Skip to content

Commit cade085

Browse files
cschanhniemcschanhniem
authored
evidence: source inventory for hypertension-control-global (#41)
literature-scout source-inventory task. Adds 6 dated, classified evidence records (usable/limited/rejected) with measured-BP discipline: - NCD-RisC Lancet 2021 pooled analysis (usable backbone) - WHO STEPS microdata (usable) - DHS Program biomarkers (limited) - GBD 2021 high-SBP burden (limited) - PURE study JAMA 2013 (limited) - self-report bias study (rejection rule for self-report-only surveys) Also fixes a rotted URL in the pre-existing who-hypertension-2023 record (404 -> stable WHO IRIS handle) and refreshes its access date. Updates datasets.md and evidence.md to match. Satisfies done_condition: >=4 sources classified with explicit reasons covering measure family, timing, geographic grain, and linkage relevance.
1 parent 2f00ca7 commit cade085

3 files changed

Lines changed: 163 additions & 10 deletions

File tree

‎problem-packs/public-health/hypertension-control-global/datasets.md‎

Lines changed: 21 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -1,15 +1,29 @@
11
# Dataset Inventory
22

3-
| Source | Grain | Status | Use |
4-
| ---------------------------- | ------------ | ------ | ------------- |
5-
| Primary survey data | Sub-national | Usable | Core analysis |
6-
| Satellite/environmental data | Variable | Usable | Risk context |
7-
| Population data | 100m | Usable | Denominator |
3+
Dated source inventory for hypertension detection and control gaps in LMICs. Classification reflects fitness for a sub-national detection-treatment-control cascade analysis. Inventory compiled 2026-06-18.
4+
5+
## Sources
6+
7+
| Source | Grain | Measure family | Status | Use |
8+
| ---------------------------------------- | --------------------------------- | --------------------------------------- | -------- | ---------------------------------------------- |
9+
| WHO Global Report on Hypertension (2023) | Global / national | Modeled cascade benchmark | Usable | Global benchmark for diagnosis/control rates |
10+
| NCD-RisC pooled analysis (Lancet 2021) | National, ages 30-79 | Measured-BP cascade by country and sex | Usable | Cross-country comparison backbone |
11+
| WHO STEPS microdata (NCD Microdata Repo) | National; limited sub-national | Measured-BP survey microdata | Usable | Country cascade estimates from raw survey data |
12+
| DHS Program biomarkers | National + georeferenced clusters | Measured-BP prevalence (subset) | Limited | Supplementary prevalence where BP was measured |
13+
| GBD 2021 high systolic BP burden (IHME) | National + some subnational | Modeled attributable burden (DALYs) | Limited | Burden context, not a cascade measurement |
14+
| PURE study (JAMA 2013) | Community cohorts, ages 35-70 | Measured-BP cascade, non-representative | Limited | Historical cross-income comparison only |
15+
| Self-report-only household surveys | Varies | Self-reported diagnosis | Rejected | Underestimates measured prevalence (see below) |
16+
17+
## Classification Reasons
18+
19+
- **Usable** sources provide directly measured blood pressure and a detection/treatment/control cascade at national or finer grain.
20+
- **Limited** sources are usable only for a narrower purpose: DHS BP coverage is survey-specific and cascade items are inconsistent; GBD outputs are attributable burden, not measured control; PURE is non-representative community cohorts from 2003-2009.
21+
- **Rejected**: self-report-only surveys underestimate measured prevalence by ~20% (women) to ~55% (men) per a large Indian comparison study (SSM - Population Health, 2022), so they cannot anchor prevalence or cascade estimates.
822

923
## Required Properties
1024

11-
Source methodology, data year, geographic grain, license.
25+
Source methodology, data year, geographic grain, measure family (measured vs self-reported), license, and detection/treatment/control coverage.
1226

1327
## Rejection Rule
1428

15-
Rejected if methodology, grain, date, or license unverifiable.
29+
Rejected if methodology, grain, date, or license is unverifiable, or if hypertension status is self-reported rather than measured.

