Trial design
The COR-PRIM study was a prospective, randomised, parallel, and single centre study (NCT01462799) involving 157 patients with CHD in a primary care setting in south-eastern Sweden. The COR-PRIM study tested the hypothesis that one year of problem-based patient education improves a patients’ empowerment and self-efficacy, to change self-care significantly compared to one year of standardised home-sent patient information. Recruitment began in November 2011 and tests were performed at baseline[24] and one, three, and five years, which will be finalised in 2020, after completion of a one-year problem-based learning programme for patients with CHD. Here, we report only data from baseline and one-year follow-up, which was finally collected in 2015.
Patients
According to the COR-PRIM study,[18] the following inclusion criteria were used: (i) CHD verified by myocardial infarction and/or Percutaneous Coronary Intervention (PCI) and/or coronary artery by-pass surgery (CABG) within 12 months before planned start of the intervention irrespective of age; (ii) stable cardiac condition; (iii) optimised cardiac medication not substantially changed the previous month; and (iv) completed hospital heart school at one of the identified six primary health care centres. The following exclusion criteria were used: (i) planned CABG or other conditions demanding continuing cardiologic care such as on-going contact with heart failure clinic; (ii) life-expectancy ≤ 1 year; (iii) documented psychiatric disease impeding cooperation with other people; (iv) obvious abuse of alcohol or narcotics; and (v) inability to read or communicate in Swedish.
Randomisation and procedures
During a follow-up at hospital, nurses informed patients about the study and those patients who agreed to receive further information were contacted by the principal researcher by letter and telephone. Some patients voluntarily stated their reasons for not participating in the study – e.g., difficulties leaving work/home; long distance to the primary health care centres; and disapproval of group activity.
The patients who agreed to participate, returned informed consent document and who had completed the baseline questionnaires were randomised (1:1 ratio) to either the experimental group (received problem-based learning) or the control group (received patient information leaflets delivered to the home address) before any intervention. The randomisation was carried out with sealed unmarked opaque envelopes and were assigned by an administrator in a room separated from the research and intervention area. We used a block of 18 study numbers that were blindly allocated to either problem-based learning or home-sent patient information[25]. The envelopes contained a unique number starting from number 1 that were hand-picked by an administrator who was blinded during this procedure.
Conventional care and interventions
All patients were offered a brief conventional cardiac rehabilitation programme at an outpatient clinic in hospital care that included counselling visits with a nurse and a cardiologist about 1 month and 6-12 months after discharge respectively; physical exercise 1-2 times per week, for 3-4 months and diet counselling. Additionally, patients were offered a day long heart school primarily focussing on CHD, medication, physical exercise, and diet. If the patient’s condition was stable, he or she was referred to a general practitioner in primary care. The conventional care provided and the design of the interventions described below and in a previous paper[18]. Here, we provide a short description of the interventions.
Home-sent patient information group
The patients in the home-sent patient information group served as controls. After receiving conventional care, 6-9 patients per group met directly after randomisation in the primary health care centres to discuss self-care goals and follow-ups during the study year. Predetermined written patient information was provided[23] at this meeting to support self-care as suggested in brochures produced, for example, by The Swedish Heart and Lung Foundation. Next, the patient information was mailed to the patient’s home address at the same times as the problem-based meetings (explained below) during the study year. Finally, after one year of intervention the patients were invited to a focus-group interview to share their beliefs about their performance of self-care; experiences of the study materials and participating in the study.
Problem-based learning intervention group
The patients in the experimental group (6-9 patients/group) started the one-year problem-based learning intervention at a primary health care centre. Each meeting was for two hours. The problem-based intervention was completed one year after start. The goal was to improve self-care by strengthening patient empowerment with a focus on understanding cardiac symptoms e.g. angina pectoris, swelling legs and dyspnoea, medications, and the health benefits associated with lifestyle changes regarding diet, physical activity, and mental health including depression, anxiety and fear. Nurses who were trained to tutor the patients in the problem-based learning process[26] were fundamental to this study. The nurses took part in a training session for two days given by the project team and later the nurses were tutored monthly by the authors, AKK and PT to discuss and develop their work. The training session included learning about tutoring, in problem-based learning regarding self-directed learning and problem-solving, to help patients to formulate issues and self-care goals. The patients used scenarios as triggers e.g. pictures, texts and concreate materials as a starting point for learning during the meetings. Moreover, the nurses supported the patients to choose learning materials and challenged patients to choose evidence-based literature. Resource professionals (e.g., physician and dietician) were invited to discuss questions not solved by the patients. Patients’ relatives were also invited to the meetings. During the final meeting, follow-up focus-group interviews were performed to collect data about patients’ beliefs about their performance of self-care and about their experiences of participating in the study. For further descriptions, see Kärner et al[18].
