Sage Journals: Discover world-class research

Abstract

Longevity and lower incidence of chronic disease of Seventh-day Adventists in North America are well documented. Seventh-day Adventists, in general, follow healthy lifestyle compared to the general population even though there is some variation among themselves. The objective of our study was to assess the degree to which differences in adherence to the healthy lifestyle impacts self-reported health status. Our sample consisted of 58,866 individuals in the Adventist Health Study-2 cohort recruited between 2002-2007 in the United States and Canada who answered the baseline questionnaire. In this study, we used a framework developed by Grossman to estimate health outcomes relative to behavior. Our data were based on self-reported health status, demographic and lifestyle questions, and ordered probit technique was used to measure the health investment equation. Our findings showed that lifestyles aligned with health principles of this group, particularly in nutrition, exercise, and restraint from tobacco, were associated with a higher reported health status. The predictions based on variables such as age and education were also validated in this group. Interaction between variables including race and gender show results similar to other findings. Our study shows that practices consistent with the group’s norms have higher probability of reporting excellent health.

Keywords

lifestyle health behavior Seventh day Adventists health status

“The strengths of our study include the large sample size of close to 59,000 participants from all 50 states in the United States and Canada.”

Introduction

Seventh-day Adventists are renowned to have the highest longevity in the United States. The city of Loma Linda where they make up a very large proportion of the population is the only Blue Zone in North America. The current proportion of Seventh-day Adventists in the United States population is less than .5% while they make up about 36% of the population in Loma Linda, California.

Today, many Seventh-day Adventists still adhere to most of the healthy lifestyle practices dating back to the 19^th century which have been scientifically validated over time and include other current advances in medicine in their various healthcare institutions that they operate in the United States and around the world. These institutions are open to all and pass on some of their practices to those who were not familiar with them and find the new information useful. Their focus was through health promotion which discouraged the use of tobacco, alcohol, and flesh meat, while advocating a plant-based diet and physical exercise. Recent research has corroborated the importance of lifestyle factors in their contribution to better health and a longer lifespan.^1,2

Many health studies have been conducted about Adventists from various perspectives, but to our knowledge, none has been done using a framework or model by Grossman.³ The Grossman model used in our study treats health status as a form of capital or health stock which either appreciates or depreciates depending on what the inputs are. It suggests that a body is a form of durable equipment that wears out over time, that can be maintained with appropriate health behavior, but that can also be damaged by poor maintenance and subsequently be put out of use, resulting in death.

The purpose of this study was to determine the extent to which participants in the Adventist Health Study-2 who adhere to the healthy lifestyle principles relates to the health status they report. The importance of this study is that it could demonstrate that when good health behavior is adopted on many fronts, there are significant health benefits to individuals and communities, translating into physical well-being, productive lives, and lower healthcare costs at the societal level. If findings affirm this plausibility, it could help explain why Adventists on average have a higher life expectancy relative to the general population in the United States as well as a lower incidence of lifestyle related chronic diseases.

Materials and Methods

Study Participants

All study participants for this study were from the Adventist Health Study (AHS-2): Connecting Lifestyle to Disease and Longevity, which is a follow-up on a previous research on Adventists whom studies have shown to live longer than the general population. The AHS-2 is a longstanding, large population-based prospective cohort study comprised of over 96,000 individuals. From 2002 to 2007, study participants were recruited from all 50 states in the United States as well as Canada who were 30 years or older at enrollment, mostly White (65%) and Black (27%) adults. At baseline, study participants completed a validated questionnaire on detailed demographic, health, and lifestyle characteristics including dietary practices, physical activity levels, smoking, and drinking behaviors. The purpose of investigation was to assess the association of diet with certain cancers and other health outcomes.

Hence, the AHS-2 data pool is the most comprehensive available for this type of project. The profile of study participants is reported elsewhere.⁴ Those with extreme responses were excluded from this study along with those who were neither White nor Black race as these two races represented majority of study participants. The analytic data sample size was 58,866 individuals who had answered all the relevant questions.

The Dependent Variable

The dependent variable of the stock of health capital was measured by the categorical health rating technique where individuals assessed their own health status based on the following scales: excellent, good, fair, and poor.

The Independent Variables

Income

The level of income was expected to be correlated with wages. The model assumes that wages have two effects on medical care use. On one hand, higher wages will increase demand for health capital. On the other hand, the increased opportunity cost of not working in order to seek medical care will reduce demand. The net effect is in this case ambiguous. It could be negative or positive.

Concerning the level of income, the respondents were divided into two groups: low income individuals and high-income individuals. The questionnaire divided people into 8 different ranges of personal income. The majority of responses were skewed toward the lower income distribution. We chose to put the lower three categories, which together cover a household income of $30,000 or less into the low-income category and the rest in the high-income category.

