Preventive interventions to improve older people’s health outcomes: systematic review and meta-analysis

Background Systematic reviews of preventive, non-disease-specific primary care trials for older people often report effects according to what is thought to be the intervention’s active ingredient. Aim To examine the effectiveness of preventive primary care interventions for older people and to identify common components that contribute to intervention success. Design and setting A systematic review and meta-analysis of 18 randomised controlled trials (RCTs) published in 22 publications from 2009 to 2019. Method A search was conducted in PubMed, MEDLINE, Embase, Web of Science, CENTRAL, CINAHL, and the Cochrane Library. Inclusion criteria were: sample mainly aged ≥65 years; delivered in primary care; and non-disease-specific interventions. Exclusion criteria were: non-RCTs; primarily pharmacological or psychological interventions; and where outcomes of interest were not reported. Risk of bias was assessed using the original Cochrane tool. Outcomes examined were healthcare use including admissions to hospital and aged residential care (ARC), and patient-reported outcomes including activities of daily living (ADLs) and self-rated health (SRH). Results Many studies had a mix of patient-, provider-, and practice-focused intervention components (13 of 18 studies). Studies included in the review had low-to-moderate risk of bias. Interventions had no overall benefit to healthcare use (including admissions to hospital and ARC) but higher basic ADL scores were observed (standardised mean difference [SMD] 0.21, 95% confidence interval [CI] = 0.01 to 0.40) and higher odds of reporting positive SRH (odds ratio [OR] 1.17, 95% CI = 1.01 to 1.37). When intervention effects were examined by components, better patient-reported outcomes were observed in studies that changed the care setting (SMD for basic ADLs 0.21, 95% CI = 0.01 to 0.40; OR for positive SRH 1.17, 95% CI = 1.01 to 1.37), included educational components for health professionals (SMD for basic ADLs 0.21, 95% CI = 0.01 to 0.40; OR for positive SRH 1.27, 95% CI = 1.05 to 1.55), and provided patient education (SMD for basic ADLs 0.28, 95% CI = 0.09 to 0.48). Additionally, admissions to hospital in intervention participants were fewer by 23% in studies that changed the care setting (incidence rate ratio [IRR] 0.77, 95% CI = 0.63 to 0.95) and by 26% in studies that provided patient education (IRR 0.74, 95% CI = 0.56 to 0.97). Conclusion Preventive primary care interventions are beneficial to older people’s functional ability and SRH but not other outcomes. To improve primary care for older people, future programmes should consider delivering care in alternative settings, for example, home visits and phone contacts, and providing education to patients and health professionals as these may contribute to positive outcomes.

using the original Cochrane tool.Outcomes examined were healthcare use including admissions to hospital and aged residential care (ARC), and patient-reported outcomes including activities of daily living (ADLs) and self-rated health (SRH).

Results
Many studies had a mix of patient-, provider-, and practice-focused intervention components (13 of 18 studies).Studies included in the review had low-to-moderate risk of bias.Interventions had no overall benefit to healthcare use (including admissions to hospital and ARC) but higher basic ADL scores were observed (standardised mean difference [SMD] 0.21, 95% confidence interval [CI] = 0.01 to 0.40) and higher odds of reporting positive SRH (odds ratio [OR] 1.17, 95% CI = 1.01 to 1.37).When intervention effects were examined by components, better patient-reported outcomes were observed in studies that changed the care setting (SMD for basic ADLs 0.21, 95% CI = 0.01 to 0.40; OR for positive SRH 1.17, 95% CI = 1.01 to 1.37), included educational components for health professionals (SMD for basic ADLs 0.21, 95% CI = 0.01 to 0.40; OR for positive SRH 1.27, 95% CI = 1.05 to 1.55), and provided patient education (SMD for basic ADLs 0.28, 95% CI = 0.09 to 0.48).Additionally, admissions to hospital in intervention participants were fewer by 23% in studies that changed the care setting (incidence rate ratio [IRR] 0.77, 95% CI = 0.63 to 0.95) and by 26% in studies that provided patient education (IRR 0.74, 95% CI = 0.56 to 0.97).

