STRATEGIES AGAINST FLU EMERGENCE

Health Care Utilization Survey in East Jakarta and Bogor District Indonesia

July 2, 2012

This publication was produced for review by the United States Agency for International

Development. It was prepared by Development Alternatives, Inc.

STRATEGIES AGAINST FLU EMERGENCE PROJECT

STRATEGIES AGAINST FLU EMERGENCE

Health Care Utilization Survey in

East Jakarta and Bogor District

Indonesia

Title:

Health Care Utilization Survey in

East Jakarta and Bogor District

Indonesia

Sponsoring office: USAID/Indonesia Office of Health

Contracting officer's technical representative: Artha Camellia

Contract number: AID-EDH-I-00-05-00004-00 Order number: AID-497-TO-11-00001 Contractor: DAI DAI project number: 1001470

Submitted: July 2, 2012

The views expressed in this publication do not necessarily reflect the views of the United States

Agency for International Development or the United States Government.

7/11/2012

1

Health Care Utilization Survey inEast Jakarta and Bogor District

Indonesia

1

Strategies Against Flu Emergence - DAIJohns Hopkins University Center for Communication Programs

Centers for Disease ControlWorld Health Organization

University of Indonesia Center for Health Research

Funded by USAID

June 2012

This study is made possible by the support of the American People through the United States Agency for International

Development (USAID.) The contents of this study are the sole responsibility of DAI and do not necessarily reflect the

views of USAID or the United States Government.

2

AcknowledgementsThis study was conducted as part of the USAID-funded Strategies Against Flu Emergence (SAFE) project

(contract # AID-EDH-I-00-05-00004-00), in support of the Government of Indonesia’s efforts to reduce the

impact of avian influenza (H5N1) on humans and animals. Many people contributed to the design

implementation and analysis that went into this research.

Special acknowledgment of Douglas Storey, Yunita Wahyuningrum and Jennifer Kreslake with JHU∙CCP who

led the design of the study, supervised field work and conducted analysis of the findings.

Thanks to Timothy M. Uyeki, Aaron D. Storms, Kathryn Lafond, Danielle A. Iuliano and Yekti Praptiningsih at

CDC Atlanta and Jakarta; and Graham Tallis and Oratai Rauyajin at WHO for substantial input into the design

of the study and the survey instrument. Special thanks to Aaron D. Storms for assistance with the sample

design.

At USAID/Jakarta, Artha Camellia, Bambang Heryanto and Kendra Chittenden helped shape the direction of

the study and provided careful insights into the interpretation of the results.

At the University of Indonesia Center for Health Research (PPK-UI), Christiana Rialine, Fitra Yelda and their

team coordinated all the field work and data management.

Thanks to the DAI and JHU∙CCP staff who also helped shape the study, develop the instruments and manage

the day to day implementation of this study: Maria I. Busquets, Basil Safi, Rekha Lal, Heri Haerudin and

Gabrielle Hunter.

Finally, a special thank you to the local governments of East Java and the district of Bogor for allowing us to

conduct the interviews and data collection.

7/11/2012

2

3

Part A:

STUDY DESIGN & FIELD WORK

4

INTRODUCTION

Infection from H5N1, the highly pathogenic avian influenza (AI) virus, results in high cases fatality

rates. Indonesia has the highest number of confirmed human cases of AI and one of the highest case

fatality rates in the world, 83 percent as of 29 May 2012.1 This high case fatality rate is widely attributed

to delays in care seeking, diagnosis and initiation of treatment for respiratory disease. Respiratory

disease and influenza-like illnesses (ILIs) are extremely common in Indonesia and experts estimate that

the actual number of H5N1 cases is several times higher than the confirmed total with many cases

unidentified, misidentified, or unreported. The western half of Java accounts for more than 68 percent

of all human cases in Indonesia. While H5N1 is not readily transmitted among humans, the virus is

endemic in animal populations in Indonesia, raising the possibility that H5N1 could at some point evolve

into a form more easily transmissible between humans, causing a pandemic that could kill millions. Both

direct and indirect exposure to live and domesticated birds, to poultry waste and to poultry in wet

markets is extremely common throughout Indonesia.

The USAID-funded SAFE project is designed to reduce this risk by simultaneously working to (1)

improve biosecurity practices in the poultry industry, thereby reducing bird to bird transmission and (2)

improve hygiene and poultry handling practices among the general public, thereby reducing bird to

human transmission, while 3) encouraging rapid care seeking and faster initiation of appropriate

treatment as early as possible after the onset of symptoms of respiratory disease.

Under the umbrella of the USAID Strategies Against Flu Emergence (SAFE) project and in

conjunction with the Centers for Disease Control Atlanta and Jakarta offices and the World Health

Organization Indonesia, a community-based household survey was conducted in East Jakarta and Bogor

District in West Java.

1 http://www.who.int/influenza/human_animal_interface/H5N1_cumulative_table_archives/en/index.html

7/11/2012

3

5

OBJECTIVES

The survey was designed to generate estimates of the seasonal influenza disease burden

and to determine the proportion of persons with ILI that seek care, their understanding of signs

and symptoms that indicate the need for care and decision-making about when and where to

seek care for respiratory illness. The findings from the HUS will be used in conjunction with

enhanced surveillance data collected in a separate study by CDC/Jakarta for the purpose of

developing disease burden estimates for seasonal influenza among East Jakarta district residents

who present as outpatients with ILI or as hospitalized patients with SARI (e.g. pneumonia).2

This survey also aims to understand perceptions about exposure to birds and of the risk of

H5N1 transmission. In addition, findings of the HUS will be used to inform preventive education

strategies at the community level to reduce bird to human transmission of H5N1 virus and to

reduce delays in care seeking that can result in higher than necessary mortality rates.

Through a competitive procurement process, Pusat Penelitian Kesehatan Universitas

Indonesia/PPK-UI (Center for Health Research at the Faculty of Public Health, University of

Indonesia) was selected as the research agency to conduct the fieldwork for HUS in East Jakarta

and Bogor District.

2 These calculations to be provided by CDC/Atlanta; analysis is in progress.

Fig. 1:EastJakarta

Fig. 2: Bogor

6

METHODOLOGY

Data collection was accomplished through a face-to-face survey in

households throughout East Jakarta and Bogor District in West Java.

The survey was administered primarily in Bahasa Indonesia (translated

from English). About 60 percent of the interviews in West Java were

conducted in Bahasa Sunda where Sundanese is the primary language.

Study sites

The study was conducted in the East Jakarta

District of DKI Jakarta and in the Bogor District in West

Java. In East Jakarta, study sites included the seven sub-

districts of Matraman, Pulogadung, Duren Sawit, Kramat

Jati, Pasar Rebo, Ciracas and Makasar, where enhanced

surveillance sites monitored by CDC/Jakarta are located

(community health centers (puskesmas) of Matraman,

Pulogadung, Duren Sawit and Kramat Jati, and the six

hospitals of Persahabatan, Budi Asih, Pasar Rebo,

Harapan Bunda, Islam Pondok Kopi and Haji Pondok

Gede.

In Bogor District, seven sub-districts were

randomly selected (Cijeruk, Cileungsi, Gunung Putri,

Citeureup, Ciampea, Cibinong and Rancabungur).

7/11/2012

4

Sampling Method

To estimate the proportion of persons living in East Jakarta that were hospitalized for respiratory illness in the prior year, we searched the literature for similar estimates from different international settings (including Kenya,3 Thailand, 4, 5 and Guatemala6) and spoke to experts conducting similar activities in Bangladesh and El Salvador. Estimates varied from site to site, and we selected a conservative estimate of 0.7 percent.

To estimate the proportion of persons hospitalized for respiratory illness (Step 1) at the East Jakarta sites, we looked at the pneumonia reports posted in the Provincial Health office website that detail the number of pneumonia cases (used as a surrogate for respiratory illness) that were admitted to each hospital in the district). These data indicated that 67 percent of hospitalized pneumonia cases in East Jakarta for the previous year were hospitalized at any one of our sites.

From this we were able to estimate that the proportion of persons living in East Jakarta that were hospitalized for respiratory illness at any one of our sites is 0.44 percent (67 percent of 0.7 percent).

3 Breiman RF, Olack B, Shultz A, Roder S, Kimani K, Feikin DR, et al. Healthcare-use for major infectious disease syndromes

in an informal settlement in Nairobi, Kenya. J Health Popul Nutr. 2011 Apr;29(2):123-33.4 Chamany S, Burapat C, Wannachaiwong Y, Limpakarnjanarat K, Premsri N, Zell ER, et al. Assessing the sensitivity of

surveillance for pneumonia in rural Thailand. Southeast Asian J Trop Med Public Health. 2008 May;39(3):549-56.5 Jordan HT, Prapasiri P, Areerat P, Anand S, Clague B, Sutthirattana S, et al. A comparison of population-based pneumonia

surveillance and health-seeking behavior in two provinces in rural Thailand. Int J Infect Dis. [Comparative Study]. 2009

May;13(3):355-61.6 Lindblade KA, Johnson AJ, Arvelo W, Zhang X, Jordan HT, Reyes L, et al. Low usage of government healthcare facilities for

acute respiratory infections in Guatemala: implications for influenza surveillance. BMC Public Health. 2011 Nov 24;11(1):885.

7

East Jakarta City

7 sub-districts 7 sub-districts

9 clusters (RW) per

kelurahan

9 clusters (RW) per

desa

Bogor District

20 households per

cluster (RW)

20 households per

cluster (RW)

Probability proportionate

to size (PPS)

Simple Random Sampling (SRS)

SAMPLING DESIGN

SRS

Purposive

sampling

Sampling Method (cont.)

In order to ensure that we detect a prevalence level of at least 0.4 percent (expected SARI cases hospitalized in our facilities) but up to 0.7 percent (highest value that we could find of SARI cases hospitalized in one of our facilities) with an alpha of 0.05 and power of 0.85, we would need survey data from 4,997 persons.

Based on data from the most recent Indonesia Demographic and Health Survey (2007) indicating that average household size is approximately 4 persons per household, this would require a sample of at least 1,250 households.

The resulting sample design is summarized in Figure 3.

Fig. 3: Sample Design

In each selected RW, list of households was obtained from either head of the RW or RT. Lists were validated and adjusted according to field conditions.

In each household, the interviewer asked to speak with the person who was most knowledgeable about the health condition of household members or was the main caregiver for household members.