Education

It has been suggested that more educated people are more efficient producers of health;⁵ they know how to better access and make the best use of the available information to improve their health status. It was expected that education levels would be positively associated with a better health status.

Age

The model assumes that beyond a certain age, the rate of depreciation will increase and thus lower the level of health capital stock. The Grossman model treats this variable as exogenous and thus beyond the control of the individual. Investment in health capital, however, will counteract this and delay the time when the level of capital is at its minimum, precipitating death.

Gender

It was assumed that the rate of health capital depreciation would be greater for males given the higher average longevity of females, at least within a certain age bracket.

Race

Blacks were expected to report a lower health status based on other observations in the United States due to a variety of reasons, including social-economic conditions. Health capital depreciation rate is higher in this group relative to Whites.

Marital Status

It was expected that individuals who were married would report to have higher probability of excellent health compared to individuals who were single, because they would presumably enjoy greater emotional well-being which would translate into better physical health.

Physical Activity

In the book, Ministry of Healing by Ellen G. White, the natural remedies essential for good health are listed as follows: pure air, sunlight, abstemiousness, rest, exercise, proper diet, the use of water, trust in divine power.⁶ Our study focused on two of the above 8 principles of health, namely, exercise and diet. In our empirical estimation, individuals were divided into three groups: those who exercise more than 60 minutes a week, those who exercise less than 60 minutes a week, and those who do not exercise at all. It was expected that a higher amount of physical activity would be associated with a higher level of health capital and contribute to an improved physical well-being.

Dietary Patterns

A description of each dietary regime is described below:

Non-vegetarians consume meat and fish combined once a week or more, and meat once a month or more; semi-vegetarians consume meat once a month or more, and meat and fish combined less than once a week; pesco-vegetarians consume meat less than once a month and fish once a month or more; lacto-ovo-vegetarians consume meat/or fish less than once a month, and dairy and/or eggs once a month or more; vegans consume meat, fish, eggs, and dairy less than once a month. It was expected that health status would be affected by the dietary regime. The less meat consumption and other animal products, the higher the expected health status.

Alcohol Use

Alcohol consumption categories included those who never used alcohol, drank alcohol in the past for 0-9 years, 10-19 years, 20+ years, and current users. It was expected that there would be an inverse relationship between health status and the number of years of drinking alcohol. Higher alcohol consumption would be associated with lower health capital through a lower return on investment in health capital.

Smoking

This category comprises of people who never smoked, those who quit smoking 20-29.9 years ago, quit 10-19.9 years ago, 5-9.9 year ago, quit 1-4.99 years ago, and those who quit less than a year ago. It was expected there would be an inverse relationship between health status and the number of years of smoking. Smoking may increase utility to the smoker in the short run but have a negative impact on the level of health capital in the long run.

Membership Duration

The study members represented both lifelong Adventists as well as adult converts. It was assumed that those who were baptized at the age of 20 or more were of the latter category. The longer the duration and adherence to the denominational lifestyle principles, the better health status expected.

The most relevant explanatory variables relative to the Adventist lifestyle are those in the range from C.7-C.11.

Statistical Analysis

A probit model was used to generate regression results, which consisted of the threshold coefficients as well as the probabilities of reporting one’s health as either poor, fair, good, or excellent given the attributes associated with a specific respondent. We focused on the “excellent health” responses and their probabilities for each of the independent variables in this study; however, the probabilities of the other levels of health status were also recorded.

Because the dependent variable consisted of a categorical health status rating, with ordered outcomes, the Ordinary Least Squares (OLS) estimation technique would result in some shortcomings and the preferred approach is the ordered probit model which uses maximum likelihood estimation. The model assumes an unobservable latent variable h_i*, which represents the health status of an individual on a continuous scale. There are unknown thresholds for each individual that mark the change from one level of h* to the next. The model assumes a linear relationship between the latent variable (h_i*) and the independent variables such that

h_{i} * = x^{'} β + e_{i}, e_{i} \sim N (0, σ^{2})

(1)

The coefficients βs measure the marginal effects of the independent variables x_i on the latent variable h*. A positive coefficient suggests an increase in the probability of a higher status of health due to a small change in the explanatory variable. By assuming a cumulative density function of a normal distribution (Ф), and normalizing the mean and variance of ε to 0 and 1, probabilities for an individual in each category is computed and reported in Table 3.

Where appropriate, interactions were estimated to see how the combined effects reveal something of significance given the intersection of those variables, such as gender or race in connection with marital status, exercise, alcohol use, and their impact on self-reported health status. This enabled us to compare Adventists among themselves based on different demographic and lifestyle choices. The change in the probability of a particular health status was estimated using R.⁷

This study was approved by the Institutional Review Board of Loma Linda University (IRB #48134). Written, signed, and dated informed consent forms were obtained from all study participants.

