Table 3

Characteristics and quality assessment of included studies

AuthorYear of publicationCountry study was conductedType of study, sample size, genderOutcomeExposure characteristicsEffect size/measure (aOR or
p value)		Overall quality and capacity to determine if cell spatial density is associated with outcome*
Cell spatial density values analysed (m2 per person)Cell design type
Hussain et al 36 2003PakistanCross-sectional study, 425, 100% maleMycobacterium
tuberculosis		>5.6 vs ≤5.6Barracks (term not defined and may refer to single and/or multiple person cells aOR 2.6; 95% CI 1.6 to 4.3 Fair
Medium to high risk of:

  • exposure misclassification

  • confounding

Potential for:

  • recall bias

Aguilera et al 39 2016ChileCross-sectional study, 418, 77% maleLatent tuberculosis infectionUnclear if continuous or dichotomised variable used.Not reported aOR 3.5; 95% CI: 1.1 to 11.5) Fair
Medium to high risk of:

  • exposure misclassification

Potential for:

  • recall bias

Hoge et al 35 1994USACohort study, 46, 92% malePneumococcal disease2.9 vs (4.2 and 2.6);Single versus (4 persons and dormitory) aOR 2.0; 95% CI 1.1 to 3.8 Poor to fair
Medium to high risk of:

  • exposure misclassification

  • associations due to chance

Potential for:

  • recall bias

  • confounding

13.0 vs (2.9, 4.2 and 2.6);Single versus (single, 4 persons and dormitory) P<0.001
6.8 vs (2.9, 4.2 and 2.6)Not reported versus (single, 4 persons and dormitory) P<0.001
Oninla and Onayemi37 2012NigeriaCross-sectional study, 305, 97% maleInfectious dermatoses0.9 vs 2.4Single versus dormitories P=0.03 Poor
High risk of:

  • associations due to chance

  • confounding

  • selection bias

Potential for:

  • recall bias

  • exposure misclassification

Oninla et al 38 2013NigeriaCross-sectional study, 305, 97% maleInfectious and non-infectious dermatoses0.9 vs 2.4Single versus dormitories P<0.001 Poor
High risk of:

  • associations due to chance

  • confounding

  • selection bias

Potential for:

  • recall bias

  • exposure misclassification

McCain et al 27 1980USACross-sectional study, 289, 100% maleContagious illness reporting at clinic4.5/5.6 vs 4.6/5.5Singles/singles versus cubicles/dormitories P<0.05 Fair
Medium to high risk of:

  • outcome misclassification

Potential for:

  • associations by chance

  • confounding

  • recall bias

  • selection bias

Gaes40 1982USACross-sectional study, 352, 100% maleContagious illness reporting at clinic4.0 to 8.2Singles and cubiclesnsPoor to fair
Medium to high risk of:

  • outcome misclassification

Potential for:

  • confounding

  • selection bias

  • *Assessment guided by the NHMRC’s checklist to critically appraise aetiology or risk factor studies.32 Checklist items used to guide the assessment included: exposure misclassification, outcome misclassification, selection bias, confounding and chance.

  • Bold, statistical significance; ns, not statistically significant.

  • aOR, adjusted odds ratio.