Use of a discrete choice experiment approach to elicit patients’ preferences for hip fracture rehabilitation services as part of a feasibility study

Abstract
Background
Hip fracture, a common health issue in older age, is associated with serious consequences for the individual, such as mortality and frailty. Multidisciplinary rehabilitation has been proposed as a method to help older people recover after a hip fracture. Previous research exploring patients' preferences for hip fracture rehabilitation has been conducted in Australia. However, little research has been done in a UK context to inform rehabilitation services.
Methods
As part of a larger feasibility study conducted in North Wales, a discrete choice experiment (DCE) questionnaire was administered to patients who recently had a hip fracture and were randomised to receive either usual care or a newly developed multidisciplinary rehabilitation intervention. DCEs are a stated preference technique for eliciting individuals' preferences about goods and services. Individuals make trade-offs between choices, and these trade-offs provide information about individuals' relative preferences for different attributes of a service. Findings from a systematic review, patient focus groups, and a health-care professional survey conducted to develop the intervention were used to design the attributes and levels for the DCE questionnaire. Participants who consented to take part in the larger feasibility study were asked to complete the DCE questionnaire at their 3 month follow-up.
Findings
32 (65%) of 49 participants completed the DCE questionnaire at follow-up. 22 participants (69%) were women, and ten (31%) were men. The pilot DCE demonstrated that the method could be used with participants with a mean age of 79 years (SD 7·54). Logit regression (Stata, version 10.1) indicated participants' preferences for increased time with the health-care professional who delivers rehabilitation (β=–0·005, 95% CI 0·000 to −0·010), and a physiotherapy or occupational therapy assistant to deliver the rehabilitation sessions (β=–0·596, −0·862 to −0·445).
Interpretation
Previous DCEs conducted in Australia focused on patient outcomes such as pain and dose of therapy. The design of the DCE described here used attributes associated with service configuration, which could have the potential to inform service implementation and assist service design, incorporating the preferences of patients. It could also provide lessons for future DCEs conducted with similar populations.
Funding
National Institute for Health Research Health Technology Assessment (NIHR HTA) programme (ref 11/33-03).