Correlates and participation in community-based exercise programming for cancer patients before and during COVID-19

Abstract

Purpose

COVID-19 pandemic restrictions ceased the opportunity for face-to-face group exercise classes with at risk populations, such as cancer patients, forcing an adaptation to online exercise programming. The purpose of this study was to compare the attendance rates and correlates of face-to-face exercise programming pre-COVID-19 to online programming delivered during the first year of pandemic restrictions.

Method

The sample was comprised from 1189 patient records between 2018 and 2021. Data analysis was based around the three primary research questions: (i) whether the volume of attendance in online exercise programming differed from the previous face-to-face programming; (ii) whether there were any differences in attendee demographics between online and face-to-face classes; and (iii) whether there were specific correlates of online attendance that can inform future exercise programming.

Results

Class attendance increased significantly when online exercise classes were introduced during the first year of the pandemic when compared to face-to-face attendance the prior years (p < .01). Multiple demographic findings were also observed including age, gender, and geographic differences.

Conclusion

While COVID-19 has effected the ability to deliver face-to-face exercise programs for cancer patients, online programming has proved a promising delivery model with greater geographical reach. The approach, however, has gender and age differences in program attendance so targeted programming to reach specific cancer patient demographics may need attention. These results add to the continuing research in the area of online exercise and online programming strategies offering an effective option for cancer patients to achieve targeted exercise prescription.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00520-023-07725-3.

Introduction

Consistent moderate to vigorous intensity physical activity (MVPA) is now proven for its role in the prevention and treatment of many cancers. Specifically, MVPA has been shown to reduce the relative risk of certain cancer development by anywhere between 10 and 50% depending on the cancer diagnoses [1]. In terms of cancer survivorship, regular MVPA improves cancer patients’ health related quality of life (HRQoL) during treatment and improves functional mobility [1, 2]. Based on this evidence, the recommendation that persons with cancer should rest and avoid exercise is counterproductive; regular MVPA is an important prescription in most cases of supportive cancer care [3, 4].

Due to the possible debilitating nature of cancer and the multitude of additional treatment side effects, specific considerations need to be addressed when prescribing exercise for cancer patients. Physical and psychological limitations related to cancer diagnosis as well as treatment related side effects can greatly effect one’s ability to participate in regular and sustained physical activity [5–7]. It is important for fitness professionals and program coordinators to understand these challenges to be able to create safe and effective exercise programs [8]. Exercise recommendations for persons with cancer closely follows the 2020 Canadian 24-h movement guidelines [9, 10] of achieving at least 150 min/week of MVPA and several hours of light intensity physical activity. Regular exercise is an effective approach for achieving these MVPA guidelines and certain exercise recommendations have even been identified to help mitigate specific treatment related side effects [9]. Further, group exercise settings have been a recommendation for persons with cancer for its ability to provide beneficial outcomes in HRQoL and overall fitness [11].

Despite the well-recognized benefits of regular physical activity (PA) on chronic disease and general health, most Canadians do not meet the recommended weekly amount of PA. This issue is further pronounced in persons with cancer [12, 13]. Adherence to physical activity guidelines range between 17 and 47% in cancer patient samples [14]. Prescribed exercise programs among persons with cancer have also proven challenging, specifically due to treatment related barriers [15]. Exercise programming shows an attendance rate of approximately 71% in supervised exercise sessions and adherence rates of 87% in home-based exercise programs for breast cancer patients [5, 16, 17]. Thus, persons with cancer are willing to participate in cancer-related exercise programs; however, many care institutions fail to provide these services [18].

Due to the growing body of evidence supporting the efficacy of exercise for cancer patients during and after treatment, there has been a clear need for the increase in development of community supportive care programs such as the “LIVESTRONG at the YMCA” exercise classes [19]. However, face-to-face recreational programming was drastically affected by the COVID-19 pandemic [20], which ushered in a rapid switch from traditional exercise programming to online program delivery. This dramatically affected the public’s ability to achieve adequate PA levels due to stay at home orders and the inability to run traditional exercise programs [21]. The barriers to safe and effective program implementation are apparent, however online approaches to exercise program delivery may present unique advantages, particularly in reach and ease of access. This allows patients in more remote locations the ability to participate in high quality, specialized programing [22].

