Abstract
Purpose:
To assess the mental health conditions, as indicated by mental health service contact in adolescents and young adults (AYAs) diagnosed with cancer in New South Wales (NSW) and associations with cancer mortality.
Methods:
In 3998 NSW AYAs diagnosed with cancer in 2005–2017, mental health service contacts were obtained from hospital inpatient records and specified medical and pharmaceutical insurance claims. Odds of postcancer mental health contact were assessed by precancer mental contacts using logistic regression adjusted for sociodemographic and cancer characteristics. The risk of cancer-specific mortality related to postcancer mental health contacts was estimated using competing risk regression.
Results:
The prevalence of mental health service contacts in the 5 years postcancer diagnosis was 27.0%, higher than the corresponding precancer prevalence of 21.4%. The most common mental health conditions were depression and anxiety. The odds of having a mental health contact postcancer diagnosis were higher in patients with a precancer mental health service contact (adjusted odds ratio 5.69, confidence intervals [95% CIs]: 4.90–6.75). The 5-year cancer-specific survival was 87.9% (95% CI: 85.8–89.8) for patients with a mental health service contact postcancer, which was lower than the 93.9% (95% CI: 93.0–94.7) for patients without this contact. The subhazard ratio (SHR) for cancer mortality in patients having mental health service contact postcancer diagnosis was 1.67 (95% CI: 1.29–2.15), adjusted for sociodemographic characteristics, cancer stage, and precancer mental health status.
Conclusion:
The prevalence of mental health service contact increased after a cancer diagnosis. Mental health care should be a continued priority for AYA cancer patients, particularly for high-risk groups.
Introduction
Adolescence and young adulthood (AYA) are critical life stages characterized by significant changes in brain development, psychological and cognitive capacity, social functioning, and evolving personal and role identity.1,2 They are also a time of vulnerability for experiencing mental health conditions. Of people experiencing any mental health condition, roughly one-third are reported to have had an onset before the age of 14 years, and nearly two-thirds before the age of 25 years. 3
World Health Organization (WHO) has estimated the prevalence of mental health conditions for ages 10–19 years at 14.0%. 4 The Australian National Survey of Mental Health and Wellbeing reported a higher 12-month prevalence of mental conditions (indicated by anxiety, mood changes, and substance use) of 31.5% at ages 16–24 years. 5 Critical psychosocial development during AYA years may be disrupted by a cancer diagnosis, with potential lifechanging effects on the emotional and mental well-being of AYAs and their families. AYAs with cancer have unique care needs due to this critical period of psychosocial, sexual, educational, and emotional development. 6
It is encouraging that 5-year survival from cancer has increased to 90% in AYAs in Australia. The extent of any ongoing impact on mental health status is uncertain, although there is some limited international evidence that mental health challenges can remain for some survivors.7–9 Understandably less research in AYA cancer patients has been directed at mental health conditions than at cancer treatment and survival. 10 Several small clinical studies using questionnaires have shown elevated mental health concerns.
For example, a state-wide cross-sectional survey in Victoria of 53 cancer patients aged 16–30 years reported a prevalence of distress of 25% using the Beck Depression Inventory-Fast Screen. 11 Another study included 288 AYA patients aged 15–29 years from the United Kingdom, United States, Canada, and Australia, indicating that 47% experienced distress using the distress thermometer. 12 A cross-sectional study including 196 AYA cancer patients from 18 sites across Australia found that 48% scored above a designated clinical threshold for post-traumatic stress symptoms. 13
It is difficult to derive summary statistics and draw conclusions regarding the prevalence of mental health conditions in AYA cancer survivors due to methodological differences, potential for wide geographic variation, and uncertain generalizability. Population-based evidence remains limited. A Canadian study reported that 61% of cancer patients diagnosed in 1991–2013 at age 15–39 years experienced psychiatric disorders in the 3 years after cancer diagnosis, with anxiety, depression, and trauma-related and stress-related disorders being the most common. 14 The relevance of this finding to Australian AYAs is unknown.
