Measurement Based Care

​Measurement based care is generally perceived to improve the outcomes of various psychiatric conditions by routinely bringing the same assessments to everyday treatment world the as those used in the clinical trials of the interventions being prescribed. When scientists talk of studies of clinical interventions the first two terms you need to know are independent variables and dependent variables. An independent variable is the intervention; it does not depend on any other action other than being given to the subject (patient). The dependent variable is what is measured and hypothesized by the investigator to be a result of the intervention (independent variable). For example, if a patient is given Prozac (fluoxetine) then the investigator hypothesizes there should be a reduction in depression. This begs the question; how do we know the depression has been reduced? Better yet, how much depression was present at the beginning of treatment and how much depression is remaining after a period of treatment? A carpenter would not cut a board without first measuring the board to see if the board were long enough to yield his desired length and then measure at least twice to cut his desired length. Yet, everyday physicians and clinicians of all stripes that treat mental illness make metaphorical cut after cut without knowing much of anything except the name of the metaphorical saw he or she is using. In the United States only about 18% of psychiatrists and 11% of psychologists routinely administer rating scales to their patients. The clinicians that use clinical judgement alone detect only 21% of their patients that have deterioration and who have increased symptom severity. The clinicians detect even fewer patients who are not deteriorating but are simply not improving. When the clinician fails to detect a deterioration or even more commonly a failure to respond optimally to treatment, then the patient continues to be ill, or at least more ill than necessary, due to lack of adjustment in the treatment plan. In addition to the lack optimal patient outcomes there is also a lack of optimal clinician outcomes. Clinicians that do not observe the systematic results of their interventions have no way of knowing that the interventions were not successful and to what level they were not successful. The clinician does not get the feedback that allows him or her to adjust treatment strategies and grow in clinical knowledge and skill. Measurement based care (mbc) has been defined as “enhanced precision and consistency in disease assessment, tracking, and treatment to achieve optimal outcomes”. Symptom rating scales are the starting point for the evaluation of clinical effectiveness. For the mbc to be effective the symptom feedback must be clinically actionable. Perhaps we should coin a phrase here: “you don’t need to measure what you can’t treat, and you can’t treat what you can’t measure”. For the rating scale to be useful (valid) it must be a current assessment, it must be interpretable, and it must be readily available during the clinical encounter. Screening tools are now mandated in many electronic health records as part of the Byzantine Federal regulations promulgated upon the medical community. Most commonly, the PHQ-9 is given at a yearly screening exam. This is a 9-symptom rating scale of depressive symptoms. It is not a diagnosis, that requires clinical judgement, in addition to point-in-time data. Screening alone is insufficient to improve outcomes without system in place to monitor treatment response. Mbc requires that symptom severity must be assessed frequently to be effective. This assessment must be done in concurrently with the clinical encounter. For example, if the rating scale is administered quarterly in the first month of the quarter but the patient is seen quarterly in the third month of the quarter then the rating scale is useless. (I am sure it is only a matter of time before some insurance company or managed care company comes upon this bright idea to institute this useless strategy under the heading of quality improvement). I have been practicing management-based care in my clinic for the last 30 years. My goal from the outset was to be able to demonstrate the effectiveness of the treatments rendered. Of course, there are numerous variables other than the independent variable of the medication prescribed that may impact the treatment outcome (in statistics, we call these confounds). These confounds only apply to comparisons between patients, within the same patient the confounds remain the same and the comparison of the same rating scale outcomes over time and interventions is both reliable and valid. There are many rating scales available for clinical use and are simple and easy to administer or have the patient self-administer. I will review these, perhaps, in the next blog. At my office we use the SCL-90 for the diagnostic screening, although I am thinking of changing to the Mini-Neuropsychiatric Assessment, and the Basis-32 and a Review of Systems (ROS) both at baseline and with each follow-up visit. The Basis-32/ROS covers all the symptoms on the MADRS, PHQ-9, and GAD-7, and most of the symptoms on the Ham-D. In addition, it has a robust section for social functioning assessment. If the need arises for further clarification, I will complete a MADRS, AIMS, PANSS or Columbia Suicide Scale. I hope this has given you some background information of why you and I do what we do at your clinical visit. References available upon request. The author does not wish to represent the findings of others as his.