Skip to main content

Table 5 Predictors of discovering at least one variant classified as likely pathogenic or pathogenic in univariate and multivariable analyses

From: Predictors of the utility of clinical exome sequencing as a first-tier genetic test in patients with Mendelian phenotypes: results from a referral center study on 603 consecutive cases

Predictor

Univariate analysis

Multivariable analysis†

Beta

(SE)

Odds ratio

(95% CI)

P value*

Beta

(SE)

Odds ratio

(95% CI)

P value‡

Suspected diagnosis subgroup, Lysosomal disorder

2.15

(1.1)

8.55

(0.99–73.84)

2.01 × 10−2

2.47

(1.11)

11.81

(1.35–103.33)

0.03

Suspected diagnosis subgroup, Hypercholesterolemia

1.28

(0.33)

3.61

(1.9–6.86)

5.62 × 10−5

1.42

(0.33)

4.14

(2.15–7.96)

 < 0.0001

Pediatric care department

0.55

(0.21)

1.73

(1.15–2.6)

8.29 × 10−3

0.67

(0.22)

1.96

(1.28–2.99)

0.002

Suspected diagnosis subgroup, cholestatic disorder

 − 1.93

(1.05)

0.15

(0.02–1.14)

1.84 × 10−2

Not retained§

  

Suspected diagnosis group, Dyslipidemia

0.56

(0.26)

1.75

(1.06–2.89)

2.89 × 10−2

Not used in the model||

  

Adult care department

 − 0.55

(0.21)

0.58

(0.38–0.87)

8.29 × 10−3

Not used in the model||

  
  1. 95% CI 95% confidence interval; Beta beta coefficient; SE standard error
  2. *Univariate logistic regression analysis
  3. Cox & Snell R2, 0.07; Nagelkerke R2, 0.10; Percent of cases correctly classified, 67%; AUROC, 0.629 (95% CI, 0.581 to 0.675)
  4. Multivariable logistic regression analysis using the stepwise method
  5. § Not retained in the multivariate logistic regression model
  6. || Items not used in multivariable logistic regression analysis to avoid collinearity with other items reported in the multivariable model (e.g., ‘Adult care department’ vs. ‘Pediatric care department’)