Effect of an integrated care model for pre-frail and frail older people living in community

Abstract

Objectives

to evaluate the effect of an integrated care model for pre-frail and frail community-dwelling older people.

Design

a quasi-experimental design.

Setting and participants

we enrolled people aged ≥60 years from a community care project. An inclusion criterion was pre-frailty/frailty, as measured by a simple frailty questionnaire (FRAIL) with a score of ≥1.

Methods

we assigned participants to an intervention group (n = 183) in which they received an integrated intervention (in-depth assessment, personalised care plans and coordinated care) or a control group (n = 270) in which they received a group education session on frailty prevention. The outcomes were changes in frailty, individual domains of frailty (‘fatigue’, ‘resistance’, ‘ambulation’, ‘illnesses’ and ‘loss of weight’) and health services utilisation over 12 months. Assessments were conducted at baseline and at the 12-month follow-up.

Results

the mean age of the participants (n = 453) at baseline was 76.1 ± 7.5 years, and 363 (80.1%) were women. At follow-up, the intervention group showed significantly greater reductions in FRAIL scores than the control group (P < 0.033). In addition, 22.4% of the intervention and 13.7% of the control participants had reverted from pre-frail/frail to robust status, with the difference reaching significance when the intervention was compared with the control group (odds ratio [OR] 1.6, 95% confidence interval [CI] 1.0–2.4) after adjustments for age, sex, living arrangement/marital status and hypercholesterolemia. For individual domains of frailty, the adjusted OR for improved ‘resistance’ was 1.7 (95% CI 1.0–2.8). However, no effects were found on reducing use of health services.

Conclusion

the integrated health and social care model reduced FRAIL scores in a combined population of pre-frail/frail community-dwelling older people attending older people’s centres.

ageing, frailty, integrated care, person-centred, personalised care plans, older people

Key points

• The integrated health and social care model reduced frailty score in a combined population of pre-frail/frail community-dwelling older people attending older people’s centres.

• The model of care supported social care providers to deliver services based on older people’s needs.

• The model of care oriented services towards community based through a coordinated use of existing community resources and practices.

Introduction

Frailty is a geriatric syndrome characterised by decreases in the functional reserve that places older people at risk of falls, disability, hospitalisation and death [1]. Therefore, effective care models should be identified to prevent or delay the progression of frailty.

Recognition and management of older people with frailty in community settings have emerged as an effective way to improve outcomes for people with frailty [2,3]. A number of interventional trials (e.g. strength training and protein supplementation) have also been shown to be an effective strategy to reduce frailty and improve functional capacity in older people [4–6]. However, translating interventional results that are effective into community programs entail complex changes in organisation and practices. To the best of our knowledge, there are still relatively few community-based interventions designed to explicitly focus on frailty prevention, although the number of older people with frailty is expected to increase [7]. On the other hand, integrating self-management support in a coordinated primary care supported by assessments and interventions may prevent or even reverse frailty. The hypothesis is that helping frail older people to acquire knowledge and skills for managing frailty has the potential to enable them to take better care of themselves, receive services tailored to their needs, remain independent for longer and rely less on secondary/tertiary healthcare. The World Health Organization (WHO) underpins the importance of supporting older people to self-manage their health conditions by providing peer support, training, information and advice [8,9], and it suggests changing the focus of health care from an emphasis on acute conditions to one that promotes integrated care resulting in healthy ageing [10]. However, there have been few primary care services that are provided through an integrated approach for frail older people, although there have been some attempts in the USA, Canada and some European countries [11,12].

With the ageing of the population, an increasing trend in frailty is to be expected in Hong Kong [7], which urges integrated primary care for frail older people. Therefore, we conducted a quasi-experimental study to examine the effectiveness of an integrated care model supported by frailty assessment, personalised care plans and coordinated care services as arranged by community centres for older people in pre-frail and frail older people.

Methods

Study design

This study is a part of an ongoing community primary care project (Jockey Club Community eHealth Care Project) commenced since 2016 for Chinese people aged 60 years or older who are members of older people’s community centres across 18 districts of Hong Kong. The aims of the project are to empower older people on self-care health management, promote older people’s centres as the first point of contact for addressing health and social needs for older people and ultimately, to facilitate the development of a medico social integrated model of primary care for older people. In this study, 38 older people’s centres were selected from a list of 40 that joined the project between 2016 and 2018. These centres were located across 17 districts and were heterogenetic in terms of funding mode (sub-vented vs. self-financing), service provision level (district vs. neighbourhood) and centre attendance (average 348, range 65–1,100).