8

7/11/2012

5

INSTRUMENT DEVELOPMENT

The instrument was developed in collaboration with USAID/Jakarta, CDC/Atlanta and Jakarta, WHO/Indonesia and SAFE. The questionnaire was designed to measure the following:• frequency of influenza-like illness in the previous two weeks,• proportion of cases of ILI that sought care at health care facility, • frequency of SARI (among all members of selected households) in the previous year, • proportion of cases of SARI that were admitted to any hospital and at enhanced

surveillance hospitals, • knowledge of risk factors for human infection with H5N1 virus, including direct contact

with sick/dead poultry, and indirect exposure (e.g. visiting a wet poultry market – also referred to commonly as traditional or live bird market),

• knowledge about human illness with H5N1 virus and health-care seeking behaviour for family member with ILI and exposure to sick and dying poultry,

• exposures to poultry, including slaughtering of sick or dying poultry, through animal husbandry practices (for residents who keep live birds), through visits to wet poultry markets and other indirect sources of contact, and through food handling practices with emphasis on handling and preparing poultry products and

• media habits, information sources and exposure to messages related to AI.

9

INSTRUMENT DEVELOPMENT (cont.)

The instrument was translated into Bahasa Indonesia before being pre-tested with sample household members in Depok and Bogor Municipality in December 2011.

Pre-testing of the instrument was conducted by researchers from PPK-UI, in the Cimanggis sub-district of Depok City and in the North Bogor sub-district of Bogor City. These two areas were selected based on their similarity to the field sites where data collection would actually be conducted, but neither one was selected for the actual study.

Pre-testing was designed to confirm the wording, flow and time spent for each respondent, and to ensure that respondents could answer each question. The pre-testing results were then sent to all partners for input. Detailed feedback on HUS pre-tests, including suggested revisions to specific questions were addressed.

10

7/11/2012

6

DATA COLLECTION MANAGEMENT

Prior to data collection activities, SAFE and PPK-UI obtained ethical clearance for the survey implementation from the Research Ethics Committee at the Faculty of Public Health of the University of Indonesia on December 21, 2011 and from the CDC Atlanta Institutional Review Board on February 8, 2012.

A three-day training program for field personnel, including field coordinators and interviewers, was conducted on 24-26 January 2012. It was attended by 32 interviewers, four field coordinators, SAFE and CDC Jakarta staff and researchers from PPK/UI.

Data collection was conducted from 8 February – 1 March 2012. In total, 2520 respondents were interviewed (1260 respondents in each district/city). Those households contained a total of 11,328 regular residents, 5,535 in East Jakarta and 5,793 in West Java for a mean household size of 4.4 and 4.6, respectively.

On average, 12 percent of households that were approached refused to participate (19 percent in East Jakarta City and 4 percent in Bogor District. On average, the interview time per respondent was approximately 45 minutes.

After all completed questionnaires were checked and edited, field coordinators selected 5 percent (n=120) of the questionnaires randomly and conducted a spot-check. This was done by revisiting the respondents and re-interviewing them on some important key questions on the same day or the day after the original interview. No major problems/discrepancies were identified by field coordinators.

11

DATA ENTRY MANAGEMENT

Questionnaires were double entered into a database using Epi-Data software. When any discrepancies were found, the data were crosschecked against the original questionnaires.

Construction of population and normalized weight variables

The final dataset contains two weighting variables: population weight and normalized weight. Different methods were applied to obtain the weight values in East Jakarta City and Bogor District, according to the sampling design used in each area.

• In East Jakarta City, the calculation of weights took into account the selection of sub-districts and households in each RW (both using the simple random sample method).

• In Bogor District, the calculation of weights took into account only the selection of households in the RW, since the selection of sub-districts was made proportional to the population size in each sub-district (PPS method).

When the population weight is used, the weighted proportion reflects the proportion from the total population (N= 7,100,632); while normalized weight reflects the proportion only from the total sample (N=2520). However the proportion should remain identical when either population or normalized weight is applied.

The normalized weight was used in statistical analyses.

12

7/11/2012

7

13

Household Characteristics

14

Table 1: Household Member Characteristics

by Survey Area

Household member

characteristics

East

Jakarta

(n=5535)

Bogor

(n=5793)

Total

(n=11328)

Gender

Male

Female

49.1

50.9

51.3

48.7

50.5

49.5

Age categories

Under 5

5-14

15-49

Over 50

8.3

17.0

54.2

20.5

9.9

22.0

54.7

13.3

9.3

20.3

54.5

15.8

Marital status

Single

Married

Divorced

Widowed

Separated

46.8

47.4

0.8

4.6

0.5

49.8

45.8

1.1

3.2

0.1

48.7

46.3

1.0

3.7

0.2

Highest education achieved

No school

Some primary

Completed primary

Completed lower secondary

Completed upper secondary

Academy

University

12.1

12.8

12.9

14.5

35.0

5.2

7.1

17.1

21.6

28.1

14.7

14.6

1.3

2.6

15.4

18.5

22.7

14.8

21.8

2.7

4.2

Household Characteristics

There were some differences in household composition between East Jakarta and Bogor due to the fact that East Jakarta is predominantly urban while West Java is predominantly rural (Table 1).

Households in East Jakarta had slightly more women than men, while Bogor had slightly more men than women.

Households in East Jakarta were a little older on average with fewer members in the Under 5 and 5 to 14 year old age categories and more in the Over 50 age group compared to Bogor.

East Jakarta households were also better educated on average with more than twice as many members having completed upper secondary, academy or university level education compared to Bogor.

7/11/2012

8

15

Characteristics of Respondents

(member most knowledgeable about household health)

Table 2: Characteristics of respondents

Characteristic

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Gender

(female %)

76.5 79.4 78.3*

Age

(mean years)

45.3 40.1 42.0*

Married

(%)

83.6 89.2 87.1*

Completed high

school

(%)

43.0 20.1 28.6*

Table 2 summarizes the

characteristics of the main

respondents (the member of each

household who was most

knowledgeable about the health

of other family members).

More than three-fourths were

female, with an average age of 45

in East Jakarta and 40 in West

Java.

More than 8 out of 10 were

currently married. In East Jakarta,

43 percent had completed at least

high school compared to 20

percent in West Java.

16

Table 3:

Household possessions by study area

Household possessions

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Color TV

Black & white TV

VCR

Video CD player

Stereo

CD stereo

Laser disc player

Karaoke machine

Keyboard instrument

Camera

Video camera

Personal computer

Air conditioner

Fixed line telephone

Mobile phone

Refrigerator or freezer

Microwave

Washing machine

Motorcycle

Automobile

Own (vs rent) home

Fan

Gas cooker

Sewing machine

Water pump

Rice cooker

Radio

Bicycle

98.6

0.9

3.2

75.5

12.1

13.9

2.7

10.1

4.1

31.9

9.0

42.3

16.4

35.3

94.1

84.9

11.8

52.2

79.1

15.7

72.2

95.2

93.1

22.7

71,0

86.4

59.9

64.2

88.0

1.4

2.0

53.8

21.1

8.0

1.5

6.5

4.2

14.0

3.9

17.9

7.0

10.1

87.6

55.0

5.0

24.4

61.5

8.9

82.7

53.5

90.5

12.9

59.9

77.9

34.0

51.1

92.0

1.2

2.4

61.8

17.8

10.2

1.9

7.8

4.2

20.6

5.8

27.0

10.4

19.4

90.0

66.0

7.5

34.7

68.0

11.4

78.8

68.9

91.4

16.5

64.0

81.1

43.5

55.9

Mean number of

possessions

13.1 10.0 11.1

Household Amenities

Households were asked to indicate which physical amenities they possessed as a proxy for socio-economic status. A list of 33 common household appliances or possessions was used (Table 3).

Ownership of a color television was nearly universal in East Jakarta and nearly 90 percent in Bogor. Over 40 percent of households in East Jakarta had a computer of some kind, compared to 18 percent in Bogor, while mobile phone ownership was 94 percent and 88 percent, respectively. Overall, more than two-thirds of households had a motorcycle and 11 percent owned an automobile (about twice as many in East Jakarta compared to Bogor). Sixty-four percent of households in East Jakarta owned a radio compared to 34 percent in Bogor.

7/11/2012

9

17

INPATIENT CARE AMONG

HOUSEHOLD MEMBERS

18

Households with Hospitalized Members in the Past Year

Table 4:

Percent of HHs with one or more members

hospitalized in past year

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

One person 197

15.6%

150

11.9%

347

13.8%

More than

one person

13

1.0%

9

0.7%

22

0.9%

Of the 11,328 people

enumerated in 2,520 households

only 375 (14.9 percent of

households) were reported to

have been hospitalized in the 12

months preceding the survey.

More households in East Jakarta

reported a hospitalized member

compared to Bogor.

Overall, 13.8 percent of

households reported one

hospitalized member while 1

percent reported multiple

hospitalizations.

7/11/2012

10

19

Reported Symptoms of Hospitalized Household Members

Table 5:

Symptoms of hospitalized household members

Number hospitalized

n

(n=2520)

%

(100%

)

One person (n=347)

Fever/chills

Cough

Difficult breathing/shortness of breath

Any of those symptoms

98

17

15

119

26.2

4.6

4.0

32.8

More than one person (n=38)

Fever/chills

Cough

Difficult breathing/shortness of breath

Any of those symptoms

5

1

0

6

13.2

2.6

0.0

15.8

Of the households

reporting a hospitalized

member in the past year

(n=347), 32.8 percent

reported that the

symptoms of the

hospitalized person

included one or more of

the symptoms associated

with avian influenza: fever

and/or chills, cough and

difficulty breathing or

shortness of breath. Most

of that was accounted for

by fever/chills. In

households with morethan one hospitalization (n=38), 15.8 percent of the second person hospitalized

were reported to have an avian flu-like symptom, most of which was again

accounted for by fever/chills. Only two of the hospitalized patients were

diagnosed with seasonal flu; none were diagnosed with avian influenza.

20

MORTALITY

7/11/2012

11

21

Deaths in the Household and Symptoms

Table 6:

Household deaths and symptoms of deceased in past year

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Number of deceased

Any

One

More than one

60 (4.8%)

53 (4.2%)

7 (0.1%)

29 (2.3%)

26 (2.1%)

3 (.002%)

89 (3.5%)

79 (3.1%)

10 (.004%)

Symptoms of deceased

Fever/chills

Cough

Difficult breathing/

shortness of breath

Pneumonia

1 (0.03%)

0 (0.0%)

2 (0.3%)

1 (0.03%)

0 (0.0%)

0 (0.0%)

2 (0.07%)

1 (0.03%)

1 (0.01%)

0 (0.0%)

4 (0.04%)

2 (0.02%)

Overall, there were 89

deaths reported in the

past 12 months; 3.5

percent of households

reported one or more

deaths.