Conclusion
Preventive primary care interventions are beneficial to older people's functional ability and SRH but not other outcomes.To improve primary care for older people, future programmes should consider delivering care in alternative settings, for example, home visits and phone contacts, and providing education to patients and health professionals as these may contribute to positive outcomes.

Introduction
Reliable research evidence is one of the key elements of evidence-based decision making. 1 With multimorbidity on the rise, particularly for older people, 2,3 the need to generate relevant evidence to guide management of older primary care patients [4][5][6] and the use of broader outcome measures that matter to older people, [5][6][7] such as maintaining independence and quality of life (QoL), is increasingly emphasised in the literature.
Many intervention trials set in primary care have tested the effectiveness of preventive rather than disease-specific approaches relating to the care of older people, and numerous systematic reviews related to this topic have been published.[12][13][14] These reviews often report effects according to the primary intervention focus or what is thought to be its active ingredient -for example, multidimensional assessment and subsequent management; 8,9,14,17 health promotion and disease prevention programmes; 15,16,18 or preventive home visiting. 19,20However, preventive strategies and intervention programmes typically include multiple intervention components to address diverse objectives for change.The most recent Medical Research Council (MRC) guidance 21 emphasises how complexity may arise from the content of interventions (for example, having several potentially interacting intervention components RESEARCH | e209 and mechanisms of change) or the context within which the intervention is implemented.
Summarising evidence with sufficient detail to guide development of intervention programmes can be particularly challenging for complex interventions.As a case in point, reviews on preventive primary care interventions that focus on multidimensional assessment have found benefits, such as reduced admissions to hospitals and nursing homes, 8 and better functional ability. 8,9,17In contrast, interventions that include multidimensional assessments as an intervention component but are primarily focused on a change in the setting of care delivery (preventive home visiting) or enabling patients to improve their health (health promotion and disease prevention programmes) have been less successful in demonstrating consistent effects on healthcare use 18,19 and functioning. 19,209][20] It is for this reason that the aim of the current study was to examine the overall impact of primary care interventions and to identify key intervention components that contribute to an effect on healthcare use, functional ability, and QoL.

Method
In addition, inadequate intervention reporting is a long-standing issue, [22][23][24][25][26] possibly even more so for trials pre-dating reporting tools and guidance such as the 2008 MRC framework.The authors of the current study therefore chose to exclude earlier studies as the aim was to examine specific intervention components in this review.

Study selection
Titles and abstracts were independently examined by two reviewers using pre-defined eligibility criteria.The authors selected reports on intervention trials aimed at reducing admissions to hospital and ARC placement or improving functional ability and QoL, and where: • the sample consisted mainly of people aged ≥65 years -for studies that included adults aged <65 years, a mean age of ≥70 years was required or a separate reporting of outcomes by age group; • the intervention was delivered in primary care; and • the study had a general focus, that is, it was applicable to the general older primary care population and was not disease specific.
Studies were excluded that: • were not RCTs; • focused primarily on a pharmacological or psychological intervention; or Effects were coded as an intervention benefit where studies reported statistically significantly fewer admissions to hospital or ARC, better functioning, or higher QoL in favour of the intervention group; otherwise, they were coded as having nil effect or favouring the control.P-values <0.05 were considered statistically significant.Effect estimates were independently extracted from studies by the assigned reviewers.For studies that examined healthcare use, the proportion admitted to hospital, the total number of hospital admissions, the total number of days in hospital, and the proportion admitted to ARC over the follow-up period were extracted.For studies that examined functional ability, basic and extended activities of daily living (ADLs) scores on follow-up were extracted.For studies that examined QoL, outcome measures extracted were the proportion reporting positive self-rated health (SRH), that is, those selecting response options good and above, the SRH score for studies that treated SRH as a continuous variable, QoL index scores, physical QoL domain scores, and mental QoL domain scores on follow-up.

