Jung-won Lim, PhD

1
Recruitment and Measurement Challenges in Ethnic
Minority Cancer Survivorship Research
Jung-won Lim, PhD Assistant Professor Mandel
School of Applied Social Sciences Case Western
Reserve University
Presented at MSASS Research Colloquium
February 15, 2012
2
Trends in Female Breast Cancer Incidence
Mortality Rates by Ethnicity
Incidence Rates, US, 1975-2008
Mortality Rates, US, 1975-2007
American Cancer Society (2011)
3
5-year Breast Cancer Survival Rates
Note. Survival rates are based on patients
diagnosed between 2001 and 2007 and followed
through 2008. American Cancer
Society, Surveillance Research (2011)
4
Project 1Communication and Coping in Asian
American Families with Breast Cancer Survivors
Funded by NIH/NCI R03 (1R03CA139941) PI
Jung-won Lim
5
Research Rationale

• How individuals and their families communicate
  and cope
• with stressful experiences influence
  outcomes.
• Increasing number of Chinese- and
  Korean-Americans as
• well as Breast Cancer Survivors in U.S.
• Chinese- and Korean-American studies under the
  
• umbrella category of Asian studies
• Similar geographical background but, different
  cultures,
• languages, attitudes, and histories

6
The Purpose of the Study

1. To examine family communication, coping, and
   HRQOL for Chinese- and Korean-American breast
   cancer survivors (BCS)
2. To investigate the pathways of family
   communication and coping, and other possible
   predictors influencing HRQOL for Chinese- and
   Korean-American BCS
3. To explore the dyadic relationships in family
   communication and coping between BCS and their
   families within the family system.

7
Asian American Breast Cancer Survivor Studies
Studies Topic Samples Recruitment
Ashing-Giwa (2003) Breast cancer experiences (Qual) Asian-American BCS (n34) Korean (n10), Chinese (n11) Cancer support groups, hospitals and community health clinics (LA)
Giedzinska (2004) HRQOL English speaking Asian-American BCS (n77) Tumor registry listings, listings from doctors office, hospital logs (LA, Washington DC)
Ashing-Giwa (2004) Kim (2006) HRQOL Asians-American BCS (n206) Chinese (n85), Filipino (n39), Korean (n29), Japanese (n26) Community agencies and support groups, fliers (LA)
Lim (2008) Social support and HRQOL Korean-American BCS (n51) Community agencies, support groups, fliers (LA)
Yoo (2009) Emotion work involved in self disclosing Asian-American BCS (n52) (included those speaking in Chinese or English) Cancer Center clinic, cancer organizations, Northern California Cancer Registry (SF)
Yi (2011) English proficiency, symptoms, and QOL Vietnamese- (n25), and Chinese-Americans (n72) BCS Advertisements in Asian-American media, Fliers (Houston)
Lim (2011) Cultural Health beliefs (mixed) Asian-American BCS (n206) Korean-American BCS (n11) Community agencies, support groups, fliers (LA)
8
Research Methods I

• Target population Chinese- and Korean-American
  BCS and their
• family
  members
• Eligible Survivors
• 1. Women who are within 1-5years of a
  breast cancer diagnosis
• 2. Cancer stage 0-III
• 3. Have completed active treatment (e.g.,
  chemotherapy, radiation)
• 4. Have not been diagnosed with another
  type of cancer
• 5. Age more than 18 years at the time of
  study
• 6. Korean- and Chinese Americans
• 1) Born in the US after parents
  immigration ( 2nd generation)
• 2) Immigrated from Korea or China to US
  (1 or 1.5 generation)
• 3) Able to speak Chinese or Korean
  and/or English

9
Research Methods II

• Research Design Cross-sectional and sequential
  mixed methods

Methods Samples Samples Samples Samples
Questionnaires Chinese BCS Chinese BCS Korean BCS Korean BCS

4 Focus groups (based on family communication score) High Low High Low
Unmatched Family Chinese BCS family Chinese BCS family Korean BCS family Korean BCS family
4 Focus groups survey (based on BCS family communication score) High Low High Low
10
Research Methods III

• Study Site Southern California (mainly LA
  County)
• ? Of all Asian Americans living in LA
  county, Chinese-Americans were
• 29 and Korean-Americans were 16,
  indicating a greater percentage
• compared to other areas.
• Recruitment
• Mixed Sampling Method
• 1) California Cancer Surveillance Program
• LA County, Orange County, and San
  Diego Regions
• 2) Hospital registry City of Hope
• 3) Community agencies including support
  groups