To minimise the group differences, recruited centres were pair-matched according to funding mode, service provision level, centre attendance and geographical location. In each pair, the first centre able to identify an older person with pre-frailty/frailty who had a score of 1 or higher on a simple frailty questionnaire (FRAIL) was allocated to an intervention group while the other centre was allocated to a control group. From each centre, older members identified with pre-frailty/frailty were recruited until a quota of 30 had been reached. Frailty status, other geriatric syndromes and health services utilisation were assessed for both groups at baseline (September 2016–February 2018) and at the 12-month follow-up (September 2017–February 2019). All these centres had implemented a frailty screening (within the last 3 months before study commencement). The assessors were blinded to the allocation of intervention. The project was approved by the Survey and Behavioural Ethics Committee that is free of author and institutional identification. All participants provided written (or witnessed, if illiterate) informed consent.

Intervention

Participants in the intervention group received an integrated intervention consisting of in-depth assessment, personalised care plans and coordinated care (see Appendix 1).

In-depth assessment

Individual interviews were conducted to identify problems or needs, using an electronic questionnaire. Common geriatric syndromes have been taken into account, which included ‘yes/no’ questions to the following domains (chewing difficulties, vision impairment, hearing impairment, sarcopenia, memory complaints, self-rated health, psychological well-being, incontinence, instrumental activities of daily living impairment and polypharmacy), followed by in-depth assessment and checking of prescribed medications. The assessments were conducted by a nurse with experience working in geriatric wards or a health worker holding a degree in a health-related field and trained with geriatric care and case management skills with topics covering frailty and fall prevention, nutrition for better health, simple cognitive assessment and management of mobility impairment and medication issue.

Personalised care plans

A follow-up session of personalised care planning was provided to deal with the identified problems or needs, which was conducted by the nurse or the trained health worker. Problems or needs associated with frailty were listed in a report per participant. Personalised regimes and recommendations including possible actions and interventions to reduce frailty were discussed based on the problems or needs identified. Depending on the functional levels and risk of falls of each participant, different types of physical exercises consisting of aerobic, resistance, flexibility and balance components based on the protocol of a local validated frailty prevention program [6] were introduced. Dietary advice was provided based on the Asia-Pacific Clinical Practice Guidelines for the management of frailty containing two nutritional recommendations, that adults with frailty with unintentional weight loss be screened for reversible causes and considered for protein and caloric supplementation/food fortification, and vitamin D be prescribed for those found to be deficient [13]. Frailty-related problems and medication management were discussed with the participants using a resource kit providing them knowledge about the problems and recommendations to improve their conditions. A practical guide booklet to prevent frailty was provided to each participant. The nurse and the health worker also motivated participants to participate in group programs matching their suitable options within the community. A service directory of contacts of health professionals was provided when necessary.

Coordinated care

Social care providers of the older people’s centres were invited to observe the assessment and caring planning sessions, after which the nurse or the health worker liaised with the social care providers on the status/conditions of the participant so that they could follow the participant throughout to ensure that they receive continued support or care. The report for each participant and the resource kit were provided to social care providers for reference. Further consultations with the participants were provided on an as-needed basis.

Control

Participants in the control group received a group-based education session on frailty prevention provided by the nurse or the trained health worker.

Outcome measures

The primary outcome was frailty, which was measured using the FRAIL scale [14,15]. It included five domains: fatigue, resistance, ambulation, illnesses and loss of weight. The FRAIL scores range from 0 to 5, with 0–1 point for each domain and a total score of 0 represents robust, 1–2 as pre-frail and 3–5 as frail. The secondary outcomes were the five domains in the FRAIL scale and the use of health services (hospitalisation, specialist outpatient clinic and general outpatient clinic) in the previous 12 months.

Statistical analysis

Continuous variables were tested for the normality of the distribution using the Kolmogorov–Smirnov test. Categorical variables were described as frequencies and percentages. Comparisons between baseline and follow-up measures were performed by pair t-tests or non-parametric Wilcoxon signed tests for continuous variables or McNemar tests or non-parametric Wilcoxon signed ranks tests for categorical variables. The mean differences of the outcome measures between the intervention and control groups were examined using independent t-tests or non-parametric Mann–Whitney tests. The percentage differences of outcome measures between the intervention and control groups were examined using binary logistic regressions. Odds ratios (ORs) with 95% confidence intervals (CIs) were reported. Models were adjusted for covariates (age, sex, living arrangement/marital status and presence of hypercholesterolemia). The variance inflation factor (VIF) was used to examine the degree of multi-collinearity among the independent variables for regression models. Multi-collinearity among variables was defined as a VIF ≥10 [16]. All analyses were carried out using the Window-based SPSS Statistical Package (v. 25.0; SPSS, Chicago, IL) and P values of <0.05 were considered statistically significant.

