Figure. Framework for de-implementation in cancer care delivery. Adapted with permission from Norton WE et al. J Clin Oncol. 2019;37(2):93-96.⁷

Overuse is common in health care. Defined by the Institute of Medicine in 1998 as “a health care service [that] is provided under circumstances in which its potential for harm exceeds the possible benefit,”¹ overuse has been estimated to cost the US health care system over $100 billion annually, distracting funds from higher value, evidence-based care delivery.² The extent of unnecessary tests and procedures drove the Choosing Wisely campaign’s efforts in partnership with over 80 societies to address over 700 tests and treatments that were unnecessary or overused.³

There are many reasons we overuse services ranging from diagnostic and prognostic uncertainty, outdated practice styles, and insufficient evidence, to fear of missing a diagnosis, defensive medicine (ie, avoiding malpractice), patient preference, and even revenue generation in fee-for-service systems.⁴ Indeed, indication creep, where evidence from clinical trials drives interventions and services into broader, untested populations typically with less severe disease, is also at play.⁵ From off-label drug use to excessive imaging, indication creep can foster overuse, limiting benefits to those with milder disease and exposing patients not included in trials to harms.

Addressing overuse is complicated. While Choosing Wisely popularized overuse awareness and in partnership with the AUA defined 15 mostly don’t do items (see Table), how best to accomplish not doing these services was missing.³ When “less is more,” doing less can be challenging, even for those with the best of intentions. Moreover, some things are easier to do less often, while others are more challenging, especially when we are used to doing them. For example, using bone scans to stage low-risk prostate cancer used to be widespread, even after it was deemed overuse. While some providers were able to stop ordering, others needed more time and motivation, creating opportunities to test strategies to help do less.⁶

