Table 1 Description and the results of articles reviewed

Wong TH (2017) [26]

Retrospective cohort

From 1992 to 2014

Survival of patients with head and neck squamous cell carcinoma (HNSCC).

Patients with HNSCC in National Cancer Center database. Mortality information matched from Singapore Death Registry. Non-residents were excluded from analysis.

Sample size

686 patients were analysed, 84 (11%) of them were rental housing residents.

Primary outcome

• All-cause mortality:

Patients living in postal codes with a room index < 3.0* had the worst survival [median, 28 months, CI 21–48 months] compared to those staying in larger housing sizes (higher room index) and owner occupied.

Secondary outcome

Disease stage at presentation:

Patients living in a lower room index postal code were not more likely to present with advanced disease.

*majority of residents lived in rental housings with a monthly household income <S$1500

Patients with HNSCC living in smaller, higher-subsidy housings have poorer survival despite no apparent delays in presentation.

Poor

Wee LE (2013) [13]

Prospective, Interventional

From 2009 to 2011

Hypertension management and lifestyle changes following screening for hypertension.

Residents > 40 y/o.

two public rental housing precincts.

Site A (Western Singapore) and Site B (Eastern Singapore).

Intervention:

6-month community-based intervention comprising access-enhanced screening component and follow-up (outreach) component.

Sample size

577 residents

Participation rate

83.2% (577/693)

Follow up rate (follow up for 1 year)

80.9% (467/557)

Prevalence rate

Baseline: 60.4% (282/467)

Known hypertension n = 179, Newly diagnosed n = 103

Primary outcome

Untreated and uncontrolled hypertension

• Baseline:

Known hypertension and not treated n = 48

Untreated hypertension 53.5% (151/282)

Known hypertension and treated n = 131

Uncontrolled hypertension despite treated 54.2% (71/131)

• Post intervention:

Untreated hypertension 48.3% (73/151)

Started treatment n = 78

Uncontrolled hypertension 47.0% (70/149)

Secondary outcome BP screening

• Baseline:

No hypertension n = 185

Did not screen 52.4% (97/185)

• Post intervention

Did not screen 31.9% (57/185)

Reason for not going for screening and starting treatment.

Cost (for test and further treatment) and misperceptions were common barriers.

An access-enhanced intervention had some success in improving hypertension management; however, it was less successful in improving cardiovascular risk management, amongst newly diagnosed hypertensives in the rental housing community.

–

Wee LE (2012) [14]

Prospective and Qualitative

From Jan 2009 to June 2010

Individual and neighbour-hood social factors of hypertension management.

Residents ≥40 y/o.

6 blocks of a socially integrated housing precinct.

3 blocks public rental flats vs 3 blocks of owner-occupied public housing flats. (3 vs 3)

Located in Taman Jurong (Western Singapore).

Sample size

710 residents

Participation rate

78.9% (710/900)

Rental: 90.0% (359/400) VS Owned: 70.2% (351/500)

Prevalence rate

64.2 (456/710)

Rental: 63.5% (228/359) VS Owned: 65.0% (228/351)

Follow up rate: NA

Primary outcome

• Awareness

Rental: 61.8% (141/228) VS Owned: 83.3% (190/228)

• On treatment

Rental: 69.5% (98/141) VS Owned: 85.3% (162/190)

• BP under control

Rental: 43.9% (43/98) VS Owned: 66% (107/162)

Secondary outcome

Independent factors associated with hypertension awareness, treatment and control in rental housing (lower SES) community.

• Awareness higher among:

Diabetics (adj OR 6.51, CI 2.59–16.37, p < 0.001)

Dyslipidemics (adj OR 6.74, CI 2.74–16.59, p < 0.001)

≥60 years (adj OR 3.08, CI 1.61–5.91, p < 0.001)

Regular access to a doctor (adj OR 5.63, CI 1.43–22.14, p < 0.013).

• Treatment more likely among:

≥60 years (adj OR 2.33, CI 1.08–5.01, p = 0.031)

Treatment less likely among:

Need financial aid (adj OR 0.39, CI 0.18–0.83, p = 0.016).

• Controlled BP less likely among:

Employed (adj OR 0.13, CI 0.04–0.41, p < 0.001).

Hypertension management (awareness, treatment and control) in those of rental housing community (lower SES) is poorer than in those of owner occupied housing community (higher SES).

