Mobility matters: a protocol to improve mobility and reduce length of stay in hospitalised older adults

Abstract

Functional decline in hospitalised older adults aged 65 and above is a significant clinical problem. Despite its adverse outcomes, the problem of not mobilising older adult patients in clinical settings remains. Existing evidence suggests that a mobility protocol can be effective in addressing this concern. The aims of this quality improvement project were to determine whether a nurse-driven, multidisciplinary collaborative mobility protocol would increase the daily out-of-bed episodes, improve mobility level and reduce hospital length of stay (LOS). A nurse-driven mobility protocol was implemented in three phases. This five-component protocol included mobility assessment using the Johns Hopkins Highest Level of Mobility (JH-HLM) scale, documentation of mobility score, implementation of out-of-bed activities three times per day, communication of mobility score and goal at daily huddle, and indication of mobility score and goal on the board at patient’s bed. Data were collected before and after the implementation. 142 patients were recruited from an acute geriatric unit. There were 72 patients from the pre-implementation group and 70 patients from the post implementation group. Comparing the pre-implementation and post implementation groups, the mean out-of-bed episodes per patient day increased from 0.80 to 3.59 (p<0.001). JH-HLM scores at discharge with ambulation status increased from 51.4% to 71.4% (p<0.001). Patients had improved JH-HLM scores with a median 2.00 (B 2.00, 95% CI 1.35 to 2.65, p<0.001) higher at discharge in the post implementation group after adjusting for score at admission. Increased mobility did not lead to any fall incidents. The mean hospital LOS was reduced from 15.67 (SD 11.30) days to 13.07 (SD 7.18) days (p=0.069). In conclusion, the implementation of a nurse-driven mobility protocol resulted in increased frequency of out-of-bed episodes and improved mobility, and reduction in LOS.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Previous studies have consistently shown that hospitalised older adults experience significant declines in mobility, leading to increased length of stay (LOS), functional decline and healthcare costs. Despite the importance of mobility promotion, it lacks standardised protocols that integrate nurses, as part of a multidisciplinary team, to work alongside physiotherapists in promoting mobility locally.

WHAT THIS STUDY ADDS

• Our study demonstrates the effectiveness of a mobility protocol in improving mobility and reducing LOS in hospitalised older adults. The protocol, which includes standardised mobility assessments, goal-setting and progressive mobility interventions, was found to be feasible and acceptable to nursing staff.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings have important implications for practice, as they highlight the critical role of nurses in promoting mobility in hospitalised older adults. The mobility protocol developed in this study can be adapted and implemented in various healthcare settings, potentially improving outcomes for older adults. Future research should focus on exploring the long-term effects of mobility protocol on patient outcomes.

Introduction

Problem description

Functional decline in hospitalised older adults aged 65 and above is a significant clinical problem. It is associated with falls, frailty, difficult transition from hospital to home and increased mortality.¹ Functional decline, defined as the inability to perform independent self-care activities or worsening in self-care skills, may develop in one-third of hospitalised older adults.^{2 3} In addition to physiological changes of ageing, there are many causes of functional decline among hospitalised older adults, leading to loss of activities of daily living function despite treatment of acute illnesses.⁴ Prolonged bedrest or immobilisation is a major factor with 16%–33% hospitalised older adults reported to have reduced mobility levels, performing very little or no physical activities^{5 6} and spending 83% of their time lying in bed for those who are able to ambulate independently.⁷

The evidence suggests that nursing interventions such as integrating patient mobility as part of the care process^{8 9} may ameliorate risk factors leading to immobilisation. In addition, mobility interventions to avoid decline can improve mobility levels and reduce hospital length of stay (LOS) for older adults.⁸ However, the practice of promoting physical mobility is lacking and inconsistent among nurses working with older adults across all disciplines. Currently, there is neither a standardised nursing guideline on mobility assessment and intervention in daily practice nor an integration of mobility needs of the older adults as part of the care goals. Hence, there is a need for a nurse-driven standardised mobility protocol to promote mobility and reduce the development of functional decline for hospitalised older adults.

