Denver HANDS Pilot Program Final Report 2013 The Denver-based HANDS Program was led by Dr. Comilla Sasson. The team included community liaisons Merida Carmona, Christofer Medina, Mandy Medrano, Charlene Barrientos-Ortiz, Lorenzo Ramirez, Beverly Tafoya-Dominguez, Councilman Paul Lopez and May Tran. From the University of Colorado, our team included Thomas Califf, Ariann Nassel and Nhat-Tuan Tran. Finally, from Denver Health and Hospital Authority, Dr. Chris Colwell, Dr. Jason Haukoos, Dr. Kevin McVaney and Dr. Ricardo Padilla. Our community partner organizations include: Center for AfricanAmerican Health, Servicios de la Raza, 2040 Partners for Health, and the American Heart Association Mile High Chapter. Funding was provided by the Colorado Clinical and Translational Sciences Institute and the American Heart Association.
BACKGROUND Traditional methods for increasing CPR (cardiopulmonary resuscitation) use generic training programs that are employment, school, or event-based. However, these complex approaches to CPR training has not been successful in communities of African-Americans, Latinos, those with limited English proficiency, and the poor, groups with a high incidence of OHCA (out-of-hospital cardiac arrest) and low survival. Potential reasons for the low rate of CPR in these communities include poor accessibility to CPR training and training programs that are neither culturally-sensitive nor tailored to the cultural and socioeconomic characteristics of these groups. An alternative approach to increase community CPR is to develop tailored CPR training programs and to target training to “high-risk” neighborhoods with a high incidence of OHCA and low bystander CPR. RATIONALE Previous research has shown residents who live in neighborhoods that are primarily Latino, African-American, or poor are more likely to have a cardiac arrest. In addition, when they experience a cardiac arrest, they are less likely to receive CPR and survive. Therefore, these high-risk neighborhoods (where the incidence of OHCA is high and the prevalence of bystander CPR provision is low) are an important target for public health interventions to help reduce disparities in bystander CPR and to ultimately improve OHCA survival. However, no prior research has been conducted in cardiac arrests to see if a systematic approach may be successful in reducing health disparities. We present the results of a novel conceptual framework, methodology, and pilot implementation trial for the HANDS (identifying High Arrest Neighborhoods to Decrease disparities in Survival) Program. The HANDS Program challenges the traditional paradigm of blanketing a city with CPR training, and moves toward the ultimate goal of focusing scarce public health resources for CPR training in the geographic locations in which it is most needed. METHODOLOGY Before we can change the paradigm of CPR training in the community, we first identified high-risk neighborhoods in which the relative incidence of cardiac arrest was high, and the prevalence of bystander CPR was low (Phase 1 Step 1). We then conducted focus group and interviews of high-risk neighborhood residents to understand barriers and facilitators to learning and performing CPR (Phase 1 Step 2). We then used this information to design and implement a community-based intervention to increase bystander CPR in these neighborhoods (Phase 2). Finally, we evaluated our results and the real-world effectiveness of the program (Phase 3). The Phase 1 following report outlines our successful completion of these three phases (Figure 1). Phase 2
Phase 3
Figure 1: Three Phases of the HANDS Program
Phase 1 Step 1: Location of High-Risk Neighborhoods In the Denver-based HANDS Program pilot, we used data from the Denver subset of the Cardiac Arrest Registry to Enhance Survival (CARES) database. CARES links a standard set of data elements from three sources: 911 call centers, EMS providers and receiving hospitals. The registry was established in 2005 and includes 40 major U.S. cities and 6 U.S. states, with a catchment area of more than 64 million people. Using the Denver subset, we geocoded cardiac arrest address data into census tracts and then added census tract socio-econo-demographic data, obtained from the 2010 Census, to each event. Adjusted cardiac arrest incidence and proportion of arrests receiving bystander CPR were calculated for each census tract. Using three spatial analysis techniques, we identified tier 1 and tier 2 areas with relatively high rates of cardiac arrest, and relatively low prevalence of bystander CPR. Tier 1 neighborhoods were identified in all three spatial analysis methods, while tier 2 neighborhoods were identified in two out of three spatial analysis methods. Figure 2 shows the ten high-risk neighborhoods that were identified using this approach. Figure 2: Map of Denver's High-Risk Neighborhoods
