Evidence Based Practice Week 7 - Alternative Research

Introduction

• Single-subject and qualitative research are useful for supporting clinical decision-making, though not as common as other research types in physical therapy literature.

Single Subject Research

• Single-subject design (SSD) involves intensely following one participant.
• Variables of interest are repeatedly measured during a baseline period before the intervention begins to establish typical values for each outcome measure.
• Data collection continues periodically during the intervention period.
• An additional measurement period may occur after intervention without treatment.
• Changes in variables during treatment are compared to baseline and post-treatment periods.
• SSD generates abundant data from one participant, but can also use and analyze many participants separately.
• Case studies differ from SSDs by being retrospective and detailing the characteristics of a single case and its intervention but lack controlled manipulation or experimental controls.
• Case studies lack systematic control, implementation, and evaluation of treatment.
• Randomized Controlled Trials (RCTs) can be implemented using SSDs by grouping multiple single-subject studies with similar participant characteristics.
• In an RCT with single subjects, one subject is randomly assigned across subjects, similar to a group design.
• A crossover design can be used with a group of single-subject studies in which a participant is randomly assigned to treatment A followed by treatment B, with measurements taken after each treatment.

Determining the Applicability of SSR

• Applicability depends on the relevance of the participant characteristics to your own patient.
• High inter-subject variability in large samples can reduce the likelihood of finding significant differences between interventions.
• In SSD, subjects act as their own controls; baseline and intervention periods are compared to detect treatment effects.
• The trade-off between SSDs and group designs is that single participant treatment responses are not as powerful as a group's responses.

Determining the Quality of SSR

• Measurement must be valid and reliable, controlling for other factors that might affect outcomes during the study.
• Control of experimental procedures is a characteristic of all quality research, regardless of design.

Measurement

• SSR measures are taken repeatedly on a single participant over several days or weeks before intervention.
• This allows investigators to measure natural fluctuations in the outcomes of interest.
• An important consideration for SSR is the possible impact of multiple measures on the outcome of interest; repeatedly practicing a measure may alter a participant’s performance.
• Blinding of the evaluators is often challenging in an SSD, but it is critical to the rigor of the study.
• A therapist from another clinic can act as an independent evaluator, when outcomes are measured.

Intervention

• In SSD, intervention is specific to the participant.
• Participant characteristics and the specific intervention are typically more fully described, often making it easier to determine if the study outcomes are appropriate for your patient and if the intervention is feasible in your clinic.
• Alternative group designs create individual treatment for participants in the group and systematically measure the progress of each participant.
• These studies are costlier and typically take longer to complete, but the do address the issue that no one treatment fits all.
• SSD should be rigorously controlled including monitoring the patient for all possible contributions to outcomes during the study period.
• An SSD research article should report the relevant demographic and clinical characteristics of the participant as well as the results of the outcomes measured in the study and often reported in tables.
• Analysis of the results of a single-subject study should include graphs and statistics.
• Both visual inspection and statistical analysis are essential in the analysis of the data from single-subject studies.
• Visual analysis is necessary, but not sufficient for analysis of SSR and studies should report from both types of analyses.

Visual Inspection of Results

• The celeration line quantifies trends as a “best-fit” line through the data, assuming the data fits a linear function.
• Celeration lines assists in describing the increasing and decreasing trends in the data.
• It is created by computing the median value for each phase of the study beginning with the baseline phase.
• Slope of the celeration line conveys the amount and rate of increase or decrease in the data.
• The slope value represents the rate of change during the baseline week, before treatment begins.

Statistical Analysis of Results

• The "significance" of the results is expressed with terms other than p-values, which are the common expression of statistics for group designs.
• Data from SSD studies should be processed with regard to the common features of these data sets, including dependency of data and the possibilities of extreme values in baseline data.

Detrending Data and Serial Dependency

• Because of the natural trend that would be expected throughout the study, it must be removed or analyzed within the subsequent treatment phase which can be achieved through various methods or by fitting trend lines to specific phases of the data set.
• Repeated measures create dependency in the data, termed serial dependency where data points from the same person are positively correlated, also termed autocorrelation.
• If autocorrelation values remain significant, then further statistical analysis will be problematic.

Two-Standard-Deviation Band

• A statistical test used in combination with the celeration line with the data normally distributed and without significant autocorrelation coefficient.
• The mean and standard deviation for the baseline data are computed, and parallel lines representing +2SD and –2SD are drawn through the data graphs of the baseline and treatment phases.
• By convention, if two data points are above or below the 2SD bands in the treatment phase, the change is considered statistically significant.

Qualitative Research

• Focuses on questions of experience, culture, and social/emotional health.
• Study designs facilitate understanding of processes experienced by individuals, groups, or cultures.
• Allows researchers to answer important questions that cannot be answered by quantitative designs.
• One purpose is to generate hypotheses.

Appraising Qualitative Research

• 5 general designs used:
• Ethnology.
• Phenomenology.
• Grounded theory.
• Biography.
• Case study.

Phenomenology

• Useful to study experiences in life.
• Often possible to generate insights from personal perspective into themes that may not be revealed with quantitative research or through clinical treatment.

Ethnology

• Typically used to study questions in anthropology.
• Questions of group behavior, including daily life patterns, attitudes, and beliefs during rehabilitation, could be studied with this design.

Grounded Theory

• Used to construct or validate a theory.
• Researchers focus on core processes in a given life experience and develop a theoretical model to represent the processes.
• Models can generate a hypothesis that could be tested with a quantitative design.
• Qualitative research includes an in-depth study of a specific topic, typically with a small and select group of participants.
• Random selection of participants is not a part of qualitative study sampling, rather, participants are selected specifically because they can contribute to the understanding of a topic (purposive sampling).

“Snowball” Sampling

• The researcher identifies a few participants and then asks their families if they can refer other potential participants or families.

Triangulation

• Is the use of different perspectives to study the identified process.
• Member checking verifies that the investigator has adequately represented the participant’s contributions to the question under study in which, the investigator may share preliminary drafts of the interpretation of the participant’s interview and then revise the draft with further comment from the participant.
• Triangulation in collecting the data is then used in the analysis process to determine the cohesiveness of the perspectives.
• The results of a qualitative study should be reviewed in light of the overall purpose of the study.

Explore single-subject research designs (SSRD) in physical therapy, contrasting them with case studies. Understand how SSRD uses repeated measurements during baseline, intervention, and post-treatment phases to assess treatment effects, providing in-depth data for clinical decisions.