Prediction models for diagnosis and prognosis of covid-19 infection: systematic review and critical appraisalBMJ 2020; 369 doi: https://doi.org/10.1136/bmj.m1328 (Published 07 April 2020) Cite this as: BMJ 2020;369:m1328
- Laure Wynants, assistant professor1 2,
- Ben Van Calster, associate professor2 3,
- Marc M J Bonten, professor4 5,
- Gary S Collins, professor6 7,
- Thomas P A Debray, assistant professor4 8,
- Maarten De Vos, associate professor2 9,
- Maria C Haller, medical doctor10 11,
- Georg Heinze, associate professor10,
- Karel G M Moons, professor4 8,
- Richard D Riley, professor12,
- Ewoud Schuit, assistant professor4 8,
- Luc J M Smits, professor1,
- Kym I E Snell, lecturer12,
- Ewout W Steyerberg, professor3,
- Christine Wallisch, research fellow10 13 14,
- Maarten van Smeden, assistant professor4
- 1Department of Epidemiology, CAPHRI Care and Public Health Research Institute, Maastricht University, Peter Debyeplein 1, 6229 HA Maastricht, Netherlands
- 2Department of Development and Regeneration, KU Leuven, Leuven, Belgium
- 3Department of Biomedical Data Sciences, Leiden University Medical Centre, Leiden, Netherlands
- 4Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht University, Utrecht, Netherlands
- 5Department of Medical Microbiology, University Medical Centre Utrecht, Utrecht, Netherlands
- 6Centre for Statistics in Medicine, Nuffield Department of Orthopaedics, Musculoskeletal Sciences, University of Oxford, Oxford, UK
- 7NIHR Oxford Biomedical Research Centre, John Radcliffe Hospital, Oxford, UK
- 8Cochrane Netherlands, University Medical Centre Utrecht, Utrecht University, Utrecht, Netherlands
- 9Department of Electrical Engineering, ESAT Stadius, KU Leuven, Leuven, Belgium
- 10Section for Clinical Biometrics, Centre for Medical Statistics, Informatics and Intelligent Systems, Medical University of Vienna, Vienna, Austria
- 11Ordensklinikum Linz, Hospital Elisabethinen, Department of Nephrology, Linz, Austria
- 12Centre for Prognosis Research, School of Primary, Community and Social Care, Keele University, Keele, UK
- 13Charité Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, Berlin, Germany
- 14Berlin Institute of Health, Berlin, Germany
- Correspondence to: L Wynants
- Accepted 31 March 2020
Objective To review and critically appraise published and preprint reports of prediction models for diagnosing coronavirus disease 2019 (covid-19) in patients with suspected infection, for prognosis of patients with covid-19, and for detecting people in the general population at risk of being admitted to hospital for covid-19 pneumonia.
Design Rapid systematic review and critical appraisal.
Data sources PubMed and Embase through Ovid, Arxiv, medRxiv, and bioRxiv up to 24 March 2020.
Study selection Studies that developed or validated a multivariable covid-19 related prediction model.
Data extraction At least two authors independently extracted data using the CHARMS (critical appraisal and data extraction for systematic reviews of prediction modelling studies) checklist; risk of bias was assessed using PROBAST (prediction model risk of bias assessment tool).
Results 2696 titles were screened, and 27 studies describing 31 prediction models were included. Three models were identified for predicting hospital admission from pneumonia and other events (as proxy outcomes for covid-19 pneumonia) in the general population; 18 diagnostic models for detecting covid-19 infection (13 were machine learning based on computed tomography scans); and 10 prognostic models for predicting mortality risk, progression to severe disease, or length of hospital stay. Only one study used patient data from outside of China. The most reported predictors of presence of covid-19 in patients with suspected disease included age, body temperature, and signs and symptoms. The most reported predictors of severe prognosis in patients with covid-19 included age, sex, features derived from computed tomography scans, C reactive protein, lactic dehydrogenase, and lymphocyte count. C index estimates ranged from 0.73 to 0.81 in prediction models for the general population (reported for all three models), from 0.81 to more than 0.99 in diagnostic models (reported for 13 of the 18 models), and from 0.85 to 0.98 in prognostic models (reported for six of the 10 models). All studies were rated at high risk of bias, mostly because of non-representative selection of control patients, exclusion of patients who had not experienced the event of interest by the end of the study, and high risk of model overfitting. Reporting quality varied substantially between studies. Most reports did not include a description of the study population or intended use of the models, and calibration of predictions was rarely assessed.
