This cross-sectional study was conducted between February 2020 and April 2021 in a 16-bed recovery rehabilitation ward attached to a secondary emergency hospital in Tokyo, Japan, which consisted of seven inpatient departments, 18 outpatient departments, and five 180-bed wards.

The applicable diseases, services, and duration of stay for convalescent rehabilitation wards are based on the medical fee determined by the government. The convalescent rehabilitation fee in the medical fee system ranges from 1 to 6, depending on the facility’s standards. Differences in facility standards include the number of staff, availability of rehabilitation services on holidays, percentage of severe disease patients, and achievement quotas. Hospitalization fee-1 has the highest allocation of human resources and the highest achievement quota. The facility standard in this study was Hospitalization fee-5.

The facilities also provide treatment by specialists in each department for each of the complex diseases during hospitalization in the convalescent rehabilitation wards. For example, respiratory disease is treated by respiratory specialists, cardiovascular diseases by cardiology specialists, and gastrointestinal diseases by gastroenterology specialists, while liver, renal, allergic, and metabolic diseases including diabetes are treated by general internal medicine specialists, and mental disorders by psychosomatic medicine and clinical psychologists, forming a team for each patient. Pain is treated immediately after admission and continuously with oral and transdermal analgesics depending on the pain level and chronic disease status by an orthopedic surgeon and/or internist. Patients with severe conditions requiring treatment, such as patients with respiratory, hepatic, or renal disorders and patients with New York Heart Association (NYHA) functional classification class-III or higher [17] are treated in the acute care unit, not transferred to the convalescent rehabilitation ward. The heart failure patients included in this study were those who were diagnosed with heart failure in other departments and transferred to the convalescent rehabilitation unit, and due to the conditions of the convalescent rehabilitation hospital, NYHA class-II patients were of the majority. Patients were assessed by the attending physician and those with stable heart failure symptoms and able to tolerate practicing ADL were allowed to be transferred to the convalescent rehabilitation ward for the primary purpose of practicing ADL, even if their subjective symptoms were categorized as NYHA class-III. Patients received personalized one-on-one rehabilitation with a therapist for 1–2 hours per day during the rehabilitation hours. Personalized rehabilitation includes basic action practice, strength training, ADL, gait practice, and endurance exercise. In the presence of many comorbidities, the guidelines followed the general principle of exercise prescription, which states that “a combined exercise regimen consisting of aerobic exercise, resistance training, balance training, and flexibility exercises should be considered to improve exercise capacity and physical function [18],” under the judgment and direction of the attending physician.

This facility surveyed physical activity in patients aged 75 years and older who were hospitalized for musculoskeletal diseases and examined the data on physical activity at discharge.

The participants were subacute musculoskeletal patients admitted to the recovery rehabilitation ward. Subacute phase in this study is defined as the period from the day of transfer to the rehabilitation ward to the day of discharge. The patient’s transfer was permitted by the attending physician based on a combination of factors, including removal of the indwelling bladder catheter, no signs of wound infection, and the ability to take the required amount of food orally. Musculoskeletal disease is defined as the applicable conditions for admission to the convalescent rehabilitation ward, as follows: fractures of the femur, pelvis, vertebra, hip joints, knee joint, and hip joint replacement or knee joint replacement. The inclusion criteria for participants were as follows: (1) ≥75 years of age; (2) those who provided consent to participate in the study; and (3) a musculoskeletal disease applicable for admission to the convalescent rehabilitation ward. The exclusion criteria were as follows: (1) an admission diagnosis of a condition other than musculoskeletal disease; (2) comorbid disorders, long-term fever symptoms, and limitations in therapeutic or medical activities due to lower limb load restriction; (3) a mental state that would impair decision-making regarding participation in the study, including delirium; and (4) unstable mental or physical conditions and a high possibility of a sudden change in short-term physical conditions, judged by the physician upon admission into the ward.

The required sample size was set at 80 participants using G*power (https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower) as follows: Wilcoxon–Mann–Whitney test (two groups), two tails, effect size 0.5, α err prob 0.05, power (1-β err prob) 0.5, and an allocation ratio of 2/1.

This study was approved by the Institutional Review Board of The Anti-Tuberculosis Association, Shin-Yamanote Hospital (No. 19001) and by the research ethics committee of the Tokyo Metropolitan University, Arakawa Campus (No. 20038). Participants were informed of the content and purpose of the study verbally and in writing, and they provided written informed consent.

Potential confounders included age, sex, body mass index (BMI), length of hospital stay, ability to perform ADL, type of musculoskeletal disease diagnosed on admission, medical history, and mobility. Predictors included the presence of a history of heart failure. Outcomes were bout duration and the total time of physical activity classified by activity intensity. For patients with heart failure, the data on cardiac function variables were collected as patient characteristics.

