The relationship between the external and internal validity of randomized controlled trials: a sample of hypertension trials from China

Article type
Authors
Zhang X1, Wu Y1, Kang D1, Liu X1, Peng L1, Hong Q1
1West China Hospital, China
Abstract
Objectives:
To explore the relationship between the external and internal validity of hypertension randomized controlled trials (RCTs) conducted in China.

Methods:
Comprehensive literature searches were performed in MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL) and others, and advanced search strategies were used to locate hypertension RCTs conducted in China between 1996 and 2009. The risk of bias in RCTs was assessed by a modified scale, and then studies with three or more grading scores were included for the purpose of evaluating external validity. A data extraction form that included four domains and 25 items was used to explore the external validity and relationship with internal validity. Univariate and multivariate analyses were done using SPSS software, version 21.0 (SPSS, Chicago, IL).

Results:
Two-hundred and twenty-six (226) hypertension RCTs were included for final analysis. RCTs conducted in university affiliated hospitals (P value < 0.001) or secondary/tertiary hospitals (P value < 0.001) scored higher for internal validity. Multi-center studies (median = 4.0, interquartile range (IQR) = 2.0) scored higher for internal validity than single-center studies (median = 3.0, IQR = 1.0) (P value < 0.001). Funding-supported trials had better methodological quality (P value < 0.001). In addition, reporting of inclusion criteria also leads to better internal validity (P value 0.004). Multivariate analyses indicated that sample size, industry-funding, the reporting of quality of life and a university affiliated hospital setting had statistical significance (P value < 0.001, P value < 0.001, P value 0.001, P value 0.006, respectively).

Conclusions:
Several components of the external validity of RCTs related to the internal validity do not stand in easy relation to each other. Regarding the poor reporting, other possible links between two variables need to trace in the future methodological researches.