Table 3 Showing the results of studies examining interventions to increase immunization uptake in urban poor and slum communities in low and middle-income countries

Uddin, 2010 [88]

Bangladesh (Dhaka)

Multi-C

Extended hours,

provider training,

screening tool,

community group.

Before (bf) and after (aft)

Children 12–23 months in a Dhaka slum.

529 before,

526 after.

Fully immsd increased from 43% to 99% (p < 0.000)

Increases seen across range of individual vaccines, and in both children of working and non-working mothers.

Hayford, 2014 [87]

“

“

“

Economic evaluation

“

Cost of $20.95 per fully immsd child

Total cost for intervention for 1 year $18,300.

Pradhan, 2012 [81]

India

(Patna)

Multi-C

Outreach services,

additional staff, task shifting, link workers,

geographic monitoring, community involvement, additional supervision.

Before and after

All eligible living in slums in Patna

Estimated eligible population immsd before and after intervention.

BCG 29% bf, 64% aft

DPT1 28% bf, 62% aft

DPT3 21% bf, 49% aft

MCV 23% bf, 51% aft

TT1 15% bf, 22% aft

TT2 10% bf, 28% aft

Increases observed across a range of vaccines in a population of approx. 25,000 children, however statistical significance not reported.

Agarwal, 2008 [77]

India

(Indore)

Multi-C

Increasing awareness & demand, improve vaccine supply and accessibility, community links.

Before and after

Children aged 12–23 months living in 79 slum areas

Eligible children within estimated 150,000 study population.

Fully immsd increased from 32% to 72% between 2003 (bf) and 2006 (aft).

Primarily descriptive, with limited detail on outcomes achieved and no statistical analysis.

Khan, 2006 [85]

Pakistan (Karachi)

Multi-C

Information, education, community involvement, intensive vaccination campaign.

Cluster RCT

Children aged 2 to 16 years in squatter settlements.

21,059 children in 60 clusters.

Achieved 74% coverage in previously unvaccinated population.

Purpose of study was to test effectiveness of typhoid polysaccharide vaccine, using Hep A as a control, but also reported on campaign design.

Poulos, 2004 [96]

India

(New Delhi)

Multi-C

Mass vaccination vs. school campaign vs. targeted campaign at 2–5 year old children.

Economic evaluation

All people within a slum area in new Delhi

26 clusters

Assuming cost of $1 per typhoid vaccine:

Mass campaign = $50 per case avoided, School campaign = $41, and targeted to young children = $14.

Examination of the economic benefits of typhoid vaccine campaigns from a societal perspective using different methods in a slum area.

Mbabazi, 2012 [92]

Kenya

(Nairobi and Nyanza/Western Provinces)

Multi-C

House to house canvassing, community mobilization, mobile phone documentation, web application monitoring.

Cross-sectional (post-hoc data only)

Children in high-density urban poor communities.

164,643 households, with 161, 695 children

Post campaign monitoring found measles coverage of 96% reported (92% confirmed).

Additional strategy as part of a mass measles campaign to increase coverage of the campaign. 75% households reported acceptance of supplemental measles vaccination prior to intervention.

Lhamsuren, 2012 [95]

Mongolia (Ulaanbaatar)

Multi-C

Reaching Every District (RED) strategy [104].

Cross-sectional & qualitative.

At risk children in urban poor communities, focusing on one district of 22,726 people)

3126 at risk children under 15 years old in selected community.

Immsd an additional 477 at risk children (15% of total number of eligible children).

Detail provided on barriers to imms services. Cost in study district = $14,166, which also included other maternal and child health interventions.

Igarashi, 2010 [74]

Zambia

(Lusaka)

Outreach

Growth Monitoring Plus (GMP): outreach vaccination and other child health services into under-served slum areas using community volunteers.

Interrupted time series

Children in four slum area, split into 2 primary areas and 2 areas with 2-year time-lag.

1128 (584 in primary site, 544 in time-lag intervention site)

Full imms coverage increased from 52.6% at baseline to 68.8% at final measurement in primary area (p < 0.001) and 43.1% to 56.7% in lag area (non-significant).

Length of residence was significantly longer in primary area, which may explain the differential effect. Frequency of attendance at GMP+ services significantly associated with higher coverage (OR 1.27, P < 0.001)

Sasaki, 2011 [90]

“

“

“

Before and after cross-sectional surveys

Children in one slum area of 48.798 people (one of the 4 areas included in the above study).

280 sampled households.

DPT3 increased 75.7% (bf) to 87.3% (aft) and measles from 66.8% (bf) to 76.1% (aft).

Closer distances to imms service points significantly associated with higher coverage, and impact of this reduced with GMP+ outreach services.

