Effects of providing manuscript editing through a combination of in-house and external editing services in an academic hospital

Abstract

Background: English editing services are effective for improving manuscript quality as well as providing learning opportunities for non-native English-speaking authors. Herein, we describe the effects of a combined system of in-house and external editing services for handling large volumes of editing requests and providing personalized editing service in academic hospitals.

Methods: We established the Scientific Publications Team (SPT), an in-house editing team in Asan Medical Center in Seoul, Korea. The SPT is composed of two professional editors who manage editing requests sent to external companies while also providing one-on-one in-house editing services. We gathered author satisfaction data from 936 surveys between July 2017 and December 2018 and analyzed the number of editing requests and research publications by segmented regression analysis of interrupted time series data.

Results: The SPT processed 3931 editing requests in 2017-2018, which was a marked increase compared with prior to its establishment (P = 0.0097). The authors were generally satisfied with the quality of editing services from both in-house and external editors. Upon conducting regular quality control, overall author satisfaction with one external company gradually increased over the course of one year (P for trend = 0.086). Author satisfaction survey results revealed that overall satisfaction of editing service was most strongly correlated with how well the edits conformed to the authors' intentions (R = 0.796), and was only weakly correlated with quick turnaround time (R = 0.355). We also observed a significant increase in the trend of the number of research publications (P = 0.0007) at one year after the establishment of the SPT.

Conclusion: Providing a combination of in-house and external editing services resulted in high author satisfaction and subsequent hospital-wide increases in manuscript writing and publication. Our model system may be adapted in academic hospitals to better address the editing needs of non-native English-speaking researchers.

Fig 1. Manuscript editing workflow.

Boxes with bold texts indicate the routine workflow in processing editing requests. Boxes in faded colors indicate processes that are infrequently carried out.

Fig 2. Trend in the number of editing requests.

Trend of editing requests from AMC researchers from 2011 Q1 to 2018 Q4. (A) Quarterly number of editing requests. Numbers on top of the bars indicate the actual number of editing requests made during the respective quarters. Bars in orange indicate the period after which the Scientific Publications Team was established. (B) Trend in the number of editing requests. Asterisks (**) indicates a significant (P = 0.0097) difference in the slope of the editing request trend after intervention (establishment of the Scientific Publications Team; January 2017). (C) Number of requests that asked for in-house editing. Numbers on top of the bars indicate the actual number of editing requests that specifically asked for one of the in-house editors. Data from June 2018 was omitted because the in-house editors were on leave for most of the period, during which the number of editing requests was 2.

Fig 3. Overall satisfaction scores.

Overall satisfaction scores of the EECs and the SPT from author satisfaction surveys gathered from July 2017 to December 2018. Overall satisfaction scores from (A) blind surveys and (B) non-blind surveys. Numbers on top of the bars indicate the mean values, and numbers in parentheses indicate the number of survey results. Error bars indicate standard deviation from mean. * P < 0.05 vs. SPT, # P < 0.05 vs. EEC 4; ## P < 0.01 vs. EEC 4; **** P < 0.0001 vs. SPT. EEC, external editing company; SPT, Scientific Publications Team.

Fig 4. Specific question scores.

Scores of specific questions (items #2 –#7 in Table 1) from blind survey results. Questions that showed (A) strong, (B) moderate, and (C) weak correlation with overall satisfaction score. Column bar graphs show the mean scores of each question (numbers on top of the bars); error bars indicate 95% confidence range in “Re-request” and “Recommend,” and standard deviation from mean in others. Scatter plots show the correlations between overall satisfaction and the specific questions. Trend between two variables are indicated by the solid lines. Shaded areas between dotted lines indicate 95% confidence range. * P < 0.05 vs. SPT, # P < 0.05 vs. EEC 4; **, ##, †† P < 0.01 vs. SPT, EEC 4, and EEC 2, respectively; ##, P < 0.001 vs. EEC 4; **** P < 0.0001 vs. SPT.