Table 2

Quality assessment and risk of bias in full reports of studies meeting eligibility criteria for the systematic review

First author, year publishedTotal *RoB †QoRQoSD
Badiani, 201713 50% (10/20)50% (3/6)57% (4/7)43% (3/7)
Bramley, 201412 63% (12/19)40% (2/5)71% (5/7)57% (4/7)
Dhillon, 200122 29% (5/17)100% (3/3)14% (1/7)57% (4/7)
Marvin, 201123 65% (11/17)67% (2/3)71% (5/7)71% (5/7)
Williams, 199424 35% (6/17)100% (3/3)43% (3/7)43% (3/7)
Average (mean) percentage48%71%51%54%
  • *Quality assessment was measured using the Axis tool, developed by Downes et al. 17 Depending on the study design, not all RoB items were relevant (ie, three RoB items pertain to non-response bias, and three of the studies did not recruit study participants since their aim was to only assess PMHS enquiries. Additionally, the study by Bramley 201412 recruited study participants with no non-responders, thus rendering one item obsolete). This accounts for the different maximum total scores and RoB scores across studies.

  • †The risk of bias items were reversed, so that the reported percentages reflect the amount of potential bias in each study. However, the Axis total score was calculated without reversing the risk of bias items, to ensure that the reported total score percentages reflect the amount of positively coded items in the tool. This accounts for the discrepancy between the total score and the sum of the subscales for each study.

  • PMHS, patient medicines helpline services; QoR, quality of reporting score (out of a maximum score of 7); QoSD, quality of study design score (out of a maximum score of 7); RoB, risk of bias score (out of a maximum score of 6).