Results

Between 2016 and 2018, we screened 792 people and 347 (43.8%) of them were identified with pre-frailty and 106 (13.4%) with frailty. These 453 people were recruited and assigned to either the intervention (n = 183) or the control group (n = 270) based on the older people centre’s assignment as previously described. The mean age of the participants at baseline was 76.1 ± 7.5 years, 80.1% were women—45.9% married—and 76.4% had below secondary education levels. There were no significant differences between the two groups except for living arrangement (P = 0.006) and presence of hypercholesterolemia (P = 0.002) (Table 1).

Table 1

Baseline characteristics of participants.

	Intervention group (n = 183)	Control group (n = 270)
	Mean ± SD/n (%)	Mean ± SD/n (%)	Pa
Demographics
Age at the baseline visit, years	76.3 ± 7.1	75.9 ± 7.8	0.606
Women	146 (79.8)	217 (80.4)	0.877
Education (<secondary)	139 (76.0)	207 (76.7)	0.861
Married or cohabiting	74 (40.4)	134 (49.6)	0.054
Living alone	72 (39.3)	73 (27.0)	0.006
Self-reported medical disorders
Hypertension	104 (56.8)	133 (49.3)	0.113
Hypercholesterolemia	25 (13.7)	69 (25.6)	0.002
Diabetes	38 (20.8)	62 (23.0)	0.580
Heart diseases	24 (13.1)	35 (13.0)	0.962
Stroke	7 (3.8)	10 (3.7)	0.947
Frailty
FRAIL score, median (quartile: 25–75%)	2.0(1.0–3.0)	2.0(1.0–2.0)	0.099
Robust	0	0	0.241
Pre-frail	135 (73.8)	212 (78.5)
Frail	48 (26.2)	58 (21.5)
Other geriatric syndromes
Chewing difficulties	82 (44.8)	104 (38.5)	0.182
Vision impairment	51 (27.9)	74 (27.4)	0.914
Hearing impairment	35 (19.1)	56 (20.7)	0.674
Sarcopenia	35 (19.1)	54 (20.0)	0.818
Memory complaints	148 (80.9)	225 (83.3)	0.472
Poor self-rated health	25 (13.7)	35 (13.0)	0.830
Poor psychological well-being	38 (20.8)	42 (15.6)	0.154
Incontinence	9 (4.9)	10 (3.7)	0.527
Instrumental activities of daily living	61 (36.1)	78 (30.7)	0.248
Prescribed medications ≥5	48 (26.2)	56 (20.7)	0.594
Use of health servicesb
GOPC visit	147 (80.3)	207 (76.7)	0.355
SOPC visitc	157 (85.8)	236 (88.1)	0.480
Hospitalisation	46 (25.1)	72 (26.7)	0.716

GOPC, general outpatient clinic; SOPC, specialist outpatient clinic.

^aBetween-group differences (independent t-tests/Chi-squared tests).

^bThe reported percentages are self-reported attendance of the health services in the past 12 months.

^cMissing data: Instrumental activities of daily living (intervention group: n=14, control group: n=16); SOPC (control group: n = 2).

At the 12-month follow-up, the mean change in FRAIL score was reduced for the intervention group (−0.4, P < 0.001) and for the control group (−0.2, P = 0.011). Between-group difference was significant (P = 0.033). In addition, 22.4% of the intervention group had reverted from pre-frail/frail to robust status (P < 0.001), whereas only 13.7% reverted to robust status in the control group (P < 0.001). Among the domains in the FRAIL scale, ‘fatigue’, ‘resistance’ and ‘ambulation’ were the domains that showed a significant improvement at the 12-month follow-up in the intervention (P < 0.05) but not in the control participants. No significant effects were observed for the use of health services (Table 2).

Table 2

Levels of frailty, the different domains of frailty and use of health services at baseline and at follow-up.

