Doctors with chronic illness in Egypt often self-treat and skip medications

In a finding that challenges one of medicine’s most comfortable assumptions, researchers at Luxor University in Egypt have shown that physicians suffering from chronic diseases are no better at taking their prescribed medications than ordinary patients with comparable conditions. The study, published in Discover Mental Health, compared 250 practicing physicians who live with chronic illnesses against 250 age-, sex-, and disease-matched non-physician controls, all recruited from the same university outpatient clinics in Upper Egypt. Adequate self-reported adherence was recorded in 28.4 percent of physicians and 31.2 percent of controls, a difference that proved statistically indistinguishable (p = 0.557). Even more striking, when the investigators linked pharmacy refill records and pill-count documentation for a consenting subset of participants, the objective measures told the same story: median proportion of days covered (PDC) and the proportion of participants reaching the accepted PDC threshold of 0.80 or higher were comparable between the two groups. Medical knowledge, it seems, does not translate automatically into medication-taking behavior.

The research team, led by Mahmoud Ahmed Bekiet of the Psychiatry Department and Ahmed Jado Nabih Ali of the Internal Medicine Department at the Faculty of Medicine, Luxor University, designed the study as a cross-sectional comparative observational investigation reported according to the STROBE guidelines. Participants were adults aged 18 or older diagnosed with at least one chronic illness requiring long-term pharmacological treatment, recruited during routine outpatient visits at internal medicine and specialty chronic disease follow-up services. Controls were selected using a 1:1 frequency-matching strategy based on age, sex, and disease category, which minimized confounding from the clinical type of chronic illness and the associated medication regimens. Based on preliminary power calculations assuming a two-sided alpha of 0.05, the team estimated that a minimum of 178 participants per group would be required, and they inflated the target to 250 per group, for a total of 500 participants, to compensate for incomplete responses and missing questionnaire data.

The primary outcome, medication adherence, was assessed with the 8-item Morisky Medication Adherence Scale (MMAS-8), a validated self-report instrument scored according to established guidelines and categorized using standard cutoffs to define adequate versus inadequate adherence. To strengthen the robustness of the assessment, secondary instruments were also deployed, including the Adherence to Refills and Medications Scale and the General Medication Adherence Scale. Crucially, for participants who consented to pharmacy linkage, the researchers calculated the proportion of days covered over a fixed six-month observation window, defined as the number of days on which medication was available divided by the total observation period, with PDC values of at least 0.80 considered adequate in line with conventional pharmacoepidemiologic standards. Concordance between the subjective MMAS-8 results and the objective PDC data was tested using Cohen’s kappa and correlation analysis, and the results were sobering: agreement between self-reported and objective adherence was poor in both physicians and controls, a finding the authors describe as a central reason why adherence research demands multimethod assessment.

Beyond adherence itself, the study probed the psychological and behavioral terrain surrounding medication use. Self-treatment behavior was quantified with the Self-Administration Scale for Self-Medication Practices, a structured tool in which higher scores reflect greater engagement in self-directed medication behaviors, such as adjusting doses or managing symptoms without formal clinical supervision. Depressive symptoms were measured with the Patient Health Questionnaire-9, anxiety with the Generalized Anxiety Disorder-7, and burnout with a standardized burnout scale. Covariates included age, sex, number of chronic conditions, disease duration, number of prescribed medications, and polypharmacy status defined as five or more medications, alongside education level, monthly income, and health insurance status. Continuous variables were compared with Mann-Whitney U tests because of non-normal distributions, categorical variables with chi-square or Fisher’s exact tests, and a Bonferroni correction was applied to multiple comparisons of baseline characteristics to guard against type I error.

The multivariable logistic regression analysis delivered the study’s most consequential result: physician status was not independently associated with adequate adherence after adjustment for demographic, clinical, and psychological variables. Instead, two other factors emerged with statistical significance. A higher number of prescribed medications was independently associated with greater odds of adequate adherence in both the overall and physician-only models, and greater engagement in self-treatment behaviors also predicted higher odds of adequate self-reported adherence. This relationship with polypharmacy, while counterintuitive at first glance, is consistent with a growing body of literature suggesting that a larger medication burden can make illness more salient, prompt more frequent follow-up visits, and impose more organized routines. Recent syntheses of polypharmacy interventions, including drug reviews and deprescribing programs, have found that such measures may improve prescribing appropriateness without necessarily improving clinical outcomes, underscoring that medication count is less a measure of over-treatment than a marker of disease burden and care intensity.