Table 1 lists the socio-econo-demographic data for each neighborhood as compared to the mean for the City and County of Denver. The neighborhoods were comprised primarily of Latino and African-American residents with a median household income ranging from $17,373 to $42,396. Adjusted Cardiac Arrest CPR Neighbor- Tract Incidence Prevalence hood Name ID (per 1000) (%) Tier 1 West Colfax 7.02 0.72 7.69 Lincoln Park 19.01 0.65 16.67 23 Whittier 0.74 0 Five Points 24.03 0.87 9.09 Skyland 36.03 1.06 14.29 NE Park Hill 41.02 0.95 14.29 Tier 2 Ruby Hill 14.01 0.63 9.09 Lincoln Park 18 0.61 0 36.02 Clayton 0.57 0 Montbello 83.06 0.61 9.09 0.18 Denver County 26.17
Median % of High Median Age, Household School Grad % Hispanic years Income ($) or higher
% White
% African American
61.5 51.0 55.6 62.5 39.5 30.4
7.4 24.0 32.1 25.9 52.9 59.1
38.4 24.6 11.5 13.3 8.8 15.6
27.7 27.3 32.4 32.5 37.9 34.0
31,380 17,373 39,432 31,195 42,396 28,867
74.8 69.3 90.2 84.2 92.3 76.6
63.1 77.1 44.8 38.8 52.7
3.2 5.8 34.5 38.7 10.4
44.4 19 26.9 29.8 31.5
30.6 29.3 30.1 28.9 33.7
30,470 27,083 36,167 38,728 49,091
58.5 80.0 63.3 64.6 85.1
Table 1: Characteristics of High-Risk Neighborhoods
Phase 1 Step 2: Barriers and Facilitators to Learning and Performing CPR Historically, CPR training has been based on convenience and is typically offered at workplaces, schools, and community events. This CPR training is not targeted based on needs or tailored to specific social or cultural groups that are less likely to do CPR, such as Latinos, persons with limited English proficiency, and the poor. This traditional approach to CPR training fails to consider: (1) who is getting training, (2) the setting in which the training occurs, and (3) how the training is being delivered. No previous studies had been conducted to understand how these three contributing factors, particularly in certain racial/ethnic and socioeconomic groups, at the individual and neighborhood levels, impacted a person’s likelihood of learning and performing CPR. Table 2: Demographics of Focus Groups and Key Informant Interviews Characteristic
Count (n)
Percentage
7
11%
20 – 39
28
44%
40 – 59
18
29%
9
14%
47
75%
Age (years)a < 20
60 + Gender b Female Race/Ethnicity b Latino
55
88%
Black/African-American
1
2%
White
4
6%
Other
3
5%
30 6
47% 10%
13
21%
Completed College
5
8%
Master’s Degree/Doctorate Degree
9
14%
< 10,000
10
16%
10,000 – 20,000
14
22%
20,000 – 30,000
5
8%
30,000 – 50,000
9
14%
50,000 – 100,0000
9
14%
Education (highest level) b Primary school-Some High School Completed High School Some College
Annual Household Income ($/yr)
100,000 – 200,000 Unknown
a b
4
6%
13
20%
As a result, we used community-based participatory research with qualitative methods to understand the underlying cause of health disparities in cardiac arrest and provision of bystander CPR in the target neighborhoods, and how to design a tailored, culturally-sensitive neighborhood-based intervention that could then be implemented in the highrisk neighborhoods identified in Step 1. We conducted six focus groups and nine key informant interviews in our high-risk Denver neighborhoods, oversampling Latinos to better understand the challenges they may face in learning and performing bystander CPR. Demographics of the 64 participants are listed in Table 2. All participants received a CPR Anytime Kit to take home.
Of the 64 total participants who filled out the pre-focus group survey, only 62 provided an age. Only 63 of the 64 participants provided a gender, race/ethnicity and education.
Six main thematic areas were identified as barriers to calling 911 for residents: cost incurred by victim/victim’s family, undocumented status, fear of getting involved, unsure of what an emergency situation is, language barriers, and cultural issues. Eight main thematic areas were identified as barriers to performing CPR: legal ramifications, fear of doing CPR incorrectly, fear of breathing into someone else’s mouth, fear of hurting victims, being too old to do effective CPR, panic, undocumented status and a lack of resources on CPR in Spanish. We identified significant cultural and language barriers to calling 911 and performing CPR. This research laid the foundation for the next phase in our research, to work directly with residents from these neighborhoods to design and implement a communitybased intervention to overcome the identified barriers.
Phase 2: Implement Targeted, Culturally-Sensitive CPR Training Program We conducted a pilot, community-based trial in Denver’s high-risk neighborhoods to test our targeted, intervention strategy. A total of 344 participants were recruited from target neighborhoods during the 12-week study period. We conducted 12 trainings across the Denver area. Participants completed a pre-test survey to assess baseline knowledge of CPR. A group hands-only CPR training lasting 1 hour and was conducted with the CPR Anytime Kit (Figure 3), which includes an educational DVD and hands-on practical skills training. Participants then completed a survey to assess their post-training knowledge and were asked to use the kits to train others. A $10 incentive was given to participants. At 2-4 weeks, participants were asked to return their data collection sheet on the family and friends they had trained, as well as a pre and post-test survey that was given to five family and friends to evaluate if they felt comfortable performing hands-only CPR having learned it from a study participant using the CPR Anytime Kit.