Age, sex, BMI, length of hospital stay, the ability to perform ADL, musculoskeletal disease diagnoses, cardiac function variables, medical history, and mobility at discharge were collected from patient medical records. BMI was calculated by dividing the weight (kg) by height² (m²). To assess the ability to perform ADL, the Functional Independent Measure (FIM) was selected as the evaluation index. FIM compares movement ability objectively independent of the patient’s use of assistive devices, such as wheelchairs, walkers, canes, and railings, or communication skills, such as hearing, vision, verbal comprehension, and non-verbal comprehension. FIM quantifies the amount of ADL assistance provided to the patient and consists of 13 motor and five cognitive items on a 7-point scale ranging from 7 to 1 each, for a total score of 126 to 18 [19,20]. The data on FIM scores were collected from the medical records of entry and discharge conferences and were assessed by the ward nurses. “FIM at entry” was assessed within 24 hours of entry to the ward, and “FIM at discharge” was assessed the day before discharge. Preadmission indoor mobility was examined after admission based on the Living Space Assessment (LSA) [21]. The patients’ level of independence according to the LSA level-3 categories “been to a neighborhood other than their yard or apartment during the past four weeks,” and “needs help from others” were classified in our study as “personal assistance,” while “equipment only,” and “no equipment or personal assistance” were classified as “unassisted.” For indoor mobility at the time of discharge, locomotion was assessed under the FIM motor domain and “walking” or “wheelchair” scores were collected for assessment. Walking with a cane or walker was classified as “walking,” and those who were unable to walk 50 m (FIM walking score≤4) were classified as “wheelchair users” regardless of whether they were independent wheelchair users. The musculoskeletal diseases that caused hospitalization were identified and recorded upon admission to the convalescent rehabilitation wards. A medical history of diabetes was included as patient characteristic variable because its incidence correlates inversely with the amount of physical activity [22]. The diagnosis of diabetes mellitus was based on hemoglobin A1c levels, ongoing treatment, and diagnostic history in the medical record. History of cerebrovascular disease was included as a patient characteristic variable because its sequelae are considered to increase the risk for reduced physical activity and ADL impairment [23]. The diagnosis of cerebrovascular disease was based on the medical history of previous stroke and cerebral hemorrhage, regardless of the severity of sequelae or the regular prescriptions. All study participants with a history of cerebrovascular disease were in the living phase of the disease, more than 6 months after disease onset.

Heart failure in this study was defined as having a diagnosis of heart failure, being on cardiovascular medication, and subjective symptoms with NYHA functional class-II or higher [17] prior to the onset of musculoskeletal disease: NYHA class-I, no symptoms with normal physical activity and function status; NYHA class-II, mild symptoms with normal physical activity, comfortable at rest, slight limitation of functional status, clinically walking more than two blocks on the level and climbing more than one flight of ordinary stairs, and performing to completion of any activity requiring 5 metabolic equivalents (METs); NYHA class-III, moderate symptoms with less than normal physical activity, comfortable only at rest, marked limitation of functional status, clinically walking one to two blocks on the level and climbing one flight of stairs, and patient can perform to completion of any activity requiring >2 METs; and NYHA class-IV, severe symptoms with features of heart failure with minimal physical activity and even at rest, severe limitation of functional status [17,24]. The research facility is a specialized rehabilitation ward for musculoskeletal diseases, and as a rule, does not accept patients in NYHA class-III or higher.

Participants with the diagnosis of heart failure among patients with musculoskeletal disease were assigned to “with heart failure” group. Participants who were never diagnosed with heart failure, regardless of how many diseases they had, were assigned to the “without heart failure” group. Cardiac function indices of participants with heart failure were collected for presenting participant characteristics, as follows: brain natriuretic peptide (BNP) (pg/mL) which reflect worsening or improving hemodynamics, left ventricular ejection fraction (LVEF) (%), E velocity divided by A-wave velocity (E/A ratio), average E velocity divided by mitral annular e’ velocity (average E/e’), septal e’ velocity (cm/s), and tricuspid regurgitation (TR) velocity (m/s) as an echocardiographic parameter when investigating the cause of heart failure [25]. BNP test and echocardiography are not part of the usual practice for musculoskeletal disease. They are performed only when risk screening for cardiac function is appropriate at the start of musculoskeletal disease treatment. Subsequent echocardiographic testing was not performed. The conditions of participants with heart failure are assessed by BNP levels and clinical findings as needed.

A total of 84 participants were included in the study, including 25 with heart failure. No one failed to finish the study due to exacerbation of heart failure during the study period. The number of participants in each phase of the study is shown in the flowchart in Fig. 1. The descriptive statistics for all participants were as follows: age, 85 years (80.8–89.0 years) (median [interquartile range, 25th percentile–75th percentile]); female patients, 63 (75.0%), BMI, 20.5 (18.3–24.0) kg/m²; duration of stay in the recovery rehabilitation ward, 30.5 (20.0–50.3) days; FIM at entry, 85.5 (72.0–96.0) points; FIM at discharge, 108.0 (96.5–117.0) points; number of patients with postoperative femur fracture, 38 (45.2%); number of patients under treatment for diabetes mellitus, 27 (32.1%); number of patients with a history of cerebrovascular disease, 12 (14.3%); number of patients with independent indoor mobility prior to admission due to musculoskeletal disease, 74 (88.1%); and number of patients discharged ambulatory, 78 (91.7%; Table 1).

