Community-Based Intervention to Improve Cardiometabolic Targets in Patients With Stroke
A Randomized Controlled Trial
Background and Purpose—Many guidelines for secondary prevention of stroke focus on controlling cardiometabolic risk factors. We investigated the effectiveness of a management program for attaining cardiometabolic targets in survivors of stroke/transient ischemic attack.
Methods—Randomized controlled trial of survivors of stroke/transient ischemic attack aged ≥18 years. General practices were randomized to usual care (control) or an intervention comprising specialist review of care plans and nurse education in addition to usual care. The outcome is attainment of pre-defined cardiometabolic targets based on Australian guidelines. Multivariable regression was undertaken to determine efficacy and identify factors associated with attaining targets.
Results—Overall, 283 subjects were randomized to the intervention and 280 to controls. Although we found no between-group difference in overall cardiometabolic targets achieved at 12 months, the intervention group more often achieved control of low-density lipoprotein cholesterol (odds ratio, 1.97; 95% confidence interval, 1.18–3.29) than controls. At 24 months, no between-group differences were observed. Medication adherence was ≥80% at follow-up, but uptake of lifestyle/behavioral habits was poor. Older age, being male, being married/living with partner, and having greater functional ability or a history of diabetes mellitus were associated with attaining targets.
Conclusions—The intervention in this largely negative trial only had a detectable effect on attaining target for lipids but not for other factors at 12 months or any factor at 24 months. This limited effect may be attributable to inadequate uptake of behavioral/lifestyle interventions, highlighting the need for new or better approaches to achieve meaningful behavioral change.
Clinical Trial Registration—URL: http://www.clinicaltrials.gov. Unique identifier: ACTRN12608000166370.
Survivors of stroke or transient ischemic attack (TIA) are at great risk of having severe or fatal secondary vascular events.1,2 Intensive management of modifiable cardiometabolic risk factors could reduce the risk of recurrent vascular events by up to 80% for 10 years.3 Major cardiometabolic factors include blood pressure (BP), lipids, blood glucose, blood creatinine, body weight, and smoking.4 Proven strategies for controlling cardiometabolic factors include pharmacological and lifestyle/behavioral interventions.5 Therefore, recommendations in guidelines largely focus on these strategies to achieve cardiometabolic targets (eg, BP <120/80 mm Hg).5,6
Cardiometabolic targets are often difficult to attain in clinical practice. For example, in highly advanced and accessible healthcare settings in Europe, targets were poorly achieved in at-risk individuals with diabetes mellitus7 and high BP.8 Similarly, attainment of targets is often poor in people with stroke.9 In Australia, there has been significant progress in the uptake of therapies for secondary prevention of stroke.10 However, it is unclear whether this translates to attaining cardiometabolic targets.
In Australia and countries with similar healthcare systems, primary care providers are well placed to implement strategies for managing chronic diseases, such as diabetes mellitus and hypertension.11 Strategies include ongoing management of risk factors and monitoring adherence to treatment.11 However, there is limited evidence for the effectiveness of these strategies in the management of stroke. We investigated the effectiveness of an individualized management program for improving attainment of cardiometabolic targets, based on Australian guidelines, in community-dwelling survivors of stroke/TIA. We hypothesized that survivors of stroke/TIA who receive an individualized management program will have better attainment of cardiometabolic targets at 12 and 24 months than those undergoing usual care.
Trial Design and Subjects
The STANDFIRM (Shared Team Approach between Nurses and Doctors for Improved Risk factor Management) is a cluster randomized controlled trial of secondary prevention in people with stroke/TIA. The full trial protocol, including sample size calculation, has been published previously.12 Briefly, subjects were recruited from 4 tertiary hospitals in Melbourne, Australia, between January 2010 and December 2015. Adults aged ≥18 years were eligible if they were hospitalized for stroke/TIA and were living within 50 km of recruitment hospitals. We excluded people participating in another trial, admitted from/discharged to a nursing home, or with a rapidly deteriorating health condition. Potentially eligible subjects were identified during hospitalization by research nurses and stroke physicians and informed about the trial. Final consent was obtained at a baseline in-home visit (median 10 weeks post-discharge). Ethics approval was obtained (HREC 2011000331).
Randomization and Blinding
Subjects were randomized into study groups, using a computer-generated, blocked procedure, stratified by recruitment hospital. This method ensured equal allocation to groups within each hospital. Randomization was clustered by general practice to reduce between-group contamination. As part of the consent process, subjects were informed that they would receive treatment A or B with no details about what each entailed to avoid response bias. Similarly, outcome assessors, specialists, and general practitioners (GPs) were blinded to treatment assignment.
