Days of Antibiotic Spectrum Coverage (DASC) as a Metric for Evaluating the Impact of Prospective Audit and Feedback (PAF) against Long-Term Broad-Spectrum Antibiotic Use

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for conducting this interesting research. I have few comments to be addressed:

1- Figure 1 needs more clarification. The two groups should be identified on the diagram.

2- Please include a comparison between the two groups regarding age and severity of disease or comorbidity score.

3- Were the infection treatment guidelines the same when used in the two periods. Please state this in the methods.

4- Please elaborate more on the meaning of ASC scores.

Author Response

Reviewer 1

 

1- Figure 1 needs more clarification. The two groups should be identified on the diagram.

 

→Thank you for your suggestions. The group name for each group was noted, and the items to be compared were adjusted to the same height.

 

2- Please include a comparison between the two groups regarding age and severity of disease or comorbidity score.

 

→Thank you for your suggestion. We collected data regarding age, sex, and the Charlson Comorbidity Index (CCI) to compare the background of patients between the two periods. Between the two periods, there was a significant difference in terms of sex, although no significant difference was observed in terms of age and CCI.

 

3- Were the infection treatment guidelines the same when used in the two periods. Please state this in the methods.

 

→Thank you for your suggestion. The JAID/JSC Guide to Infectious Disease Guidelines in Japan was not changed between the two periods. This has been added to the Material and Methods accordingly (lines 219–220).

4- Please elaborate more on the meaning of ASC scores.

 

→Thank you for your suggestion. We have added the explanation in the revised manuscript accordingly.

The ASC score is system used to evaluate the spectrum of antibiotics. The score categorizes clinically important bacterial pathogens into two domains: wild-type organisms without acquired resistance and commonly isolated microorganisms with specific mechanisms of acquired resistance. The wild-type category includes 11 organism groups, whereas the acquired resistance category includes 5. The ASC score was added based on the number of activities of each antibiotic against the categorized 16 strains. This explanation has been added to the Methods section (lines 253–259).

Reviewer 2 Report

Comments and Suggestions for Authors

Authors used a panel data analysis to examine the effect of introduction of “Hierarchical Medical System” on subjective self-rated health among 77,763 residents in China. However, this article has not fully answered some of the questions due to insufficient description and inadequate statistical analysis.

First, authors suggest “131 patients were excluded for the following reasons: planned termination, appropriate selection, poor status, and unknown causative bacteria” (L69), but they do not exclude these persons in baseline period, which may lead to biased results. Authors should add sensitivity analysis in which these persons are included.

Second, authors do not adjust potential confounding factors such as age and gender in this study, which may lead to biased results. Authors should add results after adjustment of potential confounding factors.

Finally, authors described the following sentences without citation “To overcome the limitation of DOT, Kakiuchi et al. proposed a new metric for antimicrobial consumption: days of antibiotic spectrum coverage (DASC).” (L54), but it is difficult for readers to judge it without references as evidence for each description. Authors should add references for these descriptions.

 

Minor comments

Table 1. “0.75 b Others” may be typo.

 

Author Response

Authors used a panel data analysis to examine the effect of introduction of “Hierarchical Medical System” on subjective self-rated health among 77,763 residents in China. However, this article has not fully answered some of the questions due to insufficient description and inadequate statistical analysis. First, authors suggest “131 patients were excluded for the following reasons: planned termination, appropriate selection, poor status, and unknown causative bacteria” (L69), but they do not exclude these persons in baseline period, which may lead to biased results. Authors should add sensitivity analysis in which these persons are included.

 

→Thank you for your valuable comments. We compared only the patients for the target intervention including 156 vs. 100 patients.

The pre-revised result showed a statistical decrease in DASC/patient and DASC/DOT for all antimicrobial agents, but the revised comparison showed no difference. The discussion has been revised to reflect the change in the results.

Second, authors do not adjust potential confounding factors such as age and gender in this study, which may lead to biased results. Authors should add results after adjustment of potential confounding factors.

 

→Regarding potential confounding factors, we collected age, sex, and Charlson Comorbidity Index (CCI) to compare the background of patients between two periods. Between the two periods, there was a significant difference in terms of sex, although no significant difference was observed in terms of age and CCI.