How this fits in
Study investigators were contacted to obtain additional information when necessary.Estimates were transformed into odds ratios (ORs), incidence rate ratios (IRRs), ratios of durations, and standardised mean differences (SMDs), as appropriate.

Quality assessment
Assigned reviewers used the Cochrane Handbook for Systematic Reviews of Interventions tools for assessing risk of bias (original tool) and quality of evidence. 27

Data synthesis
Data on intervention effects were synthesised narratively and through meta-analysis.Narrative synthesis of intervention effects by intervention domains and components involved vote counting based on direction of effect and statistical significance, which is further described in Supplementary Information S2.Random-effects meta-analysis was performed to obtain pooled intervention effects for similar outcome measures.To adjust for clustering inherent in the design of some trials, the authors of the current study inflated standard errors of estimates from cluster RCTs by multiplying to the square root of their respective design effects.Pooled effects were also examined according to the most common intervention components.Confidence intervals (CIs) were constructed using a 95% confidence level.R and Microsoft Excel were used for overall management of data and analysis.

Results
The search resulted in 8612 publications.The initial screening yielded 135 eligible publications; agreement was 97.6% (4180 of 4283 publications).Eighteen studies from 22 publications remained for data extraction after examining full reports.  Reass for exclusion are outlined in the flow diagram (Figure 1).Table 1 presents the characteristics of 18 studies (total of 22 publications) included in the review.  Mosttudies were conducted in Europe and North America.The median sample size of included studies was 743 (interquartile range [IQR] 267-2057), and the median duration of follow-up was 12 months (IQR 12-21).][37][38][39][40]42,43,[45][46][47]49 Five studies included financial interventions such as provider incentives. 30,31,34,36,37,42,43one of the studies used regulatory interventions such as changes to professional licensure or medical liability.Of the three outcomes of interest for this review, QoL was more frequently reported (17 studies); the exception was Schmidt-Mende et al. 44 When Cochrane intervention taxonomies were used to code study interventions, the reviewed studies included 28 of 53 possible intervention components.Supplementary Table S1 shows the number of studies that included a specific intervention component.The most common intervention components were changes to the setting or site of service delivery such as home visits or phone contacts (14 studies); education for health

Research
Heterogeneity   Overall, studies included in the review were assessed to have low-to-moderate risk of bias (see Supplementary Figure S1).Across the risk assessment domains, incomplete outcome data had the largest proportion of studies categorised as high risk.The median attrition rate of included studies was 23% (IQR 11%-35%).Most studies reported results for multiple outcomes of interest (17 of 18 studies); the exception was Schmidt-Mende et al. 44 None of the studies that examined healthcare use (0 of 13 studies) and less than half of studies that examined functional ability (five of 15 studies) and QoL (six of 17 studies) reported differences favouring the intervention.When intervention effects on outcomes according to intervention domains (see Supplementary Table S2) and components (see Supplementary Table S3) were examined, differences in outcomes were generally nil or favouring the control.

Meta-analysis of intervention effects on older people's outcomes
All studies that examined healthcare use as an outcome (13 of 13 studies) provided sufficient information to compute ORs, IRRs, or ratios of durations.Effect estimates in Figure 2 are grouped into the following outcome measures: proportion admitted to hospital (low-to-moderate quality evidence), frequency of hospital admissions (moderate-quality evidence), length of stay in hospital (moderate-quality evidence), and proportion admitted to ARC (moderateto-high quality evidence).Heterogeneity in studies examining frequency of hospital admissions was statistically significant.Pooled intervention effects are represented as diamonds in Figure 2. No significant difference in measures of healthcare use were observed.
In studies that examined functional ability (15 studies), 11   examining extended ADLs was statistically significant.Figure 3 shows that basic ADL scores were higher in intervention participants (combined SMD 0.21, 95% CI = 0.01 to 0.40).There was no overall benefit to extended ADLs in intervention participants compared with controls.
Fifteen of 17 studies that examined QoL provided sufficient information to compute ORs or SMDs.Pooled intervention effects for the proportion reporting a positive SRH (low-to-moderate quality evidence), SRH treated as a continuous score (moderate-to-high quality evidence), overall QoL index score (low-quality evidence), physical QoL domain score (very-low-quality evidence), and mental QoL domain score (very-low-quality evidence) are shown in Figure 4.
Heterogeneity was statistically significant in studies that treated SRH as a continuous score.Compared with controls, the odds of reporting a positive SRH was 17% higher in the intervention participants (combined OR 1.17, 95% CI = 1.01 to 1.37).An intervention benefit was not observed for the other QoL measures.

