음악중재효과 평가를 위한 통계적 표본크기 결정 방법 고찰-비교임상시험 사례 중심

초록

Many randomized clinical trials in a range of fields demonstrating the effectiveness of music as an adjuvant method of treatment have been published but there are cases where questions have been raised on the reliability and replicability of such study results due to the study designs not properly taking into account requisite sample size. For the reason that there are no consistent standards to be employed when deducing subjective variables such as emotion, stress, pain as the effect of music, there are disagreements on proving causation in terms of selection of music, the seriousness of patients’ illness and sample size. This study aims to discuss statistical methods for determining sample size when designing randomized clinical trials to show the effects of music. Finally, the designs and analytical processes of the selected studies, statistical hypotheses, types of primary endpoints (countable vs categorical), test statistics, α-error, β-error(power), effective size were analyzed against the current principles for clinical trials published by the regulatory agencies. This study shows the necessity of robust scientific process in randomized clinical trials and presents a statistical formula for determining sample sizes.

키워드