Available knowledge and rationale

In this project, mobility is defined as a patient getting out of bed, including mobility events such as ambulating, standing and sitting out of bed.⁸ Previous studies on the effectiveness of mobility interventions in this cohort have either focused solely on ambulating the patient,10,13 or on multi-component mobility protocol which includes assessing mobility, setting mobility goals and getting patients out of bed.^{1 8 14 15} The studies focusing on ambulation programme reported improvements in ambulation frequency,^{10 12} walking distance,^{10 12} reduced hospital LOS¹¹ and reduced functional decline.¹³ Similar improvements were also found in the multi-component mobility protocol studies.^{1 8 14 15}

Validated mobility assessment tools designed for nurses to lead and which can be adopted for use in acute hospitals include the de Morton Mobility Index,¹⁶ Banner Mobility Assessment Tools,¹⁷ Activity Measure for Post-Acute Care 6-Clicks¹⁸ and Johns Hopkins Highest Level of Mobility (JH-HLM) scale.¹⁹ Among these, the JH-HLM scale is particularly favourable for adoption due to its excellent inter-rater reliability and test-retest reliability to be used by nurses,¹⁹ and its minimal training requirements and ease of use.²⁰ Evidence suggests that such a multi-component mobility protocol is best driven by nurses in collaboration with other healthcare professionals,^{1 14} as nurses are already involved in the daily care of the patients and hence suited to perform mobility assessment and administer mobility interventions.²⁰

Aims

The purpose of this quality improvement (QI) project was to promote mobility and prevent functional decline in hospitalised older adults aged 65 and above using a nurse-driven mobility protocol. The project’s aims were to increase the daily out-of-bed episodes, to improve or maintain the mobility level and to reduce hospital LOS in this cohort of patients.

Methods

Context and sample

The project centre is a tertiary hospital located in the eastern region of Singapore with a 1000-bed capacity. The study site was a geriatric unit with a capacity of 32 beds for patients aged 65 and above under the care of a multidisciplinary team. The geriatric unit is designed to have wider corridors, terraces, dining areas and therapy gardens, features which are conducive for promotion of mobilisation. The Iowa model²¹ of evidence-based practice was used to guide the implementation. A mobility workgroup was formed consisting of an advanced practice nurse as project lead, geriatricians, physiotherapists (PTs) and nursing mobility champions, which allows the team to enhance multidisciplinary collaboration and leverage the expertise from geriatricians and PTs to standardise the nursing mobility practice. Patients were recruited in both pre-implementation and post implementation phases in the designated geriatric unit from September 2021 to January 2022. Patients were excluded if they fulfilled any of the following criteria: (a) being haemodynamically unstable with blood pressure below 90/50 mm Hg or heart rate above 100 beats/min¹⁴; (b) being out of the unit for a procedure for more than 4 hours; (c) being medically unstable as assessed by the geriatrician in charge; (d) being bedbound with JH-HLM score of 1 at premorbid mobility level.

Interventions

Based on the literature,^{1 8 14 15 22} the project team designed a nurse-driven mobility protocol comprising five components: (a) introduction of the JH-HLM scale for daily mobility assessment, (b) documentation of JH-HLM scores on the mobility chart two times per day, (c) implementation of out-of-bed activities three times per day based on the identified mobility level, (d) communication of JH-HLM score and goal during daily multidisciplinary huddle and (e) indication of JH-HLM score and goal on the board at the head of patient’s bed for information sharing between nurses and the multidisciplinary team.

Outcome measures and data collection

This QI project was a prospective study of patients in pre-implementation group without mobility protocol and post implementation group with mobility protocol. There were three phases in total. Phase I (pre-implementation phase) was to establish the baseline occurrence of daily out-of-bed episodes, mobility level and LOS. Phase II was a 10-day training phase which involved teaching the nurses and healthcare team on the use of the JH-HLM scale for mobility assessment and the mobility protocol. Prior to the training and post implementation of the mobility protocol, nurses’ perception was surveyed in five statements with five response options (‘strongly agree’, ‘agree’, ‘neutral’, ‘disagree’ or ‘strongly disagree’). Phase III (post implementation phase) included the implementation of the mobility protocol and assessment of daily out-of-bed episodes, mobility level and LOS after implementation.

