Analysis of secondary data sources (such as cohort studies, survey data, and administrative records) has the potential to provide answers to science and society's most pressing questions. However, researcher biases can lead to questionable research practices in secondary data analysis, which can distort the evidence base. While pre-registration can help to protect against researcher biases, it presents challenges for secondary data analysis. In this article, we describe these challenges and propose novel solutions and alternative approaches. Proposed solutions include approaches to (1) address bias linked to prior knowledge of the data, (2) enable pre-registration of non-hypothesis-driven research, (3) help ensure that pre-registered analyses will be appropriate for the data, and (4) address difficulties arising from reduced analytic flexibility in pre-registration. For each solution, we provide guidance on implementation for researchers and data guardians. The adoption of these practices can help to protect against researcher bias in secondary data analysis, to improve the robustness of research based on existing data.

Self-perceptions of own social position are potentially a key aspect of socioeconomic inequalities in health, but their association with mortality remains poorly understood. We examined whether subjective social status (SSS), a measure of the self-perceived element of social position, was associated with mortality and its role in the associations between objective socioeconomic position (SEP) measures and mortality. We used Cox regression to model the associations between SSS, objective SEP measures and mortality in a sample of 9972 people aged ≥ 50 years from the English Longitudinal Study of Ageing over a 10-year follow-up (2002-2013). Our findings indicate that SSS was associated with all-cause, cardiovascular, cancer and other mortality. A unit decrease in the 10-point continuous SSS measure increased by 24 and 8% the mortality risk of people aged 50-64 and ≥ 65 years, respectively, after adjustment for age, sex and marital status. The respective estimates for cardiovascular mortality were 36 and 11%. Adjustment for all covariates fully explained the association between SSS and cancer mortality, and partially the remaining associations. In people aged 50-64 years, SSS mediated to a varying extent the associations between objective SEP measures and all-cause mortality. In people aged ≥ 65 years, SSS mediated to a lesser extent these associations, and to some extent was associated with mortality independent of objective SEP measures. Nevertheless, in both age groups, wealth partially explained the association between SSS and mortality. In conclusion, SSS is a strong predictor of mortality at older ages, but its role in socioeconomic inequalities in mortality appears to be complex.

Although high body-mass index (BMI) is associated with increased risk of developing colorectal cancer (CRC), many CRC patients lose weight before diagnosis. BMI is often reported close to diagnosis, which may have led to underestimation or even reversal of direction of the BMI-CRC association. We aimed to assess if and to what extent potential bias from prediagnostic weight loss has been considered in available epidemiological evidence. We searched PubMed and Web of Science until May 2022 for systematic reviews and meta-analyses investigating the BMI-CRC association. Information on design aspects and results was extracted, including if and how the reviews handled prediagnostic weight loss as a potential source of bias. Additionally, we analyzed how individual cohort studies included in the latest systematic review handled the issue. Overall, 18 reviews were identified. None of them thoroughly considered or discussed prediagnostic weight loss as a potential source of bias. The majority (15/21) of cohorts included in the latest review did not exclude any initial years of follow-up from their main analysis. Although the majority of studies reported having conducted sensitivity analyses in which initial years of follow-up were excluded, results were reported very heterogeneously and mostly for additional exclusions of 1-2 years only. Where explicitly reported, effect estimates mostly increased with increasing length of exclusion. The impact of overweight and obesity on CRC risk may be larger than suggested by the existing epidemiological evidence.

Several systematic reviews and meta-analyses have summarized the association between sedentary behavior (SB) and cancer. However, the level of evidence and the potential for risk of bias remains unclear. This umbrella review summarized the current data on SB in relation to cancer incidence and mortality, with a particular emphasis on assessing the risk of bias. We searched PubMed, Web of Science and Cochrane Database for systematic reviews and meta-analyses on the association between SB and cancer incidence and mortality. We also searched for recent observational studies not yet included in existing meta-analyses. We re-calculated summary risk estimates for cancer incidence and mortality using random effects models. We included 14 meta-analyses covering 17 different cancer sites from 77 original studies. We found that high SB levels increase the risk for developing ovarian, endometrial, colon, breast, prostate, and rectal cancers, with relative risks of 1.29 (95% confidence interval (CI) = 1.08-1.56), 1.29 (95% CI = 1.16-1.45), 1.25 (95% CI = 1.16-1.33), 1.08 (95% CI = 1.04-1.11), 1.08 (95% CI = 1.00-1.17), and 1.07 (95% CI = 1.01-1.12), respectively. Also, we found an increased risk of cancer mortality of 1.18 (95% CI = 1.09-1.26). Most associations between SB and specific cancer sites were supported by a "suggestive" level of evidence. High levels of SB are associated with increased risk of several types of cancer and increased cancer mortality risk.

