Cancer patients participating in a lifestyle intervention during chemotherapy greatly over-report their physical activity level: a validation study

The present validation study is part of the I CAN-study [19]; a 12-month prospective feasibility intervention with the aim to increase population based adherence to healthy lifestyle behaviors including diet, PA, mental stress management and smoking cessation. The intervention was delivered to 100 cancer patients undergoing chemotherapy with curative or palliative intent through: 1) a grouped start-up course with patients and nearest relatives, 2) an information binder with recommendations, recipes and tips on how to manage possible disease and treatment-related symptoms themselves, and 3) monthly counseling with a lifestyle supervisor with recommendations individualized to the patients’ abilities, barriers and preferences. The study is described in detail elsewhere [19].

From the 100 I CAN participants, 16 participants dropped out before enrolment in the present validation study. The remaining 84 participants received a SWA and completed the IPAQ-sf. Of these, 18 were eliminated due to 1) SWA malfunction (n = 14) or 2) not sufficient wearing-time of the SWA (n = 4). In total 66 participants had valid registrations on the SWA and were included to the present validation study. There were no significant differences in self-reported MVPA between participants vs. non-participants at baseline (p = 0.414). Participants wore the SWA for an average of 3.6 days and 23.7 h^.day⁻¹ of all valid days. Demographic, medical and physical characteristics of the participants (n = 66) and non-participants (n = 34) are presented in Table 1.

In the present study, cancer patients participating in a comprehensive lifestyle intervention while undergoing chemotherapy over-reported time in MVPA compared to SWA, when completing the IPAQ-sf. Specifically, these cancer patients over-reported their moderate-to-vigorous PA by nearly 100 min^.day⁻¹. No differences in over-reporting were observed between patients undergoing chemotherapy with curative or palliative intent.

Almost 90 % of the participants in the present study perceived themselves as meeting the PA guidelines of 150 min^.wk⁻¹ of MPA while less than 50 % actually met the PA guidelines according to the objective measures. Our findings are supported by previous studies, where more than half of healthy adults included perceived themselves as accommodating the PA guidelines [24, 25]. Objective measures in the same population revealed that only 15 % were reaching the guidelines [26]. The higher percentage of participants self-reporting accommodating the PA recommendations observed in the present study compared to previous studies [26] may be a result from drawing our participants from a lifestyle intervention focusing on the participants’ diet, mental stress and smoking cessation in addition to their PA. In addition to measuring the participants’ lifestyle behaviors, participants also received recommendations on how to maintain or adhere to a healthy lifestyle during their cancer treatment. It is a well-known phenomenon that individuals participating in a study often improve aspects of their behavior as a response to being studied or they believe they have improved more than what they actually have (the Hawthorne effect) [27‐29]. In the present study, the IPAQ-sf identified significantly more participants as accommodating the PA guidelines of 150 min^.wk⁻¹ of MVPA vs. the SWA. Six of the participants reported from 420 to 840 min^.wk⁻¹ of MVPA, while less than 60 min^.wk⁻¹was registered on their SWA. These findings are not only statistical significant, but also of great clinical importance. Health care professionals should take these findings into consideration when delivering PA recommendations in this population and if using IPAQ-sf as an assessment tool in the clinic. In terms of this knowledge, individualized PA recommendations can be delivered with barriers such as fatigue, feeling sick, loss of interest and nausea, in mind [13, 30], so that cancer patients undergoing oncologic treatment can harvest the known health benefits from adhering to the developed PA guidelines [6].

Participants in the present validation study self-reported being in MVPA 662 min^.wk⁻¹ more than what was objectively registered by the SWA; an over-reporting that is supported in the literature [9, 31‐33]. Of the 844 min^.wk⁻¹ MVPA reported on the IPAQ-sf in the present study, participants reported 459 min^.wk⁻¹ as walking activities. As reported previously, IPAQ-sf may over-report MPA because it includes walking at any intensity [33]. In the present study, IPAQ-sf over-reported time spent on walking by 283 min^.wk⁻¹ or 1.6 times compared to MPA obtained by the SWA. When including MPA from the IPAQ-sf in the analysis, IPAQ-sf over-reported walking by 602 min^.wk⁻¹ or 3.4 times compared to MPA data on the SWA. To include time spent on walking in the MVPA but not differentiate the intensity of the walking may be a potential source of over-reporting in cancer patients undergoing chemotherapy, and time spent on walking was thus both included and excluded to MVPA in the present study. A significant amount of the walking performed by these patients may be objectively registered as light intensity [34], while being experienced and self-reported as moderate [35]. The IPAQ-sf defines MPA as activities that make you breathe somewhat harder than normal [17]. Importantly, cancer patients dealing with disease- and treatment-related side-effects such as reduced physical capacity, fatigue, pain, depression and anxiety [36‐39] may feel short of breath at a much lighter intensity than they have previously or compared to the population which the questionnaire is developed for. Consequently, the experienced side-effects are a great source of over-reporting since the patients might experience and report the PA as moderate, while the SWA assesses the PA as light intensity. This is important to have in mind when using self-reports in cancer patients; many PA self-reports, like the IPAQ-sf, are developed for use in a healthy population. Gil-Rey et al. [39] thus suggest cancer-specific PA guidelines to maximize the health benefits in this population.