Conclusion Prediction models for covid-19 are quickly entering the academic literature to support medical decision making at a time when they are urgently needed. This review indicates that proposed models are poorly reported, at high risk of bias, and their reported performance is probably optimistic. Immediate sharing of well documented individual participant data from covid-19 studies is needed for collaborative efforts to develop more rigorous prediction models and validate existing ones. The predictors identified in included studies could be considered as candidate predictors for new models. Methodological guidance should be followed because unreliable predictions could cause more harm than benefit in guiding clinical decisions. Finally, studies should adhere to the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guideline.
The novel coronavirus disease 2019 (covid-19) presents an important and urgent threat to global health. Since the outbreak in early December 2019 in the Hubei province of the People’s Republic of China, the number of patients confirmed to have the disease has exceeded 775 000 in more than 160 countries, and the number of people infected is probably much higher. More than 36 000 people have died from covid-19 infection (up to 30 March 2020).1 Despite public health responses aimed at containing the disease and delaying the spread, several countries have been confronted with a critical care crisis, and more countries will almost certainly follow.234 Outbreaks lead to important increases in the demand for hospital beds and shortage of medical equipment, while medical staff themselves could also get infected.
To mitigate the burden on the healthcare system, while also providing the best possible care for patients, efficient diagnosis and prognosis of the disease is needed. Prediction models that combine several variables or features to estimate the risk of people being infected or experiencing a poor outcome from the infection could assist medical staff in triaging patients when allocating limited healthcare resources. Models ranging from rule based scoring systems to advanced machine learning models (deep learning) have been proposed and published in response to a call to share relevant covid-19 research findings rapidly and openly to inform the public health response and help save lives.5 Many of these prediction models are published in open access repositories, ahead of peer review.
We aimed to systematically review and critically appraise currently available prediction models for covid-19, in particular diagnostic and prognostic models for the disease. This systematic review was carried out in collaboration with the Cochrane Prognosis Methods Group.
We searched PubMed and Embase through Ovid, bioRxiv, medRxiv, and arXiv for research on covid-19 published after 3 January 2020. We used the publicly available publication list of the covid-19 living systematic review.6 This list contains studies on covid-19 published on PubMed and Embase through Ovid, bioRxiv, and medRxiv, and is continuously updated. We validated the list to examine whether it is fit for purpose by comparing it to relevant hits from bioRxiv and medRxiv when combining covid-19 search terms (covid-19, sars-cov-2, novel corona, 2019-ncov) with methodological search terms (diagnostic, prognostic, prediction model, machine learning, artificial intelligence, algorithm, score, deep learning, regression). All relevant hits were found on the living systematic review list.6 We supplemented this list with hits from PubMed by searching for “covid-19” because when we performed our intitial search this term was not included in the reported living systematic review6 search terms for PubMed. We further supplemented the list with studies on covid-19 retrieved from arXiv. The online supplementary material presents the search strings. Additionally, we contacted authors for studies that were not publicly available at the time of the search,78 and included studies that were publicly available but not on the living systematic review6 list at the time of our search.9101112
We initially searched databases on 13 March 2020, with an update on 24 March 2020. All studies were considered, regardless of language or publication status (preprint or peer reviewed articles). We included studies if they developed or validated a multivariable model or scoring system, based on individual participant level data, to predict any covid-19 related outcome. These models included diagnostic and prognostic models for covid-19, or those aiming to identify people at increased risk of developing covid-19 pneumonia in the general population. No restrictions were made on the setting (eg, inpatients, outpatients, or general population), prediction horizon (how far ahead the model predicts), included predictors, or outcomes. Epidemiological studies that aimed to model disease transmission or fatality rates, diagnostic test accuracy, and predictor finding studies were excluded. Titles, abstracts, and full texts were screened in duplicate for eligibility by pairs of independent reviewers (from LW, BVC, and MvS), and discrepancies were resolved through discussion.
