Hypothesis Testing Example - GSK Respiratory Research

GSK Respiratory Research

Title: Does a 10-valent pneumococcal-Haemophilus influenzae protein D conjugate vaccine prevent respiratory exacerbations in children with recurrent protracted bacterial bronchitis, chronic suppurative lung disease and bronchiectasis: protocol for a randomised controlled trial.

Objective: Our primary objective is to determine the clinical efficacy of PHiD-CV in reducing respiratory exacerbations in children aged 18 months to 14 years with recurrent PBB, CSLD or bronchiectasis.

Year: 2013

Source: Trials

Link: https://trialsjournal.biomedcentral.com/articles/10.1186/1745-6215-14-282

Clinical Area: Respiratory

Sample Size Section in Paper/Protocol:

“Prospective data collected in Brisbane children with non-CF bronchiectasis indicate a mean annual incidence of 2.1 (standard deviation 1.046) exacerbations requiring hospital clinic attendance or hospitalisation. PHiD-CV efficacy against exacerbations in children with bronchiectasis is unknown. However, based on the aforementioned pilot data and AOM data from the POET study (30% reduction) [33], our trial is powered to detect a 30% reduction from 2.1 to 1.47 exacerbations in the 12 months following the second vaccine dose. Assuming a Poisson distribution, 93 children per group will provide 90% power (α = 0.05, two-sided) to detect a 30% reduction in exacerbations in the PHiD-CV group and 80% power to detect a 25% reduction. Assuming a 10% loss to follow-up, we will recruit 206 children (103 per group). Differences of less than 20% are unlikely to change clinical practice without additional supporting evidence.”

Summary of Necessary Parameter Estimates for Sample Size Calculation:

Parameter	Value
Significance Level (1-Sided)	0.025
Control Group Rate	2.1
Rate Ratio	0.7
Exposure Time (years)	1
Allocation Ratio	1
Power	90%
Expected Dropout	1%

Step 1:
Select the Test for the Ratio of Two Incidence Rates using the Poisson Model from the Select Test Design & Goal window.

This can be done using the radio buttons or alternatively, you can use the search bar at the end of the Select Test Design & Goal window.

Step 2:
Enter the parameter values for sample size calculation taken from the study protocol.

Step 3:
Select Run to solve for sample size.

This gives a sample size of 93 per group, as per the study protocol.

Step 4:
Once the calculation is completed, nQuery Advanced provides an output statement summarizing the results. It States:

Output Statement:

In a study testing ratio of independent two incidence rates under a Poisson model rate, a sample of 93 subjects in group 1 observed for 1 time periods and a sample of 93 subjects in group 2 observed for 1 time periods achieves 90.14% power to detect an incidence rate ratio (γ₁/γ₀) of 0.7 (assuming the incidence rate ratio is 1 under the null hypothesis) when the mean incidence rate for group 1 is 2.1 at the 0.025 significance level using a one-sided test. This was done under the assumption that the test statistic would be the ln(CMLE).

Sample Size Determination Example

AREA: Means | Respiratory

GSK Respiratory Research

Title: Does a 10-valent pneumococcal-Haemophilus influenzae protein D conjugate vaccine prevent respiratory exacerbations in children with recurrent protracted bacterial bronchitis, chronic suppurative lung disease and bronchiectasis: protocol for a randomised controlled trial.

Email me this example in PDF format

Email me this example in PDF format

Free 14 Day Trial

LOCATION