Examination of intervention effects by common components
Table 2 summarise the effects of studies that included common intervention components.Intervention participants of studies that changed the setting of care, for example, by doing home visits or phone contacts, had 23% fewer hospital admissions (combined IRR 0.77, 95% CI = 0.63 to 0.95), higher (that is, better) basic ADL scores (combined SMD 0.21, 95% CI = 0.01 to 0.40), and higher odds of reporting a positive SRH (combined OR 1.17, 95% CI = 1.01 to 1.37).In studies that had educational components for patients, intervention participants had 26% fewer hospital admissions (combined IRR 0.74, 95% CI = 0.56 to 0.97) and were found to have higher basic ADL scores (combined SMD 0.28, 95% CI = 0.09 to 0.48).
In studies that included educational components for health professionals, intervention participants had higher basic ADL scores (combined SMD 0.21, 95% CI = 0.01 to 0.40) and higher odds of reporting a positive SRH (combined OR 1.27, 95% CI = 1.05 to 1.55) (Table 2).,48

Summary
The current study sought to identify the key components of successful interventions for older people set in primary care.Overall, it was found that participants of preventive primary care interventions had better functional ability and were more likely to report positive SRH than controls.In addition to better patient-reported outcomes, fewer hospital admissions were observed in studies that changed the care setting and those that provided patient education.Studies that provided education to health professionals had some benefits to patient-reported outcomes but not healthcare use.

Strengths and limitations
Many studies have investigated the impact of complex primary care interventions on older people's outcomes, but attention towards identifying intervention components contributing to intervention success has been lacking. 50,51Previous reviews generally report effects according to what authors consider to be the intervention's active ingredient -the ingredient identified could, understandably, differ based on trialists' and reviewers' background and research interests.The strength of the present review is in using the intervention taxonomies of two Cochrane review groups to determine intervention components regardless of whether they were active ingredients, and found many studies having a mix of patient-, provider-, and practice-focused intervention components.
In synthesising data, vote counting was performed based on direction of effect and statistical significance, a useful but limited approach as it does not account for the size of study samples or the magnitude of effects.However, the current study also went beyond narrative synthesis and combined effect sizes through meta-analysis, albeit there was heterogeneity in estimates for some outcomes.An attempt was made to identify how common intervention components may be contributing to positive outcomes, but subgroups of older primary care patients were not examined as a source of heterogeneity.This is a limitation of the present review as several studies targeted frail older people at risk of adverse health outcomes.It is possible that combined effects are closer to the null because of these studies, as older people at earlier stages of frailty or disability are more likely to show greater intervention benefits from preventive interventions. 52,53rlier studies also tend to show greater benefits, 8 thus the exclusion of studies published before 2009 is another important limitation.The original, 27 rather than the updated, 54 Cochrane tool was used for assessing risk of bias.As risk is assessed at the study level in the original tool and at the results level in the updated tool, it is possible that some results may have been assigned a higher risk of bias than warranted. 55 was noted that some studies described their interventions more extensively than others -it is possible that some intervention components might have been missed.It also cannot be excluded that there was publication bias, but it is reassuring that the estimates produced in the current study are based on trials with generally nil effects, that is, less than half of the studies reported an intervention benefit.