‎problem-packs/public-health/hypertension-control-global/evidence.json‎

Lines changed: 118 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -5,16 +5,131 @@
55
"claim": "WHO estimates 1.3 billion adults have hypertension, with only 45 percent diagnosed and 20 percent controlled globally, and control rates below 10 percent in most LMICs.",
66
"evidence_type": "primary-source",
77
"source": {
8-
"title": "Global Report on Hypertension",
9-
"url": "https://www.who.int/publications/i/item/9789240081060"
8+
"title": "Global report on hypertension: the race against a silent killer",
9+
"url": "https://iris.who.int/handle/10665/372896"
1010
},
1111
"source_date": "2023-09-19",
12-
"access_date": "2026-06-07",
12+
"access_date": "2026-06-18",
1313
"method": "Reviewed WHO hypertension report, detection and control cascade data.",
1414
"limitations": [
1515
"Diagnosis data depends on survey frequency.",
1616
"Treatment-control data from routine systems is incomplete."
1717
],
1818
"confidence": "high"
19+
},
20+
{
21+
"id": "ncd-risc-hypertension-trends-2021",
22+
"problem_id": "public-health/hypertension-control-global",
23+
"claim": "A pooled analysis of 1201 population-representative studies with measured blood pressure (104 million participants, 184 countries) estimates age-standardized hypertension prevalence, detection, treatment, and control for adults aged 30-79 by country and sex for 1990-2019, and is usable as the cross-country comparison backbone for this pack.",
24+
"evidence_type": "peer-reviewed-study",
25+
"source": {
26+
"title": "Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019: a pooled analysis of 1201 population-representative studies with 104 million participants",
27+
"url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8446938/"
28+
},
29+
"source_date": "2021-09-11",
30+
"access_date": "2026-06-18",
31+
"method": "Read the open-access NCD-RisC article and its methods section; confirmed it uses only measured (not self-reported) blood pressure and reports the full detection-treatment-control cascade by country, sex, and year.",
32+
"limitations": [
33+
"Estimates are at national level for ages 30-79 only; the paper does not provide sub-national grain.",
34+
"Where country data are sparse the cascade values are modeled, widening uncertainty for many low-income countries.",
35+
"Cut-off is 2019, so it predates recent program scale-up."
36+
],
37+
"confidence": "high"
38+
},
39+
{
40+
"id": "who-steps-microdata",
41+
"problem_id": "public-health/hypertension-control-global",
42+
"claim": "The WHO NCD Microdata Repository distributes nationally representative STEPS survey microdata with directly measured blood pressure, supporting country-level detection, treatment, and control estimates and limited sub-national breakdowns where sample design allows.",
43+
"evidence_type": "dataset",
44+
"source": {
45+
"title": "WHO NCD Microdata Repository — STEPS collection",
46+
"url": "https://extranet.who.int/ncdsmicrodata/index.php/collections/STEPS"
47+
},
48+
"source_date": "2024-01-01",
49+
"access_date": "2026-06-18",
50+
"method": "Opened the STEPS collection in the WHO NCD Microdata Repository and confirmed it lists country STEPS datasets with measured blood pressure variables and downloadable microdata under request-based access.",
51+
"limitations": [
52+
"Surveys are cross-sectional and run infrequently, so a country may have only one usable year.",
53+
"Sampling is designed for national estimates; sub-national power is uneven and often insufficient.",
54+
"Field BP measurement protocol (single visit, device) differs from clinical diagnosis, so control status is approximate.",
55+
"Source date is approximate — the collection is continuously updated rather than published once."
56+
],
57+
"confidence": "high"
58+
},
59+
{
60+
"id": "dhs-program-biomarkers",
61+
"problem_id": "public-health/hypertension-control-global",
62+
"claim": "The DHS Program measures blood pressure as a biomarker for women and men aged 15 and over in selected surveys, providing an additional source of measured-BP prevalence with georeferenced clusters in some LMICs.",
63+
"evidence_type": "dataset",
64+
"source": {
65+
"title": "The DHS Program — Biomarkers",
66+
"url": "https://www.dhsprogram.com/What-We-Do/Biomarkers.cfm"
67+
},
68+
"source_date": "2024-01-01",
69+
"access_date": "2026-06-18",
70+
"method": "Reviewed the DHS biomarkers page and confirmed blood pressure is among collected biomarkers for ages 15+; cross-checked that BP is a recent and survey-specific addition, not universal across all DHS rounds.",
71+
"limitations": [
72+
"Blood pressure is collected only in a subset of DHS surveys and countries, not all rounds.",
73+
"Treatment and control cascade items are not consistently captured, so DHS is stronger for prevalence than for control.",
74+
"Source date is approximate — this is a continuously maintained program page, not a dated publication."
75+
],
76+
"confidence": "medium"
77+
},
78+
{
79+
"id": "gbd-2021-high-sbp-burden",
80+
"problem_id": "public-health/hypertension-control-global",
81+
"claim": "GBD 2021 (IHME) provides modeled estimates of deaths and DALYs attributable to high systolic blood pressure by country, sex, age, and selected subnational units for 1990-2021, usable as a burden-context layer but not as a direct detection-treatment-control measurement.",
82+
"evidence_type": "dataset",
83+
"source": {
84+
"title": "Global Burden of Disease Study 2021 (GBD 2021) Burden and Strength of Evidence by Risk Factor 1990-2021",
85+
"url": "https://ghdx.healthdata.org/record/ihme-data/gbd-2021-burden-by-risk-1990-2021"
86+
},
87+
"source_date": "2024-05-18",
88+
"access_date": "2026-06-18",
89+
"method": "Opened the GHDx record for the GBD 2021 risk-factor burden product and confirmed it covers high systolic blood pressure attributable burden with subnational coverage for a subset of countries.",
90+
"limitations": [
91+
"Outputs are attributable burden (deaths, DALYs), not measured detection, treatment, or control rates.",
92+
"Values are modeled estimates with uncertainty intervals, not survey observations.",
93+
"Subnational estimates exist only for a subset of countries."
94+
],
95+
"confidence": "medium"
96+
},
97+
{
98+
"id": "pure-study-hypertension-2013",
99+
"problem_id": "public-health/hypertension-control-global",
100+
"claim": "The PURE study reported hypertension prevalence, awareness, treatment, and control across high-, middle-, and low-income countries using measured blood pressure in 142042 adults aged 35-70 from 628 communities, documenting markedly lower control in lower-income settings.",
101+
"evidence_type": "peer-reviewed-study",
102+
"source": {
103+
"title": "Prevalence, Awareness, Treatment, and Control of Hypertension in Rural and Urban Communities in High-, Middle-, and Low-Income Countries (PURE)",
104+
"url": "https://pubmed.ncbi.nlm.nih.gov/24002282/"
105+
},
106+
"source_date": "2013-09-04",
107+
"access_date": "2026-06-18",
108+
"method": "Read the PubMed record and abstract to confirm cohort size, age range, country mix, and that the cascade was based on measured blood pressure.",
109+
"limitations": [
110+
"Sample is community cohorts, not nationally representative, so country estimates are not generalizable.",
111+
"Age range is restricted to 35-70 years.",
112+
"Data were collected 2003-2009, so figures are dated relative to current control programs."
113+
],
114+
"confidence": "medium"
115+
},
116+
{
117+
"id": "self-report-hypertension-bias-2022",
118+
"problem_id": "public-health/hypertension-control-global",
119+
"claim": "Self-reported hypertension diagnosis underestimates measured prevalence by roughly 20 percent in women and 55 percent in men in a large Indian survey, which is the documented basis for rejecting self-report-only household surveys as a prevalence source for this pack.",
120+
"evidence_type": "peer-reviewed-study",
121+
"source": {
122+
"title": "Inconsistency in prevalence of hypertension based on self-reports and use of standard tests: Implications for large scale surveys",
123+
"url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC9547289/"
124+
},
125+
"source_date": "2022-10-03",
126+
"access_date": "2026-06-18",
127+
"method": "Read the open-access article and its results to confirm the direction and magnitude of self-report bias against standard measured tests, then applied it as the rejection rule for self-report-only sources.",
128+
"limitations": [
129+
"Quantified bias is from India and may differ in magnitude elsewhere.",
130+
"The study justifies rejecting self-report for prevalence; it is not itself a usable cross-country data source.",
131+
"It does not address treatment or control measurement specifically."
132+
],
133+
"confidence": "high"
19134
}
20135
]