11
Challenges in Asian American Recruitment

• 1. Cultural context
• language, health beliefs, attitudes toward
  disclosure,
• acculturation
• 2. Socio-political context
• poverty level, health care system,
  mistrust,
• fear of being exploited
• 3. Personal level
• access to research participation,
  awareness
• 4. Study-related context
• target population, the inclusion criteria
  

12
Challenges in Asian American Recruitment How to
Overcome Challenges?

• 1. Use of Cancer Registry, if possible
• 2. Collaborating with community
  organization/hospital
• City of Hope, St. Vincent and Hollywood
  Presbyterian Medical Centers, Herald Cancer
  Association
• 3. Employing and training culturally
  sensitive/bilingual recruiter
• 4. Ethnic matching of interviewers Chinese,
  Korean
• 5. All study documents are available in Chinese,
  Korean, and English.
• 6. Participating in community events or support
  groups
• Chinese conferences and support groups,
  Korean support groups

13
Translated Study Fliers
14
Recruitment Outcomes
Final response rates of 25.3 of the identified
samples 63.2 of the accessible samples.
15
Differences by Study Participation
Characteristics Of all accessible people Of all accessible people Inaccessible people (n371) T a F b
Characteristics Respondent (n157) Non-respondent (n91) Inaccessible people (n371) T a F b
Characteristics Mean (SD) Mean (SD) Mean (SD) T a F b
Cancer Stage 1.7 (0.8) 1.7 (1.0) 1.4 (0.9) 0.02ns 1.99ns
Years since Dx 3.9 (1.9) 4.0 (1.5) 4.3 (1.2) -0.27ns 1.34ns
Age 55.9 (10.4) 57.3 (14.6) 59.1 (12.4) -0.67ns 3.21
N () N () N () X2
Language Chinese Korean English 67 (71.3) 21 (22.3) 6 (6.4) 28 (62.2) 9 (20.0) 8 (17.8) - - - 4.34ns -
Note. nsnot significant plt0.05 a two group
differences including all accessible people only
b three group differences including all people
16
Demographic Characteristics
Variables Chinese (n86) Korean (n71)
Age (mean SD) 55.2 (9.7) 53.9 (9.7)
Length of U.S (mean SD) 23.5 (12.2) 23.9 (9.9)
Household income Under 25K 25K 45K 45K 75K Over 75K 35 (40.7) 13 (15.1) 14 (16.3) 18 (20.9) 22 (33.8) 12 (18.5) 12 (18.5) 19 (29.2)
Education lt high school high school graduated gt High school 16 (18.6) 11 (12.8) 59 (68.6) 2 (2.8) 14 (19.7) 55 (77.5)
Employment status Unemployed / homemaker Employed 42 (48.8) 40 (46.5) 28 (40.0) 42 (60.0)
Marital Status Married Others 69 (80.2) 17 (19.8) 56 (78.9) 15 (21.1)
Note. plt0.01
p lt .01 p lt .001
17
Medical Information
Variables Chinese (n86) Korean (n71)
Cancer stage 0 I II III 10 (11.6) 22 (25.6) 45 (52.3) 9 (10.5) 1 (1.4) 34 (48.6) 23 (32.9) 12 (17.1)
Surgery (Yes) Axillary node dissection (AND) Lumpectomy Mastectomy 39 (45.3) 45 (52.3) 48 (55.8) 22 (31.0) 37 (52.1) 35 (49.3)
Radiation (Yes) 45 (55.6) 41 (58.6)
Chemotherapy (Yes) 61 (72.6) 44 (62.0)
Hormonal therapy (Yes) 58 (68.2) 41 (58.6)
Years since diagnosis (Mean, SD) 3.24 (1.8) 3.94 (1.4)
of comorbidities (Mean, SD) 3.80 (3.2) 2.99 (3.4)
Note. plt0.05 plt0.01
p lt .05
18
Measurements

• Quality of Life Outcome 1) FACT-B 2) MOS SF-36
• Psychological Distress BSI-18 (Derogatis, 2000)
• Family communication and coping
• 1) F-COPES (McCubbin, et al, 1981)
• 2) FACES IV Family Communication Scale
  (Olson, et al, 2004)
• 3) Family Avoidance of Communication about
  Cancer (FACC) Scale (Mallinger, Griggs,
  Shields, 2006)
• Social Support MOS Social Support Survey
  (Sherbourne, 1991)
• Acculturation Asian American Multidimensional
  Acculturation Scale
• (Chung,
  2004)

