Table 2. Description of the included studies and pooled analyses.
| Review category | Author and year (Ref)* | Intervention | Study population (Sex) | Number participants randomized (Institutions) | Follow-up duration (Months) | Number of Hip Fractures in Treatment (T) and Control (C) groups (Percentage) | Pooled Odds Ratio (OR) of Hip Fracture in Intervention Group Compared to Controls † (Number of participants and trials) |
| Vitamin D or calcium | Chapuy 1994 [25] | Vitamin D3 800 IU+1.2 g calcium daily | All from nursing homes (All female) | 3270 (180) | 36 | T: 138/1634 (8.4%) C: 184/1636 (11.2%) | Oral vitamin D: Pooled OR 0.86 (95% credibility interval, 0.74, 0.98) (Pooled n = 12,875 in 7 trials) |
| Chapuy 2002 [27] | Vitamin D3 800 IU+1.2 g calcium daily | All from nursing homes (All female) | 583 (55) | 24 | T: 27/393 (6.9%) C: 21/190 (11.1%) | ||
| Flicker 2005 [28] | Vitamin D2 10,000 IU/week or 1,000 IU per day+0.6 g calcium daily | Nursing home and other institutions (Male and female) | 625 (149) | 24 | T: 3/313 (1.0%) C: 7/312 (2.2%) | ||
| Law 2006 [29] | Vitamin D2 2.5 mg every 3 months (equivalent to 1,100 IU/day) | Nursing home and other institutions (Male and female) | 3717 (118) | 10 (Median and mean) | T: 24/1762 (1.4%) C: 20/1955 (1.0%) | ||
| Lyons 2007 [30] | Vitamin D2 2.5 mg three times per year (About 833 IU/day) | Nursing home and other institutions (Male and female) | 3440 (314) | 36 | T: 127/1725 (7.4%)‡ C: 126/1715 (7.3%)‡ | ||
| Meyer 2002 [31] | Vitamin D3 400 IU daily in 5 ml cod liver oil | All from nursing homes (Male and female) | 1144 (51) | 24 | T: 50/569 (8.8%) C: 47/575 (8.2%) | ||
| Sato 2005 [32] | Vitamin D2 1,100 IU daily | Chronic geriatric hospitalization (Females, post-stroke with hemiplegia) | 96 (1) | 24 | T: 0/48 (0%) C: 4/48 (8.3%) | ||
| Sato 2005 (Sunlight) [33] | Outdoor sunlight exposure in clear weather for 15 minutes daily 1.2 g calcium daily | Chronic geriatric hospitalization, Alzheimer's Disease (all females) | 264 (1) | 12 | T: 2/132 (1.5%) C: 9/132 (6.8%) | Sunlight exposure: Pooled OR 0.43 (95% confidence interval 0.10, 1.83) (n = 522 in 2 trials)λ | |
| Sato 2003 (Sunlight) [34] | Outdoor sunlight exposure in clear weather for 15 minutes daily | Chronic geriatric hospitalization, post-stroke hemiplegia (Males and females) | 258 (1) | 12 | T: 1/129 (0.8%) C: 6/129 (4.7%) | ||
| Drug therapies | Greenspan 2002 [35] | Alendronate 10mg orally daily+Vitamin D 400 IU daily (type not specified) +Calcium carbonate (OsCal 500, if dietary calcium <1500 mg/day) | Nursing home and other institutions (All female) | 327 (25) | 24 | T: 2/163 (1.2%)ϕ C: 4/164 (2.4%)ϕ | OR 0.50 (95% confidence interval 0.09, 2.75) λϕ (n = 327 in 1 trial) |
| Alternative medicines | Inkovaara 1975 [36] | Sodium monofluorophosphate (25 mg of fluorine daily for first 5 months, then 25mg twice a week) | Municipal home for the aged (Males and females) | 460 (1) | 8 | T: 8/237 (3.4%) C: 5/223 (2.2%) (Femoral) | OR 1.52 (95% confidence interval 0.49, 4.73) λϕ (n = 460 in 1 trial) |
| Exercise, behavioral or multimodal | Becker 2002 [37] | Multimodal intervention: Staff and resident education, balance and resistance training, hip protectors | All in nursing homes (Males and females) | 981 (6) | 12 | T: 17/509 (3.3%) C: 15/472 (3.2%) | Relative Risk 1.11 (95% confidence interval 0.49, 2.51) λ∂ (Adjusted for cluster randomization) (n = 981 in 1 trial) |
| Cox 2003 [38] | Multimodal intervention including staff education and individual risk assessment with feedback and advice to physicians | Nursing home and other institutions (Male and female) | 6229 (58 PCO clusters) | 12 | T: Not reported/3476 C: Not reported/2753 | Relative Risk 0.86 (95% confidence interval 0.63, 1.18) λ∂£ (Adjusted for cluster randomization) (n = 5637 residents in 1 trial) | |
| Jensen 2002 [39] | Multimodal intervention: Staff education, environmental modification, exercise program, drug modification (for falls prevention), select specialist referral, hip protectors | All in nursing homes (Males and females) | 402 (9) | 11 week intervention with 34 week follow-up | T: 3/194 (1.5%) C: 12/208 (5.8%) | Unadjusted OR 0.25 (95% confidence interval 0.05, 1.13)£ λ∂ OR Adjusted for Baseline Variables 0.23 (95% confidence interval 0.06, 0.94)£ ∂μ (All adjusted for cluster randomization) (n = 384 in 1 trial) | |
| Sakamoto 2006 [40] | Exercise: Daily supervised uni-pedal standing balance exercise | Nursing home and other institutions (Male and female) | 553 (32) | 6 | T: 1/337 (3.0%) C: 1/216 (4.6%) | Not significant (p>0.999) λ∂£ (n = 527 in 1 trial) | |
| Hip protectors | Sawka 2007 (Systematic review, data from 4 trials) [22] (Ekman 1997, Harada 2001, Jantti 1998, Meyer 2003) | Hip protectors (Shields on both hips) | All nursing home residents (Harada - all female, others - male and female) | 1,922 (Ekman – 4, Harada – 6, Jantti - 1, Meyer 42) | Range 11–15 | Ekman: T: 4/302 (1.3%) C: 17/442 (3.8%) Harada: T: 1/88 (1.1%) C: 8/76 (10.5%) Jantti: T: 1/36 (2.8%) C: 7/36 (19.4%) Meyer: T: 21/459 (4.6%) C: 42/483 (8.1%) | Hip Protectors Pooled OR 0.40 (95% credibility interval 0.27, 0.56) (Pooled n = 2,594 individuals in 5 trials) |
| Koike 2009 [45] | Hip protectors (shields on both hips) and leaflet on fracture prevention | All nursing home residents (or equivalent)‡ (All female) | 672 | 12–26σ | T: 9/345 (5.5%) C: 39/327 (11.9%) |
*Ref is the reference number listed at the end of this paper.
†Intention to treat analysis unless otherwise indicated.
‡The primary author confirmed through e-mail correspondence that all residents had onsite 24-hour per day nursing care available.
σTime from allocation to treatment to end of follow-up period (allocation in January, 2004 and March, 2005 to end of follow-up in March, 2006).
λ A Bayesian random effects meta-analysis could not be performed for this data.
ϕCalculated.
∂Results reported in the primary paper.
£The treatment effect reported by the primary author is not an intention to treat analysis as the number randomized is higher than the number for which the statistical test was performed.
μThe baseline variables adjusted for in this model included: Mini-Mental Status Examination score, Barthel index score, physical restraints and delirium, sex and history of falls, and age.