Data extraction of included articles was done by two independent reviewers (from LW, BVC, GSC, TPAD, MCH, GH, KGMM, RDR, ES, LJMS, EWS, KIES, CW, and MvS). Reviewers used a standardised data extraction form based on the CHARMS (critical appraisal and data extraction for systematic reviews of prediction modelling studies) checklist13 and PROBAST (prediction model risk of bias assessment tool).14 We sought to extract each model’s predictive performance by using whatever measures were presented. These measures included any summaries of discrimination (the extent to which predicted risks discriminate between participants with and without the outcome), and calibration (the extent to which predicted risks correspond to observed risks) as recommended in the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) statement.15 Discrimination is often quantified by the C index (C index=1 if the model discriminates perfectly; C index=0.5 if discrimination is no better than chance). Calibration is often quantified by the calibration intercept (which is zero when the risks are not systematically overestimated or underestimated) and calibration slope (which is one if the predicted risks are not too extreme or too moderate).16 We focus on performance statistics as estimated from the strongest available form of validation. Any discrepancies in data extraction were resolved by LW and MvS. The online supplementary material provides details on data extraction. We considered aspects of PRISMA (preferred reporting items for systematic reviews and meta-analyses)17 and TRIPOD15 in reporting our article.
Patient and public involvement
It was not appropriate or possible to involve patients or the public in the design, conduct, or reporting of our research. The study protocol and preliminary results are publicly available on https://osf.io/ehc47/ and medRxiv.
We retrieved 2690 titles through our systematic search (fig 1; 1916 on 13 March 2020 and 774 during an update on 24 March 2020). Two additional unpublished studies were made available on request (after a call on social media). We included a further four studies that were publicly available but were not detected by our search. Of 2696 titles, 85 studies were retained for abstract and full text screening. Twenty seven studies describing 31 prediction models met the inclusion criteria and were selected for data extraction and critical appraisal.789101112181920212223242526272829303132333435363738
Twenty five studies used data on patients with covid-19 from China (supplementary table 1), one study used data on patients from Italy,31 and one study used international data (United States, United Kingdom, and China, among others).35 Based on 18 of the 25 studies that reported study dates, data were collected between 8 December 2019 and 15 March 2020. The duration of follow-up was unclear in most studies, although one reported a median follow-up of 8.4 days,19 while another reported a median follow-up of 15 days.37 Some Chinese centres provided data to multiple studies, but it was unclear how much these datasets overlapped across our 25 identified studies. One study used US Medicare claims data from 2015 to 2016 to estimate vulnerability to covid-19,8 two studies used control CT (computed tomography) scans from the US or Switzerland,1125 and one study used simulated data.18All but one study24 developed prediction models for use in adults. The median age varied between studies (from 34 to 65 years; see supplementary table 1), as did the proportion of men (from 41% to 61%).
Among the six studies that developed prognostic models to predict mortality risk in people with confirmed or suspected covid-19 infection, the percentage of deaths varied between 8% and 59% (table 1). This wide variation is partly because of severe sampling bias caused by studies excluding participants who still had the disease at the end of the study period (that is, they had neither recovered nor died).7202122 Additionally, length of follow-up could have varied between studies (but was rarely reported), and there might be local and temporal variation in how people were diagnosed as having covid-19 or were admitted to the hospital (and therefore recruited for the studies). Among the 18 diagnostic model studies, only one reported on prevalence of covid-19 infection in people with suspected covid-19; the prevalence was 19% (development dataset) and 24% (validation dataset).30 One study reported that 8% of patients had severe disease among confirmed paediatric patients with covid-19 infection.24 Because 16 diagnostic studies used either case-control sampling or an unclear method of data collection, the prevalence in these diagnostic studies might not have been representative of their target population.
Table 1 gives an overview of the 31 prediction models reported in the 27 identified studies. Supplementary table 2 provides modelling details and box 1 discusses the availability of models in a format for use in clinical practice.
Availability of models in format for use in clinical practice
Twelve studies presented their models in a format for use in clinical practice. However, because all models were at high risk of bias, we do not recommend their routine use before they are properly externally validated.
Models to predict risk of hospital admission for coronavirus disease 2019 (covid-19) pneumonia in general population
The “COVID-19 diagnosis aid APP” is available on iOS and android devices to diagnose covid-19 in asymptomatic patients and those with suspected disease.12 The “suspected COVID-19 pneumonia Diagnosis Aid System” is available as an online tool.1040 The “COVID-19 early warning score” to detect covid-19 infection in adults is available as a score chart in an article.30 A decision tree to detect severe disease for paediatric patients with confirmed covid-19 is also available in an article.24
Diagnostic models based on computed tomography (CT) imaging
Three of the seven artificial intelligence models to assist with diagnosis based on CT images are available through web applications.232629414243 One model is deployed in 16 hospitals, but the authors do not provide any usable tools in their study.33
To assist in the prognosis of mortality, a nomogram (a graphical aid to calculate mortality risk),7 a decision tree,21 and a CT based scoring rule are available in the articles.22 Additionally a nomogram exists to predict progression to severe covid-19 disease.32
Models to predict risk of hospital admission for covid-19 pneumonia in general population
We identified three models that predicted risk of hospital admission for covid-19 pneumonia in the general population, but used admission for non-tuberculosis pneumonia, influenza, acute bronchitis, or upper respiratory tract infections as outcomes in a dataset without any patients with covid-19 (table 1).8 Among the predictors were age, sex, previous hospital admissions, comorbidity data, and social determinants of health. The study estimated C indices of 0.73, 0.81, and 0.81 for the three models.
