1. Patient demographics
We enrolled 666 patients, and 658 patients were included in this analysis, as the other 6 patients had incomplete data. There was a female predominance in the study participants, with 439 (66.72%) to be females, and 219 (33.28%) to be males (Table 1). The median age was 63-year-old with a range between 26 to 100 years old. Most of the participants were over 50 years old, while those younger than 50 years old were only 18.24%.
Table 1
Demographics and Clinical Features of Patient Population
Patient Characteristics
|
N (%*)
|
Total Number of Patients
|
658
|
Gender
|
|
Male
|
219 (33.28)
|
Female
|
439 (66.72)
|
Age
|
|
Median (Range)
|
63 (26-100)
|
Age ≤ 30
|
8 (1.22)
|
Age 31-49
|
112 (17.02)
|
Age 50-69
|
355 (53.95)
|
Age ≥ 70
|
183 (27.81)
|
Cancer Types
|
|
Breast
|
218 (31.28)
|
CNS
|
7 (1)
|
GI
|
132 (18.94)
|
GU
|
62 (8.9)
|
GYN
|
79 (11.33)
|
Head-Neck
|
15 (2.15)
|
Lung
|
77 (11.05)
|
Hematological
|
73 (10.47)
|
Other
|
13 (1.87)
|
Unknown
|
0 (0)
|
Multiple Cancer Types
|
21 (3.01)
|
Disease Stage
|
|
Stage 0-1
|
122 (18.54)
|
Stage 2-3
|
297 (45.14)
|
Stage 4
|
213 (32.37)
|
Not Applicable
|
25 (3.8)
|
Unknown
|
1 (0.15)
|
* Percentage calculated with Total Number of Patients as the denominator.
|
All cancer types were represented in the participants, the most common to be breast cancer (31.28%), and some patients had multiple types of cancers. There were patients of all disease stages, 122 (18.547%) in stage 0-1, 297 (45.14%) in stage 2-3, and 213 (32.37%) in stage 4 (Table 1).
2. Immigration status and ethnicity:
Of the study population, 301 (45.74%) were US born, the rest were immigrants including 12 (1.82%) from Africa, 62 (9.42%) from Canada or Europe, 106 (16.11%) from East and South Asia, 151 (22.95%) from Latin America and the Caribbean islands, and 22 (3.34%) from Middle-Eastern countries (Table 3). English was the first language in 390 (59.27%) patients, followed by Chinese (n=87, 13.22%), Spanish (n=50, 7.6%) and Russian (n=31, 4.71%) (Table 2).
Table 2
Race and Immigration Background of Patient Population
Patient Characteristics
|
N (%*)
|
Total Number of Patients
|
658
|
Birth Place
|
|
US Born
|
301 (45.74)
|
Non-US Born
|
353 (53.65)
|
Africa
|
12 (1.82)
|
Canada or Europe
|
62 (9.42)
|
East and South Asia
|
106 (16.11)
|
Latin America and the Caribbean
|
151 (22.95)
|
Middle-Eastern
|
22 (3.34)
|
Unknown
|
4 (0.61)
|
First Language
|
|
English
|
390 (59.27)
|
Arabic
|
24 (3.65)
|
Chinese
|
87 (13.22)
|
Indian
|
9 (1.37)
|
Russian
|
31 (4.71)
|
Spanish
|
50 (7.6)
|
Multiple
|
10 (1.52)
|
Others
|
56 (8.51)
|
Unknown
|
1 (0.15)
|
Number of Years Living In The US
|
|
US Born
|
301 (45.74)
|
Non-US Born
|
347 (52.74)
|
< 2 Years
|
9 (1.37)
|
2 - <10 Years
|
38 (5.78)
|
>= 10 Years
|
300 (45.59)
|
Unknown
|
10 (1.52)
|
Race
|
|
African
|
198 (30.09)
|
Non-US Origin
|
98 (14.89)
|
US Origin
|
100 (15.20)
|
Asian
|
113 (17.17)
|
Chinese
|
87 (13.22)
|
Non-Chinese
|
26 (3.95)
|
Caucasian
|
246 (37.39)
|
Hispanic
|
80 (12.16)
|
Non-US Origin
|
32 (4.86)
|
US Origin
|
48 (7.29)
|
Middle-Eastern
|
4 (0.61)
|
Multi-Racial
|
8 (1.22)
|
Other
|
8 (1.22)
|
Unknown
|
1 (0.15)
|
* Percentage calculated with Total Number of Patients as the denominator.
