The SOLIDARITY Finland Long-COVID is the long-term follow-up of the remdesivir sub-study
of the SOLIDARITY Finland. (Following the WHO core protocol with local adjustments, the
SOLIDARITY Finland is an adaptive, randomized, open-label clinical trial evaluating the
safety and efficacy of possible therapeutic agents in hospitalized adult patients
diagnosed with COVID-19.)

Study population Adult patients with confirmed SARS-2-CoV-2 admitted to the hospital ward
or the intensive care unit (ICU) were included. Adult hospital patients, with definite
COVID-19 and, as per the responsible doctor, no contraindication to the study drugs were
entered into the online electronic data collection system and randomly allocated between
the local standard of care alone or local standard of care plus remdesivir (daily
infusion until discharge or up to ten days). Between July 2020 and January 2021, 208
patients were recruited from 13 hospitals, of which 114 were randomized to receive
remdesivir plus usual care and 94 to receive usual care only. (These patients (if alive;
n=202 or less) will be invited to participate in the SOLIDARITY Finland Long-COVID.)
During patient recruitment at the hospital phase of the study, the following data was
collected: 1) hospital and randomizing doctor; 2) confirmation that patient has provided
consent, 3) patient identifiers, age, and sex, 4) major comorbidities, 5) COVID severity.

Consent:

All patients have already consented to SOLIDARITY Finland (in-hospital study). In the
SOLIDARITY Finland Long-COVID, patients will receive, by mail, an information letter,
consent form, and questionnaires. The mail will be sent three days before the anniversary
of the randomization day during hospitalisation. They may reply by i) sending back the
completed documents, ii) scheduling a phone call with an investigator for an interview,
or iii) declining participation. If the patient does not reply in 14 days (from the day
the investigators sent the questionnaire to the patient), the investigators will send one
reminder by mail and, if there will not be a reply, then the investigators will approach
the patient via telephone at around 10 days from the reminder mail. The phone call will
be attempted twice: during and after office hours. If the patient will not answer the
phone, the investigators will also send a text message to inform them about the reason
for our attempted phone call.

Questionnaire (symptoms and characteristics):

Our multidisciplinary team of clinicians, methodologists, and patient partners developed
a questionnaire that records basic patient information. This will include i) date of
completing the questionnaire, ii) age, iii) height and weight, iv) smoking status (never,
ex-smoker, current smoker), v) possible comorbidities, and whether diagnosed before or
after COVID-19-infection (obstructive sleep apnea, stroke, coronary artery disease,
diabetes, hypertension, cancer, and any psychiatric disease), vi) employment (student,
unemployed, employed, sickness allowance, retired), vii) working capability in comparison
to the pre-COVID-19 state, viii) physician visits due to symptoms associated with
COVID-19, ix) physician-diagnosed long-COVID-19 syndrome. Questionnaires will also
document long-term symptoms. Relevant long-COVID-19 -symptoms (in total 20) were
identified from recent publications and review articles. This same questionnaire will be
used at one and two years. Exertional and cardiopulmonary symptoms potentially related to
COVID include fatigue, postexertional malaise, dyspnea during exercise, chest discomfort,
palpitations, cough, and respiratory mucous discharges. Main neuropsychiatric symptoms
potentially related to COVID include generalized fatigue, attention and memory deficits,
sleeping difficulties, depression and anxiety, dizziness, and even sensory disturbances
such as paresthesias and changes in taste or smell perceptions. Other commonly
encountered symptoms potentially related to COVID include widespread pains (muscle and
joint pains, headache), skin rash, nausea, and fever. The burden from each individual
symptom is an ordinal variable and will be graded from 0 to 3, where 0 represents no
symptom, 1 represents mild bother, 2 moderate bother, and 3 severe bother due to the
symptom. Dyspnea is assessed in accordance with the Modified Medical research council
dyspnea scale (mMRC) from 0 to 4, where 0 represents dyspnea only with strenuous
exercise, and 4 the presence of dyspnea even with mild physical activity, e.g., dressing
clothes. To capture the dimensions of recovery, the investigators will use the Core
Outcome Measure for Recovery, which has been recommended for use in COVID-19 research.

Quality of life:

The investigators have obtained permission from EuroQol to use the EQ-5D-5L questionnaire
to record the patient's QoL. EQ-5D-5L assesses the domains of mobility, self-care, usual
daily activities, general pain/discomfort, anxiety/depressions, and an overall impression
of health. The first five domains are graded from 1 to 5, while the last uses the visual
analogue scale from 0 to 100. Due to our multiethnic patient population, the
questionnaires have been translated in the following languages: Albanian, Arabic,
English, Estonian, Farsi, Finnish, Russian, Somali, and Swedish. The above-mentioned
questionnaire (Finnish language version as the original) has also been translated into
these languages.