	Intervention group (n = 183)				Control group (n = 270)
	Baseline	12 months	Difference	Pa	Baseline	12 months	Difference	Pa	Pb
	Mean ± SD/n (%)				Mean ± SD/n (%)
FRAIL score	1.94 ± 1.03	1.56 ± 1.20	−0.38 ± 1.22	<0.001	1.79 ± 0.90	1.63 ± 1.07	−0.16 ± 1.04	0.011	0.033
Frailty status
Robust	0	41 (22.4)		<0.001	0	37 (13.7)		<0.001
Pre-frail	135 (73.8)	100 (54.6)			212 (78.5)	176 (65.2)
Frail	48 (26.2)	42 (23.0)			58 (21.5)	57 (21.1)
FRAIL domains
Fatigue
Yes	119 (65.0)	102 (55.7)		0.047	178 (65.9)	174 (64.4)		0.746
No	64 (35.0)	81 (44.3)			92 (34.1)	96 (35.6)
Resistance
Yes	123 (67.2)	103 (56.3)		0.014	174 (64.4)	160 (59.3)		0.099
No	60 (32.8)	80 (43.7)			96 (35.6)	110 (40.7)
Ambulation
Yes	62 (33.9)	46 (25.1)		0.030	78 (28.9)	66 (24.4)		0.134
No	121 (66.1)	137 (74.9)			192 (71.1)	204 (75.6)
Illnesses
Yes	23 (12.6)	18 (9.8)		0.472	30 (11.1)	26 (9.6)		0.617
No	160 (87.4)	165 (90.2)			240 (88.9)	244 (90.4)
Loss of weight
Yes	28 (15.3)	17 (9.3)		0.054	23 (8.5)	15 (5.6)		0.216
No	155 (84.7)	166 (90.7)			247 (91.5)	255 (94.4)
Use of health servicesc
GOPC
Yes	147 (80.3)	140 (76.5)		0.360	207 (76.7)	217 (80.4)		0.237
No	36 (19.7)	43 (23.5)			63 (23.3)	53 (19.6)
SOPCd
Yes	157 (85.8)	164 (89.6)		0.189	236 (88.1)	243 (90.0)		0.486
No	26 (14.2)	19 (10.4)			32 (11.9)	27 (10.0)
Hospitalisation
Yes	46 (25.1)	51 (27.9)		0.583	72 (26.7)	67 (24.8)		0.596
No	137 (74.9)	132 (72.1)			198 (73.3)	203 (75.2)

GOPC, general outpatient clinic; SOPC, specialist outpatient clinic.

^aWithin-group differences (pair t-tests/Wilcoxon signed tests/McNemar tests/Wilcoxon signed ranks tests).

^bBetween-group differences (independent t-tests/Mann–Whitney tests).

^cThe reported percentages are self-reported attendance of the health services in the past 12 months.

^dMissing data: SOPC (control group, n = 2).

Compared with the control group, the intervention group had a greater improvement in frailty status, with an OR of 1.6 (95% CI 1.1–2.4). This remained significant after adjustments for age, sex, living arrangement/marital status and presence of hypercholesterolemia. Further adjustments for chronic illnesses (hypertension, diabetes, heart diseases and stroke) did not change the results (OR 1.6, 95% CI 1.0–2.4) (data not shown). For individual domains of frailty, the adjusted OR for improved ‘resistance’ was 1.7 (95% CI 1.0–2.8). No significant improvements were observed for the other domains of frailty or the health services used (Table 3).

Table 3

Odds ratio and 95% confidence interval for changes in levels of frailty, the different domains of frailty and use of health services at follow-up.

