall event | N/Ac | MA fall | Yes |
Carande-Kulis (2015) [55] | Individual | N/A | MA fall | No |
Comans (2009) [59] | Individual | N/A | Any fall | Yes |
Frick (2010) [66] | Individual | N/A | Hip fracture | No |
Hektoen (2009) [67] | Fall event | N/A | Any fall | Yes |
Howland (2015) [71] | Individual | N/A | MA fall | Yes: targeted recurrent fall |
Ippoliti (2018) [72] | Fall event | N/A | Hip fracture | Yes |
Ling (2008) [75] | Individual | N/A | Any fall | No |
Miller (2011) [77] | Individual | N/A | MA fall | No |
Poole (2014) [83] | Fall event | N/A | Hip fracture | Yes |
Fall event | N/A | Any fall | Yes | |
Velde (2008) [91] | Fall event | N/A | Any fall | Yes |
Wu (2010) [93] | Individual | N/A | Any fall | Yes: targeted recurrent fall |
Static modeld | ||||
Agartioglu (2020) [50] | Individual | N/A | Any fall | No |
Albert (2016) [51] | Individual | N/A | Any fall | Yes |
CSP (2016) [56] | Individual | N/A | MA fall | Yes |
Fall event | N/A | Any fall | Yes | |
Hirst (2016) [69] | Individual | N/A | Fractures | No |
McLean (2015) [76] | Individual | N/A | Any fall | Yes: Adjusted risk |
PHE (2018) [85] | Fall event | N/A | Any fall | Yes |
Smith et al. (2016) [88] | Individual | N/A | MA fall | No |
Cohort-level Markov model | ||||
Alhambra-Borras (2019) [52] | Individual | 1 year | Compositee | Yes: Compositee |
BODE3 models | Individual | 1 year | MA fall | No |
Individual | 1 year | Any fall | No | |
Eldridge (2005) [63] | Individual | 1 year | Any fall | No |
Farag (2015) [64] | Individual | 1 year | Any fall | No |
Franklin (2019) [65] | Individual | 1 year | Any fall | Yes |
Honkanen (2006) [70] | Individual | 1 year | Hip fracture | No |
Johansson (2008) [73] | Individual | 1 year | Hip fracture | No |
Lee (2013) [74] | Individual | 1 month | Any fall | Yes |
Moriarty (2019) [79] | Individual | 1 year | MA fall/Hip fracture | No |
OMAS (2008) [81] | Individual | 1 year | MA fall | No |
Poole (2015) [84] | Individual | 1 year | MA fall | No |
RCN (2005) [34] | Individual | 1 year | MA fall | No |
Tannenbaum (2015) [89] | Individual | 6 months | Any fall | Yes |
Turner (2020) [90] | Individual | 1 month | MA fall/Hip fracture | Yes |
Individual-level Markov model (microsimulation) | ||||
Hiligsmann (2014) [68] | Individual | 6 months | Fractures | Yes |
Mori (2017) [78] | Individual | 1 year | Fractures | No |
Nshimyumukiza (2013) [80] | Individual | 1 year | Fractures | No |
Zarca (2014) [94] | Individual | 3 months | Hip fracture | Yes |
Falls risk factors
Falls health consequences
Study label (n = 46)b | Non-MA or non-injurious fall | MA or injurious fall | Fracture | Fatal fall | Fear of falling | Fall-induced LTC admission | Excess mortality |
---|---|---|---|---|---|---|---|
Agartioglu (2020) [50] | ˟ | Injury | Mixc | ||||
Albert (2016) [51] | ˟ | MA | |||||
Alhambra-Borras (2019)d [52] | Com | Com | Com | ||||
Beard (2006) [53] | MA | ||||||
BODE3 models | MA | ˟ | |||||
Carande-Kulis (2015) [55] | MA | ˟ | |||||
CSP (2016) [56] | ˟ | MA | ˟ | ||||
˟ | MA | Mix | ˟ | ˟ | ˟ | ||
Comans (2009) [59] | ˟ | MA | |||||
˟ | MA | ||||||
Eldridge (2005) [63] | ˟ | MA | Hip | ˟ | ˟ | ˟ | ˟ |
Farag (2015) [64] | ˟ | MA | ˟ | ˟ | |||
Franklin (2019) [65] | ˟ | MA | ˟ | ˟ | ˟ | ||
Frick (2010) [66] | Hip | ˟ | ˟ | ||||
Hektoen (2009) [67] | ˟ | Injury | Mix | ||||
Hiligsmann (2014) [68] | Mix | ˟ | |||||
Hirst (2016) [69] | Mix | ˟ | |||||