First, the daily out-of-bed episodes were recorded using a self-designed mobility chart at phase I and prospective documentation review of mobility chart at phase III. Second, the measurement of the mobility level was obtained using the JH-HLM score on admission and on discharge through observation at phase I and prospective chart review of the JH-HLM documentation at phase III. The permission to use the JH-HLM scale for mobility assessment was obtained.¹⁹ The JH-HLM is an 8-point ordinal scale that represents mobility level, where 1=lying, 2=turn self/activity, 3=sit at edge of bed, 4=transfer to chair/commode, 5=standing for 1 min, 6=ambulating 10+ steps, 7=ambulating 7.5 m and 8=ambulating 75 m.⁸ With this scale, nurses rated patients daily on their highest level of mobility⁸ and JH-HLM scores were collected to assess the improvement of mobility levels for patients at admission versus discharge for both pre-implementation and post implementation groups.

Statistical analysis

Descriptive statistics of demographic variables and outcomes were reported as numbers and percentages for nominal data, mean±SD for continuous data or median and IQR for ordinal data. To determine if there were differences in outcomes between patients in the pre-implementation and post implementation groups, the mean out-of-bed episodes per patient day and LOS were modelled using linear and negative binomial regression, respectively. JH-HLM scores at discharge and ambulating level of mobility at discharge were modelled using quantile and negative binomial regression, respectively, with JH-HLM score at admission as a covariate.

To examine if the differences in outcomes were independent of age, gender, ethnicity and living status, the project team adjusted for these variables in multivariable regression models. Box-Tidwell tests assessed the linearity assumption of linear regression. Robust SEs accounted for heteroscedasticity in linear regression models. Quantile-quantile (q-q) plots of model residuals against the expected order statistics of the standard normal distribution assessed the plausibility of the normality of errors assumption for linear regression. Where there was overdispersion in the count data, negative binomial regression was used. There was no missing data. Unstandardised beta coefficients, ORs, incidence rate ratios (IRRs) and the corresponding 95% CIs were reported. Statistical tests were two-sided with a 0.05 significance level. All statistical analyses were conducted using Stata V.15.0 (College Station, Texas: StataCorp).

Results

Participant characteristics

A total of 142 patients from the geriatric unit were included in this project based on the inclusion and exclusion criteria. All recruited patients remained throughout the implementation period with all daily mobility assessments completed. There were 72 patients included in the pre-implementation group and 70 patients in the post implementation group. The median age of the patients was 85 (81–90) years, with 28.2% male and 71.8% female. Patients’ characteristics are summarised in table 1. There was no fall incident reported in the recruited patients.

Table 1. Demographic characteristics of 142 patients.

Characteristic	Pre-implementation(n=72)	Post implementation(n=70)
	Patients (n) or median (IQR)	Patients (n) or median (IQR)
Median age at admission in years (IQR)	84 (81–89.5)	86 (81–90)
Gender, n (%)
Male	29 (40.3)	11 (15.7)
Female	43 (59.7)	59 (84.3)
Ethnicity, n (%)
Chinese	56 (77.8)	58 (82.9)
Malay	10 (13.9)	10 (14.3)
Indian	3 (4.2)	1 (1.4)
Caucasian	2 (4.2)	1 (1.4)
Living status, n (%)
Lives alone	5 (6.9)	3 (4.3)
Lives with partner/spouse	10 (13.9)	4 (5.7)
Lives with partner/spouse  and children	12 (16.7)	27 (38.6)
Lives with children	44 (61.1)	36 (51.4)
Nursing home resident	1 (1.4)	0 (0.0)
Inpatient falls, n (%)	0 (0.0)	0 (0.0)

Out-of-bed episodes per patient day

The mean out-of-bed episodes per patient day in the pre-implementation and post implementation groups were 0.80 (SD 0.24) and 3.59 (SD 1.34), respectively. Simple linear regression showed a mean 2.80 (B 2.80, 95% CI 2.48 to 3.12, p<0.001) higher number of out-of-bed episodes per patient day in the post implementation group compared with the pre-implementation group (table 2). This was consistent with quantile regression, which showed a median 2.65 (B 2.17 to 3.14, p<0.001) higher number of out-of-bed episodes per patient day in the post implementation group.

Table 2. Differences in out-of-bed episodes, mobility score and hospital length of stay between the pre-implementation and post implementation group in 142 patients.