In rental housing community, awareness was higher among those with diabetes, dyslipidaemia, those ≥60 years and those with regular access to doctors.

Treatment was more likely among those ≥60 years, but less likely among those needing financial aid. Control was less likely in the employed.

–

Qualitative interview:

Reasons for not going on hypertensive treatment or participating in regular hypertension screening/monitoring were explored.

In the rental housing community:

Unknown hypertensives who did not go for regular BP screening in the past 1 year (n = 141), the top three reasons:

Too busy to go/no time.

Testing too expensive.

Cost of further treatment, if positive, too expensive.

Known hypertensives who were not monitoring their BP regularly (n = 64), the top three reasons were similar:

Too busy to go/no time.

Monitoring too expensive.

Cost of further treatment, if positive, too expensive.

Reasons for not taking BP medications ≥90% of the time among known hypertensives (n = 43):

30.2% (13/43) did not think that the medicine would benefit them.

25.6% (22/43) had problems with the cost of chronic medication.

11.6% (5/43) preferred to take non-Western medication.

Financial barriers need to be addressed for the rental housing community.

Wee LE (2014) [17]

Cross-sectional

From Jan to Feb 2012.

Individual and area-level socio-economic status and their association with depression.

(GDS-15 ≥ 5)

Residents ≥60 y/o.

2 integrated public housing precinct.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(7 vs 2)

Sample size

559 residents

Participation rate

61.5% (559/909)

Site A: 61.3% (236/385) VS Site B: 61.6% (323/524)

Rental: 63.7% (398/625) VS Owned: 56.7%, (161/284) (p = 0.0473)

Prevalence rate

22.9% (128/559)

Rental: 26.2% (104/397) VS Owned: 14.8% (24/164)

Follow up rate: NA

Primary outcome

Prevalence rate as above.

Secondary outcome

Living in a rental housing (lower SES) community was independently associated with depression (adj OR 1.68, CI 1.02–2.84, p = 0.049]

Not being married (adj OR 2.27, CI 1.35–3.70), falls (adj OR 2.72, CI 1.59–4.67) and poorer social network (adj OR 3.70, CI 1.96–7.14) were associated with depression.

Other independent factors associated with depression in rental housing community:

Falls (adj OR 2.72, CI 1.59–4.67, p < 0.001)

Visual impairment (adj OR 2.37, CI 1.28–4.39. p = 0.006)

Residing in rental housing was independently associated with depression after controlling for individual SES.

Poor

Wee LE (2012) [18]

Cross-sectional

From Jan to Feb 2012.

Individual and area Level socio-economic status and Its association with cognitive function and cognitive impairment.

(MMSE < 24)

Residents ≥60 y/o.

2 integrated public housing precinct.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(7 vs 2)

Sample size

558 residents

Participation rate

61.4% (558/909)

Site A: 61.0% (235/385) VS Site B: 61.6% (323/524)

Rental: 63.7% (397/625) VS Owned: 56.7%, (161/284) (p = 0.0473)

Prevalence rate

23.3% (130/558)

Rental 26.2% (104/397) VS Owned 16.1% (26/261)

Newly diagnosed in rental housing: 96.2% (100/104)

Follow up rate: NA

Primary outcome

Prevalence rate as above.

Secondary outcome

Living in a rental housing community was independently associated with cognitive impairment (adj OR 5.13, CI 1.98–13.34, p = 0.001).

Living in rental housing is independently associated with cognitive impairment. Many of them with cognitive impairment were undiagnosed prior.

Poor

Wee LE (2016) [19]

Cross-sectional

From 2009 to 2014

In 2012, a separate study in Site A and Site B focused on those aged ≥60 years old was conducted.

Chronic pain in a low socio-economic status population.

Residents 40–59 y/o.

5 integrated public housing precincts.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(4 vs 5)

Site C (Eastern Singapore)

(3 vs 5)

Site D (Eastern Singapore)

(2 vs 1)

Site E (Central Singapore)

(2 vs 1)

Sample size

40–59 y/o. 2037 residents

≥ 60 y/o. 559 residents

Participation rate

40–59 y/o. Rental: 72.0% (936/1300) VS Owned: 61.2% (1101/1800)

≥ 60 y/o. Rental: 63.7% (397/625) VS Owned: 56.7% (162/284)

Prevalence rate

40–59 y/o. Rental: 14.2% (133/936) VS Owned: 14.4% (158/1101)

≥ 60 y/o. Rental: 13.4% (53/397) VS Owned: 13.0% (21/162)

Follow up rate: NA

Primary outcome

Prevalence rate as above.