Of those households,

only a very small

percentage (less than 0.1

percent) reported that the

deceased had

experienced avian

influenza-like symptoms,

including pneumonia.

None of the HHs with

multiple deaths reported death from the same symptoms.

Of the deceased, 49 percent were hospitalized within 30 days prior to death

and of those who were hospitalized, 95 percent sought some other care before

being admitted to the hospital (not shown).

22

HOUSEHOLD EXPOSURE TO POULTRY

7/11/2012

12

23

Keeping Live Poultry

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

HH keeps any kind of

birds (self-report)

18.4 32.4 27.3*

HH has birds

(observed)

16.5 31.6 26.0*

HH keeps chickens

(self-report)

8.3 27.1 20.1*

Where birds are kept

(among HH with

birds)*

Inside the home

Outside the home

Both inside and

outside

(n=154)

9.3

79.7

11.0

(n=463)

11.1

83.1

5.8

(n=617)

10.7

82.2

7.1

Households in rural West Java are

more likely to keep poultry

compared to households in East

Jakarta.

Direct observation by the

interviewers confirmed this pattern.

The type of birds kept was more

likely to be chickens in Bogor

compared to East Jakarta.

Households in Bogor were more

likely to keep birds inside the family

living area of the house, although

only about 10% of households

overall reported this practice.

24

Caging Practices

East

Jakarta

(n=154)

Bogor

(n=462)

Total

(n=617)

Birds allowed to roam freely (self-report) 12.6 35.5 29.5*

Caging practices (self-report)

Always caged

Always free

Sometimes free

Away from house

69.3

3.1

26.1

3.9

21.2

14.8

63.1

0.9

32.7*

11.9*

54.0*

1.4*

Observed conditions

Poultry observed in house

Birds kept in closed building or fenced area

Other animals can come in contact

Poultry feces on ground

Water and soap available near poultry area

18.9

81.4

23.8

26.8

20.1

35.5

55.8

67.6

70.0

8.4

31.6*

61.7*

57.5*

60.1*

11.1*

Other higher risk conditions were also more commonly observed by interviewers

in Bogor than in East Jakarta: interviewers were more likely to observe birds in

the house and poultry feces on the ground in Bogor and less likely to observe

poultry enclosures and the availability of soap and water near poultry areas.

Compared to East

Jakarta, households in

Bogor were more likely

to report allowing their

birds to roam free

(36% vs 13%,

respectively) and 63%

of Bogor households

reported letting their

birds range at least

sometimes, compared

to 25% in East

Jakarta.

7/11/2012

13

25

Exposure through Poultry Care: Who USUALLY cares for poultry?

* Significant difference between E. Jakarta & Bogor

Generally, the head of household was reported to be in charge of poultry rearing

in household that keep birds, but in Bogor the female spouse was more likely to

bear responsibility for this job compared to East Jakarta. Small percentages of

children were also responsible.

26

Exposure through Poultry Care: Who EVER cares for poultry?

When asked about who in the household EVER cares for poultry, it became clear

that exposure to household poultry was widespread in the family. Two thirds of

female spouses, one quarter to one third of male children and about 10% of

female children provided some care. Other adults in the household help with

poultry rearing, too.

* Significant difference between E. Jakarta & Bogor

7/11/2012

14

27

Exposure through Commercial Activity

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Involved in some commercial poultry work 9.3 8.9 9.0

Selling live poultry 0.7 4.2 2.9*

Selling eggs 8.3 5.7 6.7*

Trading/transporting eggs or poultry 0.8 0.3 0.5

Slaughtering 0.2 0.3 0.2

Less than 10% of households indicated that they were involved in some kind

of commercial poultry work.

More reported selling poultry in Bogor, while more reported selling eggs in

East Jakarta.

28

Exposure through Consumption of Poultry Products

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Anyone eaten poultry in past 7

days*

78.5 69.2 72.6

Anyone eaten raw poultry products

(mostly eggs) in past 7 days

9.3 6.8 7.7

Main source of poultry products*

Supermarket

Wet market

Neighbor/friend/relative

Warung

Local farmer

Mobile vendor

Own production

Other (restaurant, slaughterer,

agent)

3.4

54.6

1.6

21.3

1.0

14.3

0.2

2.9

1.2

30.4

1.9

17.9

6.1

36.2

2.2

3.9

2.0

39.6

1.8

19.3

4.2

28.1

1.5

3.6

Potential exposure to

H5N1 through

consumption of

poultry products is

high. Nearly ¾ of

HHs reported

consuming poultry

products in the past

week and 8%

(slightly more in East

Jakarta than in

Bogor) reported

consuming raw

poultry products—

typically uncooked

eggs. Households

rely more on wet markets as a source of poultry products in E. Jakarta, while

households rely more on mobile vendors as a source in Bogor.

7/11/2012

15

29

Exposure Through Consumption of Eggs

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Anyone purchased eggs in past 7 days* 84.5 80.8 81.7

Main source of eggs*

Supermarket

Wet market

Neighbor/friend/relative

Warung

Local farmer

Mobile vendor

Own production

Other (restaurant, slaughterer, agent)

5.0

19.6

0.7

65.8

0.3

1.1

0.9

5.3

1.8

7.0

0.6

79.3

1.1

5.5

0.3

4.0

3.0

11.7

0.7

74.3

0.8

3.9

0.5

4.5

Over 80% of households purchased eggs in the past week. Warungs are major

source of eggs, somewhat more in Bogor than in Jakarta, while wet markets

are an important source of eggs in East Jakarta.

30

Exposure through the Wet Market

East Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Ever Past 7

days

Ever Past 7

days

Ever Past 7

days

Does anyone in HH go to wet market?1, 2 93.6 62.9 39.8 39.8 48.3 48.3

Who?

Head of household2

Spouse1, 2

Male child

Female child2

Other male adult1

Other female adult2

56.3

75.9

16.0

17.5

10.5

22.4

25.7

46.9

4.4

5.9

2.6

10.2

54..4

72.5

16.0

20.0

12.4

15.0

18.9

25.9

3.2

4.5

3.0

4.3

55.1

73.8

16.0

19.1

11.7

17.7

21.4

33.6

3.6

5.0

2.9

6.5

1 Ever—district difference is significant2 Past 7 days—district difference is significant

Wet markets pose potentially the greatest risk of exposure to H5N1. Nearly half of all households reported that someone in the family went to a wet market in the past week. This was higher in East Jakarta where nearly 2/3 of households said that someone went to the wet market in the past week; 94% reported EVER going to the wet market. Most often it is the female spouse who goes to the wet market, followed by the head of household (55%), but significant proportions of male (16%) and female (19%) children are also exposed.

7/11/2012

16

31

Type of Wet Market Exposure in Past 7 Days

Type of exposure

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Been exposed to live birds

Handled live birds

Handled slaughtered poultry

Exposed to wild birds or bird feces

Visited traditional market

Bought poultry meat at traditional market

42.5

6.2

40.8

16.9

33.1

36.4

21.9

2.7

12.4

52.3

14.8

16.6

31.4*

4.0*

22.9*

39.2*

21.6*

23.9*

Types of exposure in the wet market varied somewhat between Bogor and East

Jakarta. In Bogor, those going to the wet market were more likely to be

exposed to wild birds or bird feces, while in East Jakarta exposure to live birds,

handling of slaughtered poultry and buying poultry meat were more common.

32

POSSIBLE EXPOSURE TO AI

7/11/2012

17

33

Household Exposure to Avian Outbreaks

Type of exposure

East

Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Own chickens:

Any chickens fallen sick past 6

months?

How many died? (Half or more)

Any sudden deaths past 6

months?

8.0

0.7

11.2

21.9

5.5

16.9

31.4*

4.8*

16.0

Chickens in neighborhood:

Any chickens died past 6

months?

How many died? (Half or more)

Were any culled?

7.4

0.0

1.1

32.1

2.7

6.6

22.0*

2.3

5.8

The HUS asked a

series of questions

about deaths of

chickens in their own

flock or in the

neighborhood, which

could indicate an AI

outbreak. Households

in Bogor were more

likely to report that

some of their birds had

died and that birds had

died in their

neighborhood.

They were also more likely to report that the mortality involved half or more of their

own flock; 6% in Bogor reported those higher mortality rates, which was significantly

higher than in East Jakarta.

34

In an effort to examine

links between

hospitalization and wet

market exposure, the

survey asked whether

those hospitalized with

any of the main AI

symptoms (fever, cough

or shortness of breath) in

the past year had been

exposed to slaughtered

poultry at the wet market

in the past week. Of

those hospitalized for AI

symptoms, 39% reported

wet market exposure

while 28% reported no

wet market exposure.

Hospitalization and Exposure to Wet Markets

7/11/2012

18

35

Similarly, those

hospitalized for AI

symptoms in the past year

were more likely to have

shopped at a traditional

market in the past week.

Although these questions

used a different time frame

(past year, past week), the

correlation suggests a

pattern worthy of further

investigation.

Hospitalization and Exposure to Wet Markets

36

KNOWLEDGE & ATTITUDES

ABOUT AI & ILI

7/11/2012

19

37

General Knowledge of AI Risk

General knowledge of avian influenza is high:

• 97.2% of respondents say they have heard of AI

• 84.4% of those who know about AI know that

humans can get AI/flu burung

38

Knowledge of AI

Transmission

ModesHowever, specific knowledge of

how AI is transmitted was

superficial: 47% said AI was

transmitted through contact with

sick birds, 31% said through

general contact with poultry and

13% said through contact with

dead chickens.

Overall, no other mode of

transmission was mentioned by

more than 10% of respondents,

although residents in East Jakarta

were more likely to mention

contact with live chickens and

feces at the wet market and

consuming undercooked sick

chickens or eggs.