There is little research informing online exercise programming for persons with cancer. The purpose of this study was to compare the attendance rates and correlates of face-to-face exercise programming pre-COVID-19 to online programming delivered during the first year of pandemic restrictions from a community-based supportive cancer organization. Specific research questions included (i) whether the volume of attendance in online exercise programming differed from the previous season-matched attendance of face-to-face programming; (ii) whether there were any differences in attendee demographics (regional and cancer specific) between online and face-to-face classes; and (iii) whether there were specific correlates (demographic, regional, and cancer specific) of online attendance that can inform future exercise programming.

Methods

Design

This study used retrospective patient data from InspireHealth during April 2018 to March 2021. “Face-to-face” exercise programming was based on data from April 2018 to March 2020. “Online” exercise participation was based on data during April 2020–March 2021 in order to provide a clear delineation of data pre- and during pandemic. Other possible contributing factors were controlled between exercise models to strengthen possible findings. These factors included cost, advertisement, class style, and duration which all remained unchanged between face-to-face classes and online.

The organization involved in this research was InspireHealth, a not-for profit cancer care organization funded by the provincial Ministry of Health as well as grants and community donations. Prior to the COVID-19 pandemic, InspireHealth offered traditional face-to-face exercise programs at three centers in large cities in British Columbia. Due to COVID-19 pandemic restrictions in March 2020, InspireHealth immediately adopted an online approach to these exercise programs. The exercise sessions consisted of 60-min classes including resistance training, balance, core strengthening and cardiovascular focused exercises often in a circuit format.

Participants and procedures

Prior to data collection, a data agreement was drafted and signed by all parties involved with the data share. The research was approved by the University of Victoria Ethics office (20–0590) on February 18, 2021. An outline data sheet with requested variables and compatible formatting was provided for data population. Consultations with InspireHealth liaisons were held to create a suitable variable list for InspireHealths’s electronic medical records (EMR).

Data were extracted and compiled by InspireHealth staff from participants’ EMR on April 9th, 2021. Extracted variables included patient file number, date of birth, gender, postal code, cancer diagnosis, city of residence, registration date, class type code, class description, and class attendance duration. Data were extracted for participants who had their initial exercise consultation at InspireHealth between 2018 and 2020 which consisted of 31,336 individual exercise sessions, from 1189 eligible participants ranging across British Columbia.

Measures

Primary outcome

Number of total classes attended was the primary outcome. Face-to-face exercise attendance was gathered manually with patients checking-in upon arrival. During online programming, patients would register in advance and on the date of the exercise class, a staff member would check attendance.

Predictor/correlate variables

Predictor/correlate variables included medical, sociodemographic, and regional data. All variables were collected from participants by InspireHealth staff on the initial consultation date or registration within their program. This was performed with registration forms completed by patients online or by the InspireHealth staff on behalf of the patient.

Cancer diagnosis codes were categorized by the top 4 most common cancer diagnoses in Canada as reported by the Canadian Cancer Society in 2021. Skin and blood related cancers were also included due their large prevalence in our sample. All other forms were categorized as “other”. Cancer diagnoses were collected and categorized at InspireHealth using the International Classification of Diseases, 9th Revision (ICD9).

General demographics were pulled from InspireHealth’s EMR. These included sociodemographic variables of gender, age, registration date, date of initial participation, and regional variables via city of residence. Regional variables were categorized using the BC Ministry of Health, Health Boundaries.

Three demographic regions were assigned based on the three locations of InspireHealth centres: Lower Mainland, Vancouver Island, and Interior. The Lower Mainland variable was comprised of the Vancouver Central Health Authority and the Fraser Health Authority, the Vancouver Island variable was comprised of the Vancouver Island Health Authority, and the Interior variable was comprised of the Interior Health Authority. Regional variables were further sub-categorized into “within district” and “outside of district” based on their distance away from the nearest InspireHealth center. Participant’s city of residence that was a 60-min commute or longer via personal transportation was considered “outside of district”. This was determined using Google Maps to measure the travel time from the city in question and the nearest InspireHealth center at approximately 4:30 pm on April 28th, 2021.