Secondary outcome

In the rental housing community, unemployment was associated with chronic pain (adj OR 1.92, 95%, CI 1.05–2.78, p = 0.030)

Among the elderly, dependency in instrumental activities of daily living (iADLs) was associated with chronic pain (adj OR 2.38, CI 1.11–5.00, p = 0.025), as well as female gender, being single, and having higher education (all p > 0.05).

There was no difference in pain prevalence between the rental housing community and owner-occupied community.

In rental housing community, chronic pain associated with unemployment and functional limitation.

Poor

Wee LE (2017) [20]

Cross-sectional

and Qualitative

From 2009 to 2014

Health screening participation and its association with chronic pain.

Residents 40–59 y/o.

5 integrated public housing precincts.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(4 vs 5)

Site C (Eastern Singapore)

(3 vs 5)

Site D (Eastern Singapore)

(2 vs 1)

Site E (Central Singapore)

(2 vs 1)

Sample size

40–59 y/o. 2037 residents

Participation rate

40–59 y/o. Rental: 72.0% (936/1300) VS Owned: 61.2% (1101/1800)

Prevalence rate

40–59 y/o. Rental: 14.2% (133/936) VS Owned: 14.4% (158/1101)

Follow up rate: NA

Primary outcome

In the rental-housing community, chronic pain was associated with higher participation in screening for:

Diabetes (adj OR 2.11,CI 1.36–3.27, p < 0.001)

Dyslipidemia (adj OR 2.06, CI 1.25–3.39, p = 0.005)

Colorectal cancer (adj OR 2.28,CI 1.18–4.40, p = 0.014)

Cervical cancer (adj OR 2.65,CI 1.34–5.23, p = 0.005)

Breast cancer (adj OR 3.52,CI 1.94–6.41, p < 0.001)

This association was not present in the owner-occupied community.

Secondary outcome: NA

Chronic pain was associated with higher cardiovascular and cancer screening participation in the rental housing community.

Poor

Qualitative interview:

General attitudes towards screening tests; and how their pain might affect their attitudes to screening participation.

Three main themes emerged from the analysis of the link between chronic pain and screening participation:

Pain as an association of “major illness”.

Screening as a search for answers to pain.

Labelling pain as an end in itself.

To those living in rental housing, disease only occurs when symptoms manifest, such as chronic pain. There is a possibility that chronic pain may present as the “hidden agenda”.

Wee LE (2012) [15]

Prospective, Interventional

From Jan 2009 to May 2011

Screening for cardio-vascular disease risk factors at baseline and post-intervention.

2 integrated public housing precinct.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(7 vs 2)

Intervention:

6-month community-based intervention comprising access-enhanced screening component and follow-up (outreach) component.

Sample size

1081 residents

Participation rate

78.2% (1081/1383)

Site A: 83.4% (832/998) VS Site B: 64.6% (249/385).

Prevalence rate: NA

Follow up rate: NA

Primary outcome

Baseline screening participation for:

Hypertension. Rental 41.7% (150/360) VS Owned 54.1% (139/257)

Diabetes. Rental 38.8% (177/456) VS Owned 59.6% (254/426)

Dyslipidaemia. Rental 30.8% (128/416) VS Owned 50.2% (165/329)

Post intervention:

Hypertension. Rental 99.2% (357/360) VS Owned 96.9% (249/257)

Diabetes. Rental 45.2% (206/456) VS Owned 67.6% (288/426)

Dyslipidaemia. Rental 37.0% (154/416) VS Owned 58.4% (192/329)

Secondary outcome

Living in a rental housing community (adj RR 0.61, CI 0.37–0.99, p = 0.048) and having hypertension (adj RR 0.45, CI 0.18–0.98, p = 0.049) was associated with lower participation in screening for diabetes and dyslipidaemia respectively.

Employment (adj RR 1.57, CI 1.03–2.60, p = 0.040) and Chinese ethnicity (adj RR 1.84, CI 1.00–3.43, p = 0.050) was associated with higher participation in screening for diabetes and dyslipidaemia respectively.

Those living in rental housings had lower participation in screening

before intervention, when compared against those living in owner-occupied housings.

Post intervention, participation rates for all three screening modalities rose significantly by similar proportions in both rental and owner-occupied community (all p < 0.001).