Mode of transmission mentioned

East

Jakarta

(n=1260)

Bogor

(n-1260)

Total

(n=2520)

Contact with…

Sick chickens

Sick chicken feces

Sick fighting cocks

Dead chickens

Sick wild birds

Sick chicken meat at home

Sick chicken meat wet market

Live chickens or feces at wet market

A person who has AI

With poultry in general

Feathers

Saliva

43.3

0.6

2.0

14.7

8.5

5.6

1.1

10.8

7.4

32.1

6.8

5.3

49.5

0.5

0.5

13.1

1.0

5.2

0.3

2.7

8.6

29.3

6.7

2.7

46.6

0.5

1.2*

13.8

4.5*

5.4

0.6*

6.5*

8.1

31.3

6.8

4.5

Feeding or caring for sick chickens 2.4 1.4 1.8

Eating…

Undercooked sick chicken or eggs

Raw chicken products

Sick chicken meat

Poultry in general

12.3

2.6

5.3

3.7

6.6

1.8

10.7

5.3

9.3*

2.2

6.8

6.8

7/11/2012

20

Symptom mentioned Seasonal flu Flu burung

Difficult/fast breathing* 2.8 13.2

Tight chest/congestion* 0.4 2.4

Fever 66.1 59.3

Cough* 33.9 11.4

Muscle ache* 2.4 1.5

Sore throat* 16.7 6.1

Stuffed/blocked nose* 29.6 5.6

Runny nose* 34.4 8.3

Sneezing* 44.7 7.5

Earache/ear discharge 0.1 0.1

Rash 0.5 1.0

Nausea/vomiting* 1.2 2.8

Seizures* 0.1 1.4

Knowledge of AI vs Seasonal Flu Symptoms

39

Regarding knowledge of symptoms of AI, the HUS data suggest that people may recognize some differences between AI and seasonal influenza.

They are more likely to correctly associate difficulty breathing and chest congestion with AI, while considering cough, nasal congestion and sneezing as more common in seasonal influenza.

Unfortunately, only about 15% associate difficulty breathing with AI, even though it is one of the most important signs besides fever.

40

Perceived Severity of AI and Seasonal Flu

mean s.e. p

Likelihood of dying from seasonal influenza

Likelihood of dying from AI

1.61

3.75

0.06

0.04

Difference between AI

and seasonal flu

is significant.

AI is perceived to be a more serious illness than seasonal influenza with

people rating AI more than twice as high on a six-point perceived severity

scale.

7/11/2012

21

41

CARESEEKING KNOWLEDGE & BEHAVIOR

42

Preferred Service Delivery Sites

East Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Men Women Men Women Men Women

Preferred source of services

Public hospital

Private hospital

Sub-district puskesmas

Village puskesmas

Other government clinic

Private clinic

Pharmacy/shop

Traditional healer/dukun

Midwife

Other

27.1

17.9

26.0

27.3

1.8

44.4

12.9

8.0

2.4

17.1

25.3

18.7

31.6

32.2

0.9

41.8

11.2

6.1

12.2

13.2

9.0

14.7

26.4

23.1

0.6

30.9

8.7

4.1

9.5

42.8

8.6

15.6

28.8

30.4

0.8

28.5

6.0

6.7

30.1

34.0

15.6

15.9

26.2

24.9

1.1

35.8

10.2

5.5

6.9

33.5

14.8

16.7

29.8

31.1

0.8

33.4

7.9

6.5

23.5

26.4

General preferences for service providers favor the sub-district puskesmas,

village puskesmas, private clinics and midwives.

7/11/2012

22

43

Requesting ServicesPercent

Seasonal fluEver asked provider to test you?Ever been told by provider they wanted to test you?Ever heard of vaccine?Where would you go to get it? (of those who know of vaccine)

Public hospitalPuskesmasPrivate hospitalPrivate clinicPrivate practitionerOther (mostly Puskesmas or Posyandu)

1.10.9

16.5

29.52.3

18.86.98.6

49.1

Flu burungEver asked provider to test you?Ever been told by provider they wanted to test you?Ever heard of Tamiflu?Is it effective? (among those who know of Tamiflu)

0.30.9

12.730.5

Knowledge of testing and treatment for both seasonal flu and AI are low. Only 1% or less of respondents said they had ever requested or been offered testing for either illness. 17% reported knowledge of seasonal flu vaccines, while 13% had ever heard of Tamiflu, but only 30% of those who knew about Tamiflu said they thought it was an effective treatment.

44

Experience of Acute Illness in Past 2 WeeksEast

Jakarta Bogor Total

Has anyone in the household been seen by a health care provider for acute illness in the past two weeks?

26.4 27.5 27.1

Symptoms of patients with acute illnessFever / feverishness / chills

CoughChest pain

Difficulty breathing/shortness of breath

38.436.52.12.3

46.138.93.34.1

40.238.02.93.5

What was the diagnosed condition?Seasonal flu 30.5 36.5 34.5

Was patient tested for…Seasonal flu

Flu burung0.9

0.5

0

0

0.4

0.2

Acute illness was fairly common in the households in the study area. About 27% reported that someone in the household had seen a provider due to acute illness in the past two weeks. The symptoms that prompted care seeking were predominantly fever (40%) and/or cough (38%). Around 3% mentioned chest pain or difficulty breathing. In about 1/3 of the cases, the provider diagnosed the condition as seasonal flu but almost never performed a test to confirm a diagnosis.

7/11/2012

23

45

Care seeking NormsWhat should you do if you/someone in

your family had…?

Fever

& cough

Fever, cough &

difficult breathing

You Family You Family

Bring to health center immediately 13.7 18.7 31.1 30.7

Bring to doctor immediately 14.9 23.0 47.4 49.6

Give medicine at home 14.3 12.4 4.6 4.2

Give fluids at home 4.9 2.3 0.6 0.2

Buy treatment from pharmacy or shop 20.1 19.0 7.5 7.0

Contact kader 0.06 0.12 0 0

Traditional treatment or get from dukun 3.2 1.8 1.4 1.3

Other (mostly use OTC treatment) 28.9 22.7 7.4 7.0

To explore care-seeking norms, the HUS asked respondents what should be done in case of fever & cough or fever, cough & difficulty breathing. People were more likely to say they should seek care immediately from the puskesmas or doctor if the more severe symptoms occurred. Thiswas true for both self and family members.

There was some evidence of preference for early self-treatment (giving medicines or fluids at home, seeking a remedy from a pharmacy or dukun), particularly when the symptoms were a little less severe.

46

Response to Acute Illness Events in Past 2 Weeks

East Jakarta

(n=1260)

Bogor

(n=1260)

Total

(n=2520)

Fever &

cough

Fever,

cough,

difficulty

breathing

Fever &

cough

Fever,

cough,

difficulty

breathing

Fever &

cough

Fever,

cough,

difficulty

breathing

Has anyone in the household

suddenly fallen ill with…

13.9 1.6 20.5 4.4 18.1 3.3

What did you do?

Waited

Gave medicine/fluids at home

Went to a pharmacy/shop

Went to private clinic

Went to healer/dukun

Went to private hospital

Went to puskesmas

Went to sub-district puskesmas

2.0

47.9

16.7

10.4

0.9

1.0

10.9

10.4

0

37.6

20.2

21.9

3.5

5.4

2.1

13.9

2.7

29.4

19.4

7.2

2.8

3.2

9.4

12.5

2.7

14.8

22.6

15.1

0

3.6

7.8

12.6

2.5

34.6

18.6

8.1

2.3

2.6

9.8

11.9

2.2

18.7

22.2

16.3

0.6

3.9

6.9

12.9

Regarding actual careseeking behavior, of those who suddenly fell ill with fever

and cough, more than half chose to self-treat at home or use remedies from a

pharmacy. Over 40% chose to self-treat even when the symptoms included

shortness of breath.

7/11/2012

24

47

Choice of Healthcare Provider for Fever, Cough, or Difficult Breathing

Of households where someone experienced an acute respiratory event, 62% (n=445) sought treatment (not shown).

Among households that sought treatment from a healthcare provider for these symptoms, private clinics were most commonly used (27%). Sub-district puskesmas (18%), village puskesmas (14%), and private doctors (13%) were used fairly commonly. It is notable that midwives are also visited for these symptoms (9%), slightly more so than private hospitals (7%) or public hospitals (4%).

Total(weighted)

%

Sources of professional care for household members seeking treatment for fever, cough, or difficult breathing in previous 2 weeks

First care-seeking household member (n=390)Private clinic

Sub-district puskesmas

Village puskesmas

Private doctor

Midwife

Public hospital

Private hospital

Nurse

Traditional healer

Second care-seeking household member (n=55)Private clinic

Sub-district puskesmas

Village puskesmas

Private doctor

Midwife

Public hospital

Private hospital

Nurse

Traditional healer

105

69

51

50

38

15

30

4

4

9

12

8

9

2

3

0

1

0

26.8

17.7

13.0

12.8

9.9

4.0

7.6

1.0

1.1

17.2

21.8

15.8

17.1

4.3

5.7

0.0

1.90.0

48

Choice of Healthcare Provider for other Acute SymptomsIn the prior two weeks, slightly fewer than 40% of respondents who sought care for an acute health event did so for symptoms that were not characteristic of AI (diarrhea, sore throat, nasal congestion, pneumonia or respiratory infection, headache, body ache, and others).

The choice of healthcare facility was similar to facilities used by patients with potential AI symptoms. The most popular source of care for these acute symptoms was a private clinic (20%, n=294 patients). As with the other sample, there was little difference in popularity between sub district puskesmas (16%), village puskesmas (15%), and private doctors (15%). Among these patients, midwives used by fewer patients (7%) than public or private hospitals (14% and 8%, respectively).

Total(weighted)

%

Sources of professional care for household members seeking treatment for other acute symptoms in previous 2 weeks

First care-seeking household member (n=273)Private clinic

Sub-district puskesmas

Village puskesmas

Private doctor

Midwife

Public hospital

Private hospital

Nurse

Traditional healer

Second care-seeking household member (n=21)Private clinic

Sub-district puskesmas

Village puskesmas

Private doctor

Midwife

Public hospital

Private hospital

Nurse

Traditional healer

55

46

40

40

22

38

23

14

3

6

0

4

4

0

2

0

2

1

20.1

17.0

14.6

14.6

7.9

13.9

8.3

5.0

1.0

27.6

0.0

19.0

18.2

0.0

8.9

0.0

7.9

4.0

7/11/2012

25

49

E.

Jakarta

(%)

Bogor

(%)

p

Reason for choosing specific healthcare

facility for fever/cough (n=190)

Quality of medicine

Accessibility of facility

Cost of services

Cost of medication

Hours/days of service

Quality of providers

Waiting time

Privacy

Availability of medicine in facility

How patients are treated

50.0

30.3

18.2

14.1

10.1

6.3

4.1

3.1

2.3

2.1

27.8

25.3

15.8

16.1

3.3

2.3

0.5

2.9

0.8

0.0

0.04

-

-

-

-

-

0.03

-

-

-

Important Factors in Choosing Healthcare Facility

Quality, accessibility, and cost were the most frequently cited reasons for choosing a facility among those households that sought treatment for patients with fever and cough.

