Scheduled and Breakthrough Opioid Use for Cancer Pain in an Inpatient Setting at a Tertiary Cancer Hospital

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Introduction: I recommend adding a section on digital health interventions, such as using mobile apps for pain management, to provide a modern context to the evolving landscape of pain management.

Research Design: Highlighting the retrospective design's strengths, particularly its ability to analyze long-term trends through existing medical records, would clarify the choice of methodology.

Methods: A more detailed explanation of the decision-making process for opioid prescription adjustments, possibly including standardized pain assessment tools, could enhance this section.

Results: An expansion on the clinical significance of the findings, such as the impact of regimen adjustments on breakthrough pain episodes, would provide deeper insights.

Discussion and Conclusions: Suggesting future research directions, like investigating genetic factors in opioid efficacy and discussing the implications for clinical practice, such as guidelines on opioid rotation strategies, would be valuable additions.

I hope these suggestions, with practical examples, will help refine the manuscript.

Author Response

Reviewer 1

1. Introduction: I recommend adding a section on digital health interventions, such as using mobile apps for pain management, to provide a modern context to the evolving landscape of pain management.

Based on the reviewers comments we have now added a details of using digital health interventions in the context of pain management (page2, lines, 59-67).

“In the ambulatory setting pain assessment can be accomplished using pain dairy’s or more recently has been accomplished using digital health applications such as mobile apps for pain.  These mobile phone apps help in the regular pain assessments, provide timely feedback to patients and their clinicians, facilitate patient education, and help the physician to make timely medication changes and improve patient-physician communication.[17] However, there are still challenges which prevent routine use of mobile phone apps. Some of the main barriers are socioeconomic status, data protections, and evidence-based app validation.[18]”

2. Research Design: Highlighting the retrospective design's strengths, particularly its ability to analyze long-term trends through existing medical records, would clarify the choice of methodology.

Based on the reviewers comments we have now added details and rationale for the choice of retrospective design in this study (page 2&3, lines 95-101).

“For the conduct of this study whose goal was it capture the opioid prescription use of scheduled and breakthrough opioids in routine clinical practice at a tertiary cancer setting, we have used a retrospective design due to its ability to analyze long-term trends through existing medical records. This design was chosen to capture the real time practice patterns in cancer hospital settings. The retrospective design was also chosen as it was more feasible to obtain the outcomes aimed in our study than a prospective design due to resources, time and expense. [34]”

3. Methods: A more detailed explanation of the decision-making process for opioid prescription adjustments, possibly including standardized pain assessment tools, could enhance this section.

Based on the reviewers comments we have now added details of the process of supportive care consultation and decision-making process for opioid prescription adjustments (page 3&4, lines 106-158).

4. Results: An expansion on the clinical significance of the findings, such as the impact of regimen adjustments on breakthrough pain episodes, would provide deeper insights.

Based on the reviewer comments we have now revised the discussion section (page 9, lines 291-303; 372-375).

5. Discussion and Conclusions: Suggesting future research directions, like investigating genetic factors in opioid efficacy and discussing the implications for clinical practice, such as guidelines on opioid rotation strategies, would be valuable additions.

Based on the reviewers comments we have now added a paragraph in the discussion section suggesting future research directions (pages 9-10, lines 337-366).

“ Future studies are needed to optimize the use of opioids (scheduled and breakthrough) prescriptions to improve pain control and thereby overall patients’ quality of life. Some of the interventions to consider includes use of patient’s genetic data such single nucleotide polymorphisms of candidate genes such as inflammatory genes to determine sensitivity of a particular opioid type for a given patient, dose needed for a opioid to be effective, and assess the risk opioid related side effects.[48] In a recent study by our team, we assessed the genetic factors associated with pain severity, daily opioid dose, and pain response to opioids. The results of the study suggest single nucleotide polymorphisms of OPRM1, COMT, NFKBIA, CXCL8, IL-6, STAT6, and ARRB2 genes were significantly associated with pain severity, opioid daily dose and pain response.[48] Similar findings were found in other studies investigating the use of pharmacogenomics for personalized pain management.[49] Further studies are needed.

In recent years there has been increased use of Artificial Intelligence (AI) to provide patient care as well as in pain research. AI may have a potential to provide better pain control as in traditional pain assessment and management methods there is a high likelihood of variability of patient reported pain scores, perception of pain by different individuals, and algorithms for pain management which includes the use of opioids and other pain interventions. AI technologies such as machine learning, deep learning, and natural language processing have been used for pain assessment, surveillance and monitoring and opioid misuse risk prediction use. However, there is limited published research in the use of AI for pain management. [50-52] Future studies are needed in patients with cancer pain using AI and these should utilize the recent advances in pain assessments such as facial image analysis.[50] Better cancer pain management may be facilitated by using predictive clinical decision systems which incorporate patients’ clinical data, patient data obtained from wearable devices which assess pain, sleep and activity, and biomarkers as discussed above such as single nucleotide polymorphisms or various candidate genes which may predict the sensitivity to certain opioids, response rates of a given pain type to an opioid or other pain treatment. However, the use of AI may only supplement clinician decision making processes for the management of cancer pain rather than replace them due to their inherent limitations.”

Reviewer 2 Report

Comments and Suggestions for Authors

The authors try to compare opioid prescribing patterns. I do not understand how the authors defined the categories pre-supportive care and supportive care.

