Nurse-led, telephone-based secondary preventive follow-up benefits stroke/TIA patients with low education: a randomized controlled trial sub-study

This was an open randomized (1:1) controlled trial sub-study and a post-hoc analysis.

The primary outcome was the mean difference in SBP between participants with high and low education (as defined below) 12 months after hospital discharge. We analyzed the following variables as secondary outcomes: the mean differences in diastolic blood pressure (DBP) and LDL-C levels between education groups at 12 months and changes in the SBP, DBP, and LDL-C levels between baseline and 12 months within each education group. We analyzed all outcome variables separately for the intervention and control groups. BP was measured in the sitting position after 5 min of rest. LDL-C was calculated from serum concentrations of cholesterol and fasting triglycerides by using the Friedewald formula. Based on a power calculation, study groups of 180 participants—standard deviation (SD) 19 mm Hg, mean SBP 140 versus 135 mm Hg, two-tailed alpha 0.05, and power 80%—were necessary to reliably detect a difference of 5 mm Hg in the mean SBP.

The study included participants from the NAILED stroke risk factor trial who remained in the trial and had measurement data available from the 12-month follow-up (n = 771). Recruitment for the NAILED stroke risk factor trial occurred at Östersund Hospital, which is the only hospital in the county of Jämtland, Sweden. In 2010, the county of Jämtland had 126,691 inhabitants, 35.0% of whom lived in the centrally located city of Östersund, 23.0% in villages of 500–4000 inhabitants, and the remaining 42.0% in the more sparsely populated surroundings. The mean age of the county’s inhabitants was 43.1 years and 50.0% were female. Among inhabitants 65 years or older, 46.3% had completed no more than 9–10 years of school and the remaining 53.7% had completed higher levels of education [25].

In Sweden, public inpatient and outpatient health care is provided to all citizens at a low cost but is not entirely free of charge. There is also a national system for stepwise subsidizing of medical drugs [26]. During the inclusion period, the hospital stroke unit generally treated admitted stroke and TIA patients, or another hospital care unit did if no bed was available in the stroke unit. The treating physicians initiated secondary preventive treatment in-hospital and, in some cases, followed by a return visit to the hospital out-patient clinic. Thereafter, primary health care was the principal responsibility of the secondary preventive follow-up. In primary health care, patients with a previous stroke or TIA are not routinely called for regular control of BP and blood lipids unless they are also diabetic. Medical prescriptions are renewed at the patient’s request and patients are welcome to book an appointment with a doctor if they want to discuss secondary preventive treatment further. At many primary health-care centers, patients can also visit at any time during opening hours for a self-measured BP check-up; results are later reviewed by a nurse or physician or both.

Since 1962, compulsory schooling in Sweden has consisted of 9–10 years of elementary school followed by 2–4 years of optional upper secondary school/vocational school leading to a vocational degree or qualifying for college and university studies. All levels of public school are free of charge. During the preceding three decades, the school system was more complex; compulsory school ranged from 6 to 8 years and was free of charge, but qualification for upper secondary school and university required graduation from a partly parallel school system. Altogether, 9–10 years of school was required to proceed to upper secondary school and thereafter university [27].

The NAILED stroke risk factor trial is an open, population-based, randomized controlled trial testing the hypothesis that nurse-led, telephone-based secondary preventive follow-up is more effective than standard-care follow-up in regard to achieving improved modifiable risk factor levels after stroke or TIA. The trial inclusion period was from January 1, 2010, to December 31, 2013, and follow-up is ongoing. Mooe et al. previously published the methods of the NAILED stroke risk factor trial [23]. Briefly, the trial recruited consecutive patients admitted to Östersund Hospital because of stroke or TIA (Fig. 1). Patients who were physically and cognitively able to participate in the study follow-up provided informed consent prior to inclusion. Exclusion criteria were participation in concurrent trials and/or an inability to participate because of impaired hearing, aphasia, cognitive impairment, or severe, often terminal, disease. After random assignment (allocation ratio 1:1; allocation sequence computer-generated in blocks of four; stratified for sex and the degree of disability according to the modified Rankin Scale, mRS), participants underwent nurse-led, telephone-based follow-up (intervention) or follow-up in accordance with local standard procedure (that is, mainly within primary health care (control)). A study nurse telephoned all participants for follow-up at 1 and 12 months after hospital discharge. The study nurses made repeated attempts to get in contact with participants who did not answer the first telephone call. Before each follow-up occasion, the participants had their BP measured and a blood sample for lipids taken at their closest health-care facility. Intervention follow-up included information about measurement results, lifestyle counseling (physical activity, diet, and smoking cessation), and assessment of pharmacological treatment. Assessments of lipid lowering treatment were restricted to participants with ischemic stroke or TIA. Participants in the intervention group who did not meet treatment targets for BP (<140/90 mm Hg) or LDL-C (<2.5/1.8 mmol/L) underwent pharmacological titration with repeated measurement and adjustment every 4 weeks until they achieved the treatment targets or they could achieve no further improvement. A study physician consulted on all pharmacological adjustments. Owing to an update to the local guidelines, the LDL-C treatment target for patients with diabetes changed during the course of the study. Consequently, for participants with diabetic mellitus who had their 1-month follow-up after March 31, 2013, (n = 20), the treatment target was LDL-C less than 1.8 mmol/L, but it was less than 2.5 mmol/L for the rest of the population.

