Gender differences in the factors predicting initial engagement at cardiac rehabilitation

Introduction

Attendance at and completion of cardiac rehabilitation (CR) programmes is poor worldwide. In the UK, recent analysis of National Audit of Cardiac Rehabilitation (NACR) data shows that uptake of CR among eligible patients is currently 50%.1 Although this places the UK in the top 2% of countries in Europe,2 uptake still remains below national recommendations of 65%–70%.1

The CR pathway of care for patients following a cardiac event involves six stages (figure 1), each of which is vital for the achievement of meaningful clinical outcomes.3 Barriers to engagement, attendance and adherence within the pathway have been widely studied and shown to include patient-level factors (eg, illness perception, beliefs about treatment, social support, family responsibilities, work constraints); service-level factors (eg, programme accessibility, travel time, referral); sociodemographic factors (eg, older age, female, ethnic minority, low education levels, comorbidities); and psychological factors (eg, depression, anxiety).4–9

A range of strategies to increase the number of people participating in CR have been developed with these barriers in mind, such as motivational communications by nurse liaisons, therapists or peers; early appointments after discharge; gender-tailored CR; and intermediary rehabilitation programmes for older people.10 Although there is currently only weak evidence that these interventions are effective at improving participation, tailored approaches which aim to address social factors and patient-identified barriers have been recommended as the most likely to yield benefit.6 10

It is increasingly recognised that a ‘one-size-fits-all’ model will not be effective in the future,11 and that CR needs to be ‘rebranded and re-invigorated’ as a more tailored, person-centred intervention in order to reach a larger patient population.12 Gender-tailored CR interventions have been recommended as holding particular promise for improving uptake.13 14 A recurrent theme across qualitative studies of gender and CR experience is that women and men hold divergent views on their rehabilitation needs and their preferences on how exercise, group interaction and emotional support aspects of programmes are delivered.15 Gender has been shown to be a key variable in self-management decisions and preferences in a range of long-term conditions, including whether to attend CR-related support interventions.16 However, where gender-related barriers and solutions to CR attendance have been considered in the extant literature, women have tended to be the focus. Women’s lesser participation in CR programmes is widely recognised and has been extensively reviewed,15 17–21 with multilevel barriers including non-referral, lower education level, lack of social support and high burden of family responsibilities cited as key factors associated with poorer uptake.13 17 18 20 22–26

To date, research on the factors associated with men’s participation at CR has received little attention. Although men are more likely to take up CR than women, and be included in trials of CR effectiveness,27 male participation also remains suboptimal.1 More can be done to optimise uptake in both men and women if interventions to improve participation can be designed to address each group’s specific barriers. To this end, we undertook an analysis to determine whether there are gender differences in the patient-level and service-level factors that predict CR engagement, defined as attendance at the initial CR baseline assessment following referral. Our hypothesis was that the factors that predict men’s engagement with CR are different from those which are associated with women’s engagement.

Methods

Results

The analysis sample included 95 638 patients. The baseline characteristics of male and female participants (both engaged and not engaged in CR) are presented in table 2. Figure 2 presents a flow diagram of the patients within study time period (those who were referred to CR with an acceptable completeness of data) and the gender split of included patients.

Table 2 shows the differences between those patients receiving an initial assessment and those who did not. The total number of patients included in the regression analysis was 81 938 (59 232 male and 22 706 female); 13 700 were excluded. The regression model population included 45 086 patients who were processed using multiple imputation to account for missing values. The final models for each gender are shown in the online supplementary material.

Table 3 presents the results from the regression. This shows that for all included predictors, the relationship between predictors and likelihood for engagement is the same for both men and women. The results from the regression presented an OR of 0.983–0.986 for 1-year increase in age. If we consider a 10-year increase this effect becomes 14%–17% less likely to attend, so a patient who is 65 years old in comparison to a 55-year-old patient, accounting for other factors, is 14%–17% less likely to achieve engagement.

If patients are from a higher socioeconomic area, as presented as quintile IMD, they are increasingly more likely to initially engage with CR. If patients identify as non-single, they are 29%–47% more likely to attend engagement than those who are single (men—OR partner 1.332 and previous partner 1.295, p<0.05; women—OR partner 1.329 and previous partner 1.470, p<0.05). There were no significant interactions between the factors included in the regression for either gender.