Honkanen (2006) [70] | Hip | ˟ | ˟ | ||||
Howland (2015) [71] | MA | ||||||
Ippoliti (2018) [72] | Hip | ||||||
Johansson (2008) [73] | Hip | ˟ | |||||
Lee (2013) [74] | ˟ | MA | ˟ | ||||
Ling (2008) [75] | ˟ | MA | ˟ | ||||
McLean (2015) [76] | ˟ | Injury | Mix | ||||
Miller (2011) [77] | ˟ | MA | |||||
Mori (2017) [78] | Mix | ˟ | ˟ | ||||
Moriarty (2019) [79] | MA | Hip | ˟ | ˟ | |||
Nshimyumukiza (2013) [80] | Mix | ˟ | ˟ | ||||
OMAS (2008) [81] | MA | Mix | ˟ | ˟ | |||
Poole (2014) [83] | Hip | ˟ | |||||
Poole (2015) [84] | MA | ˟ | ˟ | ||||
PHE (2018) [85] | ˟ | MA | Mix | ˟ | ˟ | ˟ | |
RCN (2005) [34] | MA | Hip | |||||
˟ | MA | ||||||
Smith (2016) [88] | MA | Mix | |||||
Tannenbaum (2015) [89] | ˟ | MA | Mix | ˟ | ˟ | ˟ | |
Turner (2020) [90] | MA | Mix | ˟ | ||||
Velde (2008) [91] | ˟ | MA | |||||
Wu (2010) [93] | ˟ | MA | |||||
Zarca (2014) [94] | Hip | ˟ |
Health utilities
Fall-related economic consequences
Narrative synthesis: falls prevention intervention features
Intervention access pathway
Intervention type (total N) | Reactive | N | Proactive | N | Self-referred | N | Unclear | N |
---|---|---|---|---|---|---|---|---|
Exercise (33) | 0 | 9 | 10 | 14 | ||||
HAM (11) | 2 | 2 | Wilson (2017) [92] | 1 | 6 | |||
Medication review and modification (10) | 0 | 4 | 0 | 6 | ||||
Cataract surgery (5) | 0 | 2 | 0 | 3 | ||||
Vitamin D supplement (11) | Nshimyumukiza (2013) [80] | 1 | 7 | 0 | 4 | |||
Other single-component (6) | Day (2009) [60] – cardiac pacing | 1 | 0 | Farag (2015) [64] – non-specific intervention | 1 | 4 | ||
MF int. and MRA (17) | 2 | 5 | Albert (2016) [51] | 1 | 9 | |||
MC int. (7) | Howland (2015) [71] | 1 | 0 | 3 | 3 | |||
Total (101) | 7 | 29 | 16 | 49 |
Falls risk screening
Intervention resource-use and cost
Intervention efficacy
Study label (n = 46)a | Intervention type | Efficacy (main model) fall-related event | Data source type | Efficacy (incidence) metric | Effectiveness periodb (model time horizon) |
---|---|---|---|---|---|
Agartioglu (2020) [50] | HAM | Any fall (any fall) | External meta-an. and internal RCT | RR (risk) | 1 year (1 year) |
Albert (2016) [51] | Multifactorial int. | Any fall (any fall) | Internal non-randomised | RR (risk) | 1 year (1 year) |
Alhambra-Borras (2019) [52] | Group exercise | Compositec (composite) | Internal quasi-experiment | RR (risk) | 1 year (lifetime) |
Beard (2006) [53] | Multifactorial int. | Hospital fall (hospital fall) | Internal quasi-experiment | RaR (rate) | 5-year sustainability (5 years) |
Boyd (2020) [54] | Expedited cataract surgery | Any fall (MA fall) | External RCT | RR (risk) | 1 yeard (lifetime) |
Carande-Kulis (2015) [55] | Multiple types | Any fall (MA fall) | External RCTs | RR or RaR (risk) | 1 year (1 year) |
CSP (2016) [56] | Physiotherapy | Any fall (MA fall) | External meta-an. | RaR (risk) | 1 year (1 year) |
Church (2011) [57] | Multiple types | Any fall (any fall) | External meta-an. | RaR (risk) | Efficacy durability differ by int. type (10 years) |
Church (2012) [58] | Multiple types | Any fall (any fall) | External meta-an. | RaR (risk) | Efficacy durability differ by int. type (lifetime) |
Comans (2009) [59] | Multifactorial int. | Any fall (any fall) | External RCT | RaR (risk & rate) | 1 year (1 year) |
Day (2009) [60] | Multiple types | Any fall (any fall) | External RCTs | RaR (rate) | Efficacy durability same as model time (1, 2 or 5 years) |