	Intervention group (n = 183)	Control group (n = 270)	Model 1		Model 2		Model 3		Model 4
			Control group as reference		Control group as reference		Control group as reference		Control group as reference
Change from baseline	n (%)	n (%)	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Frailty status
Improved	61 (33.3)	64 (23.7)	1.61 (1.06–2.44)	0.025	1.61 (1.06–2.45)	0.024	1.55 (1.02–2.38)	0.042	1.59 (1.04–2.43)	0.032
Not improved	122 (66.7)	206 (76.3)
FRAIL domains
Fatigue
Improved	41 (22.4)	45 (16.7)	1.44 (0.90–2.32)	0.128	1.44 (0.90–2.31)	0.132	1.35 (0.84–2.19)	0.216	1.39 (0.86–2.24)	0.179
Not improved	142 (77.6)	225 (83.3)
Resistance
Improved	40 (21.9)	38 (14.1)	1.71 (1.05–2.79)	0.032	1.71 (1.05–2.79)	0.032	1.71 (1.03–2.82)	0.037	1.70 (1.03–2.80)	0.038
Not improved	143 (78.1)	232 (85.9)
Ambulation
Improved	32 (17.5)	33 (12.2)	1.52 (0.90–2.58)	0.119	1.52 (0.89–2.57)	0.123	1.52 (0.89–2.61)	0.127	1.53 (0.89–2.61)	0.124
Not improved	151 (82.5)	237 (87.8)
Illnesses
Improved	18 (9.8)	20 (7.4)	1.36 (0.70–2.66)	0.362	1.38 (0.71–2.70)	0.342	1.30 (0.65–2.59)	0.460	1.35 (0.68–2.68)	0.398
Not improved	165 (90.2)	250 (92.6)
Loss of weight
Improved	19 (10.4)	20 (7.4)	1.45 (0.75–2.80)	0.270	1.45 (0.75–2.80)	0.271	1.45 (0.74–2.83)	0.279	1.53 (0.78–3.00)	0.212
Not improved	164 (89.6)	250 (92.6)
Use of health servicesa
GOPC
Reduced	25 (13.7)	24 (8.9)	1.62 (0.90–2.94)	0.111	1.62 (0.89–2.94)	0.113	1.56 (0.85–2.86)	0.149	1.55 (0.85–2.84)	0.157
Not reduced	158 (86.3)	246 (91.1)
SOPCb
Reduced	7 (3.8)	14 (5.2)	0.72 (0.29–1.82)	0.490	0.72 (0.28–1.82)	0.486	0.72 (0.28–1.85)	0.494	0.73 (0.29–1.87)	0.514
Not reduced	176 (96.2)	254 (94.8)
Hospitalisation
Reduced	24 (13.1)	31 (11.5)	1.16 (0.66–2.06)	0.602	1.18 (0.67–2.09)	0.566	1.18 (0.66–2.11)	0.587	1.14 (0.64–2.03)	0.663
Not reduced	159 (86.9)	239 (88.5)

GOPC, general outpatient clinic; SOPC, specialist outpatient clinic.

Model 1: crude; model 2: adjusted for age and sex; model 3: adjusted for variables in model 2, living arrangement and presence of hypercholesterolemia; model 4: adjusted for variables in model 2, marital status and presence of hypercholesterolemia.

^aThe reported percentages are self-reported attendance of the health services in the past 12 months.

^bMissing data: SOPC (control group, n = 2).

Discussion

In this study, a model of care of older people integrating screening, in-depth assessment, personalised care plans and coordinated care was implemented in community centres for older people in Hong Kong, results of which demonstrated that the model of care succeeded in reducing frailty scores in a combined population of pre-frail/frail community-dwelling older people attending older people’s centres.

Compared with those in the control group, pre-frail and frail older people in the intervention group had significantly higher odds for improved frailty status at the 12-month follow-up. The results are likely explained by some features of the intervention. First, in-depth assessments were conducted by a geriatric nurse or a health worker trained with geriatric assessment and case management skills. This determined the efficacy to design personalised care plans and provide self-management support, contributing to the improvement. Second, the provision of personalised care plans, using a people-centred approach with the older people’s preferences considered during the process could facilitate early interventions for frailty (e.g. participation in an exercise program) and its associated issues uncovered with the in-depth assessment. Particularly, a significant improvement in ‘resistance’, a domain in the FRAIL scale was observed in the intervention group. This improvement may be related to the personalised care planning process during which the nurse or the health worker demonstrated to participants some exercises matching their functional level, provided them with dietary advice to prevent muscle loss, and motivated them to participate in group programs. Third, coordinated care that involved an interdisciplinary team of members (e.g. nurses, health workers, social workers and centre staff members) was provided in community-based settings, which can facilitate continuity of care for community-dwelling older people. Therefore, our findings suggest that a medico social integrated and people-centred care in community settings may be an important element of a successful care system for preventing and delaying the progression of frailty. Nevertheless, a modest improvement in frailty was also observed among the control group receiving a group-based education session on frailty prevention. It is speculated that education increases the understanding of frailty, promotes early intervention and therefore, improves outcomes in the control group.