Telephone contact with participants in the control group did not include any lifestyle counseling or changes to pharmacological treatment. BP and LDL-C measurement results were forwarded to the patient’s general practitioner for assessment.

All data presented in this report were from the NAILED stroke risk factor trial. The NAILED study team collected patient characteristics at baseline, including age, education level, functional level according to the mRS, cardiovascular risk factors, and medical history, in-hospital through patient interviews and a review of medical records. The study team also measured weight and height to calculate body mass index (BMI).

We used a dichotomized classification of education level. We defined low education as not more than 10 years of formal education and high education as the completion of more than 10 years of formal education.

The qualifying events, prior vascular events, and comorbid conditions refer to diagnoses made by clinical physicians. Stroke included both ischemic and hemorrhagic events with the exception of subarachnoid hemorrhage. We defined prior ischemic heart disease as including previous acute myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting or a combination of these.

The proportion of participants considered to have a high education level decreased gradually with increasing age (Fig. 2). The pattern was the same for both men and women (data not shown). Consequently, participants with low education were older than participants with high education (Table 1). Compared with participants with high education, those with low education were more likely to be included in the study because of stroke and a higher proportion had an mRS score of at least 3. In addition, we found an association between low education and a higher prevalence of comorbid conditions and cardiovascular risk factors, including previous diagnosis of ischemic heart disease, previous stroke, congestive heart failure, hypertension, and diabetes mellitus.

Baseline SBP and LDL-C levels measured 1 month after hospital discharge did not differ according to education level (Table 2). Pharmacological treatment at the same time point is shown in Table 1. Treatment with antihypertensives was more common among participants with low education, whereas the proportions prescribed warfarin, antiplatelet drugs, or lipid-lowering agents were not different.

The association between education level and BP and LDL-C levels differed according to secondary preventive follow-up and age group. At 12 months, controls with low education had higher SBP than controls with high education (mean difference 4.2 mm Hg, 95% confidence interval (CI) 0.8 to 7.6; Table 2). After stratification for age, a significant difference in SBP between the high- and low-education subgroups remained only for those not more than 70 years of age (Fig. 3). In the intervention group, we found no association between education level and BP at 12 months (Table 2), regardless of age group (Fig. 3).

In regard to the change in BP between 1 and 12 months, SBP and DBP improved in the intervention group regardless of education level (Table 2). Mean SBP improved in controls with high education, but we found no change among controls with low education.

LDL-C at 12 months did not differ according to education level in the control group (Table 2), regardless of age group. However, we found an association between low education and an increase in LDL-C between 1 and 12 months (Table 2). We observed the opposite situation in the intervention group; an association of low education with lower LDL-C at 12 months (mean difference 0.2, 95% CI 0.1 to 0.4) resulted from a decrease in LDL-C in the low-education subgroup between 1 and 12 months (Table 2). After stratification for age, the difference at 12 months was significant only for participants in the intervention group not more than 70 years of age (Fig. 3). Notably, we found similar LDL-C levels across all groups with high education at 12 months and levels did not improve among highly educated participants during the first year, regardless of secondary preventive follow-up (Table 2 and Fig. 3).

Adjusting for sex, BMI, type of qualifying event, medical history, and functional level according to mRS score did not change the education level differences in SBP, DBP, and LDL-C levels at 12 months (data not shown).