Are We Misusing Beck's Depression Inventory?
A measure meant to be a screening tool is increasingly being used for diagnoses
The year was 1961. Psychiatry and psychotherapy were strongly under the influence of psychoanalysis and its teachings. Patients were assessed, diagnosed, and treated through psychoanalytic therapy.
But at the same time, an undercurrent of discontent was brewing. Clinical practitioners were tired of the slow and cumbersome process of psychoanalysis which, at times, took years to reach a breakthrough.
There were also concerns about the low reliability of psychiatric diagnoses among the practitioners. A person who had manic-depression as per one doctor had psychotic depression as per another, and a personality disturbance as per another.
There was very little in terms of objectivity when it came to psychiatric diagnoses and treatment. Clinical depression was no different.
An Inventory for Measuring Depression
It was in this context that Aaron Beck, a psychiatrist disillusioned with psychoanalysis, proposed Beck’s Depression Inventory. It was supposed to be an objective scale to measure depression in patients and populations.
Unlike psychoanalysis, which relied on unseen desires and drives, BDI used a patient’s behavior, thoughts, and emotions to assess them for depression. It might seem obvious to us now, but at the time, it was a breath of fresh air against the cumbersome psychoanalysis.
BDI contains 21 questions about a person’s mood, disturbances in thoughts, and behavior. The person answers these questions on a scale of 0-3. 0 being the lowest severity and 3 being the highest.
The inventory solved a lot of problems for clinical practitioners.
It solved the problem of variation in diagnosis and provided a stable measure of depression. It was easy to use and since it gave a numerical score at the end, it allowed one to assess the efficacy of therapy and other clinical interventions.
Naturally, BDI became super popular. It is popular to this day, although, the latest updated version is called BDI-II.
But there is a problem with how we use it. Beck recognized it 26 years after the original BDI was first released.
The majority of clinical practitioners and researchers have still not had that realization.
From Screening to Diagnosis
Ever since BDI was released, it has seeped deeper and deeper into clinical and research practices. Today, it is common to use BDI to assess the prevalence of depression in a given population.
This is how it usually goes:-
You find a sample you want to assess the presence of depression in and you give them the BDI to fill out. It usually takes 1-2 minutes per person. The responses are combined and you get a final score for each individual. This score is then classified into 1) no depression 2) mild depression 3) moderate and 4) severe depression.
You now know how prevalent depression is in the sample you tested.
Only, that is not how one should be using the BDI, as per Beck himself.
Multiple method assessment is strongly recommended before employing the label "depressed" and implying a nosologic category.
-Kendall, Hollon, Beck, Hammen, and Ingram (1987)
This came after researchers in the USA consistently administered BDI to their samples once and selected depressed and non-depressed individuals from that one score.
Deardorff and Funabiki (1985) previously suggested that the best way to understand depression in a sample is to administer the BDI twice with a gap of two weeks and conduct a diagnostic interview in between.
BDI was never meant to diagnose people with depression by itself. It was meant to help clinicians understand the changes in their clients over time. Beck recommended using ‘dysphoric’ over “depressed” for those with high scores on the BDI.
I believe this can be generalized to most self-report inventories. The availability of these inventories online, for free, has meant that people have been increasingly diagnosing themselves with disorders without understanding the caveats of these tools.
Complexities of People
To defend clinicians and researchers, there has been an increasing interest in accessing mental healthcare services and conducting studies on the same. In such a situation, they are expected to assess a large number of people in a limited time.
The BDI offers a possible way out but we forget the limitations of the same.
In an ideal world, mental disorders would be just as straightforward as the items on a checklist. They would have the same manifestations in everyone and they would be truly absent in people who score 0 on the BDI.
Sadly, we don’t live in an ideal world. People, and the disorders they have, are a lot more complex. We are still trying to understand how mental disorders show themselves. Relying too strongly on one scale for such a complex phenomenon might be a mistake in the long run.
And that is it for the week! I had a different topic planned for this week but a conversation on LinkedIn with someone who is not familiar with the field changed that quickly. I find it surprising how often MHPs use tools that are meant to help them diagnose as the sole diagnostic criteria. Is there something incomplete in how our MHPs are being trained or is it the lack of time that is making them use such shortcuts?
What do you think?
You can let me know by answering this email :)
Until next time,
Arjun
Hi Arjun,
Beautifully written article and a breath of fresh air with all the psychobabble floating around on the internet. I would also like to add some more detail here.
While you have pointed out the strong influence of Psychoanalysis on Mental health treatment in the 1960s, rating scales were being used to assess patients before BDI as well. Ham-D(developed in the late 1950s) has been the most widely used clinician administered rating scale, specially for assessing the effectiveness of anti-depressants. Initially BDI relied on theory of Cognitive distortions unlike HAM-D which was developed mainly using widely reported symptoms of depression at the time. The Ham-D was quite influential in pushing clinicians in looking towards symptoms-their pattern and severity.
Both BDI and HAM-D represent different perspectives of depression in a different era. So while objectivity still hadn't entered the picture in the strictest sense of the word, there was still more structure than the varied diagnoses being used prior.
Contrasting the use of BDI with the subjectivity of psychoanalytic diagnoses may therefore paint a slightly distorted picture. Rating scales and their purported objectivity have been long questioned by both statisticians and traditional psychoanalysts with varying reasons. Today even Psychoanalysts have embraced empiricism and it is sad that a lot of professionals in the field themselves don't try to understand the difference among assessment, measurement and diagnosis. In that aspect, your question towards the training of MHPs is pertinent. A stronger emphasis towards conceptual clarity in statistics for MHPs might foster a nuanced understanding of using rating scales in treatment and research( and consequently more cautious, I suppose).
Again, much appreciation for your efforts in writing this blog.
Best,
Aditi
My ug research was done using BDI. Although the data collection had wildly unfair means iykuk. Even i always thought self repport scales are not for diagnosis or even mental state assessment that much as we are complex. It is also fairly easy to be dishonest and deceptive in a self report scale.