Hip fractures are an increasingly frequent consequence of falls in older adults and are becoming a significant concern. It is estimated that 30%–60% of older adults who suffer a hip fracture experience permanent limitation to mobility or to their general level of independence (Dyer et al., 2016).
Previous studies suggest that older adults recovering from a hip fracture undertake few physical activities and exhibit sedentary behavior over prolonged daytime periods (Ekegren et al., 2018; Fleig et al., 2016; Resnick et al., 2011; Taraldsen et al., 2013; Zusman et al., 2018, 2019). High levels of sedentary behavior are associated with a reduction in muscle mass and strength (Gianoudis et al., 2015), increased risk of falls (Thibaud et al., 2012), and even mortality (Chau et al., 2013). Reducing sedentary behavior and encouraging regular physical activities can help preserve an acceptable level of mobility and independence among older adults, and is especially important among older adults recovering from a hip fracture, since activity also increases the likelihood of recovery (Talkowski et al., 2009).
The first step to overcoming this problem is a better understanding of physical activity and sedentary behavior. Physical activity refers to any body movement that raises energy expenditure above resting levels and is often categorized by intensity (Caspersen et al., 1985). Sedentary behavior is defined as behavior resulting in energy expenditure ≤1.5 metabolic equivalents (METs) while in a sitting, reclining, or lying posture (Tremblay et al., 2017). Although physical activity and sedentary behavior share some attributes, each should be considered a distinct domain. However, an increase in physical activity does not necessarily result in a reduction of sedentary behavior (Prince et al., 2014). For example, an older adult recovering from hip fracture might receive 30 min of therapy during the morning, but spend the rest of the day sitting. In recent times, the use of wearable activity sensors such as accelerometers have made it possible to obtain a reliable, objective representation of physical activity and sedentary behavior (Hart et al., 2011; Klenk et al., 2019). Nonetheless, it is important to properly assess and interpret accelerometer measurements, an issue that is particularly challenging in the case of sedentary behavior as there are numerous ways, ranging from simple to complex, to assess this behavior (Boerema et al., 2020). Recent literature supports a focus on the pattern of accumulation of sedentary behavior, the main benefit of which is sensitive quantification of (changes in) sedentary behavior (Boerema et al., 2020; Chastin et al., 2015).
A second important step to improve recovery after hip fracture is the identification of subgroups of older adults defined by levels of physical activity and sedentary behavior. This is important because it enables us to provide tailored interventions that are likely to be more effective in improving the chance of recovery (White et al., 2020).
Therefore, the aims of this study were (a) to quantify physical activity and sedentary behavior using an accelerometer and (b) to identify groups based on movement patterns and correlate functional and mental characteristics in older adults recovering from hip fracture.
Methods
Design and Population
This study was part of the inception cohort-based study HIP CARE (Hip fractures: Inventorization of Prognostic factors and their Contribution towArds Rehabilitation in older pErsons; NTR NL7491). The goal of the HIPCARE study was to determine functional recovery, quality of life, and health care use during the first year after a hip fracture and was initiated in 2018 (van der Sijp et al., 2021). HIPCARE study participants are older adults (≥70 years) who in most cases were admitted to geriatric rehabilitation facilities in the Netherlands with a proximal femoral fracture. This is a single-center study in which patients are followed from one single hospital, to multiple regional geriatric rehabilitation facilities and home. Patients with high energy trauma or pathological fractures were excluded.
In the current cross-sectional cohort study, between January 2019 and March 2020, a selection of eligible patients of the HIPCARE cohort received an accelerometer for 7 days after an outpatient check-up 3 months after surgery. The goal was to obtain descriptive insights in the activity of the independent mobile patients of the HIPCARE cohort included. Patients were instructed to wear the accelerometer on their waist 24 hr/day for seven consecutive days. For the reliable estimation of movement variables, we only included patients who wore the accelerometer for at least 13 hr for a minimum of 2 days (van Schooten et al., 2015). As the accelerometer could not clearly distinguish between sleeping and sedentary behavior, accelerometer data between 23:00 p.m. and 7:00 a.m. were excluded to avoid misclassification of sedentary behavior as sleep.
Assessments
Baseline characteristics were assessed during admission for surgery and comprised age, sex, and body mass index, as well as general health status using the American Society of Anesthesiologists classification (Parenti et al., 2016). The following assessments were registered during the outpatient check-up 3 months after surgery. Cognition was evaluated using the 6-Item Cognitive Impairment Test (6CIT; range 0–28, lower scores indicate better cognitive functioning; O’Sullivan et al., 2016). Activities of daily living (ADL) functioning was measured using the Katz Index of Independence in ADL (range, 0–6, higher scores indicate better ADL functioning; Katz et al., 1970). Mobility was assessed using The Parker Mobility Score (range, 0–9, higher scores indicate better mobility; Parker & Palmer, 1993), Short Physical Performance Battery Living (SPPB, range, 0–12, higher scores indicate better lower extremity; Guralnik et al., 1994), Functional Ambulation Classification (FAC, ranges from 0: non-functional walking to 5: independent walking outside; Holden et al., 1984), and Timed Up & Go test (TUG, lower scores indicate better mobility; Podsiadlo & Richardson, 1991). Fear of falling was evaluated using the Falls Efficacy Scale International (FES, range 16–64, higher scores indicate greater fear of falling; Jørstad et al., 2005). Evaluation of hip fracture was assessed using the Harris Hip Score (HHS), which is a disease-specific measure for measuring outcomes after hip arthroplasty and includes the domains pain, function, deformity, and range of motion (range 0–100, higher scores indicate better functioning; Harris, 1969). Quality of life was measured using the Dutch version of EuroQol (EQ-5D-5L) and the Visual Analog Scale of EuroQol (EQ-5D-5L VAS; Group, 1990).