Outcomes
The primary outcome was patient empowerment to reach self-care goals one year after randomisation. The questionnaires used to assess the primary and secondary outcomes are briefly presented here and more thoroughly elsewhere[24]. Patient empowerment was assessed using the Swedish-Coronary Empowerment Scale 10 which was developed to survey patient empowerment in patients with CHD. Swedish-Coronary Empowerment Scale 10, based on the Swedish Diabetes Empowerment Scale 23, is a valid and reliable tool for assessing patient empowerment in patients with diabetes mellitus. The scale consists of four empowerment subscales: goal achievement, self-awareness, stress management and readiness to change. The Cronbach's α‐coefficient for the total Swedish Diabetes Empowerment Scale 23 ranged from 0.68 to 0.91[27-29]. The Swedish Diabetes Empowerment Scale 10, a shortened version of the Swedish Diabetes Empowerment Scale 23, was also found to be reliable with the Cronbach's α‐coefficient value α = 0.84. After securing approval from the author of the scale, we replaced the word ‘diabetes’ with ‘coronary heart disease’. Swedish-Coronary Empowerment Scale 10 has four subscales: 1) Goal achievement and overcoming barriers to goal achievement; 2) Self-awareness; 3) Managing stress; and 4) Assessing dissatisfaction and readiness to change. The items are scored on a five-point Likert scale ranging from strongly agree (1) to strongly disagree (5). A higher value means a stronger patient empowerment[30]. Secondary outcomes were self-efficacy in general,[31] healthy diet,[32] and physical exercise[33, 34].Self-efficacy was assessed using the General Self-efficacy Scale, which uses a four-point Likert scale ranging from not at all true (1) to exactly true (4). A higher score indicates a higher general self- efficacy. The General Self-efficacy Scale has been confirmed to be highly reliable, stable, and valid. The internal consistency of the scale was excellent with the Cronbach's α‐coefficient value α = 0.88[35, 36]. Healthy diet was assessed using the Nutrition Self-efficacy Scale and physical exercise using the Physical Exercise Self-efficacy Scale, both which use a four-point Likert scale ranging from very uncertain (1) to very certain (4). A higher score indicates a higher Nutrition/Physical Exercise self-efficacy. The Nutrition Self-efficacy Scale and the Physical Exercise Self-efficacy Scale are assessed to be reliable and valid tools, with internal consistency values were α = 0.87 and 0.88 respectively[32]. Physical exercise was also assessed using Stages of Change Scale[33].Well-being[37] was assessed using the Cantril Ladder of Life, a single-item indicator with a ladder of steps numbered from zero at the bottom to 10 at the top. Zero means the worst possible life, and 10 the best. The patients also answer on which step they stood one year ago, one which step they stand at present and they are asked to predict on which step they will stand one year in the future. Cantril’s Ladder of Life is used in large populations and validity and test-retest coefficients of 0.70 have been reported [38] . This ladder has also been used in patients ≥ 65 years recovering from an acute coronary event [39]. The EuroQoL 5-dimensions scale,[40] is a reliable and valid tool for use in patients with CHD. The Cronbach's α‐coefficient value of the EuroQoL 5-dimensions scale indicated an acceptable internal consistency with a value of α = 0.73. Discriminative validity analyses have confirmed that this scale distinguished well between patient groups with a different age, gender, or educational level. Self-rated health was measured by a Visual Analogue Scale within the EuroQoL 5-dimensions scale. This scale makes scores of 0-100, with higher scores indicating a better overall quality of life [41]. The Visual analogue Scale is easy to administer and produce stable intraclass correlations score (0.79) showing acceptable reliability, and satisfactory validity in patients with acute coronary syndrome[42]. Blood pressure, BMI, waist size, and blood tests were used at follow-up to measure effects of self-care. Primary and all secondary outcome data were included in the intention-to-treat analysis, which means that [43] all randomised patients in the groups which they were randomly assigned to were included in the analysis, regardless from deviation from protocol.
Statistical analysis
Sample size justification
Sample size calculation was based on the estimation that patients with diabetes mellitus, also a life-long disease, who reported poor patient empowerment scored on average 3.0 of the Swedish Diabetes Empowerment Scale Those who scored 3.6 or more were considered reporting good patient empowerment [29]. The difference between those reporting poor and good patient empowerment (0.6) was considered as a clinically relevant estimation of effect size and has been used for sample size calculation. Accordingly, at a significance level of α = 0.05 and a power of 1-β = 0.80 this generated the required sample size in each group of at least 63 patients. We recruited 157 patients to compensate for withdrawals.
Primary analysis: intention-to-treat
The intention-to-treat analysis included patients who fulfilled all inclusion criteria. Missing values in the primary outcome The Swedish-Coronary Empowerment Scale 10 have been substituted by the typical mean for the sample. Continuous data are presented as means +SD or as median and interquartile range. Between-group differences were tested using independent-sample Student’s t-test for numerical variables or Mann-Whitney U test for non-normal distributed variables. For categorical variables, Fisher’s exact test was used. All statistical tests were two-sided with a significant level of P ≤ 0.05. The data were analysed using SPSS (IBM® SPSS Statistics, Version 23).