The Association of Context with Reported Self-Efficacy for Cancer-Preventive Behaviors and Perceived Cancer Risk in U.S. Adults from the Midlife in the United States (MIDUS) Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for the opportunity to review this manuscript, which describes a study examining neighborhood quality and conditions with perceived cancer risk and self-efficacy. This is a good use of secondary dataset, however I have a few issues/comments which need to be addressed.

 1. Last line of page 1 and first two lines of page 2 need to be fixed – incomplete sentence. “Social scientists have identified that beliefs of personal attributes as dynamic, and malleable, are associated with perceptions of personal agency of changing current circumstances via effortful behavior. [9] which have been applied to health domains.”

2. There is a lack of citations and detail about the health belief model. Page 2, lines 6-8: before this paragraph a brief introduction about health belief model (HBM) and constructs would be beneficial. Both perceptions of cancer risk (as perceived susceptibility) and self-efficacy are components of the HBM and including a paragraph about the model and the need to look at modifying factors to individual constructs is needed. Citation: Champion, V. L., & Skinner, C. S. (2008). The health belief model. Health behavior and health education: Theory, research, and practice, 4, 45-65.

3. MIDUS should be Midlife in the United States – please remove development from the title.

4. Why was wave 3 used? Were survey items available in other waves? If so justification is needed for the use of only this wave.

5. Income and marital status should be added as a covariates in the model. MIDUS assess both at each wave and are included.

6. Please put possible scores (min and max) and what higher scores indicate (more or less of measure) for all surveys used. These makes interpreting means for participants easier for readers.

7. Little’s test for missing completely at random (MCAR) is needed. If data are MCAR then listwise deletion and complete case analysis is justifiable to use. If data are not MCAR (p<.05 for little) than listwise deletion can result in biased results.

8. Covariates paragraph: Please add alpha for positive relations with others. Also, why was only one dimension of psychological well-being used instead of all six dimensions. This should be clarified, and language should be used carefully to indicate whether a single dimension was used or all six dimensions.

9. Standardized betas need to be reported. Can remove or keep unstandardized values, but standardized values should be reported.  

10. Table 3, significant values need to be bolded to stay aligned.

Author Response

"Please see the attachment."

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting study that is focused on a relatively new area of research. Considering the introduction, strong statistical methods used, and results presented, I was expecting much more in the discussion and definitely missed the “conclusions and recommendations” section.

Following are some comments and suggestions that could increase the quality of the manuscript and contribution of your study in the advancement of this topic in the scientific literature.

Participants: As you did not state the inclusion criteria for your sample, it is necessary to clarify the definition of “midlife” and the age range of participants in your study. According to the reference #40, MIDUS 2 respondents were aged 35 to 86, and according to ref #46, midlife participants were 40-59 years old (national sample).

Diversity of sample population: According to your discussion, “This study examined associations among multilevel determinants of health behaviors… in a diverse sample” (p.7) and also you state that “this is the first study to examine perceived neighborhood condition and cancer risk perception and cancer prevention self-efficacy using a large diverse national sample” (p8). According to your data (Table 1) your sample is majority Whites (82%), followed by Blacks (13.4%). It is recommended to instead of a “diverse sample” write “a majority White sample”, so readers do not have wrong expectations about the results.  

It calls the attention that the “Black” participants in your study may be those from the Milwaukee African American Sample, that were added to the original MIDUS database (p.3) just to increase “diversity”? As you are controlling all results by race, you could be missing important differences in this group. Perhaps conducting a separate analysis using only data from this second group will enhance your results, more specifically because you did not include any reference about publications regarding this group.

It would be also great to include in the demographics (table 1) the distribution of your sample by different USA regions and/or include a paragraph clarifying the “national” representation of this sample.

Perceived cancer risk (PCR): Although you adjusted for age, race/ethnicity, and family history of cancer, there is a lack of discussion about evidence of relationship between high risk of cancer for elders, blacks, and those with family history. Also, there is a lack of discussion that, regardless of neighborhood perceptions, education may influence knowledge about cancer risk, and, as result, perception about cancer risk. Running a separate analysis for each one of these variables vs PCR may highly enhance your discussion and results

Self-efficacy for cancer prevention: Although you adjusted for education, there is a lack of discussion about evidence of relationship between educational level and self-efficacy. It also called my attention that you did not include income in the demographics. There is a lack of discussion that, regardless of neighborhood perceptions, income may influence access to resources, and as result, self-efficacy.

Perceived quality of neighborhood: When looking at the items to measure this construct, it called my attention that the questions refer more to “safety” than “quality”. In the sources you cite (#41 and # 42), they refer to “perceived trust in neighbors” and “perceived neighborhood safety”. The name you chose “quality” give the perception that it includes different variables. According to the U.S. Census Bureau, the definition of Neighborhood Quality is proxied by crime, pollution, and access to food, although they are also considering the inclusion of other variables such as poverty, employment, and housing conditions (https://www.census.gov/content/dam/Census/library/working-papers/2020/demo/SEHSD-WP2020-08.pdf). It is highly recommended to review the name of your construct and/or add this topic in the discussion.

