In a recent study posted to the medRxiv* preprint server, researchers evaluated retrospective EQ-5D-5L data collection among coronavirus disease 2019 (COVID-19) patients.
Background
Studies have highlighted the effect of SARS-CoV-2 infection that extends past clinical results. To fully understand the impact of COVID-19, it is necessary to assess the effect on the quality of life (QoL). The EuroQoL Group 5 dimension and 5 level (EQ-5D-5L) is widely employed for evaluating health-related QoL (HRQoL) and generating utilities for the computation of quality-adjusted life years (QALYs).
In infectious disease patient-reported outcome (PRO) research, EQ-5D-5L data collection, both retrospectively and prospectively, has been utilized to identify the baseline health status pre-infection as well as morbid health status during the infection course. However, previous investigations have not validated the retrospective EQ-5D-5L data collection.
About the study
In the present study, researchers employed the PRO information obtained in COVID-19 studies that assessed the validity of retrospective EQ-5D-5L data collection in establishing pre-COVID-19 infection status.
In the PRO research of COVID-19 patients, participants were enrolled among adult outpatients in the United States having at least one self-reported symptom and a positive severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) reverse transcription polymerase chain reaction (RT-PCR) test. A prior study involved an analytical cohort restricted to unvaccinated or BNT162b2-vaccinated people. All recruited patients were included to evaluate the validity of the retrospective data-collecting procedure.
EQ-5D-5L assessed QoL across five distinct dimensions, including mobility, regular activities, self-care, discomfort/pain, and depression/anxiety, and five levels, including none, minor, moderate, severe, and extreme issues/inability. The five domains were transformed into the Utility Index (UI) via US-based weights. On the recruitment day, about three days after testing SARS-CoV-2 positive, individuals answered the EQ-5D-5L questionnaire twice, including an altered version having all questions framed in the past tense to evaluate HRQoL pre-SARS-CoV-2 baseline and a version framed in the present tense to evaluate current HRQoL.
Results
Almost 10.2% of the 676 participants were aged at least 65 years or over, 73.2% were women, and 71.9% were White. Eligible participants had reported rates of 8.6% for chronic lung disease or asthma, 4.7% for diabetes, and 11.2% for hypertension. A worse visual analog scale (VAS) was related to a higher degree of disease severity in every EQ-5D5L domain for retrospectively obtained pre-COVID-19 baseline along with standard collection on the third day after testing positive for COVID-19.
For pre-SARS-CoV-2 infection mobility, the team noted that the average EQ-VAS score for those who reported none was 88.5, mild was 74.9, moderate was 57.7, and severe issue was 50.0. None of the participants reported the inability to move. Significant differences existed between the mean EQ-VAS scores with respect to none, mild, and moderate/severe problems.
This study’s participants were predominantly women, Caucasian, and had fewer chronic conditions than the general US population. Both the pre-SARS-CoV-2 infection baseline average UI was 0.924% while the average VAS was 87.4%. In addition, fewer issues were observed in all five domains, with prevalence rates of 7.1%, 6.7%, 7.9%, 27.3%, and 43.5% with respect to the present study group versus 28.4%, 6.5%, 24.7%, 51.0%, and 38.4% related to mobility, usual activity, self-care, discomfort/pain, and depression/anxiety, respectively.
In the model that predicted EQ-VAS using a retrospective assessment variable called RETRO, UI, and their interaction, the measured coefficient associated with the UI-by-RETRO interaction was -4.2. This showed that retrospective estimation did not significantly affect the extent of the correlation between UI and EQ-VAS.
Comparing the EQ-5D-5L to the US population norms, the team used the matching-adjusted indirect comparison (MAIC) approach to match the proportions of age category, sex, hypertension, and diabetes observed in the present sample to the US norms. The weighted pre-COVID-19 baseline average for UI reduced to 87.0, while VAS decreased to 0.922. Both were far above the US population norms. With the inclusion of the fraction of individuals having mobility challenges, the sample size was decreased to 291.
The baseline weighted average of UI prior to COVID-19 was 0.866, which did not show significant variation from the US population average of 0.851. The average weight of VAS was 84.6, which was greater than the US average of 80.4.
Three days after the COVID-19 test positivity, the UI was 0.808%, and VAS was 73.33%. Cohen’s d value for UI was 0.68, and VAS was 1.01, suggesting a moderate-to-large effect on UI and a high effect on VAS compared to the baseline evaluation. UI and VAS had Cohen’s d values of 0.21 and 0.44, respectively, showing a modest to moderate effect on UI and a minimal to moderate effect on VAS relative to US population norms. However, only 17.2% of the COVID-19 group experienced mobility problems, compared to 25.2% of the US population. After adjusting for percentages of age, sex, hypertension, and diabetes, 19.0% of the population reported mobility difficulties, which was much lower than 25.2%. It was determined that UI and VAS effect sizes (ES) were 0.15 and 0.39, respectively.
Conclusion
The study findings showed that the retroactively obtained pre-COVID-19 EQ-5D-5L was acceptable with respect to the US population norms. It is also adequately aligned in comparison to the standard EQ-5D-5L collection for COVID-19. Collecting pre-COVID-19 EQ-5D-5L data permitted direct examination of COVID-19’s impact on health-related QoL. Future research that specifically compares traditional prospective evaluation with retrospective evaluation of the EQ-5D-5L prior to COVID-19 is encouraged.
*Important notice
medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.
