Use of Internet Technology among Older Adults in Residential Aged Care Facilities: Protocol for a Systematic Review and Meta-Analysis

Round 1

Reviewer 1 Report

This article describes the protocol for a systematic review (SR) of Internet use among older adults in residential age care facilities (RACF). The motivation for this review originates from continuous growth in Internet use among this specific group of older adults. Whereas the majority of previous research either focused on the entire older population or community-dwelling older adults, the number of studies examining the RACF context has increased recently. Therefore, now it is the time for a systematic appraisal of the accumulating evidence.

Overall, I think the proposed SR is relevant and the method appropriate. I’d like to point to some issues and questions, and suggest possible improvements as follows:

1) The question addressed by the SR should be specified more clearly. While the method section states “Internet usage” and “factors associated with Internet usage”, the abstract does not mention “factors” but “patterns” of use. I suppose the main interest is in the evidence for factors explaining Internet use (e.g., through testing of hypotheses in explanatory modeling, or qualitative studies) rather than the prevalence rates of Internet use. Prevalence rates (percentages, means, standard deviations) are of little interest and cannot be compared across different settings (time, region, study population). The list of items presented in the Data Extraction section should be revised accordingly.

2) The review should be limited to a much shorter time span. Including studies from the 1990s and 2000s is useless because of outdated technology and extremely low adoption rates. Study results from the “old days” provide little insights into technology adoption to inform policy makers and providers of digital health services. I would suggest to begin in 2010 instead of 1990. Please note that Facebook, for instance, had only 100 million users in 2008, when the first Android-based mobile phone became available. The first iPad was introduced in 2010. Adoption among older adults was very slow compared to younger generations, and uptake in RACF even smaller. The technological advances should be considered in the review.

3) Exemplar studies should be referenced in the Introduction section, which would better illustrate the relevance and scope of the review (going beyond the Australia/New Zealand context). Reference No. 5 should be removed because it is not relevant (outdated technology, see above).

4) Information sources: Is the inclusion of nine databases necessary? For instance, Scopus has greater coverage than Web of Science; hence, the latter could be removed. PubMed includes some of the other databases. The selection of sources should be justified at least.

5) Search strategy: Because the theme is Internet use, terms that are not related to Internet technology should be removed from the query (which would also reduce the number of non-relevant entries). The current list of terms is too long and quite different from terms used in previous reviews. I would drop the following terms: information technology, ICT (which is too generic, might include desktop computers with no Internet); computer communication network (same as above, term not used in the older adults literature; computers, handheld, wireless technology, computer, cell phone (all not specific to the research question); text messaging, SMS (does not use Internet technology but GSM); wifi, wireless (again, not specific).

6) Search strategy: “Older adults” must be added because it is the most used term in the literature. I suggest to replace “older people” in the title and throughout the manuscript.

7) Risk of Bias Assessment: The references no. 20 and 21 should be removed because they are too specific and the retraction problem is of minor relevance for the review.

8) Meta-analysis: The article should provide a better justification for this step. For explanatory research using regression modeling (which I expect to be present in studies), meta-analysis is hardly possible because of heterogeneity in the research models and study settings. The text in Section 2.7.1 is quite generic and should be aligned to the research question.

Author Response

We have attached the point by point response to reviewers' comments as a word document.

Author Response File: Author Response.pdf

Reviewer 2 Report

The overall manuscript is relevant for existing literature and has great potential. However, I am concerned about the quality of the analyses because the manuscript—despite being set up for a protocol—lacks scientific rigor and is, in my opinion, send in a premature state. The criteria are listed for a systematic review, but the criteria are absent for the meta-analysis. You will find articles without a control group, unclear information, or any other indication of poor quality. With its current form, I cannot tell if you have a plan ready. Without these, the quality of your methodological is lacking. I miss crucial details.

Comments for author File: Comments.pdf

Author Response

We have attached the point by point response to reviewers' comments as a word document.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for your efforts in addressing the earlier comments and suggestions. Overall, I believe the quality of the manuscript has significantly improved. Let me point to a few remaining issues though:

Research question: Thank you for including the factors associated with internet use in your analysis and list of data items. Having said this, I still think that this statement should be much more specific. First, I suppose that most of studies to be assessed aimed at enhancing the understanding of internet use, thus belong to explanatory research (rather than predictive or design-oriented research). The current manuscript does not refer to “explanation” at all. I would revise some phrases in the abstract, introduction and throughout in this respect. Second, in extracting data from studies, it is not sufficient to identify factors associated with internet use (in terms of statistical significance). What is needed is to identify *all* factors that have been tested (usually through including them in a linear regression model), mark those factors that were associated (based on statistical significance), note the strength of association (usually through regression coefficients as a measure of effect size and their CI; care should be taken whether the coefficients are unstandardized or standardized), check whether interactions have been tested, and report the explanatory power of the tested model (usually through adjusted r-squared) for each study (if applicable).

Exemplar studies: My earlier comment was related to the specific target group, thus older adults living in RACF. Thus, I would drop the two studies that included mostly community-dwelling older adults [4, 5], and add one or two RACF studies to the second paragraph, e.g., https://doi.org/10.1080/03601277.2017.1326224

Information sources: I understand your reasoning for selecting the various databases and administering a very comprehensive query term including rather generic terms. My only concern was that the great number of duplicate records and non-relevant articles might cause to much effort on the part of the coders (I agree that it does not affect the final set of articles deemed eligible).

Author Response

Dear Editor,

We would like to thank our peer-reviewer for their extensive review of our manuscript and feedback.

Attached is our point by point response to our reviewers' feedback. 

Regards,
Sandesh

Author Response File: Author Response.pdf

Reviewer 2 Report

Well done with the addition of the information about the meta-analysis. 

Author Response

Thank you.