Comparison with existing literature
The impact of complex interventions set in primary care on older people's outcomes has been extensively researched in recent decades.[12][13][14]18 In the present review, an intervention benefit to older people's SRH and functional ability was found. Poor SRH in l Older people with limitations in ADLs have a higher risk for admissions to hospital, 65 although it remains unclear whether the functional benefit observed in the current review is clinically meaningful. The authrs of the current review were unable to observe any study-level benefit for other patient-reported outcomes or for healthcare use.
The context within which these studies were conducted may provide some insight into the current findings.That is, considering improvements in usual primary care for older people 8 and broader system-wide service delivery, 66 it can be difficult to demonstrate success in improving outcomes.It is also possible that the scope for outcome improvement Extended ADLs 10 0.04 (-0.05 to 0.14) 8 0.08 (-0.01 to 0.18) 7 0.03 (-0.07 to 0.13) 9 0.02 (-0.07 to 0.10) 7 0.02 (-0.12 to 0.15) 6 0.00 (-0.16 to 0.15) 6 0.09 (-0.05 to 0.22)

Quality of life
Per cent reported positive SRH, OR QoL index score 6 0.09 (-0.07 to 0.25) 6 0.09 (-0.07 to 0.25) 4 0.17 (-0.05 to 0.38) 5 0.09 (-0.09 to 0.26) 5 0.02 (-0.08 to 0.12) 5 0.02 (-0.08 to 0.12) 4 0.16 (-0.11 to 0.42) Physical domain score 5 -0.03 (-0.12 to 0.05) 4 -0.07 (-0.17 to 0.02) 4 -0.03 (-0.12 to 0.07) 5 -0.03 (-0.12 to 0.05) 3 is not as large as would be expected given the current standard of routine careover a third of the interventions were tested in countries known for their strong primary care tradition, for example, the Netherlands, the UK, and New Zealand.Questions remain whether complex interventions would yield substantial benefits in other contexts, for example, in low-and middle-income countries where the organisation and delivery of primary care has much room for improvement.
The current study found other positive outcomes when intervention effects were examined according to common intervention components that potentially improve older patients' engagement with primary care.Previous reviews of home visiting programmes, that is, change in care settings, have found no difference in the number of people admitted 19,67 and frequency of admissions, 68 and mixed effects in functional ability, 19,20,67,69,70 but in the current study it was noted that there were fewer admissions to hospital in the intervention participants of studies that provided care in an alternative setting and there were benefits to functional ability and SRH.In the current study both home visits and phone contacts were considered as alternatives to the usual clinic setting; both could increase frequency of engagement with primary care by removing the need for patients to travel to GP practices.This review was restricted to interventions that were integrated with primary care by excluding standalone interventions based in the community -this is a possible explanation for the inconsistency of the current findings with previous reviews.
Patient education was another component also associated with fewer admissions to hospitals and better functional ability in intervention participants.Improving patients' knowledge and skills [71][72][73] is a common theme in discussions about increasing levels of patient engagement, particularly for multimorbid older people.It would be interesting to know how these components, which could increase patients' frequency and level of primary care engagement, will influence outcomes when delivered as part of community-based rather than GP practice-based complex interventions.
A previous Cochrane review has shown that clinicians' adherence to recommended practice can be improved by providing them with educational activities 74 such as meetings, outreach, audit, and feedback.In the present review, the authors note that having health professional education as an intervention component positively impacts on functional ability and SRH in older people.Taking a programme theory perspective, 21

Figure 2 .Figure 3 .
Figure 2. Effect of preventive primary care interventions for older people on healthcare use (n = 13).Data in the final column in brackets are 95% confidence intervals.aHigh and unclear risk of bias in domains assessed are marked using the conventional style: -for high risk of bias and ?for unclear risk (>1 -or ?reflects bias in >1 domain).1P = provider-focused domain only.2P = provider-and practice-focused domains for van Hout et al (2010);48 provider-and patient-focused domains for Blom et al (2016). 283P = includes patient-, provider-, and practice-focused intervention components.%, LTC home = proportion admitted to long-term care home.dur = duration.ed = education.F = financial components.IRR = incidence rate ratio.MDT = multidisciplinary team.OR = odds ratio.

