2  Materials and methods

2.1  Study design and participants
For this single-center retrospective study, 449 hospitalized patients with laboratory confirmed COVID-19 admitted to the West District of Wuhan Union Hospital between January 25 and March 15, 2020 were screened using the following inclusion and exclusion criteria. The inclusion criteria included: (1) patients aged between 18 and 80 years; and (2) absolute counts of peripheral blood lymphocytes < 1.1 × 109/L within 24 h of hospital admission. The exclusion criteria included: (1) duration of hospital stay < 14 days or death during hospitalization; (2) pregnant or lactating women; (3) deterioration of the condition requiring ICU care within 14 days of admission; and (4) loss of consciousness during hospitalization. A total of 153 patients met these criteria and were included in the study (Fig. 1 ). The diagnosis, clinical classification, treatment, and discharge criteria of COVID-19 were based on the “Diagnosis and Treatment Protocol for Novel Coronavirus Pneumonia (6th edition)” issued by the National Health and Family Planning Commission of China (NHPFC, 2020). COVID-19 was confirmed by real-time RT-PCR for throat-swab specimens from the upper respiratory tract. All the included patients had recovered from COVID-19.
Fig. 1 Study flow diagram. ALC, absolute lymphocyte count; COVID-19, coronavirus disease-2019; ICU, intensive care unit; PSQI, Pittsburgh Sleep Quality Index; RCSQ, Richards-Campbell sleep questionnaire.
The study was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (Permission number: 0173). Informed consent for this retrospective study was waived due to the rapid emergence of this infectious disease.

2.2  Data collection
We reviewed the hospital's electronic medical records, nurse records, laboratory findings, and imaging examinations for all patients with confirmed COVID-19. All data were checked by two researchers.
We collected data on age, sex, body mass index, education levels, marital status, comorbidities (chronic respiratory disease, cardiovascular and cerebrovascular diseases, and gastrointestinal, endocrine, urologic, and nervous system diseases), current smoking status, time from symptom onset to admission, signs and symptoms at disease onset, chest imaging abnormalities (unilateral or bilateral distribution of patchy shadows or ground glass opacity), vital signs on admission (respiratory rate, percutaneous oxygen saturation, heart rate, systolic blood pressure, and body temperature), laboratory parameters within 24 h of admission (blood routine, blood biochemistry and electrolytes, cardiac biomarkers, and coagulation parameters), treatment (oxygen therapy, vasoconstrictive agents, antiviral therapy, antibiotics, corticosteroids, immunoglobulin, and immunoregulatory therapy), and complications during hospitalization (acute kidney injury (AKI), acute liver dysfunction, acute cardiac injury, hyperglycemia, and hospital-acquired infection (HAI)). AKI was identified based on the definition in the Kidney Disease: Improving Global Outcomes statement (KDIGO, 2012). Cardiac injury was diagnosed when the levels of serum hypersensitive cardiac troponin I exceeded the upper limit of normal (ULN), measured in the laboratory of the Wuhan Union Hospital. Acute liver dysfunction was defined as an ALT ≥ 1 × ULN. HAI was defined as infection acquired>48 h after hospital admission. For patient(s) with deterioration of the condition requiring ICU care, the subsequent data regarding laboratory parameters, treatment, and complications were not included in the final analysis.

2.3  Sleep assessment
The Richards-Campbell sleep questionnaire (RCSQ), a widely used subjective survey instrument (Nagatomo et al., 2020, Richards et al., 2000, Simons et al., 2018), was administered to assess sleep quality. Sleep assessment was performed by telephone within 30 days after hospital admission. The overall sleep quality was assessed within the first week (score one), second week (score two), and third week (score three) after admission using the RCSQ, since not all patients were able to accurately assess their daily sleep quality. The RCSQ assesses the following five items on a 0–100 mm visual analog scale: perceived sleep depth, sleep latency, frequency of awakenings, latency after awakenings, and sleep quality. The overall RCSQ score is the average value of these five items, with higher scores indicating better sleep quality. Self-reported factors associated with disruptive sleep were also recorded.
After assessment with the RCSQ, sleep quality within two or three weeks after hospital admission was further confirmed using the Chinese version of the Pittsburgh Sleep Quality Index (PSQI), which is reliable and valid for the Chinese population (Liu et al., 1996). The PSQI includes seven component scores: subjective sleep quality, sleep duration, sleep latency, habitual sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. The sum of the seven component scores (each ranging from 0 to 3) provides the global sleep quality score (ranging from 0 to 21), with a score greater than 7 indicating poor sleep quality (Liu et al., 1996). Patients were divided into two groups: good-sleep group (at least two RCSQ scores ≥ 70, and PSQI ≤ 7) and poor-sleep group (at least two RCSQ scores ≤ 50, and PSQI > 7). Patients with deterioration of their health condition requiring ICU care, or those with a duration of hospital stay < 21 days were assigned to one of the two groups based on the overall quality of their sleep within the first two weeks after hospital admission: patients with two RCSQ scores ≥ 70 and PSQI ≤ 7 were assigned to the good-sleep group and those with one RCSQ score ≤ 50 and PSQI > 7 were assigned to the poor-sleep group.

2.4  Outcomes
The primary outcomes were: (1) recovery rate based on ALC and deterioration based on an increased NLR on day 7, 14, and 21 after admission; (2) need for ICU care within 14 days of hospital admission.
The absolute counts of peripheral blood lymphocytes and neutrophils within 24 h of hospital admission were set as baseline value. The formula for the recovery percentage or deteriorative percentage is described as:ALCorNLRday7,day14orday21--ALCorNLRbaselineALCorNLRbaseline×100%
The secondary outcomes were: (1) HAI; and (2) total length of hospital stay.

2.5  Statistical analysis
Categorical variables were described as number (%). Continuous variables were described using mean (SD) if they were normally distributed, or median (interquartile range, IQR) if they were not. Continuous variables were compared using independent group t tests when the data were normally distributed; otherwise, the Mann-Whitney test was used. Proportions for categorical variables were compared using the χ2 test, Yates’ continuity corrected χ2 test, or Fisher’s exact test. A two-sided P value < 0.05 was considered statistically significant. The data collected were all analyzed using SPSS version 20.0 software (SPSS, Tokyo, Japan).