Results and Discussion

As shown in Table 1, the study participants were 65.3% female with the mean age of 60.2 at enrollment. Additionally, 73.8% of our study members were White (non-Hispanic) and 26.2% of study members were Black/African-Americans. Married people represented 72.40% and single people accounted for 27.60% in our study. Our study population is unique in that only 1% of our study participants reported current smoking and 86.5% of the cohort had never smoked. Almost two thirds of the total (64.7%) were never users of alcohol. About 35.3% reported using any type of alcohol in their lifetime and only 10% reported current alcohol use. Among the latter group, the amount of alcohol consumption was very low. The AHS-2 cohort consisted of participants with a wide range of dietary patterns; vegans (∼8%), lacto-ovo-vegetarians (∼33%), pesco-vegetarians (∼10%), semi-vegetarians (∼6%), and 44% non-vegetarians. About 67.5% were lifelong members of the Adventist church.

Table 1.

Characteristics of Study Participants in the Adventist Health Study-2 (N = 58,866).

Characteristics			n	%
Self-reported health status	Poor		745	1.3
	Fair		7418	12.6
	Good		34831	59.2
	Excellent		15872	27.0
Low income	No		25560	43.4
Low income	Yes		33306	56.6
Education	High school or less		16999	28.9
	Some college		22872	38.9
	College graduate or above		18995	32.3
Age	30-39		6430	10.9
	40-49		13498	22.9
	50-64		20783	35.3
	65+		18155	30.8
Gender	Female		38412	65.3
Gender	Male		20454	34.7
Race	White		43469	73.8
Race	Black		15397	26.2
Marital status	Married		42639	72.4
Marital status	Single, divorced, widowed		16227	27.6
Exercise	None		11633	19.8
	0-60 min/wk		21208	36.0
	>60 min/wk		26025	44.2
Dietary pattern	Vegan		4691	8.0
	Lacto-ovo-vegetarian		19220	32.7
	Pesco-vegetarian		5565	9.5
	Semi-vegetarian		3293	5.6
	Non-vegetarian		26097	44.3
Alcohol use	Never		38062	64.7
	Former	Drank for 0-9 years	7750	13.2
		Drank for 10-19 years	3825	6.5
		Drank for 20+ years	3321	5.6
	Current		5908	10.0
Smoking	Never		50901	86.5
	Former	Quit <1 year ago	388	0.7
		Quit 1-4 years ago	729	1.2
		Quit 5-9 years ago	949	1.6
		Quit 10-19 years ago	1122	1.9
		Quit 20-29 years ago	4219	7.2
	Current		558	0.9
Membership duration	Lifelong		39728	67.5
Membership duration	Adult convert		19138	32.5

Table 2 shows the coefficients and the Z values for each of the independent variables; however, it is the associated probabilities of reporting different states of health that is of interest, particularly the category of excellent health whose values are reported in Table 3.

Table 2.

Ordered Probit Maximum Likelihood Estimation Results (Dependent Variable: Health Status).

Variables	Coefficients	Std Error	Z Values	Pr (>\|z\|)
High income	Ref
Low income	−.30	.011	−27.44	<2e-16
H.S. or less	−.10	.012	−7.79	6.78E-15
Some college	−.17	.012	−13.75	<2e-16
Bachelor’s degree	Ref
Age 30-39	.22	.017	12.94	<2e-16
Age 40-49	.14	.013	10.97	<2e-16
Age 50-64	Ref
Age 65+	−.20	.012	−15.93	<2e-16
Male	.10	.011	9.28	<2e-16
Female	Ref
White	Ref
Black	−.35	.012	−29.98	<2e-16
Married	Ref
Not-married	−.12	.011	−10.75	<2e-16
Exercise none	Ref
Exercise 30-60 min/wk	.27	.014	20.2	<2e-16
Exercise 60+ min/wk	.57	.013	42.87	<2e-16
Vegan	.37	.019	19.51	<2e-16
Lacto-ovo-vegetarian	.25	.012	20.74	<2e-16
Pesco-vegetarian	.19	.017	10.81	<2e-16
Semi-vegetarian	.09	.021	4.2	2.71E-05
Non-vegetarian	Ref
Alcohol current	.15	.018	8.16	3.35E-16
Alcohol former	−.04	.013	−3.32	.0009
Alcohol never	Ref
Never smoked	Ref
Quit smoking 20-29.9 yrs	−.10	.02	−5.2	2.02E-07
Quit smoking 10-19.9 yrs	−.30	.035	−8.6	<2e-16
Quit smoking 5-9.9 yrs	−.32	.038	−8.34	<2e-16
Quit smoking 1-4.9 yrs	−.39	.043	−9.06	<2e-16
Quit smoking <1 year	−.53	.058	−9.22	<2e-16
Current smoker	−.51	.049	−10.46	<2e-16
Lifelong adventist	Ref
Converted adventist	−.08	.012	−7.05	1.84E-12

Table 3.

