Skip to main content

Table 5 Summary of findings for KQ1a on the benefits and harms of screening compared with no screening

From: Screening for the primary prevention of fragility fractures among adults aged 40 years and older in primary care: systematic reviews of the effects and acceptability of screening and treatment, and the accuracy of risk prediction tools

Study approach

Population; studies; sample size

Follow-up (y)

Assumed population riske

Absolute effects

Certaintyf

What happens?

Hip fractures

All eligible / offer-to-screen

Females 45–54 y [90]

The evidence from 1 RCT (n=2979) is very uncertain.

VERY LOWa-d

Very uncertain

Females 68–80 y

1 RCT; 34,229 [5]

5

Study data:

25 per 1000

0.3 fewer in 1000

(4.2 fewer to 3.9 more)

LOWa-c

May not reduce

General:

20 per 1000

0.2 fewer in 1000

(2.4 fewer to 2.2 more)

Acceptors of screening

Females 45–54 y [90]

The evidence from 1 RCT (n=2604) is very uncertain.

VERY LOWa-d

Very uncertain

Offer-to-screen in selected populationg

Females ≥65 y

3 RCT+1 CCT; 43,736 [4,5,6, 68, 91]

3 to 5

Study data:

31 per 1000

6.2 fewer in 1000

(9.0 fewer to 2.8 fewer)

MODERATEc

Probably reduces

General:

20 per 1000

4.0 fewer in 1000

(5.8 fewer to 1.8 fewer)

Males ≥70 y [68]

The evidence from 1 CCT (n=1380) is very uncertain.

VERY LOWa-d

Very uncertain

Clinical fragility fractures

All eligible / offer-to-screen

Females 45–54 y [90]

The evidence from 1 RCT (n=2979) is very uncertain.

VERY LOWa-d

Very uncertain

Females 68–80 y

1 RCT; 34,229 [5]

5

Study data:

100 per 1000

1.0 fewer in 1000

(8.0 fewer to 6.0 more)

LOWa-c

May not reduce

General:

168 per 1000

1.7 fewer in 1000

(13.4 fewer to 10.1 more)

Acceptors of screening

Females 45–54 y [90]

The evidence from 1 RCT (n=2604) is very uncertain.

VERY LOWa-d

Very uncertain

Offer-to-screen in selected populationg

Females ≥65 y

3 RCT; 42,009 [4,5,6, 91]

3 to 5

Study data:

84 per 1000

5.9 fewer in 1000

(10.9 fewer to 0.8 fewer)

MODERATEc

Probably reduces

General:

168 per 1000

11.8 fewer in 1000

(21.8 fewer to 1.7 fewer)

All-cause mortality

All eligible / offer-to-screen

Females 45–54 y

1 RCT; 4800 [90]

9

Study data: The evidence is very uncertain.

VERY LOWb,d

Very uncertain

General:

3 per 1000

No difference in 1000

(0.8 fewer to 1.1 more)

LOWb,d

May not reduce

Females 68–80 y

1 RCT; 34,229 [5]

5

Study data:

118 per 1000

3.5 fewer per 1000

(9.4 fewer to 3.5 more)

LOWb,d

May not reduce

General:

57 per 1000

1.7 fewer per 1000

(4.6 fewer to 1.7 more)

Offer-to-screen in selected populationg

Females ≥65 yh

2 RCT+1 CCT; 26,511 [4, 6, 68]

3 to 5

Study data:

89 per 1000

No difference in 1000

(7.1 fewer to 5.3 more)

MODERATEd

Probably does not reduce

General:

57 per 1000

No difference in 1000

(4.6 fewer to 5.1 more)

Serious adverse events

Offer-to-screen in selected populationg

Females 70–85 y [6]

The evidence from 1 RCT (n=12,483) is very uncertain.

VERY LOWa,b,d

Very uncertain

Health-related quality of life/Wellbeing

All eligible / offer-to-screen

Females 45–54 y [90, 92]

9 (self-rated health)

2 (SF-36)

NA

The evidence from 1 RCT (n=2979) is very uncertain.

VERY LOWa-c

Very uncertain

Offer-to-screen in selected populationg

Females 70–85 y

1 RCT; 10,661 [6]

5

NA

SF-12 (range 0–100):

Mental health: MD −0.30, 95% CI −0.86 to 0.26

Physical health: MD 0.30, 95% CI −0.21 to 0.81

EuroQol-5D (range 0–1):

MD 0, 95% CI −0.07 to 0.07

LOWa,b

May be little to no difference

Overdiagnosis

Offer-to-screen in selected populationg

Females 70–85 y (1 RCT; 6,233) [6] 14.4 × (100 − 17.9) /100 = 11.8% overdiagnosed

Females 65–90 y (1 RCT; 5575) [4] 25.4 × (100 − 23.9) / 100 = 19.3% overdiagnosed (selected higher-risk population)

Among those considered at high risk

Females 70–85 y (1 RCT; 3064) [6] 29.3 × (100 − 17.9) / 100 = 24.1% overdiagnosed

  1. CCT Clinical controlled trial, CI Confidence interval, RCT Randomized controlled trial, MD Mean difference, NA Not applicable, y years
  2. aRisk of bias; binconsistency; cindirectness; dimprecision
  3. e Study data refers to the median control events rates across trials, which is the main analysis. A sensitivity analysis used the effects without screening for the general risk population in Canada, estimated from PRIOR et al. (Bone. 2015;71:237-43) based on 10-year follow-up
  4. f When our assessment of the certainty of evidence fell between levels, we assigned the level that best represented our actual certainty
  5. g Selected population defined as those who completed the initial risk assessment tool (as part of 2-step screening). This population may be more accepting of screening and have higher compliance than the general (intention-to-screen) population
  6. h This analysis included 1379 men from Kern 2005, representing 5.4% of the total sample