Discussion

Gender differences in cardiovascular disease incidence, treatment, risk factor management and rehabilitation are observed worldwide.29 The need for better understanding and response to these differences is widely acknowledged.30 To our knowledge, the current study is the first quantitative investigation into gender differences in patient-level and service-level determinants of initial engagement with CR after referral (attendance at an initial baseline assessment). Our analysis has shown that there are no important differences in the factors that predict men’s and women’s likelihood of initial engagement with CR.

Early engagement with the CR pathway is vital for ensuring programme uptake and adherence and the achievement of meaningful clinical outcomes.1 3 Attendance at an initial assessment enables personalised goals to be identified and a tailored care plan to be agreed that meets individual needs, participation preferences and choices.1 3 Our findings provide a steer to health professionals on which patients are less likely to attend an initial assessment so that strategies aiming to optimise engagement can be adapted and tailored.31 The baseline characteristics of our nationally representative cohort of patients showed engagement with CR to be suboptimal in both men (66.5%) and women (62.4%). Regression analysis identified seven factors associated with a reduced probability of engagement, and eight factors associated with increased probability.

Consistent with the wider literature on factors associated with CR participation,4–9 32 increasing age, being South Asian or non-white ethnicity (other than Black) and being single were all associated with a reduced likelihood of attending an initial assessment in both men and women. Age was coded as continuous and showed a 10-year increase in age resulted in 14%–17% reduced likelihood of engagement in both genders.

Other factors identified as significant predictors of non-engagement in both men and women included having hypercholesterolaemia/dyslipidaemia (OR 0.78 men/0.79 women) and no history of a cardiac event (OR 0.83 men/0.78 women). If patients were from a lower socioeconomic area, as presented as quintile IMD, they were also increasingly less likely to engage with CR, with each group from 10%–13.8% going from 1 to 2 and 81%–82% from 1 to 5 (p<0.05).

Factors including having a partner, referral by general practitioner and having a PCI, CABG or other treatment were all predictive of attendance at initial assessment in both men and women. We identified only one factor where a significant difference between men and women was observed. Having diabetes was associated with an 11% reduced likelihood of engagement in men; an association that was not identified in women and the factor was removed from the final model. Previous qualitative research has found that men and women tend to report similar reasons for initial non-participation in CR32 and our findings corroborate this. Other than diabetes, it was striking that all factors identified as predicting the likelihood for engagement/non-engagement in the current study were the same for both men and women.

Aligned with the extensive literature on non-participation,4–9 28 the overwhelmingly important determinant of engagement that was observed in both men and women in the current study was receiving an invitation to attend an assessment session (OR 4.223 men/4.033 women; p<0.05). This suggests that efforts to enhance initial engagement should primarily focus on ensuring all patients receive an invitation to attend an assessment, with particular attention paid to patients who are single, older, non-English speakers and from lower socioeconomic groups.

Gender differences in the factors influencing attendance at and adherence to CR are well established in the literature,6 and targeted and tailored interventions to promote access have been recommended on this basis.15 17–19 33 We therefore hypothesised that there would be gender differences in the factors predicting initial engagement with CR after referral. The unexpected finding that there are no important gender differences at this important initial stage of the rehabilitation pathway suggests that a more nuanced view of how gender interplays with CR participation is required.

The decision to attend CR is multifactorial and intertwined with social contexts that afford inequitable opportunities for access.15 The current study adds weight to findings from recent systematic reviews and meta-analyses which indicate gender by itself may not always be the main determinant of CR uptake,34 and differences between men and women may diverge over the period of rehabilitation.35 This suggests that gender-specific strategies aiming to improve participation may be appropriate at some stages of the patient care pathway (eg, improving adherence) but not all (eg, promoting initial engagement). Further research into gender differences in determinants of CR attendance and adherence is warranted to fully determine this.

Conclusions

This is the first study into gender differences in the predictors of CR engagement, providing new insights into the factors that lead men and women to attend their initial CR baseline assessment (CR engagement) in a nationally representative cohort of patients. The key findings from the study have shown that there are no important differences in the factors that predict men’s and women’s likelihood of initial engagement with CR. Consideration of gender differences in predictors of CR uptake should probably be more nuanced and informed by the stage of the patient care pathway.