br As seen in other studies investigating
As seen in other studies investigating exercise barriers among dif-ferent study populations, bad weather seems to be a universal barrier to outdoor PA (Baert et al., 2015; Cheville et al., 2012; Sheill et al., 2018). In our study, physical limitations caused by aging or comorbidities, and a wide range of external circumstances were also barriers to exercise. The findings of physical limitations and comorbidities as barriers to PA are in accordance with the results from a systematic review conducted by Bauman et al. (2012). Factors such as health status, self-eﬃcacy and previous PA were identified as consistent correlates with PA. Similarly, other studies investigating factors associated with regular exercise specifically in older adults found that perceived physical frailty and poor health are negatively associated with PA (Rhodes et al., 1999). When engaging older patients with cancer in exercise programs, it 4091-99-0 is important that HCPs take the individual patient's health status into consideration. Thus, exercise programs must be composed in a way that allows individual adaptations to ensure feasibility, comfort and safety.
In the current study, life events that occurred in older age were reported as barriers to PA, as these events often aﬀected daily life routines. This is in accordance with prior research that has shown that major life events may accelerate feelings of getting older, and thus change self-eﬃcacy towards PA (Kenter et al., 2015). Moreover, our results confirm that prior exercise habits and peoples’ sociocultural understandings of PA throughout life aﬀects later motivation and self-eﬃcacy for PA (Kenter et al., 2015). Mapping previous experiences with PA and previous life events can assist HCPs to identify individual sup-port needs prior to and during exercise engagement.
Social support from family and friends was identified as a motivator for PA. In addition, social interactions were highly valued and moti-vating in relation to structured exercise, and therefore group-based exercise was preferred. Many preferred to exercise with people who were at the same physical level as themselves, or with those who they had something in common with. These findings are consistent with the results from a qualitative synthesis performed by Franco et al. (2015) that explored older people's perspectives on PA participation. In this synthesis, informants in 65% of the included studies valued social contact during exercise and/or preferred group-based activities. In contrast to our results, Franco et al. (2015) also found that informants in 22% of the studies reported a lack of confidence or feelings of awkwardness in social exercise settings. This was particularly found in relation to PA that involved people from diﬀerent age groups, physical capabilities or cultural backgrounds (Franco et al., 2015). In our study, the informants expressed that exercise programs should be tailored to European Journal of Oncology Nursing 41 (2019) 16–23
individual needs and limitations. Hence, it is possible that feelings of inability or awkwardness could be reduced if individual adjustments were made.
In line with prior studies focusing on preferences for exercise among older people (Jancey et al., 2009; Janssen and Stube, 2014), the in-formants in our study preferred to engage in activities that were fa-miliar to them. This may be due to memories and values that have been attached to the activities through life, and thus make the activities meaningful. Known activities may also provide confidence and increase self-eﬃcacy. Even if age and cancer-related declines in heath may in-hibit older patients to engage in some former activities, patients’ pre-ferences and elements of their former activities could still be integrated in programs to increase short and long-term motivation.
4.1. Strengths and limitations
The sample size of the study was determined by ‘information power’ as described by Malterud et al. (2016), and was influenced by a focused study aim, the specificity of the informants and type of analysis (Malterud et al., 2016). MKM who had experience in conducting re-search interviews and had clinical nursing experience in treatment and care for patients with cancer, conducted all interviews. The 23 included informants who were selected with a purposeful sampling strategy that ensured equal representation of females and males, diﬀerent cancer diagnoses, treatments, and a variety of ages, provided rich and nuanced descriptions to cover the study aim.
In conducting a study that focuses on a relatively specific research area such as PA, there is a risk of selection bias (Norris, 1997), as pa-tients who have an interest in this topic (i.e. active and/or experienced exercisers) naturally would be more willing to participate. In case of selection bias, there is a possibility that the informants’ statements would not reflect the perceptions of patients whom we encounter in clinical practice. In this study, the risk of selection bias is estimated to be low as only two patients declined participation. In addition, data on current PA levels showed that we included both active and inactive informants.