Statistical analysis

Once data had been compiled, it was transposed into IBM SPSS (v 27) statistical software for analysis. To examine whether the volume of attendance to InspireHealth’s exercise programs changed when comparing online programming to the previous seasons of traditional face-to-face programming, aggregates were compared using paired sample t tests. Three 12-month periods were created; the 2018/2019 and 2019/2020 periods were both face-to-face programming years and the 2020/2021 period was when online programming was introduced. These analyses were compared by a full 12-month period matched by month as well as quarterly periods. The face-to-face year groupings were tested against the online year and were also tested against each other.

To evaluate whether attendee demographics, cancer diagnosis, and regional profiles differed between online and face-to-face classes, participant attendance data was coded into five variables: “pre-COVID participants 2018/2019 only”, “pre- COVID participants 2019/2020 only”, “During COVID participants only”, “participation in both years” and “2018–2020 pre- COVID participants”. The “pre-COVID participants only” variables included participants who had participated in at least one of InspireHealth’s exercise sessions during their respective year periods. The “During-COVID participants only” variable included participants who had participated in at least one of InspireHealth’s exercise sessions between April 2020 and March 2021. The “participation in both years” grouping consisted of participants who had attended as least one exercise session in both the April 2018–March 2020 period and the April 2020–March 2021 period. Finally, the “2018–2020 pre-COVID participants” group consisted of participants who had attended at least one class between April 2018–March 2019 and April 2019–March 2020.

Attendance data was also split into different groupings based on individual participant attendance volume. These groupings were created to identify participants who were achieving the average recommended attendance for InspireHealth’s classes of 2 or more classes per week. This grouping was created by averaging the number of classes attended in an individual’s highest 6 months of participation and highest 10 months of participation. This 10-month (rather than 12 months) average was used as a pragmatic assessment of regular attendance to allow for vacation or any normal interruptions among participants.

A grouping of individuals who had an average participation of > 1 class per week and < 2 classes per week was created for both 6- and 10-month periods. This resulted in 4 groupings based on attendance volume: (i) 6-month average of > 1 class per week; (ii) 6-month average of > 2 classes per week; (iii) 10-month average of > 1 class per week; (iv) 10-month average of > 2 per week. The analysis for our second research question was performed on the total data set as well as these groupings.

One-way ANOVAs were used to assess differences in age between year groupings and the three attendance pattern groupings using Tukey’s HSD post hocs. Chi-square tests were then performed to explore the relationship between non-continuous variables (gender, demographic, cancer diagnosis) and year groupings.

For the analysis of our third research question, only participants who were grouped in the “During COVID only” variable were analyzed. Pearson correlations were used to assess the relationship between age, gender, and average online attendance. One-way ANOVA tests were run to evaluate the relationship between average online attendance, regional variables, and cancer diagnosis using Tukey’s HSD post hocs. Regression models were also created to predict average online attendance while controlling for adjacent variables.

Results

Exercise program attendance

There was a significant increase in program attendance over the 12 month period for the during-COVID exercise programming (2020/2021: M = 1360.58; SD = 281.76) when compared to pre-COVID programming in the prior 2 years (2018/2019: M = 398.92; SD = 75.61 and 2019/2020: M = 377.50; SD = 75.48). This significant finding was also present in the multiple quarterly analyses (see Table 1). To help control for possible unrelated program growth, subsequent pre-COVID years were also tested against each other, and showed no significant difference.

Table 1.