–

Wee LE (2012) [16]

Prospective, Interventional

From Jan 2009 to May 2011

Socio-economic factors affecting colorectal, breast and cervical cancer screening at baseline and post-intervention.

2 integrated public housing precinct.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(7 vs 2)

Intervention:

6-month community-based intervention comprising access-enhanced screening component and follow-up (outreach) component.

Sample size

1081 residents

Participation rate

78.2% (1081/1383)

Site A: 83.4% (832/998) VS Site B: 64.6% (249/385).

Prevalence rate: NA

Follow up rate: NA

Primary outcome

Baseline screening participation for:

Colorectal cancer. Rental 7.7% (33/427) VS Owned 16.6% (66/397)

Cervical cancer. Rental 20.4% (44/216) VS Owned 41.9% (93/222)

Breast cancer. Rental 14.3% (46/321) VS Owned 15.9% (48/302)

Post intervention:

Colorectal cancer. Rental 19.0% (81/427) VS Owned 28.2% (112/397)

Cervical cancer. Rental 25.4% (55/216) VS Owned 47.3% (105/222)

Breast cancer. Rental 17.4% (56/321) VS Owned 16.9% (51/302)

Secondary outcome

Factors associated with cancer screening in rental housing community:

Males (adj OR 2.11, CI 1.01–4.42) and those overweight (adj OR 2.76, CI 1.32–5.75) was associated with higher participation in colorectal cancer screening.

The employed (adj OR 1.56, CI 1.03–2.35) and those of higher educational status (adj OR 1.96, CI 1.27–3.02) was associated with higher participation in breast cancer screening.

Cost was a major factor in the low-SES community, especially for pap smears/mammograms. Misperceptions and lack of time/awareness were also important.

Those living in rental housings had lower participation in colorectal and cervical cancer screening before intervention, when compared against those living in owner-occupied housings.

Post intervention, participation rates rose for most screening modalities in both communities (all p ≤ 0.001), except for breast cancer in the owner-occupied community.

(p = 0.250).

–

Ng CW

(2012) [21]

Cross-sectional

From Jun to Oct 2009

Characteristic associated with non-willingness to participate in health promotion programmes

Residents ≥18 y/o.

4 blocks of 1 to 2 rooms housing estate.

Located in Toa Payoh.

Sample size

974 residents

Participation rate

79.9% (778/974)

Prevalence rate: NA

Follow up rate: NA

Primary outcome

36.1% (281/778) of residents were willing to participate in at least one health promotion programme (health screening, talk or workshop).

Older residents aged 45–64 years (OR 0.52, CI 0.35–0.76, p = 0.001) and more than 65 years (OR 0.44, CI 0.29–0.66, p < 0.001) were less likely to participate than their younger counterparts (18–44 years). Malays (OR 1.84, CI 1.27–2.68, p = 0.001) were more likely than Chinese to participate, and residents who do not exercise (OR 0.57, CI 0.42–0.78, p < 0.001) were less likely to participate than residents who exercise (regularly/occasionally).

Secondary outcome Reasons for non-willingness to participate were ‘not interested’ and ‘no time’.

Residents living in 1 to 2 room housing had low rate of participation in health promotion programme.

Older residents and those who do not exercise had lower rate of participation as well.

Poor

Wee LE (2011) [22]

Cross-sectional,

Interventional

From Jan 2009 to May 2010

The effect of neighbour-hood socio-economic status and a community-based program on multi-disease health screening.

Residents ≥40 y/o.

6 blocks of a socially integrated housing precinct.

(3 vs 3)

Located in Taman Jurong (Western Singapore).

Intervention:

6-month community-based intervention comprising access-enhanced screening component and follow-up (outreach) component.

Sample size

707 residents

Participation rate

78.6% (707/900)

Rental: 89.0% (356/400) VS Owned: 70.2% (351/500)

Prevalence rate: NA

Follow up rate: NA

Primary outcome

Baseline screening participation for:

Hypertension. Rental 35.8% (77/215) VS Owned 52.2% (84/161)

Diabetes. Rental 35.0% (98/280) VS Owned 66.0% (190/288)

Dyslipidaemia. Rental 26.2% (70/267) VS Owned 53.1% (119/224)

Colorectal cancer. Rental 6.0% (15/251) VS Owned 17.0% (49/288)

Post intervention:

Hypertension. Rental 98.6% (212/215) VS Owned 100.0% (161/161)

