Bridging the COVID-19 knowledge gap among cancer survivors: The need of the hour

The original article “Awareness about COVID-19 among cancer patients attending the radiation oncology department: A prospective observational study - ACROS” published in Cancer Research, Statistics, and Treatment, is interesting and provides significant insights.[1] The study aimed to assess the awareness of COVID-19 in terms of knowledge, attitude, and practice among cancer survivors visiting the Oncology Department of a medical college.[1] The authors have attempted to address an important domain, as the COVID-19 pandemic has led to a decline in healthcare-seeking behavior among cancer survivors.[2] I would like to provide several observations and suggestions that could improve both the internal validity (scientifically sound methodology with steps taken to minimize bias and confounding factors) and external validity (the generalizability of the study findings), and may be useful to other researchers as well. To begin with, the authors mentioned that they used a structured questionnaire in the study.[1] However, in the actual sense, it was a semi-structured questionnaire, as it included two open-ended questions (although, they were not included in the analysis). I also noted that the authors had used the abbreviation “ACROS” for the questionnaire, but no full form has been added.[1] Moreover, surprisingly, the term has not been used in the entire methodology section. As a general rule, titles should avoid abbreviations unless they are widely recognized. I appreciate that the authors conducted a pilot study to test the questionnaire before administering it to the study participants; however, it was not specified whether the findings from the pilot study were included in the final analysis. Ideally, the results of the pilot study should be excluded from the main study findings.[3] Additionally, I have a question regarding how the authors used an English-language questionnaire with a Bengali-speaking population. In such cases, the questionnaire should ideally be translated into the local language (viz. Bengali, and possibly Hindi – if any participants spoke Hindi) using a reverse translation method, to ensure linguistic and conceptual validity. This process would complete the validation of the questionnaire.[4] At present, the validation process is incomplete. The authors categorized responses on knowledge, attitude, and practice as “adequate knowledge” if participants answered a specific number of questions correctly in the respective domains. Was this an arbitrary classification, or was it based on a standard classification used elsewhere? If it was the latter, the authors should have cited a reference, and if it was the former, the authors should have written the term “arbitrary” (as it was just for this study). In addition, the authors mentioned that one category was “adequate knowledge”, what were the other categories? This has not been mentioned and neither the scoring required for the participants was mentioned. It is also mentioned that the authors did not include responses to open-ended questions in the analysis. I feel it was a missed opportunity, as these responses could provide in-depth knowledge about what participants are feeling, and it would have been a welcome approach to triangulate the study findings and enhance their validity.[5] These responses could have been presented in the form of a table, which would have added much weight to the study findings. I was also not convinced by the justification for not using a fixed sample size. For any study conducted for the first time, authors can take the P value as 50% and accordingly determine the estimated sample size, after adding 10% as a non-response error. This would have significantly increased the internal validity of the study. In continuation, the use of convenience sampling (non-random sampling) generally dilutes the strength of a study, and I strongly feel as authors had access to many patients, they could have easily opted for some form of random sampling. As the sample size in the current study was 280, I am not convinced why the authors used the Kruskal-Wallis test and Spearman’s correlation tests, which are non-parametric tests, and are typically used when the data does not follow a normal distribution. Additionally, it was quite surprising that authors used binary regression when it was not mentioned in their study objectives. Regression analysis is used to predict something and in the present study just after reading the objectives of the study, it was uncalled for and not suitable. In Table 2, the authors mentioned about Chi-square test, it should have been mentioned in the statistical analysis section as well. To summarize, the study adds meaningful information about the existing knowledge, attitudes, and practices regarding COVID-19 among patients with cancer. However, I strongly believe that the external validity of the study could have been significantly improved, if the authors had strengthened their methodology section. Financial support and sponsorship Nil. Conflicts of interest There are no conflicts of interest.