Difference between “pre” and “during” aggregate group program attendance

Grouping 1	Pre-COVID 2018/2019 mean (SD)	Pre-COVID 2019/2020 mean (SD)	t	Effect size D (point estimate)	95% CI lower (effect size d)	95% CI upper (effect size d)
12-months	398.92 (75.61)	377.50 (75.48)	1.49	.421	− .171	1.016
Quarter 1	370.33 (7.10)	334. 67 (34.02)	1.53	.886	− .562	2.215
Quarter 2	469.33 (55.75)	482.67 (56.89)	− .514	− .297	− 1.428	.898
Quarter 3	399 (61.21)	344.33 (47.44)	3.80	2.20	− .083	4.458
Quarter 4	357 (115.05)	348.33 (49.24	.207	.119	− 1.031	1.243
Grouping 2	Pre-COVID 2018/2019 mean (SD)	During-COVID 2020/2021 mean (SD)	t1	Effect size D (point estimate)	95% CI lower (effect size d)	95% CI upper (effect size d)
12-months**	398.92 (75.61)	1360.58 (281.76)	− 9.69	− 2.798	− 4.070	1.502
Quarter 1*	370.33 (7.10)	1698.67 (92.43)	− 26.96	− 15.567	− 29.934	− 2.412
Quarter 2	469.33 (55.75)	1054.67 (199.57)	− 4.03	− 2.327	− 4.699	.044
Quarter 3*	399 (61.21)	1196.33 (162.216)	7.88	− 4.548	− 8.855	− .488
Quarter 4*	357 (115.05)	1492.67 (108.10)	− 8.92	− 5.152	− 10.002	− .612
Grouping 3	Pre-COVID 2019/2020 mean (SD)	During-COVID 2020/2021 mean (SD)	t	Effect size D (point estimate)	95% CI lower (effect size d)	95% CI upper (effect size d)
12-months**	377.50 (75.48)	1360.58 (281.76)	− 9.61	− 2.860	− 4.155	− 1.542
Quarter 1*	334. 67 (34.02)	1698.67 (92.43)	− 19.2	− 11.085	− 21.340	− 1.671
Quarter 2 *	482.67 (56.89)	1054.67 (199.57)	− 4.41	− 2.547	− 5.102	− .017
Quarter 3*	344.33 (47.44)	1196.33 (162.22)	− 9.55	− 5.514	− 10.691	− .683
Quarter 4*	348.33 (49.24)	1492.67 (108.10)	− 12.61	− 7.281	− 14.059	− 1.014

^*p < .05

^**p < .001

Demographic and cancer diagnosis comparisons between attendees of online and face-to-face programming

There was a significant difference in the average age of the participant when reviewing the data across the different exercise programs (pre-COVID, during-COVID, attendees of both; see Table 2: F4,1184 = 4.38, p = 0.002; η2 = 0.015). However, upon further analysis it appears that only participants who participated in > 1 class/week, when averaged over 6 months, showed a significant difference (see Table 2; F4,142 = 3.74, p = 0.006; η2 = 0.095). Tukey’s HSD post-hoc analysis showed that although there was a younger age demographic for the online classes compared to pre-COVID 2018–2019 (p = 0.008), there was no difference in age when compared to pre-COVID 2019–2020.

Table 2.

Age differences by year and attendance grouping

Variable	(1) 2018/2019 Face-to-face	(2) 2019/2020 Face-to-face	(3) 2020/2021 Online	(4) Both Face-to-face and online	(5) 2018–2020 Face-to-face	F	η2	Post hoc
All cases	(n = 292)	(n = 212)	(n = 409)	(n = 174)	(n = 102)
Age Mean (SD)	61.98 (12.10)	59.38 (12.40)	59.75 (12.36)	61.62 (10.86)	64.27 (13.75)	4.38	.015	5 > 3; 5 > 2
10-month average: > 1 class attended/week	(n = 21)	(n = 10)	(n = 43)	(n = 21)	(n = 34)
Age Mean (SD)	64.10 (12.16)	61.20 (9.38)	59.47 (12.75)	57.90 (11.89)	60.71 (16.01)	.659	.021	N/A
10-month average: > 2 class attended/week	N/A	N/A	(n = 99)	(n = 111)	(n = 42)
Age Mean (SD)			64.29 (9.40)	64.05 (9.93)	66.48 (10.94)	.508	.008	N/A
6-month average: > 1 class attended/week	(n = 38)	(n = 17)	(n = 50)	(n = 18)	(n = 24)
Age Mean (SD)	63.50 (10.65)	62.29 (10.66)	58.02 (12.22)	53.06 (11.43)	64.58 (13.78)	3.74	.095	1 > 4: 5 > 4
6-month average: > 2 class attended/week	N/A	N/A	(n = 126)	(n = 129)	(n = 65)
Age Mean (SD)			63.45 (10.42)	63.18 (10.48)	64.28 (13.45)	.302	.004	N/A

Different year groupings also differed by gender. This was seen in the overall analysis (χ2 = 32.06, p < 0.001), and between the pre-COVID and during-COVID groupings (see Table 3).