Caregivers in East Jakarta were significantly more selective based on the perceived quality of medicine; half of the respondents provided this reason, compared to 28% in Bogor. Waiting time was significantly more important to East Jakarta respondents (4.1%) than respondents in Bogor (0.5%).

50

Symptomatic Households Not Seeking Care

E.

Jakarta

(%)

Bogor

(%)

p

Reasons for not seeking care at a facility for

fever/cough (n=228)

Did not feel they were sick enough

Received care at home

Felt patient was getting better on their own

Too expensive (can’t afford)

Didn't’have the time

No transportation

Too far to travel

Patient is too old to seek care

Work wouldn’t allow it

Nobody to take care of children

54.6

28.7

22.4

5.5

2.9

0.0

1.8

1.4

0.9

0.0

45.5

5.6

17.2

15.7

1.8

3.1

2.4

0.0

0.0

0.0

-

<0.001

-

0.03

-

-

-

-

-

-

Among households that reported NOT seeking care for members with fever, cough and difficulty breathing, the largest proportion (roughly half) said they didn’t think the patient was sick enough to need professional care.

Two reasons for not seeking care varied significantly by district. In East Jakarta, significantly more respondents chose to treat the patient at home (28.7%) compared to 6% in Bogor. A significantly higher proportion of respondents in Bogor said they did not seek treatment because they could not afford it (16%), compared to almost 6% in East Jakarta who reported cost as a reason.

7/11/2012

26

51

Waiting Time Before Seeking Facility-Based CareEast Jakarta

(n=46)

Bogor

(n=145)

Total

(n=191)

How long did you wait?

Less than 1 day

One to two days

Three to four days

More than 4 days

Never went to a facility

24.0

59.4

13.0

0.6

3.0

16.6

42.9

15.0

1.2

24.3

18.4

46.8

14.5

1.1

19.2

Why did you wait more than two days?

Didn’t think it was a serious illness

Thought it was a common illness

(n=7)

31.5

31.2

(n=24)

31.7

23.3

(n=31)

31.6

25.1

Why did you wait less than 2 days?

Thought it was a dangerous disease

Feel it is important to go to facility

(n=42)

19.7

63.5

(n=95)

45.6

40.6

(n=137)

37.7

47.6

Other reasons for waiting more than two days (open response)

Tried home/over-the-counter treatment first; illness didn’t respond after 2-3 days (n=6)

Waiting until they had enough money (n=4)

Logistics (transportation, availability) (n=3)

Patient is too young (<6 months old) (n=1)

Regarding waiting time to seek care for acute respiratory symptoms, more than 65% said they sought

care within 2 days, but 20% (mostly in Bogor) said they never went to a facility for care. Reasons for

waiting longer than two days were low risk perceptions (common illness, not serious) while reasons for

seeking care quickly were higher risk perception (thought it was dangerous) or a careseeking norm (felt

it was important to seek care).

52

Careseeking Knowledge: Wait or Not?E. Jakarta

(mean)Bogor (mean)

p

If fever or cough, best to seek immediate care.

(0=completely disagree; 5=completely agree)

Households with previous sudden fever/cough

Households with no previous sudden fever/cough

3.95

3.86

4.23

4.13

0.01

<0.001

Best to wait a few days upon falling ill before seeking

treatment; might take care of itself.

(0=completely disagree; 5=completely agree)

Households with previous sudden fever/cough

Households with no previous sudden fever/cough

1.73

1.65

1.64

1.90

-

<0.001

Previous household experience with sudden onset of fever/coughing and urban city each significantly

predicted respondents’ recognition of the need for immediate care. Overall, households that had

previously experienced sudden fever and coughing were more likely to say that it was best to seek

immediate care (mean=4.15) compared to households with no previous experience with these

symptoms (mean=4.03, p=0.02). Mean scores were higher in Bogor than in East Jakarta.

Households in East Jakarta) that had never experienced sudden fever/coughing were least likely to say

that it was best to seek immediate care for these symptoms, whereas households in Bogor that had

experienced sudden fever/cough were most likely to say it was best to seek immediate care.

7/11/2012

27

53

Careseeking Knowledge: Safe Waiting Time

E. Jakarta(mean)

Bogor(mean)

p

Number of days it is safe to wait after falling ill

Households with previous sudden fever/cough

Households with no previous sudden fever/cough

1.07

1.02

1.61

1.74

<0.01

<0.001

Number of days since falling ill it is NOT safe

(dangerous) to wait

Households with previous sudden fever/cough

Households with no previous sudden fever/cough

2.07

2.01

2.62

2.78

<0.01

<0.001

Despite expressing stronger beliefs in the importance of seeking immediate care, the amount of time

perceived to be safe to wait before seeking care was significantly higher among Bogor households

compared to those in E. Jakarta, both among households that had experienced fever/cough and those that

had not. Households in each district, regardless of previous experience with fever/cough, perceived a safe

window of time to be under 48 hours.

On average, households in Bogor (especially those with no previous experience with sudden fever/cough)

believed that it only became dangerous to wait to seek care until approximately 65 hours had passed after

developing symptoms. In contrast, households in E. Jakarta were significantly more accurate in identifying

the 48-hour time period (citing approximately 49 hours as being the point at which it becomes dangerous

to wait to seek care).

54

TOWARD BEHAVIORAL MESSAGING

What predicts knowledge, risk perceptions,

efficacy and careseeking?

7/11/2012

28

55

What Makes AI Salient?Recognizing Types of Exposure that Require Vigilance

Generally speaking, people are unlikely to practice protective health behaviors if a particular

health threat is not salient. Knowledge of and risk perceptions related to AI are known to

predict lower risk poultry rearing, poultry handling and hygiene behaviors, so it is likely that

they predict careseeking behavior as well. This section examines relationships between AI

knowledge, risk perceptions, efficacy and careseeking behavior.

Because there were no identified cases of AI in this sample, analysis of reactions to actual AI

cases was not possible.

As a proxy, we examined patterns of potential exposure in households that had ever or never

previously experienced possible symptoms of AI (sudden onset of fever/coughing, difficulty

breathing), then analyzed household characteristics, attitudes, risk perception and efficacy

associated with careseeking.

First, we examined patterns of potential exposure that result from routine household

shopping and poultry handling practices in East Jakarta and Bogor. In both districts, the only

factor that was different between households with and without sudden onset of fever and

cough was contact with another symptomatic patient within the past 7 days (not a known

exposure pathway for AI).

See tables on the following two pages.

56

Sudden

fever/cough

(ever)

No sudden

fever/cough

(ever)

p

Household raises:

Chickens

Ducks

Turkey, geese, entog

Birds

Poultry indoors:

Poultry kept indoors at any time (caged or uncaged)

Poultry roam freely indoors

In past 7 days:

Exposed to birds at wet market

Handled live birds at wet market

Handled slaughtered birds at wet market

Exposed to wild birds/bird feces

Visited traditional market with live poultry

Bought poultry meat at traditional market

In contact with symptomatic person

(%)

5.6

1.1

0.0

10.8

26.4

6.5

49.1

4.7

43.1

24.0

36.2

41.8

9.1

(%)

8.9

1.0

0.1

11.5

19.6

12.2

47.6

6.3

40.5

15.9

32.9

35.8

2.9

-

-

-

-

-

-

-

-

-

-

-

-

<0.01

Presence of Potential Exposures in Households

by Previous Experience with Symptoms:

East Jakarta

7/11/2012

29

57

Sudden

fever/cough

(ever)

No sudden

fever/cough

(ever)

p

Household raises:

Chickens

Ducks

Geese

Turkey

Birds

Poultry indoors:

Poultry kept indoors at any time (caged or uncaged)

Poultry roam freely indoors

In past 7 days:

Exposed to birds at wet market

Handled live birds at wet market

Handled slaughtered birds at wet market

Exposed to wild birds/bird feces

Visited traditional market with live poultry

Bought poultry meat at traditional market

In contact with symptomatic person

(%)

24.5

2.7

1.3

0.1

6.7

23.2

48.8

25.3

3.4

14.6

54.8

36.2

19.4

17.1

(%)

27.7

1.8

0.3

0.0

6.2

15.5

32.5

21.1

2.5

11.9

51.8

32.9

15.9

8.6

-

-

-

-

-

-

-

-

-

-

-

-

-

<0.01

Presence of Potential Exposures in Households

by Previous Experience with Symptoms:

Bogor

Effects of Knowledge and Risk Perceptions on Poultry

Keeping Practices

58

Although there were no significant differences in the occurrence of AI symptoms by

potential sources of H5N1 exposure, there was significant variation in knowledge of AI

and risk perception among households with different poultry-keeping practices. These

relationships often remained when controlling for indicators of socioeconomic status

(educational level, wealth as indicated by total household goods, and district).

In this sample, educational levels and number of household goods were lower among

households that owned chickens compared to households without chickens (p<0.001),

as well as among those that allowed poultry to roam indoors versus households that

owned poultry, but did not allow free indoor roaming (p<0.001). Socioeconomic

characteristics did not vary significantly between households that kept poultry in the

household (i.e., including caged) compared to those that only kept poultry outside the

house (not shown).

7/11/2012

30

59

Effect of Owning Chickens on AI Awareness & Knowledge

HHs with

Chickens

(n=406)

HHs

without

Chickens

(n=2114)

p

General Awareness

Awareness of AI (%)

Awareness that humans can get AI (%)

Knowledge of transmission modes

Contact with sick chicken

Chicken feces

Visiting wet market

Contact with sick wild birds

Handling sick chicken in wet market

Undercooked sick chicken/eggs

Humans (misconception)

92.2

85.5

50.1

1.4

0.0

1.4

0.0

10.0

2.9

95.0

93.8

49.2

0.4

0.8

5.2

0.8

5.6

9.1

0.01

<0.001

n.s.

n.s.

n.s.

0.003

0.04

0.03

0.01

Comparing households

that raised chickens (the

most common type of fowl

raised by this sample) to

those without chickens

revealed significant

differences in knowledge,

but chicken ownership was

not a strong predictor of

aspects of care-seeking or

risk perception.

Households that raised

chickens had significantly

lower awareness of AI as well as lower awareness that humans could contract AI (when controlling for

socioeconomic variables, p=0.001). Awareness of some transmission vector differed slightly (but

significantly). Notably, households without chickens incorrectly identified human transmission as an

exposure pathway for AI more than households with chickens while households with chickens were

more likely to mention consuming undercooked poultry products from sick chickens as a transmission

vector.