|
Table 3
Social Economic Data of Patient Population
|
Patient Characteristics
|
N (%*)
|
Total Number of Patients
|
658
|
Educational Status
|
|
Grade School or Less
|
76 (11.55)
|
Less than Grade School
|
20 (3.04)
|
Grade School
|
56 (8.51)
|
High School
|
295 (44.83)
|
College or Graduate
|
285 (43.31)
|
College/University
|
211 (32.07)
|
Graduate School
|
74 (11.25)
|
Unknown
|
2 (0.3)
|
Marital Status
|
|
Married
|
325 (49.39)
|
Not Married
|
319 (48.48)
|
Divorced
|
66 (10.03)
|
Separated
|
12 (1.82)
|
Single
|
175 (26.6)
|
Widowed
|
66 (10.03)
|
Unknown
|
14 (2.13)
|
Zip Code (Poverty %)
|
|
Brooklyn
|
578 (87.84)
|
< 20% below poverty
|
225 (34.19)
|
20% - < 25% below poverty
|
215 (32.67)
|
25% - < 30% below poverty
|
138 (20.97)
|
Unknown
|
1 (0.15)
|
Non-Brooklyn
|
79 (12.01)
|
Anxiety Level
|
|
0
|
2 (0.3)
|
1-3
|
401 (60.94)
|
4-6
|
126 (19.15)
|
7-10
|
121 (18.39)
|
Unknown
|
8 (1.22)
|
* Percentage calculated with Total Number of Patients as the denominator.
|
Table 4
Univariate Analysis of Selected Patient Population
|
Patient Characteristics
|
All Patients, N (%)*
|
CAM Users, N (%)**
|
Univariate p-value
|
Total Number of Patients
|
658
|
435 (66.11)
|
|
Gender
|
|
|
|
Male
|
219 (33.28)
|
119 (54.34)
|
1.1322E-05
|
Female
|
439 (66.72)
|
316 (71.98)
|
Baseline
|
Age
|
|
|
|
< 38
|
32 (4.86)
|
15 (46.88)
|
0.0215
|
≥ 38
|
626 (95.14)
|
420 (67.09)
|
Baseline
|
Cancer Types
|
|
|
|
Breast and GYN
|
296 (44.98)
|
214 (72.3)
|
0.00252
|
Not Breast or GYN
|
362 (55.02)
|
221 (61.05)
|
Baseline
|
Unknown***
|
0 (0)
|
0 (0)
|
|
Race
|
|
|
|
African, Non-US and US Origin
|
198 (30.09)
|
144 (72.73)
|
0.0371
|
Asian, Chinese and Non-Chinese
|
113 (17.17)
|
62 (54.87)
|
0.1144
|
Caucasian and Others
|
266 (40.43)
|
169 (63.53)
|
Baseline
|
Hispanic, Non-US and US Origin
|
80 (12.16)
|
59 (73.75)
|
0.0928
|
Unknown***
|
1 (0.15)
|
1 (100)
|
|
Birth Place
|
|
|
|
US Born and Others
|
397 (60.33)
|
264 (66.5)
|
Baseline
|
East and South Asia
|
106 (16.11)
|
57 (53.77)
|
0.0161
|
Latin America and the Caribbean
|
151 (22.95)
|
113 (74.83)
|
0.0608
|
Unknown***
|
4 (0.61)
|
1 (25)
|
|
First Language
|
|
|
|
English
|
511 (77.66)
|
351 (68.69)
|
Baseline
|
Chinese
|
96 (14.59)
|
48 (50)
|
0.000489
|
Spanish
|
50 (7.6)
|
35 (70)
|
0.84855
|
Unknown***
|
1 (0.15)
|
1 (100)
|
|
Asian
|
|
|
|
Asian, Chinese
|
87 (13.22)
|
43 (49.43)
|
0.04
|
Asian, Non-Chinese
|
26 (3.95)
|
19 (73.08)
|
Baseline
|
Hispanic
|
|
|
|
Hispanic, Non-US Origin
|
32 (4.86)
|
24 (75)
|
1.00
|
Hispanic, US Origin
|
48 (7.29)
|
35 (72.92)
|
Baseline
|
Marital Status
|
|
|
|
Married
|
325 (49.39)
|
199 (61.23)
|
0.0102
|
Non-Married
|
319 (48.48)
|
226 (70.85)
|
Baseline
|
Unknown***
|
14 (2.13)
|
10 (71.43)
|
|
Educational Status
|
|
|
|
Graduate School
|
74 (11.25)
|
55 (74.32)
|
0.117
|
Less Than Graduate School
|
582 (88.45)
|
379 (65.12)
|
Baseline
|
Unknown***
|
2 (0.3)
|
1 (50)
|
|
|
658 (100)
|
435 (66.11)
|
|
* Percentage calculated with Total Number of Patients as the denominator.