Table 3.

Participant demographics based on year and attendance grouping

Variable	(1) 2018/2019 Face-to-face	(2) 2019/2020 Face-to-face	(3) 2020/2021 Online	(4) Both Face-to-face and online	(5) 2018–2020 Face-to-face	χ2	p
All participants	(n = 292)	(n = 212)	(n = 409)	(n = 174)	(n = 102)
Sex % Female All cases 2018/19 v 2019/20 2018/19 v 2020/21 2018/19 v both 2018/19 v 2018–2020 2019/20 v 2020/21 2019/20 v both 2019/20 v 2018–2020 2020/21 v both 2020/21 v 2018–2020 2018–2020 v both	84.2	86.8	93.9	94.8	80.4	32.06 .636 17.40 11.73 .805 9.00 7.12 2.18 .196 18.52 14.25	< .001 .425  < .001  < .001 .370 .003 .008 .140 .658  < .001  < .001
Cancer diagnosis % Lung % Breast % Colorectal % Prostate % Skin % Blood % Other All cases Breast v aggregate*	3.8 49.3 5.1 4.1 2.1 9.9 22.6	3.8 56.6 4.2 2.4 1.4 8.0 22.6	2.4 57.9 4.4 2.4 2.0 10.0 20.5	4.6 56.9 3.4 2.9 1.7 12.1 18.4	5.9 43.1 8.8 4.9 2.9 11.8 21.6	23.10 5.15	.284 .273
Demographic % lower mainland-distant proximity % lower mainland–close proximity % Vancouver Island-distant proximity % Vancouver Island–close proximity % interior–distant proximity % Interior–close proximity All cases Lower mainland Vancouver Island Interior	0.7 49.3 5.1 35.3 1.4 8.2	0.0 40.1 5.7 46.7 1.4 6.1	3.7 47.9 9.5 24.2 5.6 9.0	1.1 57.5 1.1 35.1 0.6 4.6	0.0 61.8 0.0 33.3 0.0 4.9	99.46 17.75 34.57 9.90	< .001 .001  < .001 .042
10-month average: > 1 class attended/week	(n = 21)	(n = 10)	(n = 43)	(n = 21)	(n = 34)
Sex % Female All cases 2018/19 v 2019/20 2018/19 v 2020/21 2018/19 v both 2018/19 v 2018–2020 2019/20 v 2020/21 2019/20 v both 2019/20 v 2018–2020 2020/21 v both 2020/21 v 2018–2020 2018–2020 v both	85.7	80.0	90.7	95.2	82.4	3.97 .164 .360 1.11 .107 .925 1.80 .029 .404 1.17 1.94	.563 .686 .549 .293 .743 .336 .180 .865 .525 .279 .164
Cancer diagnosis % Lung % Breast % Colorectal % Prostate % Skin % Blood % Other All cases Breast v aggregate*	0.0 57.1 4.8 4.8 0.0 4.8 28.6	0.0 60.0 0.0 0.0 0.0 10.0 30.0	9.3 37.2 4.7 2.3 2.3 16.3 27.9	9.5 57.1 0.0 0.0 0.0 9.5 23.8	5.9 44.1 8.8 11.8 2.9 5.9 20.6	18.35 4.17	.786 .383
Demographic % Lower mainland-distant proximity % lower mainland–close proximity % Vancouver Island-distant proximity % Vancouver Island–close proximity % Interior–distant proximity % Interior–close proximity All cases Lower mainland Vancouver Island Interior	0.0 47.6 0.0 33.3 4.8 14.3	0.0 60.0 0.0 40.0 0.0 0.0	0.0 58.1 7.0 14.0 9.3 11.6	0 66.7 4.8 28.6 0 0	0.0 73.5 0.0 26.5 0 0	23.76 N/a 7.07 .442	.095 N/a .132 .506
10-month average: > 2 class attended/week	N/A	N/A	(n = 99)	(n = 111)	(n = 42)
Sex % Female All cases 2018/19 v 2019/20 2018/19 v 2020/21 2018/19 v both 2018/19 v 2018–2020 2019/20 v 2020/21 2019/20 v both 2019/20 v 2018–2020 2020/21 v both 2020/21 v 2018–2020 2018–2020 v both			93.9	93.7	78.6	10.71 .467 .717 .663 .188 .193 .202 .804 .005 7.33 7.44	.030 .490 .397 .415 .664 .660 .653 .370 .941 .007 .006