60

Effect of Owning Chickens on Care-Seeking Attitudes

and BeliefsHHs with

Chickens

(n=406)

HHs without

Chickens

(n=2114)

p

Care-Seeking

Reported likelihood of seeking care for fever/sore throat/difficult

breathing AND in contact with birds in prior 2 weeks (mean)

Reported likelihood of seeking care for fever/sore

throat/difficulty breathing AND been to wet market (mean)

Number of days respondent feels it’s safe to wait before seeking

care after someone falls ill (mean)

Number of days respondent feels it becomes dangerous if care-

seeking is delayed (mean)

3.47

3.41

1.66

2.68

3.69

3.61

1.41

2.43

0.02

0.02

0.053

0.056

In households with chickens, respondents were less likely to say they would seek care for

fever/cough/difficulty breathing even in the patient had been in recent contact with birds or at the the

wet market. They were more likely to say they would wait longer to seek care. However, none of those

differences were significant after controlling for socioeconomic status.

7/11/2012

31

61

Among all households with poultry (n=617), those that kept their poultry inside were less

likely to seek care fever/cough compared to households that kept their poultry outside,

even when controlling for wealth, education, and district (p=0.03) (not shown).

Nearly sixteen percent (15.8%) of households with poultry kept indoors sought care when

household members had fever/cough, compared to 45.2% of households that kept their

poultry outside the home. There were no other significant differences in knowledge, risk

perception, or care-seeking between households that raised poultry outside the house only,

compared to those that kept poultry inside the house (not shown).

However, these associations are particularly strong in households where poultry is allowed

to roam freely inside the household, instead of being caged. Those who allow poultry into

the house are less knowledgeable about some risk factors, less likely to believe that

immediate careseeking for fever/cough was necessary, less likely to believe they could

protect their family and seek necessary care, more likely to rely on traditional forms of

treatment and less likely to seek professional care. This suggests some complacency

regarding risk among those who live in closer proximity to birds.

Effects of Poultry Keeping Practices on Careseeking

62

Indoor Roaming Associated with

Lower Knowledge about AIAllow

Poultry to

Roam

Indoors

(n=156)

No

Indoor

Roamin

g

(n=458)

p

General awareness

Awareness of AI (%)

Awareness that humans can get AI (%)

Knowledge of transmission modes

Contact with sick chicken

Chicken feces

Visiting wet market

Handling sick chicken

Feeding/caring for sick chicken

Undercooked sick chicken/eggs

91.0

81.5

50.1

1.4

0.0

0.5

0.3

1.3

95.0

88.7

49.2

0.4

0.8

4.5

3.2

8.5

n.s.

0.05

n.s.

n.s.

n.s.

0.04

0.04

0.03

The complacency

hypothesis is supported

by evidence that

households that allowed

indoor roaming of poultry

(30% of households with

poultry) were less likely to

know that humans could

contract AI less likely to

mention that handling,

feeding, or caring for sick

chickens, and eating

undercooked sick chicken

or eggs were risk factors

for contracting AI.

After controlling for socioeconomic status (education, wealth, and district), awareness that humans

can get AI, as well as feeding or caring for sick chickens, no longer predicted letting poultry roam

indoors, but knowledge about handling sick chickens (OR=0.09, p=0.05) and knowledge that

undercooked sick chicken or eggs (OR=0.17, p=0.04) could transmit AI remained significant.

7/11/2012

32

63

Indoor Roaming Associated with

Lower Risk Perception and Lower Self-EfficacyAllow

Poultry to

Roam

Indoors

(n=156)

No Indoor

Roaming

(n=458)

p

Risk Perception / Self-Efficacy

(0=completely disagree; 5=completely agree)

Belief that immediate treatment is necessary for fever/cough (mean)

Confidence in protecting self/family from AI (mean)

Confidence in seeking medical care when needed (mean)

3.82

3.25

3.61

4.13

3.68

3.97

0.02

0.02

<0.001

Similarly, households that allowed indoor roaming of poultry had significantly weaker beliefs in the necessity of

immediate treatment for fever/cough, and were significantly less confident about their ability to seek medical care

when needed. Despite the fact that the majority of respondents in both groups (68% with indoor roaming and 77%

with no roaming, n.s.) directly cared for their own poultry, those in households with indoor roaming still expressed

lower confidence that they could protect their household from AI.

After controlling for socioeconomic status (education, wealth, and district), confidence remained significantly lower

among households with indoor roaming that they could protect themselves and their family from AI (p=0.03) and

that they could seek medical care when needed (p=0.001). These households also continued to express lower

agreement with the importance of seeking immediate care for fever/cough when controlling for socioeconomic

status (p=0.01).

64

Indoor Roaming Associated with

Lower Likelihood of Seeking CareAllow Poultry to

Roam Indoors

(n=156)

No Indoor

Roaming

(n=458) p

Care-Seeking (0=highly unlikely; 5=highly likely)

Reported likelihood of seeking care for fever/sore

throat/difficulty breathing (mean)

Reported likelihood of seeking care for fever/sore

throat/difficult breathing AND in contact with birds in

prior 2 weeks (mean)

3.48

3.31

3.66

3.62

n.s.

0.03

Careseeking intentions in the event of fever, sore throat and difficulty breathing did not significantly differ

according to indoor roaming in the household except among those that contact with birds in the previous

two weeks; this was significantly compared to other households that also owned poultry (contained or

kept outside). However, these differences were no longer significant after controlling for socioeconomic

status.

Households that allow indoor roaming of poultry were much more likely to use traditional treatment or

obtain treatment from a dukun compared to households that did not allow indoor roaming (OR=17.6,

p=0.014). The association was even stronger when controlling for district (OR=22.3, p<0.001).

Household wealth clearly plays a role in the likelihood of protective behaviors such as poultry keeping

practices and careseeking, but it does not appear to be the only factor affecting careseeking.

7/11/2012

33

65

Confidence about Self-Protection and Caregivers by

Previous Hospitalization Experiencemean s.e. p

Confidence in protecting self from AI (range 0-5, disagree to agree)

Among households (hh) with…

Hospitalizations for AI symptoms

Hospitalizations for other symptoms

No hospitalizations

Multiple members hospitalized

Among hh with any deaths

3.62

3.36*

3.72

3.62

3.16

3.62

0.03

0.04

0.09

0.04

0.49

0.16

0.02

If have AI, confident health care provider will give proper care

(0=completely disagree, 5=completely agree)

Among households with…

Hospitalizations for AI symptoms

Hospitalizations for other symptoms

No hospitalizations

Multiple members hospitalized

Among hh with any deaths

3.76

3.73

3.82

3.75

3.94

3.65

0.03

0.11

0.08

0.03

0.26

0.13

ns

Households where someone had been hospitalized with AI-like symptoms were less confident

about self-protection, perhaps they have experienced illness directly and realize that infection is

possible. Hospitalization did not affect confidence in caregivers.

66

Perceived Susceptibility by Previous

Hospitalization Experiencemean s.e. p

Likelihood of contracting AI

Among households…

Observed by surveyor to own poultry

Observed by surveyor to have poultry in house

Where >50% of flock had died from illness

That allow poultry to roam indoors

Among respondents who say AI is contracted via exposure to:

Contact with sick chicken

Contact with chicken feces

Handling chicken in wet market

1.66

1.75

1.77

1.78

1.85

1.70

1.32

1.09

0.05

0.08

0.12

0.19

0.12

0.06

0.39

0.30

ns

0.06

Consistent with earlier analyses, households observed to have chickens or to have them in the

house or that had previously experienced a sudden episode of mortality among their poultry

were not significantly more likely to believe they were susceptible to AI. Those who knew that AI

could be transmitted through contact with chicken feces or by handling chickens in the wet

market felt marginally less susceptible to AI, suggesting that they may have a false sense of

safety about buying poultry at the wet market.

7/11/2012

34

67

Careseeking Self-Efficacy by Poultry Ownershipmean s.e. p

Confident can seek care if needed

Among households…

Observed by surveyor to own poultry

Observed by surveyor to have poultry in house

That own poultry (self-report)

With no poultry (self-report)

That allow poultry to roam indoors

3.87

3.78

3.99*

3.87

3.61**

0.07

0.10

0.03

0.07

0.11

ns

0.04

<0.001

Confident can tell when need to seek care

Among households…

Observed by surveyor to own poultry

Observed by surveyor to have poultry in house

That own poultry (self-report)

With no poultry (self-report)

That allow poultry to roam indoors

3.82

3.78

3.83

3.89

3.62**

0.05

0.07

0.05

0.03

0.10

Ns

ns

<0.01

Poultry owners were more likely to express confidence in their ability to seek care if

needed, while those who allow poultry to roam indoors were less confident about seeking

care and knowing when to seek care.

68

Self-efficacy about Recovery and Self-Protection by Poultry Ownership

mean s.e. p

People with AI can recover if brought to puskesmas quickly

Among households…

Observed by surveyor to own poultry

Observed by surveyor to have poultry in house

That own poultry (self-report)

With no poultry (self-report)

That allow poultry to roam indoors

3.44

3.73*

3.44

3.38

3.52

0.06

0.10

0.06

0.04

0.12

<0.001

Confident can protect self from AI

Among households…

Observed by surveyor to own poultry

Observed by surveyor to have poultry in house

That own poultry (self-report)

With no poultry (self-report)

That allow poultry to roam indoors

3.56

3.30*

3.55

3.64

3.25*

0.07

0.15

0.07

0.04

0.15

0.03

0.02

Households observed to have poultry in the house were more confident than other groups in

their ability to recover from AI if brought to the puskesmas quickly. But they were LESS confident

about self-protection, suggesting that they don’t think there is much they can do to prevent

infection and choose to rely instead on rapid treatment if they become infected. As earlier

analysis showed, however, this group is also MORE likely to rely on traditional remedies and less

likely to seek care at all (see Slides 60 and 64).

7/11/2012

35

Other Careseeking Predictors: HH Characteristics

69

Among households that

had experienced fever and

cough, there were no

significant differences in

demographics (age of

respondent, education,

wealth, district),

knowledge of AI, or risk

perception between those

that sought treatment and

those that did not.