Day (2010) [61] | Tai Chi | Any fall (any fall) | External meta-an. | RaR (rate) | 1 year (1 year) |
Deverall (2018) [62] | Multiple exercise types | Any fall (MA fall) | External meta-an. | RaR (risk) | Varying persistence (25 years) |
Eldridge (2005) [63] | FRAT; balance and gait int. | Any fall (any fall) | External meta-an. | RR (risk) | Not specified (lifetime) |
Farag (2015) [64] | Unspecified | Any fall (any fall) | Assumption | RR (risk) | Not specified (lifetime) |
Franklin (2019) [65] | Multiple types | Any fall (any fall) | External meta-an. and RCTs | RR and RaR (risk & rate) | 1 year (2 years) |
Frick (2010) [66] | Multiple types | Any fall (hip fracture) | External meta-an. | RR (risk) | 1 year (1 yeare) |
Hektoen (2009) [67] | Home exercise | Any fall (any fall) | External RCT | RaR (rate) | 1 year (1 year) |
Hiligsmann (2014) [68] | Vit. D + calcium supplement | Mix fracture; (mix fracture) | External meta-an. | RR (risk) | 6 yearsf (lifetime) |
Hirst (2016) [69] | Buprenorphine vs. Tramadol | Mix fracture (mix fracture) | External surveys | OR (risk) | 1 year (1 year) |
Honkanen (2006) [70] | Hip protector | Hip fracture (hip fracture) | External RCT | RR (risk) | Varying persistence (20 years) |
Howland (2015) [71] | Matter of Balance lay-led | MA fall (MA fall) | External RCT | RR (risk) | 1 year (1 year) |
Ippoliti (2018) [72] | Multifactorial int. | Hip fracture (hip fracture) | Policy variable | RaR (rate) | 3 years (3 years) |
Johansson (2008) [73] | Multifactorial int. | Hip fracture (hip fracture) | Internal quasi-experiment | RaR (risk) | 1 year (lifetime) |
Lee (2013) [74] | Vit. D screening & supplement | Any fall (any fall) | External meta-an. | RR (risk) | 2.5 years (3 years) |
Ling (2008) [75] | HAM | Any fall (any fall) | External RCT | RR (risk) | 1 year (1 year) |
McLean (2015) [76] | Exercise | Any fall (any fall) | Internal RCT | RR (risk) | 1.5 years (1.5 years) |
Miller (2011) [77] | Matter of Balance lay-led | Any fall (any fall) | Policy variable | RR (risk) | 2 years (2 years) |
Mori (2017) [78] | Exercise & bisphosphonate | Mix fracture (mix fracture) | External meta-analyses | RR or RaR (risk) | 1/2 year maintenance (lifetime) |
Moriarty (2019) [79] | Withdrawal of PIP mediations | MA fall/Hip fracture (MA fall/hip fracture) | External RCTs | RR (risk) | Lifetime persistence (35 years) |
Nshimyumukiza (2013) [80] | Exercise, Vit. D + calcium & osteoporosis int. | Mix fracture (mix fracture) | External meta-an. & surveys | RR (risk) | Lifetime sustainability (lifetime) |
OMAS (2008) [81] | Multiple types | Any fall (MA fall) | Internal meta-an. | RR (risk) | Lifetime persistence for 1st year adherers (lifetime) |
Pega (2016) [82] | HAM | Any fall (MA fall) | External meta-an. | RaR (risk) | Lifetime or 10-year efficacy (lifetime) |
Poole (2014) [83] | Vit. D supplement | Hip fracture (hip fracture) | External meta-an. | HR (rate) | 1 year (1 year) |
Poole (2015) [84] | Vit. D supplement | Any fall (MA fall) | External meta-an. | RR (risk) | 5 years maintenance (5 years) |
PHE (2018) [85] | Multiple types | Any fall (any fall) | External meta-an. and RCTs | RaR (rate) | 2 years (2 years) |
RCN (2005) [34] | Multiple types | Any fall (MA fall) | External meta-an. | RR (risk) | Not specified (lifetime) |
Expedited cataract surgery | Any fall (any fall) | Internal RCT | RaR (rate) | Lifetime efficacy durability (lifetime) | |