The participants did not systematically over-report their PA levels when completing the IPAQ-sf; in other words, higher physical activity levels from the SWA was not associated with more over-reporting of MVPA reported from the IPAQ-sf. Our findings are in contrast to the findings of Johnson-Kozlow et al. [31], who revealed larger over-reporting in the breast cancer survivors who reported the highest PA levels on the long form of the IPAQ. Reasons for our conflicting findings may be the use of different accelerometer in the two studies and the use of the long versus the short form of the IPAQ in the study of Johnson-Kozlow and colleagues [31] and the present study, respectively. The long form of the IPAQ gives a total of 35 examples of PA across different activity domains, i.e. recreational sports, leisure time and housework PA. This provides many opportunities for “forward telescoping”; a recall bias occurring if the activities were recalled as taking place during the same seven-day period which is monitored, but actually took place previously [40]. Secondly, while the participants were undergoing chemotherapy in the present study and may not have been able to perform much vigorous PA, the participants in the study of Johnson-Kozlow et al. [31] were cancer survivors two years post diagnosis. Recall of vigorous PA is more likely to be subject to “forward telescoping” on PA self-reports, since these activities are often easier to remember due to strong, distinct physiological signals [40].

Importantly, IPAQ has been criticized for being complicated to complete [41] due to difficulties in remembering which activities they performed the past seven days, to distinguish the intensity of the different activities and last but not least trying to identify whether or not the activities lasted ≥ 10 min [42]. These aspects may be even harder to recognize for cancer patients experiencing disease- and treatment-related side effects such as pain, mental stress, cognitive difficulties and fatigue [13, 14, 37]. When comparing the IPAQ-sf to the SWA, activities lasting ≥ 10 min from the SWA was applied in the analyses, which may lead to great over-reporting if the activity lasted < 10 min. Thus, post hoc PA comparisons assessed by the IPAQ-sf and minute-by-minute SWA data were conducted in the present study (data not shown). Analyses revealed that IPAQ-sf still significantly over-reported VPA but IPAQ-sf now significantly under-reported MPA by 44 % compared to the SWA. The difficulties in completing the IPAQ-sf are reasonably clear. Importantly, there are currently no gold-standard for quantifying PA [43]; however, accelerometers are a precise and valid tool and thus commonly used in validating self-reports of PA [32]. Consequently, it is of concern that both the evidence regarding PA in cancer patients and the PA recommendations in this population is developed on the basis of self-reported PA data, which in turn has impact on the validity of those recommendations [6]. Another question that arises in this context is why self-reports only addresses PA in bouts >10 min, when there is rapidly growing evidence on health benefits in shorter bouts of PA [44, 45]. This is; however, an unexplored field in cancer patients, which needs further investigation, especially since activities of longer duration may be hard to complete for cancer patients with impaired physical capacity due to the oncologic treatment.

There are strengths and limitations of the present study. To our knowledge, this is the first time the IPAQ-sf has been compared to an objective PA monitor in cancer patients undergoing chemotherapy with either curative or palliative intent. One key aspect when designing the present study was to make the objective registrations feasible to the patients. Many different guidelines regarding days wearing the activity monitors are previously provided, with a minimum wearing time of four days recommended [46, 47]. One weakness of the present study is that only 66 % of the eligible patients were included, despite wearing the SWA for only five days, due to voluntary dropout, SWA malfunction or insufficient wearing time, leading to a higher inclusion of breast cancer patients who were either married or living together with someone compared to the total sample. These observations may indicate that a shorter wearing time is preferred in this population with regard to feasibility. Secondly, the participants were instructed to remove the SWA during water-based activities and activities such as swimming were thus not recorded. However, cancer patients at our clinic were advised to refrain from activities such as swimming in public pools due to reduced immune function and increased infection risk during chemotherapy. Further, the present study, as previous lifestyle interventions, is limited by including the healthier and fitter participants compared to the population from which they are drawn [19, 30]. Unfortunately, no Bland-Altman plot was calculated for VPA since only three of the participants had bouts of ≥10 min of VPA recorded on the SWA. Further, the present study is limited by the amount of questions asked. The present validation study was part of a comprehensive lifestyle intervention, which focused on the participants' diet, mental stress, smoking cessation and quality of life in addition to their PA level. The large amount of questions asked, perhaps in combination with side effects from the disease and its treatment, may have affected the accuracy of the participants’ answers.

Based on our findings, cancer patients participating in a lifestyle intervention while undergoing chemotherapy grossly over-report their PA level from the past seven days when using the IPAQ-sf. Thus, the IPAQ-sf seems to be insufficient when assessing PA level in cancer patients undergoing oncologic treatment. Activity monitors or other objective tools should be considered in this population as an attempt to bridge the gap between how physical active the cancer patients perceive themselves and how physical active they actually are, in order to provide each patient with correct and individualized PA recommendations.