10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article

Adverse Mental Health Outcomes in a Population-Based

Cohort of Survivors of Childhood Cancer
Paul C. Nathan, MD, MSc 1,2; Alex Nachman, MSc1; Rinku Sutradhar, PhD 2,3,6; Paul Kurdyak, MD, PhD3,4;
Jason D. Pole, PhD5,6; Cindy Lau, MPH3; and Sumit Gupta, MD, PhD 1,2

BACKGROUND: The elevated risk for physical late effects in childhood cancer survivors (CCS) is well documented, but their risk for
mental health problems is less well described. METHODS: The authors assembled a cohort of all 5-year CCS who were diagnosed
before age 18 years and treated in an Ontario pediatric cancer center between 1987 and 2008. Patients were matched to population
controls and linked to health administration databases. The authors calculated rates of mental health care visits (family physician, psy-
chiatrist, emergency department, hospitalization) and the risk for a severe mental health event (emergency department, hospitaliza-
tion, suicide). Outcomes were compared using recurrent event and survival analyses. RESULTS: Compared with 20,269 controls, 4117
CCS had a higher rate of mental health visits (adjusted relative rate [RR], 1.34; 95% confidence interval [CI], 1.12-1.52). Higher rates
were associated with female gender (RR, 1.39; CI, 1.10-1.75; P 5 .006) and being diagnosed at ages 15 to 17.9 years (compared with
ages 0-4 years: RR, 1.81; 95% CI, 1.17-2.80; P 5 .008). Cancer type, treatment intensity, and treatments targeting the central nervous
system were not significant predictors. Survivors were at increased risk for a severe event compared with controls (adjusted hazard
ratio, 1.13; 95% CI, 1.00-1.28; P 5 .045). CCS who were diagnosed with cancer at age 4 years or younger were at greatest risk: 16.3%
(95% CI, 13.2%-19.8%) had experienced a severe event by age 28 years. CONCLUSIONS: CCS experienced higher rates of mental
health visits and a greater risk for a severe event than the general population. Survivors of adolescent cancer have a higher rate of
mental health visits overall, whereas survivors of cancer before age 4 years have a markedly elevated risk of severe events. Cancer
2018;124:2045-57. V C 2018 American Cancer Society.

KEYWORDS: cancer survivors, health services, mental health, pediatric cancer.

INTRODUCTION
Over 80% of children with cancer will become long-term survivors.1,2 Consequently, there are over 400,000 childhood
cancer survivors (CCS) in the United States, many of whom will experience chronic morbidity as a result of their treat-
ment.3 Research and program development has focused on the physical sequelae of cancer, with less attention to its psy-
chological impacts. Because mental illness often has its first onset during adolescence,4 CCS may be vulnerable to the
trauma of their cancer experience and the impact of therapy on their mental health. Although some studies have reported
higher rates of suicidal ideation, post-traumatic stress disorder, and symptoms of depression and anxiety compared with
siblings,5-10 others have not indicated an elevated risk for adverse outcomes.11-14 Several studies have demonstrated supe-
rior mental health in CCS compared with controls.15-17
Some of the inconsistencies in these observations may be because of the relative advantages and limitations of differ-
ent methods for assessing mental health outcomes. For example, large cohort studies that use self-report questionnaires
often have a high number of nonresponders,5,18 some of whom may be at elevated risk of poor mental health.19 Self-
report may not accurately reflect the true incidence of psychiatric symptoms because it is susceptible to social desirability
response bias.20,21 Research linking cancer survivors to their health care administrative records can provide objective data
on whole populations, but such studies use contact with the health care system or consumption of psychotropic medica-
tion as a surrogate for poor mental health outcomes and thus do not directly assess mental health. One such population-
based study from Denmark demonstrated that, among 7085 CCS, there was an elevated risk of hospital contact for a
Corresponding author: Paul C. Nathan, MD, MSc, Division of Hematology/Oncology, The Hospital for Sick Children, 555 University Avenue, Toronto ON M5G
1X8, Canada; paul.nathan@sickkids.ca
1
Division of Hematology/Oncology, The Hospital for Sick Children, Toronto, Ontario, Canada; 2Institute of Health Policy, Management, and Evaluation, University
of Toronto, Ontario, Canada; 3Institute for Clinical Evaluative Sciences, Toronto, Ontario, Canada; 4Centre for Addiction and Mental Health, Toronto, Ontario, Can-
ada; 5Pediatric Oncology Group of Ontario, Toronto, Ontario, Canada; 6Dalla Lana School of Public Health, University of Toronto, Ontario, Canada

The opinions, results, and conclusions reported in this article are those of the authors and are independent from the funding sources. No endorsement by the
Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be concluded. Parts of this material are
based on data and information compiled and provided by the Canadian Institutes of Health Information. However, the analyses, conclusions, opinions, and state-
ments expressed herein are those of the authors and not necessarily those of the Canadian Institutes of Health Information.
Additional supporting information may be found in the online version of this article.

DOI: 10.1002/cncr.31279, Received: October 20, 2017; Revised: December 20, 2017; Accepted: January 4, 2018, Published online February 22, 2018 in Wiley
Online Library (wileyonlinelibrary.com)