The carcinogenicity of opium consumption was recently evaluated by a Working Group convened by the International Agency for Research on Cancer (IARC). We supplement the recent IARC evaluation by conducting an extended systematic review as well as a quantitative meta-analytic assessment of the role of opium consumption and risk for selected cancers, evaluating in detail various aspects of study quality on meta-analytic findings. We searched the published literature to identify all relevant studies on opium consumption and risk of selected cancers in humans through 31 October, 2022. Meta-relative risks (mRRs) and associated 95% confidence intervals (CIs) were estimated using random-effects models for studies of cancer of the urinary bladder, larynx, lung, oesophagus, pancreas, and stomach. Heterogeneity among studies was assessed using the I2 statistic. We assessed study quality and conducted sensitivity analyses to evaluate the impact of potential reverse causation, protopathic bias, selection bias, information bias, and confounding. In total, 2 prospective cohort studies and 33 case-control studies were included. The overall pooled mRR estimated for 'ever or regular' versus 'never' use of opium ranged from 1.50 (95% CI 1.13-1.99, I2 = 0%, 6 studies) for oesophageal cancer to 7.97 (95% CI 4.79-13.3, I2 = 62%, 7 studies) for laryngeal cancer. Analyses of cumulative opium exposure suggested greater risk of cancer associated with higher opium consumption. Findings were robust in sensitivity analyses excluding studies prone to potential methodological sources of biases and confounding. Findings support an adverse association between opium consumption and cancers of the urinary bladder, larynx, lung, oesophagus, pancreas and stomach.

Several studies indicate that prognosis for survival in Duchenne muscular dystrophy (DMD) has improved in recent decades. However, published evidence is inconclusive and some estimates may be obsolete due to improvements in standards of care, in particular the routine use of mechanical ventilatory support in advanced stages of the disease. In this systematic review and meta-analysis (PROSPERO identifier: CRD42019121800), we searched MEDLINE (through PubMed), CINAHL, Embase, PsycINFO, and Web of Science for studies published from inception up until December 31, 2018, reporting results of life expectancy in DMD. We pooled median survival estimates from individual studies using the median of medians, and weighted median of medians, methods. Risk of bias was established with the Newcastle-Ottawa Scale. Results were stratified by ventilatory support and risk of bias. We identified 15 publications involving 2662 patients from 12 countries from all inhabited continents except Africa. Median life expectancy without ventilatory support ranged between 14.4 and 27.0 years (pooled median: 19.0 years, 95% CI 18.0-20.9; weighted pooled median: 19.4 years, 18.2-20.1). Median life expectancy with ventilatory support, introduced in most settings in the 1990s, ranged between 21.0 and 39.6 years (pooled median: 29.9 years, 26.5-30.8; weighted pooled median: 31.8 years, 29.3-36.2). Risk of bias had little impact on pooled results. In conclusion, median life expectancy at birth in DMD seems to have improved considerably during the last decades. With current standards of care, many patients with DMD can now expect to live into their fourth decade of life.

Numerous epidemiologic studies and a few systematic reviews have investigated the association between occupational solar exposure and basal cell carcinoma (BCC). However, previous reviews have several deficits with regard to included and excluded studies/risk estimates and the assessment of risk of selection bias (RoSB). Our aim was to review epidemiologic studies with a focus on these deficits and to use meta-(regression) analyses to summarize risk estimates.