Predicted Probabilities of Being in Excellent Health.

Variables	Probabilities
High income	30.3
Low income	20.8
Education high school diploma or less	24.6
Education some college	22.4
Education bachelor’s degree or above	27.7
Age 30-39	32.2
Age 40-49	30.0
Age 50-64	24.8
Age 65+	19.0
Female	23.6
Male	26.8
White	27.7
Black	17.3
Married	25.8
Not married	22.0
Exercise 60+ min/wk	32.2
Exercise <60 min/wk	22.3
Exercise none	15.0
Vegan	32.7
Lacto-ovo-vegetarian	28.5
Pesco-vegetarian	26.4
Semi-vegetarian	23.4
Non-vegetarian	20.7
Alcohol never	24.6
Quit alcohol 20+ yrs ago	23.7
Quit alcohol 10-19 yrs ago	23.8
Quit alcohol 0-9 yrs ago	22.9
Alcohol current	29.4
Smoked never	25.7
Quit smoking 20-29.9 yrs	22.5
Quit smoking 10-19.9 yrs	17.0
Quit smoking 5-9.9 yrs	16.6
Quit smoking 1-4.9 yrs	14.9
Quit smoking <1 year	11.8
Smoke current	12.3
Life-long adventist	25.6
Converted adventist	23.0

Income

Those who were included in the high income had a probability of 30.3% of reporting excellent health, while those in the low-income group had a probability of 20.8%. It appeared that the net effect of higher income resulted in higher demand for health capital. In comparison with the general U.S. population, there is no evidence that the income distribution of Adventists is more favorable to the latter to provide a better outcome in terms of health advantage. Examining the interaction between race and low income (data not shown), there was no statistically significant difference between Blacks and Whites in the income category in terms of their reported excellent health status.

Education

The Grossman model assumes that there is a positive relationship between the level of education presumably because more educated people are more efficient producers of health. In our study, comparing between those who have a high school diploma or less with those who have some college degree, and then those who have a bachelor degree or above, the probabilities of reporting of excellent health were, respectively, 24.6%, 22.4%, and 27.7%. Those who had a bachelor degree or above reported higher probability of being in excellent health, however, those who had a high school diploma or less had a higher probability of reporting excellent health compared to those who had some college but did not get a bachelor degree. One possible explanation may be that the latter group is more likely to occupy positions that involve a more sedentary lifestyle compared to the former that is more likely to be engaged in physical labor.⁸

We compared the educational attainment of Adventists with that of the rest of the general population in the United States for the age of 25 years and above. For the AHS-2 participants, the age range was 30 and above. In the general population, about 27% have earned a Bachelor degree or above⁹ while in the AHS-2 sample it was about 32.3%. The difference in educational achievement may partly explain the health advantage of Adventists.

Age

As stated earlier, it is assumed that aging is associated with health capital depreciation. The data from the AHS-2 supported that hypothesis. The age brackets were arranged as follows: 30-39, 40-49, 50-64, and 65 and above. The probabilities of reporting “excellent health” by individuals in these brackets were, respectively, 32.2, 30.0, 24.8, and 19.0%. The older the individual, the less likely one is to report being in excellent health. Yet, when we broke down the age distribution among Blacks and Whites, Blacks on average were younger. This difference in age distribution still did not compensate for other obstacles for reporting a more favorable health status for Blacks.

Gender

It is well-known that across the globe, women live longer than men and this is true among the AHS-2 population as well. This does not necessarily mean that women are on average healthier than men in every age group. In this study, men had a higher probability of reporting excellent health, 26.8% for men relative to 23.6% for women, the opposite of what we had expected because of the longer longevity of women. The average age difference could not explain it, because the average age of women was 58.9 and that of males was 58.0. However, when we compared the educational levels, female college graduates were 34.3% and for males 47.7%. As stated earlier, it is assumed that the higher the educational level achieved, the more efficient the production of health capital, possibly explaining this discrepancy.

Race

The probability of Whites reporting an excellent health status was 27.7% as opposed to 17.3% for Blacks. The disparity in health status between the two races is a well-documented phenomenon in the U.S. population at large, mainly reflecting the socio-economic differences between them, which create unequal access to health care.

Marital Status

Some findings generally associate being married with happiness; however, they caution that this does not apply to unhappily married people.^10,11 It is not clear what the linkage is between health and marital status, especially with changing marital norms in the past several decades. Tatangelo, et al.¹² reviewed 33 articles on gender, marital status, and longevity. They concluded that men derive more benefits from marriage than women and that the quality of marriage matters, women being more negatively affected by a poor marriage. Robards, et al.¹³ caution, however, that married and single individuals cannot be treated as homogeneous groups and that the protective effect of marriage is not that straight forward.