Outcome	Pre-implementation (n=72)	Post implementation(n=70)	B coefficient or OR or incidence rate ratio (95% CI)	P value
	Mean±SD or median (IQR) or patients (n) (%)	Mean±SD or median (IQR) or patients (n) (%)
Mean out-of-bed episodes per patient day, mean±SD	0.80±0.24	3.59±1.34	2.80 (2.48 to 3.12)	<0.001
JH-HLM scale (1–8)
Score at discharge, median (IQR)	6 (4–7)	7 (5–7)	2.00 (1.35 to 2.65)	<0.001
Ambulating (≥6) at discharge, n (%)	37 (51.4)	50 (71.4)	19.97 (7.19 to 55.50)	<0.001
Hospital LOS in days, mean±SD	15.67±11.30	13.07±7.18	0.83 (0.69 to 1.01)	0.069

Mean out-of-bed episodes, JH-HLM scores at discharge, ambulating mobility level at discharge and hospital LOS were modelled using linear, quantile, logistic and negative binomial regression, respectively. Coefficients for JH-HLM score and ambulating mobility level at discharge were adjusted for JH-HLM score at admission.

JH-HLMJohns Hopkins Highest Level of MobilityLOSlength of stay

JH-HLM score at discharge

The median (IQR) JH-HLM scores at discharge were 6 (4–7) and 7 (5–7) in the pre-implementation and post implementation groups, respectively, while the median change scores (IQR) at admission versus discharge were −1 (−2 to 0) and 1 (0 to 2), respectively. Quantile regression showed a median 2.00 (B 2.00, 95% CI 1.35 to 2.65, p<0.001) higher JH-HLM score at discharge in the post implementation group after adjusting for JH-HLM score at admission (table 2). After adjusting for age, gender, ethnicity and living status, multivariable quantile regression showed a median 1.65 (B 1.65, 95% CI 1.10 to 2.21, p<0.001) higher JH-HLM score at discharge in the post implementation group (table 3).

Table 3. Multivariable regression of out-of-bed episodes, mobility score and hospital length of stay in 142 patients.

Outcome	Adjusted B coefficient or OR or incidence rate ratio (95% CI)	P value
Mean out-of-bed episodes per patient day	2.68 (2.37 to 2.99)	<0.001
JH-HLM scale (1–8)
Score at discharge	1.65 (1.10 to 2.21)	<0.001
Ambulating (≥6) at discharge	14.51 (4.44 to 47.38)	<0.001
Hospital LOS in days	0.93 (0.76 to 1.14)	0.472

Mean out-of-bed episodes, JH-HLM scores at discharge, ambulating mobility level at discharge and hospital LOS were modelled using linear, quantile, logistic and negative binomial regression, respectively. All models adjusted for age, gender, ethnicity and living status. Coefficients for JH-HLM score and ambulating mobility level at discharge were additionally adjusted for JH-HLM score at admission.

JH-HLMJohns Hopkins Highest Level of MobilityLOSlength of stay

Ambulating mobility level (JH-HLM≥6) at discharge

Table 4 presents JH-HLM mobility scores in both pre-implementation and post implementation groups. There were 50 (71.4%) out of 70 patients in the post implementation group and 37 (51.4%) of 72 patients in the pre-implementation group who had an ambulating mobility level (JH-HLM≥6) at discharge. Logistic regression showed that patients in the post implementation group had 19.97 (OR 19.97, 95% CI 7.19 to 55.50, p<0.001) times the odds of an ambulating mobility level at discharge compared with patients in the pre-implementation group after adjusting for JH-HLM score at admission (table 2). After additionally adjusting for age, gender, ethnicity and living status, which are potential confounders, multivariable logistic regression showed that patients in the post implementation group had 14.51 (OR 14.51, 95% CI 4.44 to 47.38, p<0.001) times the odds of an ambulating mobility level at discharge (table 3).

Table 4. Mobility score at discharge during the pre-implementation and post implementation phase.