Figure 3: CPR Anytime Kits
Phase 3 Outcome 1: Metrics of Program Success Twelve sites hosted trainings in the community for the Denver HANDS pilot program. These sites and participants were recruited by our community liaisons. Table 3 shows the names of the training sites, number of people trained, percentage of packets returned at the 2-4 week mark (response rate), and the total number of family and friends (F&F) trained. Table 3: HANDS Training Sites and Participants Site Name Quang Minh Vietnamese Temple SGI Culture Center, Downtown Vietnamese Central Baptist Church Center for African American Health Fair NE Park Hill Neighborhood Association Vietnamese Nhu Lai Buddhist Temple First United Hmong Alliance Center for African American Health ESRI (Business) Queen of Vietnamese Martyrs Catholic Church Boulder Chinese EF Church New Hope Baptist Church
Location Type/Category Church
Participants 22
Packets Returned 18
Response Rate (%) 81.8
Church
13
6
46.2
55
4.2(5.8)
0-15
Church
9
7
77.8
21
2.3(1.9)
0-5
Health Fair
58
20
34.5
95
1.6(2.9)
0-15
Neighborhood Association Church
27
8
29.6
54
2.0(4.3)
0-18
22
12
54.5
76
3.5(4.4)
0-16
Church
16
10
62.5
48
3.0(3.4)
0-10
Health Liaisons
34
21
61.8
132
3.9(4.9)
0-16
9
3
33.3
17
1.9(3.0)
0-8
Church
67
24
35.8
87
1.3(2.2)
0-10
Church
45
25
55.6
108
2.4(2.5)
0-7
Church
22
8
36.4
69
3.1(7.3)
0-34
Business
Family &Friends Trained Total Mean (SD) Range 124 5.6(9.1) 0-43
Figure 4 shows the distribution of CPR Anytime Trainees (orange) and the family and friends (purple) trained during the implementation phase.
Figure 4: Distribution of Trainees and Family/Friends
Participants were Asian (50.0%), Black (35.6%), female (68.0%), had completed high school (26.8%), and had an annual income of less than $30,000 (37.1%). Because Phase 1 Step 2 had heavily represented Latinos, we worked closely with the Asian community to target residents of Asian and Black race living in high-risk neighborhoods. After the CPR intervention, the mean number of questions answered correctly on CPR knowledge increased (Table 4). The majority of participants (84.6%) felt comfortable performing hands-only CPR after the intervention. 154 (44.8%) participants returned information, with an additional 886 friends and family trained. Of the 886, 466 participants returned pre-/post-knowledge survey information. Table 4: Pre and Post-Test Survey Results
Question What does CPR stand for? How many compressions should be given, on average, during a one minute period? What are the correct steps for providing Hands-Only CPR? It is appropriate to use Hands-Only CPR in what situations? When pushing on the chest to provide Hands-Only CPR, one should push to what deptht? Mean Score (± SD) ‡ p < 0.05 * p < 0.001
Pre Survey (N=344) n (%) 205 (60.0)
Post Survey (N=344) N (%) 287 (83.4)*
67
(19.5)
311
(90.4)‡
177
(51.5)
317
(92.2)*
28
(8.1)
128
(37.2)*
111 (32.3) 1.71 ± 1.31
320 (93.0)* 3.96 ± 1.07†
A total of 1,230 people were trained in hands-only CPR (average: 3.6 people/kit). At $27 per CPR Anytime kit, this came to a cost $7.56 for each person trained. Participants demonstrated increased knowledge of CPR and enthusiasm to train others (Table 5). We found that this type of targeted CPR training program was feasible, with 73.2% of the family and friends feeling somewhat to very comfortable in performing hands-only CPR after learning it from their family or friend. Family and friends also increased their knowledge after learning hands-only CPR from an average of 2.78 [SD 1.49] to 4.11 correct answers [SD 0.69] out of 5 questions. Table 5: Characteristics of Family and Friends (n=466)
Characteristic Age (mean (years) [SD]) (n=448) Race/Ethnicity (n=455) Asian Black White Hispanic/Latino Other Unknown Gender (n=455)
Total n (%) 40.0 [20.0] 246 122 44 28 13 2
(54.1) (26.8) (9.7) (6.1) (2.9) (0.4)
Female
246
(54.1)
Male
209
(45.9)
Education Some High School or less
108
(23.7)
Completed High School
81
(17.8)
Some College
91
(21.6)
Completed College
92
(20.2)
Master’s Degree
47
(10.3)
Doctorate Degree
19
(4.2)
Unknown
10
(2.2)
Less than $10,000
48
(13.9)
$10,001-$20,000
38
(11.0)
$21,001-$30,000
42
(12.2)
$30,001-$50,000
70
(20.2)
$50,001-$100,000
61
(17.7)
$100,001-$200,000
36
(10.5)
3
(0.8)
46
(13.7)
Yes
143
(31.4)
No
274
(60.2)
38
(8.4)
Annual Income
$200,001 and greater Unknown Prior CPR Course
Missing Comfortable Doing Hands-only CPR Post-Training Somewhat/Very Comfortable
333
(73.2)
Neutral
69
(15.2)
Somewhat/Very Uncomfortable
52
(11.4)
1
(0.2)
Missing Number of Pre-Test Questions Correct (mean [SD])
2.78 [1.49]
Number of Post-Test Questions Correct (mean [SD])
4.11 [0.69]
