Impact of military service on physical health later in life: a qualitative study of geriatric UK veterans and non-veterans

Objectives Military veterans often experience physical health problems in later life; however, it remains unclear whether these problems are due to military service or are a feature of the ageing process. This study aimed to explore veteran and non-veteran perceptions of the impact of their occupation on their physical well-being later in life. Design Semi-structured qualitative interviews analysed using thematic analysis. Setting Interviews were conducted face-to-face in participants’ homes or via telephone. Participants 35 veterans (≥65 years), 25 non-veterans (≥65 years) were recruited, as well as a close companion of all participants for triangulation (n=60). Results Most veterans reported good physical health later in life which they attributed to the fitness they developed during military service. However, several veterans described challenges in maintaining their desired level of physically activity due to new commitments and limited sports facilities when they left service. Fewer non-veterans had experienced work-related fitness activities or exercise in their civilian jobs. Ongoing physical health difficulties, such as deafness, were perceived to be due to exposure to workplace hazards and appeared more common in veterans compared with non-veterans. Veterans also described greater reluctance than non-veterans to seek medical treatment for physical health difficulties, which could be challenging for close companions who had to provide informal care. Conclusions Military service was largely perceived to be beneficial for physical well-being; although when occupation-related physical health problems were experienced, many veterans were unwilling to seek treatment. These findings may inform clinicians of the needs of older veterans and highlight potential barriers to care.

This manuscript presents the results of a qualitative study examining the impacts of military service on physical health among older veterans. The study has an impressive sample size for a qualitative work and gathered information from both comparative (non-veteran) and corroborative samples (caregivers). However, while the research question is an interesting and important one and the paper has potential, there are some questions, particularly around the methods and the manner of presenting the results, which need to be addressed. METHODS: • First, the authors need to be sure they are following the COREQ guidelines for reporting on qualitative research. In particular, justification for why a qualitative approach was chosen for answering this question and how the sample size was determined need to be provided. • Likewise, justification for the groups chosen for the interviews needs to be provided. Why were veterans separated into those with and without mental health diagnoses? This does not seem to be central for the research question being addressed, and should be explained further. • On page 7 in the interview procedure section, the authors state that thematic saturation was achieved. How was this determined?
• Providing more context of the larger study would be helpful. It is unclear from the procedures (pg 7) what type of clinical context participants were recruited from, and how this may have influenced the study sample. What criteria did clinic staff use to identify potential participants for the purposive sample? • Page 8, Qualitative Analysis: More detail is needed in regards to the analysis process-how were themes identified? How many individuals participated in the analysis process? What was the method for the peer debriefing? More detail is needed.

RESULTS: •
The organization of the results section makes it difficult to follow. While the authors repeatedly emphasize comparisons between veterans and non-veterans, the quotations provided in the text are often drawn from only one of these groups, which makes it difficult for the reader to evaluate the differences. The strength of qualitative methods is the ability to demonstrate the nuance and complexity of a phenomenon, but this is lost as these results are presented. The authors might consider presenting the results separately for veterans and non-veterans, or providing a table of themes, with a column of illustrative quotations for each group and then comparing them afterwards, rather than switching back and forth within the text. • The authors need to be careful also of the language used throughout the results section, which implies that the differences were quantified and that differences between non-veterans and veterans were measurable. • Finally, the value of the caregiver data for this paper is unclear, and seems to add more confusion, rather than help corroborate the experiences. The caregiver experiences might best be presented as a separate paper, as I am sure this data is rich on its own. • Since veterans were divided into those with and without mental health problems, were any differences in the themes noted in these groups? (This speaks further to the justification for why these were separated.)
Recommendation that we include a justification for using a qualitative approach and how the sample size was determined Thank you for this review point. We have now justified the use of a qualitative approach, to allow for the exploration of issues most salient to participants, and our sample size on pages 6 and 7.

2.
Request for additional information about the recruitment of veterans with and without mental health problems.
It has been well established in previous studies that individuals with mental health problems often have comorbid physical health problems. The inclusion of veterans with mental health problems was actively sought in order to investigate whether older veterans with mental health problems also had poorer physical health and if this was believed to be linked to their occupation. We have made this point clearer on page 7. We acknowledge as a limitation in the discussion that the inclusion of nonveterans with mental health problems would also have been useful in clarifying this issue (page 16).

Query regarding how thematic saturation was determined
We describe the topic of thematic saturation on page 7 and describe how this was achieved. For clarity, saturation was determined when researchers, in the reviewing of successive data, did not observe any additional themes emerging.

4.
Query regarding the context of the larger study Due to word limit restrictions, we are not able to describe the larger scale project in detail. However, we have included a reference to the full study report should readers be interested. We anticipate that over time we will produce a number of papers from the larger study but considered this one as being particularly of interest and thus we set out to publish this first.

5.
Query regarding the clinical context that participants were recruited from and the criteria used to identify potentially eligible participants by the clinical care team We describe on page 8 that participants were recruited from National Health Service (NHS) GP surgeries, walk in centres, community groups or mental health services. We detail that the clinical care team identified potentially eligible participants using study inclusion/exclusion criteria (page 8)

6.
Suggestion that additional information regarding the process of thematic analysis is included Thank you for this request. We have now listed the steps undertaken throughout the thematic analysis process, the primary researcher who proposed initial codes and themes, and how discussions about the findings with co-authors were incorporated as part of the peer debriefing process on pages 8 & 9.
7. Suggestion that we reorganise the presentation of participant quotes to improve clarity for readers.
We thank the reviewer for this feedback and have revised the results section with all participant quotes now listed in Table 2. We have ensured that additional non-veteran quotes are listed. We feel these changes have significantly improved the clarity of the results section for readers, with comparisons between veterans and non-veterans more clearly portrayed. We hope that the reviewers are satisfied with our response but are happy to attempt to clarify further if that is felt to be necessary.

8.
Recommendation that we revise our language throughout to remove references to quantified differences We have tempered our language throughout the results section to more clearly describe the observed thematic differences in experiences between veterans and non-veterans.

9.
Query whether the experiences of close companions detracts from the aim of the study to explore the impact of occupation on wellbeing later in life.
We have given this point a great deal of consideration. Having considered our findings, we believe that the inclusion of close companions' data serves to enrich and triangulate the views of veterans/non-veterans about the diverse impact of military and non-military occupations on physical health. Data from close companions also illustrates the widespread effects that physical health problems can have not only for the employee themselves, but on their family, and how these effects manifest in later years. We have amended the manuscript to more effectively portray the views of close companions and hope that our changes better delineate their experiences. Our department has done quite a bit of work on the impact of military service on the families of veterans, and those who are still serving, and consider this to be a topical and important subject.

10.
Query whether there were differences between veterans with and without mental health diagnoses.
Few thematic differences were found between veterans with and without mental health problems. To more clearly detail this, we have added a statement on page 12.

11.
Suggestion regarding the presentation of data in Table 1 We have amended an editorial error to reflect that 85% of close companions were female. All veteran participants were no longer serving in the Armed Forces and those who had civilian occupations worked in these professions post-service. For the sake of clarity, we have provided the total sample size of each subsample in the heading column of Table 1.

12.
Recommendation that minor editorial lapses are addressed We have amended editorial lapses throughout the manuscript and thank the reviewer for noting the presence of these.