No differences were found for all variables of participant characteristics between the with heart failure and without heart failure groups: age, sex, BMI, days from onset of musculoskeletal disease to ward transfer, length of ward stay, FIM at entry, type of orthopedic disease, diabetes mellitus, cerebrovascular disease, and mobility at preadmission and discharge (p=0.44, p=0.28, p=0.47, p=0.16, p=0.29, p=0.08, p=0.10, p=0.32, p=0.77, p=0.14, and p=0.11, respectively; Table 1). Similarly, no difference was found in FIM at discharge between the with heart failure group (108.0 [86.0–113.0] points) and without heart failure group (109.0 [97.5–117.0] points; p=0.28; Table 1). There was no statistical significance in the results of multivariate analysis of the difference in FIM score change between entry and discharge with and without heart failure (two-way ANOVA, p=0.40; Table 1). Risk factors affecting physical activities were age and FIM at discharge (Table 2). Clinical data on cardiac functions for participants with heart failure BNP decreased from 124.5 (72.6–221.8) pg/mL at admission to 76.4 (42.7–182.5) pg/mL at discharge, but not significantly (p=0.11). LVEF was 67.0% (60.0%–72.0%; Table 3). Fourteen of the 25 patients had diastolic dysfunction and three had systolic and diastolic dysfunction. Other patients with heart failure had atrial fibrillation and tachyarrhythmia and were on drug therapy. Patients underwent outpatient treatment at home and had stable cardiac function. The study defined heart failure by NYHA classification based on subjective symptoms and drug therapy, the NYHA classification II and III and comprised 20 (80.0%) and 5 (20.0%) patients, respectively. Five participants of NYHA class-III were diagnosed with occasional complaints of palpitations and fatigue, but usually had only mild subjective symptoms.

Descriptive statistics for all physical activities were 180 (130–220) seconds for a single continuous period of light-intensity physical activity or higher, 15 (10–20) seconds of moderate-to-vigorous-intensity physical activity or higher, 14,913.4 (10,569.2–19,502.5) seconds total time of light-intensity physical activity or higher, and 633.4 (354.6–927.5) seconds total time of moderate-to-vigorous-intensity physical activity or higher (Table 4). Descriptive statistics for physical activity in the group with heart failure were 130 (120–200) seconds for a single continuous period of light-intensity physical activity or higher, 10 (10–20) seconds for moderate-to-vigorous-intensity physical activity or higher, 11,296.6 (8,603.3–16,186.7) seconds total time of light-intensity physical activity or higher, and 383.3 (296.7–613.3) seconds total time of moderate-to-vigorous-intensity physical activity or higher (Table 4). Descriptive statistics for physical activity in the without heart failure group were 190 (150–220) seconds for a single continuous period of light-intensity physical activity or higher, 20 (10–20) seconds of moderate-to-vigorous-intensity physical activity or higher, 15,273.3 (12,270.0–20,546.7) seconds total time of light-intensity physical activity or higher, and 730 (538.4–1,071.7) seconds total time of moderate-to-vigorous-intensity physical activity or higher (Table 4).

Between the with heart failure and without heart failure groups, differences were found in single continuous period of light-intensity physical activity or higher (p=0.004), moderate-to-vigorous-intensity physical activity or higher (p=0.019), total activity time of light-intensity physical activity or higher (p=0.005), and moderate-to-vigorous intensity physical activity or higher (p<0.001). The heart failure group had a shorter duration of physical activity for all variables in univariate analysis (Table 4). Age and FIM were selected as risk factors of physical activity according to the results of multiple regression analysis (Table 3). The single continuous period for light-intensity physical activity or higher was shorter in the heart failure group in the multivariate analysis with age and FIM as covariates (Table 4).

We used the longest continuous time per hour with an epoch length of a 10-s bout as the single continuous period in the analysis; therefore, the results may differ depending on how the single continuous period is defined. The accuracy of the accelerometer may vary with activity outcomes such as step, type of activity, limb position, and in populations with low mobility. The validity and reliability of non-free-living activities, such as exercises during rehabilitation period, remains unclear. Therefore, researchers should be cautious when applying the activity monitor to new populations or activities where the accuracy of the device has not been specifically tested. One selection bias of this study was that the facility conducting this research was situated within a general hospital. At the time of a scheduled admission or an emergency admission, patients with complex diseases other than musculoskeletal disease may intentionally select a hospital that can provide more comprehensive care. This may have induced bias in participants’ medical history in this study. As for the other medical history (such as chronic obstructive pulmonary disease, liver disease, renal disorder, dementia, and joint disease), it was assumed that a certain number of patients had been diagnosed or had completed or discontinued treatment at other hospitals in the past or had not undergone checkups and had not been diagnosed. Our medical records did not cover this disease entirely; these missing data may have provided the information bias and were not included in the medical history variable. But these diseases are considered to be an urgent issue to be investigated. For physical activity in older adults, environmental factors should also be considered, such as area of residence, facility equipment, family members living together, and their caregiving ability, in addition to individual factors, such as motivation and life history.