19
Challenges in Measurements

• 1. Language and utility of the instruments
• Non-English speaking individuals are
  excluded systematically because of
• difficulty and costs associated with
  translation and administration
• 2. Language is a proxy of Socio-economic status
  and acculturation
• 3. No normative survivorship data on Asian
  subgroups
• 1) Translation of English-version to Chinese and
  Korean (Pilot testing)
• 2) Statistical Approaches
• Reliability tests, Measurement
  invariance tests

20
Reliability (Cronbachs alpha)
Variables Chinese (n86) Korean (n71)
FACT-G Physical wellbeing Social wellbeing Emotional wellbeing Functional wellbeing 0.86 0.90 0.80 0.95 0.90 0.89 0.77 0.91
SF-36 Physical functioning Role-physical Bodily pain General health perception Social functioning Mental health Role-emotional Vitality 0.87 0.83 0.87 0.81 0.81 0.86 0.81 0.86 0.85 0.90 0.89 0.82 0.82 0.85 0.91 0.81
BSI-18 Somatization Depression Anxiety 0.85 0.91 0.91 0.78 0.85 0.88
Family Communication General Cancer specific 0.96 0.95 0.96 0.93
Acculturation 0.96 0.81
Social support 0.97 0.98
21
Family Crisis Oriented Personal Scales (F-COPES)
a Reliability (Cronbachs alpha)
Subscales Ethnicity Ethnicity Language Language Language
Subscales Chinese (n86) Korean (n71) Chinese (n76) Korean (n66) English (n15)g
1. Acquiring social support b 2. Reframing c 3. Seeking spiritual support d 4. Mobilizing family to acquire and accept help e 5. Passive appraisal f 0.83 0.85 0.91 0.69 0.39 0.90 0.87 0.87 0.73 -0.01 0.84 0.84 0.92 0.68 0.38 0.90 0.88 0.85 0.74 -0.12 0.86 0.81 0.88 0.66 0.52
Note. a The F-COPES asks respondents to choose
the answer that best describes how they and their
families cope with problems and difficulties b
the familys use of emotional and tangible
support from extended family, friends, and
neighbors (9-item) c the familys ability
to redefine stress/situation (8-item) d the
familys participation in and emotional support
acquired from their church (4-item) e the
familys use of community resources to cope with
their problems (4-item) f the familys inactive
coping style in managing problems (4-item) g
Chinese10 and Korean5.
22
Studies including F-COPES
Studies Sample-Ethnicity Reliabilities Excluded scales
5 sub-scales Leske (2010) Martin (2010) Stroup (2007) Churchill (2010) Montaguino (2004) Chui (2007) Shin (2004) Jonker (2009) Bagarozzi (1985)a Black/White (n33) White/black (n109) White(n222) White (n129) White (n18) Chinese in HK (n133) Korean (n48) Black/White (n34) N100 a0.82 a0.62-0.87 a0.32-0.83 a0.82 a0.49-0.83 a0.73 a0.63-0.83 a0.37-0.88 a0.34-0.72 2 items in the reframing a0.37 (PA) Abbreviated ver.
4 sub-scales Maupin (2010) Plunkett (1999) McKelvey (2002)a Rungreangk (2002) White/Black (n151) White (n77) White/Black (n105) Thai (n108) a0.65-0.86 a0.70-0.84 a0.64-0.82 a0.71-0.91 PA (alt0.60) PA (a0.28) PA (a0.13) PA (a0.26)
3 sub-scales Manning (2011) Shin (2010) Multiethnic (n195) Korean (n219) a0.78-0.89 a0.81-0.60 PA/Mobilizing PA/RE/SS
Note. a Factor analysis has been conducted.
23
Correlation Coefficients between Items in the
Passive Appraisal Subscale
Items 1. Watching TV 2. Knowing luck plays a big part in how well we are able to solve family problems 3. Feeling that no matter what we do to prepare, we will have difficulty handling problems 4. Believing if we wait long enough, the problem will go away
Chinese-Americans Chinese-Americans Chinese-Americans Chinese-Americans Chinese-Americans
1 1
2 0.107 1
3 0.080 0.252 1
4 -0.073 0.269 0.142 1
Koreans-Americans Koreans-Americans Koreans-Americans Koreans-Americans Koreans-Americans
1 1
2 0.069 1
3 -0.090 0.086 1
4 0.079 -0.102 -0.050 1
Note. plt0.05
24
Measurement Invariance Procedures

