Article type
Year
Abstract
Background: Well designed N-of-1 trials provide a valuable alternative methodology for evaluating the therapeutic efficacy of pharmacological or behavioural interventions when randomized controlled trials (RCTs) are not feasible.
Objectives: To evaluate methodological design, analysis and meta-analysis used in N-of-1 trials. Search strategy: We searched the following databases from their inception to August 2007: MEDLINE, PsycINFO, ERIC, Sociology Abstracts, PsycBITE, AMED, EMBASE, CINAHL, and the Cochrane Database of Systematic Reviews. Reference lists of included studies were checked. Selection criteria: Original studies specifically dealing with methodological design and statistical analysis of N-of-1 trials. Reports of N-of-1 trials themselves were not eligible for this review. Data collection and analysis: Two reviewers independently assessed the suitability of studies for inclusion. For each paper, two authors independently extracted content information for 64 predefined items specifically related to design, methodology, statistical analysis and meta-analysis.
Results: We identified 95 papers that satisfied the inclusion criteria. Of these, 51 were general reviews discussing strengths, limitations and sources of bias of basic N-of-1 designs (e.g., A-B-A-B, multiple baseline). Nine of these specifically reviewed randomisation test procedures as an adjunct for increasing methodological robustness. An additional 44 papers discussed specific data analysis techniques. Of these, 22 dealt with statistical analysis techniques exclusively, 3 with visual analysis exclusively, and 11 with comparison of statistical and visual analysis, highlighting the strengths and weaknesses of each approach. Eight papers discussed methods for quantitative synthesis of data.
Conclusions: Although consensus exists about aspects of basic design and analysis of data for N-of-1 trials to ensure internal and external validity, issues related to generalisability of results and meta-analysis of data remain contentious.
Objectives: To evaluate methodological design, analysis and meta-analysis used in N-of-1 trials. Search strategy: We searched the following databases from their inception to August 2007: MEDLINE, PsycINFO, ERIC, Sociology Abstracts, PsycBITE, AMED, EMBASE, CINAHL, and the Cochrane Database of Systematic Reviews. Reference lists of included studies were checked. Selection criteria: Original studies specifically dealing with methodological design and statistical analysis of N-of-1 trials. Reports of N-of-1 trials themselves were not eligible for this review. Data collection and analysis: Two reviewers independently assessed the suitability of studies for inclusion. For each paper, two authors independently extracted content information for 64 predefined items specifically related to design, methodology, statistical analysis and meta-analysis.
Results: We identified 95 papers that satisfied the inclusion criteria. Of these, 51 were general reviews discussing strengths, limitations and sources of bias of basic N-of-1 designs (e.g., A-B-A-B, multiple baseline). Nine of these specifically reviewed randomisation test procedures as an adjunct for increasing methodological robustness. An additional 44 papers discussed specific data analysis techniques. Of these, 22 dealt with statistical analysis techniques exclusively, 3 with visual analysis exclusively, and 11 with comparison of statistical and visual analysis, highlighting the strengths and weaknesses of each approach. Eight papers discussed methods for quantitative synthesis of data.
Conclusions: Although consensus exists about aspects of basic design and analysis of data for N-of-1 trials to ensure internal and external validity, issues related to generalisability of results and meta-analysis of data remain contentious.