Physicians perceive 17% of primary care clinic patients as difficult, according to a systematic review and meta-analysis published in the Annals of Internal Medicine. In the study, which pooled data from 45 studies across 10 countries, investigators found that patients with depression, anxiety, personality disorders, and chronic pain were more likely to be labeled "difficult" by their physicians—and that less experienced providers rated more encounters as challenging.
The 17% prevalence estimate emerged from 10 studies using the Difficult Doctor-Patient Relationship Questionnaire (DDPRQ-10) among primary care clinic patients. A higher prevalence was observed among patients with medically unexplained symptoms (33%) and those with chronic pain receiving opioids (41%).
Lead study author Jeffrey L. Jackson, MD, MPH, of the Clement J. Zablocki Veterans' Administration Medical Center and the Medical College of Wisconsin, and colleagues identified multiple patient characteristics associated with being perceived as difficult. Patients with depression, those with anxiety, and those with chronic pain were about twice as likely to be rated as difficult, and patients with personality disorders were more than twice as likely to be perceived as difficult.
Female patients were initially found to be slightly more likely to be perceived as difficult compared with male patients; however, when adjusting for depression, the investigators found that the association disappeared, suggesting that the apparent sex difference reflected higher depression prevalence among female patients rather than an independent effect.
Provider experience emerged as a significant factor. Physicians perceiving patient encounters as difficult had approximately 3.5 fewer years of experience. Residents were more likely to rate patients as difficult compared with attending physicians who had completed training. One study found that providers with greater empathy were less likely to rate patients as difficult.
In addition, patients emerging from difficult encounters had measurably worse outcomes. They were nearly twice as likely to report unmet expectations following the visit and lower satisfaction with their provider.
The investigators searched MEDLINE, Web of Science, SciELO, ProQuest, Theses, Scopus, PsychInfo, Cochrane Central Register of Controlled Trials, Global Index Medicus, and EMBASE from inception through July 7, 2025. Among the 2,835 articles identified, plus 28 from citation review, 43 unique articles presenting results from 45 studies met the inclusion criteria. Most studies (72%) were conducted in primary care settings, with the US contributing 24 studies. Twenty-eight studies used validated instruments to assess difficulty, predominantly the DDPRQ-10.
Study quality varied substantially. Just 47% used random or consecutive sampling, 35% performed sample size calculations, and 14% adequately described their analytic approach. The investigators found no evidence of publication bias for patient characteristics or outcomes. However, outcome reporting bias affected several secondary analyses; when adjusted for this bias, the association between female sex and difficulty became insignificant, and the effect of provider experience was eliminated.
The investigators noted that prevalence estimates using methods other than the DDPRQ-10 ranged widely—from 2% to 26%—and were too heterogeneous to pool. No evidence of temporal change in the percentage of patients perceived as difficult was found.
The study authors stated: "Providers perceived 17% of clinic patients as difficult. Patients perceived as difficult were more likely to have depression, anxiety, a greater number of symptoms, personality disorders, or chronic pain. Less experienced providers were more likely to judge patients as difficult. Patients from difficult encounters had more unmet visit expectations and less satisfaction."
Regarding the clinical implications, they wrote: "Patients considered difficult were more likely to have chronic pain, personality disorders, and mental health disorders—chronic problems for which nonpsychiatric physicians often have limited training and few efficacious treatments." They added: "We suspect the ability to successfully treat these patients might change physician's perception of them from 'difficult' to 'rewarding.'"
In an accompanying editorial, Cédric Lemogne, MD, of the Université Paris Cité, and Pascal Cathébras, MD, of the CHU de Saint-Étienne, drew parallels between patients labeled "difficult" and those with persistent physical symptoms—somatic symptoms lasting months that "are only partially explained by identifiable pathophysiologic processes." They observed that the profile emerging from the meta-analysis was "strikingly similar" to that of patients with persistent physical symptoms, noting that both groups presented challenges to the traditional biomedical model.
The editorial authors cited long COVID as an example of how encounters become difficult "when routine investigations are normal, despite substantial symptom burden." They noted that cognitive-behavioral interventions can improve outcomes in some patients, emphasizing: "Such interventions do not deny biology; they explicitly target brain mechanisms that shape symptom perception."
Addressing the implications of diagnostic terminology, the editorial authors stated: "The labels we choose are indeed never neutral. To call someone a difficult patient is not only to describe an encounter but also to shape how responsibility, stigma, and therapeutic possibilities are distributed between patient, clinician, and health care system."
Drs. Lemogne and Cathébras argued that because difficulty "emerges in the space between patient and physician, rather than solely from the patient, it is also potentially modifiable by better training in communication about persistent physical symptoms, integration of psychosomatic expertise into general medical settings, organizational support to reduce time pressure and isolation, and explicit attention to clinicians' own emotional responses."
They concluded: "Difficult patients challenge our care models, our skills, and our capacity for empathy. [Dr.] Jackson and colleagues have provided an important map of where and with whom difficulty tends to arise. The next step is to use that map not to avoid these patients but to better prepare ourselves to care for them."
Disclosures are available online.
Source: Annals of Internal Medicine and Editorial