Factors found to be associated with a higher prevalence of mental health conditions in AYA cancer survivors have included older age at diagnosis, female sex, precancer mental health visits, and cancer type.15,16 Few AYA-specific studies have analyzed psychiatric disorders in greater depth using consistent and standardized methods. 15 Psychosocial care is a national priority for improving AYA cancer outcomes in Australia. An age-specific AYA Psychosocial Oncology Screening Tool has been developed and used to address psychosocial concerns in a timely and comprehensive manner.17,18
Still, the prevalence of mental health conditions has not been examined at a population level in Australian AYA cancer patients using well-defined methodologies. Better evidence on prevalence is needed to support prioritization of mental health care. This knowledge gap can be addressed using mental health service contacts drawn from New South Wales (NSW) population-based linked data. NSW is the largest state covering approximately 32% of the Australia-wide population.
The significance of prioritizing mental health care is supported by the evidence that poor mental health in AYA cancer patients is associated with increased risk of both hospital admission and extended length of hospital stay, as suggested from a national AYA population data in the United States. 19 In addition, mental disorders in AYA cancer patients are more likely to last lifelong based on a synthesis of evidence, 20 and poor mental health is associated with unhealthy lifestyles that include increased consumption of sugar and decreased vegetable intake. 21 These results call for timely identification, preventive efforts, and psycho-oncological intervention for psychological comorbidity.
The purposes of this study were (1) to assess the prevalence of mental health service contacts among AYA cancer patients after cancer diagnosis; (2) to identify groups with elevated odds of a mental health service contact by sociodemographic and cancer characteristics, and by precancer mental health contacts; and (3) to investigate associations of mental health service contacts with cancer mortality in the AYA population. We chose ages 15–24 years as the AYA range for consistency with national reporting practices in Australia. 7
Methods
Study cohort
The study population comprised AYA patients in NSW diagnosed with invasive cancer in 2005–2017, as recorded in the population-wide NSW Cancer Registry. Registry data are collected by law under provisions of the NSW Public Health Act 2010 and the Cancer Institute (NSW) Act 2003, and processed in accordance with international registry standards. 22 The registry is administered by Cancer Institute NSW, the NSW Government agency responsible for monitoring cancer control and supporting service planning, delivery, and evaluation.
The registry records primary cancer topography site; morphology; diagnosis date; residential area (to infer sociodemographic characteristics); and death dates and causes, using data provided by pathology and hematology laboratories, the NSW Registry of Births, Deaths, and Marriages, and the National Death Index. 23
Mental health recording and data sources
Mental health service contact data for the 5-year periods before and after cancer diagnosis were extracted from the NSW Admitted Patient Data Collection (APDC) and universal health insurance claims (i.e., the Medical Benefits Schedule [MBS] and Pharmaceutical Benefits Scheme [PBS]). Cancer diagnosis referred to the first primary cancer diagnosis as second cancer diagnosis was beyond the 5-year observation period in this study based on the result from the national data. 7
The APDC covers all hospital separations from public and private hospitals and day procedure centers in NSW. It includes information on primary diagnoses and up to 50 supplementary diagnoses extracted from clinical records by trained coders and classified using the Australian Modification of the International Classification of Diseases, Tenth Edition (ICD-10-AM). The ICD codes of F01–F99, X60–84, R45.81, and Y87 were used, including all APDC diagnosis fields for 2000–2022. 24
These codes for mental disorders included mental and behavioral disorders attributed to psychoactive substance use; psychotic, anxiety, and stress-related disorders; behavioral syndromes associated with physiological disturbances and physical factors; disorders due to developmental and childhood illness; disorders of adult personality and behavior; and intentional self-harm and suicidal ideation. Mental disorder diagnosis was categorized as depression/anxiety, self-harm and suicidal ideation, developmental and childhood mental illness, or other mental health conditions (which included less frequent conditions such as psychotic disorders and eating disorders).
Universal health insurance data from the Medical Benefits Schedule (MBS) and PBS were extracted on medical services and medications, respectively, for 2000–2022.25,26 Mental health contacts in this study included having subsidized mental health-related services provided by psychiatrists, general practitioners, clinical psychologists, other psychologists, and allied health professionals. 27 PBS claims also were used as contacts when there was receipt prescriptions for antipsychotics, anxiolytics, antidepressants.28,29
A mental health service contact was defined for this study as having at least one mental disorder diagnosis recorded on the APDC and/or at least two mental health care episodes indicated by MBS data and/or PBS data, as in previous Australian research. 30
Data linkage
Linkage was performed by the Centre for Health Record Linkage for NSW-based data sets using probabilistic matching and similarly by the Australian Institute of Health and Welfare for Commonwealth-based data. After linkage, deidentified data were stored and analyzed in the Secure Unified Research Environment facility, a purpose-built remote-access data repository and computing environment. 31
Other study factors
Sociodemographic factors investigated included age, gender, country of birth, residential remoteness, and socioeconomic disadvantage. Cancer factors included diagnostic year, cancer type, and stage (degree of spread) at diagnosis for solid tumors. These data items were retrieved from the NSW Cancer Registry. Information on death dates and causes was obtained from the NSW Cancer Registry.