Cancer May 1, 2018 2045


Original Article
mental disorder compared with their siblings.22 However,
studies like this may be hindered by incomplete treatment
data or an inability to capture outpatient visits or care
from nonmedical providers of mental health care, such as
psychologists, social workers, and complementary or
alternative medicine care providers. A complete under-
standing of mental health outcomes in CCS likely
requires data from multiple types of studies, including
administrative data that capture both inpatient and outpa-
tient health care use.
Canadian health care is delivered through a provin-
cial, single-payer, universal insurance system. Canada’s
most populous province, Ontario, maintains a registry of
all children treated for cancer at the province’s 5 tertiary
pediatric centers. This registry captures detailed diagnos-
tic and treatment data. These can be linked to health care
administrative records maintained by the province’s
health care system, facilitating the capture of all outpatient
and hospital encounters with the health care system for a
mental health complaint in this population-based cohort
of CCS to assess the risk for adverse mental health
outcomes.
MATERIALS AND METHODS
Study Population
After obtaining Research Ethics Board approval, we iden-
tified eligible survivors using the Pediatric Oncology
Group of Ontario’s Networked Information System
(POGONIS), which has captured all pediatric (ages 0-
17.9 years at diagnosis) cancers treated at any of Ontario’s
5 pediatric cancer centers since 1985. Approximately
96% of children ages 0 to 14.9 years and 46% of adoles-
cents ages 15 to 17.9 years receive cancer care in a pediat-
ric center and thus are documented in POGONIS.23
Individuals in POGONIS were considered eligible for
this analysis if they had been diagnosed with a first cancer
between January 1, 1987 and December 31, 2008; had
survived greater than 5 years since their last pediatric can-
cer event (the latest of their initial diagnosis or a relapse or
subsequent malignant neoplasm [SMN] before age 18
years); had reached age 16 years (the age at which adminis-
trative data can be used to capture all types of mental
health visits) before the end of the follow-up period
(December 31, 2013); and had provided at least 1 year of
follow-up after reaching 5 years from their last pediatric
cancer event. For each survivor, we randomly selected 2 to
5 controls from the general population matched by birth
year and month, sex, and residential location (using postal
code) at the time of their cancer diagnosis.
2046
10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Outcome Definition
We deterministically linked survivors and controls to
administrative health databases held at the Institute for
Clinical Evaluative Sciences using their unique health
card number, which was encoded to ensure anonymity.
The administrative health databases accessed were the
Registered Persons Database, the Ontario Health Insur-
ance Plan Claims Database (OHIP), the National Ambu-
latory Care Reporting System, the Canadian Institutes of
Health Information Discharge Abstract Database the
Ontario Mental Health Reporting System (OMHRS),
the Office of the Registrar General-Deaths, and the
Ontario Cancer Registry.
Because the OMHRS database only captures psychi-
atric admissions in patients aged 16 years or older, mental
health visits were documented starting at an index date
(the later of age 16 years or 5 years after the last pediatric
cancer event) until the earliest of the end of study follow-
up, relapse or an SMN during adulthood, or death. Con-
trols were assigned the same index date and follow-up
period as their corresponding survivor. Visits to a family
physician for a mental health complaint and any visit to a
psychiatrist were identified from OHIP physician billings
using a validated algorithm.24 Emergency department
(ED) visits for a mental health indication were identified
using data from OHIP and the National Ambulatory
Care Reporting System, and hospitalizations were
obtained from the Canadian Institutes of Health Informa-
tion Discharge Abstract Database and OMHRS. Suicides
were identified from the cause of death documented in
the Office of the Registrar General-Deaths. A mental
health visit was defined as a visit to a family physician, psy-
chiatrist, or ED or a hospitalization. A severe mental health
event was defined as an ED visit, hospitalization, or sui-
cide. ED visits and hospitalizations contributed to both
the mental health visits and severe mental health event
outcomes. If more than 1 severe event occurred within 24
hours, then the event was classified as the most severe of
an ED visit, a hospitalization, or a suicide. To determine
whether any broad groups of psychiatric diagnoses (eg,
anxiety or mood/affective disorders) contributed dispro-
portionately to severe events, we subclassified ED visits
and hospitalizations into the following diagnostic subcate-
gories using International Statistical Classification of
Diseases-10th Revision and Diagnostic and Statistical Man-
ual of Mental Disorders, Fourth Edition codes (Supporting
Table 1; see online supporting information): substance-
abuse, psychotic disorders, mood/affective disorders, anx-
iety disorders, and selected disorders of adult personality
and behavior.25
Cancer May 1, 2018
 10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Mental Health Outcomes in CCS/Nathan et al

TABLE 1. Description of Survivors and Controls

No. (%)

Survivors, Controls,
Variable N 5 4117 N 5 20,269 P

Sex .97
Male 2234 (54.3) 11,005 (54.3)
Female 1883 (45.7) 9264 (45.7)
Age at diagnosis, y 1.00
0-4 1347 (32.7) 6645 (32.8)
5-9 1052 (25.6) 5161 (25.5)
10-14 1181 (28.7) 5818 (28.7)
15-18 537 (13.0) 2645 (13.0)
Income quintile at diagnosis < .001
1 (Lowest) 724 (17.6) 4177 (20.6)
2 825 (20.0) 4072 (20.1)
3 826 (20.1) 3992 (19.7)
4 855 (20.8) 4032 (19.9)
5 (Highest) 838 (20.4) 3730 (18.4)
Missing 49 (1.1) 266 (1.3)
Follow-up time from index: Median [range], y 7.5 [1.0-21.9] 7.5 [1.0-21.9] .97
Total person-years of follow-up 35,027 172,409
Cancer classification: ICCC-3
Acute lymphoblastic leukemia 1087 (26.4)
Other leukemias 189 (4.6)
Hodgkin lymphoma 386 (9.4)
Other lymphomas 307 (7.5)
Central nervous system tumors 808 (19.6)
Neuroblastoma 144 (3.5)
Retinoblastoma 92 (2.2)
Renal tumors 240 (5.8)
Hepatic tumors 39 (0.9)
Bone tumors 202 (4.9)
Soft tissue sarcomas 256 (6.2)
Germ cell tumors 164 (4.0)
Other epithelial tumors 170 (4.1)
Unspecified malignancies 33 (0.9)
Hematopoietic stem cell transplantation
No 3854 (93.6)
Autologous 114 (2.8)
Allogenic 149 (3.6)
Surgery
No 1868 (45.4)
Yes 2249 (54.6)
Chemotherapy
No 1031 (25.0)
Yes 3021 (73.4)
Unknown 65 (1.6)
Cranial radiation
No 3230 (78.5)
Yes 887 (21.5)
Relapse or SMN during childhood
No 3776 (91.7)
Yes 341 (8.3)
ITR-3
1 413 (10.0)
2 1887 (45.8)
3 1323 (32.1)
4 488 (11.9)
Unknown 6 (0.2)
High-dose methotrexate
No 3229 (78.4)
Yes 823 (20.0)
Unknown 65 (1.6)
Corticosteroid
No 2367 (57.5)
Yes 1685 (40.9)
Unknown 65 (1.6)