ResearchFigure 4 .
Figure 4. Effect of preventive primary care interventions for older people on quality of life (n = 15): a) single item and index scores; b) multidomain scores.Data in the final column in brackets are 95% confidence intervals.a High and unclear risk of bias in domains assessed are marked using the conventional style: -for high risk of bias and ?for unclear risk (>1 -or ?reflects bias in >1 domain).1P = provider-focused domain only.2P = provider-and practice-focused domains for van Hout et al (2010); 48 provider-and patientfocused domains for Blom et al (2016). 283P = includes patient-, provider-, and practice-focused intervention components.Cantril = Cantril's Ladder.CASP-19 = Control, Autonomy, Self-Realization and Pleasure scale.d = SMD, standardised mean difference.ed = education.EQ-5D = EuroQOL five dimensions summary index for quality of life.F = financial components.MDT = multidisciplinary team.OR = odds ratio.SF = Short Form Health Survey.SRH = self-rated health.WHOQOL-BREF = brief version of the World Health Organization Quality of Life assessment tool.1P = provider-focused domain only.2P = provider-and practice-focused domains for van Hout et al (2010); 35 provider-and patient-focused domains for Blom et al (2016). 28

Table 1 . Characteristics of studies included in the review (n = 18) Characteristic Studies, n a
a Data are n unless otherwise specified.b Multiple classification allowed.IQR = interquartile range.

28 (0.09 to 0.48)
The effect of community-based health management on the health of the elderly: a randomized controlled trial from China.BMC Health Serv Res 2012; 12: 449.30.Dapp U, Anders JA, von Renteln-Kruse W, et al.A randomized trial of effects of health risk appraisal combined with group sessions or home visits on preventive behaviors in older adults.J Gerontol A Biol Sci Med Sci 2011; 66(5): 591-598.31.Friedman B, Wamsley BR, Liebel DV, et al. et al.Pre frail 80: multifactorial intervention to prevent progression of prefrailty to frailty in the elderly.J Nutr Health Aging 2018; 22(10): 1266-1274.33.Godwin M, Gadag V, Pike A, et al.A randomized controlled trial of the effect of an intensive 1-year care management program on measures of health status in independent, community-living old elderly: the Eldercare project.Fam Pract 2016; 33(1): 37-41.34.Hogg W, Lemelin J, Dahrouge S, et al.Randomized controlled trial of anticipatory and preventive multidisciplinary team care: for complex patients in a community-based primary care setting.Can Fam Physician 2009; 55(12): e76-85.35.Hoogendijk EO, van der Horst HE, van de Ven PM, et al.Effectiveness of a Geriatric Care Model for frail older adults in primary care: results from a stepped wedge cluster randomized trial.Eur J Intern Med 2016; 28: Peterson MJ, Pieper CF, et al.The Veterans Learning to Improve Fitness and Function in Elders Study: a randomized trial of primary care-based physical activity counseling for older men. .Neumann L, Dapp U, von Renteln-Kruse W, Minder CE.Health promotion and preventive care intervention for older community-dwelling people: long-term effects of a randomised controlled trial (RCT) within the LUCAS cohort.J Nutr Health Aging 2017; 21(9): 1016-1023.43.Ploeg J, Brazil K, Hutchison B, et al.Effect of preventive primary care outreach on health related quality of life among older adults at risk of functional decline: randomised controlled trial.BMJ 2010; 340: c1480.Effects of preventive home visits by district nurses on self-reported health of 75-year-olds.Prim Health Care Res Dev 2016; 17(1): 56-71.