Ambulatory sensitive hospitalisations among people accessing mental health and addiction services: a retrospective cross-sectional study using national population data

People experiencing mental health and substance use conditions have a higher likelihood of physical health problems and mortality than the general population.^1,2 This disparity arises from multiple interlinking factors including healthcare access and quality, stigma and discrimination and wider socio-economic determinants of health.^1,2 Research shows lower rates of access to screening, prevention and treatment for non-communicable diseases such as cardiovascular, metabolic and respiratory conditions for people with diagnosed mental health and addiction (MHA) conditions.^1,3 Siloed physical and mental health services may also impede people’s access to appropriate support.¹ Such barriers can lead to complications and hospitalisations for otherwise preventable conditions.

Ambulatory sensitive hospitalisations (ASHs) are hospital admissions for conditions potentially preventable via timely and effective routine or preventative primary (ambulatory) care.⁴ In Aotearoa New Zealand, there are 28 ambulatory care–sensitive conditions defined for adults based on the International Classification of Diseases 11th revision (ICD-11) including hypertension, myocardial infarction and complications of established conditions like diabetes.⁵

High ASH rates are linked with poor primary healthcare access and quality, even when controlling for socio-demographic factors and health status.^4,6 This supports the use of ASH rates as a broad indicator of healthcare access (i.e., the adequate supply of appropriate, timely services). Understanding the conditions underlying these hospitalisations is also important for preventative actions.

In Aotearoa, ASH rates are reported annually for people aged 0–4 and 45–64 to monitor primary care access and quality.⁵ To date, no research has examined ASH rates among MHA service users in Aotearoa. Given evidence of higher ASH rates and psychological distress among Māori and Pacific peoples, a focus on ethnic disparities among MHA service users is warranted.^7,8

This study aims to better understand and describe ASH rates for people accessing specialist MHA services in Aotearoa. It aims to:

1.      identify common ASH conditions for people accessing MHA services, and

2.      compare ASH rates for MHA service users to the total population.

Method

Data

This study uses routinely collected data from the Programme for Integration of Mental Health Data (PRIMHD) for specialist MHA service use and the National Minimum Dataset (NMDS) for hospital admissions provided by the Ministry of Health – Manatū Hauora. PRIMHD and NMDS datasets were linked using an encrypted version of the National Health Index (NHI) database to create an anonymised dataset with unique study identifiers.

Published ASH data for the whole population aged 45–64 by ethnic group reported by the Ministry of Health were used for the comparison group.⁵

Analyses follow a repeated cross-sectional design across 1 July 2012–30 June 2018.

MHA cohort

Service use is a proxy to identify people experiencing MHA issues. The MHA cohort was drawn from a larger predefined cohort of people who had at least three in-person contacts with MHA services between 1 July 2001 and 30 June 2018. People with primary diagnoses of non-MHA conditions (e.g., dementia, intellectual disabilities, developmental disorders), with no other MHA diagnosis and people with gender recorded as “other” (n=18) or “unknown” (n=76) were excluded from this cohort.

For this study, MHA cohorts were created for each year of whole-population hospitalisation ASH data between 1 July 2012 and 30 June 2018 to enable comparison to whole-population data. Individuals were included if they:

• were aged 45–64 within the study period, and
• had at least one in-person specialist MHA service contact in the prior 3 years.

Figure 1 illustrates how the MHA cohort was selected.

View Figure 1, Table 1–4.

Comparison population

The total Aotearoa population is the comparison group, retrieved from published ASH data for people aged 45–64 for each financial year (July–June) in the study period, stratified by 5-year age groups and prioritised ethnicity. Stratified reported numbers for each year were pooled over the 6-year study period. Rates were calculated and pooled using annual Statistics New Zealand population projections.

Outcome variable

The main outcome variable is ASH rate per 100,000 person-years. ASH events for the MHA cohort are defined as:

• primary hospital diagnosis matching ICD-11 ambulatory sensitive condition codes;⁵
• occurring within 2 calendar years of MHA service contact; and
• occurring between 1 July 2012 and 30 June 2018.

For example, an individual with MHA service contact in July 2012 would have any ASH event between then and 2015 included.

Person-years were calculated using the total number of participants multiplied by the number of study years each person met inclusion criteria. This reflects the number of at-risk years each person contributes data for. For example, one individual aged 60 in July 2012 with ongoing MHA contact within the study period contributes 5 person-years (until they turn 65).

Rates of ASH admissions for specific conditions were also examined and categorised using ICD-11 chapters.⁵

Other variables

Gender, age and ethnicity were sourced from NHI data on 14 September 2021.