Table. AUA Choosing Wisely Guidance Addressing Overuse

	Recommendation	Discussion
1	A routine bone scan is unnecessary in men with very low-risk or low-risk prostate cancer.	Very low-risk or low-risk patients (defined by using commonly accepted categories, such as AUA guidelines) are unlikely to have disease identified by bone scan.
2	Don’t prescribe testosterone to men with erectile dysfunction who have normal testosterone levels.	While testosterone treatment is shown to increase sexual interest, there appears to be no significant influence on erectile function, at least not in men with normal testosterone levels.
3	Don’t order creatinine or upper-tract imaging for patients with BPH.	When an initial evaluation shows only the presence of LUTS, if the symptoms are not significantly bothersome to the patient or if the patient doesn’t desire treatment, no further evaluation is recommended.
4	Don’t treat an elevated PSA with antibiotics for patients not experiencing other symptoms.	It had previously been suggested that a course of antibiotics might lead to a decrease in an initially raised PSA and reduce the need for prostate biopsy; however, there is a lack of clinical studies to show that antibiotics actually decrease PSA levels.
5	Don’t routinely perform ultrasound on boys with cryptorchidism.	Ultrasound has been found to have poor diagnostic performance in the localization of testes that cannot be felt through physical examination.
6	Don’t prescribe antimicrobials to patients using indwelling or intermittent catheterization of the bladder unless there are signs and symptoms of urinary tract infection.	Antibiotics in the absence of signs and symptoms (which may include fever; altered mental status or malaise with no other cause; floor pelvic pain; flank or suprapubic tenderness; hematuria; dysuria; urinary urgency or frequency; and, in spinal cord injury patients, increased spasticity, autonomic dysreflexia, or sense of unease) are not efficacious and risk inducing resistance to antimicrobials.
7	Don’t obtain CT scan of the pelvis for asymptomatic men with low-risk or very low-risk clinically localized prostate cancer.	CT scan of the pelvis is very unlikely to provide actionable information in men with low-risk prostate cancer (one commonly accepted definition of low-risk prostate cancer is Gleason score <7, PSA <10.0 ng/mL, tumor stage of T2 or less, very low-risk is Gleason score <7, PSA <10 ng/ml, tumor stage T1-T2a, less than 34% of biopsy cores positive, no core with more than 50% involved, and PSA density of <.15 ng/mL/cc).
8	Don’t remove synthetic vaginal mesh in asymptomatic patients.	There is no clear benefit to mesh removal in the absence of symptoms, and mesh removal in this circumstance exposes the patient to potential complications such as bladder injury, rectal injury, and fistula formation.
9	Offer PSA screening for detecting prostate cancer only after engaging in shared decision-making.	Shared decision-making (between health care provider and patient and, in some cases, family members) is an excellent strategy for making health care decisions when there is more than 1 medically reasonable option.
10	Don’t diagnose microhematuria solely on the results of a urine dipstick (macroscopic urinalysis).	Microhematuria is defined only on urine microscopy: 3 or more red blood cells per high-powered field on microscopy of a properly collected urinary specimen.
11	Don’t treat low-risk clinically localized prostate cancer (eg, Gleason score is <7, PSA <10.0 ng/mL, and tumor stage T2 or less) without discussing active surveillance as part of the shared decision-making process.	The ultimate choice of treatment should be based on shared decision-making and individualized to the patient’s disease characteristics, his overall health, and his personal preferences.
12	Don’t treat uncomplicated cystitis in women with fluoroquinolones if other oral antibiotic treatment options exist.	Due to serious potential side effects associated with the use of fluoroquinolone antibiotics, these drugs should not be prescribed as first-line therapy for uncomplicated cystitis in women.
13	Don’t continue opioid analgesia beyond the immediate postoperative period; prescribe the lowest effective dose and number of doses required to address the expected pain.	The use of opioid analgesia for pain is often appropriate in surgical patient care.
14	Don’t obtain urine cytology or urine markers as a part of the routine evaluation of the asymptomatic patient with microhematuria.	Insufficient evidence exists for the use of urine cytology and urine markers in the routine evaluation of the asymptomatic patient with microhematuria, including BTA assays, NMP assays, and FISH assays to detect chromosomal alterations.
15	Don’t routinely use CT to screen pediatric patients with suspected nephrolithiasis.	Given the link between radiation exposure from CT in children and increased cancer risk, imaging test selection should adhere to the principle of ALARA to minimize radiation exposure.
Abbreviations: ALARA, as low as reasonably achievable; BPH, benign prostatic hyperplasia; BTA, bladder tumor antigen; CT, computed tomography; FISH, fluorescence in situ hybridization; LUTS, lower urinary tract symptoms; NMP, nuclear matrix protein; PSA, prostate-specific antigen.

Developing and testing strategies and interventions to address overuse, and in effect “do less,” is central to the science and practice of de-implementation. Also termed de-adoption and de-intensification, de-implementation can be considered a counterbalance to overuse. Illustrated in the Figure is a systematic framework for de-implementation in cancer care delivery, though it can be extrapolated to overuse in general.⁷ Starting with the strength of evidence for a given overuse intervention (ie, no longer considered effective, for example, routine cytology in asymptomatic microhematuria) the magnitude of the overuse problem can be characterized, barriers and facilitators to reducing/replacing/removing/restricting the practice can be explored, and multilevel strategies (eg, provider, clinic) can be developed and tested. Strategies might include best practice alerts, audit and feedback, pre-authorization, including formulary restriction, shared decision-making, and academic detailing to audit overuse practices, and guide improvement strategies.

However, compared to overuse literature, the de-implementation literature and evidence-base is only emerging. How best to “do less” needs further study, including through rigorous clinical trials. For example, we are investigating doing less ADT overuse in monotherapy in localized prostate cancer and biochemically recurrent, nonmetastatic disease, comparing behavioral theory-informed de-implementation strategies in a cluster randomized trial.⁸ We need others to look across and beyond AUA’s Choosing Wisely recommendations to help our field build the evidence base for doing less, yet doing more for our patients.

Conflict of Interest Disclosures: Dr Skolarus is supported by NCI R37 CA222885.