Movement variables were measured using the Dynaport MoveMonitor (Dynaport MoveMonitor, McRoberts BV), which is an accelerometer that records acceleration in a triaxial direction. Based on the measured accelerations, the DynaPort MoveMonitor classifies three postures (lying down, sitting, and standing) and four movements (walking, cycling, climbing stairs, and shuffling). Physical activity is quantified as movement intensity (average body acceleration during a specific activity), which can be subdivided into the activity levels light, moderate, or vigorous, based on METs (Haskell et al., 2007). Movement variables included in this study were mainly derived from a recent review on this topic (Boerema et al., 2020) and are described in further detail in Box 1.
Types of Movement Variables
| Physical activity variables | |
| Steps | Total steps per day (mean steps/day) |
| Light activities | Time spent in light activities below 3 METs (mean hours/day; Haskell et al., 2007) |
| Moderate activities | Time spent in moderate activities above or equal to 3 METs and below 6 METs (mean hours/day; Haskell et al., 2007) |
| Vigorous activities | Time spent in vigorous activities above or equal to 6 METs (mean minutes/day; Haskell et al., 2007) |
| Sedentary behavior variables | |
| Sedentary behavior | A minimal duration of 1 min in consecutive lying or sitting (mean hours/day) |
| Sedentary bouts ≥60 min per day | Time spent in sedentary bouts (uninterrupted periods of sitting and lying down) equal or above 60 min. Provides an indication of time spent in prolonged sedentary behavior (median number of bouts/2 days) |
| Half-life bout duration (W50%) | A weighted median bout duration in which the bout duration above and below half of all sedentary time is accumulated. Provides a good indication of centrality given the distribution of bout length (minutes; Chastin & Granat, 2010; Chastin et al., 2015) |
| Alpha | A scaling parameter that provides an indication of the distribution of sedentary bouts. A lower alpha indicates that sedentary time is largely accumulated in long bouts (unit-less variable; Chastin & Granat, 2010) |
Statistical Analysis
Principal Component Analysis
In preparation for the cluster analyses, a principal component analysis (PCA) was performed to reduce the number of dimensions of the included movement variables described in Box 1 while maintaining maximum information (Von Luxburg, 2010). Movement variables were standardized using z scores. Prior to analysis the Kaiser–Meyer–Olkin measure was used to assess the suitability of the overall PCA model. Individual movement variables with at least one correlation coefficient greater than .3 and a (Kaiser–Meyer–Olkin) measure greater than .6 were included in the PCA (Statistics, 2015). Components with eigenvalues ≥1 were used for extraction. A Pearson’s product–moment correlation was run to assess the relationship between the extracted components.
Cluster Analysis
“Components” extracted from the PCA were used to identify different movement pattern groups using k-means clustering. Due to the exploratory nature of the present study, the optimal number of clusters was determined using Silhouette analysis (Rousseeuw, 1987).
Normality of data was tested using the Shapiro–Wilk test. Differences between patient characteristics and movement variables were evaluated using one-way analysis of variance for normally distributed data and are presented as means with SDs (±). The Kruskal–Wallis test was used for nonnormally distributed data and is presented as medians with interquartile range. Baseline data and assessments outcomes 3 months after surgery were compared for all patients included in the HIPCARE study, and for groups identified through cluster analysis. Data were analyzed with SPSS (version 25.0).
Physical activity levels over 1 day were visualized in multiple series line graphs (Microsoft Excel) for individual clusters of each group’s mean percentage activity for each 60-min period.
Results
Fifty-six eligible patients agreed to additional data collection 3 months after surgery, which was 27% of patients from the original HIPCARE study between January 2019 and March 2020. Forty-three patients had sufficient accelerometer wear time of 2 days or more and were included in the analysis. Patient’s characteristics are described in detail in Table 1. The median (interquartile range) age of patients was 81 (interquartile range 75–89), and 29 patients (67%) were female. Regarding patient’s fractures, 22 (51%) had a femoral neck fracture, 19 (44%) had a pertrochanteric femoral fracture, and two (6%) had a subtrochanteric fracture. Surgical treatments included osteosynthesis/internal fixation (29 patients, 67%) or a prosthesis/arthroplasty treatment (14 patients, 33%). Except for age (75 vs. 81, p ≤ .01) and The Parker Mobility (5.8 vs. 5.7, p = .02), there were no significant differences at baseline between patients included in the current study and patients in the overall HIPCARE study regarding demographics (sex, p = .34; body mass index, p = .59; American Society of Anesthesiologists classification, p = .82; postoperative discharge location, p = .14) and assessments 3 months after surgery regarding mobility (FAC, p = .09; TUG, p = .21), fear of falling (FES, p = .90), physical function (KATZ-ADL, p = .23; SPPB, p = .83), hip fracture (HHS, p = .45), cognitive function (6CIT, p = .09), or quality of life (EQ-5D-5L, p = .43; EQ-5D-5L VAS, p = .23).
