0:05
Today, I am going to introduce you to a Mixed-Method Approach,
and discuss its importance to Healthcare Quality Improvement.
This entire course is focused on the use of data in Healthcare Quality Improvement.
But when we talk about data,
what is usually our buying is numbers or are quantitative data.
For example, the mortality rate of a hospital,
the rate of healthcare associated infections,
the usage of laboratory test,
or the score of patients satisfaction.
This is because most of us are trained in positivists or post-positivists tradition.
What is usually missing is
the qualitative data such as the transcripts of interviews with clinicians,
the diaries we kept throughout the process of implementation of the intervention,
audio or video recordings of a patient-clinician encounters,
or notes and pictures we took during observations of a clinical process.
A mixed method actually recognized that
both quantitative and qualitative data has their values
and limitations and that combining these two types of
data can greatly improve the breadths and depths of our findings.
So when we say we use a mixed-method approach we usually talk about one study
or multiple phases of a project and was in this scope we sorely collect,
analyze, and integrate both quantitative and qualitative data.
The key of a mixed-method approach is of course
the mixing of quantitative and qualitative data,
and the mixing can occur at different stages of a project or study.
First, the mixing can occur at the design stage and we
will review the different mixed-methods designs in the following slides.
Second, the mixing can occur at the data collection stage.
For example, we can collect the one type of
data and use the data to inform the collection of other types of data,
or we can use a single data collection tool
to collect both quantitative and qualitative data.
Third, the mixing can occur at the data analysis phase for example,
we can quantitatively analyze the qualitative data we collect from interview.
Finally, the mixing can occur at the data interpretation stage which means we
draw conclusions by comparing or
integrating the quantitative and qualitative data we collected.
Okay so here are the different mix and methods design.
So the first one is convergent parallel design.
With this design, we concurrently collect
both quantitative and qualitative data and makes the results during interpretation.
So, one example will be,
in many of our studies we try to evaluate
the usability of medical device or electronic tool.
And in these projects we usually collect both quantitative data on
user performance and qualitative data
on user feedback to identify those usability issues.
In addition to the convergent parallel design,
we can also use the qualitative and quantitative data in a sequential manner.
There are two types of sequential design.
One is sequential explanatory design.
With this design, we collect and analyze the quantitative data first
followed by the collection and analysis of the qualitative data.
For example, in one of our projects,
we try to understand that use of black cultures in the diagnosis of sepsis patients.
So we first collect clinical data on the use of
black cultures and quantitative data show overuse or black cultures in practices.
Then we conduct qualitative interview with clinicians to
understand why and when they order black culture to explain the quantitative data.
Another sequential design is sequential exploratory design.
With this design, we collect and analyze
qualitative data followed by the collection and analysis of quantitative data.
For example, in one of our projects,
we conduct qualitative interviews with a small group of environment care associate
who perform patient room Cloney in the hospital to understand challenges to their work.
Based on the interview data we collect,
we develop a survey and conduct survey with a larger group of
EVC associates to prioritize areas for improvement.
The last mixed-method design is called embedded design.
So with this design,
we collect additional qualitative data within
the primary quantitative design or collect
additional quantitative data was in the primary qualitative design.
The embedded design is similar to the convergent parallel design.
The differences is the weight given to the quantitative and qualitative data.
While the embedded design emphasize one types of data over the other types of data,
the convergent parallel design put equal weight on both types of data.
This is a diagram from a systematic review of
mixed-methods research in human factors in healthcare.
From this diagram, we can see the importance of mixed-method approach has been
increasingly recognized and more and more studies have applied a mix method approach.
Now why a mixed-method approach is important to quality improvement?
This is because throughout the process of
quality improvement we need to answer a number of questions.
Some of these questions need to be answered by quantitative data.
Some of these questions need to be answered by qualitative data and some of
these questions needs to be answered by both quantitative and qualitative data.
We will review these questions in details in the following video.
Matt AustinAssistant Professor
