Methods
In this retrospective cohort study, patients aged between 8 and 18 years
with a confirmed diagnosis of SCD, who received care at the Emma
Children’s Hospital, Amsterdam University Medical Centers (Amsterdam
UMC) were eligible for enrollment. As part of standard care procedures,
children routinely filled out PROMs prior to the regular visits through
the online KLIK PROM portal (https://www.hetklikt.nu/). The answers were
returned to the clinician and discussed during consultation. One of the
measures is the Pediatric Quality of Life Inventory 4.0 Generic core
scales (PedsQLTM 4.0). Every child (and/or the parent)
was requested to complete the PedsQL at least once a year. Patients were
included in this study if they had filled out at least two PedsQLs
between January 2012 and September 2021. Data of patients were excluded
if patients were on a chronic
transfusion therapy or if permission for use of data for research
purposes had been denied. This project was reviewed by the Medical
Research Ethics Committee of the Amsterdam UMC, and conducted according
to the Declaration of Helsinki.
Study measures
Demographics and disease-related characteristics including age, gender,
SCD genotype, medication use, and hospitalization for VOC were collected
from the electronic health records (EHR). Details of hospitalization for
VOC one year prior to HRQOL measurement were collected from the EHR as
well including length of stay (LOS), analgesic use and complications
during hospital admission such as development of acute chest syndrome.
With this information, the severity of each hospital admission was
scored as mild (1 point), moderate (2 points) or severe (3 points). Mild
severity was defined as a LOS shorter than 4 days; moderate severity was
defined as a LOS between 4 through 6 days; and severe was classified as
the presence of an acute chest syndrome independent of LOS, or a LOS
more or equal to 7 days.
Self-reported HRQoL was collected by completion of the PedsQL through
the online KLIK PROM portal. The PedsQL consists of 23 items covering
the following 4 subscales: physical, emotional, social and school
functioning.30,35 The questions address the preceding
week and could be answered with the following options: never (0), almost
never (1), sometimes (2), often (3) and almost always (4). The answer
‘0’ is converted into the score of 100, ‘1’ into 75, ‘2’ into 50, ‘3’
into 25 and ‘4’ into 0. Then, a mean score was computed for each of the
23 items, and transformed to a scale that ranges from 0 to 100. Both a
total score as well as a score for each HRQoL subscale was calculated. A
psychosocial score was created from the emotional, social and school
functioning subscales of the PedsQL. A weighted average of the different
PedsQL subscale scores was calculated to derive the total HRQoL score.
The higher the calculated scores, the higher the perceived
HRQoL.36 Previous studies confirmed the validity and
reliability of the PedsQLTM for the measurement of
HRQOL in the Dutch population.30 To evaluate the
reliability of the PedsQL versions in our study population, we
calculated internal consistency estimates (Cronbach’s α). Estimates of
0.70 or greater were considered good, while estimates between 0.60 and
0.70 were considered moderate and estimates below 0.60 were considered
poor.37
Statistical analyses
All data was transformed and analyzed in the Statistical Package for the
Social Sciences (SPSS) version 28.0. Descriptive data were generated for
all variables to describe the study population and its characteristics.
Categorical variables were presented as absolute numbers with
corresponding percentages. Means and standard deviations (SDs) were
calculated for continuous variables that were normally distributed.
Medians with interquartile ranges (IQRs) were calculated for values that
were not normally distributed. The period of follow-up was expressed in
patient-years by summarizing the duration of follow-up in days for all
patients divided by the number of days in a year (365.25). Baseline
characteristics and mean HRQoL scores were compared between the
hospitalization and no hospitalization for VOC group using the
chi-squared test or the unpaired t-test. For analysis purposes, genotype
was categorized in two groups: a clinically mild (HbSC and
HbSß+-thalassemia) and clinically severe genotype (HbSS and
HbSß0-thalassemia).
To assess the impact of the occurrence of hospitalization for VOC on
HRQoL over time, linear mixed model (LMM) analysis was used. The
dependent variables were HRQoL scores (both the total and the subscale
scores), and the independent variable was hospitalization for a VOC. In
all analyses, patients were considered as non-hospitalized if they had
not been admitted to the hospital in the last 12 months. Sensitivity
analyses were performed to determine whether the time period between
hospital admission and HRQoL measurement, affected HRQoL. For this, the
analyses were repeated for the time periods 3, 6, 9 and 12 months after
hospitalization. The impact of frequency of hospital admission in 12
months on HRQoL was evaluated using LMM analyses as well. To assess the
impact of the severity of a hospitalization for VOC on HRQoL over time,
a subanalysis with LMM was performed within the hospitalization group
using the calculated severity score. All analyses were adjusted for age
and SCD genotype. The estimates of the LMM analyses were reported as
regression coefficient (β). Significance levels were set at 5% and all
estimates were reported with 95% confidence interval (CI).