Smith (2016) [88] | Risk prediction; Multifactorial int. | Any fall (MA fall) | External meta-an. | RaR (risk) | 1 year (1 year) |
Tannenbaum (2015) [89] | Insomnia treatments | Any fall (any fall) | External surveys | OR (risk) | Not specified (1 or 5 years) |
Turner (2020) [90] | Sedative withdrawal | Hip/non-hip fracture (MA fall, hip/non-hip fracture) | External RCT | RaR (risk) | 1 year (1 year) |
Velde (2008) [91] | FRID withdrawal | Any fall (any fall) | Internal non-randomised | RaR (rate) | 1 year (1 yeare) |
Wilson (2017) [92] | HAM | Any fall (MA fall) | External meta-an. | RaR (risk) | Lifetime or 10-year efficacy (lifetime) |
Wu (2010) [93] | Multifactorial int. | Any fall (any fall) | External meta-an. | RR (risk) | 1 year (1 year) |
Zarca (2014) [94] | Vit. D screening & supplement | Vit. D level (vit. D level) | External meta-an. and RCT | Otherg | Varying persistence (lifetime) |
Wider health effects of interventions
Narrative synthesis: evaluation methods
Model validity
Assessing parameter uncertainty
Scenario analyses
Narrative synthesis: evaluation outcomes
Study label (n = 12)a | Target population; Analysis; Perspective | Intervention [Comparator] | Evaluation outcomesb | Methodological caveats |
---|---|---|---|---|
Church (2012) [58] | CD adults aged 65+; CUA/CEA; Public sector | (a) General population – Group exercise; Home exercise; Tai Chi; Multi-component int.; Multifactorial int.; Multifactorial risk assessment; (b) High-risk population – Group exercise; HAM; Multifactorial int. [NR; Cross comparisons] | Ratio: (a) General – Tai Chi ICER US$38,735 per QALY vs. NR; other interventions dominated; (b) High-risk – Group exercise ICER US$44,633 per QALY vs. NR; HAM ICER US$50,696 per QALY vs. NR; Multifactorial int. dominated. Aggregate: Reports incremental cost, no. of falls avoided and QALY gain per intervention, but all interventions have same reachc (including those targeting high-risk and specific subgroups), and hence cannot compare aggregate impacts. Parameter uncertainty: DSA – int. cost and efficacy had largest impact on group exercise ICER. PSA – CEAC. Scenarios: No fear of falling had the largest impact on group exercise ICER among parameter changes. | Recurrent falls not characterised; Unclear falls risk progression;d Unclear intervention reach;e Unclear how high-risk subgroup identified; Mismatch between falls incidence and efficacy metrics |
Deverall (2018) [62] | CD adults aged 65+; CUA; Public sector, Societal | Exercise – (i) Peer-led group exercise; (ii) Home exercise; (iii) Commercial exercise [NR] | Ratio: (i) Peer-led group exercise ICER US$6323 per QALY vs. NR; (ii) Home exercise ICER US$5486 per QALY vs. NR; (iii) Commercial exercise ICER US$46,733 per QALY vs. NR. Aggregate: For base case, home exercise generated 47,100 additional QALYs at incremental cost of US$225 m relative to NR; this compares to 42,000 and 42,300 QALYs for group exercise and commercial exercise at US$426 m and US$1550 m incremental cost relative to NR, respectively. Hence, home exercise dominated group and commercial exercises. Parameter uncertainty: DSA – efficacy and utility decrement had largest impact on ICER. PSA – 95% UI; CEAC. Scenarios: Subgroup analyses showed higher ICERs for Maori and men; equity analyses showed higher ICERs can be mainly attributed to their shorter life expectancies.e | Routine data lacks individual identifiers;f Recurrent falls not characterised; Unclear falls risk progression;d No background transition in health utilities;g Includes comorbidity care costs; Mismatch between falls incidence and efficacy metrics; No tiered threshold for evaluating societal outcomes;h No scenario estimating equity-efficiency trade-off.e |