‎problem-packs/public-health/hypertension-control-global/evidence.md‎

Lines changed: 24 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -6,6 +6,30 @@
66

77
WHO estimates 1.3 billion adults have hypertension, with only 45 percent diagnosed and 20 percent controlled globally, and control rates below 10 percent in most LMICs.
88

9+
### ncd-risc-hypertension-trends-2021
10+
11+
NCD-RisC pooled 1201 measured-BP studies across 184 countries to estimate the detection-treatment-control cascade for ages 30-79 by country and sex, 1990-2019. Usable as the cross-country backbone; national grain only and modeled where data are sparse.
12+
13+
### who-steps-microdata
14+
15+
The WHO NCD Microdata Repository distributes STEPS survey microdata with directly measured blood pressure. Usable for country cascade estimates; sub-national power is uneven and surveys are infrequent.
16+
17+
### dhs-program-biomarkers
18+
19+
The DHS Program measures blood pressure for ages 15+ in selected surveys, with georeferenced clusters. Limited: BP coverage is survey-specific and treatment/control items are inconsistent.
20+
21+
### gbd-2021-high-sbp-burden
22+
23+
GBD 2021 (IHME) supplies modeled deaths and DALYs attributable to high systolic blood pressure, with some subnational coverage. Limited: it is attributable burden, not a measured detection-treatment-control cascade.
24+
25+
### pure-study-hypertension-2013
26+
27+
PURE reported the measured-BP cascade across income groups in 142042 adults aged 35-70 from 628 communities. Limited: non-representative community cohorts, restricted ages, data from 2003-2009.
28+
29+
### self-report-hypertension-bias-2022
30+
31+
A large Indian comparison found self-report underestimates measured hypertension by ~20% (women) to ~55% (men). Basis for rejecting self-report-only household surveys as a prevalence source.
32+
933
## Evidence Quality Rule
1034

1135
Evidence accepted only when source, method, limitations, and confidence are explicit enough for a reviewer to attack.

0 commit comments

Comments
 (0)