Cancer May 1, 2018 2047


Original Article
TABLE 1. Continued
Variable
Cyclophosphamide equivalent dose, mg/m2
0 2203
1-3999
4000-7999
>8000
Unknown
Treatment era
1987-1993
1994-2001
2002-2008
Abbreviations: ITR-3, Intensity of Treatment Rating Scale, third edition; ICCC-3, International Classification of Childhood Cancer, third edition; SMN, second
malignant neoplasm.
Definition of Covariates
Variables that were considered potential predictors of out-
come included sex, age at diagnosis, income quintile
(using census data to assign an average neighborhood
household income to the postal code at diagnosis), cancer
diagnosis,26 and the occurrence of a relapse or SMN
before age 18 years. Receipt of chemotherapy agents with
plausible central nervous system (CNS) effects also was
considered; we recorded exposure to corticosteroids used
as chemotherapy agents and high-dose methotrexate (1
g/m2 per dose). We captured receipt of alkylating
agents,27 cranial radiation, and hematopoietic stem cell
transplantation (HSCT). The overall intensity of treat-
ment was classified using the Intensity of Treatment Rat-
ing Scale.28 We defined 3 treatment eras based on year of
diagnosis.
Statistical Analysis
We generated descriptive statistics to compare the distri-
bution of baseline characteristics between survivors and
controls (Table 1). Frequencies and percentages summa-
rized categorical variables, and medians and minimum-
maximum ranges summarized continuous variables. We
calculated the crude rate of visits by survivors and controls
as well as the crude relative rate.
An individual-level analysis modeled the association
between survivorship and the rate of mental health visits.
Because visits can occur numerous times for each individ-
ual over the course of observation, Andersen-Gill recur-
rent event multivariable regression models were used,29,30
adjusting for prior mental health visits. A robust sandwich
variance estimation approach accounted for the matched
design. The Anderson-Gill model also was used to deter-
mine predictors of the rate of mental health visits only
2048
10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
No. (%)
Survivors, Controls,
N 5 4117 N 5 20,269 P
(53.5)
321 (7.8)
1084 (26.3)
444 (10.8)
65 (1.6)
1520 (36.9)
2003 (48.7)
594 (14.4)
among survivors. To illustrate the outcome rate over time,
nonparametric analyses were performed by estimating the
mean cumulative functions,31,32 which provided an esti-
mate of the mean cumulative number of mental health
visits over time. Cox proportional hazard regression was
used to model the time to a first severe mental health
event, and the cumulative incidence function approach
was used to determine the risk of having a severe mental
health event. We decided, a priori, to include age, sex, and
income quintile in both multivariable regression models
in addition to any variables in the univariate analyses that
were significant at P < .05. All analyses were performed
using SAS statistical software (version 9.3; SAS Inc,
Cary, NC).
RESULTS
Of 7884 children who received treatment between 1987
and 2008, 4117 were eligible for inclusion (Fig. 1). Of
these, 3992 children (95.7%) could be matched to 5 pop-
ulation controls, and the remainder could be matched to
2 to 4 controls (Table 1). Both groups were followed for a
median of 7.5 years from their index date (minimum-
maximum range, 1.0-21.9 years).
Table 2 presents the crude rates of mental health
care visits to family physicians, psychiatrists, and all visits
combined. Survivors had a higher rate of visits to family
physicians (unadjusted rate ratio [RR], 1.32; 95% confi-
dence interval [CI], 1.25-1.39) and psychiatrists (unad-
justed RR, 1.56; 95% CI, 1.41-1.72). After adjusting for
sex, income quintile, age category and a history of prior
events, survivors had an elevated rate of any mental health
care visit (family physician, psychiatrist, ED, or hospitali-
zation) compared with controls (RR, 1.34; 95% CI, 1.12-
1.52) (Fig. 2A).
Cancer May 1, 2018

Figure 1. Cohort creation is illustrated. SMN indicates subse-
quent malignant neoplasm.
Although survivors individually did not have an ele-
vated risk for ED visits, hospitalizations, or suicide (Table
3), their risk for experiencing any severe mental health
event was significantly increased. After adjusting for sex,
income quintile, and age category, the hazard ratio (HR)
for a severe event in survivors versus controls was 1.13
(95% CI, 1.00-1.28; P 5 .045) (Fig. 2B). Within subca-
tegories of psychiatric diagnoses, anxiety disorders were
the most prevalent among both groups (Table 4). How-
ever, the only category in which survivors were at signifi-
cantly greater risk was psychotic disorders (HR, 1.78;
95% CI, 1.09-2.89).
Table 5 displays the recurrent event regression analy-
sis for the rate of mental health visits among survivors. In
the multivariable model, survivors who were female and
aged 15 years or older at diagnosis had a significantly
higher rate of mental health visits than males or those who
were younger at diagnosis. A comparison of the
Cancer May 1, 2018
10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Mental Health Outcomes in CCS/Nathan et al
cumulative function of visits by time from the index date,
stratified for age group at diagnosis, is illustrated in Figure
2C. None of the variables income quintile at diagnosis,
cancer type, relapse or SMN during childhood, cranial
radiation, HSCT, high-dose methotrexate, or treatment
era were associated with the rate of mental health visits.
We also assessed the risk factors for a severe event in
survivors (Table 6). Survivors who had received cranial
radiation had a reduced risk of a severe event compared
with those who did not. Survivors who were in the highest
income quintile at diagnosis were significantly less likely
to experience a severe event than those in the lowest quin-
tile. Survivors who were diagnosed at ages 5 to 9, 10 to
14, or 15 years were less likely than those diagnosed at
ages 0 to 4 years to experience a severe event, although this
difference was only statistically significant in the groups
ages 5 to 9 and 10 to 14 years (Fig. 2D). Among those sur-
vivors who were ages 0 to 4 years at diagnosis, 131 had a
severe event during study follow-up. By age 28 years, the
cumulative incidence of a severe event in this subgroup
was 16.3% (95% CI, 13.2%-19.8%). The most frequent
diagnostic subcategories were anxiety (41.2%), substance
abuse (34.4%), and mood/affective disorders (24.4%).
Psychotic disorders and personality disorders occurred in
less than 10% of these survivors. Among survivors who
were diagnosed at ages 0 to 4 years, only being in the high-
est income quintile at diagnosis was associated with a
reduced risk of a severe event (Table 7).
DISCUSSION
In this provincial cohort of over 4000 CCS, survivors had
a 34% higher rate of medical visits for a mental health
complaint than the general population. More than 40%
of survivors had at least 1 visit, and 90% of encounters
were with family physicians and psychiatrists. Visits were
significantly more frequent among females and survivors
of adolescent cancers. Studies have demonstrated that
adolescents are vulnerable to psychological challenges
after a cancer diagnosis. Counseling and psychological
support are among the most common needs expressed by
these survivors,33 who often report symptoms of post-
traumatic stress, anxiety and depression, and fears of can-
cer recurrence.34 Many adolescents who require mental
health services after cancer do not receive them,35,36 sug-
gesting that the documented rate of visits underestimates
their true burden of psychological morbidity.
Although ED visits and hospitalizations were less
common than visits to a family physician or psychiatrist,
survivors had a 13% increase in their risk for a severe
event. Suicides were rare, a finding that mirrors other
2049
 10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article

