It’s a number we see repeatedly in the popular press. In case you missed the reference, 10.8 is the percent of psychiatrists who saw all of their patients for psychotherapy according to a study published by Mojtabai and Olfson in the Archives of General Psychiatry in August 2008. It represents a decrease from 10 years ago and it is said to prove that psychiatry has come to be about nothing more than prescribing medicines.
I’ve been wondering about this statistic since I first saw it – does it really mean that psychiatrists don’t do psychotherapy anymore? Is the number of psychiatrists who see all of their patients for psychotherapy even relevant to the question of whether psychiatrists practice psychotherapy? I wanted to understand this particular statistic – the press hurls it around to discount the work that we do – and I found that I couldn’t understand the details of the original study in a way that made sense to me.
I decided to ask Dr. Ramin Mojtabai if he would explain the study to me and we met for lunch. He is an animated and energetic man, and conversation came easily. He listened while I explained my fascination with this single statistic and he talked openly about his research.
Using information Ramin gave me, along with statistics from the Centers for Disease Control’s website, I now have a sense of how “10.8 percent” was derived.
The National Medical Ambulatory Care Survey queries randomly selected office-based physicians every year. The physicians, or their office staff, are asked to submit data for patient visits during a one-week period of the year. According to the NMACS website, each year 3,000 physicians submit information on approximately 30 patient visits. For 2007, the number of physicians who actually participated was about half this, or just over 1,500 and the patient visits were closer to an average of 25 per practice.
Specially trained interviewers visit the physicians prior to their participation in the survey in order to provide them with survey materials and instruct them on how to complete the forms. Data collection from the physician, rather than from the patient, provides an analytic base that expands information on ambulatory care collected through other NCHS surveys. Each physician is randomly assigned to a 1-week reporting period. During this period, data for a systematic random sample of visits are recorded by the physician or office staff on an encounter form provided for that purpose. Data are obtained on patients’ symptoms, physicians’ diagnoses, and medications ordered or provided.
If you’d like to see the forms that are used, you can download them at http://www.cdc.gov/nchs/ahcd/ahcd survey instruments.htm.
For this study, Mojtabai and Olfson took the data that had already been gathered for psychiatrists. For each year, an average of 75.6 psychiatrists were surveyed, for a total of 756 psychiatrists for the years 1996-2005. All visits were not included, as the survey uses “a systematic random sample of visits,” and it is not possible to know if psychiatrists submitted data sheets for every patient contact. The study included data from an average of 19 visits per psychiatrist. Data was collected by visit, so theoretically, a psychoanalyst who sees the same patient five times a week would be submitting the same data for a few patients, while a psychiatrist with a high volume practice would only have a fraction of his interactions included, and psychotherapy visits might be missed.
To get to the statistic of 10.8%, the researchers looked at the responses to two questions on the form. In the area for “Non-Medication Treatment” there was an option to check “psychotherapy” under “List all ordered or provided at this visit.” (The bolding is on the form, it is not my intention to add emphasis). The second item they looked at was “Time spent with Provider” to be listed in minutes.
Mojtabai and Olfson did a simple cut to define “psychotherapy” for the purpose of this study. If the box for psychotherapy was checked and the time spent with the physician was more than 30 minutes, it was deemed to be psychotherapy. The authors were attempting to capture a traditional psychotherapy session. If a psychiatrist’s data sheets all included psychotherapy as a checked-off treatment, and all lasted over 30 minutes, then allof his patients were seen for psychotherapy. This percentage has dropped from 19.1% in 1996 to 10.8% in 2005. If you’re wondering, as I was, the percentage of surveyed psychiatrists who see some of their patients for psychotherapy sessions lasting over 30 minutes is 59.4%, which represented a non-significant decrease from 10 years earlier.