Understanding the Challenges of HPV-Based Cervical Screening: Development and Validation of HPV Testing and Self-Sampling Attitudes and Beliefs Scales

Round 1

Reviewer 1 Report

This is a well-written paper about creating a scale to measure attitudes and beliefs about HPV DNA testing via self-sampling. Minor revisions are recommended before accepting this manuscript.  The revisions are listed below (including minor edits).

·       Measures: Please include a brief description of why you selected the Theory of Planned Behavior and Health Belief Model to guide the selection of your scales.

·       Please briefly describe how you identified the seven individuals who helped ensure comprehension and clarity of your survey items.

·       Why were the 203 responses included?  Were they missing data?  What percent of items were incomplete (if so)?

·       Discussion: It would be interesting to see how the authors intend to use this scale.  Do they plan on using this before a self-sampling intervention – to gauge interest?  Or, will it be used to explain why intervention participants chose to/not to screen via self-sampling?

Minor edits:

Page 2, Line 49 – Make international health organizations plural

Page 2, Line 68 – add to the sentence ending in accurate beliefs and positive attitudes something along the lines of “of this screening method”

Page 2, Line 70 - hyphenate HPV-based

Page 2, Lines 71-75 – Maybe change to “Barriers to HPV testing have included increased screening intervals and …”

Page 4, Line 173 – “confirmatory” does not need to be capitalized

Appendices/Results:

·       Appendix A is Missing Item 45 (that you kept instead of 31)

·       The numbering may be slightly off in either the Results or Appendix A/D. For example – the paper has that items 35 and 43 performed similarly. Appendix A has these questions as Items 34 and 42 – but the items are correctly numbered in Appendix D.

Author Response

This is a well-written paper about creating a scale to measure attitudes and beliefs about HPV DNA testing via self-sampling. Minor revisions are recommended before accepting this manuscript.  The revisions are listed below (including minor edits).

We appreciate the reviewer’s positive feedback regarding our manuscript and their time in providing thoughtful recommendations.

• Measures: Please include a brief description of why you selected the Theory of Planned Behavior and Health Belief Model to guide the selection of your scales.

We agree with the reviewer that describing why these two frameworks were used is important. A previous systematic review conducted by our team (https://doi.org/10.1016/j.ypmed.2018.08.034) identified that many of the factors that influence women’s acceptability of HPV-based cervical screening can be mapped onto constructs from the Health Belief Model and Theory of Planned Behavior, We have updated the manuscript accordingly:

“To identify items to include in the survey, an extensive literature review of existing measures was conducted which identified 13 relevant scales. A pool of 781 items was then created from these existing measures as well as from items and themes identified in the literature (cite Tatar and McBride). Informed by the results of a systematic review conducted by our team (18) that used  the Health Belief Model (HBM) and Theory of Planned Behavior (TPB) to map factors associated with HPV test acceptability, these items were then categorized into two potential scales relating to attitudes and beliefs about HPV testing and self-sampling”

• Please briefly describe how you identified the seven individuals who helped ensure comprehension and clarity of your survey items.

We agree with the reviewer that this is an important detail to include. The following addition has been made (italics and underline):

“To ensure comprehension and clarity, items were tested and revised in cognitive interviews in both English and French with seven Canadians who met the study’s inclusion criteria. Participants were recruited using advertisements placed on relevant Canada-based social media groups (e.g., “Montreal Moms”, “McGill Psychology Students Association”).”

• Why were the 203 responses included?  Were they missing data?  What percent of items were incomplete (if so)?

We thank the reviewer for bringing attention to this important detail. The 203 responses were excluded as they were identified as being potentially ‘inattentive’ or ‘unmotivated’ according to methods for identifying careless responses. These methods included an “instructed response” in which a question is used that specifies a participant should choose a specific response option to verify their attention (with those who answer otherwise being identified as potentially inattentive), the identification of “straight-liners”, or those who answer with one response option across a broad selection of questions, and by identifying extreme outliers in response time (in our study, the high and low 2.5% of response times). The use of these methods is important to ensure high data quality from survey responses and is detailed in full in our previously published protocol referenced in the paper (https://www.researchprotocols.org/2022/6/e38917). To clarify further, the following change (in italics and underline) has been made to the sentence in question:

“In total, 1230 participants completed the survey and, after data cleaning methods were applied to identify potentially inattentive or unmotivated respondents [41], 203 responses were excluded (see protocol paper [28] for a detailed overview of data cleaning methods).”