Dietary patterns in childhood have been associated with child neurodevelopment and cognitive performance, while the underlying neurobiological pathway is unclear. We aimed to examine associations of dietary patterns in infancy and mid-childhood with pre-adolescent brain morphology, and whether diet-related differences in brain morphology mediate the relation with cognition. We included 1888 and 2326 children with dietary data at age one or eight years, respectively, and structural neuroimaging at age 10 years in the Generation R Study. Measures of brain morphology were obtained using magnetic resonance imaging. Dietary intake was assessed using food-frequency questionnaires, from which we derived diet quality scores based on dietary guidelines and dietary patterns using principal component analyses. Full scale IQ was estimated using the Wechsler Intelligence Scale for Children-Fifth Edition at age 13 years. Children with higher adherence to a dietary pattern labeled as 'Snack, processed foods and sugar' at age one year had smaller cerebral white matter volume at age 10 (B = -4.3, 95%CI -6.9, -1.7). At age eight years, higher adherence to a 'Whole grains, soft fats and dairy' pattern was associated with a larger total brain (B = 8.9, 95%CI 4.5, 13.3), and larger cerebral gray matter volumes at age 10 (B = 5.2, 95%CI 2.9, 7.5). Children with higher diet quality and better adherence to a 'Whole grains, soft fats and dairy' dietary pattern at age eight showed greater brain gyrification and larger surface area, clustered primarily in the dorsolateral prefrontal cortex. These observed differences in brain morphology mediated associations between dietary patterns and IQ. In conclusion, dietary patterns in early- and mid-childhood are associated with differences in brain morphology which may explain the relation between dietary patterns and neurodevelopment in children.

In the last decade, many studies have examined associations between poor psychosocial work environment and depression. We aimed to assess the evidence for a causal association between psychosocial factors at work and depressive disorders. We conducted a systematic literature search from 1980 to March 2019. For all exposures other than night and shift work and long working hours, we limited our selection of studies to those with a longitudinal design. We extracted available risk estimates for each of 19 psychosocial exposures, from which we calculated summary risk estimates with 95% confidence intervals (PROSPERO, identifier CRD42019130266). 54 studies were included, addressing 19 exposures and 11 different measures of depression. Only data on depressive episodes were sufficient for evaluation. Heterogeneity of exposure definitions and ascertainment, outcome measures, risk parameterization and effect contrasts limited the validity of meta-analyses. Summary risk estimates were above unity for all but one exposure, and below 1.60 for all but another. Outcome measures were liable to high rates of false positives, control of relevant confounding was mostly inadequate, and common method bias was likely in a large proportion of studies. The combination of resulting biases is likely to have inflated observed effect estimates. When statistical uncertainties and the potential for bias and confounding are taken into account, it is not possible to conclude with confidence that any of the psychosocial exposures at work included in this review is either likely or unlikely to cause depressive episodes or recurrent depressive disorders.

Dietary trans fatty acids (TFAs) are primarily industrially produced and remain abundant in processed food, particularly in low- and middle-income countries. Although TFAs are a cause of adverse cardiometabolic outcomes, little is known about exposure to TFAs in relation to brain development. We aimed to investigate the effect of maternal TFA concentration during pregnancy on offspring head growth in utero and during childhood. In a prospective population-based study in Rotterdam, the Netherlands, with 6900 mother-child dyads, maternal plasma TFA concentration was assessed using gas chromatography in mid-gestation. Offspring head circumference (HC) was measured in the second and third trimesters using ultrasonography; childhood brain morphology was assessed using magnetic resonance imaging at age 10 years. We performed regression analyses adjusting for sociodemographic and lifestyle confounders and instrumental variable (IV) analyses. Our IV analysis leveraged a national policy change that led to a substantial reduction in TFA and occurred mid-recruitment. After adjusting for covariates, maternal TFA concentration during pregnancy was inversely related to fetal HC in the third trimester (mean difference per 1% wt:wt increase: - 0.33, 95% CI - 0.51, - 0.15, cm) and to fetal HC growth from the second to the third trimester (- 0.04, 95% CI - 0.06, - 0.02, cm/week). Consistent findings were obtained with IV analyses, strengthening a causal interpretation. Association between prenatal TFA exposure and HC in the second trimester or global brain volume at age 10 years was inconclusive. Our findings are of important public health relevance as TFA levels in food remain high in many countries.