Ansari, 2010 [78]

India

(Aligarh City)

Outreach

Immunisation outreach camps

Cross-sectional

Children <5 in slum areas of Aligarh City

2531

13,989 vaccines administered to the 2531 children attending services.

Population coverage not reported. Significantly lower imms coverage observed in female children attending.

Prabhakaran, 2014 [80]

India

(New Delhi)

Outreach

Mobile health clinic

Economic Evaluation

Children <5 years living in Dakshinpuri extension resettlement colony.

1583 children attending the clinic in 1 year.

1583 children received 8488 vaccinations through the service at a cost of 66.14 Rupees per vaccine (US$0.10)

Imms services provided alongside a range of other health services.

Uddin, 2016 [89]

Bangladesh

(Dhaka)

Reminder/ recall

SMS (text) message reminders managed using mTika software system.

Before and after

Both children aged 0–11 months in hard-to-reach rural areas and urban street children in slums.

2823 urban street children (intervention 518 bf and 520 aft, with 1785 controls)

Urban intervention: 40.7% (bf) to 57.1% (aft) compared to controls 44.5% (bf) to 33.9% (aft).

Adjusted OR = 3.0 (95% CI 1.4–6.4). Decline in control population imms coverage also noted in rural population (not shown here).

Kazi, 2014 [83]

Pakistan

(Karachi)

Reminder/ recall

SMS (text) message monitoring of immunisation activities.

Cross-sectional

20 clusters of 200–250 households in 3 high polio risk areas.

28 households with children under 5 in each cluster.

Coverage in population who replied to messages was 74.5% (95% CI 71.6–77.4) which was very similar to result found using WHO lot quality assurance sampling.

Text messages sent to parents to confirm if immunization staff had visited and vaccinated child. If no answer, follow-up phone calls made.

Domek, 2016 [94]

Guatemala

(Guatemala City)

Reminder/ recall

SMS (text) message reminders

RCT pilot

Children aged 8–14 weeks presenting at a clinic serving a low-income population.

321 (160 intervention, 161 control)

Higher proportion in intervention group completed series (84.4% vs 80.7%), which was not statistically significant.

Three reminders sent to intervention parents. Loss to follow-up 25 in intervention and 31 in control groups. Demonstrated feasibility.

Mukanga, 2005 [91]

Uganda

(Mulago II Parish, outside Kampala)

Reminder/ recall

Child health cards

Cross-sectional

Children aged 0–24 months

260 households

66% children had child health cards, and were almost 10 times more likely to be fully immsd than those without (OR = 9.55, 95% CI 3.19–29.45)

Children born in a health facility were 4 times more likely to have a health card than those born at home.

Owais, 2011 [84]

Pakistan

(Karachi)

Education

Education session delivered by community health workers using pictorial message.

RCT

Mothers of children sampled from 5 low-income areas of Karachi

366 mother-infant pairs (179 in intervention, 178 in control group finally snslysed)

Increase in DPT3/Hep B coverage of 39% in intervention group (RR = 1.39, 95% CI: 1.06–1.81)

27% (312) of sampled mothers declined to participate.

Anjum, 2004 [86]

Pakistan

(Karachi)

Education

Education messages provided to mothers by medical students

Randomized, controlled before and after

People living in Sikanderabad squatter settlement in Karachi.

317 households (110 intervention 207 controls)

Follow-up 4 years post intervention found significant increase in full imms intervention area (46.5% bf, 75% aft, p < 0.005) and no sig difference in control area. However, no test for interaction performed, compromising the statistical results.

The results are limited by poor study design. Although uptake of imms services significantly increased in both intervention and control households the groups were poorly matched with very divergent baseline service use, preventing firm conclusions being drawn about intervention effectiveness.

Shei, 2014 [93]

Brazil

(Salvador)

Incentives

Conditional cash transfer (Bolsa Familia) to people on low-incomes.

Nested cross-sectional survey within prospective cohort

Bolsa Familia recipients in a slum area of 14,000 people.

1266 children from 3000 randomly sampled households (841 beneficiaries, 425 controls)

Recipient children under 7 years 2.8 times more likely to attend services for vaccination (OR 2.8, 95% CI 1.4–5.4, P = 0.002).

Also increased odds of attending for growth monitoring (OR 3.1, P < 0.001) and health checks (OR = 1.6, p = 0.061). No impact on children older than 7.

Chandir, 2010 [82]

Pakistan

(Karachi)

Incentives

Food/medicine coupon incentives worth US$2 for follow-up DTP vaccinations.

Quasi-experimental with non-simultaneous controls

11 sub-districts, including middle and very-low income households.

2561 intervention, 2051 controls.

Completion of DTP 3 vaccinations higher in intervention group (Adjusted RR 2.2, 95% CI 1.95–2.48, p < 0.001)

Significant loss in the control cohort, only 847 (41% enrolled) entering final analysis.