JH-HLM category	Pre-implementation (n=72)	Post implementation (n=70)
	Patients (n) (%)	Patients (n) (%)
Walk (JH-HLM≥6)	37 (51.4)	50 (71.4)
Stand (JH-HLM=5)	7 (9.7)	9 (12.9)
Chair (JH-HLM=4)	23 (31.9)	11 (15.7)
Bed (JH-HLM≤3)	5 (6.9)	0 (0.0)

JH-HLMJohns Hopkins Highest Level of Mobility

Hospital LOS

The mean hospital LOS was 13.07 (SD 7.18) days in the post implementation group and 15.67 (SD 11.30) days in the pre-implementation group. As there was evidence of overdispersion, negative binomial regression was used to model hospital LOS. Negative binomial regression did not show evidence of a difference in incidence rate of LOS between the two groups (IRR 0.83, 95% CI 0.69 to 1.01, p=0.069) (table 2). After adjusting for age, gender, ethnicity and living status, multivariable negative binomial regression did not show evidence of a difference in incidence rate of LOS between the two groups (IRR 0.93, 95% CI 0.76 to 1.14, p=0.472) (table 3).

Nurses’ perception

A total of 33 nurses were surveyed. The survey findings (figure 1) demonstrated an improvement in the nurses’ impression in terms of promoting mobilisation will be more work for nurses and increased mobilisation of patients will increase the fall risks. Our interventions led to 91% of the nurses feeling more confident in mobilising the patients (figure 1). It also led to enhanced documentation of mobility activities in 94% of the nurses.

Figure 1. Survey responses from nurses before and after implementation.

Feedback from patients and staffs

The results show that 92.9% of patients and 90% of staff were satisfied with the implementation of mobility protocol. Most patients (84.3%) felt that mobility protocol was helpful in their recovery.

Discussion

Statement of principal findings

The results showed that the mean out-of-bed episodes per patient day for older adults were increased after the implementation of a nurse-driven mobility protocol. In addition, ambulation episodes also increased significantly. Furthermore, the findings revealed that the median JH-HLM score at discharge was higher and the median change in JH-HLM score between hospital admission and discharge was improved after implementation. These improvements were statistically significant and positively impacted the mobility level of older adults without leading to any fall incidents. The reduction in the mean of LOS in the post implementation group was statistically insignificant.

Interpretation within the context of the wider literature

Our findings are in keeping with previous studies which reported that mobility protocols were effective in achieving increased ambulation frequency, enhanced out-of-bed episodes, improved mobility level and decreased hospital LOS in older adults.^{8 15 22 23} Several factors may have contributed to these positive outcomes, such as mobility assessment, goal setting, daily review and documentation, and communication. The project team used a structured implementation model to apply these strategies in the daily nursing care. Our results showed that the hospital LOS was reduced after the implementation of mobility protocol. Similar results were reported in prior studies.^{1 8 14 15} Although it was statistically insignificant, it has demonstrated its potential impact on patient outcomes. The project team postulates two factors that contributed to a reduced LOS. First, the out-of-bed episodes were increased significantly after implementation, which might have reduced the complications associated with immobility, as shown in previous studies.^{24 25} Second, the project might have had a positive impact on the process of care coordination which was depicted in prior research.⁸ Importantly, this protocol incorporated mobility goal setting aligned with the identified JH-HLM mobility level into daily multidisciplinary huddle, which expedited the discharge planning process and determined appropriate disposition, which was reported in a prior study.^{8 14} In this project, there was an indication of better social support network provided by family as 90% of patients from the post implementation group lived with children and/or their spouses or partners. However, it is unlikely the factor for reduced LOS as multivariable negative binomial regression did not show evidence of a difference in incidence rate of LOS after adjusting for living status.

We have also addressed barriers such as perception of increased workload and increased risk for falls when promoting mobility,^{26 27} by providing nurses with training sessions prior to the implementation increased their understanding and acceptance of the mobility protocol. It was also important to ensure adequate chairs were made available at the bedside for patients. These key strategies greatly enabled the mobility protocol to blend into the nursing routine and facilitate the out-of-bed activities by nurses. As such, getting patients out of bed became a less burdensome task for the nurses. Additionally, the nursing mobility champions were actively engaged daily in the ward to walk the talk and conduct a timely audit.

In addition, the use of the JH-HLM scale provided a standardised language to communicate the mobility level among various members of the team. Furthermore, it documented individual patient mobility levels and tracked its improvements during hospital stay. Notably, the project team found that no fall incidents were reported during the implementation which is consistent with prior studies.^{8 14} This is important as it provides reassurance to the nurses that promoting mobility does not necessarily lead to falls.