Perhaps the most provocative finding concerns self-treatment. Controls actually scored higher than physicians on the self-medication scale, indicating that non-physician patients were more engaged in self-directed medication behaviors. Yet higher self-treatment scores were positively associated with adequate adherence in the multivariate models. The authors interpret this cautiously: self-treatment may reflect active self-management and high engagement, such as purposeful regimen planning and symptom-based adjustment, which can coexist with routine adherence rather than undermine it. They are quick to note that this does not mean self-treatment is appropriate or safe. The broader evidence on self-management interventions indicates that patient agency improves outcomes and self-efficacy only when it is anchored in professional guidance, systematic follow-up, and system-level supports, and the SAS-SMP instrument itself cannot distinguish structured self-management from informal dose adjustments or potentially unsafe practices.

Equally notable is what the study failed to find. Neither PHQ-9 depression scores, nor GAD-7 anxiety scores, nor burnout measurements, nor physician-specific workload variables such as weekly working hours, number of night shifts per month, or specialty type showed independent associations with adherence. The authors urge caution in interpreting these null results. Mental health can influence adherence through motivation, executive function, and perceived necessity of medication, but when multiple determinants act concurrently and adherence is self-reported, effects may be diluted. Even when PHQ-9 and GAD-7 scores were analyzed by clinically meaningful severity categories, no association with adherence appeared, which the researchers attribute either to the inherent limitations of cross-sectional design in capturing dynamic psychological influences or to the possibility that adherence in this cohort is driven more by behavioral and contextual factors than by symptom severity.

The context in which physicians manage their own health has changed markedly in recent years. Large national surveys conducted during the COVID-19 pandemic documented substantial shifts in physician burnout and work-life integration, painting a picture of a professional environment capable of degrading routine preventive care, follow-up persistence, and long-term health behavior maintenance. Physicians are highly health-literate and enjoy favorable access to medical knowledge, yet time constraints, symptom normalization, professional stigma, and fragmented care pathways can obstruct continuous self-management when the clinician becomes the patient. The methodological literature reinforces this concern: social desirability and professional identity may bias self-reports among physicians, and a COSMIN-informed systematic review of adherence instruments in cardiovascular disease and type 2 diabetes has highlighted inconsistent measurement quality among widely used patient-reported measures, recommending careful attention to instrument selection and reporting.

The clinical implications are direct. Clinicians should not assume that physician-patients are inherently adherent to long-term therapy, and physicians living with chronic illness may require the same adherence screening and follow-up support as any other patient group. Educational knowledge alone is insufficient, and structured self-management support along with stigma-free access to care should be promoted for healthcare professionals. The authors also emphasize that interventions work best as combinations of education, follow-up, reminders, structured self-management support, and care coordination, with digital tools offering benefits in facilitating routines, access, and engagement, though their effectiveness depends heavily on implementation and patient circumstances.

The study carries limitations that temper, though do not erase, its message. The cross-sectional design precludes causal inference, and reverse causation remains possible, since individuals more engaged in their care may both adhere more consistently and report greater self-directed medication practices. Self-report instruments are vulnerable to recall and social desirability bias, objective measures such as PDC and pill counts are indirect proxies that do not confirm actual ingestion, and the pharmacy-linked subset was slightly older and reported lower adherence than non-consenters, introducing potential selection bias. Residual confounding from unmeasured factors such as health literacy, self-efficacy, illness perceptions, and continuity of care cannot be excluded, and the single-institution setting in Upper Egypt may limit generalizability to other healthcare systems and physician populations. Still, the investigators frame their findings as hypothesis-generating and call for longitudinal studies with finer characterization of self-management behaviors. Their bottom line stands: medication adherence among physicians with chronic illness is broadly comparable to that of matched non-physician patients on both self-reported and objective measures, and knowledge, by itself, is not enough.