Phase 3 Outcome 2: Program Evaluation We evaluated the HANDS Program using the RE-AIM measures (Table 6). The RE-AIM framework has been used extensively for evaluating the effectiveness of communitybased interventions and is well-vetted in the program implementation literature. This program evaluation framework has been designed to increase the speed in which promising, evidence-based community interventions are translated into the real-world. This systematic evaluation framework concentrates on five major areas: reach to the intended target population, efficacy/effectiveness, adoption, implementation, and maintenance of the intervention over time. The answers to the program evaluation for the Denver-based HANDS Program are provided in italics. Table 6: RE-AIM Program Evaluation Results Evaluation Measure REACH: Extent of Representativeness of Participants
EFFECTIVENESS: Short Term Impact/ Outcomes for Participants
Proposed Strategies to Enhance Overall Impact A. Clearly defined target population and numbers of people eligible for the intervention B. Major barriers and facilitators to CPR intervention identified in Step 2 to develop targeted intervention
A. Tailored messaging approach based on community recommendations
B. Reinforced CPR intervention with direct training and take-home kit
ADOPTION: Interface Between Researchers and Potential Program Settings
IMPLEMENTATION: Fidelity or Intervention Integrity MAINTENANCE: Both Individual Participant and Program/Setting Level
C. Multiple outcome measures to triangulate outcome effect A. Settings for intervention based on community input B. Intervention can be easily conducted in multiple settings with few resources needed C. Commitment to piloting intervention from key community-organizations
A. Meet with community settings where intervention was held and participants who completed the CPR intervention to understand how to make intervention better A. Reduce level of resources needed for starting and maintaining this type of project B. Formalize relationships with national organizations to replicate this intervention in other communities C. Work with existing network of community organizations locally to develop sustainability plan
Outcome Measures to Evaluate Impact -Proportion of residents from high-risk neighborhoods who participated in CPR intervention (n=344) -Number of settings willing and unwilling to host intervention (No settings were unwilling to host the intervention) -Number of residents from high-risk neighborhoods who participated (n=344) -Post-intervention survey assessing the participants satisfaction with the training, knowledge acquisition and recommendations for improvement (Increase in knowledge acquisition from 1.71 ± 1.31 to 3.96 ± 1.07 out of possible 5 questions) -In-person follow-up at 2-4 weeks to understand how participants felt training family and friends (80% of family and friends feel comfortable performing handsonly CPR) -See Outcome 1: Metrics for Program Success Section Above -Record which settings were wiling/unwilling to participate (N/A) -Record which settings initially participated and then dropped out and why (N/A) -Number of formal local collaborations established by the end of the intervention period (continuation of 2 major community-based programs in stroke and hypertension in same populations) -Record number of participants who complete the follow-ups (n=186) -Creation of a “How-to-Guide” for the entire intervention that can then be “taken off the shelf” and used in other communities (in process) -Number of formal collaborations built with national organizations (discussions in place with 2 national organizations for formal collaboration) -Number of formal local collaborations established by the end of the intervention period (discussions in place with 2 local organizations for formal collaboration)
CONCLUSION There are large health disparities in the provision of bystander CPR in underserved populations. We have implemented the pilot Denver HANDS Program, and have had outstanding results, not only from the numbers of people trained, but more importantly, the relationships that have been built with this program. This final report outlines the rigorous, step-wise approach to changing the paradigm of community bystander CPR training from a one-size fits all approach to one that is targeted where the need is the greatest. Future work is already being done in four other cities to see if the results we have achieved in Denver are able to be replicated in new locations. We hope that the novel HANDS Program will begin to address and correct the disparities we see daily for OHCA victims in the U.S.
