Real-time Positioning Among Nursing Home Residents Living With Dementia: A Case Study

Dementia contributes to the development of pressure injuries (PrIs). PURPOSE: This study describes the real-time body positions of 2 nursing home (NH) residents, residing in the United States and living with dementia, to inform development of PrI prevention strategies tailored to individual risk profiles. METHODS: As part of a larger study, eligible residents were fitted with a triaxial accelerometer sensor placed on the anterior chest to monitor body positions 24-hours daily through a 4-week monitoring period. The current study used an observational, prospective design during routine repositioning events for 2 residents. A convenience sample of 2 residents from a single NH wing who were considered moderately at risk for PrI development (Braden Scale score 13–14) with a Brief Interview for Mental Status score in the severely impaired range were selected based on nursing staff recommendation. RESULTS:  Sensor data showed that both residents, although “chairfast” according to the Braden Scale, spent <5% in an upright position and the great majority of time reclining at an angle <50%. One (1) resident demonstrated a persistent side preference. CONCLUSIONS: Wearable sensors are not a long-term solution for protecting those with dementia from PrI formation but do provide a crude picture of overall body positions throughout the 24-hour day that may inform individualized PrI prevention strategies. Studies including large samples of NH residents living with dementia are warranted.

Dementia is a degenerative disorder characterized by a decline in mental ability severe enough to impact daily life. The decline of neurophysiologic function in dementia contributes to the risk of developing pressure injuries (PrIs).¹ PrIs are defined as localized damage to the skin and/or underlying tissue as a result of pressure or pressure in combination with shear.^2,3 Specifically, alterations in cognitive, behavioral, perceptual-motor, autonomic, and sensory function resulting from dementia act synergistically to increase risk for PrI. Fluctuation of these functions can affect an individual’s ability to position his or her body—either independently or with assistance—to relieve pressure on the skin and underlying tissue. Currently, there are no studies describing body positions and repositioning among nursing home (NH) residents living with dementia (LWD). In this study, we described the real-time body positions of 2 NH residents LWD over multiple weeks to determine how PrI prevention strategies might be tailored to each unique individual risk profile.

Dementia as a PrI risk factor. The incidence of PrIs varies across skilled nursing settings. Although many factors put NH residents at higher risk,⁴ a mainstay of PrI prevention is offloading or redistributing pressure by changing body position (repositioning). Confirming that a sufficient turn angle is achieved during repositioning events is important to make sure the pressure is offloaded from the affected section of the body, which then restores blood flow to the affected tissue.

NH residents with dementia are particularly vulnerable to PrI development.⁵ Dementia imposes cognitive limitations—including those of learning, memory, language, and executive functions—as well as behavioral, perceptual-motor, autonomic, and sensory limitations.⁶ Together, these factors may increase PrI risk.⁷ For example, expressive language problems may impair communication with care providers when moving to or from a painful position, and receptive language problems may weaken understanding of repositioning instructions. Similarly, memory alterations may challenge the ability to recall the last occurrence of repositioning, and impaired executive functioning may affect basic hygiene, including skin care.⁸ Behavioral disturbances, such as apathy, anxiety, agitation, aggression, or irritability—prevalent among 60%–90% of people with dementia—may present challenges to self or staff-assisted repositioning, as well as increase friction and shearing forces.^9–11 Altered sensory perception may decrease a resident’s ability to localize pain¹² or to express restlessness,¹³ leading to friction and shear. Alterations in proprioception can affect balance,¹⁴ which can decrease mobility and increase risk for fall-related injuries, thereby raising the risk of PrIs. Further complicating issues for those LWD include the following: perceptual motor problems, associated with reduced mobility¹⁵ and affecting the muscles used to swallow, possibly leading to malnutrition and dehydration¹⁶; autonomic nervous system dysfunction increasing PrI risk by impairing blood pressure and heart rate, thus reducing tissue perfusion; and increased sweating and bowel and bladder incontinence, resulting in skin exposure to excessive moisture.^17,18 

Dementia-associated challenges in body positioning. One challenge of assessing the health of NH residents with neurophysiologic decline affecting their ability both to change positions and to offload pressure, is the absence of consistent measures of body position over time. Positioning occurs over a 24-hour period, may be self-initiated or assisted, and observation with the human eye may not accurately capture some aspects of positioning, such as the number of degrees of tilt obtained in a turn/repositioning event. Current published studies describing positioning and/or movement in people LWD have relied on human observations,¹⁵ retrospective electronic health record data,¹¹ and interviews,⁹ or have used sensors to observe individuals as they performed a specific task under tightly controlled conditions for a short period.¹⁴ None of these methods adequately captures body positioning occurring continuously in real-world conditions. Continuous use of sensors (triaxial accelerometer) would allow a greater degree of precision, but no studies have employed sensors to describe real-time body positions among NH residents LWD over extended periods of time. 

