Discussion

In the first decade of life, VOCs are the most common complication of SCD,38 and associated with an impaired HRQoL.39,40 Generally, SCD is associated with a high health care utilization including emergency room visits and hospitalizations due to VOCs.41 With age, the frequency of hospitalization for VOC increases.39,42This study shows that occurrence of hospitalization for VOC has a negative effect on total HRQoL, evident up to 12 months after hospitalization in children with SCD. This effect was most pronounced if HRQoL was measured closely after hospitalization and decreased over time independently of the frequency of hospitalization. As the number of hospitalization increases, the negative effect on total HRQoL significantly increases as well.
Our study is in line with previous research involving the impact of hospitalization for VOC on HRQoL in children with SCD.43 Brandow et al. prospectively found a decrease in HRQoL after hospitalization for VOC, but this effect seemed to disappear after one week.43 The authors compared HRQoL during hospitalization for VOC with control patients in steady state. The discrepancy with our study could be attributed to the cross-sectional design of their study and the comparison with a separate control group. Instead, we used repeated measures over time to compare HRQoL between patients and within patients with and without hospitalization for VOC. Using repeated measures in the same patients allows us to monitor changes of HRQoL over time, while controlling for factors that cause variability. In our longitudinal analyses, although the effect of occurrence of hospitalization on HRQoL was greatest after 3 months, this effect was not significant most probably due to lack of power. Nevertheless, the trend up to 12 months is evidently present, and decreasing over time.
In addition, previous studies have showed an inverse correlation between self-reported HRQoL and the frequency of hospitalizations and emergency department visits, comparable to our study.44-48 As the effect of frequency of hospitalization on HRQoL in our study does not look clinically relevant, it will become relevant as healthcare utilization increases with age. Accordingly, Campball et al. reported the higher the frequency of VOCs, the lower the self-reported total HRQoL in children with SCD.42
Dampier et al. found in regression models that mainly occurrence of pain-related acute care visits was responsible for negatively affected HRQoL subscales.17 All HRQoL subscales are found to be affected in patients with SCD compared to healthy norms,24,44,49 but the subscale physical functioning seems to be the most affected across previous studies.25,39 This is line with our study, reinforcing the fact that hospitalization has a negative impact mainly on the physical functioning lasting for months. School functioning was only significantly affected after 12 months, so it seems that the effect of hospitalization on school functioning only manifests over time. Even though the reliability of the PedsQL for school functioning was poor in our study, hospitalization for VOC does result in school absenteeism, which in turn can lead to poor school performance or academic functioning.50-53 As a matter of fact, the frequency of hospitalization has been reported to be a predictor of poor HRQoL with regard to school functioning in children with SCD.50
In several studies, SCD severity and complications were associated with worse HRQoL.54 In the association between hospitalization for VOC and HRQoL, the severity of hospitalization has not been taken into account before. According to our subanalysis, severity of hospitalization could have affected total HRQoL, but only until 3 months after hospitalization. After this time, severity does not seem to affect the total HRQoL score. However, these results were not significant, as this analysis was conducted within a subgroup of patients with 45 completed PedsQLs.

Strengths and limitations

An important strength of this study is its longitudinal nature with data covering the last 10 years with at least two measurements of HRQoL per patient. Also, our study population with SCD is closely monitored at our SCD comprehensive care center, as each patient is requested to annually fill out PROMs before a regular visit as part of standard care.
There are also limitations that should be taken into account when interpreting the data. First of all, patients with SCD who have mainly VOCs managed at home, are not taken into account in the analyses, while these VOCs burden patients as well. This also applies to patients with other SCD-related complications such as chronic pain, that may have a harmful effect on aspects of HRQoL. We did exclude patients on a chronic transfusion therapy as they were more likely to have less hospitalizations for VOC, while carrying the burden of frequent blood transfusions. Second, the PedsQL was found to have moderate/good reliability for all subscales, except for school functioning. Therefore, the PedsQL results for school functioning should be interpreted with caution. Third, patients without hospitalization completed the PedsQL more often than the hospitalized patients. Hospitalized patients may experience a higher disease burden or stress causing them to have less time for completion of PedsQLs. Another explanation is, when a patient is hospitalized, while having a visit scheduled at the outpatient clinic, this visit is not automatically rescheduled to after hospitalization.
Half of the eligible patients (or caregivers) did not give consent to use the HRQoL data for scientific research. Moreover, not all children were able to fill out the PedsQL regularly. For instance, children and/or caregiver(s) who lack digital literacy, did not complete the PedsQL via the online platform. However, we are facilitating and offering technical support at the outpatient clinic for these patients and their caregivers to overcome this problem. All of the above could have caused selection bias of the study population affecting the generalizability. In addition, our dataset was too limited to include socio-economic status in our analyses. Considering that socio-economic status is a well-known potential confounder of HRQoL, this may have caused a bias as well.
Based on findings of this study, there are some implications for clinical practice and recommendations for future research. Because of the negative impact of both the occurrence and frequency of hospitalizations for VOC on HRQoL, it is important to focus on the prevention of VOCs and the usage of disease-modifying drugs. In case of hospitalization, close monitoring and psychosocial support aimed at all aspects of HRQoL, but in particular physical functioning (e.g. by offering physiotherapy), should be considered. After frequent hospitalizations, there should be sufficient attention for school functioning as well. Furthermore, the KLIK PROM portal is an important tool in clinical practice to assess the need for extra psychosocial support in patients with SCD. SCD-specific symptom list in addition to the generic PROMs evaluating aspects of HRQoL could increase the reliability making research more replicable and easily comparable in this patients with SCD. Ultimately, a mixed-methods study should be performed to provide in-depth insights into how HRQoL is affected by hospitalization for VOC. This could allow healthcare providers to identify and better respond to the needs of children with SCD, who have been hospitalized for VOC, and customize their care.