AYA age at the time of cancer diagnosis was categorized as 15–19 or 20–24 years to conform with national cancer reporting by the Australian Institute of Health and Welfare. 7 Country of birth was classified as Australia, other predominantly English-speaking or non-English speaking countries, using the Standard Australian Classification of Countries (2nd edition), 32 and as detailed previously. 33 Residential area was classified as major city, inner regional, outer regional or remote/very remote, using the Australian Standard Geographical Classification Remoteness index. 34
Socioeconomic status was determined by place of residence at census collection district level and coded using the socioeconomic index for areas (SEIFA) that classifies areas by index of relative socioeconomic disadvantage quintiles. 35
Cancer type was classified using the WHO International Classification of Diseases for Oncology (ICD-O-3), 36 consistent with coding undertaken by the Australian Institute of Health and Welfare, using combination ICD-O-3 codes for topography and morphology. 7 It was grouped into 11 major groups: central nervous system, bone, soft-tissue sarcoma, germ cell, melanoma, thyroid, colorectal, leukemia, lymphoma, lip/oral/pharynx cancer, and others.7,8 Stage (degree of spread) of solid cancers was classified as localized, regional spread to adjacent organs, regional spread to regional lymph nodes, distant metastases, or unknown. 37 Diagnostic period was categorized as 2005–2009, 2010–2014, and 2015–2017. Death cause was classified as cancer death or death from another cause using WHO ICD-10 criteria. 36
Data analysis
The prevalence of having mental health contacts was calculated using the number of people with a relevant hospital diagnosis or at least two mental health services funded by MBS/PBS as the nominator and the total number of persons recorded in the data as the denominator and was expressed as a percentage. Sociodemographic and cancer items were compared by mental health service contact status using the Pearson chi-square test, substituting Fisher's exact test for small numbers.
Unadjusted and adjusted odds ratios with 95% confidence intervals (aOR, 95% CIs) for a mental health service contact within 5 years after cancer diagnosis were derived using univariable and multivariable logistic regression adjusted for sociodemographic characteristics (i.e., age, gender, country of birth, residential remoteness, and SEIFA socioeconomic disadvantage status of residential area), and cancer characteristics (i.e., diagnostic period, cancer type, and degree of spread), and whether a mental health service contact had presented in the 5 years preceding diagnosis.
Kaplan–Meier disease-specific product-limit survival estimates were calculated and compared by mental health service contact. Vital status was followed until December 31, 2022. Adjusted subhazard ratios with 95% CIs (aSHRs, 95% CIs) for cancer-specific death were calculated by mental health contact status, adjusting for sociodemographic, cancer factors, and precancer mental health status, using multivariable competing risk regression with other causes of death treated as competing events. 38
A series of sensitivity analyses were conducted: first, a small group (n = 47) without any linked APDC and MBS/PBS data was excluded from the analysis; second, MBS mental health care service was excluded; third, both mental health care service contact items from MBS and PBS medications other than antidepressant/anxiolytics were excluded, to align with hospital anxiety/depression/self-harm/suicidal ideation diagnosis, and lastly, all mental health service contacts related to MBS and/or PBS were excluded.
Data analyses used STATA release 16. 39
Results
Characteristics of the study population
Of the 3988 AYA cancer cases in NSW diagnosed with invasive cancer in 2005–2017, 62% were aged 20–24 years and 38% were aged 15–19 years at diagnosis; and 52% were men and 48% were women. Table 1 gives other socioeconomic and cancer characteristics, and the vital status at the end of 2022 of the cohort.
Population Characteristics of New South Wales Adolescent and Young Adult Cancer Patients Diagnosed in 2005–2017 by Mental Health Service Contact in the 5 Years Postcancer Diagnosis
MH service contacts operationally defined as having a relevant hospital diagnosis or inference of a mental health condition from repeated (>2 times) mental health services funded by MBS/PBS. 29
p from chi-square test (note: excluding unknown).