A second challenge to protecting health in this population is that standard nursing assessment of PrI risk currently depends on staff observation and assessment with a structured tool, such as the Braden Scale for Predicting Pressure Sore Risk (hereafter Braden Scale).¹⁹ Its activity subscale describes an individual’s best effort at physical activity. According to the subscale, a NH resident is deemed “chairfast” if the patient’s ability to walk is severely limited or nonexistent, the patient cannot bear their own weight, and/or must be assisted into a chair or wheelchair. The Braden Scale does not include a duration of upright time component, which is a key element of risk to tissues. The relationships among duration of time in a body position and PrI risk were addressed in Bergstrom’s landmark study, which concluded repositioning intervals could be extended up to 4 hours without additional risk for PrI; however, the study did not differentiate based on the activity subscale.²⁰   

Standard PrI prevention protocols, informed by Braden Scale scores, require regular repositioning to offload pressure and enable tissue reperfusion.²¹ Because individuals LWD experience changes in body positioning ability, a “one-size-fits-all” approach to repositioning may be a poor fit with the individual’s distinctive positioning profile. The purpose of this small study of 2 cases was to contribute to the evidence base by describing 2 nurse-identified NH residents LWD who were also at moderate risk for PrI according to the Braden Scale, with the ultimate goal of describing the necessary components of tailored, preventive PrI interventions for such residents. 

Design and sample. As part of a larger study (NIH R01NR016001; ClinicalTrials.gov: NCT02996331), this study of 2 cases used an observational, prospective design. The parent clinical trial compared 2-, 3-, and 4-hour repositioning schedules across 9 NHs. To be considered eligible for the clinical trial, NH residents had to 1) have viscoelastic or high-density-foam–supported surfaces, 2) be free from adhesive allergy and existing PrIs, and (3) have scored >10 on the Braden Scale.²² All NHs provided linens and a lift device in each clinical unit, along with 5 pillows per resident. Eligible residents were fitted with a triaxial accelerometer sensor placed on the anterior chest to monitor body positions 24-hours daily through a 4-week monitoring period, as well as time beyond prescribed repositioning interval that the resident remained in the same position, degree, and angle of repositioning change. Demographic characteristics (age, gender, and race/ethnicity), weekly Braden Score most current at the start date of the intervention (10–12 = high, 13–14 = moderate, 15–18 = mild, and 19–23 = low), and primary medical diagnoses (ICD-10-CM) were extracted from the electronic medical record (EMR); the latter was supplemented by information from the Minimum Data Set (MDS).

For the current study, 2 resident cases were selected based on 2 variables: the Braden Scale score extracted from the parent study and their Brief Interview for Mental Status (BIMS) score from the EMR.^19,23 BIMS is a validated, performance-based, cognitive-status screen for NH residents and was performed on all residents of the parent study. We used the BIMS score in addition to a formal ICD-10-CM dementia diagnostic code because some NH residents with clinically evident dementia did not have a formal diagnosis listed in their medical record yet. Cognitive impairment and dementia are severely underdiagnosed in the United States²⁴; among older adults LWD, fewer than half receive a dementia diagnosis,^25,26 and an even smaller proportion have a dementia diagnosis documented in their medical record.^27–30 The BIMS21 assessment evaluates the resident’s current level of orientation, ability to attend to a task, and overall ability to recall information (range, 0–15: 0–7 = severely impaired, 8–12 = moderately impaired, and 13–15 = cognitively intact).

Two (2) residents from a single NH wing considered moderately at risk for PrI development (Braden Scale score 13–14) with a BIMS score in the severely impaired range were selected for observation based on nursing staff recommendation. In prior studies conducted among NH residents, 12% met criteria for moderate risk of PrI  on Braden Scale scores³¹ and 27% were assigned severely impaired BIMS scores.²³ Resident A also had a formal ICD-10-CM diagnosis of dementia, whereas Resident B did not. A study team member observed the 2 residents twice each during standard repositioning events performed by NH nursing staff during the 4-week study. The parent study was approved by the Duke University Institutional Review Board as meeting ethics requirements for human subject research.