All subjects continued to receive the usual care provided in their general practices and stroke prevention clinic of participating hospitals.
To supplement usual care, subjects in the intervention group received a management program, comprising a standard evidence-based care plan, and 3 education sessions. The care plan was individualized to the risk factor profile of subjects and contained clear goals/targets for managing cardiometabolic risk factors (see Table 1; online-only Data Supplement), based on recommendations in Australian guidelines.5
After a comprehensive blinded assessment of risk factors at baseline, an initial care plan was developed by an unblinded intervention nurse and reviewed by independent stroke specialists. For example, in the case illustrated in the care plan provided in the online-only Data Supplement, the nurse identified problems with attaining targets for BP and lipids (page 1) based on data obtained at the baseline assessment. The specialist then checked these recommendations, made amendments if required, and provided treatment recommendations for the GP to facilitate the subject’s care.
Before sending the management plan to the GP, the intervention nurse undertook in-home visits to discuss tailored behavioral interventions and provide advice on identified care problems/needs; advice included benefits of healthy lifestyle and adherence to medications. The nurse also discussed treatment goals and targets highlighted in the care plan. Subjects were asked if there were any potential barriers to the uptake of healthy lifestyle strategies/adherence to medications. Using a standard education syllabus (see online-only Data Supplement), the nurse provided tailored information to support self-management in overcoming any identified barriers.13 Finally, the nurse organized appointments for subjects to discuss and agree on their care plan with their GPs.
The format of our care plan is similar to those routinely used for treatment of diabetes mellitus, hypertension, and other chronic conditions and so is familiar to GPs. The care plan complies with the Australian Medicare insurance scheme, and so GPs are reimbursed when used. Reimbursement is greater than standard consultations, thereby providing an incentive to use our care plan.
The process of blinded assessment, preparation of care plan, education, and arrangement of GP appointment was repeated at 3 and 12 months. At the 3- and 12-month education sessions, the intervention nurse discussed barriers/enablers to adhering to treatment targets set at earlier visits. The care plan was again revised and sent to GPs at 6 and 18 months based on information provided by subjects during telephone interviews (described below).
Baseline and Follow-Up Assessments
Details of stroke and demographic information were obtained from hospital records. Other baseline data were obtained at in-home nurse visits (median 10-week post-discharge). During baseline visits, subjects underwent blinded, standardized anthropometry, biochemical tests, and assessment of BP, disability, mood disorder, and lifestyle/behavioral habits. Assessments were repeated at 3, 12, and 24 months while brief telephone interviews were conducted at 6 and 18 months to obtain self-reported data on risk factors.
The study outcome was attainment of pre-defined targets for 6 cardiometabolic factors at 12 and 24 months. This was based on recommendations in clinical guidelines and reference values recommended in pathology guidelines in Australia (Table 1).
Baseline characteristics were compared using χ2 test (categorical variables) and Kruskal–Wallis test (continuous variables). Within-group differences in target attainment were estimated using McNemar test (categorical variables) and Wilcoxon signed-rank test (continuous variables). For between-group changes (intention-to-treat analyses), multivariable logistic (categorical variables) and Poisson (count variables) regression models were undertaken (see online-only Data Supplement for details). The robustness of the effect estimates were ascertained in sensitivity analyses, including complete case analyses and analyses restricted to subjects whose risk factor targets were not optimal at baseline.
To determine factors associated with number of targets attained at 12 and 24 months, multivariable Poisson regression models were constructed using methods similar to those stated above, except that we did not adjust for the baseline status for target attainment. All analyses were conducted using STATA IC 12.1 (StataCorp, 2012), with a 2 tailed P≤0.05.
Study Subjects and Baseline Characteristics
Between January 2010 and November 2013, 5633 patients were assessed for eligibility, of whom 2516 (45%) were eligible (Figure). Among those eligible, 563 (22%) were enrolled: 283 to the intervention and 280 to usual care.14 When compared with subjects who declined participation, those enrolled were more often men (64% versus 55%) and less often aged ≥65 years (63% versus 80%) or discharged to inpatient rehabilitation from acute care (32% versus 43%; P<0.001). Subjects were randomized at median 73 (Q1: 54, Q3: 97) days post-discharge. Baseline characteristics were similar between groups.14 Overall, the median age was 70.1 (Q1: 60.9, Q3: 78.6) years, 65% were men, and 78% had an ischemic stroke. At baseline, ≥80% of subjects were prescribed recommended medications (Table 2). Adherence to recommended lifestyle habits was high for abstaining from smoking (78%) and healthy drinking (95%) but poor for being physically active (12%) and daily consumption of vegetables (4%), fruit (47%), or salt (3%; Table 2).
