Skip main navigation

Receipt of Mental Health Treatment in People Living With Stroke: Associated Factors and Long-Term Outcomes

Originally published 2023;54:1519–1527



Untreated poststroke mood problems may influence long-term outcomes. We aimed to investigate factors associated with receiving mental health treatment following stroke and impacts on long-term outcomes.


Observational cohort study derived from the Australian Stroke Clinical Registry (AuSCR; Queensland and Victorian registrants: 2012–2016) linked with hospital, primary care billing and pharmaceutical dispensing claims data. Data from registrants who completed the AuSCR 3 to 6 month follow-up survey containing a question on anxiety/depression were analyzed. We assessed exposures at 6 to 18 months and outcomes at 18 to 30 months. Factors associated with receiving treatment were determined using staged multivariable multilevel logistic regression models. Cox proportional hazards regression models were used to assess the impact of treatment on outcomes.


Among 7214 eligible individuals, 39% reported anxiety/depression at 3 to 6 months following stroke. Of these, 54% received treatment (88% antidepressant medication). Notable factors associated with any mental health treatment receipt included prestroke psychological support (odds ratio [OR], 1.80 [95% CI, 1.37–2.38]) or medication (OR, 17.58 [95% CI, 15.05–20.55]), self-reported anxiety/depression (OR, 2.55 [95% CI, 2.24–2.90]), younger age (OR, 0.98 [95% CI, 0.97–0.98]), and being female (OR, 1.30 [95% CI, 1.13–1.48]). Those who required interpreter services (OR, 0.49 [95% CI, 0.25–0.95]) used a health benefits card (OR, 0.73 [95% CI, 0.59–0.92]) or had continuity of primary care visits (ie, with a consistent physician; OR, 0.78 [95% CI, 0.62–0.99]) were less likely to access mental health services. Among those who reported anxiety/depression, those who received mental health treatment had an increased risk of presenting to hospital (hazard ratio, 1.06 [95% CI, 1.01–1.11]) but no difference in survival (hazard ratio, 0.86 [95% CI, 0.58–1.27]).


Nearly half of the people living with mood problems following stroke did not receive mental health treatment. We have highlighted subgroups who may benefit from targeted mood screening and factors that may improve treatment access.

Approximately one-third of people living with stroke experience depression at some point in their recovery, while an estimated 18% to 24% experience anxiety.1,2 The 2 conditions are comorbid and associated with caregiver burden.3,4 Unresolved depression or anxiety may affect a range of poststroke outcomes, including recovery in activities of daily living and long-term survival.5–7 Individuals who experience poststroke depression have been shown to have increased hospitalizations and outpatient visits relative to those without a mental health diagnosis, even after accounting for prestroke medical utilization, stroke severity, and mental health-related visits.8

See related article, p 1528

Pharmacotherapy is often the first-line treatment for depression after stroke but may be associated with side effects and risks especially in those with multiple comorbidities and polypharmacy.9–11 Psychological treatment has been used to alleviate symptoms of poststroke depression or anxiety without adverse impacts.10 For depression, psychological interventions delivered in conjunction with medication seem to be more effective than either treatment alone.12,13 Clinical guidelines for the treatment and prevention of poststroke mood problems include antidepressant medications, exercise programs, and psychological therapies.14 Despite evidence of treatment effectiveness, more than two-thirds of people living with stroke reported that their psychological needs were not fully met.15–17 Most Australians with a neurological disorder reported at least 1 barrier to receiving mental health treatment,18 and treatment access may be complicated by physical, cognitive, and communication limitations.19 Our earlier research, one of the first to investigate this in people with stroke, identified older age, not feeling socially isolated, having no previous mental health treatment, no medical diagnosis of anxiety/depression, and no multidisciplinary team care arrangement plan as barriers.20 Although this study provided novel insights, the restriction of the cohort to registry participants who completed a project-specific survey may not have been reflective of the broader population. Further investigation was warranted to confirm these results at a population level and to clarify whether receipt of mental health treatment influences long-term outcomes.

The objectives of this study were to (1) identify demographic, clinical and structural factors associated with receipt of mental health treatment following stroke or transient ischemic attack in a population cohort and (2) investigate the association between receipt of mental health treatment in the 6 to 18 months following stroke and long-term outcomes, including survival and hospital utilization.


Ethics and Data Availability

This project was approved by the Monash University Human Research Ethics Committee (MUHREC/12301) and the Australian Institute of Health and Welfare ethics committee (EO2018/2/449). Approvals were also obtained from the Australian Stroke Clinical Registry (AuSCR) Research Task Group and relevant data custodians for each of the linked datasets. The AuSCR has ethics approval to use opt-out consent, whereby patients are automatically included in the registry unless they subsequently decline (opt out rate ~3%). Due to ethical and legal restrictions, patient-level data from this study cannot be shared. However, aggregated data outputs and coding that support the findings of this study are available on reasonable request from the corresponding author, following approval from the relevant data custodians.