AYA, adolescent and young adult; CNS, central nervous system; MBS, Medical Benefits Schedule; MH, mental health; NSW, New South Wales; PBS, Pharmaceutical Benefits Scheme SEIFA, socioeconomic index for areas.
Prevalence of a mental health service contact
Of the 3998 AYA cancer patients, 3951 had linked records from inpatient and/or MBS and/or PBS data (Fig. 1). Approximately 21% of all AYA cancer patients had evidence of prior mental health service contact based on hospital inpatient admissions, MBS, or PBS claims (Fig. 1). By comparison, the prevalence of a mental health service contact postcancer diagnosis was closer to 27%, including 583 new onset cases (15%) and 497 precancer cases (12%).
Population flow chart-linkage results. AYA, adolescent and young adult; MBS, Medical Benefits Schedule; NSW, New South Wales; PBS, Pharmaceutical Benefits Scheme.
Mental health conditions indicated by inpatient records increased from 6.9% in the 5 years before cancer diagnosis to 12.3% in the 5 years after the diagnosis, with a prominent increase for anxiety/depression diagnosis from 3.1% to 8.0%. Meanwhile self-harm/suicidal ideation remained at a similar level at 1.1% and 1.2%, respectively, and developmental and childhood illness remained unchanged at 0.9% (Table 2).
Prevalence of Any Mental Health Service Contact 5 Years Pre- and Postcancer Diagnosis Among Adolescents and Young Adults Diagnosed in New South Wales 2005–2017
Rate = number of persons with a record of mental disorder diagnosis from hospitalization data or at least two mental health service contacts funded by MBS or PBS divided by the total number persons recorded in the data. For example, the rate of 5-year overall precancer mental disorder diagnosis was 218 of 3149 = 6.9%.
MH service contacts operationally defined as having a relevant hospital diagnosis or inference of a mental health condition from repeated (>2 times) mental health services funded by MBS/PBS. 29
The proportion of people with at least two MBS claims doubled after cancer diagnosis, as related to mental health service contact for psychiatrist, general practitioners, clinical psychologists, and other allied health providers (Table 2). General practitioners were the most frequent provider type for MBS-funded mental health care, applying to 16% of AYAs before cancer diagnosis and 33% after cancer diagnosis. Corresponding percentages of AYAs having contacts with clinical psychologists were 16% and 20%, respectively (Table 2).
The overall percentage receiving repeated PBS medication relating to mental conditions increased from 16.8% precancer diagnosis to 41.1% postcancer diagnosis (Table 2). The most common medication claims were for antidepressants, which increased from 14.2% to 32.0%, followed by anxiolytics increasing from 4.1% to 13.8%, and antipsychotics increasing from 1.8% to 9.9%.
The diagnosis of mental health conditions peaked approximately 1 year before and 1 year after a cancer diagnosis. For example, among a total of 395 hospital episodes of mental disorders from the 210 patients (with anxiety/depression, self-harm/suicidal ideation, development and childhood illness, and others) before cancer diagnosis, nearly half (n = 186, 47%) were recorded within the first year prior; whereas after cancer diagnosis, hospital recorded mental disorders took up 46% within the first year (412 out of 887 episodes), 16% at the second year (143 out of 887), 15% (129 out of 887) at the third year, 12% (110 out of 887) at the fourth year, and reducing to 10% (93 out of 887) at the fifth year.
Although the number of mental health claims through the MBS and PBS was highest at 1 year before cancer diagnosis for the precancer period, there was an increase over time in the postcancer diagnosis period (Supplementary Table S1).
AYAs recorded as having a precancer mental health contact had approximately sixfold higher odds of a postcancer mental health contact in the unadjusted analysis. After adjusting for sociodemographic and cancer characteristics, the aOR was 5.69 (95% CI: 4.90–6.75) (Table 3). The aOR was elevated consistently across group subclassified by sociodemographic.
Odds Ratios (95% Confidence Interval) of Having a Postcancer Mental Health Service Contact by Sociodemographic and Cancer Factors Among Adolescent And Young Adult Cancer Patients Diagnosed in 2005–2017 in New South Wales
MH service contacts operationally defined as having a relevant hospital diagnosis or inference of a mental condition from repeated (>2 times) mental health services funded by MBS/PBS. 29
Adjusted ORs (95% CI) from multivariable logistic regression analysis, including age, gender, country of birth, residential remoteness, SEIFA quintile, diagnosis period, cancer type, cancer stage at diagnosis, and prediagnosis mental health status.