|
** Percentage calculated with Total Number of Patients for the subgroup at interest.
|
*** 'Unknown' values are ignored during univariate analysis.
|
Most of the immigrants have lived in the U.S. for more than 10 years (n=300, 45.59%, which was 86.46% of all immigrants), and very few (n=9, 1.37%, or 2.59% among all immigrants) were new immigrants of less than 2 years (Table 2).
Patient self-identified themselves into the following race groups: African (US Origin) (n=100, 15.20%), African (non-US Origin) (n=98, 14.89%), Asian (n=113, 17.17%), Caucasian (n=246, 37.39%), Hispanic (n=80, 12.16%), Middle-East2ern (n=4, 0.61%), multi-racial (n=8, 1.22%) and others (n=8, 1.22%) (Table 2).
3. The social economic status of the patients
The social economic status of the patients is displayed in Table 3. Regards to the education level, most of the patients received high school (44.83%), college or graduate schools (43.31%) education. Only 76 patients (11.55%) received only grade school or less education.
About half of the patients (49.39%) were married, and the other half of the patients had a status of single at the time of survey (n=319, 48.48%), including divorced (n=66, 10.03%), separated (n=12, 1.82%), single (n=175, 26.60%), or widowed (n=66, 10.03%).
A majority (n=578, 87.84%) of the patients lived in the same borough of Brooklyn where the cancer center is, while 79 (12.01%) patients resided in other boroughs. One hundred and thirty-eight (20.97%), 215 (32.67%), and 225 (34.19%) patients lived in regions with high, intermediate and low percentages of families with below poverty.
Patients were asked for their anxiety level on a 1-10 points scale. Most of the patients (n=401, 60.94%) had low level of anxiety (score 1-3), while 126 (19.15%), and 121 (18.39%) patients had intermediate (score 4-6) and high levels (score 7-10) of anxiety, respectively.
4. The difference in CAM use among different categories: demographics and clinical factors
Of the 658 patients, 435 (66.11%) patients used one or more types of CAM. The prevalence of CAM use was higher in females (71.98%) than the males (54.34%) (p=1.13x10−5). The age distribution of patients using CAM is illustrated in Figure 1. By treating age as categorical variables to be below 38 years old, and equal or above 38 years old, we found that patients ≥ 38 years old had significantly more CAM use (67.09%) than < 38 years old (46.88%) (p=0.0215) (Table 4). By dividing patients in the age groups of 31-49, 50-69 or age ≥ 70, there was no difference in CAM use (data not shown).
Patients with earlier stages of disease had numerically higher CAM use than those at stage 3-4 (Supplemental Table 1A), and patients with breast and GYN cancers had higher CAM use (72.3%, p=0.00252), consistent with the data on the higher CAM use in females (Table 3).
5. The difference in CAM use among different categories: Race/ethnicity and immigration status
CAM use appeared to have differences among different races at initial analysis (supplemental Table S1B). We then grouped them into 4 race groups of (1) African American (US born + non-US born, n= 198, 30.09%), (2) Asian (n=113, 17.17%), (3) Caucasian + others (including Middle-Eastern, Multi-Racial and Others) (n=266, 40.43%), and (4) Hispanics (n=80, 12.16%) (Table 4), and their CAM use rate were 72.73%, 54.87%, 63.53% and 73.75%, respectively. Taking Caucasians and others as baseline (CAM use 63.53%), patients of African Americans had statistically higher CAM use (72.73%) (p=0.0371). Asian patients appeared to have used CAMs less than the Caucasians, and the Hispanics appeared to have used CAMs more than the Caucasians, but the difference in those comparisons did not reach statistical significance. In this comparison, patients in each race category may be US born or non-US born.
In dividing the patients by their birth-place, there was no difference of CAM use between the US born patients (n=301, CAM use 68.77%) and all the immigrants (n=353, CAM use 64.31%) as a whole (Supplement Table S1B). However, Asian born immigrants (n=106) had statistically less CAM use (53.77%) than the US born and others (n=397, CAM use 66.50%) (p=0.0161), while the Latin-American born had a trend towards higher CAM use (n=151, CAM use 74.83%, p=0.0608) (Table 4).