Design and Setting

This is a sub-study of the PRECISE project,21 an observational cohort study using linked population data. The cohort was derived from the AuSCR (a national clinical quality registry designed to monitor and improve stroke care) and linked to a range of administrative datasets to obtain exposure and outcome variables not routinely collected within the AuSCR. Data were linked by accredited state and commonwealth data linkage units using probabilistic and deterministic methods. This manuscript follows STROBE reporting guidelines22 (Table S1).


Registrants eligible for this study were those aged 18 years and older who were admitted to participating hospitals (n=45 hospitals) between 2012 and 2016 with a diagnosis of stroke or transient ischemic attack and completed the 90 to 180 day AuSCR follow-up survey. We restricted the analysis to registrants admitted to hospitals in Victoria and Queensland, as these states have government support for participation in the registry resulting in high coverage. We also excluded registrants who, during the first 18 months poststroke/transient ischemic attack, either died, did not have a Medicare-subsidised visit with a primary care physician, or resided in permanent residential aged care or palliative care.

Data Sources

The linked dataset in PRECISE21 comprises information from the following sources:

  1. AuSCR registrants have a clinical diagnosis of stroke, and information on clinical processes of care and health outcomes are prospectively collected. Between 90 and 180 days poststroke, eligible registrants are followed up and administered a survey, which includes the European Quality of Life 5 Dimensions 3 Levels (EQ-5D-3L)23 (a scale assessing subjective quality of life that includes a question on current presence of anxiety or depression).

  2. Medicare is the Australian government’s universal healthcare system that includes rebates for primary care visits and mental health-related services.24 Rebates are available for assessment and formulation of mental health treatment plans by primary care physicians, as well as referral to and treatment sessions with appropriate specialists. Table S2 contains descriptions of Medicare items included.

  3. Pharmaceutical Benefits Scheme provides subsidised and affordable access to the majority of medications prescribed in Australia, including mental health-related prescriptions.25

  4. National Death Index lists all deaths that have occurred in Australia since 1980.

  5. Hospital admission data informs inpatient separations, including formal separations from the facility (ie, discharges, transfers, deaths) and changes in principal clinical intent within the same period of stay. Primary and secondary admission related diagnoses are coded using standardised International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM) and Australian Classification of Health Interventions.

  6. Emergency department (ED) data informs presentations to public EDs, and contains variables on triage category, type of visit, discharge destination and primary diagnosis.


Our exposure variable was the receipt of mental health treatment following stroke, as indicated by Medicare claims and pharmaceutical dispensations of mental health-related medications (ATC codes commencing with N05 and N06, excluding drugs for dementia and sleep disorders). Medicare claimable mental health services may be provided by psychiatrists, psychologists, primary care physicians, and allied mental health workers (Table S2). The exposure period was defined as the 6 to 18 months following stroke. By 6 months most people with stroke have completed their hospital treatments and are in the care of their primary care physician. This timeframe also aligns with the AuSCR follow-up period.

Factors Investigated for Association With Mental Health Treatment Receipt

Variables are detailed in Table S3 and included: Demographic factors such as sex, age at stroke and living situation; Clinical factors such as stroke severity (ability to walk on admission as a reliable proxy marker of minor stroke26,27) and type of stroke; and Structural factors such as regularity and continuity of visits to a primary care physician.

On the EQ-5D-3L, participants may respond Not/Moderately/Extremely to the question of whether they currently have anxiety or depression. Those who selected Moderately or Extremely were categorised as having self-reported anxiety/depression. This has been shown to have moderate concordance with the Hospital Anxiety and Depression Scale, which assesses feelings of anxiety and depression over the past week.28 While this is a brief measure that compounds 2 distinct constructs and does not confer a diagnosis, it is considered suitable for population level screening of mood problems.28 Comorbidities were determined using hospital ICD-10-AM coded data from all hospital utilization (ED and admission) in the 5 years prior to and including the stroke event.29 Derived measures included the following: socioeconomic status using the Index of Relative Socioeconomic Advantage and Disadvantage divided into 5 predetermined quintiles; primary care physician regularity defined as at least once every 6 months; and continuity of visits using a weighted measure where a score of >80% indicated continuity with the same primary care physician.30


Outcomes examined were survival status and hospital utilization during the 18 to 30 months following stroke. Hospital utilization included planned (eg, arranged by primary care physician) and unplanned admissions, as well as ED presentations that did not results in a hospital admission. Cumulative rates of hospital utilization were calculated per 1000 person-years to account for the competing risk of death.

Statistical Analyses

Completeness was assessed and missing data in primary datasets were imputed with data from another dataset. Multivariable logistic regressions were used to determine factors associated with receipt of mental health treatment. We used a staged approach to explore the relative contribution of the different categories of variables: (1) demographic covariates only, (2) demographic and clinical covariates, and (3) demographic, clinical, and structural covariates. This was undertaken for both forms of treatment combined as well as Medicare claims and pharmaceutical dispensing separately. Models were assessed for multicollinearity (a Variance Inflation Factor <10 indicating low multicollinearity), goodness-of-fit (X2 and Bayesian Information Criterion) and percentage of variance accounted for at each stage. A sensitivity analysis was conducted, in which those with a history of mental health treatment prestroke were excluded from these analyses.