Patient Characteristics (Mean±, Median IQR)
| Baseline | 3 months after surgery | |
|---|---|---|
| Age (y) | 81 (75–88) | — |
| Sex, female (%) | 29 (67%) | — |
| BMI (kg/m2) | 24.0 ± 3.0 | — |
| Comorbidity (ASA) | 2 (2–3) | — |
| Time since fracture (days) | 92.2 ± 6.8 | |
| Postoperative discharge location | — | |
| Home | 9 (21%) | — |
| Geriatric rehabilitation | 34 (79%) | — |
| Length of stay rehabilitation (days) | 48 (42–66) | — |
| Current level of received care | — | |
| Independent | 41 (96%) | |
| Unknown | 2 (4%) |
Note. BMI = body mass index; ASA = American Society of Anesthesiologists classification; IQR = interquartile range.
Movement Patterns
PCA revealed two components that had eigenvalues ≥ 1, which together explained 71% of the total variance. The Kaiser–Meyer–Olkin for the complete PCA model was 0.74, indicating that the model was middling (Kaiser, 1974). The first, sedentary behavior component (accounting for 58% of variance) mostly included movement variables related to sedentary behavior, with strong positive loadings of mean time spent in sedentary behavior, mean time spent in sedentary bouts ≥ 60 min per day, half-life bout duration (W50%), mean time spent in light activities, and negative loadings of mean steps per day, mean time spent in moderate activities, and alpha. Higher values in the sedentary behavior component indicate more sedentary behavior. The second, physical activity component (13% variance) included movement variables related to physical activity with strong positive loadings of mean steps per day, mean time spent in moderate activities, mean time spent in vigorous activities, and negative loading of mean time spent in sedentary bouts ≥60 min/day. Higher values on the physical activity component indicate more active behavior. Component loadings are described in Supplementary Material (available online). Three groups could be identified through k-means clustering, and the mean silhouette score for all clusters was 0.54. A scatterplot of the clusters and components can be found in Figure 1. Movement variables per group are presented in Table 2 and visualized in Figure 2.
—Scatterplot cluster analysis. The graph presents the groups in relation to the two components. The sedentary behavior component included movement variables related to sedentary behavior, for example, mean time spent in sedentary behavior, mean time spent in sedentary bouts ≥60 min per day, half-life bout duration (W50%), alpha, and mean time spent in light activates. The physical activity component included variables related to physical activity: mean steps per day, mean time spent in moderate activities, and mean time spent in vigorous activities.
Citation: Journal of Aging and Physical Activity 32, 3; 10.1123/japa.2023-0090
Movement Variables per Group (Mean ±, Median Interquartile Range)
| Group 1 (n = 17) | Group 2 (n = 21) | Group 3 (n = 5) | Total | p | |
|---|---|---|---|---|---|
| Steps | 232.6 (32.6–623.1) | 1,786.2 (920–4,319.9) | 7,392.5 (4,242.2–7,392.5) | 1,235.9 (387.4–3,034.3) | <.01*,§ |
| Sedentary behavior | 13.3 ± 0.7 | 10.5 ± 1.3 | 10.6 ± 1.5 | 11.6 ± 1.8 | <.01§,† |
| Light activities | 14.3 ± 0.7 | 12.9 ± 1.1 | 12.7 ± 0.8 | 13.4 ± 1.2 | <.01*,§ |
| Moderate activities | 0.2 (0.1–0.4) | 0.9 (0.4–1.4) | 1.9 (1.4–2.1) | 0.5 (0.2–1.2) | <.01*,§ |
| Vigorous activities | 0.3 (0.2–0.6) | 0.4 (0.2–1.0) | 3.8 (2.9–4.6) | 0.4 (0.2–1.0) | <.01§,† |
| Sedentary bouts ≥60 min/day | 24.0 (15.0–26.5) | 8.0 (3.0–14.0) | 4.0 (1.5–12.5) | 13.0 (5.0–23.0) | <.01*,§ |
| Half-life bout duration (W50%) | 124.0 (92.5–162,5) | 36.0 (30.5–60.5) | 30.0 (22.5–42.5) | 56.0 (32.0–104.0) | <.01*,§ |
| Alpha | 1.3 ± 0.1 | 1.5 ± 0.1 | 1.5 ± 0.0 | 1.4 ± 0.1 | <.01*,§ |
| Worn time | 14.6 (14.1–13.8) | 13.9 (13.3–14.6) | 14.7 (14.1–15.2) | 14.5 (13.7–13.8) | .06 |
| Sedentary behavior component | 1.0 ± 0.4 | −0.8 ± 0.5 | −0.0 ± 0.8 | 0.0 ± 1.0 | <.01*,§,† |
| Physical activity component | −0.4 ± 0.4 | −0.3 ± 0.6 | 2.3 ± 0.5 | 0.0 ± 1.0 | <.01*,§,† |
Note. Higher values in the sedentary behavior component indicate more sedentary behavior and higher values on the physical activity component indicate more active behavior.
*Statistically significant differences between Groups 1 and 2. §Statistically significant differences between Groups 1 and 3. †Statistically significant differences between Groups 2 and 3.
—Physical activity distributed over day time and visualized per group. Physical activity levels for individual clusters mean percentage activity for each 60-min period.