It is of interest to note that our empirical results based on the interaction between marital status and gender do not conform to the observation made in the review of the literature by Tatangelo et al.¹² that men enjoy greater health benefits than women. As shown in the Table 4, the gender and marital status interaction confirmed that single men reported similar probabilities having excellent health compared to married men. As far as women are concerned, a higher percentage of married women reported excellent health compared to non-married women, 25.8% vs 22%.

Table 4.

Estimated Percentages of Self-Reported Health Status by Marital Status and Sex Interaction*.

Female	Health	Estimated Percent^a	95% CI		Male	Health	Estimated Percent^a	95% CI
Marital Status	Health	Estimated Percent^a	Lower	Upper	Marital Status	Health	Estimated Percent^a	Lower	Upper
Married	Poor	.61	.55	.67	Married	Poor	.51	.46	.57
Not married	Poor	.91	.83	1.00	Not married	Poor	.52	.45	.61
Married	Fair	10.39	10.08	10.72	Married	Fair	9.40	9.07	9.75
Not married	Fair	13.08	12.63	13.53	Not married	Fair	9.53	8.83	10.28
Married	Good	63.99	63.55	64.43	Married	Good	63.13	62.65	63.61
Not married	Good	65.39	64.96	65.83	Not married	Good	63.25	62.48	64.02
Married	Excellent	25.01	24.50	25.53	Married	Excellent	26.95	26.33	27.58
Not married	Excellent	20.62	20.02	21.24	Not married	Excellent	26.69	25.28	28.15

*P < .0001 for marital status x sex interaction (likelihood ratio test).

^aPercentages were estimated from an ordered probit model including all covariates plus marital status x sex interaction.

Physical Exercise

Those who exercised for more than 60 minutes a week had a 32.2% probability of reporting that they were in excellent health while those who exercised less than 60 minutes had a 22.3% probability of reporting excellent health, and the no exercise group had a 15.0% probability of reporting excellent health.

The results of the interaction between race and physical exercise were statistically significant, Whites reporting a healthier status for the same amount of time devoted to a given activity (Table 5). On the interaction between amount of exercise and gender, there was no statistically significant difference in health status between male and female (data not shown).

Table 5.

Estimated percentages of self-reported health status by exercise level and race interaction*.

White	Health	Estimated Percent^a	95% CI		Black	Health	Estimated Percent^a	95% CI
Exercise	Health	Estimated Percent^a	Lower	Upper	Exercise	Health	Estimated Percent^a	Lower	Upper
None	Poor	1.34	1.22	1.47	None	Poor	2.32	2.07	2.60
0-60 min/wk	Poor	.62	.56	.68	0-60 min/wk	Poor	1.50	1.35	1.67
>60 min/wk	Poor	.25	.22	.27	>60 min/wk	Poor	.80	.72	.90
None	Fair	16.13	15.53	16.75	None	Fair	21.48	20.42	22.59
0-60 min/wk	Fair	10.45	10.09	10.83	0-60 min/wk	Fair	17.16	16.43	17.92
>60 min/wk	Fair	6.02	5.78	6.26	>60 min/wk	Fair	12.14	11.58	12.73
None	Good	65.81	65.38	66.23	None	Good	64.46	63.84	65.07
0-60 min/wk	Good	64.04	63.58	64.50	0-60 min/wk	Good	65.72	65.29	66.15
>60 min/wk	Good	58.09	57.56	58.61	>60 min/wk	Good	65.04	64.56	65.52
None	Excellent	16.73	16.08	17.40	None	Excellent	11.74	10.94	12.59
0-60 min/wk	Excellent	24.89	24.28	25.51	0-60 min/wk	Excellent	15.61	14.86	16.38
>60 min/wk	Excellent	35.65	35.00	36.30	>60 min/wk	Excellent	22.01	21.16	22.90

*P < .0001 for exercise level x race interaction (likelihood ratio test).

^aPercentages were estimated from an ordered probit model including all covariates plus exercise level x race interaction.

Out of the sample of 58,866 respondents, 47,233 (80%) reported that they were involved in some regular exercise; however, only 59% of the adults in the American population at large reported to exercise on a regular basis in 2007.¹⁴ This suggests that a higher percentage of Adventists are involved in some physical exercise compared to the general population and would thus benefit more in terms of health than the general population.

Dietary Patterns

As mentioned above, the dietary regime was divided into vegan, lacto-ovo-vegetarian, pesco-vegetarian, semi-vegetarian, and non-vegetarians. Vegans who avoid the consumption of any animal products such as meat and cheese had the highest report of excellent health with a probability of 32.7%, lacto-ovo-vegetarians came in second with a probability of 28.5%, pesco-vegetarians came in third with a probability of 26.4%, semi-vegetarians report excellent health with a probability of 23.4%, and finally, non-vegetarians reported excellent health with a 20.7% probability. On the interaction between gender and dietary regime, there was no significant difference in terms of reporting excellent health status between men and women. This was also true in the interaction between dietary regime and race (data not shown).