Diagnostic models to detect covid-19 infection in patients with symptoms
We identified one study that developed a model to detect covid-19 pneumonia in fever clinic patients (estimated C index 0.94)10; one to diagnose covid-19 in patients with suspected disease (estimated C index 0.97)30; one to diagnose covid-19 in patients with suspected disease and asymptomatic patients (estimated C index 0.87)12; and one to diagnose covid-19 by using deep learning of genomic sequences (estimated C index 0.98).35 A further study was developed to diagnose severe disease in paediatric inpatients with symptoms, based on direct bilirubin and alanine transaminase (reporting an F1 score of 1.00, indicating 100% observed sensitivity and specificity).24 Only one study reported assessing calibration, but it was unclear how this was done.12 Predictors used in more than one model were age (n=3), body temperature or fever (n=2), and signs and symptoms (such as shortness of breath, headache, shiver, sore throat, and fatigue, n=2; table 1).
Thirteen prediction models were proposed to support the diagnosis of covid-19 or covid-19 pneumonia (and monitor progression) based on CT images. The predictive performance varied widely, with estimated C index values ranging from 0.81 to nearly 1.
Prognostic models for patients with a diagnosis of covid-19 infection
We identified 10 prognostic models (table 1). Of these, six estimated mortality risk in patients with suspected or confirmed covid-19.71819212237 The intended use of these models (that is, when to use them, in whom to use them, and the prediction horizon, eg, mortality by what time) was not clearly described. Two models aimed to predict a hospital stay of more than 10 days from admission.20 Two models aimed to predict progression to a severe or critical state.932 Predictors included in more than one prognostic model were age (n=5), sex (n=2), features derived from CT scoring (n=5), C reactive protein (n=3), lactic dehydrogenase (n=3), and lymphocyte count (n=2; table 1).
Only two studies that predicted mortality reported a C index; these studies obtained estimates of 0.9022 and 0.98.7 One study also evaluated calibration.7 When applied to new patients, their model yielded probabilities of mortality that were too high for low risk patients and too low for high risk patients (calibration slope >1), despite excellent discrimination.7 One study developed two models to predict a hospital stay of more than 10 days and estimated C indices of 0.92 and 0.96.20 The two studies that developed models to predict progression to a severe or critical state estimated C indices of 0.95 and 0.85.932 One of these studies also reported perfect calibration, but it was unclear how this was evaluated.32
Risk of bias
All models were at high risk of bias according to assessment with PROBAST (table 1), which suggests that their predictive performance when used in practice is probably lower than that reported. Therefore, there is cause for concern that the predictions of these models are unreliable. Box 2 gives details on common causes for risk of bias for each type of model.
Common causes of risk of bias in the 19 reported prediction models
Models to predict hospital admission for coronavirus disease 2019 (covid-19) pneumonia in general population
These models were based on Medicare claims data, and used proxy outcomes to predict hospital admission for covid-19 pneumonia, in the absence of patients with covid-19.8
People without covid-19 (or a proportion of them) were excluded, altering the disease prevalence.30 Controls had viral pneumonia, which is not representative of the target population for a screening model.12 The test used to determine the outcome varied between participants,12 or one of the predictors (fever) was part of the outcome definition.10 Predictors were dichotomised, which led to a loss of information.243036
Diagnostic models based on computed tomography (CT) imaging
Generally, studies did not clearly report which patients had CT scans during clinical routine, and it was unclear whether the selection of controls was made from the target population (that is, patients with suspected covid-19).1123293336 Often studies did not clearly report how regions of interest were annotated. Images were sometimes annotated by only one scorer without quality control,2527 the model output influenced annotation,28 or the “ground truth” that was used to build the model was a composite outcome based on the same CT images used to make the prediction, among other factors.38 Careful description of model specification and subsequent estimation were lacking, challenging the transparency and reproducibility of the models. Every study used a different deep learning architecture, some were established and others specifically designed, without benchmarking the used architecture against others.