TABLE 2. Crude Rate and Rate Ratio of Mental Health Visits to Family Physicians, Psychiatrists, and All
Health Care Providers in Survivors and Controls

Crude Rate per 1000 Person-Years (95% CI)

Survivors, Controls, Unadjusted RR

Level of Care N 5 4117 N 5 20,269 (95% CI)

Family Physician 68.7 (65.4-72.1) 52.0 (50.8-53.2) 1.32 (1.25-1.39)

Psychiatrist 16.2 (14.8-17.6) 10.4 (9.9-10.9) 1.56 (1.41-1.72)
All mental health visitsa 79.5 (75.9-83.3) 57.8 (56.5-59.2) 1.38 (1.31-1.45)

Abbreviations: CI, confidence interval; RR, rate ratio.

a
Visits to family physicians, psychiatrists, emergency departments, and hospitalizations.

Figure 2. Charts illustrate (A) the average cumulative number of mental health care visits among survivors and controls as a
function of attained age, (B) the cumulative incidence of a first severe event (emergency department visit, hospitalization, or sui-
cide) in survivors versus controls as a function of attained age, (C) the average cumulative number of mental health care visits
among survivors over time stratified by age group at diagnosis, and (D) the risk of a severe event (emergency department visit,
hospitalization, or suicide) in survivors according to attained age stratified by age group at diagnosis. Shaded areas in A and B
represent 95% confidence intervals around estimates.

studies indicating that, although suicidal ideation might population-based cohort study, which demonstrated a
be more frequent in survivors,6 suicide is not.37 Our find- 38% elevation in the risk for contact with hospital psychi-
ings are consistent with those from the Danish atric departments.22 In that study, children who were

2050 Cancer May 1, 2018

 10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Mental Health Outcomes in CCS/Nathan et al

TABLE 3. Cumulative Incidence by Age 30 Years of a Severe Mental Health Event (Emergency Department,
Hospitalization, or Suicide) in Survivors and Controls

Cumulative Incidence by Age 30 Years (95% CI), %

Event Survivors Controls Risk Ratio (95% CI)

Emergency department visits 11.46 (10.16-12.84) 10.13 (9.57-10.70) 1.12 (0.99-1.27)

Hospitalizations 1.47 (1.08-1.97) 1.27 (1.09-1.47) 1.15 (0.83-1.59)
Suicide 0.13 (0.03-0.39) 0.03 (0.01-0.08) 2.92 (0.70-12.23)
Any severe event 12.00 (10.70-13.39) 10.56 (10.00-11.14) 1.13 (1.00-1.28)

Abbreviation: CI, confidence interval.

TABLE 4. Cumulative Incidence of an Emergency Department Visit or Hospitalization by Age 30 Years

According to Subcategories of Psychiatric Diagnoses in Survivors and Controls

Cumulative Incidence of Any Severe Event by

Age 30 Years (95% CI), %

Event Survivors Controls Risk Ratio (95% CI)

Substance abuse 4.24 (3.44-5.16) 3.85 (3.50-4.23) 1.07 (0.87-1.32)

Psychotic disorder 0.92 (0.57-1.42) 0.50 (0.38-0.65) 1.78 (1.09-2.89)
Mood/affective disorder 3.05 (2.36-3.86) 2.27 (2.00-2.55) 1.24 (0.96-1.59)
Anxiety disorder 6.20 (5.23-7.28) 5.44 (5.02-5.89) 1.14 (0.96-1.35)
Other personality disorder 0.09 (0.04-0.14) 0.08 (0.04-0.14) 1.39 (0.39-4.99)

Abbreviation: CI, confidence interval.

TABLE 5. Univariate and Multivariable Recurrent-Measure Models of Predictors of Mental Health Care Visit
Rates in Survivors

Univariate Model Multiple Variable Model

Variable RR 95% CI P RR 95% CI P

Sex
Male Ref Ref
Female 1.39 1.11-1.75 .005 1.39 1.10-1.75 .006
Income quintile
1 Ref Ref
2 1.15 0.78-1.69 .471 1.16 0.79-1.70 .463
3 1.09 0.76-1.56 .630 1.09 0.76-1.56 .650
4 1.06 0.74-1.53 .737 1.07 0.73-1.56 .737
5 1.31 0.90-1.91 .164 1.31 0.90-1.92 .165
Age category, y
0-4
5-9 1.22 0.93-1.62 .157 1.15 0.84-1.56 .381
10-14 1.45 1.10-1.91 .008 1.36 0.98-1.88 .068
15 1.88 1.29-2.73 .001 1.81 1.17-2.80 .008
Cancer type
ALL Ref Ref
Other leukemia 1.14 0.43-2.98 .795 0.85 0.40-1.82 .676
Hodgkin lymphoma 1.20 0.82-1.77 .348 0.96 0.61-1.50 .857
Other lymphomas 1.37 0.82-2.28 .234 1.29 0.77-2.18 .337
CNS tumors 1.30 0.93-1.81 .128 1.28 0.90-1.82 .167
Neuroblastoma 0.93 0.57-1.52 773 1.08 0.66-1.77 .749
Soft tissue sarcomas 1.04 0.62-1.72 .895 0.99 0.59-1.66 .968
Retinoblastoma 0.66 0.36-1.22 .184 0.78 0.42-1.44 .426
Renal tumors 1.15 0.74-1.78 .532 1.23 0.79-1.91 .360
Hepatic tumors 0.99 0.47-2.11 .981 1.09 0.50-2.38 .820
Bone tumors 1.28 0.79-2.05 .314 1.17 0.71-1.92 .541
Germ cell tumors 1.43 0.76-2.68 .265 1.23 0.65-2.32 .529