Strengths and limitations

Our findings demonstrate that promoting mobility through a standardised nurse-driven mobility protocol with multidisciplinary collaboration improved patient outcomes, especially the frequency of out-of-bed activities and individual patient mobility. Previous studies^{14 23 28} suggested that comprehensive mobility protocols gave the greatest benefit. With the improvement in patients’ outcomes, the satisfaction level from patients and the healthcare team was enhanced. In addition, our feedback suggested that the intervention protocol is well received by patients and staff. Through multidisciplinary collaboration, this project also fostered a culture of mobility promotion into the daily nursing care process and supported early management of mobility-related complications. With the mobility promotion in the multidisciplinary team context, it ensured mobility plans were tailored to individual patient needs and incorporated input from PTs. Additionally, JH-HLM was successfully integrated into the care process as a mobility assessment tool to standardise the quantification of mobility in clinical practice. As there are no reported studies in Singapore exploring the impact of mobility protocol in hospitalised older adults, this project provides important insight into the benefits of mobility promotion in the local context.

Our study is potentially limited in several aspects. First, this was a QI project conducted in a single geriatric unit of a large tertiary hospital with a relatively small sample size. Second, this project took place during the COVID-19 pandemic, resulting in a shorter timeframe for implementation and smaller sample size due to competing demands for clinical space and resources. Third, the use of non-randomised design cannot establish a direct cause-and-effect relationship between the mobility protocol and the outcome measures, such as improved mobility and reduced LOS. While patients’ characteristics were very homogeneous between pre-implementation and post implementation groups, and the findings were adjusted for known confounders using multivariable regression analysis, the project team could not rule out the possibility that there may be other factors contributing to this reduction in LOS. Fourth, while nurses documented the highest level of mobility and the frequency that patients were mobilised daily during hospitalisation, some aspects of mobility were not captured, including the duration of an out-of-bed mobility episode, the distance of ambulation, and the extent of mobility activities with therapists. As such, our data does not fully represent the total mobility and activity levels of the patients. This can be addressed in our future study by enhancing collaboration with PTs to advance and optimise the mobility protocol and its evaluation. In addition, this study did not measure the potential change of mobility level with the interplay of cognitive function and other medical profiles, which can be further explored. Finally, previous studies reported that some older adults were not keen to mobilise due to various factors, including concern of illness exacerbation, perception of bedrest as a vital part of hospitalisation, poor physical status due to various symptoms, fear of falling and fear of bothering the nurses.29,32 These were not explored in our study. As such, future larger prospective randomised studies with a more generalisable patient cohort of hospitalised older adults addressing the limitations above are required to confirm and extend the findings of our study.

Implications for policy, practice and research

This project has several implications for policy, practice and research. Notably, a structured nurse-driven mobility protocol was used to standardise mobility practice. This provides a standardised guideline for nurses on mobility assessment, the frequency and the type of mobility events to be provided for hospitalised older adults. Furthermore, a mobility protocol fosters multidisciplinary communication and collaboration.¹⁰ It also provides role delineation for nurses to advocate for patients and clarify the indication of bedrest orders. As such, the nurses are more likely to take steps to provide mobility activities for the older adults.²³ A mobility workgroup can be instrumental in policy making and practice standardisation for mobility promotion in future scale-up and spread. The sustainability plan for the mobility protocol needs to consider ongoing evaluation, integration of technology-based solutions to support mobility training for nurses and development of supporting staff, and continuous QI. In the future, practice standardisation across departments will need regular review of the protocol based on patient outcomes, patient and healthcare team feedback, and evidence-based practices. Additionally, we will provide continuous education and training for staff and reward nursing mobility champions to maintain their momentum and foster a culture of mobility excellence. Not to mention, it is important for the mobility workgroup takes the helm to drive the mobility promotion hospital-wide.

Future research should evaluate the effectiveness of this mobility protocol on the readmission rate among hospitalised older adults over a longer period in the local context. Economic analysis may be conducted to determine healthcare cost benefits of mobility protocol in the long term.

Conclusions

A nurse-driven mobility protocol with a multidisciplinary collaborative effort demonstrated increased patient’s out-of-bed episodes, improved mobility levels and contributed to a non-significant reduction of LOS. Incorporation of the JH-HLM scale enabled the standardisation of the method of mobility assessment and the establishment of guidelines for patient mobility promotion in the hospital.

Data availability statement

No data are available.