Diabetes. Rental 40.0% (112/280) VS Owned 66.7% (192/288)

Dyslipidaemia. Rental 30.3% (81/267) VS Owned 54.0% (121/224)

Colorectal cancer. Rental 16.3% (41/251) VS Owned 18.7% (54/288)

Secondary outcome

Living in a better-off neighbourhood was independently associated with diabetes mellitus (66% vs. 35%, adj OR 2.12, p < 0.01), hyperlipidemia (53% vs. 26%, adj OR 4.34, p < 0.01) and colorectal cancer screening (17% vs. 6%, adj OR 15.43, p < 0.01), as were individual socioeconomic factors such as employment, need for financial aid and household income.

Cost was cited more commonly as a barrier to health screening in the rental housing community.

Reasons for not participating in screening in both community, and for a majority of modalities:

Misperceptions

Lack of time

Uptake of all screening modalities significantly increased in the rental housing community post-intervention (all p < 0.05).

Poor

Wee LE (2016) [23]

Cross-sectional

and Qualitative

From 2009 to 2014

Primary care characteristic and their association with health screening.

Residents 40–59 y/o.

5 integrated public housing precincts.

Site A (Western Singapore)

(3 vs 3)

Site B (Eastern Singapore)

(4 vs 5)

Site C (Eastern Singapore)

(3 vs 5)

Site D (Eastern Singapore)

(2 vs 1)

Site E (Central Singapore)

(2 vs 1)

Sample size

1996 residents

Participation rate

64.4% (1996/3100)

Rental: 72.0% (936/1300) VS Owned: 58.9% (1060/1800)

Prevalence rate: NA

Follow up rate: NA

Primary outcome

Rental:

Regular primary care was independently associated with regular:

Diabetes screening (adj OR 1.59, CI 1.12–2.26, p = 0.009).

Hyperlipidemia screening (adj OR 1.82, CI 1.10–3.04, p = 0.023).

Proximity to primary care was associated with less participation in regular:

Colorectal cancer screening (adj OR 0.42, CI 0.17–0.99, p = 0.049)

Breast cancer screening (adj OR = 0.29, CI 0.10–0.84, p = 0.023).

Usage of subsidized primary care was only associated with increased

participation in regular:

Breast cancer screening (adj OR 2.33, CI 1.23–4.41, p = 0.009).

Owned:

Regular primary care was independently associated with regular:

Hypertension screening (adj OR 9.34 CI 1.82–47.85, p = 0.007)

Usage of subsidized primary care was associated with regular:

Diabetes screening (adj OR 2.94, CI 1.04–8.31, p = 0.042).

Proximity to primary care was associated with higher participation in regular: Colorectal cancer screening (adj OR 1.48, CI 1.01–2.21, p = 0.049).

Usage of subsidized primary care was associated with higher participation in regular:

Cervical cancer screening (adj OR 7.93. CI 1.03–62.51, p = 0.047)

Breast cancer screening (adj OR 6.02, CI 1.69–21.28), p = 0.006)

Proximity to primary care was associated with higher participation in regular:

Cervical cancer screening (adj OR 3.22, CI 1.72–5.84, p < 0.001)

Breast cancer screening (adj OR 2.22, CI 1.08–4.54), p = 0.032)

Regular primary care follow up was associated with less participation in regular:

Breast cancer screening (adj OR 0.10, CI 0.01–0.75, p = 0.025).

Secondary outcome: NA

Regular primary care was independently associated with regular participation in cardiovascular screening in both rental housing and owner occupied communities.

However, for cancer screening, in the rental housing community, proximity to primary care was associated with less participation in regular screening, while in the owner occupied housing community, regular primary care was associated with lower screening participation; possibly due to embarrassment regarding screening modalities.

Poor

Qualitative interview:

To elicit perceptions about cardiovascular disease and cancer screening.

Major themes and subthemes:

• Primary care characteristics (Barriers)

Lack of trust in healthcare system/healthcare professionals

Healthcare professional does not often discuss screening – no time

Embarrassment about screening modality

Characteristics of clinic (manpower, location, hours open)

• Knowledge (Barriers)

Not aware of screening

No need screening as healthy /not at risk

Not aware of where to go for screening

Screening may not be accurate/alternative screening methods are better

Last test normal, so no need to go again

Confusion that mammogram causes cancer

• Priorities (Barriers)

No time to go, too busy

Can spend money on other things

• Attitudes (Barriers)

Fatalism

Fear of diagnosis and/or treatment

Too old to go for screening

Traditional medicine is better

Disease not important

Patients were discouraged from screening by distrust in the doctor-patient relationship; for cancer screening in particular, patients were discouraged by potential embarrassment.