It is evident from this report that individuals who adopted some form of vegetarianism had a higher probability of reporting excellent health compared to those who did not. Our vegetarian group represented about 56% of the respondents. Reinhart¹⁵ indicates that in a 2018 Gallup poll, 5% of Americans reported that they were vegetarians, same figure as in 2012, and 3% reported to be vegan, figures that are substantially lower than those of Adventists. Thus, it can be surmised that Adventists for the most part adopt dietary habits that are healthier and conducive to good health.

Alcohol Use

The great majority of Adventists shun particularly alcohol and tobacco. Only 10% of the AHS-2 respondents indicated that they were current users of alcohol, a significantly lower proportion than the general population, where 51% reported to be current users of alcohol in the Results from the 2007 National Survey on Drug Use and Health.¹⁶

In our study, former drinkers had 22.9%-23.8% probability of reporting excellent health. Curiously, current alcohol drinkers had a higher probability of reporting excellent health (29.4%), even above those who have never consumed alcohol (24.6%). This is not an unusual finding as far as moderate drinking is concerned.

Other studies have reported that moderate alcohol consumption is believed to improve health.^17-19 More recently, however, the opinion that moderate drinking is advantageous to health has come under scrutiny and the pendulum has been moving in the opposite direction. A study by Stockwell, et al.²⁰ indicated that meta-analysis of 87 studies examining the relationship between moderate drinking and total mortality risk reproduced the J-shaped curve. However, after adjusting for abstainers who were biased toward ill-health, and quality-related study characteristics, they could not find reduced mortality risk due to moderate drinking.

In the interaction between alcohol use and race, Whites reported a better health status in all categories of alcohol consumption (Table 6). It is worth pointing out that among those who reported drinking more than 3 drinks a month, a larger percentage of Blacks used liquor than Whites (25% vs 19%), while 31% of current Black drinkers consumed wine against 41% among Whites (data not shown). Wine typically has a significantly lower concentration of alcohol compared to liquor.

Table 6.

Estimated percentages of self-reported health status by alcohol use and race interaction.

White	Health	Estimated Percent^a	95% CI		Black	Health	Estimated Percent^a	95% CI
Alcohol Use	Health	Estimated Percent^a	Lower	Upper	Alcohol Use	Health	Estimated Percent^a	Lower	Upper
Never	Poor	.50	.46	.55	Never	Poor	1.22	1.11	1.34
Current	Poor	.30	.26	.35	Current	Poor	1.12	.93	1.33
Former	Poor	.56	.50	.62	Former	Poor	1.37	1.23	1.54
Never	Fair	9.27	8.98	9.58	Never	Fair	15.38	14.82	15.96
Current	Fair	6.84	6.38	7.34	Current	Fair	14.65	13.38	16.02
Former	Fair	9.87	9.44	10.32	Former	Fair	16.41	15.59	17.26
Never	Good	62.99	62.53	63.44	Never	Good	65.79	65.36	66.21
Current	Good	59.69	58.81	60.55	Current	Good	65.72	65.27	66.18
Former	Good	63.55	63.03	64.08	Former	Good	65.78	65.35	66.20
Never	Excellent	27.24	26.70	27.78	Never	Excellent	17.61	16.97	18.28
Current	Excellent	33.17	31.88	34.48	Current	Excellent	18.51	16.91	20.22
Former	Excellent	26.01	25.21	26.83	Former	Excellent	16.44	15.56	17.36

*P < .0014 for alcohol x race interaction (likelihood ratio test).

^aPercentages were estimated from an ordered probit model including all covariates plus alcohol x race interaction.

Smoking

As stated above, Adventists are encouraged to abstain from the consumption of tobacco and only 1% reported to be current smokers compared to about 28.6% of current smokers in the U.S. population, 12 years and older according to the Results from 2007 National Survey on Drug Use and Health.¹⁶ As expected, those who reported never having smoked had the highest probability of reporting excellent health (25.7%). The probabilities consistently increased in relationship to how long study participants reported to having quit smoking. In other words, the longer the years from the time members reported having stopped smoking, the higher the probability of reporting the status of excellent health. Greater abstinence from smoking relative to the general population gives an additional edge to Adventists health-wise. The interaction between race and current smoking showed a statistically significant greater negative effect on the health status of Whites compared to Blacks (data not shown).

Membership Duration

Of those who reported on their age at which they joined the group, 68% were lifelong Adventists and the rest were adult converts. Lifelong members had a higher probability (25.6%) of reporting excellent health compared to a probability of 23.0% for adult converts. The likely interpretation is that lifelong Adventists have enjoyed the benefits of Adventist lifestyle from birth and thus would report better health compared to those who began that lifestyle later on in life.