• Test of the single-factor model, based on the
  original F-COPES model
• 1) The single-factor within-group model
• 2) The single-factor between-group model
• 2. If the single-factor within-group model
  does not fit the data, or the two models
  (unconstrained vs. constrained in the
  between-group model) significantly differ, we
  modified the original single-factor model to
  achieve fit and create an equivalent model across
  ethnicity
• A series of multi-group CFAs to determine the
  adequacy of the modified factor structure of the
  F-COPES and assess factorial invariance across
  ethnic groups
• Configural, Metric, Strong, and Strict
  invariance
• 4. Test of the validity of the modified factor
  structure of the F-COPES

25
Fit Statistics for Single Factor Models of
Original Structure in the F-COPES
Factors X2 df p CFI RMSEA Comparison Comparison Comparison
Factors X2 df p CFI RMSEA ?X2 ?df p
Social support (9-item) Chinese (within) Korean (within) Unconstrained (between) Constrained (between) 86.0 74.6 160.6 163.5 27 27 54 62 0.00 0.00 0.00 0.00 0.78 0.86 0.83 0.84 0.16 0.16 0.11 0.10 2.86 8 0.943
Reframing (8-item) Chinese (within) Korean (within) Unconstrained (between) Constrained (between) 61.7 23.4 85.0 97.4 20 20 40 47 0.00 0.27 0.00 0.00 0.84 0.98 0.90 0.89 0.16 0.05 0.08 0.08 12.33 7 0.093
Spiritual support (4-item) Chinese (within) Korean (within) Unconstrained (between) Constrained (between) 2.6 1.1 3.7 15.3 2 2 4 7 0.28 0.58 0.45 0.03 0.99 1.00 1.00 0.98 0.06 0.00 0.00 0.09 11.59 3 0.009
Mobilizing (4-item) Chinese (within) Korean (within) Unconstrained (between) Constrained (between 3.0 0.6 3.5 9.8 2 2 4 7 0.23 0.76 0.48 0.20 0.98 1.00 1.00 0.97 0.07 0.00 0.00 0.05 6.31 3 0.097
26
Fit Statistics for Single Factor Models of
Modified Structure in the F-COPES
Factors X2 df p CFI RMSEA Comparison Comparison Comparison
Factors X2 df p CFI RMSEA ?X2 ?df p
Social support (2-factor, 8-item) a Chinese (within) Korean (within) Unconstrained (between) Constrained (between) 24.3 33.9 58.2 64.5 19 19 38 44 0.18 0.02 0.02 0.02 0.98 0.95 0.96 0.96 0.06 0.10 0.06 0.06 6.25 6 0.396
Reframing (7-item) b Chinese (within) Korean (within) Unconstrained (between) Constrained (between) 35.0 12.9 47.9 56.1 14 14 28 34 0.00 0.53 0.01 0.01 0.90 1.00 0.95 0.94 0.10 0.00 0.07 0.07 8.23 6 0.222
Spiritual support (4-item) c Chinese (within) Korean (within) Unconstrained (between) Constrained (between) 0.6 0.1 0.8 7.8 1 1 2 5 0.42 0.75 0.69 0.17 1.00 1.00 1.00 0.99 0.00 0.00 0.00 0.06 7.09 3 0.069
Note. a Social support is divided by two factors
1) relatives/friend support (Items 1, 2, 5, 20,
25) and 2) neighbor support (Items 8, 10, 29),
and item 16 has been deleted b Item 13
(showing that we are strong) in the reframing
subscale has been deleted c Error variances of
item 14 (attending church services) and item
30 (having faith in God) in the seeking
spiritual support subscale have been correlated.
27
The Modified Five-Factor Model of 23-item F-COPES
X2 df CFI RMSEA
Chinese model 350 219 0.9 0.08
Korean model 354 219 0.9 0.08
Unconstrained (a) 704 438 0.9 0.06
Measurement weight (b) 733 456 0.9 0.06
Measurement intercepts (c) 932 479 0.7 0.08
Structural covariances (d) 943 494 0.7 0.08
Measurement residuals (e) 999 518 0.7 0.08
e15
e22
e6
e9
b-a ?X228.8, ?df18, pgt0.05
28
Standardized Factor Loadings of the Modified
5-factor Model of the F-COPES
Paths Chinese-Americans Chinese-Americans Korean-Americans Korean-Americans
Paths Std coefficients Standard error Std coefficients Standard error
Relatives/friends support ? item 1 0.82nt - 0.87nt -
Relatives/friends support ? item 2 0.78 0.12 0.77 0.10
Relatives/friends support ? item 5 0.52 0.15 0.81 0.11
Relatives/friends support ? item 20 0.49 0.14 0.66 0.12
Relatives/friends support ? item 25 0.60 0.12 0.63 0.13
Neighbor support ? item 8 0.73nt - 0.84nt -
Neighbor support ? item 10 0.86 0.15 0.91 0.11
Neighbor support ? item 29 0.65 0.13 0.69 0.15
Reframing ? item 3 0.77nt - 0.83nt -
Reframing ? item 7 0.74 0.15 0.69 0.15
Reframing ? item 11 0.73 0.15 0.66 0.12
Reframing ? item 15 0.51 0.15 0.50 0.12
Reframing ? item 19 0.63 0.15 0.64 0.13
Reframing ? item 22 0.41 0.14 0.67 0.13
Reframing ? item 24 0.74 0.14 0.71 0.10
Spiritual support ? item 14 0.90nt - 0.76nt -
Spiritual support ? item 23 0.94 0.09 0.87 0.19
Spiritual support ? item 27 0.72 0.08 0.82 0.17
Spiritual support ? item 30 0.83 0.08 0.74 0.16
Mobilizing ? item 4 0.74nt - 0.82nt -
Mobilizing ? item 6 0.57 0.17 0.60 0.13
Mobilizing ? item 9 0.49 0.15 0.46 0.15
Mobilizing ? item 21 0.59 0.18 0.37 0.12
29
Validation of the F-COPES Construct Validity
Equivalence
A. Unconstrained X2(34)77.9, CFI0.91,
RMSEA0.08 B. Measurement weight X2(40)84.8,
CFI0.91, RMSEA0.08 C. Structural weight
X2(42)86.0, CFI0.91, RMSEA0.08
Relatives support
e1
.76nt (.92nt)
Neighbor support
e2
Family coping
Health-related Quality of Life
.64(.74)
.33 (.22)
e5
Reframing
e3
.70(.67)
.25(.20)
Spiritual support
-.19 (-.30)
e4
-.65 (.54)
.77(.84)
Mobilizing
e5
.35(.40)
.45(.45)
Psychological distress
Family Communication
e6
-.38(-.31)
Note. Numbers outside of parenthesis are factor
loadings/correlations for Chinese-Americans and
numbers inside of parenthesis are factor
loadings/correlations for Korean-Americans
ntnot tested plt0.05 plt0.01 plt0.001.
30
Project 2 Promoting Family Resilience Among
Couples Diagnosed with Breast, Colorectal, or
Prostate Cancer Approach to Family
Communication and Coping
Funded by NINRP30 (PI Daly) Pilot PI Jung-won
Lim
31
Study Procedures
The target population is White and
African-American breast, colorectal, and prostate
cancer survivors and their partners

