Skip to main content

Table 2 Prevalence rates ratios and 95% confidence intervals of the consumption of cocaine, marijuana, cigarette, alcohol risk of consumption of psychoactive substances associated with depression adjusted for age groups of Colombian workers

From: Negative log-binomial model with optimal robust variance to estimate the prevalence ratio, in cross-sectional population studies

    

CI 95%

  

Estimator

Confusion effect percentage (%)

(Eq. 10)

Standard error estimation

Lower limit

Upper limit

Comparison of the standard error of the models with MH (Eq. 11)

BIC

Lifetime cocaine consumption prevalence

(prevalence = 1.8%)

 PR MH (age adjusted)

2.913

13.08

0.7570

0.786

12.845

  

 PR Negative Log-Binomial model

2.931

12.38

0.7185

0.717

11.985

5.358

 

 PR Robust Negative Log-Binomial model

2.931

12.38

0.7144

0.723

11.889

5.963

-43,857.1

 PR Cox/Poisson model

2.928

12.50

0.7161

0.720

11.918

5.711

 

 PR Cox/ Poisson Robust model

2.928

12.50

0.7152

0.721

11.896

5.845

-43,787.03

 PR Binomial regression

2.925

12.62

0.7128

0.722

11.851

6.201

 

 PR Robust binomial regression

2.925

12.62

0.7160

0.719

11.903

5.726

-43,602.83

 OR Age adjusted – logistic regression

2.965

 

0.7184

0.725

12.121

  

 OR Age adjusted – Robust logistic regression

2.965

 

0.7198

0.723

12.155

  

Marijuana consumption prevalence

(prevalence = 9.6%)

 PR MH (age adjusted)

3.407

2.52

0.3121

1.848

6.281

  

 PR Negative Log-Binomial model

3.444

1.42

0.3245

1.823

6.506

3.821

 

 PR Robust Negative Log-Binomial model

3.444

1.42

0.3154

1.856

6.391

1.046

-45,899.8

 PR Cox/Poisson model

3.440

1.54

0.3198

1.838

6.438

2.408

 

 PR Cox/ Poisson Robust model

3.440

1.54

0.3159

1.852

6.389

1.203

-45,541.37

 PR Binomial regression

3.435

1.69

0.3151

1.852

6.370

0.952

 

 PR Robust binomial regression

3.435

1.69

0.3164

1.847

6.387

1.359

-44,534.91

 OR Age adjusted – logistic regression

3.702

 

0.3246

1.959

6.995

  

 OR Age adjusted – Robust logistic regression

3.702

 

0.3255

1.956

7.007

  

Cigarette consumption prevalence

(prevalence = 21.3%)

 PR MH (age adjusted)

2.209

17.38

0.1913

1.518

3.214

  

 PR Negative Log-Binomial model

2.175

19.22

0.2091

1.443

3.276

8.513

 

 PR Robust Negative Log-Binomial model

2.175

19.22

0.1919

1.493

3.167

0.313

-36,500.12

 PR Cox/Poisson model

2.197

18.02

0.1991

1.487

3.247

3.918

 

 Cox/Poisson Robust model

2.197

18.02

0.1901

1.863

3.190

0.631

-35,956.34

 Binomial regression

2.225

16.54

0.1885

1.538

3.220

1.485

 

 Robust binomial regression

2.225

16.54

0.1878

1.540

3.215

1.864

-34,252.81

 OR Age adjusted – logistic regression

2.536

 

0.2105

1.679

3.831

  

 OR Age adjusted – Robust logistic regression

2.536

 

0.2127

1.671

3.848

  

Lifetime alcohol consumption prevalence

(prevalence = 85.7%)

 PR MH (age adjusted)

1.241

1.13

0.0361

1.157

1.332

  

 PR Negative Log-Binomial model

1.243

0.97

0.0872

1.048

1.475

58.601

 

 PR Robust Negative Log-Binomial model

1.243

0.97

0.0360

1.158

1.334

0.278

-45,293.58

 Cox/ Poisson model

1.242

1.05

0.0668

1.089

1.416

45.958

 

 Cox/Poisson Robust model

1.242

1.05

0.0359

1.157

1.333

0.557

-44,885.75

 Binomial regression

1.238

1.37

0.0449

1.153

1.329

19.599

 

 Robust binomial regression

1.238

1.37

0.0450

1.153

1.329

19.778

-41,961.72

 Age adjusted OR – logistic regression

2.810

 

0.1260

2.194

3.597

  

 Age adjusted OR – Robust logistic regression

2.810

 

0.1248

2.199

3.589

  

Risk of consumption of psychoactive substances (prevalence = 96.1%)

 PR MH (age adjusted)

1.086

0

0.0185

1.047

1.126

  

 PR Negative Log-Binomial model

1.086

0

0.0793

0.930

1.269

76.671

 

 PR Robust Negative Log-Binomial model

1.086

0

0.0186

1.047

1.127

0.538

-47,709.15

 PR Cox/Poisson mode

1.086

0

0.0576

0.970

1.216

67.882

 

 Cox/Poisson Robust model

1.086

0

0.0186

1.047

1.127

0.538

-47,583.75

 PR Binomial regression

No converge

      

 PR Robust binomial regression

No converge

      

 OR Age adjusted OR – logistic regression

3.462

 

0.1857

2.406

4.982

  

 OR Age adjusted – Robust logistic regression

3.462

 

0.1825

2.421

4.951

  
  1. Models were controlled by grouped age for all cases. Results are shown for the models of negative log-binomial, Cox regression with constant time, log-Poisson, log-binomial compared with MH, and unconditional binary logistic regression model – OR value
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us