Cancer diagnosis % Lung % Breast % Colorectal % Prostate % Skin % Blood % Other All cases Breast v aggregate*			2.0 57.6 4.0 3.0 2.0 13.1 18.2	4.5 52.3 5.4 3.6 1.8 11.7 20.7	7.1 45.2 7.1 2.4 2.4 14.3 19.0	22.77 3.28	.533 .513
Demographic % Lower mainland-distant proximity % Lower mainland–close proximity % Vancouver Island-distant proximity % Vancouver Island–close proximity % Interior–distant proximity % Interior–close proximity All cases Lower mainland Vancouver Island Interior			4.0 49.5 7.1 22.2 4.0 13.1	0.9 59.5 0.9 34.2 0.0 4.5	0.0 61.9 0.0 31.0 0.0 7.1	29.22 4.79 11.29 2.24	.084 .310 .024 .326
6-month average: > 1 class attended/week	(n = 38)	(n = 17)	(n = 50)	(n = 18)	(n = 24)
Sex % Female All cases 2018/19 v 2019/20 2018/19 v 2020/21 2018/19 v both 2018/19 v 2018–2020 2019/20 v 2020/21 2019/20 v both 2019/20 v 2018–2020 2020/21 v both 2020/21 v 2018–2020 2018–2020 v both	81.6	88.2	92.0	100	79.2	6.35 .380 2.14 3.79 .055 .221 2.25 .578 1.53 2.50 4.26	.174 .537 .143 .052 .815 .639 .134 .447 .216 .114 .039
Cancer diagnosis % Lung % Breast % Colorectal % Prostate % Skin % Blood % Other All cases Breast v aggregate*	0.0 55.3 2.6 2.6 2.6 10.5 26.3	0.0 47.1 0.0 5.9 0.0 5.9 41.2	2.0 42.0 12.0 2.0 2.0 12.0 28.0	0.0 61.1 0.0 0.0 5.6 16.7 16.7	8.3 45,8 0.0 4.2 4.2 12.5 25.0	21.86 2.76	.588 .600
Demographic % Lower mainland-distant proximity % Lower mainland–close proximity % Vancouver Island-distant proximity % Vancouver Island–close proximity % Interior–distant proximity % Interior–close proximity All cases Lower Mainland Vancouver Island Interior	0.0 44.7 7.9 28.9 2.6 15.8	0.0 47.1 0.0 41.2 0.0 11.8	2.0 62.0 4.0 20.0 4.0 8.0	0.0 55.6 0.0 33.3 5.6 5.6	0.0 54.2 0.0 41.7 0.0 4.2	15.92 1.52 5.09 2.43	.722 .823 .279 .657
6-month average: > 2 class attended/week	N/A	N/A	(n = 126)	(n = 129)	(n = 65)
Sex % Female All cases 2018/19 v 2019/20 2018/19 v 2020/21 2018/19 v both 2018/19 v 2018–2020 2019/20 v 2020/21 2019/20 v both 2019/20 v 2018–2020 2020/21 v both 2020/21 v 2018–2020 2018–2020 v both			92.1	93.8	80	10.39 .034 .829 1.52 .149 .208 .487 .259 .292 5.89 8.53	.034 .854 .363 .217 .700 .648 .485 .611 .589 .015 .004
Cancer diagnosis % Lung % Breast % Colorectal % Prostate % Skin % Blood % Other All cases Breast v aggregate*			4.8 52.4 4.0 3.2 2.4 13.5 19.8	5.4 52.7 4.7 3.1 1.6 11.6 20.9	4.6 44.6 9.2 6.2 3.1 10.8 20.0	13.67 1.88	.954 .759
Demographic % Lower mainland-distant proximity % Lower mainland–close proximity % Vancouver Island-distant proximity % Vancouver Island–close proximity % Interior–distant proximity % Interior–close proximity All cases Lower mainland Vancouver Island Interior			3.2 49.2 7.9 20.6 6.5 13.5	0.8 60.5 1.6 33.3 0.0 3.9	0.0 66.2 0.0 29.2 0.0 4.6	45.72 5.44 15.27 2.99	< .001 .245 .004 .225