Smoking is an established risk factor for cardiovascular disease including coronary heart disease and stroke, however, data regarding smoking and sudden cardiac death have not been summarized in a meta-analysis previously. We therefore conducted a systematic review and meta-analysis to clarify this association. We searched the PubMed and Embase databases for studies of smoking and sudden cardiac death up to July 20th 2017. Prospective studies were included if they reported adjusted relative risk (RR) estimates and 95% confidence intervals (CIs) for smoking and sudden cardiac death. Summary RRs were estimated by use of a random effects model. Twelve prospective studies were included. The summary RR was 3.06 (95% CI 2.46-3.82, I2 = 41%, pheterogeneity = 0.12, n = 7) for current smokers and 1.38 (95% CI 1.20-1.60, I2 = 0%, pheterogeneity = 0.55, n = 7) for former smokers compared to never smokers. For four studies using non-current (never + former) smokers as the reference category the summary RR among current smokers was 2.08 (95% CI 1.70-2.53, I2 = 18%, pheterogeneity = 0.30). The results persisted in most of the subgroup analyses. There was no evidence of publication bias. These results confirm that smoking increases the risk of sudden cardiac death. Any further studies should investigate in more detail the effects of duration of smoking, number of cigarettes per day, pack-years, and time since quitting smoking and sudden cardiac death.

While there is strong epidemiological evidence that circulating insulin-like growth factor-I (IGF-I) is associated with a higher risk of several cancers, little is known about its association with non-cancer outcomes. We investigated associations of circulating IGF-I with risk of 25 common conditions, other than cancer, in a large British cohort. Study participants were 318,749 middle-aged adults enrolled in the UK Biobank Study. Serum IGF-I concentration was measured in samples collected at baseline (2006-2010), and re-measured in 12,334 participants after an average of 4.3 years. We followed-up participants over an average of 11.5 years by linking to hospital admissions and mortality registries. Multivariable-adjusted Cox regressions estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for associations between circulating IGF-I and 25 common conditions, using the repeated IGF-I measurements to correct for regression dilution bias. After correction for multiple testing (P < 0.002), IGF-I was positively associated with carpal tunnel syndrome (HR per 5 nmol/l higher concentration = 1.12, 95% CI 1.08-1.16), and inversely associated with varicose veins (0.90, 0.85-0.95), cataracts (0.97, 0.95-0.99), diabetes (0.92, 0.90-0.95), and iron deficiency anaemia (0.90, 0.86-0.93). The associations for cataracts and diabetes attenuated when restricted to cases diagnosed after five or more years of follow-up, suggesting that these associations were likely affected by reverse causality. Higher IGF-I concentration might be associated with the risk for several conditions, but genetic studies are needed to clarify which associations may be causal.

Cardiovascular disease is the leading cause of death worldwide, while sudden cardiac death (SCD) accounts for over 60% of all cardiovascular deaths. Elevated blood pressure and hypertension have been associated with increased risk of SCD, but the findings have not been consistent. To clarify whether blood pressure or hypertension is associated with increased risk of SCD and to quantify the size and the shape of any association observed. PubMed and Embase databases were searched for published prospective studies on blood pressure or hypertension and SCD up to 30 April 2018. Summary relative risks (RRs) and 95% confidence intervals (CIs) were calculated using a random effects model. The meta-analysis included 2939 SCDs among 418,235 participants from 18 studies. The summary RRs were 2.10 (95% CI 1.71-2.58, I2 = 56.7%, pheterogeneity = 0.018, n = 10) for prevalent hypertension, 1.28 (95% CI 1.19-1.38, I2 = 45.5%, pheterogeneity = 0.07, n = 10) per 20 mmHg increment in systolic blood pressure (SBP) and 1.09 (95% CI 0.83-1.44, I2 = 83.4%, pheterogeneity = 0.002, n = 3) per 10 mmHg increment in diastolic blood pressure (DBP). A nonlinear relationship was suggested between SBP and SCD. The results persisted in most subgroup and sensitivity analyses. There was no evidence of publication bias. This meta-analysis found an increased risk of SCD with hypertension diagnosis and increasing SBP. Future studies should clarify the association for DBP and the shape of the dose-response relationship between blood pressure and SCD.

