The study is a non-randomised controlled trial with 12-month follow-up. Alongside this, a
health economic study and process evaluation study will be conducted in line with the
British Medical Research Council's framework for development and evaluation of complex
interventions. The study will be reported in three separate scientific papers for each
evaluation study.

This study aims to evaluate the effectiveness of a structured and group-based
rehabilitation intervention, called The Long COVID Rehabilitation Intervention, with the
goal to optimise the rehabilitation of patients experiencing long-term symptoms after
COVID-19 (long COVID).

The primary hypothesis: A structured and personalised rehabilitation intervention for
patients with long COVID is more effective evaluated on sick leave, mental fatigue and
disability compared to usual rehabilitation.

To evaluate the aim, the intervention group (patients receiving The Long COVID
Rehabilitation Intervention in Aarhus Municipality, Central Denmark Region, Denmark) will
be compared to a control group (patients receiving rehabilitation in the 18 other
municipalities in Central Denmark Region, Denmark).

The data will be handled in accordance with the World Medical Association Coherence
Declaration of Helsinki and with the General Data Protection Regulation. The study is
approved by the Danish Regional Data Protection Agency (J.nr. 1-16-02-655-20).

The intervention group (patients from Aarhus Municipality, Central Denmark Region)
received The Long COVID Rehabilitation Intervention; a rehabilitation programme with a
duration of minimum 3 months comprising 4-12 individual sessions and 1-4 group-based
courses with physiotherapists and occupational therapists. The control group (patients
from the other 18 municipalities in Central Denmark region) received the rehabilitation
interventions delivered in their home municipality which were generally less structured
with more individual sessions compared to the intervention group. Based on a systematic
identification through semi-structured interviews with representatives from all 19
municipalities in Central Denmark Region, the rehabilitation provided for long COVID
patients was described for each municipality using the TIDieR template. Throughout the
study period, it will be explored if it is possible to divide the control groups into two
subgroups depending on the rehabilitation provided in the municipalities.

With a mean Mental Fatigue Scale score of 18.6 (SD: 5.6), a minimal clinical important
difference of 2 points, a power of 80% and a two-sided significance level of 0.05, a
total study population of 375 (125 patients in each group) patients should be included.
To account for a 20% loss to follow-up, the total number is 450 patients. For long term
sick leave, a power calculation will be performed during the fall 2024 based on data from
an ongoing study.

Data are collected through questionnaires sent out electronically and through register
data. Sociodemographic characteristics (sex, age, and self-reported living status,
education, work status and occupation) and self-reported information on main symptoms and
time since infection is collected at inclusion. Study data are collected and managed
using REDCap electronic data capture tools.

Descriptive statistics will be conducted to describe sociodemographic characteristics. To
evaluate the effectiveness of the intervention, outcomes in the intervention group will
be compared to outcomes in the control group using multiple linear regression analysis,
adjusting for possible confounders chosen prior to the analyses, i.e. sociodemographic
variables.

In the economic evaluation study, a cost-utility analysis and a cost-effectiveness
analysis will be performed from a societal perspective alongside the effectiveness study.
The outcome in the cost-utility analysis will be quality-adjusted life years (QALYs)
assessed by EQ-5D-5L measured at baseline and 12 months follow-up and valued by Danish
preference weights. The outcome in the cost-effectiveness analysis will be changes in
Mental Fatigue Scale. The costs of The Long COVID Rehabilitation Intervention will be
estimated using micro-costing. Cost of usual rehabilitation will be estimated using
micro-costing when possible or based on survey information from the municipalities.
Visits to primary health care services will be extracted from the Danish National Health
Service Register for Primary Care. Secondary healthcare services will be extracted from
the National Patient Register. The valuation will be determined by the
Diagnosis-Related-Grouping reimbursement rate and the outpatient-grouping-system.
Productivity loss will be extracted from Danish Register for Evaluation of
Marginalisation. The productivity costs per patient will be calculated using the Human
Capital method.

To evaluate the cost-effectiveness of the intervention, the Incremental
Cost-Effectiveness Ratio (ICER) will be calculated. Bias-corrected and accelerated
bootstrapping with 10,000 replications will be performed to estimate 95% confidence
intervals around cost differences and the uncertainty surrounding the ICERs. Uncertainty
will be shown in cost-effectiveness plane. The cost-effectiveness acceptability curve
will be drawn to show the probability that the intervention is cost-effective at
different thresholds for willingness to pay for a gain in QALY or a clinical relevant
improvement in Mental Fatigue Scale. Sensitivity analyses will be performed to test the
robustness of the study results.

In the process evaluation study, the investigators will explore the processes of The Long
COVID Rehabilitation Intervention by evaluating implementation, mechanism of impact and
contextual factors. The study design is a concurrent mixed method study with quantitative
and qualitative data collected and analysed during a similar timeframe. Implementation of
the intervention will be investigated, focusing on fidelity, dose, adaptations and reach
using patient baseline characteristics, routine monitored data from the patient'
journals, and interviews with patients and health professionals. To evaluate mechanism of
impact and contextual factors, the investigators will use a realist evaluation approach
with pre-defined Context-Mechanism-Outcome configurations (CMOs) to explore under what
circumstances, for whom, why and how the intervention works. This will be done using
routine monitoring data from patient journals, individual interviews with patients (n=15)
and focus group interviews with health professionals (n=5). An initial programme theory
has been developed and graphically shown in a general logic model.

Quantitative and qualitative data are obtained and analysed simultaneously and linked at
multiple points. Descriptive statistics will be used to analyse fidelity, dose and reach
and will be presented with mean and standard deviation or median and percentiles as
appropriate. Interview data will be transcribed verbatim and will be analysed following
the standards for realist evaluations. The qualitative and quantitative results and most
important CMOs will be merged to support refinement of the programme theory.