Study participants were often excluded because they did not develop the outcome at the end of the study period but were still in follow-up (that is, they were in hospital but had not recovered or died), yielding a highly selected study sample.7202122 Additionally, only one study accounted for censoring by using Cox regression.19 One study developed a model to predict future severity using cross sectional data (some participants were severely ill at inclusion)37; this implies that the timing of the measurement of the predictors is not appropriate and the (unclearly defined) outcome might have been influenced by the predictor values. Other studies used highly subjective predictors,22 or the last available predictor measurement from electronic health records (rather than measuring the predictor value at the time when the model was intended for use).21
Eleven of the 27 studies had a high risk of bias for the participants domain (table 2), which indicates that the participants enrolled in the studies might not be representative of the models’ targeted populations. Unclear reporting on the inclusion of participants prohibited a risk of bias assessment in eight studies. Four of the 27 studies had a high risk of bias for the predictors domain, which indicates that predictors were not available at the models’ intended time of use, not clearly defined, or influenced by the outcome measurement. The diagnostic model studies that used CT imaging predictors were all scored as unclear on the predictors domain. The publications often lacked clear information on the preprocessing steps (eg, cropping of images). Moreover, complex machine learning algorithms transform CT images into predictors in an untransparent way, which makes it challenging to fully apply the PROBAST predictors section for such imaging studies. Most studies used outcomes that are easy to assess (eg, death, presence of covid-19 by laboratory confirmation). Nonetheless, there was reason to be concerned about bias induced by the outcome measurement in 10 studies because of the use of subjective or proxy outcomes (eg, non covid-19 severe respiratory infections).
All studies were at high risk of bias for the analysis domain (table 2). Many studies had small sample sizes (table 1), which led to an increased risk of overfitting, particularly if complex modelling strategies were used. Three studies did not report the predictive performance of the developed model, and one study reported only the apparent performance (the performance in exactly the same data used to develop the model, without adjustment for optimism owing to potential overfitting).
Four models were externally validated in the model development study (in an independent dataset, excluding random training test splits and temporal splits).7122532 However, in three of these studies, the external validation datasets are probably not representative of the target population (box 2).71225 Consequently, predictive performance could differ if the models were applied in the target population. Gong and colleagues had a satisfactory predictive performance on two unbiased but small external validation datasets.32 One study was a small (n=27) external validation that reported satisfactory predictive performance of a model originally developed for avian influenza H7N9 pneumonia. However, patients who had not recovered at the end of the study period were excluded, which led to selection bias.22 Only three studies assessed calibration,71232 but the method to check calibration was probably suboptimal in two studies.1232
In this systematic review of prediction models related to the covid-19 pandemic, we identified and critically appraised 27 studies that described 31 models. These prediction models were developed for detecting people in the general population at risk of being admitted to hospital for covid-19 pneumonia, for diagnosis of covid-19 in patients with symptoms, and for prognosis of patients with covid-19 infection. All models reported good to excellent predictive performance, but all were appraised to have high risk of bias owing to a combination of poor reporting and poor methodological conduct for participant selection, predictor description, and statistical methods used. As expected, in these early covid-19 related prediction model studies, clinical data from patients with covid-19 are still scarce and limited to data from China, Italy, and international registries. With few exceptions, the available sample sizes and number of events for the outcomes of interest were limited. This is a well known problem when building prediction models and increases the risk of overfitting the model.44 A high risk of bias implies that these models will probably perform worse in practice than the performance reported by the researchers. Therefore, the estimated C indices, often close to 1 and indicating near perfect discrimination, are probably optimistic. Five studies carried out an external validation,712222532 and only one study assessed calibration correctly.7
We reviewed 13 studies that used advanced machine learning methodology on chest CT scans to diagnose covid-19 disease, covid-19 related pneumonia, or to assist in segmentation of lung images. The predictive performance measures showed a high to almost perfect ability to identify covid-19, although these models and their evaluations also had a high risk of bias, notably because of poor reporting and an artificial mix of patients with and without covid-19.
