One of the big frontiers right now in behavioral health is measurement and quality. One of the first things people ask before they send a loved one to $40,000 per month residential treatment is what the chances are of them remaining abstinent in the long term. Residential programs used to quote success rates in the 90% range, which is dubious at best. In fact, there is still one large national residential chain that trumpets on their website, “Between 85 percent and 95 percent of all people who successfully complete drug rehab report still being abstinent from all drugs nine months after discharge.”
Most people in the field would seriously question these numbers, and long-term studies have shown that treatment success rates vary based on the evidence of the treatment. Strong evidence-based treatments, like medications for opioid use disorder, are far more likely to reduce substance use than non-evidence-based treatments like equine therapy.
But even at the high end, success rates above 60% are difficult to believe. The National Institute on Drug Abuse estimates 40-60% of people relapse after substance use treatment, a rate slightly worse than Type I diabetes but better than high blood pressure or asthma.
These numbers reference abstinence and relapse. But even if we agreed that simple abstinence was the primary goal, what substances are of interest? Only the person’s primary substance of choice, or all substances? What tests would we use, and furthermore, how would we engage someone in testing if they have moved beyond the clinical setting into community recovery? And how would we deal with inconclusive or false positive tests?
Perhaps we need a different frame of reference. Most people would agree that there are many areas of greater interest than strict abstinence. Reconnecting with loved ones, gainful employment, stable housing, and avoiding criminal justice involvement are all highly worthy goals from the individual and societal points of view. These outcomes, however, also suffer from critical flaws. First of all, they’re nonbinary and in some cases noncategorical; imagine trying to classify the quality of one’s relationships with loved ones in a meaningful, consistent fashion! Second, while some of these outcomes are measurable, they’re not measured through the usual frame of reference in health care: the electronic health record. The only way to measure these with some validity is through an interview process, which systematically biases the responses toward people who remain engaged in their recovery.
From the payer’s standpoint, ongoing physical and mental wellness are of primary interest. While it is rare for a health plan and a plan member to align their interests, this seems to be a major opportunity. Both plans and their members like to avoid ED visits and hospitalizations. ED visits come at a heavy financial cost for the plan, and they’re painful and disruptive for the member.
Thus, we need to measure behavioral health outcomes that are:
- Not strictly based on a binary model of abstinence or relapse; outcomes that take into account physical health and other outcomes of interest.
- Captured in the EHR
- Repeatable over time, including well after the acute treatment episode has completed.
A more portable, relevant way to measure behavioral health outcomes will benefit both payers and members, leading to higher quality behavioral health treatment.