^*p < .05

^**p < .001

In addition to age and gender, there were also regional differences between those who attended pre-COVID compared to during-COVID. These results were seen when analysing all cases (χ2 = 99.46, p < 0.001) with increased “OUT” group participation in all regions, as well as when analysing the Vancouver Island region separately (see Table 3). Finally, regarding exercise program delivery groups when compared by cancer diagnosis information, there were no significant differences by cancer type in our sample (p > 0.27).

Correlates of attendance in the online exercise programming

There was a significant correlation between age (r = 0.191, p < 0.001) and attendance of the online exercise sessions, but no relationship between attendance and gender, cancer diagnosis variables, or regional distribution. In a hierarchal regression analysis, age predicted average exercise attendance (β = 0.036, p < 0.001) (Table 4).

Table 4.

Hierarchical regression analysis of average attendance

Variable	Fchange	df	R2 change	β
Average attendance (Block #1)	3.84	4	.037
Age				.036**
Sex				− .008
Demographic Region				.001
Cancer Dx				.034

Note: df = degrees of freedom, β1 = standardized regression coefficient

Discussion

Due to the COVID-19 pandemic, organized exercise programs were disrupted owing primarily to their face-to-face delivery. Online exercise program delivery is an emerging area to enable persons with cancer to maintain a healthy, active lifestyle [22–24]; however, their effect on participation in these programs has received little research. Thus, our primary research question focused on the overall attendance changes between programming types. Results showed a marked increase in program attendance which could be explained by the convenience of this emerging exercise model and its wider reach. The increase in home fitness equipment and advancements in technology for online exercise delivery have made online-based home exercise a viable model for achieving adequate physical activity levels [23]. Implementation for elderly populations have seen positive impacts in multiple physical performance measures as well as perceived convenience of participation [25]. Preliminary research of online-based exercise programming has also been noted in cancer populations. Sotirova et al. [26] focused on online exercise program delivery for cancer patients and found that 90% of applicable studies showed generally high levels of adherence, at least in the short-term.

Our second research question sought to understand more about the sociodemographics of participants who attended the online sessions compared to the pre-COVID years. Our current findings are in line with Sotirova et al. [26] who also found that younger ages (age < 57 years) were more likely to adhere to online exercise programs. Over 90% of young people (over 15 years of age) report having access to a computer, but this prevalence decreases with age [27]. This would suggest that a large portion of the population, primarily composed of older adults and people of lower socioeconomic status, do not have adequate access to the internet or overall adequate digital literacy [28]. The number of different interface models on digital devices and applications makes it very difficult for older adults to accustom themselves [29, 30]. Addressing digital literacy disparities in older adults is a crucial step to implementing successful online exercise programs. The use of mixed program delivery, when possible, could also help bridge digital age divide by allowing the choice between face-to-face and online models to allow individuals to choose their preferred model [31]. Still, it should be noted that although we found a significant association between younger age and greater online exercise class attendance compared to face-to-face, the magnitude of this association was small (approximately 2 years of age difference). Whether this difference is substantial enough to resemble a difference in digital literacy between the groupings requires further exploration.

When investigating participant gender and program type, our results indicated that women made up the majority in all our groupings. Results further showed that men participated in the face-to-face classes more than the online classes. When the online exercise model was introduced, there was a significant drop-off in men. The large overall presence of women in all groups could be due to the make up of participant diagnoses, as breast cancer made up approximately 50% of each exercise group. Breast cancer is the most common type of cancer in British Columbia and affects women at a much higher rate than men as reported by HealthLinkBC in 2020 [32]. Women are also the primary participant in group exercise classes [1, 34–36], where men prefer exercise outside of organized group classes [33]. Further, program adherence is highly affected by demographic similarities [37, 38]. Due to the high proportion of InspireHealth’s classes being comprised of women participants, this factor could be attracting more women to engage in the program while contributing to the decrease in men. Gender differences could also be explained by the relationship established between participants and instructors, as individuals who are more aligned with the social identity of their instructor can be more easily motivated by them [39].