Challenges and opportunities
The main aim of prediction models is to support medical decision making. Therefore it is vital to identify a target population in which predictions serve a clinical need, and a representative dataset (preferably comprising consecutive patients) on which the prediction model can be developed and validated. This target population must also be carefully described so that the performance of the developed or validated model can be appraised in context, and users know which people the model applies to when making predictions. However, the included studies in our systematic review often lacked an adequate description of the study population, which leaves users of these models in doubt about the models’ applicability. Although we recognise that all studies were done under severe time constraints caused by urgency, we recommend that any studies currently in preprint and all future studies should adhere to the TRIPOD reporting guideline15 to improve the description of their study population and their modelling choices. TRIPOD translations (eg, in Chinese and Japanese) are also available at https://www.tripod-statement.org.
A better description of the study population could also help us understand the observed variability in the reported outcomes across studies, such as covid-19 related mortality. The variability in the relative frequencies of the predicted outcomes presents an important challenge to the prediction modeller. A prediction model applied in a setting with a different relative frequency of the outcome might produce predictions that are miscalibrated45 and might need to be updated before it can safely be applied in that new setting.1646 Such an update might often be required when prediction models are transported to different healthcare systems, which requires data from patients with covid-19 to be available from that system.
Covid-19 prediction problems will often not present as a simple binary classification task. Complexities in the data should be handled appropriately. For example, a prediction horizon should be specified for prognostic outcomes (eg, 30 day mortality). If study participants have neither recovered nor died within that time period, their data should not be excluded from analysis, which most reviewed studies have done. Instead, an appropriate time to event analysis should be considered to allow for administrative censoring.16 Censoring for other reasons, for instance because of quick recovery and loss to follow-up of patients who are no longer at risk of death from covid-19, could necessitate analysis in a competing risk framework.47
Instead of developing and updating predictions in their local setting, individual participant data from multiple countries and healthcare systems might allow better understanding of the generalisability and implementation of prediction models across different settings and populations. This approach could greatly improve the applicability and robustness of prediction models in routine care.4849505152
The evidence base for the development and validation of prediction models related to covid-19 will quickly increase over the coming months. Together with the increasing evidence from predictor finding studies53545556575859 and open peer review initiatives for covid-19 related publications,60 data registries6162636465 are being set up. To maximise the new opportunities and to facilitate individual participant data meta-analyses, the World Health Organization has recently released a new data platform to encourage sharing of anonymised covid-19 clinical data.66 To leverage the full potential of these evolutions, international and interdisciplinary collaboration in terms of data acquisition and model building is crucial.
With new publications on covid-19 related prediction models rapidly entering the medical literature, this systematic review cannot be viewed as an up to date list of all currently available covid-19 related prediction models. Also, 24 of the studies we reviewed were only available as preprints. These studies might improve after peer review, when they enter the official medical literature. We also found other prediction models that are currently being used in clinical practice but without scientific publications,67 and web risk calculators launched for use while the scientific manuscript is still under review (and unavailable on request).68 These unpublished models naturally fall outside the scope of this review of the literature.
Implications for practice
All 31 reviewed prediction models were found to have a high risk of bias, and evidence from independent external validation of these models is currently lacking. However, the urgency of diagnostic and prognostic models to assist in quick and efficient triage of patients in the covid-19 pandemic might encourage clinicians to implement prediction models without sufficient documentation and validation. Although we cannot let perfect be the enemy of good, earlier studies have shown that models were of limited use in the context of a pandemic,69 and they could even cause more harm than good.70 Therefore, we cannot recommend any model for use in practice at this point.
We anticipate that more covid-19 data at the individual participant level will soon become available. These data could be used to validate and update currently available prediction models.16 For example, one model that predicted progression to severe covid-19 disease within 15 days of admission to hospital showed promising discrimination when validated externally on two small but unselected cohorts.32 Because reporting in this study was insufficiently detailed and the validation was in small Chinese datasets, validation in larger, international datasets is needed. Owing to differences between healthcare systems (eg, Chinese and European) on when patients are admitted to and discharged from hospital, and testing criteria for patients with covid-19, we anticipate most existing models will need to be updated (that is, adjusted to the local setting).