Sought

care for

fever/cough

(n=146)

Did not seek

care for

fever/cough

(n=231) p

DEMOGRAPHICS

Age (mean)

EducationNone

Some primary school

Primary school

Junior high school

High school

Academy

University

Number of household goods (mean)

DistrictEast Jakarta (% within district)

Bogor (% within district)

41.5

7.1

14.1

33.6

17.7

20.0

5.1

2.4

9.9

40.6

37.7

39.9

5.8

17.2

26.8

14.4

24.9

5.3

5.5

10.7

59.4

62.3

-

-

-

-

Other Careseeking Predictors: Knowledge & Risk Perception

70

Sought

care

for

fever/c

ough

(n=146)

Did not

seek

care for

fever/c

ough

(n=231)

p

Knowledge

Mention fever OR cough as AI symptoms (%)

Mention fever AND cough as AI symptoms (%)

Mention fever, cough, OR difficulty breathing as

AI symptoms (%)

Risk Perception

Perceived likelihood of death from AI if infected

(mean; 0=highly unlikely, 5=highly likely)

Perceived likelihood of self/family being

infected with AI (mean; 0=highly unlikely,

5=highly likely)

Belief that recovery from AI is possible with

rapid treatment (mean; 0=strongly disagree,

5=strongly agree)

61.5

11.0

39.9

3.74

1.85

3.54

58.1

11.1

36.7

3.87

1.62

3.40

-

-

-

-

-

-

Among households that

had experienced fever

and cough, there were

no significant

differences in measures

of knowledge or risk

perception between

those that sought

treatment and those that

did not.

7/11/2012

36

Predictors of Delay in Careseeking: HH Characteristics

71

Most of respondents who visited a facility for fever/cough did so within 48 hours.

Respondents who waited longer to seek care (48 hours or more) either had slightly lower levels of education or were the most educated (i.e., university) (approaching significance, p=0.06).

Time to seek care was significantly associated with wealth; having fewer household goods was associated with longer time to care-seeking (p=0.05)

Sought

care

<48

hours

(n=136)

Sought

care >48

hours

(n=30) p

DEMOGRAPHICS

Age (mean)

EducationNone

Some primary school

Primary school

Junior high school

High school

Academy

University

Number of household goods (mean)

DistrictEast Jakarta (% within district)

Bogor (% within district)

39.7

6.3

11.2

29.0

20.3

22.5

7.0

3.8

10.3

85.9

78.5

44.6

4.4

32.5

37.6

8.0

10.3

0.0

7.1

8.7

14.1

21.5

-

0.0

6

0.0

5

-

Predictors of Delay in Careseeking: Knowledge & Risk Perception

72

Sought

care

<48

hours

(n=136)

Sought

care

>48

hours

(n=30)

p

Knowledge

Mention fever OR cough as AI symptoms (%)

Mention fever AND cough as AI symptoms (%)

Mention fever, cough, OR difficulty breathing as

AI symptoms (%)

Risk Perception

Perceived likelihood of death from AI if infected

(mean; 0=highly unlikely, 5=highly likely)

Perceived likelihood of self/family being

infected with AI

(mean; 0=highly unlikely, 5=highly likely)

Belief that recovery from AI is possible with

rapid treatment

(mean; 0=strongly disagree, 5=strongly agree)

59.4

12.9

78.8

3.76

1.95

3.42

75.0

23.6

87.8

3.77

1.87

3.95

-

-

-

-

-

-

Among households that had sought

treatment for fever and cough, there

were no significant differences in

measures of knowledge or risk

perception between those that

sought treatment and those that did

not.

Although not significant, households

that waited to seek care correctly

identified AI symptoms more often

than households that sought care in

under 48 hours, as well as

expressing a stronger belief that

recovery from AI is possible with

rapid treatment. That these

households did not seek treatment

within the recommended 48 hours

suggests that they be over-

confidence about their ability to

identify and address AI, and

determined that their episodes of

fever/cough were not cause for

alarm.

7/11/2012

37

73

TOWARD BEHAVIORAL MESSAGING:

What predicts careseeking and

waiting time to seek care?

Predictive Models

74

Multivariate Predictive ModelsIn order to identify the factors that most strongly predict timely careseeking behavior and which can be used as the basis for a careseeking communication strategy, we tested several multiple regression models. Variables in these analyses were as follows:

Highest level of education Seven levels, none to completed university

HH wealth N of possessions

District East Jakarta =1, Bogor = 2

Birds allowed to roam free in the house No = 0, Yes = 1

N of sources of exposure to poultry e.g., backyard, wetmarket. Range 0-6, mean 1.5

N of AI media message sources recalled (15 sources) e.g., television, radio. Range 0-7, mean 1.2

Belief that humans can get AI No = 0, Yes = 1

N of correct transmission vectors known (unprompted) Range 0-6, mean 0.8

Anyone in HH had an acute respiratory event in past year No = 0, Yes = 1

Perceived severity of AI (how likely that an infected person will die Range 0-5, mean 3.8

Perceived susceptibility to AI (how likely that someone in your HH

will become infected with AI)

Range 0-5, mean 1.5

Perceived self-efficacy to seek care if needed Range 0-5, mean 4.0

Perceived self-efficacy to protect self and family from AI Range 0-5, mean 3.7

Perceived self-efficacy to know when to seek care Range 0-5, mean 3.9

7/11/2012

38

75

Summary—Predictors of care-seeking*(+ positive relationship, - negative relationship)

Likelihood of seeking care if someone

suddenly develops fever, sore throat and

difficulty breathing+ Perceived self-efficacy to

seek care if needed

Number of AI messages

recalled +

Of all the predictor variables, the two that were significant were number of avian flu

messages recalled out of 15 sources that were used in previous AI prevention

campaigns (television, radio, billboard, poster, flyer/leaflet, booklet, newspaper,

community events, cadre, health facility, doctor, district health office, district animal

health office, community leader (RT/RW) and perceived self-efficacy to seek care if

needed.

Message implications: Strive for the highest reach possible with care seeking messages

and emphasize building confidence in the ability to seek care quickly as soon as

symptoms start to appear.

* Controlling for education, household wealth (possessions), district and 10 other variables.

76

Summary—Predictors of shorter waiting time to seek care*(+ positive relationship, - negative relationship)

Shorter waiting time before seeking care+

Perceived self-efficacy to

seek care if needed

District (Bogor) −

Residents of Bogor district said on average that it was safe to wait longer to seek care

after the appearance of AI-like symptoms (cough, sore throat, difficulty breathing),

while those who perceived AI to be a more severe health threat and those who

expressed greater confidence (self-efficacy) in their ability to seek care if needed gave a

shorter estimate of how long it was safe to wait to seek care after the appearance of AI-

like symptoms.

Message implications: Target residents of Bogor and emphasize how deadly AI can be,

as well as how easy it can be to seek care as soon as symptoms start to emerge.

* Controlling for education, household wealth (possessions), district and 10 other variables.

Perceived severity of AI+

7/11/2012

39

77

Summary—Predictors of self-efficacy to seek care*(+ positive relationship, - negative relationship)

Perceived self-efficacy to seek care

+

Perceived severity of AI

District (Bogor)

Residents of Bogor district expressed greater confidence in their ability to seek care as

did those with greater household wealth. Those who perceived AI to be a more severe

health threat and those who correctly knew more of the ways that AI can be

transmitted were also more likely to express higher self-efficacy to seek care when

needed.

Message implications: Target lower income families, emphasizing how serious an AI

infection can be and the ways in which AI can be transmitted.

* Controlling for education, household wealth (possessions), district and 10 other variables.

Household wealth

+ Number of correct AI

transmission vectors known

+

+

78

Summary—Predictors of self-efficacy to protect self & family*(+ positive relationship, - negative relationship)

Perceived self-efficacy to protect self & family

+

Perceived susceptibility to

AI

Education

Those with higher levels of education and those with higher care seeking self-efficacy were more

likely to say they were confident in their ability to protect themselves and their family from AI.

However, those who reported lower levels of recall of AI messages from previous campaigns and

those perceived themselves as more susceptible to AI infection were less likely to say they were

confident in their ability to protect themselves and their family from AI.

Message implications: Strive for maximum message reach. Emphasize the many ways people can

be exposed to the H5N1 virus, but how easy it can be to seek care quickly as soon as symptoms

emerge.

* Controlling for education, household wealth (possessions), district and 10 other variables.

Number of AI messages

recalled

Perceived self-efficacy to

seek care if needed

−

+

−

7/11/2012

40

79

Summary—Predictors of knowledge about AI transmission*(+ positive relationship, - negative relationship)

Number of AI transmission vectors

known+ Knowledge that humans

can get AI

Education

Knowledge of how AI is transmitted was associated with level of education, recall of AI messages

from previous campaigns, knowledge that humans can become infected with AI and perceived self-

efficacy to seek care if it is needed.

Message implications: Strive for maximum message reach. Emphasize that humans can become

infected with the H5N1 virus through a variety of contacts. Also emphasize how easy it can be to

seek care quickly as soon as symptoms emerge.

* Controlling for education, household wealth (possessions), district and 10 other variables.

Number of AI messages

recalled

Perceived self-efficacy to

seek care if needed

+

+

+

80

Summary—Predictors of perceived AI severity*(+ positive relationship, - negative relationship)

Perceived severity of AI infections+

Knowledge that humans

can get AI

District (Bogor)

Residents of Bogor, a more rural district where more outbreaks of AI have occurred compared to

East Jakarta, were more likely to perceive that AI infections pose severe health threats, as were

those who perceive themselves as more susceptible to AI infection and those who know that

humans can become infected with the H5N1 virus.

Message implications: Target East Jakarta residence where the perceived severity of AI is relatively

lower. Emphasize that humans CAN become infected with the H5N1 virus through a variety of

contacts. Also emphasize how susceptible everyone is from the almost daily contact had by most

Indonesians in western Java.

* Controlling for education, household wealth (possessions), district and 10 other variables.

Perceived susceptibility to

AI

+

+

7/11/2012

41

81

Summary—Predictors of perceived susceptibility to AI*(+ positive relationship, - negative relationship)

Perceived susceptibility to AI infections

+

Knowledge that humans

can get AI

HH wealth

Less wealthy households and those with lower levels of perceived self-efficacy to protect

themselves from AI were more likely to perceive themselves at greater risk (higher susceptibility).

Those who know that humans can become infected with AI and those who perceive AI as a severe

health threat were more likely to perceive themselves as susceptible.

Message implications: Target lower income households residence where the perceived severity of

AI is relatively lower. Emphasize that humans CAN become infected with the H5N1 virus and that

infection can have very serious consequences. Also emphasize how easy it can be to protect ones

family from infection through avoiding high risk exposures and practicing safe poultry handling

practices, as well as seeking care as soon as symptoms begin to emerge.

