Tailoring and Evaluating Treatment with the Patient-Specific Needs Evaluation: A Patient-Centered Approach

Abstract

Background: No patient-reported instrument assesses patient-specific information needs, treatment goals, and personal meaningful gain (PMG), a novel construct evaluating individualized, clinically relevant improvement. This study reports the development of the Patient-Specific Needs Evaluation (PSN) and examines its discriminative validity (ie, its ability to distinguish satisfied from dissatisfied patients) and test-retest reliability in patients with hand or wrist conditions.

Methods: A mixed-methods approach was used to develop and validate the PSN, following Consensus-Based Standards for the Selection of Health Measurement Instruments guidelines, including pilot testing, a survey (pilot, n = 223; final PSN, n = 275), cognitive debriefing ( n = 16), expert input, and validation. Discriminative validity was assessed by comparing the satisfaction level of patients who did and did not achieve their PMG ( n = 1985) and test-retest reliability using absolute agreement, the Cohen kappa, and intraclass correlation coefficients ( n = 102). The authors used a sample of 2860 patients to describe responses to the final PSN.

Results: The PSN has only 5 questions (completion time, ±3 minutes) and is freely accessible online. The items and response options were considered understandable by 90% to 92% of the end-users and complete by 84% to 89%. The PSN had excellent discriminative validity (Cramer V, 0.48; P < 0.001) and moderate to high test-retest reliability (kappa, 0.46 to 0.68; intraclass correlation coefficients, 0.53 to 0.73).

Conclusions: The PSN is a freely available, patient-centered decision support tool that helps clinicians tailor their consultations to patients' individual needs and goals. It contains the PMG, a novel construct evaluating individualized, clinically relevant treatment outcomes. The PSN may function as a conversation starter, facilitate expectation management, and aid shared decision-making. The PSN is implementation-ready and can be readily adapted to other patient populations.

Fig. 1.

Flow chart of the development (in blue) and validation (in orange) of the PSN, describing the sample and the most important goals and activities per stage.

Fig. 2.

Visual of the PSN treatment goal and PMG. In this example, the patient entered that the most important treatment goal was to improve the performance of activities. The score at baseline was 3 on a scale of 0 to 10 (high scores indicate better performance), and the patient indicated that a score of 7 is needed to become satisfied with the treatment result. After this section is filled in, the digital PSN automatically generates a statement on the treatment goal and the PMG, so the patient can check whether it is correct or needs modification.

Fig. 3.

Distribution of information need in the final version of the PSN. The patient chooses one of the following options: I do not need information; diagnosis (I have questions about the diagnosis); advice (I want to know what is the best thing to do in my situation); treatment (I have questions about the treatment); or perspective (I want to know what to expect in the future).

Fig. 4.

Goal domains chosen as most important in the final version of the PSN.

Fig. 5.

Distribution of scores on the most important goal domain at baseline (above) and the score patients reported that they needed to achieve to be satisfied with the treatment results (below) in the final version of the PSN. Not all patients want to obtain the maximum score on their most important outcome domain to be satisfied with the treatment results. The median score needed to be satisfied with the treatment result was 9 in the final version. For ease of interpretation, we converted each domain score to the same scale (ie, reversing the scores on the pain, numbness, and tingling domains).

Fig. 6.

Discriminative validity of the PMG in 1985 patients, demonstrating that patients who obtained their PMG were much more often satisfied with their treatment results compared with patients who did not obtain their PMG, with a medium to large effect size (Cramer V, 0.48; P < 0.001).