Eldridge (2005) [63] | Adults aged 65+ in community or nursing home; CUA; Public sector | Falls risk screening + multifactorial int. or exercise [UC] | Ratio: Not reported. Aggregate: Intervention reduced the number of fallers by 2.8% over one year under base case (6.5% uptake of screening). Parameter uncertainty: PSA – 40% probability intervention is cost-effective at US$41,900 (£30,000) per QALY threshold. Scenarios: 100% screening uptake would reduce number of fallers by 11.3% over one year; 100% screening uptake and 100% self-referred exercise uptake would reduce number of fallers by 15.0%; impact of uptake increase on ICER not reported. | Recurrent falls not characterised; Unclear falls risk progression;d No background transition in health utilities;g Incorporated fixed intervention costs |
Farag (2015) [64] | CD adults aged 65+ without prior fall; CUA; Public sector | Non-specific falls prevention int. with relative risk of 0.75 and per-participant cost of US$587 [NR] | Ratio: ICER of US$24,190 per QALY vs. NR Aggregate: Incremental cost and QALY gain outcomes per person can be scaled up but unclear to what extent. Parameter uncertainty: DSA – falls risk and LTC cost had largest impact on ICER. PSA – CEAC; 57% probability of being cost-effective at AUS$50,000 (US$41,809) threshold. Scenarios: e.g., variation in uptake rate had little impact on ICER | Recurrent falls not characterised; Unclear falls risk progression;d No discounting |
Honkanen (2006) [70] | Adults aged 65+ living in community at baseline; CUA/ROI; Public sector | Hip protector [NR] | Ratio: Women, baseline age 65 – intervention dominated by NR; Women, age 70 – intervention dominated by NR; Women, age 75 – ICER of US$27,006 per QALY; Women, age 80 – intervention dominates NR; Women, age 85 – intervention dominates NR; Men, age 65 – intervention dominated by NR; Men, age 70 – intervention dominated by NR; Men, age 75 – intervention dominated by NR; Men, age 80 – ICER of US$184,609 per QALY; Men, age 85 – intervention dominates NR. Aggregate: Prevented fractures, incremental cost and QALY gain outcomes per person can be scaled up but unclear to what extent. Parameter uncertainty: DSA – base case results robust. PSA – 68% probability of being cost-effective at US$50,000 (US$710042021 price) threshold for women age 75; 61% for men age 85. Scenarios: Intervention is less cost-effective for functionally dependent subgroup – e.g., intervention no longer dominant for women age 80 and 85, though still cost-effective at US$71,004 threshold (point estimates not reported) | Unclear falls risk progression;d Includes comorbidity care costs. |