Citation: Journal of Aging and Physical Activity 32, 3; 10.1123/japa.2023-0090
Group 1 (n = 17, 39.5% of all included patients) was characterized by a very low step count, a very low level of physical activity, and a very high level of sedentary behavior that was largely accumulated (long sedentary bouts). Group 2 (n = 21, 48.8%) was characterized by a moderate step count, a low level of physical activity, and a high level of sedentary behavior that was more evenly distributed across moderate sedentary bouts. Group 3 (n = 5, 11.6%) was characterized by a high step count, a moderate level of physical activity, and a moderate level of sedentary behavior that was relatively evenly distributed across shorter sedentary bouts.
Assessments
At baseline, there were no significant differences between the groups regarding mobility (The Parker Mobility, p = .22), physical function (KATZ-ADL, p = .29), quality of life (EQ-5D-5L, p = .79; EQ-5D-5L, VAS, p = .26), or cognitive function (6CIT, p = .71). Due to the assessment on the day of the hip fracture the SPPB, TUG, FES, HHS, and FAC could not be assessed at baseline. While the HHS differed significantly between all groups, there were no significant differences between the groups on specific items such as pain (HHS, domain pain; p = .06) or the locomotion functions of the hip joint (HHS, domain locomotion; p = .194) that could independently affect physical function. Furthermore, there was a small significant difference between all groups regarding type of surgical treatment (p = .03). There was no significant difference regarding postoperative discharge location (p = .07).
Assessments by group during the outpatient check-up 3 months after surgery are described in Table 3. Between all groups, there were no differences regarding physical function (KATZ-ADL, p = .19) and cognitive function (6CIT, p = .95). Except for physical function (SPPB, p = .272), mobility (FAC, p = .60), and quality of life (EQ-5D-5L, p = .28) all assessments were statistically significant different between Groups 1 and 2. Between Groups 1 and 3, all assessment were statistically different. Groups 2 and 3 differed on mobility (FAC, p = .27; TUG, p = .15; TPB, p = .79), fear of falling (FES, p = .37), and quality of life (EQ-5D-5L, p = .26; EQ-5D-5L, VAS, p = .76).
Patient Characteristics and Assessments per Group (Mean±, Median IQR)
| Group 1 (n = 17) | Group 2 (n = 21) | Group 3 (n = 5) | Total | p | |
|---|---|---|---|---|---|
| Characteristics | |||||
| Age | 81 (73–90) | 80 (73–90) | 76 (73–80) | 81 (75–88) | .43 |
| Sex, female (%) | 12 (70%) | 14 (67%) | 3 (60%) | 29 (67%) | .90 |
| Surgical treatment | .03† | ||||
| Osteosynthesis/internal | 11 (65%) | 17 (81%) | 1 (20%) | 29 | — |
| Prosthesis/arthroplasty | 6 (35%) | 4 (19%) | 4 (20%) | 14 | — |
| Postoperative discharge location | .07 | ||||
| Home | 3 (18%) | 3 (14%) | 3 (60%) | 9 | — |
| Geriatric rehabilitation | 14 (82%) | 18 (86%) | 2 (40%) | 34 | — |
| Physical function | |||||
| KATZ-ADL | 1.0 (0.0–2.0) | 0.0 (0.0–1.5) | 0.0 (0.0–0.0) | 0.0 (0.0–2.0) | .193 |
| SPPB | 5.4 ± 2.4 | 6.8 ± 3.0 | 10.8 ± 0.8 | 6.8 ± 3.1 | <.01§,† |
| Hip fracture | |||||
| HHS | 57.0 ± 12.3 | 66.7 ± 13.4 | 91.8 ± 5.8 | 66.4 ± 16.2 | <.01*,§,† |
| Mobility | |||||
| FAC | 4.0 (3.0–4.0) | 4.0 (4.0–4.5) | 5.0 (4.0–5.0) | 4.0 (4.0–4.0) | <.01§ |
| TUG | 31.9 (23.1–35.1) | 18.2 (13.1–23.1) | 10.0 (8.5–11.2) | 20.4 (12.2–35.0) | <.01*,§ |
| TPM | 4.0 (2.5–6.0) | 6.0 (6.0–6.5) | 9.0 (6.5–9.0) | 6.0 (4.0–6.0) | <.01*,§ |
| Fear of falling | |||||
| FES | 12.0 (9.5–15.0) | 9.0 (7.0–11.7) | 7.0 (7.0–7.5) | 9.5 (7.0–12.3) | <.01*,§ |
| Cognitive function | |||||
| 6CIT | 2.0 (0.0–5.5) | 2.0 (0.0–3.5) | 2.0 (0.0–8.5) | 2.0 (0.0–4.2) | .948 |
| Quality of life | |||||
| EQ-5D-5L | 0.6 (0.4–0.7) | 0.7 (0.6–0.8) | 0.9 (0.8–0.9) | 0.7 (0.6–0.8) | <.01§ |
| EQ-5D-5L VAS | 60 (50.0–72.5) | 75.0 (67.5–80.0) | 80.0 (70.0–80.0) | 70.0 (60.0–80.0) | .02*,§ |
Note. KATZ-ADL = the Katz Index of Independence in Activities of Daily Living; SPPB = Short Physical Performance Battery Living; HHS = Harris Hip Score; FAC = Functional Ambulation Classification; TUG = Timed Up & Go test; TPM = The Parker Mobility Score; FES = Falls Efficacy Scale International; 6CIT = the 6-Item Cognitive Impairment Test; EQ-5D-5l = Dutch version of EuroQol; EQ-5D-5L VAS = Visual Analog Scale of EuroQol; IQR = interquartile range.