Cancer May 1, 2018 2051

 10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article

TABLE 5. Continued

Univariate Model Multiple Variable Model

Variable RR 95% CI P RR 95% CI P

Other epithelial neoplasms 1.97 1.16-3.33 .012 1.58 0.89-2.80 .117

Unspecified malignancies 2.44 0.57-10.34 .227 2.59 0.66-10.18 .173
Relapse or SMN
No Ref
Yes 1.20 0.83-1.74 .333
CNS radiation
No Ref
Yes 0.80 0.60-1.06 .122
ITR
1 Ref
2 0.76 0.50-1.15 .194
3 0.68 0.44-1.04 .075
4 0.92 0.54-1.57 .762
Surgery
No Ref
Yes 0.91 0.72-1.15 .447
Chemotherapy
No Ref
Yes 0.79 0.62-1.01 .056
High-dose methotrexate
No
Yes 0.89 0.66-1.21 .468
Corticosteroids
No Ref
Yes 0.96 0.76-1.22 .740
Treatment era
1987-1993 Ref
1994-2001 0.91 0.71-1.16 .448
2002-2008 1.17 0.79-1.73 .435
HSCT
None Ref
Autologous 0.64 0.42-0.99 .046 0.71 0.43-1.16 .1738
Allogeneic 1.37 0.60-3.14 .455 1.68 0.93-3.04 .0867
Cyclophosphamide equivalent dose (mg/m2)
0 Ref
1-3999 0.78 0.56-1.09 .143
4000-7999 0.78 0.59-1.01 .061
8000 1.02 0.68-1.52 .931

Abbreviations: ALL, acute lymphocytic leukemia; CI, confidence interval; CNS, central nervous system; HSCT, hematopoietic stem cell transplantation; ITR,
Intensity of Treatment Rating Scale; Ref, reference category; RR, rate ratio; SMN, second malignant neoplasm.

TABLE 6. Demographic, Diagnostic, and Treatment Predictors of a Severe Psychiatric Event in Childhood
Cancer Survivors

Univariate Model Multiple Variable Model

Variable HR 95% CI P HR 95% CI P

Sex
Male Ref Ref
Female 1.19 0.96-1.47 .105 1.19 0.96-1.47 .111
Income quintile
1 Ref Ref
2 0.88 0.64-1.22 .438 0.89 0.64-1.23 .474
3 0.75 0.54-1.05 .092 0.76 0.55-1.07 .119
4 0.85 0.62-1.17 .327 0.84 0.61-1.17 .305
5 0.63 0.44-0.89 .009 0.64 0.45-0.91 .012
Age category, y
0-4 Ref Ref
5-9 0.65 0.49-0.86 .002 0.66 0.49-0.89 .006

2052 Cancer May 1, 2018

 10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Mental Health Outcomes in CCS/Nathan et al

TABLE 6. Continued

Univariate Model Multiple Variable Model

Variable HR 95% CI P HR 95% CI P

10-14 0.63 0.48-0.85 .002 0.64 0.46-0.91 .011

15 0.68 0.46-1.02 .060 0.66 0.42-1.04 .072
Cancer type
ALL Ref Ref
Other leukemia 1.26 0.72-2.20 .419 1.11 0.62-1.98 .731
Hodgkin lymphoma 1.25 0.84-1.87 .267 1.17 0.73-1.88 .511
Other lymphomas 1.19 0.76-1.86 .455 1.26 0.79-2.02 .335
CNS tumors 1.24 0.90-1.70 .190 1.16 0.81-1.67 .412
Neuroblastoma 1.74 1.00-3.03 .050 1.15 0.64-2.06 .641
Soft tissue sarcomas 1.42 0.91-2.20 .121 1.35 0.83-2.20 .226
Retinoblastoma 0.98 0.42-2.26 .959 0.65 0.27-1.54 .327
Renal tumors 1.97 1.30-2.97 .001 1.39 0.88-2.20 .157
Hepatic tumors 1.86 0.76-4.57 .176 1.38 0.55-3.46 .488
Bone tumors 0.79 0.44-1.42 .423 0.77 0.41-1.44 .416
Germ cell tumors 0.78 0.39-1.55 .482 0.72 0.35-1.48 .369
Other epithelial neoplasms 0.68 0.32-1.47 .329 0.63 0.28-1.40 .255
Unspecified malignancies 0.49 0.08-3.22 .460 0.38 0.06-2.54 .316
Relapse or SMN
No Ref
Yes 0.72 0.45-1.14 .155
CNS radiation
No Ref Ref
Yes 0.68 0.52-0.89 .005 0.73 0.55-0.98 .038
ITR
1 Ref
2 1.15 0.77-1.70 .496
3 1.14 0.76-1.71 .528
4 1.06 0.65-1.73 .823
Surgery
No Ref
Yes 1.15 0.93-1.43 .199
Chemotherapy
No Ref
Yes 1.12 0.87-1.43 .391
High-dose methotrexate
No Ref Ref
Yes 0.75 0.56-1.00 .047 0.81 0.56-1.17 .267
Corticosteroids
No
Yes 0.88 0.71-1.10 .256
Treatment era
1987-1993 Ref
1994-2001 1.07 0.86-1.33 .560
2002-2008 1.31 0.87-1.97 .204
HSCT
None Ref
Autologous 1.20 0.64-2.27 .573
Allogeneic 1.05 0.59-1.86 .880
Cyclophosphamide equivalent dose, mg/m2
0 Ref
1-3999 1.07 0.72-1.58 .756
4000-7999 0.81 0.62-1.05 .106
8000 1.00 0.72-1.39 .994

Abbreviations: ALL, acute lymphocytic leukemia; CI, confidence interval; CNS, central nervous system; HR, hazard ratio; HSCT, hematopoietic stem cell trans-
plantation; ITR, Intensity of Treatment Rating Scale; Ref, reference category; SMN, second malignant neoplasm.

younger than 10 years at diagnosis were at a particularly severe psychiatric event later in life. It is intriguing to con-
elevated risk of hospital contact. Similarly, we demon- sider whether a specific therapeutic exposure or the overall
strated that patients diagnosed at ages 0 to 4 years were intensity of cancer treatment might impact the young
one-third more likely than older children to experience a developing brain in a manner that predisposes to the

Cancer May 1, 2018 2053

 10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Original Article

TABLE 7. Demographic, Diagnostic, and Treatment Predictors of a Severe Psychiatric Event in Childhood
Cancer Survivors Ages 0 to 4 Years at Diagnosis

Univariate Model Multiple Variable Model

Variable HR 95% CI P HR 95% CI P

Sex
Male Ref Ref
Female 1.35 0.96-1.91 .08 1.33 0.94-1.88 .11
Income quintile
1 Ref Ref
2 0.72 0.44-1.19 .20 0.72 0.43-1.19 .20
3 0.62 0.36-1.05 .07 0.63 0.37-1.08 .10
4 0.72 0.44-1.19 .21 0.73 0.44-1.20 .21
5 0.52 0.29-0.92 .03 0.51 0.29-0.92 .02
Cancer type
Leukemia/lymphoma Ref Ref
Brain tumor 1.02 0.60-1.74 .95 0.84 0.47-1.50 .55
Other 1.50 1.03-2.16 .03 1.23 0.80-1.89 .35
Relapse or SMN
No Ref
Yes 0.67 0.33-1.36 .27
CNS radiation
No Ref
Yes 0.76 0.49-1.19 .23
High-dose methotrexate
No Ref
Yes 0.59 0.37-0.94 .03 0.65 0.38-1.12 .12
Treatment era
1987-1993 Ref
1994-2001 1.31 0.88-1.97
2002-2008 —a
HSCT
None Ref
Autologous 2.00 0.88-4.54 .10
Allogeneic 0.94 0.29-3.02 .92

