Latest / This Week In Wellbeing Measurement / mental health, public health, machine learning, wellbeing, quality of life

Transcript

mental health, public health, machine learning, wellbeing, quality of life · This Week In Wellbeing Measurement

• 0:00
  Jenny: If you wanted to know whether a place was really good for people, what would you count first?
• 0:05
  Davis: I'd want the obvious stuff, like safety, money, health care, but then I'd wonder who got asked, because a city can look great on paper and still feel hostile to an older person crossing the street.
• 0:16
  Jenny: Exactly, and my skeptical little bell goes off when one score claims to settle it, because a measure has to prove it fits the people and the place before anyone uses it to make decisions.
• 0:27
  Davis: But that's the hopeful part too, because better measurement starts by admitting what the old score missed, even when it's a women's empowerment tool saying, not yet, don't compare countries fairly until we've refined this for context...welcome to This Week In Wellbeing Measurement on paperboy.fm.
• 1:00
  Davis: This week got bigger fast: six hundred seventy-two query hits, all analyzed, and eighty-nine papers made the qualified set, meaning they actually fit the wellbeing measurement brief. That came from four hundred thirty-five unique authors across twenty-two countries.
• 1:18
  Jenny: The qualified count rose from sixty-eight to eighty-nine, so twenty-one more papers, or thirty point nine percent. The shape looks measurement-heavy: twenty-one surveys, sixteen qualitative studies, twelve quantitative papers, plus five RCTs, where people are randomly assigned to test a change.
• 1:38
  Davis: The search pool widened too. Query hits went from five hundred thirty-eight to six hundred seventy-two, up one hundred thirty-four, or twenty-four point nine percent, and the top themes were mental health at seven, public health at five, and machine learning at four.
• 1:54
  Jenny: That fits the episode's thread, but I'd keep the claim modest. We're seeing lots of tools for different decisions, not one master wellbeing score, and the country tags still cluster at USA six, China four, and India three, even though country coverage rose from nineteen to twenty-two.
• 2:13
  Davis: The author pool is the big jump: two hundred seventy-nine to four hundred thirty-five, up one hundred fifty-six, or fifty-five point nine percent. Inside that, eighty-six authors are first-time, meaning first-ever paper and not just new to our feed, while two hundred three are emerging and one hundred forty-six are experienced.
• 2:36
  Jenny: So the practical read is growth with a caveat. Nearly half the authors are emerging, forty-six point six percent, but this stats file has zero city coverage and zero institution coverage, so I can't yet tell whether these measures are spreading into new local settings or just new article records.
• 2:57
  Davis: Alright, let's get into the papers with a very practical one: Age-friendliness of a provincial city. Neville, Grigg, Adams, Dikken, and van Hoof look at Napier City in New Zealand, where twenty-one point one percent of residents are sixty-five or older, and ask whether a global age-friendly framework can become a local planning tool.
• 3:17
  Davis: The plain finding is that the questionnaire worked pretty well for Napier. They ended up with a twenty-three item tool covering the eight World Health Organization age-friendly domains, like housing, transport, social participation, and health services, plus one extra domain on perceived financial wellbeing.
• 3:34
  Jenny: How do we know this questionnaire is measuring age-friendliness, rather than just satisfaction with a few city services?
• 3:41
  Davis: They did it in two phases. First, they checked the wording and relevance with policy experts and older adults, which is face and content validity, meaning, does this look right and cover the right things to the people who know the place. Then they surveyed three hundred fifty-four older adults and used confirmatory factor analysis, which is a statistical test of whether the answers actually fit the structure the theory predicts.
• 4:05
  Davis: The fit and internal consistency came out strong, and the city-level pattern was useful: housing was rated most positively, while transportation, community support, health services, and financial situation were rated lowest. They also found five resident clusters, so older people in Napier weren't having one shared experience of the city.
• 4:25
  Jenny: That feels like the thread for this week right away: validated tools, local truths. I buy this as a strong Napier tool, especially with three hundred fifty-four residents and older adults involved before the stats, but one provincial city can't prove the same questionnaire works the same way across all of New Zealand.
• 4:45
  Jenny: That Napier caveat, one city can't prove the tool travels, is exactly where this next paper lives: Women’s Empowerment Metric for National Statistical Systems. Yount and colleagues built WEMNS for routine national surveys in low- and middle-income countries, so governments could track gender-related Sustainable Development Goals, the UN targets on things like equality, resources, and agency.
• 5:11
  Jenny: The useful part is that they didn't just launch a dashboard and call it done. They tested face-to-face survey modules in Bangladesh, Malawi, and Nepal, and looked at thirteen item sets covering empowerment constructs, meaning clusters of questions meant to capture things like control over choices, access to resources, and voice in decisions.
• 5:32
  Davis: If the measure doesn't travel cleanly across countries, what should national statistical offices actually do with it?
• 5:39
  Jenny: Use it carefully for learning, not ranking. They ran exploratory and confirmatory factor analysis, which means they first looked for the shape of the answers and then tested whether that shape held up, and nine of thirteen item sets showed configural invariance across gender and country, meaning people seemed to organize the ideas in a similar basic pattern.
• 6:01
  Jenny: But the harder test mostly failed. No item set reached even partial scalar invariance across countries, meaning you can't confidently say a higher score in Nepal equals the same amount of empowerment as a higher score in Malawi or Bangladesh, and the authors say the measures need refinement before valid cross-country or gender comparisons.
• 6:23
  Davis: That's rare candor, and it fits the thread: validated tools, local truths. For a statistics office, WEMNS looks promising as a way to see what empowerment means inside a country now, but not yet as a league table where countries get ranked off numbers that may not mean the same thing.
• 6:40
  Davis: That warning about not turning every measure into a league table is the bridge here, because this next paper is also in validated tools, local truths. It's called Development and Validation of the Primary Care Needs Assessment Questionnaire, and it asks how primary care can measure what people still need, not just what clinics already count.
• 6:59
  Davis: The plain finding is that the PCNA can pull out unmet needs that appointment logs miss, especially mental health needs, access to specialists, and barriers like distance, cost, limited service availability, and hard-to-navigate systems. They tested it with eight hundred seventeen people from community and primary care settings, and the final version landed on nine factors, meaning nine clusters of needs that seemed to travel together in people's answers.
• 7:25
  Jenny: What makes this different from asking patients whether they got an appointment?
• 7:30
  Davis: They built it from the ground up first, using focus groups with primary care professionals and interviews with community members, so the questions weren't only written from the clinic's point of view. Then they ran exploratory factor analysis on five hundred twenty people, which means they looked for the hidden shape in the answers, and confirmatory factor analysis on another two hundred ninety-seven, which means they checked whether that shape held up; the internal consistency was alpha equals zero point seven six, so the items hung together reasonably well. The big caveat is that the abstract doesn't give a country or detailed setting, so we shouldn't pretend this exact tool fits every health system.
• 8:07
  Jenny: I like that distinction: eight hundred seventeen people and a split validation sample make this more than a suggestion box, but it's still a local instrument before it's a universal yardstick. For a clinic manager, the takeaway is blunt: don't only count visits, count the reasons people still can't get what they need.
• 8:28
  Paperboy.fm: This is the free version of the podcast. Subscribe at paperboy.fm to access a dozen different paper review podcasts for five dollars a month.