NU713 Unit 3 Measurement Critique, ADI and PRAPARE Example

Thirty-Eight Percent Said They Needed Nothing: A Measurement Critique of a Neighborhood Deprivation Index and a Household Social Needs Screen at a Composite Brennard County Health Center

[Student Name]

Purdue University Global

NU713: Epidemiology and Social Determinants of Population Health

Unit 3 Measurement Critique

[Instructor Name]

[Date]

The health center, patients and figures are composite; numbers are bracketed placeholders.

The Puzzle

The health center began screening adults with PRAPARE, a standardized social needs assessment developed for community health centers (National Association of Community Health Centers, 2016), in [2023]. By the end of [2024], [4,860] adults had completed it. The center had also started assigning each patient the Area Deprivation Index decile of their home block group, a neighborhood measure built from census indicators of income, education, employment and housing and published through the Neighborhood Atlas (Kind & Buckingham, 2018). Leaders assumed the two would agree. They did not, and the disagreement is the subject of this critique.

The Cross-Tabulation

Patients were grouped by the state decile of their block group on the index: deciles 1 to 5, the less deprived half; deciles 6 to 8; and deciles 9 and 10, the most deprived fifth.

In deciles 1 to 5 there were [1,420] screened adults. Of these, [21] percent reported at least one need: [15] percent food insecurity, [7] percent housing instability and [5] percent transportation barriers.

In deciles 6 to 8 there were [1,690] adults. Of these, [44] percent reported at least one need: [31] percent food, [20] percent housing and [14] percent transportation.

In deciles 9 and 10 there were [1,750] adults. Of these, [62] percent reported at least one need: [45] percent food, [33] percent housing and [26] percent transportation.

The gradient runs in the expected direction, so the instruments are related. But two cells carry the argument. [Thirty-eight] percent of patients in the most deprived block groups reported no need, and [21] percent of patients in the less deprived half reported at least one, most often food insecurity. If the center used the neighborhood index alone to decide who receives help, it would reach many households that report no need and miss many that do.

The index describes where a patient lives; the screen describes what a patient reports. The two are related, but neither can stand in for the other.

What Each Instrument Measures

Construct. The Area Deprivation Index measures socioeconomic disadvantage of a place, combining census indicators into a single rank. PRAPARE measures a person's or household's reported needs, assets and circumstances, such as food, housing, transportation, stress and social isolation.

Level. The index is assigned to a block group, an area of roughly [600 to 3,000] people. PRAPARE is completed by an individual.

Data source. The index comes from American Community Survey estimates, pooled over five years. PRAPARE comes from the patient's own answers at a visit.

Recall period. The index describes a neighborhood over several years. PRAPARE asks about the past year for most items and about current circumstances for others.

Who completes it. The index requires only an address, which makes it cheap and complete for every patient with a valid one. PRAPARE requires time from staff and patients, and its completeness depends on how often it is offered and how willing patients are to answer.

Why the Mismatch Is Expected

The mismatch is not an error in either instrument. A deprived neighborhood contains households with stable incomes and adequate food, and a less deprived one contains households in crisis. Assigning a neighborhood score to a person commits the ecological fallacy, assuming that an area's average describes each resident. A study of more than 36,000 patients in a national network of community health centers found exactly this pattern: patients who reported social risks lived in all census tracts, and a tract-level deprivation score correctly classified patients with and without social risks only 48.0 percent of the time (Cottrell et al., 2020).

The screen has its own weaknesses. Patients may decline to report needs out of pride, fear of being judged, or doubt that reporting will lead to help. Some needs are seasonal and missed by a single screen. And the [38] percent reporting no need in the most deprived group may include people who have learned to manage hardship and do not describe it as a need, which is a limitation of any self-report measure.

What the Center Had Been Doing

Before this analysis, the center's outreach team had drawn its call list from the index alone, phoning patients in deciles 9 and 10 to offer help with food and utilities. Staff reported that many calls ended quickly, with patients saying they were fine, while front-desk workers at the clinic in a less deprived neighborhood kept hearing from patients who could not afford groceries at the end of the month. The cross-tabulation explained both experiences. It also showed that the call list, built to be efficient, was spending much of its effort on households that did not want help and missing others that did.

Recommendations

Each instrument should be used for what it measures.

The Area Deprivation Index should guide decisions about places and programs: where to locate a satellite clinic, which neighborhoods to prioritize for outreach, and how to adjust for neighborhood disadvantage when comparing quality results across clinics. It is complete and cheap for these purposes.

PRAPARE should guide decisions about individual patients: who receives a referral for food assistance or legal help with housing. Because individual needs cannot be read from a neighborhood score, the screen should be offered to every adult, not only to those in deprived block groups.

Where the two disagree, the disagreement is useful information. A patient in a less deprived neighborhood who reports food insecurity may be more isolated from help than one in a neighborhood with food pantries nearby. A patient in the most deprived decile who reports no need may deserve a gentle second conversation, not because the screen is wrong but because needs change.

What Neither Captures

Neither instrument captures structural conditions that shape both neighborhood and household circumstances, such as discriminatory lending, zoning or wage policy. Neither measures strengths well, although PRAPARE includes some assets. And neither tells the center whether help, once offered, reached the patient. Those questions need other measures, including follow-up on referrals, which the center does not yet track.

Conclusion

A neighborhood index and a household screen answer different questions. The health center's data show a clear gradient between them and a large mismatch at the level of individual patients. Using the index for places and the screen for people, and treating disagreement between them as a prompt rather than an error, lets each instrument do the work it was built for.

References

Cottrell, E. K., Hendricks, M., Dambrun, K., Cowburn, S., Pantell, M., Gold, R., & Gottlieb, L. M. (2020). Comparison of community-level and patient-level social risk data in a network of community health centers. JAMA Network Open, 3(10), Article e2016852. https://doi.org/10.1001/jamanetworkopen.2020.16852

Kind, A. J. H., & Buckingham, W. R. (2018). Making neighborhood-disadvantage metrics accessible: The Neighborhood Atlas. New England Journal of Medicine, 378(26), 2456-2458. https://doi.org/10.1056/NEJMp1802313

National Association of Community Health Centers. (2016). PRAPARE: Protocol for responding to and assessing patients' assets, risks, and experiences.