This comprehensive article, titled “Outcomes associated with matching patients’ treatment preferences to physicians’ recommendations: study methodology,” meticulously outlines the research methodology for a study designed to explore the association between patient treatment preferences, physicians’ recommendations, and subsequent health outcomes. Published in BMC Health Services Research in 2012, this study is grounded in the principles of patient-centered care, defined as healthcare that is “respectful of and responsive to individual patient preferences, needs and values and ensuring that patient values guide all clinical decisions”. Patient-centered care is widely recognized for its positive impact, including improved patient satisfaction, enhanced quality of life, and better treatment adherence. Matching patient preferences to the care provided is considered a fundamental aspect of this approach.
Despite the known benefits, previous research attempting to evaluate the value of aligning patient preferences with treatment recommendations has faced significant methodological limitations. Many studies relied on simplistic binary choices (e.g., preference for treatment A over B) or hypothetical scenarios, which failed to capture the strength of patient preferences, the complex trade-offs patients make when considering different treatment attributes, or the realities of real-world clinical choices. The conceptual model underpinning this study posits that incorporating patient treatment preferences into decision-making influences patient satisfaction and adherence, which are critical mediators for achieving optimal clinical and Health Related Quality of Life (HRQoL) outcomes. To address these gaps, the authors developed a robust methodology to quantify the concordance between patient preferences and physician recommendations and to assess its association with concrete treatment outcomes.
The study employs a prospective cohort design, focusing on a specific patient population: individuals with moderate to severe psoriasis attending the bi-weekly outpatient dermatology clinics at the University Medical Centre Mannheim, University of Heidelberg, Germany. This clinic serves as a regional ‘Competence Centre for Psoriasis’ and acts as a primary, secondary, and tertiary care referral center, drawing a diverse patient population with a broad spectrum of the disease. The clinic’s annual volume of 250 to 300 patients with moderate to severe psoriasis ensures a feasible setting for participant recruitment across various clinical characteristics.
Participants and Recruitment: Eligible participants were aged 18 years or older, new or established patients, and had a physician-diagnosed moderate or severe psoriasis, indicated by a Psoriasis Area and Severity Index (PASI) score of ≥ 10, involvement of specific body areas (head, palmar or plantar surfaces), or psoriatic arthritis with any skin involvement, and were on systemic anti-psoriatic therapy. These criteria were purposefully chosen to ensure the sample required diverse treatment options. Exclusion criteria included inability to complete the online survey independently or to read and understand German, the language of the survey and subsequent interviews. Recruitment involved approaching each patient meeting the criteria before their doctor’s appointment. Research team members distributed informational leaflets, assessed eligibility, and obtained informed consent. Upon consent, a unique four-digit study identification number was assigned, and two follow-up visits (t2 at 12 weeks and t3 at 24 weeks post-initial visit) were scheduled. Participants completed a self-administered online survey before their medical appointment.
Key Measures and Data Collection: The study’s primary independent variable is the Preference Matching Index (PMI), a novel metric of concordance developed through a three-step process:
- Elicitation of Patient Preferences (Step 1): This involved measuring the value patients attach to different treatment attributes when presented with various treatment options. Relevant psoriasis treatment modalities and their attributes (processes or outcomes) and attribute levels were identified using “German evidence-based guidelines for the treatment of psoriasis” and consultation with clinical experts. Examples of process attributes included ‘treatment duration,’ ‘frequency,’ ‘cost,’ ‘location,’ and ‘delivery method,’ while outcome attributes included ‘magnitude of beneficial effect,’ ‘duration of beneficial effect,’ ‘probability of side effects,’ ‘probability of beneficial effect,’ ‘reversibility of side effects,’ and ‘side effect severity’. To manage respondent burden, each attribute was limited to four categories. Patient preferences were measured using choice-based conjoint analysis (CBC) via a self-administered online survey. This method effectively simulates real-world decision-making by presenting participants with pairwise comparisons of treatment scenarios comprising random combinations of attribute categories. The software then generated “preference scores” (partworth utilities values) for each attribute level for every participant, with higher scores indicating a greater preference.
- Abstraction of Physician-Recommended Treatments (Step 2): A trained research team member abstracted actual treatment modalities recommended by physicians from participants’ medical records following their initial clinical visit (t1).