Multilevel mixed-effects Cox proportional hazards regression models31 (with level defined as patient and health region) were used to examine the association between mental health treatment and long-term outcomes. To minimize confounding by indication, 49 baseline variables were balanced using propensity score methods to derive Inverse Probability of Treatment Weights. This method ensured an even distribution of measured covariates between those who did and did not receive mental health treatment to determine the average treatment effect across the population (Table S4 contains full list of variables). Baseline covariates between those who did and did not receive mental health treatment were compared in the weighted sample for balance. An absolute standardized difference <0.1 was defined as negligible imbalance.


Of the 28 775 AuSCR registrants eligible for data linkage, 27 435 (95.3%) were successfully linked with the administrative datasets. Of these, 12 368 were included in the final PRECISE cohort and 7214 (58%) provided EQ-5D-3L data and were included in this study (refer to Figure S1 for flowchart). A further 316 (4%) were excluded from the multivariable analyses due to missing data. Comparison of characteristics between those who did and did not provide EQ-5D-3L can be found in Table S5. Those who completed the EQ-5D-3L were less likely to be socially disadvantaged, less likely to have had a history of mental health treatment, and more likely to have private health insurance. The majority of participants were born in Australia (99%) and 42% were female. The median age at stroke was 71.3 years (Q1, 61.5; Q3, 79.2). Participant demographic and clinical characteristics are shown in Table 1.

Table 1. Baseline Characteristics of Participants Who Did and Did Not Receive Mental Health Treatment

Did not receive mental health treatmentReceived mental health treatmentP value
n (%)n (%)
 Median age at stroke admission, y (Q1–Q3)*72.4 (62.6–79.8)69.6 (59.4–78.3)<0.001
 Female*1713 (38)1331 (50)<0.001
 Born in Australia*4469 (98)2646 (99)0.106
 English as first language*3869 (85)2307 (86)0.241
 Needed an interpreter*157 (3)77 (3)0.179
 Socioeconomic position*0.391
  Most disadvantaged770 (17)474 (18)
  Second most disadvantaged764 (17)454 (17)
  Third most disadvantaged981 (22)606 (23)
  Fourth most disadvantaged1045 (23)613 (23)
  Least disadvantaged979 (22)528 (20)
 Lives in regional area*1405 (31)910 (34)0.007
 Married/de facto3156 (70)1789 (67)0.019
 Living alone*1023 (23)600 (22)0.915
 Type of stroke*<0.001
  Intracerebral hemorrhage300 (7)244 (9)
  Ischemic2980 (66)1742 (65)
  Transient ischemic attack1119 (25)601 (22)
  Not determined/missing140 (3)88 (3)
 Able to walk independently on admission (stroke severity proxy)*2502 (55)**1351 (51)††<0.001
 History of mental health treatment
  Medicare uptake ≥1115 (3)391 (15)<0.001
  Pharmaceutical dispensing ≥1§307 (7)1581 (59)<0.001
  Hospital presentation ≥112 (0)43 (2)<0.001
 Mean CCI-weighted score (SD)#1.54 (1.72)1.85 (1.79)<0.001
 Self-reported anxiety/depression1271 (28)1516 (57)<0.001
 CD management plan during exposure period1881 (41)1363 (51)<0.001
 Private health insurance1916 (42)1142 (43)0.690
 Health benefits card3063 (67)1905 (71)0.001
 Mean total primary care visits prestroke (SD)10.76 (8.57)14.57 (10.08)<0.001
  Regularity3309 (73)2218 (83)<0.001
  Continuity1169 (26)590 (22)<0.001

CCI indicates Charlson Comorbidity Index; Q1, 25th percentile; and Q3, 75th percentile.

* Derived from the Australian Stroke Clinical Registry from the acute stroke event.

† Derived from the Australian Stroke Clinical Registry 90–180 d follow-up survey.

‡ Derived from Australian Medicare (primary care) claims data.

§ Derived from pharmaceutical dispensing data.

¶ Derived from poststroke hospital data.

# Derived from prestroke hospital data.

** Missing 195 data points.

†† Missing 115 data points.

Based on responses to the EQ-5D-3L anxiety/depression domain obtained at the AuSCR 90- to 180-day follow-up assessment, 39% of respondents were categorised as self-reporting anxiety/depression. Thirty-seven percent of all participants received mental health treatment, while 54% of participants who reported anxiety/depression received mental health treatment. The breakdown of the type of treatment received is reported in Table 2, with antidepressants being the most common treatment type.

Table 2. Types of Mental Health Treatment Received by People Living With Stroke

PrestrokePoststroke overallSelf-reported poststroke anxiety/depression
Received any mental health treatment2035 (28%)2531 (37%)1437 (54%)
Medicare506 (7%)720 (10%)464 (17%)
 Primary care physician395564367
 Allied Health112416
Pharmaceutical dispensing1888 (26%)2460 (34%)1401 (50%)
Received both Medicare and pharmaceutical dispensing359 (5%)505 (7%)349 (13%)

Factors Associated With Receiving Mental Health Treatment

In the staged logistic regression (Table 3), demographic factors accounted for 2% of the variance in model 1 (X2[13]=168.75; P<0.001), the addition of clinical factors (model 2) accounted for an additional 28% (X2[9]=2594.48; P<0.001), and the addition of structural factors (full model) accounted for an additional 2% of the variance (X2[6]=130.73; P<0.001). In the full model, the 23 predictor variables explained 32% of the variance in receiving mental health treatment (X2[28]=2893.96; P<0.001).

