Caregiver risk preferences for delaying loss of ambulation in Duchenne muscular dystrophy
Jason Shafrin
Senior Managing Director, Center for Healthcare Economics & Policy at FTI Consulting; Adjunct Professor, University of Southern California
That is the title of a paper published today in Current Medical Research and Opinion with co-authors Nadine Zawadzki, Moises Marin, Ivana Audhya, Lauren Sedita, Natasha Kulkarni and Alexa Klimchak. The abstract is below.
Objectives Quantify caregiver risk preferences to inform the “value of hope” for Duchenne muscular dystrophy (DMD) therapies affecting time to loss of ambulation (LoA).
Methods Caregivers (medical decision-makers) of patients with DMD were surveyed to evaluate their preferences across 2 therapies with identical expected (average) time to LoA: 1 with variable (i.e. possibly longer or shorter than average) time to LoA and 1 with fixed (i.e. certain) time to LoA. Time to LoA with the fixed therapy was altered to determine the caregiver’s indifference point. Study endpoints were (i) the share of caregivers who preferred the variable (vs fixed) time to LoA therapy; and (ii) the length of fixed time to LoA that would result in caregiver indifference between the variable and fixed therapies, calculated using parameter estimation by sequential testing. The base case examined therapy choice for a hypothetical ambulatory DMD patient aged 9 years; sensitivity analyses explored preferences for younger (aged 5) and older (aged 13) patients.
Results Among 103 caregivers surveyed, 72 (69.9%) preferred the variable time to LoA therapy for a hypothetical 9-year-old patient with DMD (p < 0.001). Caregivers were willing to give up 11.5 months (p < 0.001) of certain time to LoA for a chance of longer time to LoA. Caregivers’ preference for the variable therapy decreased with hypothetical patient age at treatment initiation, from 72.8% (75/103) for age 5 (p < 0.001) to 60.2% (62/103) for age 13 (p = 0.048).
Conclusions Caregivers of patients with DMD demonstrated risk tolerance (positive value of hope) for therapies that could delay LoA.
You can read the full article here.
Senior HEOR strategist
1wJason Shafrin - I don't think you can ascribe preference of Treatment B to mean a preference for 'hope'. It's not the distributions that I'm querying. They appear to be correctly correlated from the eteplirsen trials. I'm querying the survey choice that was presented to DMD patients (see below). Based on the calculations in my first comment below, unless patients interpret "less than 1 year" to mean '0 years exactly' and "more than 7 years" to mean 'exactly 7 years', then Treatment B has greater expected years of ambulation than A. Treatment B would be preferred by any rational actor. You've argued that preference for Treatment B is due to preference for 'hope', but there is another, more obvious explanation. Treatment B as represented in the survey is simply better in terms of outcomes than Treatment A.
Economist
2wMy leisure/unemployment time still includes this intellectual curiosity! Who is the one having the value of hope in this case: the carer or the patient? Thanks for sharing, always a pleasure reading on the broader elements of value
Senior HEOR strategist
2wThis is really interesting stuff, thanks for sharing. I was curious about the expected years of ambulation that patients might expect from the presentation of Treatment A and Treatment B in the survey. The only way I could get Treatment B to have the same expected years of ambulation to Treatment A was to assume that patients had interpreted 'Patients experience loss of ambulation in one year or less' to be 0 years (immediate loss of ambulation) and 'Patients experience loss of ambulation in more than seven years' to be 7 years exactly. If the survey respondent had any more of an optimistic reading of 'less than 1 = 0' or 'more than 7 = 7' then the expected yrs of ambulation is increased for Treatment B (see calcs below). Unless I am missing something (which is not out of the question), it is unsurprising that patients chose Treatment B, as under anything except the most pessimistic interpretations of the survey questions, Treatment B results in greater expected yrs of ambulation.