Abbreviations: CI, confidence interval; CNS, central nervous system; HR, hazard ratio; HSCT, hematopoietic stem cell transplantation; Ref, reference category;
SMN, second malignant neoplasm.
a
No patients ages 0 to 4 years who received treatment during 2002 through 2008 were eligible for study inclusion.

development of psychiatric disease later in life. We did
not identify any such exposure: the risk of severe morbid-
ity was not a consequence of cancer type, treatment inten-
sity, or exposure to chemotherapy with a possible CNS
impact. In fact, cranial radiation appeared to have a pro-
tective effect in the overall cohort, an unexpected finding
that may be because of residual confounding (clinicians
are less likely to subject younger children to cranial radia-
tion) or could be spurious. Nonbiologic causes of the risk
for severe psychiatric events in survivors of cancer at
young ages could be related to the impact of the cancer
diagnosis on parental behavior and the considerable stress
of a life-threatening illness on the child. Limited studies
have indicated that the disruption of parenting practices
during and after cancer therapy can impact the emotional
health of their children. Parents of young children with
leukemia reportedly display lax parenting practices
and overprotection, which have been associated with
subsequent childhood emotional and behavioral
difficulties.38,39 Other studies have demonstrated that
parenting stress and perceived childhood vulnerability
negatively impact adjustment in their children. Whether
such parenting practices have a long-term impact on men-
tal health has not been established and should be a focus
of future research. In the general population, adverse
childhood experiences (which are defined as stressful or
traumatic childhood events) have been linked to the devel-
opment of mental health disorders in adulthood.40 Lim-
ited literature regarding the impact of a life-threatening
childhood illness on subsequent mental health suggests
that illnesses such as cancer might impact psychiatric out-
comes in some children in a manner analogous to other
adverse childhood experiences.41
Our findings about the differential impact of age on
the rate of mental health care visits (which were more fre-
quent in patients who received treatment during adoles-
cence) and severe mental health events (which were more
common in those who received treatment during early

2054 Cancer May 1, 2018


childhood) suggest that outcomes in CCS should be con-
ceptualized as 2 separate constructs. Conflating these may
mask at-risk populations and potential mechanisms. Nei-
ther construct was associated with specific cancer diagno-
ses, treatment exposures, or the intensity of therapy;
rather, both were associated with demographic factors
such as age, sex, and socioeconomic status, suggesting that
it is the cancer experience and not the direct impact of a
specific therapy that influenced outcomes. The impact of
sex and socioeconomic status mirror the risk factors for
mental health disorders in the general population, in
which it has been demonstrated that females are at ele-
vated risk for mood and anxiety disorders42,43 and are
more likely to access the mental health care system. Lower
socioeconomic status also has been associated with an ele-
vated risk for mental health disorders,44,45 but it is unclear
whether this is a cause or a consequence of these outcomes
in the general population.46 In our analysis, being in the
highest socioeconomic stratum at the time of cancer diag-
nosis protected against severe psychiatric morbidity later
in life.
Depression and anxiety are among the more common
psychiatric outcomes observed in CCS. A population-based
study in British Columbia demonstrated a 20% increased
use of antidepressants in survivors compared with the gen-
eral population.47 The Danish group observed an elevated
risk for antidepressant prescription, particularly in children
who received more intensive therapies like HSCT.48 In
contrast, the North American Childhood Cancer Survivor
Study reported that survivors were more likely to use anxio-
lytic, sedative, and hypnotic medications but were no more
likely to use antidepressants.5 Survivors in our cohort were
not at an elevated risk for ED visits or hospitalizations for
mood/affective disorders or anxiety disorders, although it is
likely that many patients with these diagnoses would be
treated in an outpatient setting except in cases of extreme
distress. To our knowledge, we are the first group to
demonstrate a small but statistically significant elevated risk
of psychotic disorders, but this requires confirmation in
other cohorts.
Our study has numerous strengths. It assesses mental
health care in a large, well characterized, population-based
cohort of CCS and is not limited by the biases inherent to
studies that rely on self-report or clinic-based assessments
in selected subsets of survivors. Unlike the Danish study,
we were able to assess the impact of specific cancer thera-
pies. These did not modify the risk for adverse outcomes,
which is an important and novel finding. However, our
findings should be interpreted in the context of several
limitations.
Cancer May 1, 2018
10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Mental Health Outcomes in CCS/Nathan et al
First, although it has been demonstrated that physi-
cian diagnoses in health care administrative claims are a reli-
able measure of mental illness,49 the current study could
not capture patients who had developed a mental health
problem but did not access mental health care or those who
sought care from a practitioner who did not bill the provin-
cial health care system, such as a psychologist or social
worker. However, access to nonphysician mental health
professionals is quite limited within Ontario’s publicly
funded system and is accessible based on the existence of
private insurance coverage or an ability to pay. Conse-
quently, most of the contacts for mental health services are
captured in the freely available access to family physicians
and psychiatrists, because universal coverage provides access
to these providers without requiring copayment.
Second, we could only subclassify the psychiatric indi-
cation for visits to an ED and hospitalizations. There is no
validated mechanism for determining the reason for a men-
tal health visit to a family physician or psychiatrist; there-
fore, the overall prevalence of specific mental health
diagnoses in the survivor population cannot be determined.
Third, we could not directly assess the relation
between physical late effects and poor mental health out-
comes and thus cannot determine whether the increased
risk of adverse mental health outcomes is related to poor
physical health. The absence of a relation between specific
treatments or the intensity of treatment as a whole and the
risk of adverse mental health outcomes suggests that
mechanisms other than the acute toxicities of cancer ther-
apy may modulate mental health outcomes; however,
direct assessment of physical health outcomes would be
needed to definitively rule out chronic physical health
conditions as a risk factor for poor mental health.
Finally, our cohort only included about one-half of
the adolescents (ages 15-17.9 years) who received treat-
ment for cancer, because the remaining adolescents were
treated in adult centers. Medical and psychosocial sup-
ports offered to adolescents in a pediatric hospital often
differ from those offered in an adult center, and the distri-
bution of diagnoses in adolescents who use pediatric cen-
ters differs from that of diagnoses in adolescents who use
adult centers in Ontario (eg, leukemia and lymphoma
more likely to be treated in a pediatric center, whereas
melanoma and thyroid cancer are more likely to be treated
in an adult center; unpublished data). Consequently, care
should be taken in generalizing our findings to adolescents
who receive treatment in adult hospitals.
In summary, CCS demonstrated an increased rate
of mental health visits, particularly outpatient visits to
family physicians and psychiatrists. ED visits and
2055

