Table 3 Association between publication characteristics and methodological quality of SRs on total hip/knee arthroplasty

 2014

30 (47.6%)

6.40 ± 1.35

P = 0.375

 2015

33 (52.4%)

6.32 ± 1.19

 THA

25 (39.7%)

6.08 (5.46–7.09)^b

△P = 0.435

 TKA

37 (58.7%)

6.42 (5.58–7.50)^b

 THA and TKA

1 (1.59%)

–

 Chinese

15 (23.8%)

6.67 (5.17–6.08)^b

△P = 0.052

 English

48 (76.2%)

5.67 (5.67–7.50)^b

 Asia

50 (79.4%)

6.17 (5.5–7.19)^b

△P = 0.098

 America

9 (14.3%)

6.09 (4.34–7.84)^b

 Europe

4 (6.3%)

7.75 (6.37–9.5)^b

5、Number of SRs that included a PRISMA-like flow

 Included a PRISMA-like flow

53 (84.1%)

6.48 ± 1.20

P = 0.752

 Did not include a PRISMA-like flow

10 (15.9%)

5.73 ± 1.45

6、Was the SRs published in journal

 Number of SRs published in journal

59 (93.7%)

6.17 (5.50–7.50)^b

△P = 0.903

 Number of SRs not published in journal

4 (6.3%)

6.34 (5.62–7.29)^b

7、Tools for assessing risk of bias of primary studies

 Cochrane

35 (55.6%)

6.67 (5.50–7.50)^b

△P = 0.312

 Jadad scale

11 (17.5%)

5.75 (5.67–6.75)^b

 The PEDro scale

2 (3.2%)

7.46

 Other single assessment tool

7 (11.1%)

6.00 (4.5–7.00)^b

 Two or more assessment tools

6 (9.5%)

7.00 (5.63–8.13)^b

 Not reported

2 (3.2%)

5.45

Average(range)

8、Total number of authors in SRs

4.9 (1–8)

–

#Nonlinear relation

9、Number of databases searched

4.5 (1–12)

–

#Nonlinear relation

10、Median impact factor of the journal for which the included study was published

1.864 (0.293–5.228)

–

#P = 0.006^*