Developing prediction models for clinical use using logistic regression: an overview

Shipe, Maren E.; Deppen, Stephen A.; Farjah, Farhood; Grogan, Eric L.

doi:10.21037/jtd.2019.01.25

Cited by 294 publications

(203 citation statements)

References 21 publications

Supporting

Mentioning

201

Contrasting

Unclassified

Order By: Relevance

“…To evaluate the predictive value of serum thyroid hormone and TSH concentrations on survival, we performed logistic regression using the 30-day survival outcome as the dependent variable, and the serum T 4 , T 3 , fT4, and TSH concentrations as independent variables. 48,49 For this analysis, we entered serum T 4 and fT4 concentrations as continuous variables and serum T 3 and TSH concentrations as dichotomous (binary) variables (data coded 0 for undetectable concentrations; 1 for detectable concentrations). The significance of each explanatory variable was tested using the Wald test.…”

Section: Data and Statistical Analysesmentioning

confidence: 99%

Serum thyroxine and thyrotropin concentrations decrease with severity of nonthyroidal illness in cats and predict 30‐day survival outcome

Peterson

Davignon

Shaw

et al. 2020

Veterinary Internal Medicne

View full text Add to dashboard Cite

Background: In cats, nonthyroidal illness affects serum thyroid hormone concentrations. Serum thyroxine (T 4) and triiodothyronine (T 3) concentrations commonly decrease, whereas free T 4 (fT 4) concentrations vary unpredictably. Limited information exists regarding effects on serum thyrotropin (thyroid-stimulating hormone [TSH]) concentrations in cats with nonthyroidal illness syndrome (NTIS). Objectives: To characterize alterations in thyroid function that develop in cats with NTIS and to correlate these alterations with severity and outcome of the nonthyroidal illness. Animals: Two hundred and twenty-two cats with NTIS and 380 clinically normal cats of similar age and sex. Methods: Prospective, cross-sectional study. All cats had serum T 4 , T 3 , free T 4 , and TSH concentrations measured. Cats were grouped based on illness severity and 30-day survival. Results: Cats with NTIS had lower serum T 4 and T 3 concentrations than did normal cats (P < .001). Serum fT 4 and TSH concentrations did not differ between groups. Serum T 4 , T 3 , and fT 4 concentrations progressively decreased with increasing disease severity (P < .001). The 56 cats that died had lower T 4 , T 3 , and TSH concentrations than did the 166 survivors, with no difference in fT 4 concentration. Multivariable logistic regression modeling indicated that serum T 4 and TSH concentrations both predicted survival (P < .02). Conclusions and Clinical Importance: Cats with NTIS commonly develop low serum T 4 , T 3 , and TSH concentrations, the prevalence and extent of which increases with disease severity. Clinicians should consider evaluating thyroid function in cats with severe NTIS, because doing so could help determine probability of successful treatment responses before investing considerable time, effort, and finances in addressing the underlying disease.

show abstract

Section: Data and Statistical Analysesmentioning

confidence: 99%

Serum thyroxine and thyrotropin concentrations decrease with severity of nonthyroidal illness in cats and predict 30‐day survival outcome

Peterson

Davignon

Shaw

et al. 2020

Veterinary Internal Medicne

View full text Add to dashboard Cite

show abstract

“…The multivariate analysis estimates coefficients (for example, log odds or hazard ratios) for each predictor included in the final model and adjusts them with respect to the other predictors in the model. The coefficients quantify the contribution of each predictor to the outcome risk estimation 10 . The caveats to consider when assessing the results of a logistic regression analysis are well explained in Tolles et al 11 .…”

Section: Resultsmentioning

confidence: 99%

Non applicability of validated predictive models for intensive care admission and death of COVID-19 patients in a secondary care hospital in Belgium