Measures. Demographic and clinical variables for the 2 cases were extracted from electronic health record data from the parent study. These included the resident’s age, gender, race, ethnicity, Braden Scale score,¹⁹ medical diagnoses, and Braden Scale activity subscale.³¹ 

Position data were monitored and documented with the Leaf Patient Monitoring System sensor in the parent study.³² Measures of body position, including turn frequency and turn angle (upright, roll, and tilt angle combinations) at 10-second intervals, are presented in Table 1. A head elevation threshold of 50 degrees detects upright positions (sitting in bed or chair, or standing); note that “upright” indicates a true sitting position that is different from the standard 30- to 45-degree head-of-bed elevation typically defined as semi-recumbent.³³ For head elevation <50 degrees (recumbent positions in bed or Geri chair), the lateral roll-angle threshold was 20 degrees to the left or right side. For head elevation >50 degrees (seated positions), the tilt angle threshold was 10 degrees to the left or right side. To ensure fidelity, human observations were periodically used to verify the repositioning event.  

During 2 staff-assisted repositioning events for each resident, an investigator observed the event and applied an investigator-developed tool, broadly based on behavioral and psychological symptoms of dementia from the literature, to assess each resident’s responses to repositioning.^34–36 These observations were not conducted in the parent study. Because dementia progression varies among individuals and disease staging is not well-defined,³⁷ a spectrum of mild to severe behavioral psychological symptoms of dementia (BPSD) have been reported in the literature.^38–47 The investigator also observed nursing behaviors, focusing on alignment with evidence-based best practice management of BPSD.⁴⁸

Analytic strategy. Demographic and clinical data were retrieved from the parent study database and analyzed as univariate variables. Plots of position data obtained from the Leaf Patient Monitoring System³² as part of the parent study were extracted and developed as graphs across 28 days (Resident A) and 24 days (Resident B). Sensor data for Resident B were not available for days 21–26. We described the residents’ and nurses’ physical and verbal behaviors before, during, and after repositioning. 

Resident A. Resident A was a 67-year-old African American man diagnosed with vascular dementia (without behavioral disturbance), Alzheimer’s disease, and nonspecific other cognitive symptoms and of altered awareness. Resident A’s BIMS score of 5 indicated severe cognitive impairment. In addition, he was diagnosed with hypertension, an unspecified mood disorder, a contracture of the left hand, and dysphagia. His total Braden Scale score was 14, with subscale scores of 3 (Sensory), 2 (Moisture), 2 (Activity), 2 (Mobility), 3 (Nutrition), and 2 (Friction/Shear). A score of 2 in the activity subscale is chairfast, ie, the resident cannot bear his own weight and must be assisted into a chair or wheelchair. 

In the first repositioning event, Resident A turned to his right side from the back-lying recumbent position with considerable difficulty. Prior to the repositioning event, the nursing assistant tried to straighten his legs, which were flexed at the knee; however, straightening his legs resulted in Resident A immediately reverting back to bent knees and even more hip and knee contraction. The nursing assistant finally stood on the right side of the bed (from the head-of-bed position, the patient’s right side) and used her body weight in a heaving/lurching motion to turn Resident A to his right side; his entire body moved rigidly as 1 unit, with contracted arms and legs.

The second repositioning event involved the same nursing assistant moving Resident A to his left side from a recumbent position on his back. Resident A moved more flexibly when turned to the left, and his arms and legs contracted only minimally. During this second repositioning event, the nursing assistant was able to reposition Resident A in one fluid motion by using her arms only.

In both repositioning events, Resident A demonstrated what appeared to be repetitive and non-purposeful movements, ie, he rubbed his heels against the sheets in an up-and-down motion (dorsiflexion and plantar flexion) throughout the repositioning event and also moved his shoulders up and down rhythmically, causing his elbows to tap on the sheets. On his right side, Resident A continued the repetitive heel and shoulder motions even after the repositioning event was complete. On his left side, the repetitive motions stopped as soon as the repositioning event was completed. 

Resident A presented with a total of 28 days of monitored time (Table 2). Resident A was chairfast according to the Braden Scale activity subscale and spent the vast majority of his time (95%) tilted in bed or a chair <50 degrees, with only 4% upright time sitting in bed or a chair. The sensor data also exposed a distinct side preference, with only 10% of Resident A’s time spent on his right side.  

Resident B. Resident B was a 56-year-old African American woman. Although Resident B did not have a recorded diagnosis of dementia, her BIMS score was 6, indicating severe cognitive impairment. Her total Braden Scale score was 13, with subscale scores of 3 (Sensory), 2 (Moisture), 2 (Activity), 1 (Mobility), 3 (Nutrition), and 2 (Friction/Shear). Therefore, Resident B was also considered to be chairfast.