Follow-up was complete in 533 subjects (95%) at 12 months and 485 (86%) at 24 months (Figure).
Uptake of Secondary Prevention Strategies
There was no significant between-group difference in uptake of pharmacological and behavioral/lifestyle interventions at 12 and 24 months (Table 2). There were consistent declines in the proportion of controls prescribed lipid-lowering therapy and controls who were physically active, from baseline to 24 months (Table 2). Results were similar in analyses excluding deaths and losses to follow-up at any time point (Table II in the online-only Data Supplement). Approximately 10% of subjects who were advised to take prevention medications reported difficulties in following their prescriptions at 12 months. In contrast, 87% of subjects reported difficulties in carrying out advice on diet modification and 82% on improving exercise habits.
In both intervention and control groups, there was no detectable within-group difference in the median number of cardiometabolic targets achieved at 12 and 24 months relative to baseline (Table 3). For between-group analyses, we found no detectable difference in total number of cardiometabolic targets achieved at 12 and 24 months in both univariable (adjusting for baseline status for target attainment) and multivariable models (further adjusting for potential confounding factors; Table 4).
For individual cardiometabolic factors, there were declines in the proportion of controls achieving targets for BP, total cholesterol, and low-density lipoprotein cholesterol at 12 and 24 months relative to baseline (Table 3). No other within-group difference was observed at 12 and 24 months relative to baseline. In between-group multivariable analyses adjusting for baseline status for target attainment and potential confounding factors, subjects in the intervention group were more likely than controls to achieve targets for the control of low-density lipoprotein cholesterol (<2.5 mmol/L; odds ratio, 1.97; 95% confidence interval [CI], 1.18–3.29) and glycemia (hemoglobin A1c ≤ 7%; odds ratio, 4.44; 95% CI, 1.31–15.08) at 12 months (Table 4). No between-group differences were observed in the proportion of subjects attaining targets for other factors at 12 months or for any factor at 24 months. Findings of the intention-to-treat analyses (Table 4) were consistent with those of complete case analyses (Table III in online-only Data Supplement). For analyses restricted to subjects whose targets were not optimal at baseline, at 12 and 24 months, subjects in the intervention group were more likely than controls to achieve targets for the control of lipids, including total and low-density lipoprotein cholesterol (Table IV in online-only Data Supplement).
Factors independently associated with greater attainment of cardiometabolic targets at 12 months (Table 5) included being aged ≥65 years (incidence rate ratio [IRR], 1.06; 95% CI, 1.01–1.13), being men (IRR, 1.08; 95% CI, 1.02–1.14), and having greater functional ability (IRR, 1.30; 95% CI, 1.06–1.59). In contrast, having a history of diabetes mellitus was associated with poor attainment of targets at 12 months (IRR, 0.91; 95% CI, 0.84–0.98). These findings were consistent at 24 months, except for the identification of 1 additional factor, being married/living with a partner, which was associated with greater attainment of targets (IRR, 1.07; 95% CI, 1.01–1.14).
Our comprehensive care planning improved attainment of target for the control of lipids (low-density lipoprotein cholesterol <2.5 mmol/L) in survivors of stroke/TIA but not for other factors at 12 months or any factor at 24 months. Although our intervention appeared to improve the attainment of targets for the control of glycemia (hemoglobin A1c ≤7%), this finding may have been biased by the small sample of subjects whose hemoglobin A1c measures were suboptimal.
In the secondary prevention of stroke, interventions are targeted at improving adherence to treatment and changing lifestyle/behavioral habits.5,6 Pharmacological interventions appeared to be the preferred option for modifying risk in our subjects because a large proportion of controls (≥80%) were discharged on recommended medications and few reported difficulties in taking their medications. Prescription of medications remained high in controls throughout follow-up. These findings may even be underestimated as we did not account for subjects who were not prescribed medications because of valid clinical reasons.
In contrast to the use of medications, uptake of recommended risk-modifying behaviors was generally poor in our cohort and may partially explain the undetectable between-group difference in attaining targets. It was particularly surprising that our robust intervention was unsuccessful in a population of highly motivated and relatively less disabled survivors of stroke/TIA, with relatively greater potential to engage in healthy behavior. This finding highlights the difficulty in changing lifestyle/behavioral habits after a stroke/TIA. Our finding also highlights a clear need to modify existing approaches or develop new approaches to appropriately inform practice on the management of stroke.