• Discussion: It would be interesting to see how the authors intend to use this scale.  Do they plan on using this before a self-sampling intervention – to gauge interest?  Or, will it be used to explain why intervention participants chose to/not to screen via self-sampling?

We agree with the reviewer that discussing the use of this scale is important, however the purpose of this paper is to provide a sufficient overview of the scale development and validation for other researchers to use the scales in their investigations. Regarding the suggested examples, we believe the scale could be applied to either. Our research team has now completed a follow-up to this study in which the scales were administered to a population-based sample of Canadian women to identify psychosocial correlates of previous screening behaviors and future screening intentions. This is detailed in the published protocol paper referenced above. 

Minor edits:

Page 2, Line 49 – Make international health organizations plural

Page 2, Line 68 – add to the sentence ending in accurate beliefs and positive attitudes something along the lines of “of this screening method”

Page 2, Line 70 - hyphenate HPV-based

Page 2, Lines 71-75 – Maybe change to “Barriers to HPV testing have included increased screening intervals and …”

Page 4, Line 173 – “confirmatory” does not need to be capitalized

Response: All the aforementioned, minor edits have been implemented.

Appendices/Results:

• Appendix A is Missing Item 45 (that you kept instead of 31)
• The numbering may be slightly off in either the Results or Appendix A/D. For example – the paper has that items 35 and 43 performed similarly. Appendix A has these questions as Items 34 and 42 – but the items are correctly numbered in Appendix D.

We thank the reviewer for bringing these errors to our attention. We did not include initially items 28 (HPT testing) and 11 (HPV self-sampling) in Appendix A, which were the directed response attention check questions used to validate respondents’ attention for data cleaning. These items have been added to Appendix A and the item numbers the reviewer indicated were in error in the paper are now aligned with Appendix A and all other appendices. We have also added a note to each of the attention check questions in Appendix A (i.e., “[ATTENTION CHECK QUESTION]”) and included in the note of the tables that these items (item 28 in the HPV Testing items and item 11 in the HPV self-sampling items) were not used for analyses. The following addition has also been made to the ‘measures’ section (italics and underlined) to clarify why the item numbers extend to 45 while only 44 items were included in the analyses for the HPV testing attitudes and beliefs scale:

“In total, 44 items related to HPV testing attitudes and beliefs and 13 items related to HPV self-sampling attitudes and beliefs were retained for inclusion in the questionnaire. In each section we added  one attention check item to facilitate the identification of  “inattentive” responses during data cleaning. These items can be found in Appendix A.”

Reviewer 2 Report

This paper describes an in-depth analysis of a questionnaire regarding attitudes towards HPV testing and self sampling. Examining reasons for not attending cervical cancer screening and identifying women that are interested in performing self sampling but not a test performed by a medical doctor are very timely research questions as countries are starting to implement self sampling as part of their cervical cancer screening programs. 

A few modifiactions of the paper could make the findings and the implications of the findings more clear for the readers.

1) The EFA resulted in four and two factors, for the HPV testing attitiudes and beliefs scales (HTABS), and the HPV self-sampling attitudes and beliefs scales (HSABS), respectively. These were named "Personal barriers", "Social Norms", "Confidence" and "Worries" (HTABS) and "Concerns" and "Autonomy" (HSABS). Especially for a reader not so familiar with EFA, it would be good to explain that these labels are inferred based on the items (questions) in each of the factors (and that these not were question categories made upfront). I think it also would be easier to read the Results section, if a sentence or two could be spent on describing each of these factors at the beginning of this section.

2) In the conclusion it is stated "In addition, examining attitudes and beliefs about self-sampling could enable effective targeting of this screening approach that has shown promise in increasing uptake among those who were previously unengaged with screening." But nowhere in the manuscript could I find anything about how to target such women (i.e women that would perform self sampling, but not testing at a medical doctor). It is quite crucial that such a targeting is feasible, in order for the findings in this study to be of relevance. 

Some minor issues:

3) Table 1 and 2 would probably be easier to read if the left column was left aligned

4) There seems to be some footnote-marks lacking in the Tables (e.g after EGA in Table 2). 

