Management of cardiovascular disease risk in people with comorbid mental illness: A cross sectional study in Australian general practice

Psychological distress is associated with increased incidence of cardiovascular disease (CVD) and can negatively impact self-management for CVD risk factor control (Hare et al., 2014). For people with mental illness, prevention services, monitoring and treatment for chronic medical conditions, including CVD, may be suboptimal compared with the general population (Firth et al., 2019). Reasons include mental health needs taking priority over physical needs, increased complexity from managing multiple comorbidities, differential rates of access and care utilisation, prescribing medications that may increase CVD risk and a greater influence of social determinants of health (Firth et al., 2019; Kurdyak and Gnam, 2004).

Australian studies have identified the high prevalence of CVD risk factors and CVD-related premature death in people with chronic (Davidson et al., 2001) and severe (Spooner et al., 2022) mental illness, and thus the need for proactive risk monitoring and reduction. However, few studies have assessed whether cardiovascular care varies in the presence of co-morbid mental illness. We aimed to assess whether people with or at elevated risk of CVD and living with mental illness differed in (1) adherence to guideline-recommended blood pressure (BP) and lipid-lowering medications, (2) attainment of guideline-recommended BP and cholesterol targets, and (3) use of allied health services, when compared with those without mental illness.

We undertook a secondary analysis of baseline data from the Consumer Navigation of Electronic Cardiovascular Tools randomised control trial (RCT, n = 934) conducted in 24 primary care health services in Sydney, Australia (Redfern et al., 2020). Enrolled participants were at elevated risk of CVD (59.0%) or had diagnosed CVD (41.0%). Mental illness was defined as either self-reported disorder(s) at enrolment, or having been prescribed a medication for mental illness in the previous 12 months. We used Medicare Benefits Schedule (MBS) and Pharmaceutical Benefits Schedule (PBS) data obtained from Services Australia, and routinely collected data from the trial database. Adherence was defined as a proportion of days covered (PDC) of ⩾80% based on PBS dispensing data. Guideline-based targets were defined as BP ⩽ 130/80 mmHg for people with CVD, diabetes or albuminuria and ⩽ 140/90 mmHg for all others and low-density lipoprotein (LDL) cholesterol <2.0mmol/l. Service utilisation variables were MBS item numbers related to general practitioner (GP) consultations, GP management plans (GPMP) and allied health service use. Primary ethical approval for the RCT was from the Human Research Ethics Committee of the University of Sydney (Reference 2013/716). Participants provided written informed consent for data linkage.

Twenty two percent (22.3%, 208/934) of participants had a mental illness: 159/208 (76.4%) based on self-report only, 18/208 (8.7%) based on medication records, and the remainder (31/208, 14.9%) based on both these sources. Few people (13/208, 6.2%) had a major mental illness; the remainder had depression or anxiety. Compared to those without a mental illness, people with a mental illness were younger (mean age 65.9 vs 68.0 years), had lower education (33.2% vs 26.8% secondary school or lower), lower household income (30.8% vs 18.9% less than $800 per week) and more comorbidities, notably stroke (14.9% vs 7.7%) and respiratory illnesses (36% vs 15.6%). A difference in gender was noted between those with and without a mental illness (65.4% and 79.9% respectively were male, p < 0.0001).

Prescription rates in the previous 24 months were higher for people with a mental illness compared to those without for BP medications (42.8% vs 33.2% respectively, p = 0.01) and non-significantly higher for lipid-lowering drugs (27.9% vs 22.2% respectively, p = 0.09). Adherence rates to guideline-recommended medications were low overall and there were no differences between those with and without a mental illness for BP medications (50.0% vs 50.7%, p = 0.86 respectively with a PDC ⩾80%) and lipid lowering medications (38.5% vs 36.0%, p = 0.51 respectively with a PDC ⩾ 80%). There were no differences in meeting guideline targets for people with and without a mental illness (39.4% vs 38.3%, p = 0.77 respectively for BP targets; 29.3% vs 27.1%, p = 0.53 for LDL targets). The proportion of people meeting both BP and LDL targets concurrently was low overall (9.1% for those with a mental illness and 11.2% for those without, p = 0.41).

Median GP visits were marginally higher in those with mental illness compared to those without (16 visits vs 14 visits, p < 0.01 in the previous 24 months). There was greater use of psychology services in those with a mental illness compared to those without (12.6% vs 1.3% respectively, p < 0.01). There were no significant differences in use of other allied health services between those with and without mental illness. There was no difference between participants with and without mental illness for receipt of (49.5% vs 45.8% respectively, p = 0.36) or review of (25.3% vs 20.2% respectively, p = 0.13) a GPMP.

Previous research has noted that provision of and adherence to recommended treatments for cardiometabolic risk factors is inequitable for those with mental illness (Firth et al., 2019; Mitchell et al., 2012). Our study did not find any evidence of this; indeed, there was some evidence of increased guideline-recommended prescribing for BP medications. People with a mental illness were younger, had higher rates of comorbidities and increased markers of socio-economic disadvantage. Despite these disparities, the lack of difference in CVD management is encouraging, although it should be stressed that these management gaps are large across the population (Hespe et al., 2020) and may disproportionately affect those with mental illness (Firth et al., 2019). People with mental illness had slightly higher GP visit frequency overall, a greater use of GP management plans and used subsidised psychological services more than those without a mental illness. Increases in service utilisation may mitigate against disparities in CVD risk factor management for people with a mental illness, but this association requires further investigation, ideally with a larger longitudinal cohort.

Study strengths include the use of administrative data to assess adherence rates and care utilisation, rather than relying on self-report, and the use of two data sources to identify a mental illness. Study limitations include (1) data on private mental health and allied health services not claimed under MBS were unavailable and this may underestimate overall care utilisation and (2) given the cross-sectional nature of the study, we did not discern the timing of the onset of mental illness and CVD medication initiation. Given mental illness is highly dynamic, it is possible that adherence rates vary considerably depending on illness acuity. Third, our sample was predominantly people with depression and anxiety; there were insufficient numbers of participants with severe mental illness to look at CVD management in this group with higher risk factor prevalence (Spooner et al., 2022). The type and severity of mental illness is known to influence behaviours such as smoking rates and weight management, and specific treatments such as antipsychotic medication can themselves increase cardiometabolic disease risks (Kinley et al., 2015).

Few Australian primary care studies have assessed the impact of comorbid mental illness on CVD risk factor management or care utilisation. Although we hypothesised that those with a mental illness would have lower rates of medication adherence and of meeting management targets compared to those without mental illness, this was not observed. Increased access to GP, psychological and other health care services may be factors mitigating against inferior management practices. Further research is needed to elucidate the complex associations between comorbid mental illness and risk factor management for people with or at elevated risk of CVD.