Johansson (2008) [73] | CD adults aged 65+ (n = 5500); CUA; Societal | Multifactorial and environmental int.i [UC] | Ratio: Intervention dominates comparator Aggregate: Total int. cost of US$895,137; total costs savings of US$904,986; total QALY gain of 35.16 Parameter uncertainty: No DSA. PSA – scatter plot Scenarios: Scenarios that made intervention no longer dominant – doubled fracture risk; lower fracture cost; inclusion of net consumption care cost;j higher discount rate; no health/cost consequences of fracture beyond 1st year; 25% rise in int. cost | Unclear falls risk progression;d Includes comorbidity care costs (net consumption); Quasi-experimental study for effectiveness evidence; No tiered threshold for evaluating societal outcomes;h Internal and external validities assessed |
Nshimyu-mukiza (2013) [80] | Women aged 65+ (subgroup within women aged 40+); CUA/CEA; Public sector | Fracture risk screening + Physical activity (PA), Vitamin D & calcium and/or Osteoporosis screening & treatment [NR; Cross comparisons] | Ratio: No screening + PA dominates NR; BMD/CAROC screening + PA + Vit D & calcium produces ICER of US$57,279 relative to No screening + PA and dominates all other strategies. Aggregate: Incremental cost and QALY gain per person can be scaled up (total population reported). Parameter uncertainty: No DSA. PSA – CEAC; 75% probability that BMD/CAROC +PA + Vit D & calcium is cost-effective to No screening + PA under threshold of CAD$50,000 (US$517892021 price). Scenarios: Rankings of strategies under CUA and CEA robust under variations in single or multiple parameters. | Incorporates incoming cohorts; No background transition in health utilities;g Structural and external validities assessed |
OMAS (2008) [81] | CD adults aged 65+; CEA/ROI; Public sector | (i) Exercise; (ii) HAM; (iii) Vit D & calcium; (iv) Gait stabiliser; (v) Psychotropics withdrawal.k [NR] | Ratio: All interventions dominate NR for men and women Aggregate: Reports net cost saving per person which can be scaled up to total for each intervention subgroup at regional level Parameter uncertainty: No analysis Scenarios: No analysis | Recurrent falls not characterised; Unclear falls risk progression;d Mismatch between intervention need and falls risk;k Parameter uncertainty not assessed |
Pega (2016) [82] | CD adults aged 65+; CUA; Public sector | HAM [NR] | Ratio: HAM produces ICER of US$7155 per QALY vs. NR. Aggregate: For base case, total int. cost was US$115.2 m, total net cost vs. NR US$87.4 m and total QALY gain 34,000. Parameter uncertainty: DSA – impact on ICER not assessed, fatal falls risk and falls risk most impactful for incremental cost and QALY, respectively. PSA – 95% UI for base case ICER between below zero to US$15,901 per QALY. Scenarios: For secondary prevention scenario,l ICER was US$1591 per QALY, total int. cost US$14.2 m, total net cost vs. NR, US$4.9 m, and total QALY gain 20,100. Subgroup analyses showed higher ICERs for Maori and men; equity analyses showed higher ICERs can be mainly attributed to their shorter life expectancies.e | Routine data lacks individual identifier;f Recurrent falls not characterised; Unclear falls risk progression;d No background transition in health utilities;g Includes comorbidity care costs; Mismatch between falls incidence and efficacy metrics; Unrealistic efficacy duration; Joint parameter uncertainty not assessed; No scenario estimating equity-efficiency trade-off.e |
RCN (2005) [34] | CD adults aged 60+; CUA; Public sector | Exercise; Multifactorial intervention [NR] | Ratio: Multifactorial intervention for high-risk group dominates NR; Exercise for high-risk group produces ICER of US$18,425 per QALY relative to NR. Aggregate: Not reported. Parameter uncertainty: No DSA. PSA – scatter plot Scenarios: No analysis | Recurrent falls not characterised; Unclear falls risk progression;d Unclear intervention reach.c |