*Statistically significant differences between Groups 1 and 2. §Statistically significant differences between Groups 1 and 3. †Statistically significant differences between Groups 2 and 3.
Discussion
Principal Findings
In this study of older adults recovering from hip fracture 3 months after surgery, we identified three groups as defined by divergent levels of physical activity and sedentary behavior. Our two main findings were: (a) across all groups older adults recovering from a hip fracture engaged in very low levels of physical activity and spent an average of 11 hr/day in prolonged sedentary behavior and (b) based on movement patterns, we identified three distinct groups with substantial differences in levels of physical activity and sedentary behavior. Finally, no relationship was found between patient characteristics at baseline and movement patterns 3 months after surgery.
Comparison With Previous Studies
A unique aspect of our study was the evaluation of the pattern of sedentary behavior. Our results suggest a clear difference between the three groups in terms of the pattern of sedentary behavior, with Group 1 showing a significantly higher proportion of long sedentary bouts. While previous studies have indicated that older adults recovering from a hip fracture tend to show very little physical activity and devote a significant amount of time to sedentary behavior (Fleig et al., 2016; Zusman et al., 2018, 2019), none of these studies evaluated the pattern of sedentary behavior. By interrupting prolonged sedentary periods, associated risks can be reduced, since prolonged sedentary behavior poses a health risk independent of total sedentary time (Dunstan et al., 2012; Duvivier et al., 2017; Sardinha et al., 2015). Evaluating patterns of sedentary behavior may provide a better understanding of the effects of interventions designed to disrupt prolonged sedentary periods (Chastin et al., 2015). Furthermore during the early phase of rehabilitation planned and individually delivered comprehensive geriatric care in a geriatric hospital ward with particular focus on mobilization could improve physical activity and reduce sedentary behavior (Taraldsen et al., 2013). However, the extent to which a reduction in sedentary behavior reduces certain health risks remains to be determined (Lewis et al., 2016).
Based on accelerometery, we identified three groups of patients that differed regarding sedentary behavior and the intensity of activity. When we compared the clinically assessed data associated with these three groups, a clear pattern emerged: a low intensity of physical activity in combination with sedentary behavior correlated with lower scores for mobility, physical function, hip fracture, and quality of life, as well as a greater fear of falling. However, no significant differences were found between Group 2 and Group 3 on assessments related to mobility. This may indicate that patients in Group 2 are functionally able to increase physical activity and reduce sedentary behavior, but did not display these movement patterns. Another potential explanation may be that current assessments related to mobility are unable to account for the difference in physical activity or sedentary behavior between Groups 2 and 3. Accelerometery-based observation of objective activity intensity and sedentary behavior can, thus provide early indications of a changing health status, allowing timely tailored interventions (White et al., 2020).
While the number of patients per group varied widely, Group 3 was considerably smaller compared with Groups 1 and 2. This distribution may be attributable to the cluster technique used, which does not allow for the size of clusters. Earlier studies of physical activity and sedentary behavior variables that used similar clustering techniques also reported uneven distributions of patients per group, these studies also found that the smallest group consisted of the most physically active and least sedentary patients (Mesquita et al., 2017; Wondergem et al., 2019).
We found no relation between patient characteristics at baseline and factors including pain, locomotion functions of the hip, and movement patterns. However, we did observe a significant difference in terms of type of surgical treatment. In Group 3, relatively more patients were treated by prosthesis/arthroplasty compared with osteosynthesis/internal fixation. Previous studies reported that patients after hip fracture who were treated by osteosynthesis/internal fixation had a higher reoperation rate, higher long-term mortality, and lower quality of life after 4 months, compared with patients treated by prosthesis/arthroplasty (Gjertsen et al., 2008; Moerman et al., 2016). Furthermore, although we found no significant difference in postoperative discharge location, we observed that the number of patients discharged to geriatric rehabilitation in Group 3 was relatively lower than in Groups 1 and 2. This may indicate that patients in Group 3 had better health status postoperatively and did not need geriatric rehabilitation. As the number of patients in Group 3 is small, it is not possible at this time to draw a clear conclusion as to whether the difference in surgical treatment and discharge location has an effect on the movement patterns found. Finally, the differences found could potentially be explained by “confounding by indication”: the choice of surgical procedure is not random but related to the complexity of the fracture injury.
Nevertheless, other factors besides physical components likely impact the intensity of physical activity and sedentary behavior. Previous qualitative research identified several barriers that can constrain engagement in physical activities and encourage sedentary behavior, such as fear of falling, lack of motivation, fatigue, lack of time, or lack of knowledge (Gorman et al., 2012; Moraes et al., 2020; Nicholson et al., 2013; Salmon et al., 2003). Theory-based behavior change techniques, in combination with a stepwise approach that begins by targeting prolonged sedentary bouts, might help lower these barriers (Dogra et al., 2022; Prince et al., 2014).
In our study, “light activity” was defined as time spent in all activities below 3 METs. This meant that “sedentary behavior” also included all light activities. Other studies that have examined light activities in older adults recovering from a hip fracture chose other cut off points for levels of activity, which resulted in the exclusion of activities related to sedentary behavior (Fleig et al., 2016; Zusman et al., 2019). This difference in classification method may have impacted our results, as the amount of light activity in our study was significantly higher than in comparable studies.