Conclusions

Our study sought to determine whether there was a link between close adherence to health guidelines of Seventh-day Adventists and the reporting of excellent health status. Our study findings show that almost always, the practices consistent with the group’s recommended healthy lifestyle leads to the highest probability of reporting excellent health. When compared with nation-wide data in a number of variables where data are available and comparable, the participants in the AHS-2 show that they differ in terms of dietary practices, and other habits such as exercising, smoking, and alcohol consumption; they are more likely to participate in what is currently considered health-enhancing activities and shunning unhealthy practices.

Seventh-day Adventists have been of interest to health researchers due to the fact that they live longer than any other groups in the United States. The town of Loma Linda, the only “blue zone” in North America where people live the longest, has an unusually large population of Adventists, accounting for about 36% of its inhabitants compared to a nationwide proportion of less than .5%. This would suggest that this group has some health practices that can make a difference in health outcomes and that can provide some ideas to healthcare policymakers and researches on how other communities can benefit from their health behaviors. Given the close association between good health and longevity, this could explain the phenomenon of their highest longevity in the United States.

The strengths of our study include the large sample size of close to 59,000 participants from all 50 states in the United States and Canada. In addition, this was the first attempt at linking the Grossman model with the AHS-2 data and the findings showed consistency between the predictions made by the model and the results of our analyses. Although the participants are part of the group that encourages healthy lifestyles there is a wide degree of adopting the recommendations. Therefore, this group was ideal for this type of study by comparing health behaviors and health status outcomes.

There are limitations to our study. Our study used data on Adventists in the United States and Canada, and did not make a distinction between Americans and Canadians. Canadian Adventists represent only 4.6% of the study participants. Given the similarity in culture and standard of living between two countries, it was deemed reasonable to use the results and compare them to the general U.S. population for certain variables where data were available.

Another limitation is that the measurement of the health status in our study was based on self-reporting, but this type of categorical health measure is an accepted approach and has been used to predict mortality^21-23 and to test the Grossman model using a Swedish sample.²⁴ The explanatory variables, income and education, could have reverse effects, thus leading to endogeneity. In addition, healthy individuals are able to exercise while those who are either disabled or have serious illness might not have regular exercise routine, leading to potential bias in our estimates.

Overall, we think this study sheds some light on the benefits of lifestyle choices in terms of health status and what communities can achieve when they consistently practice health-enhancing activities and encourage others to do the same. In recent years, it has become evident that one of the major causes of diseases negatively impacting sustained longevity in developed countries is preventable detrimental health behavior.²⁵ This study validates that view. In fact, reduction in tobacco consumption in the U.S. over the years is a case in point. The National Institutes of Health²⁶ reported in 2012 that between 1975 and 2,000, a total of 800,000 deaths were averted due to a re5duction of smoking in the United States.²⁶

Further Study Directions

The cost of health care has consistently risen faster than costs in other sectors of the U.S. economy and taken a larger share of the Gross Domestic Product, a proportion that is an outlier relative to other industrialized countries without superior results in terms of health outcomes. A comprehensive study on the degree to which Adventists spend less money on health care would provide an estimate of how much society would save if a large segment of the population adopted a similar lifestyle and possibly lower the trend of ever-rising costs of health care in the United States. Health outcomes depend to a large extent on what we do or fail to do and not necessarily on factors beyond our control. Given the sustained increase in health costs and intractable health inequities in society, the old adage still provides the best advice: an ounce of prevention is worth a pound of cure.

Footnotes

Acknowledgments

We acknowledge the support of Mr. Keiji Oda for his statistical assistance.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported NIH/National Cancer Institute [1U01CA152939-01A1].

ORCID iDs

Leonard Gashugi

Jisoo Oh

References

Pan

Wang

, et al. Impact of healthy lifestyle factors on life expectancies in the US population. Circulation. 2018;138:345-355. doi:10.1161/circulationaha.117.032047

Larrick

Mendelsohn

. Finally, a regimen to extend human life expectancy. Rejuvenation Res. 2018;21:278-282. doi:10.1089/rej.2018.2088

Grossman

. On the concept of health capital and the demand for health. J Polit Econ. 1972;80:223-255.

Butler

Fraser

Beeson

, et al. Cohort profile: The adventist health study-2 (AHS-2). Int J Epidemiol. 2008;37:260-265. doi:10.1093/ije/dym165

Grossman

. The correlation between health and schooling. In: Terleckyj

, ed. Household Production and Consumption. Cambridge, MA: National Bureau of Economic Research; 1976.

White

. Ministry of Healing. Mountain View: Pacific Press Publishing Association; 2003.

R Core Team . R: A Language and Environment for Statistical Computing. Austria: R Foundation for Statistical Computing Vienna; 2019.