Our analysis of geographic region showed that online programming allowed persons with cancer living further away from InspireHealth centres to participate in the exercise classes. These findings support the efficacy of the online exercise model, because one of the most prominent predictors of adherence and recruitment to exercise interventions among persons with cancer is the travel distance to rehabilitation centres [6, 40].

Our third research question addressed the relationships between participant demographics and average online program attendance. Results showed that older participants had a higher average attendance in InspireHealth’s online exercise programs. Despite younger adults using the online class model more than the face-to-face model, older adults participating in the online classes actually attended more classes on average than the younger adults. Overall, the findings are surprising and promising as they demonstrate the ability to increase exercise participation in a population of the community that has been observed to struggle with achieving adequate PA levels. These findings could be due to the high level of age similarity in InspireHealth’s programs, in that they were skewed towards older adults [37]. InspireHealth’s exercise programs were attended by a demographic of participants with a mean age of 63 years old. As previously explained in relation to gender, the large proportion of older adults already participating may have attracted more similar-aged participants while deterring program participation in a younger demographic. Higher participation in older age groups could also be explained by simple time affordance. Older adults often do not have the same daily commitments as younger age groups.

Finally, travel distance has been reported as a prominent barrier to maintaining adequate PA levels and exercise program adherence [6, 40]. Our study found there was no significant relationship between average online attendance and proximity to the centre. Along with our findings of overall increased class attendance with online programming as compared to face-to-face, these regional results suggest that online exercise programs are able to remove the barrier of travel distance for many participants.

Despite the strengths of our research such as the large sample, objective participation data, and 3-year assessment, there are important limitations to our findings. First, due to the inability to measure severity of diagnosis or presence of secondary diseases in our data set, our null findings on cancer diagnosis are difficult to comment on. Future research into this topic could attempt to identify if cancer severity or secondary diseases associated with cancer diagnosis effect online exercise program effectiveness or attendance rates. Second, the need to condense our data into aggregates in order to assess volume changes did not allow for the direct analysis of individual program retention or adherence. Additionally, capacity differences between the online and face-to-face classes may contribute to some of the differences observed. The face-to-face Victoria and in Vancouver classes, for example, often comprised 6–18 participants with small wait-lists (approx. 2–4 people). While these class sizes v. wait lists may underrepresent the potential of the face-to-face programming, there was still far less indicators of participation interest compared to the capacity of 100 + attendees with the online model. Further, our sample consisted of a high percentage of women (~ 90%) and breast cancer patients (~ 54%). This unequal distribution of certain demographic variables could decrease our ability to generalize our findings to a wider population. Finally, while online programs clearly improve attendance, future research regarding the effectiveness of these classes for achieving to health and well-being outcomes is needed.

In summary, the online programming introduced by InspireHealth saw a significant increase in class attendance. Older participants had a higher average attendance; however, when comparing the face-to-face to online program delivery, a younger subset of our sample was seen to participate in the online programs. Women were found to be the dominant gender in both online and in-person program styles and men were found to participate significantly less when the switch to the online programs was introduced. The online programming attracted more participants from areas further from InspireHealth’s centre. All participants, regardless of distance from a centre, maintained a similar average attendance. These results add to the expanding area of research regarding the efficacy of online exercise for targeted, at-risk populations and continuing research in the area of online exercise and online programming strategies.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

Michael Smith and Ryan Rhodes wrote the main manuscript text. Michael Smith prepared all figures. All authors reviewed and approved the final manuscript.

Funding

The authors declare that no funds or other financial aid was received during the preparation of this manuscript.

Data availability

Group data can be provided upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. Research was approved by the University of Victoria ethics office (20–0590) on February 18, 2021. The authors affirm that human research participants provided informed consent for publication of data information. The data that supports the findings of this study are available from the corresponding author upon request.

Competing interests

The authors have no relevant financial or non-financial interests to disclose; however, Dr. Nette is an employee of InspireHealth.