Finally, of the two components from the PCA, the sedentary behavior component consisted almost entirely of variables related to sedentary behavior, while the physical activity component consisted of physical activity variables. This might indicate that physical activity and sedentary behavior are not interdependent, suggesting that both can be influenced independently to achieve improvements in health status (Gardner et al., 2016; Prince et al., 2014).
Strengths and Limitations
A strength of this study was the use of an accelerometer and a strict data inclusion protocol in which we only used data from patients who wore the accelerometer for at least 13 hr for a minimum of 2 days. This provided an objective, accurate, and reliable assessment of physical activity and sedentary behavior. Another strength was the comprehensive description of all groups in terms of physical activity, sedentary behavior, and other assessments. Through PCA and k-means clustering, we obtained a particularly good picture of a subgroup of older physically very inactive adults who therefore might have a higher risk of further functional decline. While we were able to objectively assess sedentary behavior using an accelerometer, our protocol did not allow us to distinguish between sedentary behavior in a sleeping or awake state. Although exclusion of data between 23:00 p.m. and 7:00 a.m. likely included the bulk of sleep data, our ability to accurately distinguish sedentary behavior from sleep was nevertheless limited and may have influenced our results. This approach was chosen because including night sleep as “sedentary behavior” would have diluted relative activity and thus, reduced sensitivity to discriminate groups based on activity. Furthermore, during this study, we did not record whether patients also received physical therapy during the accelerometer wearing period. This may have some effect on the results found, leading to a slight overestimation of the physical activity measures. Another limitation was the small sample size, partly because we had a strict inclusion protocol for the sensor data. We did not perform a sample size calculation prior to the study, as the current research question was secondary in the HIPCARE study. However, this did limit the inclusion of movement variables in the PCA. Finally, no significant differences were found at baseline between the current participants and those included in the HIPCARE study.
Conclusions
Our primary conclusion is that older adults recovering from a hip fracture indeed engage in very low levels of physical activity and spend prolonged periods of time in sedentary behavior. Second, based on actual movement patterns recorded by a wearable accelerometer, three distinct groups of older adults recovering from a hip fracture could be distinguished, each with distinct levels of intensity of physical activity and temporal sedentary behavior. Third, within these three groups, a clear association was found between a low intensity of physical activity in combination with long sedentary periods and lower scores for mobility, physical function, hip fracture and quality of life, as well as a greater fear of falling. Finally, we argue that evaluation of the pattern of sedentary behavior is essential when assessing the effectiveness of interventions aimed at reducing sedentary behavior.
Future research should focus on determining the level of reduction in sedentary behavior required to lower health risks. Classification of sedentary behavior should exclude sleep periods and light activities. Furthermore, multidisciplinary rehabilitation programs need to place greater emphasis on the contribution of sedentary behavior by not only promoting physical activity, but by also including interventions designed to reduce sedentary behavior. Those interventions should be tailored and include theory-based behavior change techniques, in combination with a stepwise approach that starts by targeting prolonged sedentary bouts.
Acknowledgments
The authors thank Miriam Haaksma and Eveline Grootendorst for their help with data analysis. Ethics Approval and Consent to Participate: The HIPCARE study and any subsequent observational study were approved by the institutional Medical Ethics Committee Southwest Holland (P18.029). Availability of Data and Materials: The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Funding: The HIPCARE study was funded by ZonMW (funding number: 839120010). Authors’ Contributions: Kraaijkamp, De Groot, Stijntjes, and van Dam van Isselt conceptualized the study. Kraaijkamp prepared the original draft, collected data, conducted the analysis, and visualization of the data. All authors read and approved the final manuscript.
References
Boerema, S.T., van Velsen, L., Vollenbroek, M.M., & Hermens, H.J. (2020). Pattern measures of sedentary behaviour in adults: A literature review. Digital Health, 6, Article 418.
Caspersen, C.J., Powell, K.E., & Christenson, G.M. (1985). Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. Public Health Reports, 100(2), 126–131.
Chastin, S.F., & Granat, M.H. (2010). Methods for objective measure, quantification and analysis of sedentary behaviour and inactivity. Gait Posture, 31(1), 82–86.
Chastin, S.F.M., Winkler, E.A.H., Eakin, E.G., Gardiner, P.A., Dunstan, D.W., Owen, N., & Healy, G.N. (2015). Sensitivity to change of objectively-derived measures of sedentary behavior. Measurement in Physical Education and Exercise Science, 19(3), 138–147.
Chau, J.Y., Grunseit, A.C., Chey, T., Stamatakis, E., Brown, W.J., Matthews, C.E., ... van der Ploeg, H.P. (2013). Daily sitting time and all-cause mortality: A meta-analysis. PLoS One, 8(11), Article 80000.
Dogra, S., Copeland, J.L., Altenburg, T.M., Heyland, D.K., Owen, N., & Dunstan, D.W. (2022). Start with reducing sedentary behavior: A stepwise approach to physical activity counseling in clinical practice. Patient Education and Counseling, 105(6), 1353–1361.
Dunstan, D.W., Kingwell, B.A., Larsen, R., Healy, G.N., Cerin, E., Hamilton, M.T., ... Owen, N. (2012). Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care, 35(5), 976–983.
Duvivier, B.M.F.M., Schaper, N.C., Hesselink, M.K.C., van Kan, L., Stienen, N., Winkens, B., ... Savelberg, H.H.C.M. (2017). Breaking sitting with light activities vs structured exercise: A randomised crossover study demonstrating benefits for glycaemic control and insulin sensitivity in type 2 diabetes. Diabetologia, 60(3), 490–498.