Original Article
hospitalizations were less common but still more likely in
survivors than in the general population. Risk was driven
by demographic factors, not treatment intensity or thera-
peutic exposures. Most striking was the observation that,
although adolescents have a higher risk for mental health
visits in general, survivors of cancer during early child-
hood are most at risk for severe mental health outcomes.
Future work should focus on screening for distress and
providing mental health resources to survivors of adoles-
cent cancer and on further exploring the mechanism for
severe psychiatric outcomes in children who receive treat-
ment for cancer at young ages.
FUNDING SUPPORT
This study was supported by a grant from the Canadian Cancer
Society Quality of Life Research (Sumit Gupta) and by the Institute
for Clinical Evaluative Sciences, which is funded by an annual grant
from the Ontario Ministry of Health and Long-Term Care.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
AUTHOR CONTRIBUTIONS
Paul C. Nathan: Conceptualization, data curation, and writing–
original draft. Alex Nachman: Conceptualization, formal analysis/
software, and writing–original draft. Rinku Sutradhar: Formal
analysis/software and writing–review and editing. Paul Kurdyak:
Conceptualization and writing–review and editing. Jason D. Pole:
Data curation and writing–review and editing. Cindy Lau: Formal
analysis/software and writing–review and editing. Sumit Gupta:
conceptualization, data curation, and writing–original draft.
REFERENCES
1. Smith MA, Seibel NL, Altekruse SF, et al. Outcomes for children
and adolescents with cancer: Challenges for the twenty-first century.
J Clin Oncol. 2010;28:2625-2634.
2. Phillips SM, Padgett LS, Leisenring WM, et al. Survivors of child-
hood cancer in the United States: prevalence and burden of morbid-
ity. Cancer Epidemiol Biomarkers Prev. 2015;24:653-663.
3. Hudson MM, Ness KK, Gurney JG, et al. Clinical ascertainment of
health outcomes among adults treated for childhood cancer. JAMA.
2013;309:2371-2381.
4. Kessler RC, Berglund P, Demier O, Jin R, Merikangas KR, Walters
EE. Lifetime prevalence and age-on-onset distributions of DSM-IV
disorders in the National Comorbidity Survey Replication. Arch Gen
Psychiatry. 2005;62:593-602.
5. Brinkman TM, Ullrich NJ, Zhang N, et al. Prevalence and predic-
tors of prescription psychoactive medication use in adult survivors of
childhood cancer: a report from the Childhood Cancer Survivor
Study. J Cancer Surviv. 2013;7:104-114.
6. Brinkman TM, Zhang N, Recklitis CJ, et al. Suicide ideation and
associated mortality in adult survivors of childhood cancer. Cancer.
2014;120:271-277.
7. Schultz KA, Ness KK, Whitton JA, et al. Behavioral and social out-
comes in adolescent survivors of childhood cancer: a report from the
Childhood Cancer Survivor Study. J Clin Oncol. 2007;25:3649-
3656.
8. Recklitis CJ, Diller LR, Li X, Najita J, Robison LL, Zeltzer LK. Sui-
cide ideation in adult survivors of childhood cancer: a report from
2056
10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
the Childhood Cancer Survivor Study. J Clin Oncol. 2010;28:655-
661.
9. Zebrack BJ, Zeltzer LK, Whitton JA, et al. Psychological outcomes
in long-term survivors of childhood leukemia, Hodgkin’s disease and
non-Hodgkin’s lymphoma: a report from the Childhood Cancer
Survivor Study. Pediatrics. 2002;110:42-52.
10. Stuber ML, Meeske KA, Krull KR, et al. Prevalence and predictors
of posttraumatic stress disorder in adult survivors of childhood can-
cer. Pediatrics. 2010;125;e1124-e1134.
11. Phillips-Salimi CR, Lommel KM, Andrykowski MA. Physical and
mental health status and health behaviors of childhood cancer survi-
vors: findings from the 2009 BRFSS Survey. Pediatr Blood Cancer.
2012;58:964-970.
12. Tai E, Buchanan GR, Townsend J, Fairley T, Moore A, Richardson
LC. Health status of adolescent and young adult survivors. Cancer.
2012;118:4884-4891.
13. Reulen RC, Winter DL, Lancashire ER, et al. Health-status of adult
survivors of childhood cancer: A large-scale population-based study
from the British Childhood Cancer Survivor Study. Int J Cancer.
2007;121:633-640.
14. Kamibeppu K, Sato I, Honda M, et al. Mental health among young
adult survivors of childhood cancer and their siblings including post-
traumatic growth. J Cancer Surviv. 2010;4:303-312.
15. Teall T, Barrera M, Barr R, Silva M, Greenberg M. Psychological
resilience in adolescent and young adult survivors of lower extremity
bone tumors. Pediatr Blood Cancer. 2013;60:1223-1230.
16. Harila MJ, Niinivirta TIT, Winqvist S, Harila-Saari AH. Low
depressive symptom and mental distress scores in adult long-term
survivors of childhood acute lymphoblastic leukemia. J Pediatr Hem-
atol Oncol. 2011;33:194-198.
17. Kazak AE, DeRosa BW, Schwartz LA, et al. Psychological outcomes
and health beliefs in adolescent and young adult survivors of child-
hood cancer and controls. J Clin Oncol. 2010;28:2002-2007.
18. Fidler MM, Ziff OJ, Wang S, et al. Aspects of mental health dys-
function among survivors of childhood cancer. Br J Cancer. 2015;
113:1121-1132.
19. Lund LW, Schmiegelow K, Rechnitzer C, Johansen C. A systematic
review of studies on psychosocial late effects of childhood cancer:
structures of society and methodological pitfalls may challenge the
conclusions. Pediatr Blood Cancer. 2011;56:532-543.
20. Bardwell WA, Ancoli-Israel S, Dimsdale JE. Response bias influences
mental health symptom reporting in public with obstructive sleep
apnea. Ann Behav Med. 2001;23:313-317.
21. van deMortel TF. Faking it: social desirability response bias in self-
report research. Aust J Adv Nurs. 2008;25:40-48.
22. Lund LW, Winther JF, Dalton SO, et al. Hospital contact for men-
tal disorders in survivors of childhood cancer and their siblings in
Denmark: a population-based cohort study. Lancet Oncol. 2013;14:
971-980.
23. Greenberg ML, Barr RD, DiMonte B, McLaughlin E, Greenberg C.
Childhood cancer registries in Ontario, Canada: lessons learned from
a comparison of 2 registries. Int J Cancer. 2003;105:88-91.
24. Steele LS, Glazier RH, Lin E, Evans M. Using administrative data
to measure ambulatory mental health service provision in primary
care. Med Care. 2004;42:960-965.
25. Khan S, Chiu M, Simpson AI, Guttmann A, Jembere N, Kurdyak
P. Use of mental health services by youths and young adults before
and during correctional custody: a population-based study. Psychiatr
Serv. 2016;67:790-793.
26. Steliarova-Foucher E, Stiller C, Lacour B, Kaatsch P. International
Classification of Childhood Cancer, third edition. Cancer. 2005;103:
1457-1467.
27. Green DM, Nolan VG, Goodman PJ, et al. The cyclophosphamide
equivalent dose as an approach for quantifying alkylating agent expo-
sure: a report from the Childhood Cancer Survivor Study. Pediatr
Blood Cancer. 2014;61:53-67.
28. Kazak AE, Hocking MC, Ittenbach RF, et al. A revision of the
intensity of treatment rating scale: classifying the intensity of pediat-
ric cancer treatment. Pediatr Blood Cancer. 2012;59:96-99.
29. Therneau TM, Grambsch PM. Modeling Survival Data: Extending
the Cox Model. New York: Springer; 2000.
Cancer May 1, 2018