Major concerns

#1. I do not understand how the authors defined the categories pre-supportive care and supportive care (Figure 1). Why the numbers are different in Figure 1? 362 patients were divided to 330 and 292 patients?

#2. If the authors used repeatedly measured data, the authors need to use statistical analysis methods for repeated measure data.

#3. I do not understand study design. Please follow the EQUATOR guidelines (https://www.equator-network.org/).

Comments on the Quality of English Language

None

Author Response

Reviewer 2

Major concerns

#1. I do not understand how the authors defined the categories pre-supportive care and supportive care (Figure 1). Why the numbers are different in Figure 1? 362 patients were divided to 330 and 292 patients?

Based on the reviewers comments we have now defined the presupportive, supportive care and early supportive care categories and added rationale for the definition (page 4, lines 159-170). We have also now clarified the numbers analyzed in various categories, i.e., data of patients seen after 72 hours after admission but prior to referral to supportive care (presupportive care, pain managed by the primary oncology team); supportive care group: data of patients seen after 72 hours after admission after referral to supportive care, and data of patients seen prior to 72 hours after admission and referred to supportive care (page 5, lines 222-230).

#2. If the authors used repeatedly measured data, the authors need to use statistical analysis methods for repeated measure data.

Based on the reviewers comments we have clarified the analysis of the data (both categorical and continuous data) and the tests used in the statistical section and in the footnote of the Tables 1-3.

#3. I do not understand study design. Please follow the EQUATOR guidelines (https://www.equator-network.org/).

 We have now clarified that we utilized the retrospective study design to descriptively understand the the frequency and prescription pattern of breakthrough and scheduled opioids and its ratio (BTO ratio) used by prior and after referral to inpatient supportive care consult (SCC) for cancer pain management. The rationale for retrospective design was its ability to analyze long-term trends through existing medical records.

We have not used designs described in https://www.equator-network.org/ which include designs for randomized controlled trials, observational studies, systematic reviews, prognostic studies, qualitative studies, care reports, qualitative studies, quality improvement studies, clinical practice guidelines, animal pre-clinical studies, and economic evaluations.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for the diligent revisions made to your manuscript following initial feedback. Upon reviewing the updated version, it is clear that you have thoroughly addressed the critical areas of concern.

The inclusion of digital health interventions within the introduction provides a modern context that enriches the paper.

Your detailed rationale for the retrospective study design, along with the expanded decision-making process for opioid prescription adjustments, adds depth and clarity to your methodology.

The extension of the results/discussion section offers greater insight into the clinical significance of your findings, and the discussion on future research directions, especially concerning genetics and AI, sets an intriguing path forward for the field.

The revisions have significantly enhanced the manuscript, and it now presents a comprehensive and up-to-date perspective on pain management in cancer care. 

Author Response

The authors sincerely thank Reviewer for their kind reviews and comments so as to improve the manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors improved the manuscript very slightly. I’m very confused by figures and tables. Do pre-supportive care and supportive care groups include same patients?

Major concerns

#1. I’m very confused by figures and tables. Do pre-supportive care and supportive care groups include same patients? Why the numbers are different in Figure 1? 362 patients were divided to 330 and 292 patients?

#2. Again, if the authors used repeatedly measured data, the authors need to use statistical analysis methods for repeated measure data.

#3. Again, I do not understand study design. Please follow the EQUATOR guidelines (https://www.equator-network.org/).

Comments on the Quality of English Language

None.

Author Response

Reviewers’ comments

#1. I’m very confused by figures and tables. Do pre-supportive care and supportive care groups include same patients? Why the numbers are different in Figure 1? 362 patients were divided to 330 and 292 patients?

 We sincerely thank for the reviewers’ comments. We have now revised the manuscript. We have revised Figure 1 and the Methods section of the manuscript (Page 5, lines 230-236), to clarify that the same patients were used in the pre-supportive care group and supportive care groups.

“Figure 1 shows the study flow diagram. 665/728 (91%) patients were evaluable. 362 patients were referred to supportive care service 72 hours after of hospitalization, and data was compared before (pre-supportive care, n= 330) and after (n=292) referral to supportive care. Due to absence of opioid use, 32 and 38 patients were excluded from analysis from pre-supportive care and supportive care groups, respectively. 355 of the 366 patients referred to supportive care service before 72 hours of hospitalization (early supportive care) were analyzed.  Eleven patients were excluded due to absence of opioid use.”

#2. Again, if the authors used repeatedly measured data, the authors need to use statistical analysis methods for repeated measure data.

We sincerely thank for the reviewers’ comments. We have now revised the manuscript and the Tables to clarify that we have used Wilcoxon signed-rank test is used to examine the change on continuous variables within a group, i.e., paired data (e.g., pre-supportive and supportive care group). Wilcoxon rank-sum test is used to examine the difference on continuous variables between groups ( i.e., for the comparison between early (before 72h of hospitalization) and late (after 72h of hospitalization) supportive care referral groups.

#3. Again, I do not understand study design. Please follow the EQUATOR guidelines (https://www.equator-network.org/).

Based on the reviewers comments we have now clarified the study design. We have now added the SQUIRE checklist as per the EQUATOR guidelines.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

The authors improved the manuscript. The authors show that BTO ratio was frequently prescribed higher than the recommended dose. I’m not sure whether the information is clinically important.