The evaluation study of the EFFICHRONIC project was a prospective cohort study with a 6-month pre-post single-group design and was conducted between January 2018 and November 2020 in five European countries [the Netherlands, Italy, the United Kingdom (UK), Spain, and France]. The EFFICHRONIC project aimed to evaluate the benefits of a CDSMP in managing and maintaining the health of citizens with a low socioeconomic status (SES) and one or more chronic conditions, as well as their caregivers (20). For full details of the study design and protocol, see Tan et al. (20). There were no major deviations from the published protocol study. ISRCTN registry number is ISRCTN70517103 and the date of registration was 20 June 2018. Participant data were collected before the start of the first workshop session at baseline and 6-month follow-up. Ethical approval was provided by the human research ethics board of the study sites (20). All participants provided written informed consent.

Participants were recruited to participate in a CDSMP intervention in their local study site (Occitanie region in France, Genoa province in Italy, Rotterdam region in the Netherlands, Principality of Asturias in Spain, and a region of London in the UK). Citizens were recruited through clinicians, public events and announcements, local patient or volunteer organizations, and community advocates. Recruitment sites were chosen based on their location in distinct environments. Vulnerability maps were constructed in three study sites (Occitanie region, Genoa province, and Principality of Asturias) based on EUROSTAT's NUTS-3 level geographical areas, in which the prevalence of the target population was high (21). For complete recruitment strategies, see Alvarez Rosete et al. (22).

The intervention targeted a vulnerable population of citizens of ≥18 years and with a low SES with one or more chronic conditions as specified in the International Classification of Primary Care (ICPC-2) present for at least 6 months (23). Caregivers of the participants with a chronic condition were also included. Citizens were only eligible to participate if they were able to comprehend the information provided in the local language and were likely to complete the 6-month study duration. Citizens were not eligible to participate when they were experiencing a crisis period, their basic housing needs were not met, they were diagnosed with severe mental health problems, or they suffered from active addictive disorders or cognitive impairment.

The CDSMP is a six-session (2.5 h weekly) community-based intervention built on the self-efficacy theory (10, 24, 25), a well-established program developed by Stanford University and assessed for over 20 years by the Self-Management Resource Center (24, 26). Each group session was led by trained (public) health professionals and trained laypersons in groups of up to 20 participants in the local language (27). Session leaders were trained by certified trainers, following the Stanford methodology (26). Topics included in the sessions were as follows: an overview of self-management and chronic health conditions, making an action plan, relaxation and management of cognitive symptoms, problem-solving, emotions (anger, fear, and/or frustration), fitness, fatigue management, healthy eating, advance directives, communication, medication, making treatment decisions, depression, informing the healthcare team, and working with healthcare professionals (12). The program focused on problem-solving, decision-making, confidence building, management of emotions, positive health and efficient communication to strengthen self-efficacy (i.e., the confidence in one's ability to accomplish a specific task or reach a goal), and managing different aspects of one's health functioning (25).

Outcomes were obtained through self-report questionnaires at the start of the first session (baseline) and 6-month follow-up. Caregivers filled out the questionnaires by themselves but not for the individuals with a chronic condition they cared for. The degree of self-efficacy was measured with the Self-Efficacy for Managing Chronic Diseases 6-item scale (SEMCD-6), with scores ranging from 6 to 60, with higher scores corresponding to higher self-efficacy (28). The following health behaviors were assessed: (1) dietary habits: two items on the intake of fruit and vegetables; (2) physical activity: six items on physical exercise (12) and one item on sedentary behavior: International Physical Activity Questionnaire (IPAQ) (29); (3) substance use: current smoking, yes/no; frequency of alcohol use, one item from the Alcohol Use Disorders Identification Test (AUDIT-C) (30); and (4) sleep and fatigue: visual analog scale (VAS), with scores ranging from 0 to 10, with higher values indicating worse sleep/more fatigue. Depression severity was assessed with the Patient Health Questionnaire 8-item scale (PHQ-8), with scores ranging from 0 to 24, with higher values indicating a higher severity, score ≥10 corresponding to current depression (31). Health-related quality of life (HR-QoL) was assessed with the 12-item Short-Form Health Survey (SF-12), with scores ranging from 0 to 100; the EuroQol-5 Dimensions-5 levels (EQ-5D-5L), using the United Kingdom value sets, with scores ranging from 0 to 1, with higher values indicating better health utility (32, 33); and EQ-VAS, with scores ranging from 0 to 100, with higher values indicating better HR-QoL.