Parisi

Mili

2020

Preprint

View full text Add to dashboard Cite

ObjectiveTo set up simple and reliable predictive scores for intensive care admissions and deaths in COVID-19 patients. These scores adhere to the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guidelines.DesignMonocentric retrospective cohort study run from early March to end of May in Clinique Saint-Pierre Ottignies, a secondary care hospital located in Ottignies-Louvain-la-Neuve, Belgium. The outcomes of the study are (i) admission in the Intensive Care Unit and (ii) death.Data sourcesAll patients admitted in the Emergency Department with a positive RT-PCR SARS-CoV-2 test were included in the study. Routine clinical and laboratory data were collected at their admission and during their stay. Chest X-Rays and CT-Scans were performed and analyzed by a senior radiologist.MethodsA recently published predictive score conducted on a large scale was used as a benchmark value (Liang score)1. Logistic regressions were used to develop predictive scores for (i) admission to ICU among emergency ward patients; (ii) death among ICU patients on 40 clinical variables. These models were based on medical intuition and simple model selection tools. Their predictive capabilities were then compared to Liang score.ResultsOur results suggest that Liang score may not provide reliable guidance for ICU admission and death. Moreover, the performance of this approach is clearly outperformed by models based on simple markers. For example, a logistic regression considering only the LDH yields to similar sensitivity and greater specificity. Finally, all models considered in this study lead to levels of specificity under or equal to 50%.ConclusionsIn our experience, the results of a predictive score based on a large-scale Chinese study cannot be applied in the Belgian population. However, in our small cohort it appears that LDH above 579 UI/L and venous lactate above 3.02 mmol/l may be considered as good predictive biological factors for ICU admission. With regards to death risk, NLR above 22.1, tobacco abuse status and 80 % of respiratory impairment appears to be relevant predictive factors. A predictive score for admission to ICU or death is desperately needed in secondary hospitals. Optimal allocation of resources guided by evidence-based indicators will best guide patients at time of admission and avoid futile treatments in intensive care units.

show abstract

“…The multivariate analysis estimates coe cients (for example, log odds or hazard ratios) for each predictor included in the nal model and adjusts them with respect to the other predictors in the model. The coe cients quantify the contribution of each predictor to the outcome risk estimation 10 . The caveats to consider when assessing the results of a logistic regression analysis are well explained in Tolles et al 11 .…”

Section: Resultsmentioning

confidence: 99%

Non Applicability of Validated Predictive Models for Intensive Care Admission and Death of COVID-19 Patients in a Secondary Care Hospital in Belgium.

Parisi

Janier-Dubry

Ponzetto

et al. 2020

Preprint

View full text Add to dashboard Cite

BackgroundSimple and reliable predictive scores for intensive care admissions and death based on clinical data are still lacking. The goal of this study is to implement such scores based on patients coming from our population catchment area and to compare them to available ones. These scores adhere to the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guidelines.MethodsMonocentric retrospective cohort study run from early March to end of May in Clinique Saint-Pierre Ottignies, a secondary care hospital located in Ottignies-Louvain-la-Neuve, Belgium. The outcomes of the study are (i) admission in the Intensive Care Unit and (ii) death. All patients admitted in the Emergency Department with a positive RT-PCR SARS-CoV-2 test were included in the study. Routine clinical and laboratory data were collected at their admission and during their stay. Chest X-Rays and CT-Scans were performed and analyzed by a senior radiologist. A recently published predictive score conducted on a large scale was used as a benchmark value (Liang score)1. Logistic regressions were used to develop predictive scores for (i) admission to ICU among emergency ward patients; (ii) death among ICU patients on 40 clinical variables. These models were based on medical intuition and simple model selection tools. Their predictive capabilities were then compared to Liang score. ConclusionsOur results suggest that Liang score may not provide reliable guidance for ICU admission and death. Predictive score based on a large-scale Chinese study cannot be applied in the Belgian population, suggesting a predominant role in genetic factors for SARS-CoV-2 susceptibility. In our small cohort, it appears that LDH above 579 UI/L and venous lactate above 3.02 mmol/l may be considered as good predictive biological factors for ICU admission. With regards to death risk, Neutrophil/lymphocyte ratio above 22.1, tobacco abuse status and respiratory impairment appears to be relevant predictive factors. Predictive score for admission to ICU or death based on clinical data are still needed in secondary hospital and hospitals in developing countries that do not have high-performance laboratories. Optimal allocation of resources guided by evidence-based indicators will best guide patients at time of admission and avoid futile treatments in intensive care units.

show abstract

Developing prediction models for clinical use using logistic regression: an overview

Cited by 294 publications

References 21 publications

Serum thyroxine and thyrotropin concentrations decrease with severity of nonthyroidal illness in cats and predict 30‐day survival outcome

Serum thyroxine and thyrotropin concentrations decrease with severity of nonthyroidal illness in cats and predict 30‐day survival outcome

Non applicability of validated predictive models for intensive care admission and death of COVID-19 patients in a secondary care hospital in Belgium

Non Applicability of Validated Predictive Models for Intensive Care Admission and Death of COVID-19 Patients in a Secondary Care Hospital in Belgium.

Contact Info

Product

Resources

About