30. Twisk JW, Smidt N, de Vente W. Applied analysis of recurrent
events: a practical overview. J Epidemiol Community Health. 2005;
59:706-710.
31. Lawless JF, Nadeau C. Some simple robust methods for the analysis
of recurrent events. Technometrics. 1995;37:158-168.
32. Sutradhar R, Agha M, Pole JD, et al. Specialized survivor clinic
attendance is associated with decreased rates of emergency depart-
ment visits in adult survivors of childhood cancer. Cancer. 2015;121:
4389-4397.
33. Galan S, de la Vega R, Miro J. Needs of adolescents and young
adults after cancer treatment: a systematic review [published online
ahead of print August 10, 2016]. Eur J Cancer Care (Engl).
10.1111/ecc.12558.
34. Barnett M, McDonnell G, DeRosa A, et al. Psychosocial outcomes
and interventions among cancer survivors diagnosed during adoles-
cence and young adulthood (AYA): a systematic review. J Cancer
Surviv. 2016;10:814-831.
35. Keegan TH, Lichtensztajn DY, Kato I, et al. Unmet adolescent and
young adult cancer survivors’ information and service needs: a
population-based cancer registry study. J Cancer Surviv. 2012;6:239-
250.
36. Dieluweit U, Seitz DC, Besier T, et al. Utilization of psychosocial
care and oncological follow-up assessments among German long-
term survivors of cancer with onset during adolescence. Klin Padiatr.
2011;223:152-158.
37. Mertens AC. Cause of mortality in 5-year survivors of childhood
cancer. Pediatr Blood Cancer. 2007;48:723-726.
38. McCarthy MC, Bastiani J, Williams LK. Are parenting behaviors
associated with child sleep problems during treatment for acute lym-
phoblastic leukemia? Cancer Med. 2016;5:1473-1480.
39. Williams LK, Lamb KE, McCarthy MC. Behavioral side effects of
pediatric acute lymphoblastic leukemia treatment: the role of parent-
ing strategies. Pediatr Blood Cancer. 2014;61:2065-2070.
Cancer May 1, 2018
10970142, 2018, 9, Downloaded from https://acsjournals.onlinelibrary.wiley.com/doi/10.1002/cncr.31279 by Cochrane Portugal, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Mental Health Outcomes in CCS/Nathan et al
40. McLaughlin KA, Greif Green J, Gruber MJ, Sampson NA,
Zaslavsky AM, Kessler RC. Childhood adversities and first onset of
psychiatric disorders in a national sample of US adolescents. Arch
Gen Psychiatry. 2012;69:1151-1160.
41. Kessler RC, McLaughlin KA, Green JG, et al. Childhood adversities
and adult psychopathology in the WHO World Mental Health Sur-
veys. Br J Psychiatry. 2010;197:378-385.
42. McLean CP, Asnaani A, Litz BT, Hofmann SG. Gender differences
in anxiety disorders: prevalence, course of illness, comorbidity and
burden of illness. J Psychiatr Res. 2011;45:1027-1035.
43. Piccinelli M, Wilkinson G. Gender differences in depression. Critical
review. Br J Psychiatry. 2000;177:486-492.
44. Sareen J, Afifi TO, McMillan KA, Asmundson GJ. Relationship
between household income and mental disorders: findings from a
population-based longitudinal study. Arch Gen Psychiatry. 2011;68:
419-427.
45. de Oliveira C, Cheng J, Vigod S, Rehm J, Kurdyak P. Patients with
high mental health costs incur over 30 percent more costs than other
high-cost patients. Health Aff (Millwood). 2016;35:36-43.
46. Kessler RC, McGonagle KA, Zhao S, et al. Lifetime and 12-month
prevalence of DSM-III-R psychiatric disorders in the United States.
Results from the National Comorbidity Survey. Arch Gen Psychiatry.
1994;51:8-19.
47. Deyell RJ, Lorenzi M, Ma S, et al. Antidepressant use among survivors
of childhood, adolescent and young adult cancer: a report of the Child-
hood, Adolescent and Young Adult Cancer Survivor (CAYACS)
Research Program. Pediatr Blood Cancer. 2013;60:816-822.
48. Lund LW, Winther JF, Cederkvist L, et al. Increased risk of antide-
pressant use in childhood cancer survivors: a Danish population-
based cohort study. Eur J Cancer. 2015;51:675-684.
49. Kurdyak P, Lin E, Green D, Vigod S. Validation of a population-
based algorithm to detect chronic psychotic illness. Can J Psychiatry.
2015;60:362-368.
2057