- Calculation of the PMI (Step 3): This involved quantifying the alignment between participants’ preferences and the physician’s recommendation. The preference scores for the participant’s most preferred treatment attributes and the physician-recommended treatment attributes were summed. The PMI was then calculated using a specific formula: (preference score for recommended treatment – preference score for least preferred treatment) / (preference score for most preferred treatment – preference score for least preferred treatment) [29, Figure 1]. This index ranges from 0 (no preference concordance) to 1 (complete preference concordance).
The dependent variables were assessed at three time points: t1 (initial visit), t2 (12 weeks post-t1), and t3 (12 weeks post-t2). The 12-week interval is considered sufficient for short-term effectiveness, while the 24-week (t3) interval better reflects sustainable treatment adherence.
- The primary dependent variable was the change in Psoriasis Area and Severity Index (PASI) score from t1 to t3. PASI is a psychometrically valid and reliable objective clinical measure of disease severity, ranging from 0 (no disease) to 72 (maximal disease), with scores ≥ 10 indicating moderate to severe disease.
- Secondary dependent variables included the change in patient-reported treatment satisfaction from t1 to t3, measured by the 14-item, psychometrically validated Treatment Satisfaction Questionnaire for Medication (TSQM). Additionally, changes in self-reported health-related quality of life (HRQoL) from t1 to t3 were assessed using the validated Dermatology Quality of Life Index (DLQI), with scores ranging from 0 to 30, where higher scores indicate greater impairment.
Potential confounding characteristics such as sex, age, partnership status, employment status, educational attainment, and net annual household income were collected via a standard German demographic questionnaire. Factors that might moderate patient satisfaction and compliance, including treatment history and disease-related factors (e.g., time since diagnosis, co-morbidities like depression), were also measured. Data was primarily collected from participants’ online survey responses and physicians’ notations in medical records.
Analytic Strategy and Hypotheses: The study aimed for a minimum sample size of 200 to detect a moderate-sized change in the primary outcome, adjusted to 240 participants to account for a 20% dropout rate. Statistical analysis involves repeated measures analysis of variance to assess changes in outcomes across time points. Separate multivariate linear regression models will evaluate the association between PMI scores and each of the three study outcomes (PASI, TSQM, DLQI), controlling for identified confounders. Sensitivity analyses will stratify by gender and patient treatment experience. The study hypothesizes a statistically significant negative association between PMI and absolute change in PASI (implying that a closer match leads to greater reduction in disease severity). Conversely, it hypothesizes a statistically significant positive association between PMI and absolute change in treatment satisfaction, and a statistically significant negative association between PMI and absolute change in HRQoL (implying a closer match leads to improved satisfaction and less impairment in quality of life).
Ethical Considerations and Broader Implications: The study protocol received approval from the ethics committee of the Medical Faculty Mannheim, University of Heidelberg, adhering to the principles of the Helsinki Declaration. To prevent the act of preference elicitation from influencing subsequent self-reported satisfaction or quality of life, patient preferences were assessed before their medical consultation, and the results were not shared with patients at any point. The authors emphasize that this detailed methodology, while demonstrated in the context of psoriasis, holds significant potential for broader application in managing other chronic diseases, fostering more meaningful shared decision-making processes. The use of conjoint analysis allows for a nuanced understanding of patient preferences, acknowledging scenarios where patients may equally value different treatment options based on shared attributes, thus providing flexibility for physicians in recommending treatments that align with patient values. Acknowledged limitations include the potential for “dominant preference” bias in conjoint analysis, where strong feelings about one attribute might overshadow trade-offs. Also, adherence was not directly measured but is inferred through the conceptual model linking satisfaction to adherence and outcomes.
In conclusion, this article offers a rigorous and innovative methodological framework for quantifying patient-physician preference concordance and evaluating its impact on critical treatment outcomes. By providing detailed insights into the design, measures, and analytical strategies, the authors aim to promote the integration of patient preferences into clinical practice, ultimately enhancing treatment satisfaction, adherence, and overall effectiveness.
Reference:
Umar, N., Litaker, D., Schaarschmidt, M. L., Peitsch, W. K., Schmieder, A., & Terris, D. D. (2012). Outcomes associated with matching patients’ treatment preferences to physicians’ recommendations: study methodology. BMC Health Services Research, 12(1), 1. https://doi.org/10.1186/1472-6963-12-1