Wee LE (2014) [24]

Cross-sectional

and Qualitative

From Jan 2009 to May 2010

Choice of primary health care source.

Residents ≥40 y/o.

6 blocks of a socially integrated housing precinct.

(3 vs 3)

Sample size

710

Participation rate

88.6% (710/800)

Rental: 89.8% (359/400) VS Owned: 87.8% (351/400)

Prevalence rate: NA

Follow up rate: NA

Primary outcome

Preferred source of medical treatment and advice

Rental:

Rely on own knowledge. 52.6% (189/359)

Alternative medicine practitioners. 29.5% (106/359)

Family/friends. 6.7% (24/359)

Western-trained doctors. 11.1% (40/359)

Owned:

Rely on own knowledge. 54.1% (190/351)

Alternative medicine practitioners. 2.0% (7/351)

Family/friends. 14.0% (49/351)

Western-trained doctors. 29.9% (105/351)

Secondary outcome

Residents staying in rental housing (compared with those staying in owner-occupied housing) were less likely:

to seek advice from Western-trained doctors

(adj OR 0.36, CI 0.21–0.61, p < 0.001)

to seek advice from family members

(adj OR 0.36, CI 0.19–0.69, p < 0.002)

They were more likely:

to turn to alternative medicine practitioners

(adj OR 14.29, CI 4.55–50.00, p < 0.001)

In rental housing community:

Unmarried were more likely to consult alternative medicine practitioners (adj OR 3.13, CI 1.41–6.67, p = 0.005)

Minority ethnicity were more likely to consult family members

(adj OR 3.23, CI 1.23–8.33, p = 0.016)

Higher household income (≥$500/month) were less likely to seek consult from anyone, relying instead on their own knowledge

(85.2%, 161/359)

With dyslipidemia were less likely to consult alternative

medicine practitioners

(adj OR 0.34, CI 0.14–0.83, p = 0.017).

Western-trained physicians are not the first choice of seeking primary care in the rental housing community.

Poor

Qualitative interviews:

To elicit perspectives on barriers/enablers that lower income patients face in seeing a Western-trained physician for primary care.

Patient and provider comments fell into the following content areas: Primary care characteristics- trust, distance, waiting time.

Knowledge- healthy, not effective, minor ailment.

Costs-for treatment, subsidies.

Priorities- busy

Attitudes- fear of diagnosis and treatment

Information sources- Media (TV, newspaper)

Self-reliance was perceived as acceptable for ‘small’ illnesses but not for ‘big’ ones.

Communal spirit was cited as a reason for consulting family/friends.

Social distance from primary care practitioners was highlighted as a reason for not consulting Western-trained doctors.

Knowledge, primary care characteristics and costs were identified as

Potential barriers/enablers.

Low LL (2016) [25]

Retrospective cohort

From Jan 2014 to Dec 2014

Housing as a social determinant of health and its association with readmission risk and increased utilisation of hospital services.

Patients who have at least one clinical encounter (admission or ED visit) to Singapore General Hospital (SGH) in 2014.

Patients, who died in 2014, are non-residents, who resided in areas where SGH is not the primary hospital or patients discharged to long-term residential care facilities were excluded.

Sample size

A total of 14,457 patients were analyzed and 2163 patients (15.0%) were rental housing residents.

Primary outcome

Readmission within 15 days associated with residence in public rental housing:

OR 1.19, CI 1.02–1.39, p = 0.029

Readmission within 30 days:

OR 1.27, CI 1.12–1.43, p < 0.001

Frequent hospital admissions:

OR 1.27, CI 1.14–1.43, p < 0.001

Frequent ED attendances:

OR 1.40, CI 1.21–1.61, p < 0.001,

Staying in public rental housing showed a 8% lower risk per one SOC visit, but the result was statistically non-significant, 0.92 (0.83–1.02), p = 0.112

Secondary outcome: NA

Patients staying in rental housings have a 19 and 27% higher odds of being readmitted within 15 and 30 days, respectively.

Patients staying in rental housings have a 27 and 40% higher risk of being a frequent hospital admitter and frequent ED attendee, respectively.

Fair