* Controlling for education, household wealth (possessions), district and 10 other variables.

Perceived severity of AI

−

+

Perceived self-efficacy to

protect self & family

−

MEDIA HABITS

& MESSAGE RECALL

82

7/11/2012

42

Total East Jakarta Bogor

TV 74.3 82.8 69.2

Radio 6.9 10.5 4.8

Newspaper 16.1 23.5 11.8

Pamphlets/Brochure 2.2 3.5 1.4

Poster 2.4 3.1 2.0

Internet Website 5.5 6.3 5.0

Private Doctor 11.9 5.3 15.8

Doctor at Health Center 22.4 28.5 18.8

Midwife 10.3 3.9 14.1

Cadre 13.8 11.4 15.2

Community meeting 1.5 3.5 0.3

Family/Relatives 7.4 9 6.5

Neighbor/Friend 21.0 24.3 19.1

Community or Religious Leader 6.8 11.9 3.8

Information Source about Health Issues

83

� Television reach is high, but interpersonal channels also matter.

Reliable source Usefulness

TV 37.4 55.2

Radio 0.7 2.8

Newspaper 0.9 7.2

Pamphlets/Brochure 0.2 0.8

Poster 0.2 0.5

Internet Website 1.4 3.6

Private Doctor 9.8 12.1

Doctor at Health Center 20.9 22.8

Midwife 4.9 7.7

Cadre 4.6 7.4

Community meeting 0.2 0.7

Family/Relatives 2.8 4.3

Neighbor/Friend 3.7 11

Community or Religious Leader 2.1 3.6

Reliable Source &

Usefulness of Health Information

84

� Trusted sources: TV and doctors

7/11/2012

43

Total East Jakarta Bogor

Never 50.1 45.3 53.0

Rarely 35.1 33.1 36.3

1-2 times per week 6.0 7.7 4.9

Three or more times

per week

8.8 13.9 5.8

Frequency of Reading Newspaper

85

Total East Jakarta Bogor

Never 69.2 66.6 70.6

Rarely 23.6 24.1 23.2

1-3 times per month 3.4 2.7 3.8

1-2 times per week 3.0 5 1.8

3-6 times per week 0.3 0.6 0.2

Everyday 0.6 1 0.4

Frequency of Reading Magazine

86

7/11/2012

44

Total East Jakarta Bogor

Never 53.7 53 54.1

Rarely 25.2 22 27

1-3 times per month 1.2 0.7 1.5

1-2 times per week 3.8 4.4 3.4

3-6 times per week 2.5 3.5 1.9

Everyday 13.7 16.4 12.1

Frequency of Listening Radio

87

Total East Jakarta Bogor

Never 3.2 0.8 4.7

Less than 1 hour per day 13.1 10.8 14.5

1-2 hours per day 28.7 26.2 30.1

2-3 hours per day 16.9 19.8 15.2

3-4 hours per day 9.6 10.5 9.1

More than 4 hours per day 28.4 31.9 26.4

Frequency of Watching TV

88

7/11/2012

45

Total East Jakarta Bogor

Never 85.3 82.6 87.1

Rarely 5.3 6.8 4.5

1-2 times per week 2 2.4 1.8

3-4 times per week 1.4 1.5 1.4

Everyday 5.8 6.8 5.2

Frequency of Using Internet

89

� SUMMARY OF BEST CHANNEL CHOICES:

� TV, radio, newspapers

� Internet may be surpassing newspapers

81.9

17

1.2

Percentage

Yes

No

Do not know of such

desease

Ever Received, Heard or Seen any Messages about Flu Burung

90

7/11/2012

46

Total East Jakarta Bogor

TV 90.4 88.6 92.3

Radio 4.1 4.3 3.8

Newspaper/Magazine/tabloid/book 10.7 14.3 6.8

Flyer/brochure/leaflet/poster/banner 3.9 5 2.7

Social network (events in the community,

cadre, RT/RW, family, neighbor, community

leader)

18.1 22.8 13

Government officer (Dinas Kesehatan dan

Dinas Peternakan, Local offices)

4.2 6.3 1.8

New media (internet website, text through

cellphone)

2.1 2.4 1.8

Puskesmas 3.3 5.4 1.9

Doctor 2 2.4 1.8

Information Sources for Flu Burung

91

� Television and social networks (interpersonal channels)

Total East Jakarta Bogor

Bury dead poultry 34.9 34 36

Report dead chicken 4.4 5 3.8

Hygiene & self-protection action 57.2 63.2 50.6

Flu Burung news 24.4 21.9 27.2

Recall of Flu Burung Message Topics

92

7/11/2012

47

29.1

70.7

0.2

Percentage

Yes

No

Don't know

Ever Heard/Seen Messages about Seasonal Influenza

93

Total East Jakarta Bogor

TV 51.1 55.9 48.3

Radio 1.8 2.6 1.3

Newspaper 4.0 5.9 2.8

Flyer/brochure/leaflet 1.9 2.6 1.5

Poster 1.2 2.6 0.2

Events in the community 0.8 2.2 0

Posyandu cadre 3 1.5 3.9

Puskesmas 6.2 8.5 4.8

Doctor 12 12.5 11.7

Health Office 1.5 1.5 1.3

RT/RW 2.2 1.5 2.6

Information Sources for Seasonal Influenza

94

7/11/2012

48

Total(n=762)

East Jakarta(n=372)

Bogor(n=390)

Avoiding close contact/keeping a distance 9 11 8

Not attending work/school 2 4 <1

Going to the doctor 16 18 13

Washing hands frequently 7 9 5

Taking medication 44 53 36

Receive influenza vaccine 1 1 1

Recall of Seasonal Influenza Message Topics

95

CONCLUSIONS &

STRATEGIC COMMUNICATION

IMPLICATIONS

96

7/11/2012

49

1. Urban/rural variations in strategy

•There are numerous differences between E. Jakarta and Bogor.

•An important messaging focus in East Jakarta should be on the

risks associated with potential exposures to H5N1 in the wet

markets, on which Jakarta residents rely for their poultry

products

•In Bogor (and in rural areas generally), there may be a sense of

false confidence stemming from constant exposure to birds and

greater exposure to outbreaks—which most people have

survived.

97

Conclusions: Strategic Implications and Priorities

2. Living Safely with Birds

•Routine daily exposure to birds is a fact of life in Indonesia, with almost

universal exposure to wet market risks among urban households.

•Most exposures to birds are not considered distinctive or noteworthy.

•Complacency regarding risks associated with bird exposures is common.

� Suggests an efficacy-based strategy framed around “living safely with

birds”.

– Flag certain types of exposure (e.g., handling live or slaughtered birds

at the wet market) as requiring vigilance.

– Educate patients to mention these distinctive types of exposure to

providers when seeking care.

– Educate providers to ask about these distinctive types of exposure

when seeing clients with respiratory illness.

98

Conclusions: Strategic Implications and Priorities

7/11/2012

50

3. Knowledge of distinctive AI symptoms

•General knowledge of AI and knowledge that AI is more

dangerous than seasonal flu are relatively high, but knowledge

of the distinctive AI symptom—sesak nafas (shortness of breath,

tightness in the chest)—is limited.

� Suggests a need to highlight sesak nafas as a danger sign

requiring prompt action.

•Do not need extensive education about biomedical distinctions;

Do need to focus on a few “differences that make a difference”.

99

Conclusions: Strategic Implications and Priorities

4. Barriers to care seeking

• Over 70% say they face no barriers to obtaining care; those

who identify barriers mention access hours and cost of

services.

• More than half report using private services (private clinics

and hospitals) because they are thought to provide accessible

quality care.

� Suggests a need to promote sources of quality care for

respiratory illness.

100

Conclusions: Strategic Implications and Priorities

7/11/2012

51

5. Response to acute events

•Most common responses to sudden onset of fever,

cough and tightness of breath are to obtain drugs from

pharmacy (22% overall), medicate/give fluids at home

(19% overall, 38% in E. Jakarta) and go to a private

clinic (16% overall, 22% in E. Jakarta).

�Suggests a need to discourage over-reliance on self-

treatment.

101

Conclusions: Strategic Implications and Priorities

6. Response timing

•65 percent report seeking care within 48 hours of onset of acute

illness episode

•24 percent report not seeking any kind of care for an acute

illness episode

•Those who perceive an AI infection as serious and feel confident

in their ability to seek care do so more quickly.

•Of those who waited less than 48 hours to seek care, HALF

(almost two-thirds in E. Jakarta) said they sought care because

they “felt it was important to go to a healthcare facility.”

� Suggests the need for a norms-based strategy: it is the right,

responsible thing to do to seek care within 48 hours if difficulty

breathing begins to occur

102

Conclusions: Strategic Implications and Priorities

7/11/2012

52

7. Communication channels

•Television has the greatest reach and has the highest reliability

and usefulness rating as a source of health information; is most

common source of flu burung and seasonal flu information.

•Interpersonal channels (doctors, neighbors, kaders) also play a

key role in the flow of health information.

– Suggests a mutually reinforcing strategy involving use of

television to inform, model positive behaviors and catalyze

community-level discussion about “safe living with birds.”

103

Conclusions: Strategic Implications and Priorities

Use television to inform about:

– Particular types of higher risk bird exposures as

transmission pathways

– Sesak nafas as a distinctive symptom requiring rapid care

seeking

– Inadequacy of self-treatment as a first response

– Two days (48 hours) as the critical response time period

– Encourage rapid care seeking as the “right thing to do”

(social norm)

– Model care seeking within 48 hours if sesak nafas begins

to develop

– Encourage interpersonal communication and social

support at the community level

Communication channels (cont.)

104

7/11/2012

53

Use community media, interpersonal outreach and signage to:

– Provide reminders at clinics and wet markets about

particular higher risk types of bird exposure

– Provide details about effective behaviors to “live safely

with birds”

– Encourage client-provider communication about higher

risk bird exposures and distinctive symptoms of AI (sesak

nafas).

Communication channels (cont.)

105

• Strive for the highest reach possible with care seeking

messages

• Take advantage of the powerful combination of “threat” AND

“efficacy”:

– Emphasize how deadly AI CAN be if treatment is not

sought as soon as possible after symptoms begin to appear

– Emphasize the many ways people can be exposed to the

H5N1 virus, given how common exposure to birds is in

Indonesian daily life

– Emphasize building confidence in the ability to seek care

quickly

General messaging implications

106