Table 2.
Summary of results
| First author, year | Outcome measured | Effect of PMWS | ||
|---|---|---|---|---|
| Positive | Non-significant | Negative | ||
| Gattrell, 2016 [15] | Adherence to CONSORT guidelines |
The proportion of articles that completely reported at least 50% of the assessed CONSORT items • With PMWS: 43/110 articles (39.1%; 95% CI 29.9–48.9) • Without PMWS: 26/123 articles (21.1%; 95% CI 14.3–29.4) |
||
| Jacobs, 2010 [16] | Logistic regression analysis showed that CONSORT items were significantly more likely to be completed in papers with a clear acknowledgement of PMWS than in those without (OR 1.44; 95% CI 1.04–2.00; p = 0.03) | |||
| Shah, 2015 [20] | 23/97 articles with PMWS (24%) had 80–100% CONSORT adherence, whereas 5/105 articles developed without PMWS (5%) had 80–100% CONSORT adherence (p < 0.0001) | |||
| Mills, 2017 [14] | Adherence to CONSORT-A guidelines | The mean proportion of CONSORT-A items reported was similar with and without PMWS (64.3% vs 66.5%, respectively; p = 0.30); PMWS was associated with a lower level of compliance with reporting of study setting (RR 0.40; 95% CI 0.23–0.70) and a higher level of adherence to disclosure of harms/side effects (RR 2.04; 95% CI 1.37–3.03) and funding source (RR 1.75; 95% CI 1.18–2.60) | ||
| Gattrell, 2016 [15] | Quality of written English |
Proportion of articles rated by all reviewers during peer review as having an acceptable standard of written English • With PMWS: 81.1% (43/53 articles; 95% CI 67.6–90.1) • Without PMWS: 47.9% (23/48 articles; 95% CI 33.5–62.7) |
||
| Gattrell, 2016 [18] | Publication in journal with an impact factor | Likelihood of publication in journal with an impact factor was significantly improved with PMWS (p = 0.001) | ||
| Mean impact factor of publication | Mean impact factor of publication was significantly improved with PMWS (p < 0.001) | |||
| Gattrell, 2017 [17] | Reporting of non-pre-specified outcomes | Articles developed with PMWS reported fewer non-pre-specified outcomes than both industry-funded (p = 0.028) and non-industry-funded articles (p < 0.01) developed without PMWS | ||
| Gattrell, 2016 [18] | Mean number of citations per year | Mean number of citations per year was not significantly improved with PMWS (p = 0.11) | ||
| Mean number of article views per year | Mean number of article views per year was not significantly improved with PMWS (p = 0.84) | |||
| Altmetric score | Altmetric score was not significantly improved with PMWS (p = 0.55) | |||
| Gattrell, 2016 [15] | Manuscript acceptance time | Time from manuscript submission to acceptance was increased with PMWS (167 days [IQR 114.5–231 days] vs 136 days [IQR 77–193 days], p < 0.01); mean number of versions submitted was unchanged | ||
| Shah, 2016 [19] | Time to publication | Time to publication from last patient visit in clinical trials was reduced with PMWS (18.6 [SD 13.2] months vs 30.8 [SD 11.7] months) | ||
| Woolley, 2006 [8] | Manuscript acceptance time | Time from manuscript submission to acceptance was reduced with PMWS (83.6 days vs 132.2 days), although this difference was not statistically significant (p = 0.053) | ||
CI confidence interval, CONSORT Consolidated Standards of Reporting Trials, CONSORT-A CONSORT for Abstracts, IQR interquartile range, OR odds ratio, PMWS professional medical writing support, RR relative risk, SD standard deviation