Social integration. The measure is confusing. Answering the question “I don’t feel I belong to anything I’d call a community” is very confusing because the question is negative and the scale moves from agree to disagree. Although, it looks that “strongly disagree (value=7) implies that “I feel I belong to a community” (higher score implies higher integration), depending of the literacy level of respondents, there may be bias when they were answering this question. In the other hand, the other 2 items “I feel close to other people in my community” and “My community is a source of comfort”, are stated in a positive way, but because of the scale, “strongly agree (value = 1)” they have to be reverse-coded. According to your ref #40, the item that was reverse-coded was the one asking “I don’t feel I belong to anything I’d call a community” what makes more sense than reversing the other two ones. Please clarify why you did not follow the same methodology here.

Models: The layout of the tables 2 and 3 is confusing since the columns and rows (quality, integration and built) are the same. Please reorganize to eliminate the empty cells, which would facilitate the visual look of the results. It is highly recommended to report the full model values in a separate table including the results for each dependent variable (perceived cancer risk, and perceived efficacy) the coefficients, SE, 95%CI, Wald X2, and p value for the intercept and each one of the explanatory variables (neighborhood quality, integration and built). It calls the attention the lack of discussion about the contributions of each explanatory variable in the model while the objective of this study was to “examine associations of neighborhood factors with cancer- risk perception and prevention self-efficacy.” (abstract).

Limitations: Considering the lapse of years in data collection, please add in the limitations that perception between participants completing surveys in different years may have been influenced along time

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Firstly, I would like to congratulate you on choosing an extremely interesting topic, namely the impact of sociodemographic characteristics and contextual factors in the mitigation of cancer burden and the development of adequate health and social policies at national levels, based on multivariate analysis of targeted populations which includes the perceived cancer risk and the self-efficacy. Cancer continues to be the second cause of death after cardiovascular events and every measure we ca propose in order to reduce its burden it is of extreme importance to the scientific community and should be of interest to all parties involved in the development of integrative health policies. The manuscript is well structured, the aim is clearly stated, the methodology is clearly presented and results are sound in my opinion.

Secondly my suggestions for the improvement of submitted manuscript are as follows:

·        There are a few mistakes of writing (i.e., on page 2 second row after reference [9] “which have been applied to health domains” is disjointed and appears after to not have a subject or active verb). Please address that.

·        As a non-native English speaker, I have to tell you that some of the phases are quite convoluted and very difficult to read. I think that, taking into account that this is an international journal with a global audience, it would be advisable to  try and simplify some of the phrases since it would be conducive to better transmitting the authors message to a larger public.

·        In my opinion the study has the limitation of not describing anything about the actual access to health care preventative measures (screening, vaccination, public health facilities, coverage of health insurance) and other social motivational facilities conducive to personal well-being and healthier behavior changes of study population (social centers, information campaigns, therapist and so on). In would have been very interesting to see if the “perceived quality of neighborhood” is in fact correlated to various social inequities. Also, there is no information on level of income of study population or perceived lacking of financial means – this can seriously affect the results for self-efficacy and can detrimentally affect the participation in preventive programs (For exemplification I suggest Simion, L.; Rotaru, V.; Cirimbei, C.; Gales, L.; Stefan, D.-C.; Ionescu, S.-O.; Luca, D.; Doran, H.; Chitoran, E. Inequities in Screening and HPV Vaccination Programs and Their Impact on Cervical Cancer Statistics in Romania. Diagnostics 2023, 13, 2776. https://doi.org/10.3390/diagnostics13172776). Such information would have been interesting and for me it would have completed and overall well-conducted study. Its effects on perceived cancer risk ad self-efficacy, and may be a future diretion of study.

Comments on the Quality of English Language

The quality of the English used in the text is adequate for scientific communication and requires only small editing corrections. The phraseology used is a bit complicated and difficult to follow - it is desirable to simplify it to facilitate the transmission of the message to an international audience.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript is greatly improved with the revisions. There are just a few minor issues with formatting and editing that I now have. 

1. I would change the title as it is a bit misleading, especially for those reading quickly. It seems perceived cancer risk and self-efficacy can prevent cancer, which is not accurate and not what the study is about. 

2. Formatting of all tables is all over the place and it makes it difficult to look at and read. 

Suggestions:

- Tables 1, 2, & 3: Left align variables in the left column for tables 1, 2, and 3 and tab over levels of categorical variables (as in Table S2)

- Table 1: Should not need to take up the entire page from right to left

- Tables 2 & 3: Instead of ‘Ref.’ I would puta dash ‘–‘ to improve readability

- Tables 2 & 3: Can use abbreviations for variables in Tables 2 and 3 to help with variable column length, will make it look neater and improve readability

- Titles for Tables 1, 3, & S1 have ‘(N=2,450)’ while title for table 2 and S2 does not comma in number ‘(N=2450)’. Pick one and be consistent

- Table S1 needs to be fixed so it is similar like the other tables and in publication format

- Table S1 needs the word Notes and explanation of * for p-values

- Table S2 seems to be going off the page on the right side

- Table S2, typo in notes at end explaining p-values: “* p < 0.05, ** p < 0.01, *** p < 0.001.0.01, *** p < 0.0”

- All tables should be proofed very carefully. For example Table 3 has female in model one as “0.14*” but is not bolded as well as reported positive relations in model one as “0.02** (0.01)” but it is not bolded. Also, family history of cancer in model one has “-.02” and is bolded with no stars. Make sure significant values are consistent.

3. Lastly, font of refences should be the same as the rest of the manuscript.

Author Response

Please see attachment.

Author Response File: Author Response.docx