A diagnosis of diabetes mellitus and prediabetes has been associated with increased risk of Parkinson's disease (PD) in several studies, but results have not been entirely consistent. We conducted a systematic review and meta-analysis of cohort studies on diabetes mellitus, prediabetes and the risk of PD to provide an up-to-date assessment of the evidence. PubMed and Embase databases were searched for relevant studies up to 6th of February 2022. Cohort studies reporting adjusted relative risk (RR) estimates and 95% confidence intervals (CIs) for the association between diabetes, prediabetes and Parkinson's disease were included. Summary RRs (95% CIs) were calculated using a random effects model. Fifteen cohort studies (29.9 million participants, 86,345 cases) were included in the meta-analysis. The summary RR (95% CI) of PD for persons with diabetes compared to persons without diabetes was 1.27 (1.20-1.35, I2 = 82%). There was no indication of publication bias, based on Egger's test (p = 0.41), Begg's test (p = 0.99), and inspection of the funnel plot. The association was consistent across geographic regions, by sex, and across several other subgroup and sensitivity analyses. There was some suggestion of a stronger association for diabetes patients reporting diabetes complications than for diabetes patients without complications (RR = 1.54, 1.32-1.80 [n = 3] vs. 1.26, 1.16-1.38 [n = 3]), vs. those without diabetes (pheterogeneity=0.18). The summary RR for prediabetes was 1.04 (95% CI: 1.02-1.07, I2 = 0%, n = 2). Our results suggest that patients with diabetes have a 27% increased relative risk of developing PD compared to persons without diabetes, and persons with prediabetes have a 4% increase in RR compared to persons with normal blood glucose. Further studies are warranted to clarify the specific role age of onset or duration of diabetes, diabetic complications, glycaemic level and its long-term variability and management may play in relation to PD risk.

Imaging plays an essential role in research on neurological diseases in the elderly. The Rotterdam Scan Study was initiated as part of the ongoing Rotterdam Study with the aim to elucidate the causes of neurological disease by performing imaging of the brain in a prospective population-based setting. Initially, in 1995 and 1999, random subsamples of participants from the Rotterdam Study underwent neuroimaging, whereas from 2005 onwards MRI has been implemented into the core protocol of the Rotterdam Study. In this paper, we discuss the background and rationale of the Rotterdam Scan Study. Moreover, we describe the imaging protocol, image post-processing techniques, and the main findings to date. Finally, we provide recommendations for future research, which will also be topics of investigation in the Rotterdam Scan Study.

The inverse association between physical activity and arterial thrombotic disease is well established. Evidence on the association between physical activity and venous thromboembolism (VTE) is divergent. We conducted a systematic review and meta-analysis of published observational prospective cohort studies evaluating the associations of physical activity with VTE risk. MEDLINE, Embase, Web of Science, and manual search of relevant bibliographies were systematically searched until 26 February 2019. Extracted relative risks (RRs) with 95% confidence intervals (CIs) for the maximum versus minimal amount of physical activity groups were pooled using random effects meta-analysis. Twelve articles based on 14 unique prospective cohort studies comprising of 1,286,295 participants and 23,753 VTE events were eligible. The pooled fully-adjusted RR (95% CI) of VTE comparing the most physically active versus the least physically active groups was 0.87 (0.79-0.95). In pooled analysis of 10 studies (288,043 participants and 7069 VTE events) that reported risk estimates not adjusted for body mass index (BMI), the RR (95% CI) of VTE was 0.81 (0.70-0.93). The associations did not vary by geographical location, age, sex, BMI, and methodological quality of studies. There was no evidence of publication bias among contributing studies. Pooled observational prospective cohort studies support an association between regular physical activity and low incidence of VTE. The relationship does not appear to be mediated or confounded by BMI.

Much theory asserts that sexual intimacy sustains mental health. Experimental tests of such theory remain rare and have not provided compelling evidence because ethical, practical, and cultural constraints bias samples and results. An epidemiologic approach would, therefore, seem indicated given the rigor the discipline brings to quasi-experimental research. For reasons that remain unclear, however, epidemiologist have largely ignored such theory despite the plausibility of the processes implicated, which engender, for example, happiness, feelings of belonging and self-worth, and protection against depression. We use an intent-to-treat design, implemented via interrupted time-series methods, to test the hypothesis that the monthly incidence of suicide, a societally important distal measure of mental health in a population, decreased among Swedish men aged 50-59 after July 2013 when patent rights to sildenafil (i.e., Viagra) ceased, prices fell, and its use increased dramatically. The test uses 102 pre, and 18 post, price-drop months. 65 fewer suicides than expected occurred among men aged 50-59 over test months following the lowering of sildenafil prices. Our findings could not arise from shared trends or seasonality, biased samples, or reverse causation. Our results would appear by chance fewer than once in 10,000 experiments. Our findings align with theory indicating that sexual intimacy reinforces mental health. Using suicide as our distal measure of mental health further implies that public health programming intended to address the drivers of self-destructive behavior should reduce barriers to intimacy in the middle-aged populations.

