Lecture Notes for Session #3: 5/31/12

Misty McIntyre Goodsell will Special Guest Lecture!

Problem Formulation

The goal for today's class is to become more familiar with the first step in the Evidence-Based Practice model:  asking well-built questions.  We'll learn to identify different types of questions and to understand what types of research provide the best kind of evidence for each.  Then, we will get some beginning practice in formulating these questions.

Questions about practice tend to come in three different types.  These three different types of questions are each answered best with very different kinds of research studies!  The type of question on which we are focusing most in this class have to do with effectiveness.  How well does a given program, intervention procedure, or policy work with what types of clients with what types of problems in what context?  There are two sub-types of effectiveness questions; those that have to do with solving problems (intervention effectiveness), which is what most social workers do most of the time, and prevention effectiveness questions, which I wish we could have more focus on, that is, what types of programs or policies work best to prevent certain types of problems from arising to begin with?  Both types of effectiveness questions are best answered with randomized studies, in which one group of individuals who have been exposed to the intervention to be tested are compared to those who have not.

A second type of question that is useful for practice has to do with assessment instrumentation.  That is, how reliable (consistent) and how valid (accurate) are particular instruments, tests, observations, or interview protocols for finding information critical to practice decision making. Again, there are two sub-types of assessment questions:  the first has to do with detection of current states or problems, and the second with prediction of future states or problems.  For example, we may want to know if a client is depressed, which is an example of current assessment, or we may want to know if a child is likely to be harmed if s/he continues to live in the home, which is an example of prediction, or risk assessment.  Both of these subtypes are best answered with studies which compare the assessment strategy against a criterion measure or a different assessment tool with the same group of people.

A final type of question has to do with background or description information about a given problem/population combination.  For example, we might want to know what types of difficulties kids aging out of the foster care system face as they try to establish themselves as independent young adults.  These questions tend to be of two types:  Quantitative/Descriptive questions tend to look at statistical associations between variables or at incidence/prevalence of occurrence associated with subgroups of the population, while Qualitative/Descriptive questions look for rich descriptions of a smaller group of people which can help develop deep theoretical understandings of issues.  Quantitative questions are best answered with detailed surveys of large, representative groups of individuals, while Qualitative questions tend, of practical necessity, to be developed with more intensive observations of smaller groups of individuals. 

Questions which are useful for evidence-informed practice need to be constructed in a way that makes them immediately useful for searching for research which answers those questions.  The anatomy of these questions makes them a little different from the way in which we normally ask questions in conversation because we want them to be sufficiently specific that they facilitate systematic information retrieval.  An acronym for these well-built questions is PICO, which as you know from your reading stands for Patient/Problem, Intervention, Comparison, and Outcome. Sometimes people add an S to the end of PICO, standing for Study Type, and others add another C to stand for Context.  We will handle these two issues in a different way.

The Late, Great, Len Gibbs came up with a different acronym for these questions, and called them "COPES," for Client-Oriented Practical Evidence Search, and this term is used quite widely in social work.  However, other helping professions have learned these as PICO questions, and I think it is important, from an inter-professional standpoint, to use the same language. Regardless, in the link to the very helpful training resource from Columbia University (BEST Training), you will notice that they use the COPES term, and if you look carefully at many of my course materials, you will see reference to COPES, although I am trying to find these references and fix them! J

There is also a link to an Evidence-Based Medicine training series, using the PICO language, which is quite good, but really medically-oriented. Hopefully, between the reading, practice in class, and the links, you will provide guidance on your first step to Evidence-Based Practice.

The discussion postings for last week and this week are designed to help orient you to asking questions in a way that will make you more of an evidence-based practitioner and which will give you practice for the first assignment.

Have a great class!