Table 3. Factors Associated With Receiving Mental Health Treatment Including Demographic, Clinical, and Structural

Stage 1: demographicStage 2: +ClinicalStage 3: +Structural
Odds ratio95% CIOdds ratio95% CIOdds ratio95% CI
Age at stroke0.99*0.98–0.990.98*0.98–0.990.98*0.97–0.98
Born in Australia1.400.90–––2.11
English as first language0.960.82–1.130.940.77–1.150.940.76–1.15
Interpreter needed0.860.63–1.170.720.49–1.070.750.50–1.12
Socioeconomic status
 Second most disadvantaged0.990.84–––1.29
 Third most disadvantaged1.030.88––1.501.200.98–1.48
 Fourth most disadvantaged1.040.88––1.501.210.97–1.51
 Least disadvantaged0.970.81––1.561.301.03–1.65
Marital status0.930.82––1.190.970.83–1.15
Not living alone1.080.94––1.441.210.99–1.45
Ability to walk on admission0.85*0.75–0.980.85*0.74–0.97
Type of stroke
 Transient ischemic attack0.67*0.52–0.870.64*0.49–0.83
Medicare mental health uptake prestroke1.96*1.49–2.581.80*1.37–2.38
Mental health medication prestroke17.70*15.21–20.6117.58*15.05–20.55
Mental health hospital admissions prestroke1.560.68–3.601.330.58–3.04
CCI weighted score1.08*1.04–1.121.05*1.02–1.09
Self-reported anxiety/depression poststroke2.68*2.36–3.042.55*2.24–2.90
CD management plan during exposure period1.17*1.02–1.34
Private health insurance1.15*1.01–1.32
Health benefits card0.930.78–1.10
Total primary care visits prestroke1.04*1.03–1.05
Regularity of primary care visits0.980.83–1.16
Continuity of primary care visits0.880.76–1.03
Bayesian Information Criterion9463.646539.236461.525
Variance accounted for2%30%32%

CCI indicates Charlson Comorbidity Index; CD, chronic disease; and Medicare, Medicare Benefits Schedule.

* Statistical significance.

In the full multivariable model (Table 3), having mental health-related pharmaceutical dispensing or Medicare claims prior to stroke were most associated with receiving poststroke mental health treatment. Other factors associated with receiving mental health treatment following stroke included: having self-reported anxiety/depression, being female, living in a regional area, having a chronic disease management plan, having private health insurance, and more primary care visits prestroke. Older age at stroke onset and having less severe stroke were associated with not receiving mental health treatment. Those who had ischemic stroke or a transient ischemic attack were less likely to receive mental health treatment than those with intracerebral hemorrhage.

Sensitivity analyses excluding participants who had a history of mental health treatment (via Medicare or pharmaceutical dispensing) yielded similar results (Table S6). However, those who had continuity of primary care visits were less likely to receive mental health treatment (OR, 0.81 [95% CI, 0.67–0.97]).

When Medicare and pharmaceutical dispensing claims were examined separately, results were similar with the exception of: those who required interpreter services, held a health benefits card (concession card in Australia), and had continuity of primary care visits were less likely to receive mental health treatment through Medicare. Having a greater number of comorbid health conditions was associated with receiving mental health-related medications through pharmaceutical dispensing (Table 4).

Table 4. Factors Associated With Accessing Mental Health Services or Medication Including Demographic, Clinical, and Structural

Accessed mental health services (Medicare)Accessed mental health medication
Odds ratio95% CIOdds ratio95% CI
Age at stroke0.95*0.94–0.960.99*0.98–0.99
Born in Australia0.940.43–2.021.340.74–2.44
English as first language0.790.59––1.24
Interpreter needed0.49*0.25–0.950.870.58–1.30
Socioeconomic status
 Second most disadvantaged1.220.90–1.651.010.81–1.27
 Third most disadvantaged1.260.94–1.691.180.95–1.46
 Fourth most disadvantaged1.220.89–1.671.210.96–1.52
 Least disadvantaged1.360.97–1.901.341.05–1.71
Marital status1.140.91–1.430.970.82–1.15
Not living alone1.020.79–1.331.160.96–1.41
Ability to walk on admission1.060.88–1.290.85*0.74–0.98
Type of stroke
 Transient ischemic attack0.700.49–1.000.64*0.49–0.84
Medicare mental health uptake prestroke4.11*3.24––1.67
Mental health medication prestroke1.80*1.47–2.1819.28*16.55–22.48
Mental health hospital admissions prestroke1.130.57––2.51
CCI weighted score0.990.94–1.051.06*1.02–1.10
Self-reported anxiety/depression poststroke2.19*1.82–2.622.45*2.15–2.79
CD management plan during exposure period1.39*1.15–1.691.140.99–1.31
Private health insurance0.890.74–1.081.16*1.01–1.33
Health benefits card0.73*0.59–0.920.960.81–1.15
Total primary care visits prestroke1.05*1.04–1.061.03*1.02–1.04
Regularity of primary care visits0.940.74––1.19
Continuity of primary care visits0.78*0.62–0.990.900.77–1.05
Bayesian Information Criterion3815.606215.42
Variance accounted for21%33%

Medicare indicates Medicare Benefits Schedule.