Dyer, S.M., Crotty, M., Fairhall, N., Magaziner, J., Beaupre, L.A., Cameron, I.D, ... Fragility Fracture Network Rehabilitation Research Special Interest Group (2016). A critical review of the long-term disability outcomes following hip fracture. BMC Geriatrics, 16(1), Article 158.
Ekegren, C.L., Beck, B., Climie, R.E., Owen, N., Dunstan, D.W., & Gabbe, B.J. (2018). Physical activity and sedentary behavior subsequent to serious orthopedic injury: A systematic review. Archives of Physical Medicine and Rehabilitation, 99(1), 164–177.e166.
Fleig, L., McAllister, M.M., Brasher, P., Cook, W.L., Guy, P., Puyat, J.H., ... Ashe, M.C. (2016). Sedentary behavior and physical activity patterns in older adults after hip fracture: A call to action. Journal of Aging and Physical Activity, 24(1), 79–84.
Gardner, B., Smith, L., Lorencatto, F., Hamer, M., & Biddle, S.J.H. (2016). How to reduce sitting time? A review of behaviour change strategies used in sedentary behaviour reduction interventions among adults. Health Psychology Review, 10(1), 89–112.
Gianoudis, J., Bailey, C.A., & Daly, R.M. (2015). Associations between sedentary behaviour and body composition, muscle function and sarcopenia in community-dwelling older adults. Osteoporosis International, 26(2), 571–579.
Gjertsen, J.E., Vinje, T., Lie, S.A., Engesaeter, L.B., Havelin, L.I., Furnes, O., & Fevang, J.M. (2008). Patient satisfaction, pain, and quality of life 4 months after displaced femoral neck fractures: A comparison of 663 fractures treated with internal fixation and 906 with bipolar hemiarthroplasty reported to the Norwegian Hip Fracture Register. Acta Orthopaedica, 79(5), 594–601.
Gorman, E., Chudyk, A.M., Hoppmann, C.A., Hanson, H.M., Guy, P., Sims-Gould, J., & Ashe, M.C. (2012). Exploring older adults’ patterns and perceptions of exercise after hip fracture. Physiotherapy Canada, 65(1), 86–93.
Group, T.E. (1990). EuroQol-a new facility for the measurement of health-related quality of life. Health Policy, 16(3), 199–208.
Guralnik, J.M., Simonsick, E.M., Ferrucci, L., Glynn, R.J., Berkman, L.F., Blazer, D.G., ... Wallace, R.B. (1994). A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and nursing home admission. Journal of Gerontology, 49(2), M85–M94.
Harris, W.H. (1969). Traumatic arthritis of the hip after dislocation and acetabular fractures: Treatment by mold arthroplasty. An end-result study using a new method of result evaluation. The Journal of Bone and Joint Surgery. American Volume, 51(4), 737–755.
Hart, T.L., Swartz, A.M., Cashin, S.E., & Strath, S.J. (2011). How many days of monitoring predict physical activity and sedentary behaviour in older adults? International Journal of Behavioral Nutrition and Physical Activity, 8(1), Article 62.
Haskell, W.L., Lee, I.M., Pate, R.R., Powell, K.E., Blair, S.N., Franklin, B.A., ... Bauman, A. (2007). Physical activity and public health: Updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation, 116(9), Article 1081.
Holden, M.K., Gill, K.M., Magliozzi, M.R., Nathan, J., & Piehl-Baker, L. (1984). Clinical gait assessment in the neurologically impaired: Reliability and meaningfulness. Physical Therapy, 64(1), 35–40.
Jørstad, E.C., Hauer, K., Becker, C., & Lamb, S.E. (2005). Measuring the psychological outcomes of falling: A systematic review. Journal of American Geriatric Society, 53(3), 501–510.
Kaiser, H.F. (1974). An index of factorial simplicity. Psychometrika, 39(1), 31–36.
Katz, S., Downs, T.D., Cash, H.R., & Grotz, R.C. (1970). Progress in development of the index of ADL. The Gerontologist, 10(1_Part_1), 20–30.
Klenk, J., Wekenmann, S., Schwickert, L., Lindemann, U., Becker, C., & Rapp, K. (2019). Change of objectively-measured physical activity during geriatric rehabilitation. Sensors, 19(24), Article 5451.
Lewis, L.K., Hunt, T., Williams, M.T., English, C., & Olds, T.S. (2016). Sedentary behavior in people with and without a chronic health condition: How much, what and when? AIMS Public Health, 3(3), 503–519.
Mesquita, R., Spina, G., Pitta, F., Donaire-Gonzalez, D., Deering, B.M., Patel, M.S., ... Spruit, M.A. (2017). Physical activity patterns and clusters in 1001 patients with COPD. Chronic Respiratory Disease, 14(3), 256–269.
Moerman, S., Vochteloo, A.J., Tuinebreijer, W.E., Maier, A.B., Mathijssen, N.M., & Nelissen, R.G. (2016). Factors associated with the course of health-related quality of life after a hip fracture. Archives of Orthopaedic and Trauma Surgery, 136(7), 935–943.
Moraes, S.A., Furlanetto, E.C., Ricci, N.A., & Perracini, M.R. (2020). Sedentary behavior: Barriers and facilitators among older adults after hip fracture surgery. A qualitative study. Brazilian Journal of Physical Therapy, 24(5), 407–414.