Saunders

McIsaac

Douillette

, et al. Sedentary behaviour and health in adults: An overview of systematic reviews. Appl Physiol Nutr Metab. 2020;45:S197-S217. doi:10.1139/apnm-2020-0272

Census US . Educational Attainment in the United States: 2007-Detailed Tables. Suitland: Census US; 2009. https://www.census.gov/data/tables/2007/demo/educational-attainment/p20-560.html. accessed September 13, 2021.

10.

Lawrence

Rogers

Zajacova

Wadsworth

. Marital happiness, marital status, health, and longevity. J Happiness Stud. 2019;20:1539-1561. Article. doi:10.1007/s10902-018-0009-9

11.

Lee

Bulanda

. Change and consistency in the relation of marital status to personal happiness. Marriage Fam Rev. 2005;38:69-84. Article. doi:10.1300/J002v38n01_06

12.

Tatangelo

McCabe

Campbell

Szoeke

. Gender, marital status and longevity. Maturitas. 2017;100:64-69.

13.

Robards

Evandrou

Falkingham

Vlachantoni

. Marital status, health and mortality. Maturitas. 2012;73:295-299.

14.

Few

. Americans Meet Exercise Targets: Self-reported rates of physical exercise show little change since 2001. Washington, DC: Gallup; 2008. https://news.gallup.com/poll/103492/few-americans-meet-exercise-targets.aspx https://news.gallup.com/poll/103492/few-americans-meet-exercise-targets.aspx. accessed September 13, 2021.

15.

Reinhart

. Snapshot: Few Americans Vegetarian or Vegan. Washington, DC: Gallup; 2018. https://news.gallup.com/poll/238328/snapshot-few-americans-vegetarian-vegan.aspx. accessed July 27, 2021.

16.

DHHS . Results From the 2007 National Survey on Drug Use and Health. Washington, DC: National Findings; 2008.

17.

Ashley

Ferrence

Room

Bondy

Rehm

Single

. Moderate drinking and health. Implications of recent evidence. Can Fam Physician. 1997;43:687-694.

18.

Ogborne

Smart

. Public opinion on the health benefits of moderate drinking: Results from a Canadian National Population Health Survey. Addiction. 2001;96:641-649.

19.

Roerecke

Rehm

. Alcohol consumption, drinking patterns, and ischemic heart disease: A narrative review of meta-analyses and a systematic review and meta-analysis of the impact of heavy drinking occasions on risk for moderate drinkers. BMC Med. 2014;12:182. doi:10.1186/s12916-014-0182-6

20.

Stockwell

Zhao

Panwar

Roemer

Naimi

Chikritzhs

. Do “moderate” drinkers have reduced mortality risk? A systematic review and meta-analysis of alcohol consumption and all-cause mortality. J Stud Alcohol Drugs. 2016;77:185-198.

21.

Torres-Collado

García de la Hera

Compañ-Gabucio

, et al. Self-reported health status and mortality from all-causes of death, cardiovascular disease and cancer in an older adult population in Spain. PLoS One. 2022;17:e0261782. doi:10.1371/journal.pone.0261782

22.

Lorem

Cook

Leon

Emaus

Schirmer

. Self-reported health as a predictor of mortality: A cohort study of its relation to other health measurements and observation time. Sci Rep. 2020;10:4886. doi:10.1038/s41598-020-61603-0

23.

Cho

Wang

Yabroff

, et al. Estimating life expectancy adjusted by self-rated health status in the United States: National health interview survey linked to the mortality. BMC Publ Health. 2022;22:141. doi:10.1186/s12889-021-12332-0

24.

Gerdtham

Johannesson

. New estimates of the demand for health: Results based on a categorical health measure and Swedish micro data. Soc Sci Med. 1999; 49: 1325-1332. DOI: 10.1016/s0277-9536(99)00206-3

25.

Rizzuto

Fratiglioni

. Lifestyle factors related to mortality and survival: A mini-review. Gerontology. 2014;60:327-335. doi:10.1159/000356771

26.

NIH . Nearly 800,000 Deaths Prevented Due to Declines in Smoking. Maryland, USA: NIH; 2012. https://www.nih.gov/news-events/news-releases/nearly-800000-deaths-prevented-due-declines-smoking. accessed September 7, 2021.

Lifestyle-Related Behavior and Self-Reported Health Status Among Seventh-Day Adventists

Abstract

Keywords

Introduction

Materials and Methods

Study Participants

The Dependent Variable

The Independent Variables

Income

Education

Age

Gender

Race

Marital Status

Physical Activity

Dietary Patterns

Alcohol Use

Smoking

Membership Duration

Statistical Analysis

Results and Discussion

Income

Education

Age

Gender

Race

Marital Status

Physical Exercise

Dietary Patterns

Alcohol Use

Smoking

Membership Duration

Conclusions

Further Study Directions

Footnotes

Acknowledgments

Declaration of Conflicting Interests

Funding

ORCID iDs

References