Data security and future use of data:

Patient information will be encrypted and held securely by the Sponsor. Those analyzing
it will use only pseudonymized data, and no identifiable patient details will appear in
publications. Data from questionnaires will also be analyzed using pseudonymized data.
The investigators have taken care to limit the questions to necessary and clinically
relevant aspects related to long-COVID.

UPDATE 02.02.2022:

Primary outcomes will comprise the following:

1. EQ-VAS

2. EQ-5D-5L, summary

3. Does the patient feel recovered from COVID-19-infection at one year or not?
(question no. 10)

4. Fatigue (questionnaire, question no. 14)

5. Exertional dyspnea (question no. 12)

To avoid problems related to multiple testing, the trial analysis will be focused on the
five most relevant outcomes (listed above). This decision was done before examining the
accumulated data, and by consulting statistical experts neither who had seen study data.

Variables collected via questionnaire:

i) Symptoms: 20 symptoms are measured as follows; each range from 0 to 3 (0 = No symptom.
1 = Symptom exists and causes small bother. 2 = Symptom exists and causes moderate
bother. 3 = Symptom exists and causes severe bother.) ii) QoL: Using the EQ-VAS and
EQ-5D-5L to compare domain-specific scores between the two treatment arms.

Additional variables (at 1 or 2 years; depending on future funding and ethical approval
decisions; currently the study has ethical approval for long-COVID and quality of life
assessments only):

- Registry data: (UPDATE 02.02.2022) Mortality data with dates are obtained from
Digital and Population Data Services Agency during early spring 2022. Later,
underlying causes of death will be obtained from Statistics Finland and classified
according to the International Statistical Classification of Diseases and Related
Health Problems, 10th revision). In line with the Finnish regulations, any consent
will not be required from the study patients to acquire information on their census
data, vital status, and causes of death for registry research purposes. Comorbidity
will be obtained from Finnish Institute for Health and Welfare THL. These will be
performed through linkage the patient national identification number to registry
records. This study is unlikely well-powered to measure changes in mortality;
however, being aware of the vital status is important for the overall follow-up of
study participants.

- Lung tests are also planned. Spirometry parameters are continuous variables as
absolute values (liters = L; liters per second = L/s), percentage of reference
values, and z-values. These include VC, FVC, FEV1, FEV1/VC, peak expiratory flow
(PEF), the maximal expiratory flow at 50 % (MEF50), and the forced expiratory time
(FET). Spirometry will be performed with a bronchodilator test to assess the changes
between baseline and after bronchodilation. Diffusing capacity parameter DLCO is a
continuous variable with ml/min/mmHg as the unit.

- Genome sequence analyses are also planned to identify potential single-nucleotide
polymorphisms that would associate with long-COVID symptoms and the effect of
remdesivir.

Subgroup analyses:

For the primary outcomes, a priori planned subgroup analysis will be performed for
whether the patient needed oxygenation at hospital admission (the investigators
hypothesize that the treatment effect is larger for those without extra oxygen than those
with extra oxygen at hospital admission).

Comparisons between the two treatment arms will be performed as follows:

- Descriptive statistics: Descriptive statistics will be presented with numbers and
percentages for categorical variables and means, standard deviation, and range for
continuous variables. In the case of clearly skewed continuous variables, they will
be presented with median, interquartile range (25th and 75th percentiles), and
range. Demographics and baseline characteristics will be presented with descriptive
statistics without any hypothesis testing.

- Continuous variables will be subject to repeated measures mixed models or
appropriate non-parametric alternatives.

- Binary response variables will be analyzed using logistic regression (possibly
adjusting for within-subject dependencies by generalized estimating equations or
mixed models) or chi-square/Mantel-Haenszel tests.

- If missing data is regarded as having a significant effect on the conclusions of the
trial, sensitivity analyses with different methods for handling missing data will be
included. The differences between respondents and non-respondents will be assessed
using available baseline variables.

Statistical analyses in more detail:

1. EQ-VAS

- Mann-Whitney test, when the treatment arm is the only variable.

- Regression analysis to adjust possible imbalances. Linear regression, if VAS
scores are not concentrated around either extreme of the scale.

2. EQ-5D-5L, summary.

- Proportions of respondents in both groups by categories (Mobility, Self-Care,
Usual Activities, Pain/Discomfort, Anxiety/Depression).

3. Does the patient feel recovered from COVID-19-infection at one year or not?
(question no. 10)

4. Fatigue (questionnaire, question no. 14)

- Logistic regression analysis. "No symptom" and "Mild symptoms" will be analyzed
together against "Moderate symptoms" and "Severe symptoms".

5. Exertional dyspnea (question no. 12) - Logistic regression analysis. mMRC categories
0 and 1 will be analyzed together against categories 2-4.