Randomized controlled studies and the treatment of middle-ear effusions and tonsillar pharyngitis: How random are the studies and what are their limitations?

Abstract

Objective To determine the applicability and external validity of randomized control studies (RCTs) in the light of patient susceptibilities and vulnerabilities to the sequelae of otitis media with effusion (OME) and tonsillar pharyngitis (T&A) and the composition of their cohorts. Study Design RCTs for OME and T&A were analyzed to determine which intrinsic and extrinsic susceptibilities to the otolaryngic disease and its sequelae were included or excluded and the composition of the cohort. Methods A Medline and a Science Direct search were performed for all RCTs concerning OME and T&A through 2007. The articles selected to be included in this study are the OME RCTs that reported language and/or hearing outcomes and the T&A RCT studies in which the outcome measure was infection, alleviation of airway obstruction, and/or quality of life. These were analyzed for their inclusion or exclusion of intrinsic and extrinsic susceptibilities to the otolaryngic disease and their sequelae and the composition of the cohort. The pertinent otolaryngologic literature was assessed to determine applicable risk factors. Results The analysis of the inclusion or exclusion of risk factors showed that for OME 0 percent were included and 16 percent were excluded, and there were no data for 84 percent. The T&A findings were similar in that 1 percent of risk factors were included, 8 percent excluded, and there were no data for 92 percent. Both the OME and the T&A cohorts data were similar in that approximately half of the suitable candidates completed the study; 25 percent of the OME and 13 percent of the T&A enrolled subjects were either withdrawn or placed into a different experimental group. On the average, with data available, it took 4.6 years to recruit the OME sample and 5.5 years to recruit the T&A sample. Conclusion The results of RCTs for these conditions are only applicable to narrowly defined and highly circumscribed populations. They cannot be generalized to the entire patient population because of their limited external validity. Future RCT protocols should be designed to control for the intrinsic and extrinsic susceptibilities that result in a propensity to acquire the disease and/or an exacerbation of the disease's sequelae. These studies would determine the most effective strategies for preventing disorders and/or their deleterious sequelae.