This study was an online vignette experiment on patients’ intention-to-change-lifestyle, using a 2-by‑2 between-subjects design. Following recommendations on conducting vignette studies [11], four written vignettes were developed to represent variations in source (cardiologist, physiotherapist) and framing (gain, loss) of brief lifestyle advice. This study protocol was approved by the Psychology Research Ethics Committee of Leiden University (2020-06-05‑A.W.M. Evers-V1-2474).

The study population consisted of members of the patient panel of Harteraad, the Dutch patient association for CVD, a voluntary database of approximately 2,600 CVD patients. Harteraad sent their panel members a survey invitation on 29 June 2020, with a two-week response period. The invitation described the study’s purpose, duration, participation procedure, data protection, and the survey link (2020 Qualtrics, Provo, UT). Informed consent was obtained on the first page of the survey. Inclusion criteria were Dutch-speaking adults (≥ 18 years old). With a priori power analysis (two-sided, α = 0.05, power = 0.80) [12], 128 participants were required to detect an effect of source and framing on intention-to-change-lifestyle with a small effect size (f = 0.25). To account for 10% of dropouts, we aimed for at least 141 participants in total.

The survey first assessed patients’ characteristics (age, sex, education level, last doctor’s visit, current lifestyle score, motivation and self-efficacy concerning lifestyle change). Afterwards, patients were instructed to imagine having a consultation with their healthcare provider [11]. Each patient was then randomly assigned one of the four vignettes (see Fig. 1) which described brief lifestyle advice communicated by a cardiologist or physiotherapist (message source), using a gain-framed or loss-framed perspective on the health-related future (message framing). Vignettes (see App. A) had similar word counts, contained supporting images, and allowed reading the message aloud [11]. Afterwards, the vignettes’ applicability and meaningfulness were assessed, together with the primary outcome measure intention-to-change-lifestyle [13, 14]. Finally, used as manipulation check, patients were asked if they recalled which messenger delivered the brief lifestyle advice and what message was emphasised. After completing the survey, a debriefing explained the study’s purpose.

Design of the study. CVD cardiovascular disease

Similar to Taylor and colleagues (2005) [15], primary outcome measure intention-to-change-lifestyle was assessed using two items, “I want to change my lifestyle or continue to maintain my healthy lifestyle.” and “I intend to change my lifestyle or continue to maintain my healthy lifestyle.”, answered via a 5-point Likert scale, ranging from Completely disagree (1) to Completely agree (5) [16].

Based on previous research [17], motivation, self-efficacy, and current lifestyle score were assessed with the questions “At this moment, how motivated do you feel to change your lifestyle or maintain your healthy lifestyle?”, “At this moment, how confident are you that you will be able to change your lifestyle or maintain your healthy lifestyle?”, and “At this moment, how would you rate your current lifestyle?”. Variables were rated on a 0–10 visual analogue scale (VAS) [18], representing a continuum between Very unmotivated/No confidence at all/Very unhealthy (0), and Very motivated/A lot of confidence/Very healthy (10). Perceived applicability and perceived meaningfulness of the vignette were assessed with the question, “Given your personal situation, how would you rate the conversation with the healthcare provider as just described?”. Variables were rated on a 0–10 VAS, representing a continuum between Not applicable at all/Not meaningful at all (0) and Very applicable/Very meaningful (10). Recall of message source was measured by asking “Which messenger were you speaking to?” with the answer options: Nurse, Cardiologist, Physiotherapist, or Mental health counsellor. Recall of message framing was measured by asking “What message did the healthcare provider emphasise?” with the answer options: a) “Positive effects of a healthy lifestyle” or b) “Negative consequences of an unhealthy lifestyle”.

Data were analysed using IBM SPSS Statistics V.25. As the primary outcome intention-to-change-lifestyle was skewed, inverse transformation of intention was used. Two-way analysis of covariance (ANCOVA) examined the effect of source and framing on intention while controlling for sex, education level, motivation, and self-efficacy [19,20,21,22]. Bonferroni post hoc analyses were used for multiple comparisons. A p-value of < 0.05 was considered statistically significant.

Descriptive statistics were used for patient characteristics. Internal reliability was checked by calculating the scale score and Cronbach’s alpha for the intention scale, consisting of both intention items (α = 0.897). Pearson’s correlations were used to assess the strength and direction of the relationship between the control variables education level, motivation, and self-efficacy, and the dependent variable intention-to-change-lifestyle. An independent t‑test was used to assess the relationship between the control variable sex and dependent variable intention-to-change-lifestyle. Chi-squared tests were used to assess differences in recall between message source and types of message framing. Independent t‑tests were used to investigate differences in perceived applicability and meaningfulness of the vignettes for both sources and types of framing. Data are reported as n (%), mean (standard deviation), or mean (95% confidence interval).