Analyses were performed with data obtained from a previously reported randomized controlled trial—Nederlands Trial Register: NTR 3045—on the effects of a 3-month Internet-assisted intervention directed at increasing daily physical activity on objectively measured physical activity and metabolic health in 60-70-year-old inactive individuals. The CONSORT-EHEALTH Checklist for this trial is included as Multimedia Appendix 1 [8]. Details on study design and intervention content have been published elsewhere [6]. In short, the study recruited inactive participants aged 60-70 years from the region of Leiden, the Netherlands. The presence of an inactive lifestyle was assessed before randomization by a self-reported physical activity questionnaire: the General Practice Physical Activity Questionnaire (GPPAQ) [9]. This yielded four categories of physical activity: inactive, moderately inactive, moderately active, and active. We defined inactive as having less than 3 hours per week of exercise and cycling combined, corresponding to the inactive, moderately inactive, or moderately active category. Participants in the active category of the GPPAQ did not meet inclusion criteria for our definition of an inactive lifestyle. Participants were considered eligible if they (1) had no history of diabetes or did not use glucose-lowering medication, (2) had no disability impeding increase in physical activity, and (3) possessed and used a personal computer with Internet connection. At baseline, participants were randomly assigned to the intervention group or to a waiting list control group by the study physician or research nurse. Participants were randomized, via computerized program, into intervention versus waiting list control groups at a ratio of 1:1, with a block size of 12; stratification was performed by sex. Concealment of treatment allocation was ensured by randomizing at the end of the first study visit, after all baseline measurements and instructions at the study center were completed. Written informed consent was obtained from all participants. The study was approved by the Medical Ethical Committee of Leiden University Medical Center, the Netherlands.

Participants in the intervention group received a commercially available Internet-based physical activity program—DirectLife (Philips, Consumer Lifestyle, Amsterdam, the Netherlands)—directed at increasing daily physical activity. The DirectLife program comprised components that are based on the stages of change and the I-Change model. Briefly, this model assumes that behavioral change is the result of individual awareness of one's behavior, motivation to change behavior and action, and taking into account the individual’s current activity level; it subsequently provides a personal goal [6,10,11]. DirectLife consists of three elements: (1) an accelerometer-based activity monitor, (2) a personal website, and (3) a personal e-coach, who provides regular updates of the individual’s physical activity status by email and gives advice on how to increase physical activity. By means of these elements, the program aims to increase awareness about one’s own physical activity behavior, to give feedback on recent actual physical activity, and to provide support to make sustainable changes in physical activity behavior. The activity monitor of DirectLife is based on the Tracmor triaxial accelerometer, and has been validated against double-labeled water for the estimation of total daily life energy expenditure [12]. The DirectLife monitor is the consumer version of the Tracmor accelerometer. Participants of the program were instructed to wear the activity monitor continuously throughout the day to measure daily physical activity. Data were uploaded through a secure Internet connection to the database of the commercial provider. After an initial 8-day assessment period starting 1 week after the study visit, in which the current level of daily activity was measured, a target was set to increase the level of daily activity during a 12-week Internet-based interactive coaching program. Personalized targets were set by the DirectLife program and were defined as the absolute increase in physical activity compared to the individual’s baseline assessment data. For the whole group, this corresponded to a mean increase of approximately 10% in daily physical activity at week 12, increasing at a linear rate per week. All participants were given the option to decrease the personalized goal, within limits (ie, minimum of 5% increase in physical activity versus 10%), or to increase their personalized end goal, dependent on physical activity level of the last week.

Participants were given a target for daily activity, which increased weekly, and data from the accelerometer were used for regular feedback. Coaching included general recommendations on physical activities and coaches were available for further questions and advice by email correspondence. The control group was placed on a 3-month waiting list after which they received access to the intervention program at the end of the study. During the trial, no specific instructions regarding daily physical activity were given to the control group.

Enrollment and follow-up took place from November 2011 to August 2012.

Normally distributed data are shown as means with standard deviation, skewed data as medians with interquartile range (IQR). Between-group differences in quality of life after 3 months were analyzed by the intention-to-treat principle with an independent samples t test. For relative change in MVPA, the Mann-Whitney nonparametric test was performed. In a secondary analysis, we included only those participants in the intervention group who successfully reached the individual, personalized end goal to increase average physical activity that was set as part of the intervention program. To investigate if an increase in physical activity was associated with an improvement in quality of life, linear regression models were used, adjusted for age, sex, and body mass index (BMI) [15]. For this purpose, physical activity was divided into tertiles based on the change in minutes spent in MVPA. All analyses were performed with SPSS version 20.0 (IBM Corp, Armonk, NY, USA). Statistical significance was accepted at P<.05.

In this randomized controlled trial assessing the effect of an Internet-based intervention on improving quality of life in inactive older adults, we found a significant improvement in quality of life in the intervention group compared to the control group for subscales on emotional or mental health and health change of the RAND-36 health survey. Improvements in quality of life were particularly observed in those participants in the intervention group who had successfully reached their personalized target with the intervention. Furthermore, our results show that more MVPA is associated with higher quality-of-life levels.

