Skip to main content

Table 1 Quality assessment and summary of findings

From: Total hip arthroplasty versus hemiarthroplasty for displaced femoral neck fractures in the healthy elderly: a meta-analysis and systematic review of randomized trials

Quality assessment Summary of findings Importance
No. of patients Effect Quality
No. of studies Design Limitations Inconsistency Indirectness Imprecision Other considerations Total hip Hemiarthroplasty Relative risk (95 % CI) Absolute
One-year mortality
6 a Randomized trials Very seriousb No serious inconsistency c Seriousd Seriousa,e Nonec 53/393 (13.5 %) 64/423 (15.1 %) RR 0.91 (0.65–1.27) 14 fewer per 1000 (from 53 fewer to 41 more) Very low Critical
13.6 % 12 fewer per 1000 (from 48 fewer to 37 more)
Revision surgery
8 c Randomized trials Very seriousb No serious inconsistency c Seriousd Seriousf Nonec 19/472 (4 %) 36/514 (7 %) RR 0.59 (0.32–1.09) 29 fewer per 1000 (from 48 fewer to 6 more) Very low Important
7.1 % 29 fewer per 1000 (from 48 fewer to 6 more)
Dislocation
6 g Randomized trials Very seriousb No serious inconsistency c Seriousd Serioush,i Nonec 33/369 (8.9 %) 14/411 (3.4 %) RR 2.53 (1.05–6.1) 52 more per 1000 (from 2 more to 174 more) Very low Important
0 % 0 more per 1000 (from 0 more to 0 more)
Major complications
5 j Randomized trials Very seriousb No serious inconsistency c Seriousd Seriousj,k Nonec 76/302 (25.2 %) 80/330 (24.2 %) RR 1.07 (0.76–1.5) 17 more per 1000 (from 58 fewer to 121 more) Very low Important
8.2 % 6 more per 1000 (from 20 fewer to 41 more)
Minor complication
5 l Randomized trials Very seriousb No serious inconsistency3 Seriousd Seriousl,m Nonec 38/302 (12.6 %) 45/330 (13.6 %) See comment 10 fewer per 1000 (from 60 fewer to 40 more) Very low Important
7 % 5 fewer per 1000 (from 31 fewer to 20 more)
  1. CI confidence interval, RR relative risk
  2. a Two out of eight studies did not adequately provide number of deaths after one-year follow-up
  3. b Allocation concealment: 3/8 study used sealed envelopes, 1/8 hospital number, 2/8 computerized, 1/8 order of admission, 1/8 did not specify blinding: none of studies blinded the patients, only 3/8 studies report on a blinded outcome assessor failure to adhere to the intention to treat principle: 5/8 studies
  4. c No explanation was provided
  5. d In the different trials, different approaches and materials, e.g. cement vs uncemented were used. This may have had some effect, e.g. pain, function or dislocation
  6. e Total (cumulative) sample (size =117) is lower than the calculated optimal information size (OIS) (64/423 = 0.15-- > needed: RR 25 %: 500)
  7. f Total (cumulative) sample (size =55) is lower than the calculated optimal information size (OIS) (36/514 = 0.07-- > needed: RR 5%: 600)
  8. g Two out of eight studies did not adequately provide information on dislocation rates
  9. h Two out of eight studies did not provide clear numbers of dislocation at all
  10. i Total (cumulative) sample (size =47) is lower than the calculated optimal information size (OIS) (14/411 = 0.03-- > needed: RR 25%: 600)
  11. j Three out of eight studies did not adequately provide information on major complications
  12. k Total (cumulative) sample (size = 156) is lower than the calculated optimal information size (OIS) (45/330 = 0.24-- > needed: RR 5%: 500)
  13. l Three out of eight studies did not adequately provide information on minor complications
  14. m Total (cumulative) sample (size =83) is lower than the calculated optimal information size (OIS) (45/330 = 0.1-- > needed: RR 5 %: 500)