Examining primary care physician rationale for not following geriatric choosing wisely recommendations

Eighty physicians above the median were emailed to participate. We interviewed initial respondents and then sent additional recruitment emails to non-responders and stopped recruiting when interviews stopped generating new themes. Fourteen physicians who were above the median for overuse (8 women, 6 men) participated in nineteen topic-specific interviews. Physicians all practiced primary care and were board-certified in either family practice, internal medicine or geriatrics. We identified 4 major themes (Table 1). These are presented herein and illustrated using representative quotes (Table 2).

We characterized primary care physicians’ perspectives for overusing tests and treatments that are not indicated in older adults, targeting physicians who were frequent high users of: PSA screening in men aged ≥ 76 years, urine testing in women aged ≥ 65 years without specific genitourinary or infectious signs or symptoms, and prescribing insulin or oral hypoglycemic medication in adults aged ≥ 75 years with diabetes and HbA1C < 7.0%. Although participants were selected for prior overuse, most were generally aware of guidelines and their recommendations. Knowledge deficits often are not the most significant driver of guideline non-adherence, and our data support this [11, 14]. Even among clinicians who deviate from guidelines frequently we find that they have an understanding of guidelines.

Deference to perceived patient preferences often influenced the decision to order testing and treatment, especially for PSA screening. Recent qualitative data suggests that in fact, many patients would be amendable to stopping cancer screening, but few recall actually discussing screening with a clinician [15]. We also found that physicians often assume a patient’s preference is to continue to be screened for prostate cancer if they had been screened in the past. Thus, physicians may not be communicating the risks of additional testing or treatment because they incorrectly assume patients want to continue their care as is. Physicians who over order may be making more assumptions about patient preferences or be less prone to challenge patient views.

Overestimation of benefits and underestimation of harms was a central theme throughout all three topics. For urinary testing, most clinicians expressed concern about missing a severe urinary tract infection without highlighting risks associated with urine testing such as overuse of antibiotics. Several studies have shown that ordering urine testing in asymptomatic women leads to unnecessary antibiotics [7, 16] and that PSA screening in asymptomatic older men can lead to negative consequences such as harms from biopsies, surgery or radiation treatment, as well as psychological harm [17]. There is also literature to support that clinicians want to feel like they have done everything for their patients [18]. Participating physicians seemed to under-weigh the risks of performing the action when discussing how they approach patient care, though acknowledged these risks when specifically asked general knowledge questions. Other studies have suggested that though clinicians have knowledge of clinical guidelines, they may have inadequate knowledge about harms of cancer screening [9, 15]. Thus, when approaching shared decision-making, they may underestimate and not fully communicate potential harms of some actions.

Finally, reluctance to change if a patient is doing well was often mentioned when asked about diabetes. Again, clinicians did not seem to recognize the harms associated with overtreatment citing the fact that if the patient was doing fine, they would not make adjustments, specifically because doing this would take time. This compares to literature citing concerns with inefficiency and inertia of previous practice [18, 19].

These data suggest that interventions to reduce overuse should not merely convey clinical recommendations but rather should target the psychological motivations that lead to overuse. Decision support tools that increase clinicians’ attention to potential downsides of an easily ordered test or a patient’s current treatment may improve care more than simple clinical guidance alone. Additionally, fostering clinician communication skills about potential harms may help reduce overutilization, especially for prostate cancer screening [20].

There are several limitations to this study. First, we included only primary care physicians in one health system, so our results may not be generalizable to other systems or types of clinicians. Second, we choose three specific overuse topics. It is unknown if results would be similar for others. Third, we did not compare differences between physicians with high and low levels of over ordering. Thus, it is unclear if there these groups have different attitudes towards overuse or if they have similar views but are applied differently. Fourth, participant variables such as years in practice or medical school graduation date were not collected. Finally, observations from physicians willing to participate in this study may not be generalizable to other physicians.

We found several reasons why primary care physicians may not follow guidelines to limit overuse of testing and treatment in older adults. Approaches to decrease unnecessary testing and treatment in older adults should account for perceived patient preferences, make the realistic harms of overtesting and overtreatment salient, and address clinical inertia among patients who are over treated but appear to be doing well.