Skip to main content

Table 3 Matched cohort study.

From: Design of cohort studies in chronic diseases using routinely collected databases when a prescription is used as surrogate outcome

 

Rate ratio

95% Confidence interval

P

Statin

0.879

0.758

1.021

0.091

Asthma COPD

0.815

0.676

0.983

0.032

Osteoarthritis

0.854

0.715

1.019

0.081

Hyperlipidemia

0.808

0.665

0.981

0.031

Hypertension

0.855

0.72

1.017

0.076

Peptic ulcer disease

0.763

0.538

1.082

0.13

OS

1.356

1.148

1.602

<0.001

DMARDs

1.545

1.282

1.862

<0.001

NSAID

0.757

0.637

0.899

0.002

 

Odds Ratio

95% confidence interval

P

Statin

1.025

0.900

1.169

0.704

Diabetes mellitus

1.369

1.107

1.690

0.004

Asthma COPD

0.829

0.703

0.977

0.026

Hyperlipidemia

1.428

1.210

1.685

<0.001

Hypertension

1.171

1.010

1.359

0.037

OS

0.386

0.332

0.450

<0.001

DMARD

0.433

0.363

0.517

<0.001

Loop diuretics

1.242

0.938

1.644

0.13

Anti-diabetics

1.895

1.334

2.691

<0.001

  1. Results from the zero inflated negative binomial model. Parsimonious model including all covariates (morbidities and medications at the index date)
  2. with p-value less than 0.1. The model has two components. Bottom: logistic model component predicting the probability of NOT having OS. Top: negative binomial model component predicting the rate of OS duration in an observation period of length 365 days for those patients who have OS.
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us