Live kidney donors are exhaustively screened pre-donation, creating a cohort inherently healthier at baseline than the general population. In recent years, three renowned research groups reported unfavourable outcomes for live kidney donors post-donation that contradicted their previous studies. Here, we compared the study design and analysis of the most recent and previous studies to determine whether the different outcomes were due to methodological design or reflect a real potential disadvantage for living kidney donors. All six studies on long-term risk after live kidney donation were thoroughly screened for the selection of study population, controls, data quality, and statistical analysis. Our detailed review of the methodology revealed key differences with respect to selection of donors and compared non-donors, data quality, follow-up duration, and statistical analysis. In all studies, the comparison group of non-donors was healthier than the donors due to more extensive exclusion criteria for non-donors. Five of the studies used both restriction and matching to address potential confounding. Different matching strategies and statistical analyses were used in the more recent studies compared to previous studies and follow-up was longer. Recently published papers still face bias. Strong points compared to initial analyses are the extended follow-up time, large sample sizes and better analysis, hence increasing the reliability to estimate potential risks for living kidney donors on the long-term. Future studies should focus on equal selection criteria for donors and non-donors, and in the analysis, follow-up duration, matched sets, and low absolute risks among donors should be accounted for when choosing the statistical technique.

The objective of this meta-analysis is to evaluate the odds of colorectal adenoma (CRA) in colorectal cancer screening participants with different body mass index (BMI) levels, and examine if this association was different according to gender and ethnicity. The EMBASE and MEDLINE were searched to enroll high quality observational studies that examined the association between investigator-measured BMI and colonoscopy-diagnosed CRA. Data were independently extracted by two reviewers. A random-effects meta-analysis was conducted to estimate the summary odds ratio (SOR) for the association between BMI and CRA. The Cochran's Q statistic and I2 analyses were used to assess the heterogeneity. A total of 17 studies (168,201 subjects) were included. When compared with subjects having BMI < 25, individuals with BMI 25-30 had significantly higher risk of CRA (SOR 1.44, 95% CI 1.30-1.61; I2 = 43.0%). Subjects with BMI ≥ 30 had similarly higher risk of CRA (SOR 1.42, 95% CI 1.24-1.63; I2 = 18.5%). The heterogeneity was mild to moderate among studies. The associations were significantly higher than estimates by previous meta-analyses. There was no publication bias detected (Egger's regression test, p = 0.584). Subgroup analysis showed that the magnitude of association was significantly higher in female than male subjects (SOR 1.43, 95% CI 1.30-1.58 vs. SOR 1.16, 95% CI 1.07-1.24; different among different ethnic groups (SOR 1.72, 1.44 and 0.88 in White, Asians and Africans, respectively) being insignificant in Africans; and no difference exists among different study designs. In summary, the risk conferred by BMI for CRA was significantly higher than that reported previously. These findings bear implications in CRA risk estimation.

Current evidence on COVID-19 prognostic models is inconsistent and clinical applicability remains controversial. We performed a systematic review to summarize and critically appraise the available studies that have developed, assessed and/or validated prognostic models of COVID-19 predicting health outcomes. We searched six bibliographic databases to identify published articles that investigated univariable and multivariable prognostic models predicting adverse outcomes in adult COVID-19 patients, including intensive care unit (ICU) admission, intubation, high-flow nasal therapy (HFNT), extracorporeal membrane oxygenation (ECMO) and mortality. We identified and assessed 314 eligible articles from more than 40 countries, with 152 of these studies presenting mortality, 66 progression to severe or critical illness, 35 mortality and ICU admission combined, 17 ICU admission only, while the remaining 44 studies reported prediction models for mechanical ventilation (MV) or a combination of multiple outcomes. The sample size of included studies varied from 11 to 7,704,171 participants, with a mean age ranging from 18 to 93 years. There were 353 prognostic models investigated, with area under the curve (AUC) ranging from 0.44 to 0.99. A great proportion of studies (61.5%, 193 out of 314) performed internal or external validation or replication. In 312 (99.4%) studies, prognostic models were reported to be at high risk of bias due to uncertainties and challenges surrounding methodological rigor, sampling, handling of missing data, failure to deal with overfitting and heterogeneous definitions of COVID-19 and severity outcomes. While several clinical prognostic models for COVID-19 have been described in the literature, they are limited in generalizability and/or applicability due to deficiencies in addressing fundamental statistical and methodological concerns. Future large, multi-centric and well-designed prognostic prospective studies are needed to clarify remaining uncertainties.