Two (2) repositioning events were observed 15 days apart. In both events, Resident B moved from the left-side position to the right-side position. Resident B’s bed was positioned with the left side of the bed (as viewed from the head of the bed) flush against the wall, with both side rails up. When asked about the rationale for bed positioning against the wall, the nursing assistant indicated that Resident B was both afraid of falling and considered a fall risk; the wall served as both a comfort and safety measure. Furthermore, a thick, heavy safety mat for fall injury prevention was placed on the floor and ran the length of the bed, which stabilized bed wheel movement during repositioning events. Given the limited nursing access to the resident for repositioning, the nursing assistant stood at the head of the bed, stooped over the resident’s head, reached for the turning sheet, and using a fluid pulling and twisting motion to create angular momentum, subsequently repositioned the resident onto another side. 

In the first repositioning event, the nursing assistant approached Resident B and spoke encouragingly. Resident B appeared anxious; her tone of voice was fearful, yet loud, and her speech was unintelligible with a worried facial expression. In a soothing tone, the nursing assistant explained her plan was to help reposition the resident and make her more comfortable. Resident B’s facial expression relaxed, and she appeared consoled by the nursing assistant’s explanation. During repositioning Resident B tensed up, with her limbs drawing toward her body, and then relaxed after the repositioning event was complete. 

During the second repositioning event, Resident B’s muscles tensed up with her limbs drawing closer to her body. Resident B grabbed the left handrail and appeared unable to let go; the nursing assistant gently unwound Resident B’s fingers in order to turn her. Grabbing the handrail did not appear purposeful. Instead, it appeared when Resident B’s fingers made contact with the rail at the beginning of the repositioning event; her arm was moving through the air, and she involuntarily grasped tightly. Resident B’s repositioning demonstrated subtle repetitive, non-purposeful movements. She moved her heels about 1 cm up and down (dorsiflexion and plantar flexion) throughout the repositioning event—a barely perceptible movement most visible after the repositioning was complete, when the sheet was replaced. 

Resident B had a total of 24 days of monitored time (Table 2). Resident B, also chairfast according to the Braden Scale activity subscale, spent 96% of her time reclining in bed or a chair (vs. 4% upright time) (Table 2). While reclining in bed or a chair, she was positioned on her back the majority of the time (44%), with similar amounts of time spent on each side (27% right recumbent vs. 25% left recumbent). 

In general, frail elderly NH residents with dementia have higher rates of functional and cognitive impairment, coinciding with difficulty in positioning the body to relieve pressure. Reduced activity, as measured by the Braden activity subscale, is also considered a strong predictor of PrI risk.⁴⁹ The 2 residents with dementia presented here were both chairfast according to the Braden Scale and had difficulty repositioning according to the nursing staff. This study is the first to use real-time data to detail everyday body positioning, thus enabling researchers to measure not only body orientations but also the durations of time spent in the upright versus recumbent, right, left, and back positions. Furthermore, the sensor data reflecting body tilt angle, position, and duration in real time broadly capture overall movement, and this has the potential to expand the evidence base related to cognitive and functional decline^50–52 directly influencing PrI risk.⁷ 

Many NH protocols encourage using the Braden Scale activity subscale to determine residents’ activity level and PrI risk. No studies to date have measured the duration of activity (sustained effort) in relation to chairfast status on the Braden Scale activity subscale. Both residents in the current case study had dementia, were considered to be chairfast (able to tolerate being upright in a chair), and were at a moderate risk for PrI development. When one considers being chairfast, it is assumed the individual can change and control body position in addition to being upright >50 degrees; however, data from the sensor system³² showed that the residents spent nearly all their time in the recumbent position (head-of-bed or chair elevation <50 degrees). In fact, <5% of time (about 1 hour/24) was spent sitting in the upright position (>50 degrees) for both of these residents. The finding that these 2 chairfast residents spent so little time upright was unexpected. It is unknown whether the residents in this case study are representative of the larger chairfast NH population (about 14% of all NH patients),³¹ and additional research is needed to determine the proportion of chairfast residents who spend the majority of their time with the back of their chair or bed lowered to <50 degrees. In the future, sensor data may be used to identify “sustained effort,” an essential concept because sustained effort (ie, endurance) is needed to build stamina and improve physical functioning.⁵³   

This case study is the first to objectively measure a side preference (Resident A), which is central to planning for PrI prevention care. For example, implications for nursing staff who care for a resident with a distinct side preference may include 1) more frequent skin inspections of the body surface that is in contact with the chair or bed, 2) a potential need for more frequent range of motion exercises, or 3) an environmental evaluation to determine whether bed placement is an issue (eg, the resident may prefer to face the television or window). In the future, periodically employing sensor data would enable nurses to identify residents’ side preferences and thereafter apply appropriate PrI prevention measures and/or environmental manipulation. Ideally, the next generation of monitoring devices would be sensitive enough to alert nursing staff of aberrant motor responses. 