* Statistical significance.

Outcomes Associated With Receipt of Mental Health Treatment

A total of 279 individuals (4%) died in the 18 to 30 months following their stroke and 9499 hospital presentations were observed: 3800 (40%) planned admissions, 3855 (41%) unplanned admissions, 1844 (19%) non-admitted ED presentations. In the multivariable analyses (Table 5), there was no significant survival benefit observed in those who received mental health treatment compared to those who did not (hazard ratio, 1.06 [95% CI, 0.82–1.38]). Results were similar when stratified by reported anxiety/depression in the 90 to 180 days poststroke.

Table 5. Hazard Ratios for Survival and Hospital Utilization Based on Mental Health Treatment Receipt, Overall and Stratified by Self-Reported Anxiety/Depression

SurvivalAll hospital utilizationUnplanned admissionsPlanned admissions
Hazard ratio (95% CI)
Overall1.06 (0.82–1.38)1.04 (1.00–1.09)1.04 (0.98–1.11)0.99 (0.92–1.06)
 Reported anxiety/depression0.86 (0.58–1.27)1.06 (1.01–1.11)*1.08 (0.99–1.18)0.95 (0.83–1.09)
 Did not report anxiety/depression1.23 (0.88–1.73)1.03 (0.97–1.10)1.01 (0.93–1.09)1.01 (0.88–1.16)

Models adjusted for IPTW and year. IPTW indicates Inverse Probability of Treatment Weights.

* Statistical significance.

There was no significant difference in hospital utilization in those who received mental health treatment compared to those who did not (hazard ratio, 1.04 [95% CI, 1.00–1.09]). However, there was a significant difference when the analysis was stratified by reported anxiety/depression in the 90 to 180 days poststroke (hazard ratio, 1.06 [95% CI, 1.01–1.11]).


Our study provides population-level evidence of how commonwealth government funded mental health treatments are used following stroke. Thirty-nine percent of participants reported anxiety/depression at 3 to 6 months poststroke. Among them, just over half received mental health treatment, predominantly antidepressant medication. Only 13% received a combination of psychological management and mood medication. Notable factors associated with treatment receipt included prestroke mental health treatment, self-reported anxiety/depression, younger age at stroke, and being female. Clinical factors explained the largest amount of variance in treatment receipt. Having a claim for a chronic disease management plan and more primary care visits prestroke also contributed. Receiving mental health treatments did not appear to reduce mortality or hospital utilization.

In line with previous research, we identified men, older adults, and those who had not previously received mental health treatment as potentially underserviced populations.32 Provision of information during early rehabilitation and consultation sessions may improve insight into mood problems and awareness of support options. Consistent with our earlier study on a smaller survey-based cohort, these population-level results highlight the pivotal role of primary care physicians in the pathway to mental health care.20 Our results suggest that access to Medicare-funded chronic disease management plans,33 designed to support collaborative care based on the patient’s needs and goals, may facilitate mental health treatment. The holistic approach of these policies may provide opportunities to discuss mental health problems and develop appropriate action plans to manage identified needs.

The main barriers specific to having a claim for Medicare-funded psychological support, independent of medication prescribing, include requiring interpreter services and having a health benefits card. Verbal communication is important for accurate identification and treatment of mental health needs during primary care visits. There is a consistent and clear association between limited language proficiency and underutilization of mental health services.34 Poor communication may explain why depressive symptoms tend to be underdiagnosed and undertreated in culturally and linguistically diverse groups, despite equal quality of care for the primary medical concern.35 Additionally, psychological therapy can be costly and may not be fully covered by Medicare, therefore, less accessible to those who require a health benefits card.

Untreated mental health problems may affect a range of long-term outcomes following stroke (eg, mortality, activity limitations), but appropriate treatment may alleviate the effect of depression on functional recovery.5–8,36,37 In our outcome analyses, treatment claims did not improve survival or reduce hospital utilization. We did not have follow-up data related to functional recovery or quality of life, which may have provided a better indication of effectiveness for these types of treatment. Another study found that people with depression following stroke had poorer recovery in activities of daily living and mobility, relative to those without depression, despite comparable stroke severity and being on antidepressant medication.38 Hospital utilization may also be indicative of severity of other medical comorbidities.