Wilson [92] | CD adults aged 65+; CUA; Public sector | HAM [NR] | Ratio: HAM produces ICER of US$4358 per QALY vs. NR. Aggregate: For base case, total int. cost was US$7.7 m, total net cost vs. NR US$6.7 m and total QALY gain 2800. Parameter uncertainty: DSA – efficacy had largest impact on base case ICER. PSA – 95% UI for base case ICER between below zero to US$12,165 per QALY. Scenarios: For secondary prevention scenario,l ICER was US$557 per QALY, total int. cost US$687,151, total net cost vs. NR, US$72,626, and total QALY gain 1420. For primary prevention scenario,m ICER was US$7633 per QALY, total int. cost US$7.0 m, total net cost US$6.6 m and total QALY gain 1520. Subgroup analyses showed higher ICER for Maori; equity analyses showed higher ICER can be mainly attributed to Maori’s shorter life expectancy.e | Routine data lacks individual identifiers;f Recurrent falls not characterised; Unclear falls risk progression;d No background transition in health utilities;g Includes comorbidity care costs; Unclear intervention reach;c Mismatch between falls incidence and efficacy metrics; Unrealistic efficacy duration; Joint parameter uncertainty not assessed; No scenario estimating equity-efficiency trade-off.e |
Zarca (2014) [94] | Adults aged 65+ without previous hip fracture; CUA/CEA; Public sector | Vitamin D – (i) Universal supplementation; (ii) Supplement then screen for calibration; (iii) Screen then supplement [NR; Cross comparisons] | Ratio: Universal supplementation was dominated by other strategies; Supplement then screen strategy produces ICER of US$7758 per QALY vs. NR; Screen then supplement strategy produces ICER of US$7307 per QALY vs. Supplement then screen and US$7605 per QALY vs. NR. Aggregate: Difficult to compare strategies without data on intervention reach.c Possible that Screen then supplement strategy has smallest reach. Estimating total cost of Screen then supplement to be US$111.7 m for 800,000 persons. Parameter uncertainty: DSA – int. cost had largest impact on ICER of Screen then supplement vs. NR. PSA – 100% probability of Screen then supplement being most cost-effective strategy at threshold of €20,000 (US$297292021 price). Scenarios: Results robust to discount rates rising from 3 to 6%. | Hospitalisation cost only; Unclear intervention reach;c Structural, external and internal validities assessed |
-
Johansson [73]: Combined multifactorial and environmental intervention for age 65+
-
RCN [34]: Multifactorial intervention for high-risk group aged 60+
-
Nshimyumukiza [80]: General physical activity promotion among women (without population-level fracture risk screening) aged 65+
-
Honkanen [70]: Hip protector use for women aged 80 or 85 at baseline and men aged 85
-
Wilson [92]: HAM for state-level population with or without MA falls history aged 65+
-
Deverall [62]: Home exercise and peer-led group exercise for age 65+
-
Pega [82]: HAM for national population with or without MA falls history aged 65+
-
Zarca [94]: Vitamin D screening followed by supplementation for age 65+
-
RCN [34]: Exercise for high-risk group aged 60+
-
Farag [64]: Non-specific intervention of US$587 per-participant cost and 25% reduction in risk for age 65+
-
Church [58]: Tai Chi for age 65+