Earlier, we found that the DirectLife intervention led to significant improvements of metabolic health and showed a significant association between an increase in minutes spent in MVPA and metabolic health outcomes [6,7]. Given the observed dose-response relationship we have shown between physical activity and the majority of subscales of quality of life, we are confident that the DirectLife program has induced an improvement in quality of life through an increase in physical activity. Our results are in accordance with results from previous experimental studies that showed that an increase in physical activity had a positive effect on quality of life in older adults [16,17]. These studies performed no subgroup analyses that included successful intervention adopters only. Evidence from a series of randomized controlled trials conducted with frail older adults to test the effects of physical activity on quality of life has indicated that there is a positive effect on emotional and social functioning in particular, and that physical activity did not exacerbate perceptions of pain [18]. Our results also confirm these findings, as the intervention particularly induced an improvement of mentally related subscales of the RAND-36 and did not lead to increased perceptions of pain. It is actually noteworthy that the effects of DirectLife induced the largest increases in subscales regarding the emotional component of quality of life. One of the explanations for this might be the type of intervention. DirectLife primarily focused on personal goal setting, aiming for an increase of perceived control, self-efficacy, and mastery, which probably will induce an improvement of mental functioning in particular [1,19,20]. Secondly, our sample contained inactive older people, who will probably more directly experience improvement of mental functioning, rather than physical functioning. As the empirical evidence regarding the mechanisms underlying the association between increased physical activity and quality of life is limited, these hypotheses remain speculations and further research on this topic is warranted.

Assuming that an increase of ≥5 points in any of the RAND-36 subscales is clinically relevant [21], the clinical relevance of the significant increases in the emotional or mental health (2.52 points) and health change (8.99 points) subscales in the intervention group is moderate to high. Baseline scores on the RAND-36 were higher than scores observed in a healthy Dutch population in a similar age category, particularly for physical functioning and limitations on usual role-related activities due to physical health problems [13]. Apparently, inactive 60-70-year-olds do not feel hampered in daily functioning because of physical limitations. For that reason, it is less likely that scores on these subscales could have increased to a higher level. This so-called ceiling effect could have contributed to the lack of intervention effect on RAND-36 subscales regarding physical functioning.

A limitation of this study arises from the representativeness of our study sample. Participation was voluntary, which unintentionally could have led to an overrepresentation of participants who are highly motivated to increase physical activity. Also, selecting participants who were able to use the Internet led to a sample with a relatively high education level. As a consequence, generalizability of the results toward the general elderly population is limited. On the other hand, our study sample contained overweight, inactive older adults with comorbidities—with exclusion of diabetes—which is representative of the general population and consequently leads to increased generalizability. Adults in our study sample are categorized as being inactive, moderately inactive, and moderately active based on the GPPAQ categories and corresponds with less than 3 hours of physical exercise and/or cycling per week. According to the official GPPAQ guidance document, the active category is taken as consistent with achieving goals set by the physical activity (PA) guidelines relating to time spent in MVPA or vigorous physical activity (VPA) [9]. It is important to mention that all other categories require a PA intervention, which was the rationale behind our choice for this GPPAQ-guided cutoff point for inclusion in the DirectLife study. The relative short duration of the DirectLife program (ie, 3 months) should be considered while interpreting the results of this study. Although the effects found on quality of life seem promising, these need to be reinforced by results from a longer-term study. In addition, we have not specifically emphasized the unraveling of determinants of quality of life in this study, such as socioeconomic status, family support, computer literacy, type of environment, and housing status. An in-depth and adequately designed determinant study would be useful to unravel the influence of these determinants on physical activity, biomedical outcomes, and quality of life.

An important strength is that we used validated instruments to measure physical activity and quality of life. However, although accelerometry is one of the state-of-the-art objective measurements of physical activity, two comments are noteworthy. First, our choices for wearing the accelerometers on the right side, as well for considering an assessment day as valid if >10 hours were registered, were arbitrary. A study by Masse and colleagues critically used the criteria of different algorithms to reduce accelerometer data on physical activity and showed that the algorithm we used—minimal amount needed: 5 days; minimal daily wear: 10 hours—will not lead to different MVPA scores in comparison to more or less stringent algorithms [22]. Second, the conversion of accelerometer data into an accurate and reliable PA measurement is an evolving topic in PA research. Whether accelerometer data should be used in the evaluation of PA guideline adherence remains under doubt [23]. In our study, still-limited information is available on the type of physical activity that was performed. Knowing whether reduced sitting time or increased cycling or walking time are responsible for improvements of metabolic outcomes and quality of life could contribute to more individually tailored advice on how to improve health and quality of life in the elderly. For example, ankle-worn accelerometers might yield data that is more predictive for other types of physical activity, such as cycling. Activity counts assessed with use of ankle-worn accelerometers are available from a subset of our study population. This topic has our attention and is one of the planned further investigations with data from the study.

Our findings show the feasibility and potential of Internet-assisted physical activity interventions in an older adult population. Based on our findings, we encourage further implementation of the DirectLife program. The DirectLife program 2.0 could function as an open-access, Web-based lifestyle advice tool that can be referred to by general practitioners and medical specialists, or can be consulted by inactive older adults. Further development of the program, however, requires reinforcement of the found effects at a longer term and optimization of the personally tailored advice by including more determinants of physical (in)activity, such as sitting time.

Our study shows that an Internet-based physical activity program was effective in improving health-related quality of life among inactive 60-70-year-olds that successfully reached their individually targeted increase in daily physical activity after 3 months, as indicated by the intervention program. Our results indicate that significant improvements in physical activity can lead to an improvement of mental functioning in particular.