5) The word "sample" in the sentence in the Introduction: "a sample of cervical screening-eligible Canadians participated in a web based survey" is a bit unfortunate as sample otherwise refers to a test samples (e.g self sample). 

Author Response

This paper describes an in-depth analysis of a questionnaire regarding attitudes towards HPV testing and self sampling. Examining reasons for not attending cervical cancer screening and identifying women that are interested in performing self sampling but not a test performed by a medical doctor are very timely research questions as countries are starting to implement self sampling as part of their cervical cancer screening programs. 

We thank the reviewer for their positive feedback regarding the purpose of our study.

A few modifiactions of the paper could make the findings and the implications of the findings more clear for the readers.

1) The EFA resulted in four and two factors, for the HPV testing attitiudes and beliefs scales (HTABS), and the HPV self-sampling attitudes and beliefs scales (HSABS), respectively. These were named "Personal barriers", "Social Norms", "Confidence" and "Worries" (HTABS) and "Concerns" and "Autonomy" (HSABS). Especially for a reader not so familiar with EFA, it would be good to explain that these labels are inferred based on the items (questions) in each of the factors (and that these not were question categories made upfront). I think it also would be easier to read the Results section, if a sentence or two could be spent on describing each of these factors at the beginning of this section.

We thank the reviewer for this important suggestion. We have made a revision to address this concern in the methods section. The following sentence has been added to explain that factor names were decided after reviewing the factor analysis results:

“Each subscale was named according to conceptual similarities of the final retained items.”

While we understand the reviewer’s suggestion of addressing the development of labels at the beginning of the results, we feel that the current order of the results facilitates an understanding of the stepwise process used for factor analysis, item reduction, and scale validation. We believe that the above added sentence in the methods and the existing discussion section should address the reviewer’s concerns.

2) In the conclusion it is stated "In addition, examining attitudes and beliefs about self-sampling could enable effective targeting of this screening approach that has shown promise in increasing uptake among those who were previously unengaged with screening." But nowhere in the manuscript could I find anything about how to target such women (i.e women that would perform self sampling, but not testing at a medical doctor). It is quite crucial that such a targeting is feasible, in order for the findings in this study to be of relevance. 

We thank the reviewer for this thoughtful comment. The reviewer correctly indicates that the sentence is missing detail about how targeting would occur and how our scales would be implicated in this. We believe the use of these scales in investigations will help to reveal attitudes and beliefs that are associated with uptake of self-sampling. Messaging interventions could be designed to address existing barriers of uptake. This would be particularly useful in populations with lower screening rates (e.g., immigrants, gender and sexual minorities, rural communities). The sentence in question has been changed as follows:

“In addition, examining attitudes and beliefs about self-sampling could inform targeted messaging strategies highlighting benefits and addressing concerns related to this novel screening approach, especially in inadequately screened populations.”

Some minor issues:

3) Table 1 and 2 would probably be easier to read if the left column was left aligned

We thank the reviewer for their comment but were not sure of what would need to be done to make this change. Should the manuscript be accepted for publication, the tables will be harmonized as per journal’s style.

4) There seems to be some footnote-marks lacking in the Tables (e.g after EGA in Table 2). 

We thank the reviewer for this observation and have updated the footnotes in tables 2 and 5.

5) The word "sample" in the sentence in the Introduction: "a sample of cervical screening-eligible Canadians participated in a web based survey" is a bit unfortunate as sample otherwise refers to a test samples (e.g self sample). 

We have removed “a sample of” in the abstract as the sentence retains its intended meaning without.

Reviewer 3 Report

They developed the scales to identify patient psychology, which is a potential barrier to the introduction of HPV self-screening, and demonstrated its effectiveness. The scale is very specific and clear, and this is even innovative; it may be very useful for the introduction of HPV self-screening, and I would like to refer to it in Japan, where the introduction of HPV self-screening has been lagging behind.

I would consider the content of this paper as sufficient to be accepted for publication in Current Oncology.

Author Response

They developed the scales to identify patient psychology, which is a potential barrier to the introduction of HPV self-screening, and demonstrated its effectiveness. The scale is very specific and clear, and this is even innovative; it may be very useful for the introduction of HPV self-screening, and I would like to refer to it in Japan, where the introduction of HPV self-screening has been lagging behind.

I would consider the content of this paper as sufficient to be accepted for publication in Current Oncology

We thank the reviewer for their kind words and endorsement of our submitted manuscript.