Health responsibilities were assessed as follows: (1) communication with healthcare professionals: three items on preparing a list of questions, asking questions, and discussing personal problems (12); (2) health literacy: two items on the Health Literacy Questionnaire (HLQ) (34); (3) healthcare utilization in the past 6 months was assessed with four items on the number of doctor visits, the number of accident and emergency department visits, number of overnight stays in the hospital, and the total number of nights spent in the hospital; (4) medication adherence: six items from the Simplified Medication Adherence Questionnaire (SMAQ) (35); and (5) perceived medical errors: three items on the understandability of a healthcare professional's explanation and prevalence of perceived medical errors from the American Association of Retired Persons (AARP) “survey beyond 50.09” questionnaire (36).

Sociodemographic characteristics were assessed using self-report questionnaires: age, sex, household composition, educational level, income, migration background, employment situation, housing situation, social relationships, and social support. The participant's household composition, income, housing situation, social relationships, and social support were measured by an adapted version of Gijón's Social-Familial Evaluation Scale (SFES) (37). The subjective improvement in the most important outcomes of the CDSMP intervention experienced by the participants was evaluated with seven items at the 6-month follow-up: change in doing at least one activity for health; not letting health problems control their life; the ability to make decisions; the ability to express themselves; in the way of communication with family, friends, and others; confidence in the health system understanding their needs; and satisfaction with the intervention as a whole.

Participant characteristics were described using mean (SD) or number of participants (%) for the total study sample. Participant's sociodemographic characteristics and health outcomes were evaluated at baseline after stratification for chronic disease status. Caregivers with a chronic condition were added to the “chronic condition group,” and caregivers without a chronic condition were added to the “no chronic condition group.” T-tests were used to compare the means for continuous variables and Pearson's chi-squared tests for categorical variables. To assess the effects of the intervention on continuous outcome measures, linear regression analyses were conducted with a change in the outcome measure between baseline and follow-up as a dependent variable and the beta of the intercept indicating the effect. For dichotomous outcome measures, paired McNemar's tests were used. Outcome analyses were stratified by chronic disease status. In addition, stratified linear regression analyses were run for country, educational level, sex, and age (age <65 years or ≥65 years). Supplementary Tables 2B–E considering 26 outcome measures, the two-sided significance threshold, after Bonferroni correction for multiple testing, was set at a P-value = 0.05/26 = 0.0019. In the protocol study, two outcome measures were considered in five countries. Analyses were conducted with SPSS version 25.0 (IBM SPSS Statistics for Windows, IBM Corp., Armonk, NY, USA).

Participants were recruited for the project between January 2018 and March 2020; 2,759 participants who fulfilled the inclusion criteria provided informed consent and started the intervention. Of the 1,693 participants who filled in the baseline questionnaire and had an available chronic disease status, 1,377 (81.3%) were citizens with a chronic condition, and 316 (18.7%) were caregivers. Participants were included in the five European study sites: the Netherlands (n = 388), Italy (n = 331), the United Kingdom (n = 345), Spain (n = 568), and France (n = 212). Table 1 shows the baseline characteristics of the study sample stratified by chronic disease status, i.e., having a chronic condition or being a caregiver without a chronic condition. Supplementary Table 1A presents the baseline characteristics of the total study population, and Supplementary Table 1B describes the baseline characteristics at the country level. Compared with caregivers without a chronic condition (the no chronic condition group), citizens with a chronic condition were, on average, less often female, reported more current depression, more often lived alone, had a lower educational status, had a lower income, less often had a migration background, and less often worked.

Of all included participants (n = 2,759), 2,277 (82.5%) attended at least four out of six sessions of the CDSMP intervention. Attending ≥4 out of six group sessions was defined as “good adherence” (38). A total of 1,844 participants (66.8% out of 2,759) took part in the baseline questionnaire, and 1,252 participants took part in the follow-up questionnaire. Four participants only took part in the follow-up questionnaire. The number of participants that completed both baseline and follow-up was 1,248. Overall, 1,152 out of 1,844 (62.5%) participants took part in both baseline and 6-month follow-up questionnaires, attended ≥4 sessions, had an available chronic disease status, and were included as the study sample for further analyses. The baseline characteristics of the 692 excluded participants were compared with the included participants. On average, the excluded participants were younger, more often female, more often a current smoker, and spent less time on aerobic physical activity; there were also significant differences in country, education level, income, and housing adaptation (see Supplementary Table 3). A flow chart of the participants is shown in Figure 1.