The design of the intervention in this study shared similarities with other community-based models developed for frail older people in western countries. For example, the intervention in this study and those in an Australian study [17], a Canadian study [2], a Dutch study [18], and a national program in France [19] included simple screening tools (either for assessing frailty or disability), in-depth/comprehensive geriatric assessments conducted by nurses/physiotherapists/trained health workers, personalised care plans and coordinated care services/interventions delivered by interdisciplinary providers. Results from the Canadian study demonstrated a significant improvement in the use of health services but did not use frailty as an outcome measure. The Dutch study did not find any effects of the model with regards to disability, and that preliminary data from the French program detected a significant improvement in polypharmacy and emergency visits but not in other outcome indicators. In this study, we did not find any significant effects of the model on the use of health services, as opposed to the Canadian study which showed a stabilised emergency room use in the experimental areas receiving a coordination-based integrated care over 4 years. The short follow-up time and the lack of provision of medical intervention—such as adjustment of medications or doctor consultation of our study—are possible explanations for the lack of reduction in the use of health services observed. Furthermore, unlike the Canadian study, a health portal facilitating communications between service providers and older members was not available in our study, although it has been recognised as an effective mechanism for supporting continuous care for frail older people [20]. Therefore, even if the intervention has positive effects on frailty, the effects on secondary and tertiary health services might be small in this study.

Results of this study also contribute to design strategies for caring pre-frail and frail older people in community settings. Indeed, integrated care is increasingly advocated as a means to develop more effective models of care and improve health outcomes. A WHO collaborating centre for frailty, clinical research and geriatric training has been formed in Toulouse, France to support WHO’s global world strategy on healthy ageing where a care model of screening, using a self-administered questionnaire and management for frail older people has been feasibly implemented in Cugnaux, France [21,22]. However, in Hong Kong, care for older people is not yet integrated, in that most of the services are designed to respond to specific diseases independently. It is likely that older people with frailty and its associated problems such as sarcopenia and memory problems are often left undetected in the health care setting. Although community centres for older people would seem well suited to provide integrated care for older people, the lack of knowledge and/or competence in geriatric care among social care providers present a significant challenge to plan and expand community older people’s services, particularly for frail older people as they have different healthcare requirements, thus, support and training that provides social care providers with knowledge in recognising and managing frailty would be beneficial. Policies are also needed to support the integration of primary health and social care services at the system level so that a truly system-wide shift can happen. In this context, this study illustrates how the health and social care services are coordinated and integrated in community settings to prevent or delay the progression of frailty.

There were some limitations in this study. First, the design of the study was not a randomised controlled trial; however, the method of pairing the centres based on funding mode, service provision level, centre attendance and location would minimise the biases due to baseline group differences. Analysis of the characteristics found that there were no significant differences between the groups with regards to demographics, medical history and geriatric syndromes except for living arrangement and presence of hypercholesterolemia, which have been adjusted in the analyses. Second, two centres (5%) which might had a less proactive attitude towards the delivery of care to frail older people declined to participate, possibly leading to a smaller effect on the change in frailty in the intervention group. Nevertheless, the pragmatic approach of this study combining assessments, care plans and coordinated care with the use of existing resources suggests that future incorporation of the integrated care model into existing practice may be feasible. It has also been suggested that pragmatic trials is a way to improve the relevance of clinical trial results to practice [23]. Third, compared to the general older population in Hong Kong, the participants have a lower educational level and fewer men participated. The results may not be applicable to the general population. Forth, the FRAIL scale was designed as a quick screening test for frailty. It may not be particularly sensitive to change. Finally, we did not estimate the cost-effectiveness of the intervention. Nevertheless, we speculate that the cost will likely be low because the intervention was provided by trained health workers based in community centres as the first step in a step care approach to primary care for older people without the participation of medical staff. However, previous cost-effectiveness studies have shown mixed results [24–27]. Further studies on the cost-effectiveness of the integrated care model are warranted.

In conclusion, this study adds evidence in support of the benefits of an integrated health and social care model to frailty prevention in the community. The model of care also supported social care providers to deliver services based on older people’s needs and oriented services towards community-based care, which would be deemed sustainable since it was delivered primarily through a coordinated use of existing resources without significant changes in infrastructure, organisation or practices. Further investigations will be performed to document the impact of the model on experience of care, self-management abilities, use of health services of pre-frail and frail older people and adaptability of the model to service providers.

Acknowledgements

We wish to thank the Hong Kong Jockey Club Charities Trust in supporting the Jockey Club Community eHealth Care Project. We also wish to extend our thanks to Ms. LY Tam, Ms. Clara Cheng and Ms. Sara Kong of CUHK Jockey Club Institute of Ageing, CUHK Stanley Ho Big Data Decision Analytics Research Centre, Jockey Club Community eHealth Care Project team, the Senior Citizen Home Safety Association and all social care providers of the participating centres for their contributions to the implementation of the intervention.

Declaration of Conflict of interest

None.

Funding

This work was supported by the Hong Kong Jockey Club Charities Trust.