A Psychological Approach to Helping Doctors Do Better at Doing No Harm

Behavioral nudges reduced doctors' overprescribing and overtesting of older patients

The medical community has been actively grappling for more than a decade with its own challenge of care overuse, the tendency among doctors to overprescribe medications and over-order tests that are outside of medical guidelines. 

Beyond the cost burden of care that is not medically necessary, overuse increases the risk of inadvertently inflicting harm. For instance, the unnecessary use of antibiotics puts patients at far greater risk of developing C. diff, a bacterial infection that is especially dangerous for older adults.  And then there’s the test that’s not medically called for that generates a false positive, and can send doctor and patient down an avoidable rabbit hole.

Since 2010, a multidisciplinary team of seven doctors and academics that includes UCLA Anderson’s Noah J. Goldstein and Craig Fox have published multiple studies documenting how behavioral nudges in field experiments reduced the overprescription of antibiotics. 

The Utility of Behavioral Prompts

Five members of that team, including Goldstein and Fox, joined with seven other researchers in a recent study that advances the potential utility of behavioral prompts in combating medical overuse. As reported in Annals of Internal Medicine, they found that behavioral nudges, paired with some educational information, were more effective in reducing the overuse of treatment for older adults than merely having the clinicians review educational material.

In a randomized controlled trial, the team first spent 12 months documenting the overuse of medical care prescribed by nearly 200 primary care and geriatric care clinicians across 30 medical practices in the Northwestern Medicine system. The trial focused on three specific issues pertinent to proper care of older patients.

• Prostate specific antigen testing for men 76 or older without a prior diagnosis of prostate cancer. Given longevity issues, testing at this life stage is considered unnecessary. 
• Urine testing for women 65 and older when there is no apparent symptom of a urinary tract infection. To the extent this practice leads to the unnecessary prescription of antibiotics, it can set off a cascade of harmful side effects.
• Continued prescription of diabetes medication (other than metformin) for patients over the age of 75 previously diagnosed with diabetes, but whose most recent testing was within the range of what is considered normal. Insulin and other diabetes medications are known to increase mortality among older adults.

All clinicians who had signed on to participate in the study were emailed a link to a short interactive tutorial that covered specific treatment guidelines for each of the three areas and reminders of risks.

Over 18 months, as clinicians used the health system’s electronic records data, half of the participants considering a treatment outside of guidelines received messaging designed to nudge them to reconsider. When an unwarranted PSA test was about to be prescribed, the pop-up pointed out that the medical norm was to not go this route for patients over the age of 75, playing to a physician’s social and reputational concerns. Clinicians considering a urine test for older women that weren’t exhibiting classic UTI symptoms were reminded what qualified as UTI red flags and the potential harm from unnecessary antibiotic treatment. For patients whose glycemic levels had fallen into the acceptable range for older adults, clinicians considering continued use of insulin and other oral medications were explicitly reminded that going this route was “associated with serious harm, including higher mortality.” 

Clinicians who chose to proceed with the PSA and urinalysis tests were prompted to write a brief justification of their prescription as an accountability nudge for the clinician to take a beat before proceeding.

Across these three different areas of overuse, the clinician group getting nudges reduced overuse more than the control group. 

A Safety Check for Patient Results

The researchers also did a safety check of sorts, with generally positive results. The nudge to reduce unwarranted urinalysis did no harm as patients spared the test didn’t have a higher probability of subsequently landing in the emergency room with a UTI or sepsis in the ensuing four weeks. Most diabetes patients whose test levels during the trial were in the normal range, and whose clinician didn’t continue prescribing insulin, also fared well, though less than 1% of those patients saw an uptick in their glycemic levels back into the too-high range.

While overuse persisted, the fact that the level of overtreatment fell among clinicians who were fed the extra messaging — without negative consequences — points to how behavioral interventions may be a useful tool in helping improve the quality of care for older adults.