Sensitivity analysis: excluding those with no hospital and MBS and PBS records (n = 47).
CI, confidence intervals; OR, odds ratio.
Odds of a mental health service contact after cancer diagnosis by sociodemographic and cancer factors
Multivariable logistic regression indicated significant (p < 0.05) differences in aORs of a mental health service contact after diagnosis, with (1) higher odds for residents of inner regional areas (aOR 1.26, 95% CI: 1.04–1.55), more recent diagnosis years (aOR 1.37, 95% CI: 1.12–1.67 for 2015–2017 period), and in those with advanced stage (degree of spread) (aOR 2.15, 95% CI: 1.57–2.94 for stage IV); and (2) lower odds for men (aOR 0.77, 95% CI: 0.65–0.90), those born in non-English-speaking countries (aOR 0.72, 95% CI: 0.54–0.76), residents of less disadvantaged areas (aOR 0.77, 95% CI: 0.60–0.99 for the fourth SEIFA quintile), and those with cutaneous melanoma (aOR 0.59, 95% CI: 0.39–0.91) or thyroid cancers (aOR 0.64, 95% CI: 0.42–0.99).
The full regression model included age, gender, country of birth, residential remoteness, SEIFA quintile, diagnosis period, cancer type, cancer stage at diagnosis, and prediagnosis mental health status. Sensitivity analyses showed consistent findings when the 47 people without linked hospital/MBS/PBS were excluded (Table 3) or when the majority of mental health contacts related to depression/anxiety were specified using hospitalization record and PBS claims or hospital diagnosis only.
Mental health service contact and 5-year cancer survival and mortality
The overall 5-year disease-specific survival was 92% (95% CI: 91.4–93.1), increasing from 89% in the 2005–2009 diagnosis period to 95% in 2015–2017. Overall cancer survival differed by residential remoteness, socioeconomic disadvantage index, cancer type, and stage.
The overall 5-year survival for patients with a mental health contact postcancer diagnosis was 87.9% (95% CI: 85.8–89.8), which was lower than the corresponding survival of 93.9% (95% CI: 93.0–94.7) (p < 0.001) for patients without it (Table 3).
Table 4 summarizes the SHRs for cancer-specific mortality associated with a mental health contact postcancer diagnosis. Cancer patients with mental health service contact after cancer diagnosis had a 67% higher risk of cancer mortality than those without mental health service contact, even after accounting for sociodemographic and cancer characteristics (aSHR 1.67, 95% CI: 1.29–2.15). The association was not changed substantially when further adjusted for precancer mental health contact. Cancers contributing to this elevation included bone cancers, soft-tissue sarcoma, lymphomas, lip, oral and pharyngeal cancers, and other rare cancers.
The Probability of 5-Year Survival (%) and Subhazard Ratios (95% Confidence Interval) of Cancer Mortality by Postcancer Mental Health Service Contact Among Adolescent And Young Adult Cancer Patients Diagnosed in 2005–2017 in New South Wales
MH service contacts operationally defined as having a relevant hospital diagnosis or inference of mental condition from repeated (>2 times) mental health services funded by MBS/PBS. 26
Probability of 5-year survival from Kaplan–Meier disease-specific product-limit estimates, deaths followed to December 31, 2022.
SHR (95% CI) from competing survival regression analysis; adjusted SHR from multivariable analysis including age, gender, country of birth, residential remoteness, socioeconomic status, diagnosis period, cancer stage for solid cancer, and precancer mental health status. Cancer type adjusted for all cancers.
SHR, subhazard ratio.
The overall unadjusted noncancer mortality was associated with postcancer mental health contact with an SHR of 1.85, 95% CI: 1.37–2.50, and with an increase in SHR indicated for soft-tissue sarcoma at 2.78, 95% CI: 1.18–6.53, but SHR elevations were not suggested after adjusting for cancer stage, along with precancer mental health status, diagnosis period, and sociodemographic predictors (Supplementary Table S2).