In examining the first language use, we grouped all patients into 3 categories: (1) English and others, (2) Chinese and Indian, and (3) Spanish speaking. The CAM use rate were 68.69%, 50.00% and 70.00%, respectively. Taking English speaking group as baseline, the Chinese speaking patients had statically lower CAM use than the English-speaking patients (p=0.000489); while the difference in CAM use between the Spanish speaking patients and the English-speaking patients was not statistically different (p=0.8485) (Table 4).
In the group of Asian patients by race (n=113) (Table 4), 87 were Chinese, and 26 were non-Chinese. The CAM use was 49.43% in Chinese and 73.08% in non-Chinese respectively (p=4.30E-02), showing the Chinese patients were unlikely to use CAM.
In the group of Hispanic patients (n=80) as shown in Table 4, 48 were of US born, and 32 were of non-US born. The CAM use was similar (72.92% vs 75%) in the two groups, without statistical difference.
The number of years living in the US by the immigrants were studied. The CAM use was compared among US born, versus new immigrants (living in the U.S. less than 2 years), immigrants living in the U.S. for 2-10 years, or more than 10 years, and there was no difference among groups (Supplemental Table S1B).
6. The difference in CAM use among different categories: Socioeconomic status
Married patients had a statistically lower CAM use (n=325, 61.23%) than the non-married (n=319, 70.85%), taken together the single status due to divorce, separation, death of the partner, or true single status (p=0.0102) (Table 4).
The association of level of education with CAM use was examined by initial statistical test of the 5 different levels (less than grade school, grade school, high school, college, graduate school), and there was no statistical difference (supplemental Table S1C). We then consolidated the patients to 2 groups: (1) graduate school, and (2) less than graduate school. The patient group with the highest level of education (graduate school) had a numerically higher rate of CAM use (74.32%) than the rest of the categories, while each of the other categories had similar CAM use rate, and there was no statistical difference either (Table 4 and Supplemental Table S1C).
The poverty levels of the patients were roughly determined according to the zip code of their residence, linked to the New York City defined percentages of families at poverty levels. Patients residing in areas with 3 different levels of poverty, high, intermediate and low, had similar likelihood of using CAM, and there was no statistical difference (Supplemental Table S1C).
Patients with low, intermediate or high anxiety levels were equally likely to use CAM, and there was no statistical difference among the 3 groups (Supplemental Table S1C).
In the multi-variate analysis, the effect of age, female sex, patients with breast and GYN cancers remained as statistically significant factors to predict higher CAM use (Supplemental Table S2).
7. CAM subtype use analysis
In this analysis, we counted the prevalence of the use of each of the 21 subtypes of CAM, in association with the different race groups, as listed in Supplemental Table S3, Supplemental Figure S1 and in Figure 2. As one patient may use more than one CAM subtypes, the total occurrence of CAM use was 435 (Supplemental Table S3). The most commonly used CAM subtype was “Prayer and Spirituality” (n=281, 64.6%), followed by “Dietary” (n=175, 40.23%), and “Herbal” (n=138, 31.72%). Other subtypes which had more than 5% uses were “Bodywork and Movement Therapy” (n=85, 19.54%), “Meditation” (n=67, 15.4%), “Guided Imagery” (n=63, 14.48%), “Physical Therapy” (n=59, 13.56%), “Relaxation Therapy” (n=46, 10.57%), “Energy Medicine” (n=32, 7.36%), “Massage Therapy” (n=31, 7.13%), Homeopathy (n=23, 5.29%) (Supplemental Table S3, Supplemental Figure S1 and in Figure 2).
Furthermore, about 84.72% of CAM users of African American patients of the combined US and Non-US origin used Prayer and Spirituality (n=122, 84.72%), which was the highest among all the ethnic groups. It was followed by Hispanics (n=42, 71.19%), Caucasians (n=84, 52.50%), and Asians (n=25, 40.32%). Chiropractic therapy was exclusively used by Caucasian CAM users (n=15, 9.38%) (Supplemental Table S3).
Prayer and Spirituality was also the most common CAM subtype among all types used by Asian CAM users (40.32%). Among all patients who used herbal medicine, there was no difference between African American, Hispanics, Caucasians and Asians (Supplemental Table S3 and Figure 2).
Of all the 435 CAM use occurrences, 150 (34.5%) patients used only 1 type, 119 (27.4%) patients used 2 types, 72 (16.6%) patients used 3 types, and 93 (21.4%) patients used equal or more than 4 types of CAM subtypes (Supplemental Table S4 and Supplemental Figure S2).