Regularly monitor the following in hospitalised patients to facilitate early recognition of deterioration and monitor for complications:[2][573]

  • Vital signs (temperature, respiratory rate, heart rate, blood pressure, oxygen saturation)

  • Haematological and biochemistry parameters

  • Coagulation parameters (D-dimer, fibrinogen, platelet count, prothrombin time)

  • ECG

  • Chest imaging

  • Signs and symptoms of venous or arterial thromboembolism.

Medical early warning scores

  • Utilise medical early warning scores that facilitate early recognition and escalation of treatment of deteriorating patients (e.g., National Early Warning Score 2 [NEWS2], Paediatric Early Warning Signs [PEWS]) where possible.[2]

  • There are no data on the value of using these scores in patients with COVID-19 in the primary care setting.[1314]

  • The sequential organ failure assessment (SOFA) score does not possess adequate discriminant accuracy for mortality prediction in patients prior to intubation for COVID-19 pneumonia.[1315]

Pregnant women

  • Monitor vital signs three to four times daily and fetal heart rate in pregnant women with confirmed infection who are symptomatic and admitted to hospital. Perform fetal growth ultrasounds and Doppler assessments to monitor for potential intrauterine growth restriction in pregnant women with confirmed infection who are asymptomatic.[838] Perform fetal growth ultrasound 14 days after resolution of symptoms.[840]

Post-discharge follow-up

  • Patients who have had suspected or confirmed COVID-19 (of any disease severity) who have persistent, new, or changing symptoms should have access to follow-up care.[2]

  • Guidelines for the respiratory follow-up of patients with COVID-19 pneumonia have been published. Follow-up algorithms depend on the severity of pneumonia, and may include clinical consultation and review (face-to-face or telephone) by a doctor or nurse, chest imaging, pulmonary function tests, echocardiogram, sputum sampling, walk test, and assessment of oxygen saturation.[1316]

  • More than half of patients discharged from hospital had lung function and chest imaging abnormalities 12 weeks after symptom onset.[1317] Pulmonary function tests may reveal altered diffusion capacity, a restrictive pattern, or an obstructive pattern.[1318]

Prognostic scores in development

  • Various prognostic and clinical risk scores are being researched or developed for COVID-19; however, further external validation across various populations is needed before their use can be recommended. The World Health Organization recommends using clinical judgement, including consideration of the patient’s values and preferences and local and national policy if available, to guide management decisions including admission to hospital and to the intensive care unit, rather than currently available prediction models for prognosis.[2]

    • A-DROP: a modified version of CURB-65 that showed better accuracy of in-hospital death prediction on admission in patients with COVID-19 pneumonia compared with other widely used community-acquired pneumonia scores.[1319]

    • APACHE II: an effective clinical tool to predict hospital mortality that performed better than SOFA and CURB-65 scores in patients with COVID-19. A score of 17 or more is an early indicator of death and may help provide guidance to make further clinical decisions.[1320]

    • CALL: a risk factor scoring system that scores patients based on four factors: comorbidities, age, lymphocyte count, and lactate dehydrogenase level. One study found that 96% of patients with low CALL scores did not progress to severe disease.[1321]

    • COVID-GRAM: a web-based calculator that estimates the probability that a patient will develop critical illness and relies on the following 10 variables at admission: chest radiographic abnormality, age, haemoptysis, dyspnoea, unconsciousness, number of comorbidities, cancer history, neutrophil-to-lymphocyte ratio, lactate dehydrogenase, and direct bilirubin. Additional validation studies, especially outside of China, are required.[1322] A retrospective cohort study was unable to fully validate the tool for predicting critical illness among hospitalised patients in Europe as it overestimated the risk in the highest-risk patients.[1323]

    • COVID-19MRS: a rapid, operator-independent clinical tool that was found to objectively predict mortality in one retrospective cohort study.[1324]

    • COVID-19 SEIMC: a mortality prediction score for predicting 30-day mortality of patients attending hospital accident and emergency departments, based on age, low age-adjusted SaO₂, neutrophil-to-lymphocyte ratio, estimated glomerular filtration rate, dyspnoea, and sex. It has been externally validated with two large datasets from patients hospitalised with laboratory-confirmed disease and shows a high degree of accuracy.[1325]

    • QCOVID: a novel clinical risk prediction algorithm to estimate the risk of hospital admission and mortality based on age, ethnicity, deprivation, body mass index, and a range of comorbidities. Population-based cohort studies have found that the algorithm performed well, showing very high levels of discrimination for deaths and hospital admissions.[1326][1327] The risk model is being used in the UK to help clinicians identify adults with multiple risk factors that make them more vulnerable to COVID-19, and inform decisions about vaccination cohorts and shielding.[1328]

    • SCARP: a novel risk calculator that provides clinically meaningful predictions of whether hospitalised patients will progress to severe illness or death based on variables that are readily available (e.g., pulse oximetry, oxygen supplementation, respiratory rate, pulse). The calculator has undergone internal and temporal validation, but further studies are required.[1329]

    • SOARS: a five-predictor risk prediction score based on demographic and clinical characteristics (i.e., SpO₂, obesity, age, respiratory rate, stroke history). May be useful to identify patients who have a low probability of mortality for outpatient monitoring and management, and could potentially be used to inform clinical triage in preadmission settings. Further validation is required.[1330]

    • 3F: a mortality prediction model based on three clinical features: age, minimum oxygen saturation, and type of patient encounter (i.e., inpatient vs outpatient and telehealth encounters). One study found that the model showed high accuracy when applied to retrospective and prospective data sets of COVID-19 patients.[1331]

    • 4C: a score developed and validated in a UK prospective cohort study of adults admitted to hospital with COVID-19. The score uses patient demographics, clinical observations, and blood parameters commonly available at the time of hospital admission, and can accurately characterise patients as being at low, intermediate, high, or very high risk of death. The score outperformed other risk stratification tools, showed clinical decision-making utility, and had similar performance to more complex models.[1332]

  • The BMJ 10-minute consultation: what is my covid risk? external link opens in a new window

Use of this content is subject to our disclaimer