Discussion
The Australian government provides funding through universal MBS and PBS health insurance for cancer care. This complements hospital and other services provided by state and territory governments. 40 Data from these services were used in this study to produce service contacts for mental health conditions in the 5 years preceding and 5 years after AYA cancer diagnosis. These service contacts were used for comparative analyses of the prevalence of mental health service contact. The breadth of population coverage of these services in aggregate from which they were drawn was close to complete such that these data are likely to be representative.
Among NSW AYA cancer patients aged 15–24 years at diagnosis in 2005–2017, 27% had mental health service contacts suggesting the presence of a mental health condition dominated by depression/anxiety (20%) in the 5 years after cancer diagnosis. This proportion is within the range reported from literature reviews of the prevalence of clinical distress from depression, anxiety, and somatic distress experienced by AYA cancer patients. 41 About 25% to 48% were reported to have elevated clinical distress in more recent studies in Australia, the United Kingdom, Canada, and the United States.11–13
These prevalence figures were lower than the 61% found in a Canadian population study of mental health conditions in AYAs that included psychiatric disorders, anxiety, depressive states, and trauma- and stress-related disorders, perhaps due to the longer follow-up period among 5-year cancer survivors in that study. 14
Although these studies were not methodologically equivalent due to differences in study periods, age profiles, study design, assessment procedures, and types of mental health conditions studied, all found the prevalence of mental health conditions in AYAs to be substantial. This study also found an increasing prevalence of mental health conditions in AYAs after a cancer diagnosis. This underscores the need for ongoing mental health surveillance along the care pathway, plus additional medical checks where indicated.17,18,42
The needs reflect in the study of the increased mental health service contacts funded by MBS and PBS during the 5-year period after cancer diagnosis could be due to personal health-seeking behavioral change, and due to the supportive and easy access to health services after cancer treatment as psychosocial care is a national priority for improving AYA cancer outcomes in Australia,17,18 and the care is outlined as an essential part of the optical care for AYA cancer patients. 43
We found that women, the Australia born, residents of regional and socioeconomically disadvantaged areas, and those with more advanced solid tumors had increased odds of mental health service contacts after AYA cancer diagnosis. This underscores the variability in the prevalence of mental health service contacts and the importance of continuing surveillance practices to inform the targeting health care services. Psychological screening during oncology visits would assist identification of patients needing referral to psychosocial services. 43
In partnership with Cancer Australia, Canteen has developed the National Service Delivery Framework for Adolescents and Young Adults with Cancer, which provides guidance in delivering psychosocial care and support along the care pathway.12,43 Evaluation of the impact of this framework will be important.
Besides sociodemographic factors, cancer types and stage (degree of spread) were associated in this study with mental health contacts post-AYA cancer diagnosis. For cancers with high survival (e.g., thyroid and cutaneous melanoma and cancers diagnosed at an early stage), the likelihood of contact with mental health care was found to be low. By comparison, brain and central nervous system malignancies were associated with an elevated likelihood, as reported in a previous systematic review. 44
The underlying processes associated with this increased likelihood were not identified, although possible causes could include the impact of cancer and its intensive treatment on physical and mental endurance. Communication between clinicians and AYA patients clearly is important to explore causes and will remain a fundamental component of care pathways.
We found precancer mental health contact strongly predicts postcancer mental health contact across all subgroups and study factors. The results are consistent with the literature reports.15,16 Precancer mental health service contact could, therefore, serve as an alert where increased emphasis on the psychosocial support of AYA patients might be required. It is also relevant that a postcancer mental health contact was associated with lower cancer survival. Other research evidence has also indicated that mental conditions among AYAs are predictors of an elevated risk of other cancers and comorbid diseases. 45
As demonstrated in this and previous studies,8,46 mental conditions or the related service contacts in the 5 years after cancer diagnosis foreshadowed an increased risk of cancer mortality. This risk persisted after adjustment for sociodemographic and cancer factors that were predictive of cancer mortality. The increased risk of mortality associated with postcancer mental health care was statistically significant in AYA patients for the following cancers: bone cancer; soft-tissue sarcoma; lymphoma; lip, oral, and pharynx cancers; and other rare cancers.
These severe or life-threatening cancers may be associated with a greater risk of psychiatric sequelae due to the increased fear of recurrence, 47 undergoing longer and aggressive treatment (e.g., potentially limb amputation and facial disfigurement), and anxiety related to repeated scans post-treatment.