A strength of this study was the large sample size and the broad range of objective and reliable population information obtained through data linkage. Limitations include the use of EQ-5D to indicate anxiety/depression, which is self-reported and not a validated measure although it has shown moderate concordance with the Hospital Anxiety and Depression Scale.28 Selection bias, due to EQ-5D completion rates, may have impacted results. Variables that could be investigated in the multivariable model were restricted based on available datasets. This meant that some factors specific to mental health treatment, such as stigma within certain groups39,40 or access to privately funded services, was not able to be accounted for. Although we accounted for confounding using Inverse Probability of Treatment Weights including 49 variables known to be associated with outcomes following stroke, we cannot discount residual confounding. Due to custodial and ethical constraints, we were unable to report and discuss whether Aboriginal and Torres Strait Islander status was associated with treatment receipt. Having a functional, quality of life, or psychological outcome measure may have provided a more sensitive and specific measure of mental health treatment effectiveness.


Our results indicate that approximately 1 in 2 people living with stroke with self-reported anxiety/depression are not receiving mental health treatment, and those who do are mostly receiving medication only. Health professionals should screen for mental health problems and introduce treatment options, with particular attention to individuals who are at risk of not receiving treatment. Future studies should investigate the lived experience of people with stroke with regards to receiving mental health treatment, which may inform practical ways to facilitate treatment receipt and improve rehabilitation and care more broadly.

Article Information


We acknowledge the Australian Stroke Clinical Registry (AuSCR) consortium (including the AuSCR Steering and Management Committees, staff from the Florey Institute of Neuroscience and Mental Health, hospital clinicians; online supplemental data) and patients who contributed data to the AuSCR. We also acknowledge staff from the state and commonwealth units who undertook the data linkage for this project and each state data collection agency that provided access to these data, including the Victorian Department of Health (source of the Victorian Admitted Episodes Dataset and the Victorian Emergency Minimum Dataset), the Centre for Victorian Data Linkage (for provision of Victorian data linkage), the Queensland Department of Health (sources of the Queensland Hospital Admitted Patient Data Collection and Emergency Data Collection), and the Australian Institute of Health and Welfare (sources of the Medicare, Pharmaceutical dispensing, National Death Index and National Aged Care Data Clearing House).

Supplemental Material

Tables S1–S6

Figure S1

Supplemental Acknowledgements

Nonstandard Abbreviations and Acronyms


emergency department


European Quality of Life 5 Dimensions 3 Levels


odds ratio


Australian Stroke Clinical Registry

Disclosures Dr Cadilhac is the current Data Custodian for the Australian Stroke Clinical Registry (AuSCR). Drs Cadilhac, Lannin, and Kilkenny are members of the AuSCR Steering or Management Committees and Dr Andrew is a member of the AuSCR Research Task Group. Dr Cadilhac reports receiving educational grants from Amgen Australia, Boehringer Ingelheim, Ipsen, Medtronic, and Shire outside the submitted work. Dr Kilkenny reports receiving educational grants from GSK and Amgen Australia outside the submitted work. Dr. Kneebone is a member of the end point review committee for the National Stroke Foundation. Dr Nelson was a member of the 2020 Novartis lipids advisory board. Dr Dalli reports receiving an educational grant from GSK outside the submitted work. All other authors report no conflicts.


For Sources of Funding and Disclosures, see page 1526.

Supplemental Material is available at

*A list of the PRECISE Investigators and staff members is available in the Supplemental Material.

Correspondence to: Nadine Andrew, PhD, Department of Medicine, Peninsula Clinical School, Central Clinical School, Monash University, Ngarnga Centre, Frankston Hospital, 2 Hastings Rd, Frankston, 3199, Victoria, Australia. Email