Age at 31 December for each fiscal year was categorised into groups: 45–49, 50–54, 55–59 and 60–64.

Gender was categorised as male or female based on NHI data.

Ethnic group as recorded on the NHI database was used. Only prioritised ethnicity was available (prioritised in order of Māori, Pacific, and non-Māori non-Pacific [nMnP]), where each person is allocated to a single group even if they identify with multiple. People with missing ethnicity data are included in “nMnP”. This method for analysis purposes does not assume the allocated ethnic group is the one people identify most strongly with.

The New Zealand Index of Deprivation 2018 (NZDep18) was used to categorise area-level deprivation, based on matching of address data in the NHI database to the national deprivation index for classification based on the 2018 Census.

Data analysis

Descriptive analyses

ASH rates are calculated for each sub-group (e.g., age, ethnicity) using total ASH events as the numerator and total person-years as the denominator.

Regression analyses

Negative binomial regression was used to examine the associations between MHA service use and ASH rates compared to the total population.

Adjusted rate ratios (ARR; 95% confidence interval [CI]) adjust for the potentially confounding effects of age and ethnicity. ARRs greater than 1 indicate an increased relative likelihood. It was not possible to adjust for deprivation as published ASH rate data were not available broken down by NZDep2018.

Analyses were performed in R version 4.0.3, using RStudio version 1.4.1103 with MASS, tidyverse, Hmisc and sjPlot packages.

Ethical approval

Ethical approval was obtained from the University of Otago Human Ethics Committee as part of a wider project entitled “Meeting physical health care needs of people with mental illness or addiction”, ethics committee reference number HD20/080.

Results

Participants

The MHA cohort (Table 1) comprised 89,547 people, reflecting 287,709 person-years.

There were 33,570 ASH events recorded for the MHA cohort between 2012 and 2018. One-sixth (16.5%) had at least one recorded ASH within this period; among them, 41.0% had at least two. The number of ASH events per individual ranged between zero and 80.

ASH rates for non-nMnP ethnic groups (predominantly New Zealand European) were lower compared to Māori and Pacific peoples. ASH rates were higher for older people and people living in more deprived areas.

Common ASH conditions

Table 2 presents primary diagnoses recorded for the MHA cohort’s ASH events. Over half were attributed to cardiometabolic and respiratory conditions: primarily chronic obstructive pulmonary disease (COPD), angina and chest pain.

Comparing MHA cohort and total population ASH rates

Table 3 shows pooled total population ASH rates alongside MHA cohort rates. In total, 268,871 ASH admissions occurred in Aotearoa between July 2012 and June 2018. Across both groups, ASH rates were lower for nMnP ethnic groups and increased with age. After adjusting for age and ethnicity, the MHA cohort’s overall ASH rate was 2.38 times higher than the total population’s.

Table 4 compares condition-specific ASH rates between the MHA cohort and total population. For all common ASH conditions except cellulitis, ASH rates were higher in the MHA cohort. After adjusting for age and ethnicity, people accessing MHA services were six times more likely to be hospitalised for epilepsy (ARR=5.96), four times for COPD (ARR=4.32) and over three times for diabetes (ARR=3.47).

Discussion

This study is the first to examine ASH rates among people accessing MHA services in Aotearoa. The MHA cohort’s ASH rates were more than twice those of the total population. Within the MHA cohort, ASH rates were higher for people aged 60–64 and living in high-deprivation areas, and lower for nMnP ethnic groups. Non-communicable diseases were the most common causes of ASH events for the MHA cohort, with angina and chest pain accounting for over one-quarter (26.0%) and COPD contributing over one-tenth (11.6%) of cases.

Alignment with previous research

Results align with international research demonstrating elevated ASH rates among people experiencing MHA issues.^9–11 For instance, an Australian study found ASH rates among people accessing mental health services were 3.6 times as high as the general population.⁹ Similarly, a Veteran Health Administration services study observed an 11% increased likelihood of ASHs among people diagnosed with bipolar disorder or schizophrenia than those with no mental health diagnosis.¹¹ Despite methodological differences, ASH rates for people experiencing MHA issues are consistently elevated.