MATCHED COUPLES
Randomly Assigned for the Interview
32
Sample Recruitment

• Study Site Cleveland, Ohio
• Recruitment
• 1) The University Hospital
• Cancer Center Registry
• 2) Metro Health for
• recruiting more
• African-Americans

Table 1. Eligible cancer survivors from 2007-2009
White Black Asian Total
Breast 966 244 12 1,222
Colorectal 357 67 3 426
Prostate 354 145 2 501
Total 1,677 456 17 2,149
Note. Patients diagnosed with cancer in Stages
I-III were included. Numbers of patients were
obtained from the Case Comprehensive Cancer
Center.
Recruitment Challenges how to approach to
African-Americans (ethnicity), male (gender),
colorectal cancer survivors (cancer type), and
couples?
As of today, a total of 73 couples completed the
survey.
33
Conclusion

• Low ethnic minority inclusion in cancer-related
  research limits the generalizability of findings
  and ascertainment of important contextual
  determinants.
• Cancer survivorship studies for ethnically
  diverse cancer survivor populations must be
  grounded and guided by appropriate, culturally
  consonant, theoretical, and practical frameworks.
  
• Studies addressing Asian-American subgroups
  specifically are required to test measurement
  invariance across the ethnicity.

34
Acknowledgements
Funding Sources NIH/NCI R03 (1R03CA139941)
NIH/NINR P30
(1P30NR011907) Principal Investigator Jung-won
Lim, PhD Co-Investigators Kimlin Ashing-Giwa,
PhD (City of Hope)
Barbara Daly, PhD, RN, FAAN (CWRU,
Nursing) Research Assistants Minso Paek
(MSASS) Megan Brawley (MSASS)
Elizabeth Glosik (MSASS) Betty Napoleon
(Nursing) Sophia Yeung (City of Hope)