Similar to our findings, others have reported limited or no benefit of interventions designed to promote behavioral change in people with stroke.15 Some authors have proposed computer-assisted strategies founded on social cognitive theory to address identified barriers to behavioral change and incorporate standard/evidence-based parameters (ie, dose, length, frequency, setting, and mode of intervention).16 Another potentially effective strategy is incorporating adoption of behavioral interventions as an important performance indicator in the management of stroke in general practice.11
In this study, we identified subgroups of people with stroke/TIA who could benefit from more targeted interventions, that is, women, people aged <65 years, those single/living alone, and those with greater functional disability, poor educational attainment, or a history of diabetes mellitus. Vascular risk factors are particularly difficult to control in the presence of diabetes mellitus as shown in a large population-based study conducted in Spain. In this study of people with diabetes mellitus, only 11% met target for managing weight, 28% for BP, and 33% for cholesterol.7 These findings, and ours, highlight the need for more aggressive treatment of vascular risk factors in survivors of stroke with diabetes mellitus.
Our study may be limited by the relatively low level of evidence for some recommendations in the guidelines for the management of stroke in Australia.5 However, these guidelines are often used by clinicians to make decisions when treating people with stroke/TIA. Another potential limitation is the possibility of incomplete blinding of subjects to treatment assignment. Incomplete blinding would have been greatly minimized by the fact that subjects were not actively informed about what the additional intervention entailed, as well as the fact that our intervention was already embedded in usual practice in Australia. Furthermore, our findings may not be generalized to the wider stroke population because only 22% of eligible subjects were enrolled. This potential selection bias was also highlighted by the differences between subjects enrolled and those who declined participation.
The strength of our study is the careful and extensive assessment of targets for cardiometabolic factors using best available evidence.5 However, our findings should be interpreted in the context that our sample appeared to be well treated and have less severe stroke, as reflected by large proportions of subjects on recommended medications or with no significant disability.14
In conclusion, in this largely negative trial, our intervention improved attainment of targets for the control of lipids in survivors of stroke/TIA after a 12-month follow-up but no detectable differences for other risk factors at 12 months or any factor at 24 months. This poor attainment of treatment targets is likely attributable to lack of improvement in adopting recommended risk-modifying lifestyle/behaviors, despite our comprehensive intervention. Thus, more effective strategies are needed to enhance uptake of lifestyle interventions in the secondary prevention of stroke, and this might be best targeted at those subgroups identified as most likely to benefit from more targeted interventions.
We appreciate the hard work of the research nurses.
Sources of Funding
This work was supported by National Health and Medical Research Council (586605, 1042600, 1064517, and 1063761).
Dr Phan received honoraria for presentations given for Bayer/Boehringer Ingelheim/Genzyme/Pfizer/Bristol-Myers Squibb. Dr Nelson is an advisory board member for Amgen and Dr Gerraty for Astra Zeneca. The other authors report no conflicts.
Presented in part at the American Heart Association/American Stroke Association Epidemiology and Lifestyle Scientific Meeting, Phoenix, AZ, March 1–4, 2016.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.117.017499/-/DC1.
- Received March 29, 2017.
- Revision received July 3, 2017.
- Accepted July 6, 2017.
- © 2017 American Heart Association, Inc.
- Hardie K,
- Hankey GJ,
- Jamrozik K,
- Broadhurst RJ,
- Anderson C
- Touzé E,
- Varenne O,
- Chatellier G,
- Peyrard S,
- Rothwell PM,
- Mas JL
- Hackam DG,
- Spence JD
- O’Donnell MJ,
- Xavier D,
- Liu L,
- Zhang H,
- Chin SL,
- Rao-Melacini P,
- et al
- 5.↵National Stroke Foundation – Australia. Clinical guidelines for stroke management 2010. https://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/cp126.pdf. Accessed May 8, 2016.
- Kernan WN,
- Ovbiagele B,
- Black HR,
- Bravata DM,
- Chimowitz MI,
- Ezekowitz MD,
- et al
- Navarro-Vidal B,
- Banegas JR,
- León-Muñoz LM,
- Rodríguez-Artalejo F,
- Graciani A
- Shah NS,
- Huffman MD,
- Ning H,
- Lloyd-Jones DM
- Cadilhac DA,
- Lannin NA,
- Anderson CS,
- Andrew N,
- Kim J,
- Kilkenny M,
- et al
- Kim J,
- Thrift AG,
- Nelson MR,
- Bladin CF,
- Cadilhac DA
- Olaiya MT,
- Joosup K,
- Nelson MR,
- Srikanth VK,
- Bladin CF,
- Gerraty RP,
- et al
- Salinas J,
- Schwamm LH