Aberrant motor responses are common in the NH population.⁵⁴ These abnormal behaviors occur without a specific purpose and can be considered disruptive (eg, picking, bouncing, tapping, or pacing). Aberrant motor responses also include rapid repetitive movements such as those observed in both residents in the current case study with heels and shoulders repeatedly moving against surfaces. The risk with these movements is that the skin is exposed to friction and shearing forces, leaving the individual at risk of a PrI. Although the sensitive detection required for this is not yet available, future iterations may be able to detect rapid repetitive movements, enabling nurses to enact preventive interventions. If it is not possible to address the cause of the aberrant motor behavior sufficiently to reduce or eliminate the rapid repetitive movements, clinicians should consider measures such as applying protective dressings on bony prominences rubbing against the bed (eg, heels, elbows). Results from these 2 case studies demonstrate that further research is needed related to overall movement, body orientation, and position duration.

Wearable sensors are not a long-term solution for protecting those with dementia from PrI formation. However, they do provide a picture of overall body positions throughout the 24-hour day and therefore may be used to facilitate repositioning compliance.⁵⁵ In the future, contactless sensors may be developed and could be used for long-term body position monitoring to protect those with dementia from the deleterious effects of immobility. This feasibility study allowed us to measure objectively the amount of time residents who need assistance getting up to a chair actually spent sitting upright. Future research is needed to establish reliability within patients across time or between patients with the same BIMS score. Validity needs to be established in terms of clinical meaning of the positional readings in this population. A study of the predictive validity of each type of positional measurement would provide needed evidence of how PrIs develop. In the clinical practice arena, nursing staff could use sensor data to track the time residents spend in an upright position to refine progressive mobility protocols.  Sensor data may also be used to ensure that the mobility protocol is tailored to the individual resident’s unique movement profile.⁵⁶ Relative to side preferences, sensor data may be useful for tailoring PrI prevention interventions such as padding and more frequent skin inspection on the dependent side, as well as their relationship to a resident’s physical surroundings (eg, the resident prefers facing a window or television). 

Limitations of this case study series include the lack of generalizability. It is possible these 2 residents were atypical of the larger population with similar characteristics. The residents were chosen as a convenience sample and therefore selection bias is also possible. Furthermore, by observing the repositioning events, we may have unintentionally changed the resident or nurse behavior because people sometimes behave differently when they know they are being watched (Hawthorne effect). Finally, the sensor technology is limited to position detection and is not yet developed enough to detect smaller movements—either aberrant or productive—that may directly affect PrI risk.

This article is the first, to the authors’ knowledge, to describe the real-time body positions of NH residents LWD over multiple weeks. Results showed that 2 residents who were deemed chairfast according to the Braden Scale activity subscale were upright only <5% of the time, and that 1 resident had a distinct side preference. Real-time sensor data may be used in the future to measure durations of activity to build stamina and improve physical functioning as well as to reduce associated risk for PrI. Sensor data may also be used to elucidate side preferences, enabling nursing staff to initiate environmental manipulation (eg, changing the location of the television) if the side preference is related to room layout, or to apply padding and more frequent skin inspection to the dependent side. Overall, results from the current case study show that sensor data may be a useful tool for developing tailored PrI interventions.

The authors thank Judith C. Hays, PhD, who assisted in writing, preparing, and critically reviewing the manuscript.

Research reported in this publication was supported by the National Institute of Nursing Research of the National Institutes of Health under Award Number R01NR016001; ClinicalTrials.gov: NCT02996331. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health [Recipient: Tracey L. Yap, PhD, RN, CNE, WCC, FGSA, FAAN].

Dr. Yap is an associate professor, Duke University School of Nursing, Durham, NC; Dr. Alderden is an assistant professor, University of Utah College of Nursing, Salt Lake City, UT; Dr. Sabol is a professor and division chair, Healthcare in Adult Populations, Duke University, Durham, NC; Dr. Horn is an adjunct professor, University of Utah School of Medicine, Salt Lake City, UT; Dr. Kennerly is a professor, East Carolina University College of Nursing, Greenville, NC. Please direct correspondence to Dr. Tracey L. Yap, PhD, RN, CNE, WCC, FGSA, FAAN, Duke University School of Nursing, 307 Trent Drive, Durham, NC, 27710; email: tracey.yap@duke.edu . TLY, JA, SMK, and VKS wrote the manuscript in consultation with SH, who reviewed the document and assisted in its refinement. All authors read and approved the final manuscript.