Attachment F – Public Comments: Brookings

Carol Graham

Leo Pasvolsky Senior Fellow,

Global Economy and Development Program

1775 Massachusetts Avenue, NW

Washington, DC 20036

telephone 202.797.6000

fax 202.797.6004

web brookings.edu

May 12, 2020

Doris Lefkowitz

Reports Clearance Officer

Agency for Healthcare Research and Quality

Dear Ms. Lefkowitz,

I write in response to the FR notice for comments on the SAQ supplement for 2021. I come from the perspective of an economist who works on well-being and public health, in both the U.S. and abroad.

I was one of the early economists to work on life satisfaction and well-being issues, beginning in the early 2000’s, as part of a small group of collaborating economists and psychologists, including Richard Easterlin, Andrew Oswald, George Akerlof, Danny Kahneman, and Arthur Stone. At the time, I was the only one of that group working on developing economies. I was able to demonstrate that the patterns in the determinants of well-being were essentially the same in both poor and rich countries and around the world. I then began to work on issues around what well-being causes (rather than what causes well-being) and wrote the first economics paper (Journal of Economic Behavior and Organization, 2004) that showed that respondents who had higher levels of well-being did better later in life, in the income, health, and social arenas.

That finding has since been confirmed by many other papers, using different methods, and is also showing up in my more recent work, matching trends in well-being and ill-being with those in deaths of despair, at the level of individuals, races, and places. We find that lack of hope is the most consistent well-being “predictor” of being in the deaths of despair category (Journal of Population Economics, 2019). In a new historical analysis, based on the PSID (Journal of Economic Behavior and Organization, 2019), we also find that respondents born from 1935-45 who reported to be optimistic in their twenties were more likely to still be alive in 2015. We also find that the only population cohort for whom optimism fell, beginning in the late 1970’s – coinciding with the first wave of manufacturing decline – were less than college educated white men, the same cohort that is over-represented in the deaths of despair today. We are in the process of building a country-wide well-being indicator, with county level data and based on both well-being and ill-being markers and premature mortality data from the CDC. We are now in the process of adding in COVID mortality rates as well, as they are likely to be higher among vulnerable cohorts of the same profile, particularly in rural areas.

I have also been involved in introducing the usage of well-being metrics into government statistics as complements to income-based measures (as in the UK and New Zealand), for the purposes of policy design, monitoring, and assessments (for a summary, please see Graham, Laffan, and Pinto in Science, 2018).

As such, I am very excited about and supportive of the new supplement that AHRQ is doing, and the inclusion of a range of measures of both ill-being and well-being (in this instance life satisfaction, which is the best marker to use if only one is going to be included in the survey).

That said, I do have two concerns about the draft survey, which I hope that you and your team will consider as you finalize the draft. The first is about the placement of Q15 on life satisfaction will lead to answers that are biased downward after respondents consider the previous questions. In my experience, life satisfaction questions should be placed early in a survey so that they are not influenced by answers to questions on specific life aspects such as job satisfaction, marital satisfaction, and economic concerns. Consequently, I ask AHRQ to consider moving the life satisfaction question to the beginning of the survey. As it stands now, it comes after several questions about insecurity and hardship, and I fear that will bias responses in an unintended manner.

Second, I am concerned that the questions about socialization do not allow respondents to differentiate between in-person socialization and virtual socialization, which still may be present in 2021. Consequently, I ask AHRQ to consider a re-phrasing of the questions in a way that takes into account the limitations on socializing that may be in place at the time due to social distancing. 

I look forward to your response and would be happy to discuss these issues further as needed. Thank you in advance for your consideration. Please feel free to reach me at cgraham@brookings.edu or (202) 527-8086.

Yours sincerely,

Carol Graham

Leo Pasvolsky Senior Fellow

The Brookings Institution/

College Park Professor

University of Maryland