  • 1. Hackett ML, Pickles K. Part I: frequency of depression after stroke: an updated systematic review and meta-analysis of observational studies.Int J Stroke. 2014; 9:1017–1025. doi: 10.1111/ijs.12357CrossrefMedlineGoogle Scholar
  • 2. Knapp P, Dunn-Roberts A, Sahib N, Cook L, Astin F, Kontou E, Thomas SA. Frequency of anxiety after stroke: an updated systematic review and meta-analysis of observational studies.Int J Stroke. 2020; 15:244–255. doi: 10.1177/1747493019896958CrossrefMedlineGoogle Scholar
  • 3. Dou D-M, Huang L-L, Dou J, Wang X-X, Wang P-X. Post-stroke depression as a predictor of caregivers burden of acute ischemic stroke patients.Psychol Health Med. 2018; 23:541–547. doi: 10.1080/13548506.2017.1371778CrossrefGoogle Scholar
  • 4. Lincoln NB, Brinkmann N, Cunningham S, Dejaeger E, De Weerdt W, Jenni W, Mahdzir A, Putman K, Schupp W, Schuback B, et al. Anxiety and depression after stroke: a 5 year follow-up.Disabil Rehabil. 2013; 35:140–145. doi: 10.3109/09638288.2012.691939CrossrefMedlineGoogle Scholar
  • 5. Chemerinski GE, Robinson TR, Kosier TJ. Improved recovery in activities of daily living associated with remission of poststroke depression.Stroke. 2001; 32:113–117. doi: 10.1161/01.str.32.1.113LinkGoogle Scholar
  • 6. Cai W, Mueller C, Li Y-J, Shen W-D, Stewart R. Post stroke depression and risk of stroke recurrence and mortality: a systematic review and meta-analysis.Ageing Res Rev. 2019; 50:102–109. doi: 10.1016/j.arr.2019.01.013CrossrefMedlineGoogle Scholar
  • 7. Bartoli F, Di Brita C, Crocamo C, Clerici M, Carrà G. Early post-stroke depression and mortality: META-analysis and meta-regression.Front Psychiatry. 2018; 9:530. doi: 10.3389/fpsyt.2018.00530CrossrefMedlineGoogle Scholar
  • 8. Ghose SS, Williams LS, Swindle RW. Depression and other mental health diagnoses after stroke increase inpatient and outpatient medical utilization three years poststroke.Med Care. 2005; 43:1259–1264. doi: 10.1097/01.mlr.0000185711.50480.13CrossrefMedlineGoogle Scholar
  • 9. Starkstein SE, Hayhow BD. Treatment of post-stroke depression.Curr Treat Options Neurol. 2019; 21:31. doi: 10.1007/s11940-019-0570-5CrossrefGoogle Scholar
  • 10. Allida S, Cox KL, Hsieh CF, Lang H, House A, Hackett ML. Pharmacological, psychological, and non-invasive brain stimulation interventions for treating depression after stroke.Cochrane Database Syst Rev. 2020; 1:eCD003437. doi: 10.1002/14651858.CD003437.pub4CrossrefGoogle Scholar
  • 11. Kim JS. Management of post-stroke mood and emotional disturbances.Expert Rev Neurother. 2017; 17:1179–1188. doi: 10.1080/14737175.2017.1395281CrossrefGoogle Scholar
  • 12. Cuijpers P, Van Straten A, Warmerdam L, Andersson G. Psychotherapy versus the combination of psychotherapy and pharmacotherapy in the treatment of depression: a meta-analysis.Depress Anxiety. 2009; 26:279–288. doi: 10.1002/da.20519CrossrefGoogle Scholar
  • 13. Mitchell HP, Veith CR, Becker JK, Buzaitis CA, Cain CK, Fruin CM, Tirschwell CD, Teri CL. Brief psychosocial–behavioral intervention with antidepressant reduces poststroke depression significantly more than usual care with antidepressant: living well with stroke: randomized, controlled trial.Stroke. 2009; 40:3073–3078. doi: 10.1161/STROKEAHA.109.549808LinkGoogle Scholar
  • 14. Stroke Foundation. Chapter 6: managing complications. in: clinical guidelines for stroke management 2017, 2017.Google Scholar
  • 15. Andrew NE, Kilkenny MF, Naylor R, Purvis T, Lalor E, Moloczij N, Cadilhac DA; National Stroke Foundation. Understanding long-term unmet needs in Australian survivors of stroke.Int J Stroke. 2014; 9:106–112. doi: 10.1111/ijs.12325CrossrefGoogle Scholar
  • 16. Walsh ME, Galvin R, Loughnane C, Macey C, Horgan NF. Community re-integration and long-term need in the first five years after stroke: results from a national survey.Disabil Rehabil. 2015; 37:1834–1838. doi: 10.3109/09638288.2014.981302CrossrefGoogle Scholar
  • 17. McKevitt C, Fudge N, Redfern J, Sheldenkar A, Crichton S, Rudd AR, Forster A, Young J, Nazareth I, Silver LE, et al. Self-reported long-term needs after stroke.Stroke. 2011; 42:1398–1403. doi: 10.1161/STROKEAHA.110.598839LinkGoogle Scholar
  • 18. Gandy M, Karin E, Fogliati VJ, Meares S, Nielssen O, Titov N, Dear BF. Emotional and cognitive difficulties, help-seeking, and barriers to treatment in neurological disorders.Rehabil Psychol. 2018; 63:563–574. doi: 10.1037/rep0000241CrossrefGoogle Scholar
  • 19. Ryan B, Bohan J, Kneebone I. Help-seeking and people with aphasia who have mood problems after stroke: perspectives of speech-language pathologists.Int J Lang Commun Disord. 2019; 54:779–793. doi: 10.1111/1460-6984.12476CrossrefGoogle Scholar
  • 20. Tjokrowijoto P, Stolwyk RJ, Ung D, Kilkenny MF, Kim J, Dalli LL, Cadilhac DA, Andrew NE. Factors associated with mental health service access among Australian community-dwelling survivors of stroke.Disabil Rehabil. 2022; 45:504. doi: 10.1080/09638288.2022.2032413CrossrefGoogle Scholar