When building a new prediction model, we recommend building on previous literature and expert opinion to select predictors, rather than selecting predictors in a purely data driven way16; this is especially true for datasets with limited sample size.71 Based on the predictors included in multiple models identified by our review, we encourage researchers to consider incorporating several candidate predictors: for diagnostic models, these include age, body temperature, and (respiratory) signs and symptoms; for prognostic models, age, sex, C reactive protein, lactic dehydrogenase, lymphocyte count, and potentially features derived from CT scoring. Predictors that were included in both diagnostic and prognostic models were albumin (or albumin/globin), direct bilirubin, and red blood cell distribution width; these predictors could be considered as well. By pointing to the most important methodological challenges and issues in design and reporting of the currently available models, we hope to have provided a useful starting point for further studies aiming to develop new models, or to validate and update existing ones.
This systematic review aims to be the first stage of a living review of this field, in collaboration with the Cochrane Prognosis Methods Group. We will update this review and appraisal continuously, to provide up-to-date information for healthcare decision makers and professionals as more international research emerges over time.
Diagnostic and prognostic models for covid-19 are available and they all appear to show good to excellent discriminative performance. However, these models are at high risk of bias, mainly because of non-representative selection of control patients, exclusion of patients who had not experienced the event of interest by the end of the study, and model overfitting. Therefore, their performance estimates are likely to be optimistic and misleading. Future studies should address these concerns. Sharing data and expertise for development, validation, and updating of covid-19 related prediction models is urgently needed.
What is already known on this topic
The sharp recent increase in coronavirus disease 2019 (covid-19) infections has put a strain on healthcare systems worldwide; there is an urgent need for efficient early detection, diagnosis of covid-19 in patients with suspected disease, and prognosis of covid-19 in patients with confirmed disease
Viral nucleic acid testing and chest computed tomography (CT) are standard methods for diagnosing covid-19, but are time consuming
Earlier reports suggest that elderly patients, patients with comorbidities (chronic obstructive pulmonary disease, cardiovascular disease, hypertension), and patients presenting with dyspnoea are vulnerable to more severe morbidity and mortality after covid-19 infection
What this study adds
Three models were identified that predict hospital admission from pneumonia and other events (as proxy outcomes for covid-19 pneumonia) in the general population
Eighteen diagnostic models were identified for detecting covid-19 infection (13 were machine learning based on CT scans); and 10 prognostic models for predicting mortality risk, progression to severe disease, or length of hospital stay
Proposed models are poorly reported and at high risk of bias, raising concern that their predictions could be unreliable when applied in daily practice
We thank the authors who made their work available by posting it on public registries or sharing it confidentially.
Contributors: LW conceived the study. LW and MvS designed the study. LW, MvS, and BVC screened titles and abstracts for inclusion. LW, BVC, GSC, TPAD, MCH, GH, KGMM, RDR, ES, LJMS, EWS, KIES, CW, and MvS extracted and analysed data. MDV helped interpret the findings on deep learning studies and MMJB and MCH assisted in the interpretation from a clinical viewpoint. LW and MvS wrote the first draft, which all authors revised for critical content. All authors approved the final manuscript. LW and MvS are the guarantors. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Funding: LW is a postdoctoral fellow of Research Foundation–Flanders (FWO). BVC received support from FWO (grant G0B4716N) and Internal Funds KU Leuven (grant C24/15/037). TPAD acknowledges financial support from the Netherlands Organisation for Health Research and Development (grant No 91617050). KGMM gratefully acknowledges financial support from Cochrane Collaboration (SMF 2018). KIES is funded by the National Institute for Health Research School for Primary Care Research (NIHR SPCR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care. GSC was supported by the NIHR Biomedical Research Centre, Oxford, and Cancer Research UK (programme grant C49297/A27294). The funders played no role in study design, data collection, data analysis, data interpretation, or reporting. The guarantors had full access to all the data in the study, take responsibility for the integrity of the data and the accuracy of the data analysis, and had final responsibility for the decision to submit for publication.
Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for the submitted work; no competing interests with regards to the submitted work; LW discloses support from Research Foundation–Flanders (FWO); RDR reports personal fees as a statistics editor for The BMJ (since 2009), consultancy fees for Roche for giving meta-analysis teaching and advice in October 2018, and personal fees for delivering in-house training courses at Barts and The London School of Medicine and Dentistry, and also the Universities of Aberdeen, Exeter, and Leeds, all outside the submitted work.
Ethical approval: Not required.
Data sharing: The study protocol is available online at https://osf.io/ehc47/. Most included studies are publically available. Additional data are available upon reasonable request.
The lead authors affirm that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
Dissemination to participants and related patient and public communities: The study protocol is available online at https://osf.io/ehc47/. A preprint version of the study is publicly available on medRxiv.
This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.