Skip to main content

Table 1 Study characteristics and results of studies investigating the association between urban and rural built environments and diabetes mellitus

From: Built environmental characteristics and diabetes: a systematic review and meta-analysis

Author

Year

Country

Country income level

Study design

Sample size

Age

Outcomea

Outcome assessmentb

Result

Adjustment for confounding

Quality statement

Urban > rural

Rural > urban

No difference

Aekplakorn et al. [89]

2011

Thailand

Upper middle

Cross-sectional

18,629

NFG: 44.3 ± 0.3 Diabetes mellitus: 54.1 ± 0.7

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex

Moderate

Agyemang et al. [90]

2016

Ghana, Netherlands, Germany, England

Lower middle and high

Cross-sectional

5659

25–70 years (NR)

T2DM prevalence

Blood sample

X

  

Age, sex, education

Moderate

Ali et al. [91]

1993

Malaysia

Upper middle

Cross-sectional

681

38.6 ± 13.7

T2DM/T1DM prevalence

Blood sample

  

X

Age

Moderate

Al-Moosa et al. [92]

2006

Oman

High

Cross-sectional

5840

24% >50 years 41% < 30 years

T2DM/T1DM prevalence

Blood sample

X

  

–

Moderate

Anjana et al. [93]

2011

India

Lower middle

Cross-sectional

13,055

40 ± 14

T2DM/T1DM prevalence

Blood sample

Southern area, western area, eastern area

 

Northern area

Age, sex

Moderate

Assah et al. [94]

2011

Cameroon

Lower middle

Cross-sectional

552

38.4 ± 8.6

T2DM/T1DM prevalence

Blood sample

X

  

–

Moderate

Attard et al. [67]

2012

China

Upper middle

Cross-sectional

NA

51 ± 0.4

T2DM/T1DM prevalence

Blood sample, self-report

X

  

Age, sex, income, region, BMI

Strong

Allender et al. [95]

2011

Sri Lanka

Lower middle

Cross-sectional

4485

46.1 ± 15.1

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex, income

Moderate

Bahendeka et al. [41]

2016

Uganda

Low

Cross-sectional

3689

35.1 ± 12.6

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex, region of residence, floor finishing of dwelling, BMI, waist circumference, total cholesterol

Moderate

Baldé et al. [96]

2007

Guinea

Low

Cross-sectional

1537

47.7 ± 12.5

T2DM/T1DM prevalence

Blood sample

X

  

Age, location, excess of waist, raised systolic BP, raised diastolic BP

Moderate

Balogun et al. [97]

2012

Nigeria

Lower middle

Longitudinal

1330

77.3 ± 0.3

T2DM incidence

Self-report

X

  

Age, sex, education

Strong

Baltazar et al. [98]

2003

Philippines

Lower middle

Cross-sectional

7044

39.0 ± 0.5

T2DM/T1DM prevalence

Blood sample

  

X

Age and sex

Moderate

Barnabé-Ortiz [99]

2016

Peru

Upper middle

Longitudinal

3123

24% < 45 years 25% >65 years

T2DM incidence

Blood sample

  

X

Sex, age, education level, SES, family history of diabetes, daily smoking, hazardous drinking, TV watching for 2+ hours per day, transport-related physical inactivity, fruit and vegetable consumption, BMI, metabolic syndrome

Moderate

Bocquier et al. [100]

2010

France

High

Cross-sectional

3,038,670

48.9 ± 18.6

T2DM/T1DM prevalence

Secondary

X

  

Age, sex

Strong

Cubbin et al. [23]

2006

Sweden

High

Cross-sectional

18,081

48% >45 years 25% < 35 years

T2DM/T1DM prevalence

Self-report

 

X

 

Age, sex, marital status, immigration status, SES composite, neighbourhood deprivation

Moderate

Christensen et al. [101]

2009

Kenya

Lower middle

Cross-sectional

1459

38.6 ± 12.6

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex

Moderate

Dagenais et al. [102]

2016

Bangladesh, India, Pakistan, Zimbabwe, China, Colombia, Iran, Argentina, Brazil, Chile, Malaysia, Poland, South Africa, Turkey, Canada, Sweden, United Arab Emirates

Lower, lower middle, upper middle and high

Cross-sectional

119,666

52 ± 9.3

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex, residency location, BMI, waist-to-hip ratio, PA levels, AHEI score, combined former and current smoking, education level, family history of diabetes, ethnicity

Strong

Dar et al. [25]

2015

India

Lower middle

Cross-sectional

3972

43% >50 years 57% 40–50 years

T2DM prevalence

Blood sample

X

  

–

Weak

Davila et al. [103]

2013

Colombia

Upper middle

Cross-sectional

1026

35% >55 years 35% < 35 years

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex, education, SES, marital status, smoking, alcohol, intake of fruit and vegetables, PA

Strong

Delisle et al. [104]

2012

Benin

Low

Cross-sectional

541

38.2 ± 0.6

Glycaemic marker: HOMA index

Blood sample

X

  

Age, sex, SES, location, diet quality, PA, alcohol, BMI

Moderate

Dong et al. [105]

2005

China

Upper middle

Cross-sectional

12,240

46.4 ± 13.9

T2DM prevalence

Blood sample

X (men)

 

X (women)

Age, sex

Moderate

Du et al. [106]

2016

China

Upper middle

Cross-sectional

3797

15% >60 years 8% 20–29 years

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex

Moderate

Esteghamati et al. [107]

2009

Iran

Upper middle

Cross-sectional

3397

23% >55 years 25% < 35 years

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex, residential area

Moderate

Georgousopoulou et al. [108]