Findings also align with research indicating high co-existing rates of mental health conditions with cardiovascular, respiratory and other non-communicable diseases.¹² Such comorbidities are often exacerbated by inadequate access to primary and preventative care; studies demonstrate people with mental health diagnoses are less likely to receive screening, early diagnosis and treatment for non-communicable diseases.^12–14 These are closely intertwined with socio-economic disadvantages and may in part account for ethnic disparities.¹⁵

Finding disparities aligns with local research showing elevated ASH rates among Māori, Pacific people and people living in high-deprivation areas.^7,16,17 These are likely attributable to multiple interlinking factors including practical barriers to accessing healthcare (e.g., transport, financial costs), racial bias contributing to unsafe clinical environments, and poor communication.^7,16,18

Clinical practice implications

Factors like diagnostic overshadowing (where symptoms are misattributed to MHA conditions rather than other co-existing issues), stigma, practitioner bias and communication difficulties with healthcare practitioners contribute to ASH rate disparities for people experiencing MHA issues.^3,19–21 Practitioner biases about people with MHA issues or diagnoses can influence recognition of physical health symptoms, screening, medication prescriptions and specialist care referrals.¹⁹ In a local study, participants reported physical symptoms being misattributed as psychosomatic or caused by MHA issues, and their mental and physical health needs being seen as competing rather than simultaneous priorities.¹⁹ Such treatment can deter people from seeking healthcare and lead to missed opportunities for timely healthcare, and thereby exacerbate health issues.

For minoritised peoples in societies with histories of colonisation, socio-economic burdens and culturally unsafe services limit healthcare access and quality, and cause unfair, preventable patterns of adverse health outcomes.^20,22 A local study highlights this pattern of privilege whereby non-Māori experiencing MHA issues are less likely to experience unfair treatment and diagnostic overshadowing in primary care than Māori.²⁰ Another study highlights the importance of culturally appropriate treatment and therapeutic relationships for Pacific peoples.²² National health strategies establish the need for tailored approaches to improve primary healthcare for Māori and Pacific peoples.²³

Improvement and integration of mental health and primary care at provider, service and system levels are needed to sustainably address the disparities found. Initiatives like introducing free influenza immunisations for MHA service users reflect an important step towards reducing health inequities.²⁴ Workforce development in cultural safety, recovery-oriented practices and developing knowledge, skills and confidence in addressing MHA issues can mitigate biases and improve outcomes.^3,19

System-level shifts are needed to enable cross-sectoral integration and support equitable health outcomes. Siloing of mental and physical healthcare, and of primary and specialist healthcare, are both drivers of inequitable health outcomes for MHA service users. Integrating primary care (screening, prevention and timely treatment) and MHA services is required to remove cross-sectoral siloing and reduce complications, hospitalisations and inequities in preventable health conditions.^1,12,25,26 Additionally, expanding availability and capacity of Kaupapa Māori and Pacific health services will enhance people’s options and access to culturally safe and holistic care. These changes require system-level shifts in service funding and regulation, and investments to ensure the sustainable provision of effective, integrated services.²⁷

Strengths and limitations

This study uses a nationally representative sample of MHA service users to examine physical health disparities over 6 years. The use of routinely collected health data demonstrates a replicable, low-cost approach to monitor changes and efficacy of interventions or policies.

This study is limited by its focus on specialist MHA service users, excluding people who may face additional barriers to accessing specialist services and those who receive mental health treatment in primary care. Findings are drawn from routine data not collected for research purposes, so potential inaccuracies in coding and clinical information are not accounted for. In particular, ethnicity data held in national health data collections are known to undercount Māori, which may impact the accuracy of ethnicity-specific rates.²⁸ Additionally, only data for MHA service use between July 2012 and June 2018 were available at the time of analyses. Examining more recent data is likely to demonstrate similar patterns but may reveal additional information. These analyses could be updated as part of routine ASH data monitoring.

Analyses focussed on people aged 45–64 with gender recorded as male or female, so findings may not be generalisable to other genders and ages. ASH data for people aged 5–44 is not reported nationally, so it was not possible to compare ASH rates at these ages, masking potentially important disparities for a significant proportion of the MHA service user population.²⁹

Regression analyses only control for age and ethnicity, limited by available total population data. Other potential confounding factors not accounted for include area-level deprivation, specific mental health conditions, health behaviours (e.g., smoking, exercise) and symptom severity, and the compounding effects of such factors with ethnicity. Including MHA service users in the total-population group may underestimate disparities. Including people not accessing any health services in the total-population denominator may have resulted in a lower estimate of the inequities experienced by MHA service users than would be found if they were compared with others accessing health services.

Conclusion

This study demonstrates the use of routine health data to examine physical health outcomes for people accessing MHA services. System-level changes are required to enable integrated physical and mental healthcare, primary and preventative care and workforce development to address disparities and improve outcomes for MHA service users.