Using Multilevel Regression and Poststratification to Estimate Physical Activity Levels from Health Surveys.
Bayesian analysis
public health surveillance
selection bias
statistical models
survey methods
Journal
International journal of environmental research and public health
ISSN: 1660-4601
Titre abrégé: Int J Environ Res Public Health
Pays: Switzerland
ID NLM: 101238455
Informations de publication
Date de publication:
13 07 2021
13 07 2021
Historique:
received:
06
05
2021
revised:
06
07
2021
accepted:
08
07
2021
entrez:
24
7
2021
pubmed:
25
7
2021
medline:
12
8
2021
Statut:
epublish
Résumé
Large-scale health surveys often consider sociodemographic characteristics and several health indicators influencing physical activity that often vary across subpopulations. Data in a survey for some small subpopulations are often not representative of the larger population. We developed a multilevel regression and poststratification (MRP) model to estimate leisure-time physical activity across Brazilian state capitals and evaluated whether the MRP outperforms single-level regression estimates based on the Brazilian cross-sectional national survey VIGITEL (2018). We used various approaches to compare the MRP and single-level model (complete-pooling) estimates, including cross-validation with various subsample proportions tested. MRP consistently had predictions closer to the estimation target than single-level regression estimations. The mean absolute errors were smaller for the MRP estimates than single-level regression estimates with smaller sample sizes. MRP presented substantially smaller uncertainty estimates compared to single-level regression estimates. Overall, the MRP was superior to single-level regression estimates, particularly with smaller sample sizes, yielding smaller errors and more accurate estimates. The MRP is a promising strategy to predict subpopulations' physical activity indicators from large surveys. The observations present in this study highlight the need for further research, which could, potentially, incorporate more information in the models to better interpret interactions and types of activities across target populations.
Sections du résumé
BACKGROUND
Large-scale health surveys often consider sociodemographic characteristics and several health indicators influencing physical activity that often vary across subpopulations. Data in a survey for some small subpopulations are often not representative of the larger population.
OBJECTIVE
We developed a multilevel regression and poststratification (MRP) model to estimate leisure-time physical activity across Brazilian state capitals and evaluated whether the MRP outperforms single-level regression estimates based on the Brazilian cross-sectional national survey VIGITEL (2018).
METHODS
We used various approaches to compare the MRP and single-level model (complete-pooling) estimates, including cross-validation with various subsample proportions tested.
RESULTS
MRP consistently had predictions closer to the estimation target than single-level regression estimations. The mean absolute errors were smaller for the MRP estimates than single-level regression estimates with smaller sample sizes. MRP presented substantially smaller uncertainty estimates compared to single-level regression estimates. Overall, the MRP was superior to single-level regression estimates, particularly with smaller sample sizes, yielding smaller errors and more accurate estimates.
CONCLUSION
The MRP is a promising strategy to predict subpopulations' physical activity indicators from large surveys. The observations present in this study highlight the need for further research, which could, potentially, incorporate more information in the models to better interpret interactions and types of activities across target populations.
Identifiants
pubmed: 34299923
pii: ijerph18147477
doi: 10.3390/ijerph18147477
pmc: PMC8304573
pii:
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Références
Epidemiol Serv Saude. 2017 Jul-Sep;26(3):605-616
pubmed: 28977184
Prev Med Rep. 2019 Nov 12;16:101000
pubmed: 31799104
Lancet. 2006 Jul 22;368(9532):299-304
pubmed: 16860699
J Sport Health Sci. 2021 Jan;10(1):55-64
pubmed: 32315605
J Phys Act Health. 2015 Mar;12(3):355-60
pubmed: 24828876
PLoS One. 2016 Jan 28;11(1):e0148038
pubmed: 26820144
Diabetologia. 2007 Sep;50(9):1832-1840
pubmed: 17641870
Front Med (Lausanne). 2018 Jan 09;4:244
pubmed: 29376052
Gerontology. 2017;63(5):443-459
pubmed: 28675889
Med Sci Sports Exerc. 1992 Jun;24(6 Suppl):S221-36
pubmed: 1625548
Am J Epidemiol. 2018 Aug 1;187(8):1780-1790
pubmed: 29635276
Br J Sports Med. 2018 Feb;52(3):149-150
pubmed: 28798040
Lancet. 2016 Sep 24;388(10051):1257-8
pubmed: 27475268
Int J Behav Nutr Phys Act. 2020 Mar 18;17(1):38
pubmed: 32183834
PLoS One. 2017 Aug 10;12(8):e0180698
pubmed: 28797037
Med Sci Sports Exerc. 2004 Jan;36(1):86-92
pubmed: 14707773
Med Sci Sports Exerc. 2013 Nov;45(11):2098-104
pubmed: 23591292
PLoS One. 2020 Jan 30;15(1):e0228373
pubmed: 31999792
Int J Behav Nutr Phys Act. 2019 Aug 20;16(1):68
pubmed: 31429772
Am J Public Health. 1998 Feb;88(2):216-22
pubmed: 9491010
Int J Epidemiol. 2001 Dec;30(6):1343-50
pubmed: 11821344
Cad Saude Publica. 2019 Oct 31;35(11):e00016319
pubmed: 31691775
Int J Public Health. 2020 Nov;65(8):1467-1476
pubmed: 32918552
J Sport Health Sci. 2022 Mar;11(2):234-243
pubmed: 33737239
Int J Public Health. 2014 Feb;59(1):197-206
pubmed: 23619721
Rev Saude Publica. 2017 Jun 01;51(suppl 1):15s
pubmed: 28591355
Am J Epidemiol. 2015 Jul 15;182(2):127-37
pubmed: 25957312
Epidemiol Serv Saude. 2020 Oct 05;29(5):e2020315
pubmed: 33027428
J Phys Act Health. 2018 Apr 1;15(4):303-314
pubmed: 29278043
Cien Saude Colet. 2014 Feb;19(2):333-42
pubmed: 24863810
Br J Sports Med. 2020 Dec;54(24):1451-1462
pubmed: 33239350
J Dent Res. 2016 May;95(5):515-22
pubmed: 26848071
J Stat Softw. 2017;76:
pubmed: 36568334
Int J Adolesc Med Health. 2018 Feb 12;31(5):
pubmed: 29432203