2017

Mediterranean islands

High

Cross-sectional

2749

75 ± 7.3

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex, BMI, physical inactivity, smoking, siesta habit, education, living alone, adherence to Mediterranean diet, GDS, number of friends and family members, frequency of going out with friends and family, number of holiday excursions per year

Moderate

Gong et al. [109]

2015

China

Upper middle

Cross-sectional

5923

38% >50 years 62% < 50 years

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex, education, PA, smoking, alcohol, BMI, triglycerides, HDL-cholesterol, hypertension

Strong

Hussain et al. [110]

2004

Bangladesh

Lower middle

Cross-sectional

6312

14% >50 years 46% < 30 years

T2DM/T1DM prevalence

Blood sample

X

  

Age, sex

Moderate

Han et al. [111]

2017

Korea

High

Longitudinal

7542

52 ± 8.8

T2DM incidence

Blood sample

X

  

Age, sex, residential area, family history of diabetes, smoking, alcohol, exercise, abdominal obesity, hypertension, high triglycerides, low HDL-cholesterol

Strong

Katchunga et al. [112]

2012

Congo

Low

Cross-sectional

699

42.5 ± 18.1

T2DM/T1DM prevalence

Blood sample

  

X

–

Moderate

Keel et al. [113]

2017

Australia

High

Cross-sectional

4836

Non-indigenous: 66.6 ± 9.7 Indigenous: 54.9 ± 8.7

T2DM/T1DM prevalence

Self-report

 

X (indigenous)

X (non-indigenous)

Age, sex, ethnicity, education, English-speaking at home, ethnicity

Moderate

Mayega et al. [114]

2013

Uganda

Low

Cross-sectional

1497

45.8% >45 years 54.2% < 45 years

T2DM prevalence

Blood sample

 

X

 

Age, sex, residence, occupation, family history of diabetes, BMI, PA level, dietary diversity

Strong

Mohan et al. [115]

2016

India

Lower middle

Cross-sectional

6853

35–70 years (NR)

T2DM/T1DM prevalence

Blood sample

X

  

Age (only women included)

Moderate

Msyamboza et al. [116]

2014

Malawi

Low

Cross-sectional

3056

12.5% >55 years 45% < 35 years

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex

Moderate

Ntandou et al. [117]

2009

Benin

Low

Cross-sectional

541

38.2 ± 10

T2DM/T1DM prevalence

Blood sample

 

X

 

Age, sex, waist circumference, education, SES, PA, micronutrient adequacy score, preventive diet score, alcohol

Moderate

Oyebode et al. [118]

2015

China, Ghana, India, Mexico, Russia, South Africa

Upper and Lower middle

Cross-sectional

39,436

47.3% >60 years 12.3% < 40Y

T2DM/T1DM prevalence

Self-report

X (pooled)

  

Age, sex, survey design, income quintile, marital status, education

Strong

Papoz et al. [119]

1996

New Caledonia

High

Cross-sectional

9390

30–59 years (NR)

T2DM/T1DM prevalence

Blood sample

X

  

Age

Moderate

Pham et al. [120]

2016

Vietnam

Lower middle

Cross-sectional

16,730

54 ± 8

T2DM/T1DM prevalence

Blood sample

X (men)

 

X (women)

Age, sex, socio-demographic factors, anthropometric measures, BP, family history of diabetes

Moderate

Raghupathy et al. [121]

2007

India

Lower middle

Longitudinal

2218

28 ± 1.2

T2DM prevalence

Blood sample

X

  

Age, sex, number of household possessions, education, PA, smoking, alcohol, parental consanguinity, family history of diabetes mellitus, body fat, BMI, waist-to-hips ratio, subscapular/triceps ratio

Strong

Ramdani et al. [122]

2012

Morocco

Lower middle

Cross-sectional

1628

54.2 ± 10.9

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex, BMI

Moderate

Sadikot et al. [123]

2004

India

Lower middle

Cross-sectional

41,270

36% >50 years 34% < 40 years

T2DM prevalence

Blood sample

X

  

Age, sex

Moderate

Sobngwi et al. [124]

2004

Cameroon

Lower middle

Longitudinal

1726

24% >55 years 28% < 35 years

T2DM/T1DM prevalence

Blood sample

X (women)

 

X (men)

Age, sex, residence, socio-professional category, alcohol, smoking, PA

Moderate

Stanifer et al. [125]

2016

Tanzania

Low

Cross-sectional

481 neighbourhoods

25% >60 years

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex

Moderate

Weng et al. [126]

2007

China

Upper middle

Cross-sectional

529

NR

T2DM/T1DM prevalence

Blood sample

  

X

Age, sex

Moderate

Wu et al. [127]

2016

China

Upper middle

Cross-sectional

23,010

40 (30.4–56.3)

T2DM/T1DM prevalence

Blood sample

  

X

Age

Moderate

Zhou et al. [128]

2015

China

Upper middle

Cross-sectional

98,658

20% >60 years 80% < 60 years

T2DM/T1DM prevalence

 

X

  

Age, sex, region

Moderate

  1. BMI body mass index, BP blood pressure, NR not recorded, PA physical activity, SES socioeconomic status, T1DM type 1 diabetes mellitus, T2DM type 2 diabetes mellitus, NFG normal fasting glucose, HOMA homeostasis model assessment, GDS geriatric depression scale 
  2. aPrevalence indicates incidence or glycaemic marker level
  3. bBlood sample: study diagnosed diabetes based on glycaemic marker or oral glucose tolerance test; secondary: from data sources such as national health survey; self-report: ever diagnosed with diabetes