  • 21. Andrew NE, Kim J, Cadilhac DA, Sundararajan V, Thrift AG, Churilov L, Lannin NA, Nelson M, Srikanth V, Kilkenny MF. Protocol for evaluation of enhanced models of primary care in the management of stroke and other chronic disease (PRECISE): a data linkage healthcare evaluation study.Int J Popul Data Sci. 2019; 4:1–14. doi: 10.23889/ijpds.v4i1.1097CrossrefGoogle Scholar
  • 22. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbrouke JP. Strobe initiative. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies.Lancet. 2007; 370:1453–1457. doi: 10.1016/S0140-6736(07)61602-XCrossrefMedlineGoogle Scholar
  • 23. Cadilhac DA, Lannin NA, Anderson CS, Levi CR, Faux S, Price C, Middleton S, Lim J, Thrift AG, Donnan GA. Protocol and pilot data for establishing the Australian stroke clinical registry.Int J Stroke. 2010; 5:217–226. doi: 10.1111/j.1747-4949.2010.00430.xCrossrefMedlineGoogle Scholar
  • 24. Department of Health. Medicare benefits schedule.2020.Google Scholar
  • 25. Department of Health. Pharmaceutical Benefits Scheme (PBS) | about the PBS.2021.Google Scholar
  • 26. Cadilhac DA, Kilkenny MF, Levi CR, Lannin NA, Thrift AG, Kim J, Grabsch B, Churilov L, Dewey HM, Hill K, et al. Risk-adjusted hospital mortality rates for stroke: evidence from the Australian Stroke Clinical Registry (AuSCR).Med J Aust. 2017; 206:345–350. doi: 10.5694/mja16.00525CrossrefGoogle Scholar
  • 27. Counsell C, Dennis M, McDowall M, Warlow C. Predicting outcome after acute and subacute stroke.Stroke. 2002; 33:1041–1047. doi: 10.1161/hs0402.105909LinkGoogle Scholar
  • 28. Thayabaranathan T, Andrew NE, Stolwyk R, Lannin NA, Cadilhac DA. Comparing the EQ-5D-3L anxiety or depression domain to the hospital anxiety and depression scale to identify anxiety or depression after stroke.Top Stroke Rehabil. 2021;1–10.Google Scholar
  • 29. Quan H, Sundararajan VM, Halfon PM, Fong A, Burnand B, Luthi J-C, Saunders LD, Beck CA, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.Med Care. 2005; 43:1130–1139. doi: 10.1097/01.mlr.0000182534.19832.83CrossrefMedlineGoogle Scholar
  • 30. Ung D, Wang Y, Sundararajan V, Lopez D, Kilkenny MF, Cadilhac DA, Thrift AG, Nelson MR, Andrew NE. Optimal measures for primary care physician encounters after stroke and association with survival: a data linkage study.Neuroepidemiology. 2022; 56:90–96. doi: 10.1159/000520700CrossrefGoogle Scholar
  • 31. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk.J Am Stat Assoc. 1999; 94:496–509. doi: 10.1080/01621459.1999.10474144CrossrefGoogle Scholar
  • 32. Ten Have M, Graaf R, Ormel J, Vilagut G, Kovess V, Alonso J. Are attitudes towards mental health help-seeking associated with service use? Results from the European study of epidemiology of mental disorders.Soc Psychiatry Psychiatr Epidemiol. 2010; 45:153–163. doi: 10.1007/s00127-009-0050-4CrossrefGoogle Scholar
  • 33. Australian Government Department of Health and Ageing. Chronic disease management (formerly Enhanced Primary Care or EPC) — GP services.Google Scholar
  • 34. Ohtani A, Suzuki T, Takeuchi H, Uchida H. Language barriers and access to psychiatric care: a systematic review.Psychiatr Serv. 2015; 66:798–805. doi: 10.1176/ Scholar
  • 35. Sorkin DH, Ngo-Metzger Q, Billimek J, August KJ, Greenfield S, Kaplan SH. Underdiagnosed and undertreated depression among racially/ethnically diverse patients with type 2 diabetes.Diabetes Care. 2011; 34:598–600. doi: 10.2337/dc10-1825CrossrefGoogle Scholar
  • 36. Gainotti G, Antonucci G, Marra C, Paolucci S. Relation between depression after stroke, antidepressant therapy, and functional recovery.J Neurol Neurosurg Psychiatry. 2001; 71:258–261. doi: 10.1136/jnnp.71.2.258CrossrefMedlineGoogle Scholar
  • 37. Desrosiers J, Noreau L, Rochette A, Bourbonnais D, Bravo G, Bourget A. Predictors of long-term participation after stroke.Disabil Rehabil. 2006; 28:221–230. doi: 10.1080/09638280500158372CrossrefMedlineGoogle Scholar
  • 38. Paolucci S, Antonucci G, Grasso MG, Morelli D, Troisi E, Coiro P, De Angelis D, Rizzi F, Bragoni M. Post-stroke depression, antidepressant treatment and rehabilitation results.Cerebrovasc Dis. 2001; 12:264–271. doi: 10.1159/000047714CrossrefGoogle Scholar
  • 39. Corrigan PW, Druss BG, Perlick DA. The impact of mental illness stigma on seeking and participating in mental health care.Psychol Sci Public Interest. 2014; 15:37–70. doi: 10.1177/1529100614531398CrossrefGoogle Scholar
  • 40. Komiti A, Judd F, Jackson H. The influence of stigma and attitudes on seeking help from a GP for mental health problems.Soc Psychiatry Psychiatr Epidemiol. 2006; 41:738–745. doi: 10.1007/s00127-006-0089-4CrossrefGoogle Scholar


eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.