Nicholson, S., Sniehotta, F.F., van Wijck, F., Greig, C.A., Johnston, M., McMurdo, M.E., ... Mead, G.E. (2013). A systematic review of perceived barriers and motivators to physical activity after stroke. International Journal of Stroke, 8(5), 357–364.
O’Sullivan, D., O’Regan, N.A., & Timmons, S. (2016). Validity and reliability of the 6-item cognitive impairment test for screening cognitive impairment: A review. Dementia and Geriatric Cognitive Disorders, 42(1–2), 42–49.
Parenti, N., Reggiani, M.L.B., Percudani, D., & Melotti, R.M. (2016). Reliability of American society of anesthesiologists physical status classification. Indian Journal of Anaesthesia, 60(3), Article 208.
Parker, M.J., & Palmer, C.R. (1993). A new mobility score for predicting mortality after hip fracture. The Journal of Bone and Joint Surgery. British Volume, 75(5), 797–798.
Podsiadlo, D., & Richardson, S. (1991). The timed “up & go”: A test of basic functional mobility for frail elderly persons. Journal of American Geriatric Society, 39(2), 142–148.
Prince, S.A., Saunders, T.J., Gresty, K., & Reid, R.D. (2014). A comparison of the effectiveness of physical activity and sedentary behaviour interventions in reducing sedentary time in adults: A systematic review and meta-analysis of controlled trials. Obesity Reviews, 15(11), 905–919.
Resnick, B., Galik, E., Boltz, M., Hawkes, W., Shardell, M., Orwig, D., & Magaziner, J. (2011). Physical activity in the post-hip-fracture period. Journal of Aging and Physical Activity, 19(4), 373–387.
Rousseeuw, P.J. (1987). Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20, 53–65.
Salmon, J., Owen, N., Crawford, D., Bauman, A., & Sallis, J.F. (2003). Physical activity and sedentary behavior: A population-based study of barriers, enjoyment, and preference. Health Psychology, 22(2), 178–188.
Sardinha, L.B., Ekelund, U., dos Santos, L., Cyrino, E.S., Silva, A.M., & Santos, D.A. (2015). Breaking-up sedentary time is associated with impairment in activities of daily living. Experimental Gerontology, 72, 57–62.
Statistics, L. (2015). Principal Components Analysis (PCA) using SPSS statistics. Statistical Tutorials and Software Guides.
Talkowski, J.B., Lenze, E.J., Munin, M.C., Harrison, C., & Brach, J.S. (2009). Patient participation and physical activity during rehabilitation and future functional outcomes in patients after hip fracture. Archives of Physical Medicine Rehabilitation, 90(4), 618–622.
Taraldsen, K., Sletvold, O., Thingstad, P., Saltvedt, I., Granat, M.H., Lydersen, S., & Helbostad, J.L. (2013). Physical behavior and function early after hip fracture surgery in patients receiving comprehensive geriatric care or orthopedic care—A randomized controlled trial. The Journals of Gerontology: Series A, 69(3), 338–345.
Thibaud, M., Bloch, F., Tournoux-Facon, C., Brèque, C., Rigaud, A.S., Dugué, B., & Kemoun, G. (2012). Impact of physical activity and sedentary behaviour on fall risks in older people: A systematic review and meta-analysis of observational studies. European Review of Aging and Physical Activity, 9(1), 5–15.
Tremblay, M.S., Aubert, S., Barnes, J.D., Saunders, T.J., Carson, V., Latimer-Cheung, A.E., ... Chinapaw, M.J. (2017). Sedentary behavior research network (SBRN)–terminology consensus project process and outcome. International Journal of Behavioral Nutrition and Physical Activity, 14(1), Article 75.
van der Sijp, M.P.L., van Eijk, M., Niggebrugge, A.H.P., Putter, H., Blauw, G.J., & Achterberg, W.P. (2021). Prognostic factors for short-term recovery of independence in a multistate model for patients with a hip fracture. Journal of American Medical Directors Association, 22(6), 1307–1312.
van Schooten, K.S., Rispens, S.M., Elders, P.J., Lips, P., van Dieën, J.H., & Pijnappels, M. (2015). Assessing physical activity in older adults: Required days of trunk accelerometer measurements for reliable estimation. Journal of Aging Physical Activity, 23(1), 9–17.
Von Luxburg, U. (2010). Clustering stability: An overview. Foundations and Trends® in Machine Learning, 2(3), 235–274.
White, N.D., Bautista, V., Lenz, T., & Cosimano, A. (2020). Using the SMART-EST goals in lifestyle medicine prescription. American Journal of Lifestyle Medicine, 14(3), 271–273.
Wondergem, R., Veenhof, C., Wouters, E.M.J., de Bie, R.A., Visser-Meily, J.M.A., & Pisters, M.F. (2019). Movement behavior patterns in people with first-ever stroke. Stroke, 50(12), 3553–3560.
Zusman, E.Z., Dawes, M., Fleig, L., McAllister, M.M., Cook, W.L., Guy, P., ... Ashe, M.C. (2019). Older adults’ sedentary behavior and physical activity after hip fracture: Results from an outpatient rehabilitation randomized controlled trial. Journal of Geriatric Physical Therapy, 42(2), E32–E38.
Zusman, E.Z., Dawes, M.G., Edwards, N., & Ashe, M.C. (2018). A systematic review of evidence for older adults’ sedentary behavior and physical activity after hip fracture. Clinical Rehabilitation, 32(5), 679–691.