 

Finally, authors described the following sentences without citation “To overcome the limitation of DOT, Kakiuchi et al. proposed a new metric for antimicrobial consumption: days of antibiotic spectrum coverage (DASC).” (L54), but it is difficult for readers to judge it without references as evidence for each description. Authors should add references for these descriptions.

 

→Regarding the reference for the study by Kakiuchi et al., reference 13 has now been cited in the sentence

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Authors revised the manuscript, but this article has not fully answered some of the questions due to insufficient description and inadequate statistical analysis.

As mentioned in the previous review, authors do not adjust potential confounding factors such as age and gender in this study, which may lead to biased results. Authors suggest “Between the two periods, there was a significant difference in terms of sex”, but they did not adjust these effects. Authors should add results after adjustment of potential confounding factors.

Author Response

Authors revised the manuscript, but this article has not fully answered some of the questions due to insufficient description and inadequate statistical analysis.

As mentioned in the previous review, authors do not adjust potential confounding factors such as age and gender in this study, which may lead to biased results. Authors suggest “Between the two periods, there was a significant difference in terms of sex”, but they did not adjust these effects. Authors should add results after adjustment of potential confounding factors.

 

→Thank you for your suggestion.

The pre-intervention patient data was reviewed and reanalyzed.

There is no significant differences in patient background between two periods.

Trends in antimicrobial use did not change significantly, with a new significant decrease in DASC/patient for all antimicrobials, but no significant decrease in DASC/DOT, and no narrowing of the antimicrobial spectrum was obtained.

Due to changes in the number of patients in the pre-intervention period, other data were also adjusted, but no differences in trends were noted.

Major revisions are indicated in red.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

Authors suggest they revised the manuscript, but this article has not fully answered some of the questions due to insufficient description and inadequate statistical analysis.

As mentioned in the previous review, authors do not adjust potential confounding factors in this study, which may lead to biased results. Authors suggest “There is no significant differences in patient background between two periods.”, but they also suggest “Hematology was the most common hospital department targeted for intervention in both periods (39.9% vs. 22.0%), with a statistically higher rate in the baseline period than in the intervention period (p < 0.05). The most common types of infection in both periods were respiratory infections, such as empyema and obstructive pneumonia (43.6% vs. 31.0%, p = 0.05). The rate of urinary tract infections (3.0% vs. 9.0%, p < 0.05) was statistically higher in the intervention period than in the baseline period.” (L84). Authors should adjust for these factors. Moreover, authors also suggest “Regarding the patient background, there was no statistically significant difference in terms of sex, age and Charlson Comorbidity Index (CCI) between the two periods.” (L82), but p-value was 0.05 for sex, which may lead to biased results. Authors should add results after adjustment of potential confounding factors.

Author Response

As mentioned in the previous review, authors do not adjust potential confounding factors in this study, which may lead to biased results. Authors suggest “There is no significant differences in patient background between two periods.”, but they also suggest “Hematology was the most common hospital department targeted for intervention in both periods (39.9% vs. 22.0%), with a statistically higher rate in the baseline period than in the intervention period (p < 0.05). The most common types of infection in both periods were respiratory infections, such as empyema and obstructive pneumonia (43.6% vs. 31.0%, p = 0.05). The rate of urinary tract infections (3.0% vs. 9.0%, p < 0.05) was statistically higher in the intervention period than in the baseline period.” (L84). Authors should adjust for these factors. Moreover, authors also suggest “Regarding the patient background, there was no statistically significant difference in terms of sex, age and Charlson Comorbidity Index (CCI) between the two periods.” (L82), but p-value was 0.05 for sex, which may lead to biased results. Authors should add results after adjustment of potential confounding factors.

 

→Thank you for your suggestion.

We performed propensity score matching to adjust for patient backgrounds.

After the propensity score matching, 86 patients were selected from each periods, and the backgrounds were well-balanced between the two matched periods.

 

Major revisions are indicated in red.

Round 4

Reviewer 2 Report

Comments and Suggestions for Authors

Authors revised the manuscript, and I have no further comment.