Table 2 presents the outcome scores of the 1,152 participants who completed both the baseline and follow-up questionnaire, attended ≥4 sessions, and had an available chronic disease status. Among participants with a chronic condition (n = 951), there were significant improvements at follow-up compared with baseline (all P < 0.002) in self-efficacy (SEMCD-6 6.7 vs. 6.4), PHQ depression scores (6.3 vs. 7.0), physical and mental HR-QoL (PCS 42.3 vs. 40.2, MCS 42.8 vs. 41.4), health utility (EQ-5D-5L 0.88 vs. 0.86), and self-rated overall health (EQ-5D-5L 67.2 vs. 63.9). In addition, a significant improvement was shown in communication with healthcare providers (2.28 vs. 2.11) and understanding information (3.10 vs. 3.02). Furthermore, within the healthcare utilization, the number of doctor visits (3.61 vs. 4.97), Accident and Emergency department visits (0.25 vs. 0.48), and total nights spent in a hospital (0.65 vs. 1.13), all measured over the last 6 months, significantly diminished. Finally, the percentage of perceived medical errors (19.6 vs. 28.7%) decreased. No significant changes were shown in dietary habits, physical activity, substance use, and sleep and fatigue. In addition, finding health information, the number of overnight hospital visits, medication adherence, perceived unclear communication with the doctor, and perceived experienced medical errors as a problem did not change.

Among caregivers without a chronic condition (n = 201), only a significant decrease in doctor visits was reported (1.54 vs. 2.25, P < 0.001). Supplementary Table 2A presents the effects of the EFFICHRONIC intervention on the total study population, and Supplementary Tables 2B–E presents the effects of the EFFICHRONIC intervention per country, educational level, sex, and age group. To determine the relevance of the statistically significant effects of the intervention, the differences in outcome scores before and after the intervention were compared with the minimal clinically important differences (MCIDs) of the respective outcome measures (39). The differences in the effect size of the depression score (PHQ-8), HR-QoL score (SF-12), health utility score (EQ-5D-5L), and overall health score (EQ-VAS) did not meet the MCIDs of these parameters (40–42). No MCIDs were available for the other parameters.

Participants who completed the questions on experienced outcome changes at follow-up (n = 1,248) reported an improvement in doing at least one activity for health (85.2%), not letting health problems control their life (80.3%), ability to make decisions (56.0%), ability to express themselves (49.3%), communication with family, friends, and others (51.1%), and confidence in the healthcare system understanding their needs (45.1%) (Supplementary Table 4). The average satisfaction score with the intervention was 8.3 ± 1.7 on a scale from 0 (lowest) to 10 (highest) (Supplementary Table 4). No adverse events were reported for the intervention.

First, a lack of a control group prevents the outcomes from being linked to the intervention since non-specific effects related to group-based intervention participation cannot be ruled out. The outcome measures were self-reported; therefore, a bias in the outcome estimates cannot be excluded and is probably less precise than objectively measured outcomes. Furthermore, the presence of selection or social desirability bias cannot be ruled out; we only have data on the sociodemographic characteristics of the participants who completed the baseline questionnaire, which makes it hard to infer the implications of our study. The intervention was targeted at a vulnerable population; however, we did not achieve in including those who are most vulnerable, for example, people who were not able to speak the local language were not able to participate. In addition, despite all efforts, many participants did not fill in one or both questionnaires and were thus excluded from analyses, resulting in a high attrition rate.

There were differences in sociodemographic variables and lifestyle factors between dropped-out participants and participants who completed the follow-up, which might have caused attrition bias. In addition, participants in the study sites were recruited in different ways. On the one hand, the heterogeneity of the study population might make it harder to infer the effects of the intervention. On the other hand, this heterogeneous study population with diverse backgrounds might enable the outcomes to reflect the general population more. As discussed above, a CDSMP intervention with a 6-week duration might have been too short for changing health behaviors since that is a complex matter, which needs lasting attention, support, and practice. Finally, the follow-up time was 6 months, which measured only “mid-term” effects.

It might be worthwhile to conduct a meta-analysis of chronic disease self-management studies to assess the overall effectiveness and determining factors, such as duration, content, and setting. We also recommend a longer follow-up of behavioral and/or lifestyle intervention studies to assess the durability of any observed outcomes and to include objective measures, such as blood pressure, blood glucose, and BMI. Efforts should be made to involve vulnerable groups in chronic disease management programs to address health disparities. Communities in “Blue zones,” worldwide geographical regions where people live longer and healthier lives than the average, show that active engagement with social surroundings, a sense of belonging and purpose in life, plays a vital role in chronic disease prevention (63).