We suspect multiple factors would have contributed to the increased mortality observed in this study and others when mental health conditions or service contacts were reported postcancer diagnosis. For example (1) mental health conditions could impact on adherence to treatment recommendations; (2) mental health treatments might interact negatively with cancer treatments 48 ; and (3) having both cancer and mental health conditions may combine to disrupt dietary practices, 21 physical activity, and sleep.49,50
In addition, anxiety and depression may activate the hypothalamic–pituitary–adrenal axis and autonomic nervous system with corresponding deregulation of circadian cortisol rhythm and release of catecholamine that affect immunity function. 51 Further research is needed to explore these and other possible mechanisms. Mental health service contact could be a proxy for residual confounding by cancer severity, such as rate of cancer progression, and treatment responsiveness of cancer. These characteristics were not measured at diagnoses and may lead to more mental health service contact and higher cancer mortality. Potential for confounding aside, results point to a need to strengthen proactive health promotion and psychosocial support for AYA cancer patients. 52
This is the first population-wide study of mental health in the AYA cancer population of NSW using linked administrative data, which explored associations of mental health service contacts with increased cancer-specific mortality at a population level independent of a variety of sociodemographic and cancer factors that were predictive of mortality. Detailed death information facilitated cancer-specific mortality analysis.
Mental health service records were from multiple sources, including hospital inpatient, and MBS and PBS records, which enabled insight into the distribution of mental health conditions. The use of psychotropic medication has been found effective in identifying people with mental disorders (e.g., depression associated with the use of antidepressants). 53 Notably, sensitivity analyses in this NSW study yielded consistent results, irrespective of the data sources used.
Study limitations should be noted. Mental health service contacts derived from our data sources likely missed some conditions where treatment was not provided through NSW hospitals or funded through the MBS/PBS. We believe this to be a relatively small number, given the range of mental health‑related services funded in Australia by the various levels of government and/or individuals with services delivered by both government and nongovernment providers. 54 Also changes in codes used for mental health services through MBS-funded and other care outlets could have reduced data consistency.
Another limitation is that closer medical surveillance postcancer diagnosis may have increased detection of mental health service contacts, artificially inflating the size of the increase in prevalence over precancer levels. In addition, increase in prevalence of mental health conditions could be due to a background of increasing mental health symptoms across the population of adolescent and young adulthood residents, a period when these symptoms tend to peak even among those without cancers.
A further issue was that hospital diagnosis codes used for mental health conditions (ICD-10: F01–F99, X60–X84, R45.81, Y87) and MBS codes were not well aligned, and matching with PBS codes proved difficult. Nonetheless, results for anxiety/depression/intentional self-harm/suicidal ideation, the most common mental illnesses in this cohort, were similar in sensitivity testing of data selection. There are limitations from reliance on health system data, including gaps in describing cancer recurrence/progression, personal behavioral and family factors, and other social factors such as socioeconomic disadvantage index derived from residential post code. 32 Further investigations could be undertaken using national census data in which individual socioeconomic factors may be applied.
Although further research is clearly needed, the mental health service contacts used in this study have suggested important relationships between mental health status and social and cancer characteristics, and likely impact on population-wide cancer mortality. This and further biomedical, health systems, and clinical research will be complementary.
Conclusion
Over a quarter (27%) of AYA cancer patients in NSW were found to have clinically diagnosed mental health conditions or can be inferred to have received mental health care from hospital inpatient and MBS/PBS data. Postcancer mental health service contact was more prevalent in women, the Australian born, and residents of regional and socioeconomically disadvantaged areas. Mental health service contact was more prevalent, as measured, after than before cancer diagnosis. Postcancer mental health service contact was associated with higher cancer-specific mortality. Although there were data limitations that detracted from the value of the study, results underscore the importance of mental health care as part of the AYA cancer care pathway and point to profiles of those most in need.
Footnotes
Acknowledgments
We thank the Centre for Epidemiology and Evidence, NSW Ministry of Health, and the Cancer Institute NSW for the use of these data.
Authors' Contributions
Study concept was contributed by M.L. and D.R. Study design was taken care by M.L., G.S., and D.R. Data acquisition was done by D.R. and T.A.O. Data analysis was carried out by M.L. Data interpretation was by M.L. and D.R. Article writing was by M.L. Review of the article was by M.L., G.S., D.R., and T.A.O. All authors read and approved the final version of the article.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
