Astin JA. Why Patients Use Alternative MedicineResults of a National Study. JAMA. 1998;279(19):1548-1553. doi:10.1001/jama.279.19.1548
From the Stanford Center for Research in Disease Prevention, Stanford University School of Medicine, Palo Alto, Calif.
Context.— Research both in the United States and abroad suggests that significant
numbers of people are involved with various forms of alternative medicine.
However, the reasons for such use are, at present, poorly understood.
Objective.— To investigate possible predictors of alternative health care use.
Methods.— Three primary hypotheses were tested. People seek out these alternatives
because (1) they are dissatisfied in some way with conventional treatment;
(2) they see alternative treatments as offering more personal autonomy and
control over health care decisions; and (3) the alternatives are seen as more
compatible with the patients' values, worldview, or beliefs regarding the
nature and meaning of health and illness. Additional predictor variables explored
included demographics and health status.
Design.— A written survey examining use of alternative health care, health status,
values, and attitudes toward conventional medicine. Multiple logistic regression
analyses were used in an effort to identify predictors of alternative health
Setting and Participants.— A total of 1035 individuals randomly selected from a panel who had agreed
to participate in mail surveys and who live throughout the United States.
Main Outcome Measure.— Use of alternative medicine within the previous year.
Results.— The response rate was 69%.The following variables emerged as predictors
of alternative health care use: more education (odds ratio [OR], 1.2; 95%
confidence interval [CI], 1.1-1.3); poorer health status (OR, 1.3; 95% CI,
1.1-1.5); a holistic orientation to health (OR, 1.4; 95% CI, 1.1-1.9); having
had a transformational experience that changed the person's worldview (OR,
1.8; 95% CI, 1.3-2.5); any of the following health problems: anxiety (OR,
3.1; 95% CI, 1.6-6.0); back problems (OR, 2.3; 95% CI, 1.7-3.2); chronic pain
(OR, 2.0; 95% CI, 1.1-3.5); urinary tract problems (OR, 2.2; 95% CI, 1.3-3.5);
and classification in a cultural group identifiable by their commitment to
environmentalism, commitment to feminism, and interest in spirituality and
personal growth psychology (OR, 2.0; 95% CI, 1.4-2.7). Dissatisfaction with
conventional medicine did not predict use of alternative medicine. Only 4.4%
of those surveyed reported relying primarily on alternative therapies.
Conclusion.— Along with being more educated and reporting poorer health status, the
majority of alternative medicine users appear to be doing so not so much as
a result of being dissatisfied with conventional medicine but largely because
they find these health care alternatives to be more congruent with their own
values, beliefs, and philosophical orientations toward health and life.
IN 1993 Eisenberg and colleagues1 reported
that 34% of adults in the United States used at least 1 unconventional form
of health care (defined as those practices "neither taught widely in U.S.
medical schools nor generally available in U.S. hospitals") during the previous
year. The most frequently used alternatives to conventional medicine were
relaxation techniques, chiropractic, and massage. Although educated, middle-class
white persons between the ages of 25 and 49 years were the most likely ones
to use alternative medicine, use was not confined to any particular segment
of the population. These researchers estimated that Americans made 425 million
visits to alternative health care providers in 1990, a figure that exceeded
the number of visits to allopathic primary care physicians during the same
Recent studies in the United States2
and abroad3,4 support the prevalent
use of alternative health care. For example, a 1994 survey of physicians from
a wide array of medical specialties (in Washington State, New Mexico, and
Israel) revealed that more than 60% recommended alternative therapies to their
patients at least once in the preceding year, while 38% had done so in the
previous month.2 Forty-seven percent of these
physicians also reported using alternative therapies themselves, while 23%
incorporated them into their practices.
When faced with the apparent popularity of unconventional medical practices
and the fact that people seem quite willing to pay out-of-pocket for these
services,1 the question arises: What are the
sociocultural and personal factors (health status, beliefs, attitudes, motivations)
underlying a person's decision to use alternative therapies?
At present, there is no clear or comprehensive theoretical model to
account for the increasing use of alternative forms of health care. Accordingly,
the goal of the present study was to develop some tentative explanatory models
that might account for this phenomenon.
Three theories that have been proposed to explain the use of alternative
medicine were tested:
Dissatisfaction: Patients are dissatisfied with
conventional treatment because it has been ineffective,5,6
has produced adverse effects,6,7
or is seen as impersonal, too technologically oriented, and/or too costly.6- 15
Need for personal control: Patients seek alternative
therapies because they see them as less authoritarian16
and more empowering and as offering them more personal autonomy and control
over their health care decisions.14,16- 19
Philosophical congruence: Alternative therapies
are attractive because they are seen as more compatible with patients' values,
worldview, spiritual/religious philosophy, or beliefs regarding the nature
and meaning of health and illness.19- 24
In addition to testing the validity of these 3 theoretical perspectives,
this study also sought to determine on an exploratory basis how the decision
to seek alternative therapies is affected by patients' health status and demographic
Participants completed an extensive mail survey that gathered information
on use of alternative health care, perceived benefits and risks of these therapies,
health beliefs and attitudes, views toward and experiences with conventional
medicine, political beliefs, and worldview. The original survey instrument
was developed by Ray,24 and the survey was
conducted through National Family Opinion, Inc, which maintains a panel of
persons who have agreed to be participants in mail surveys. This panel constitutes
a representative national sample from which subsamples can be drawn. A random
sample of 1500 individuals was drawn from this panel, with 1035 people completing
the questionnaire (a response rate of 69%).
Following Eisenberg et al,1 the dependent
variable, alternative health care use, a dichotomous measure, was operationalized
as used within the previous year of any of the following treatments: acupuncture,
homeopathy, herbal therapies, chiropractic, massage, exercise/movement, high-dose
megavitamins, spiritual healing, lifestyle diet, relaxation, imagery, energy
healing, folk remedies, biofeedback, hypnosis, psychotherapy, and art/music
therapy. Several of these treatments, however, were deemed not to be alternative
or unconventional if they were used to treat particular health-related problems:
(1) exercise for lung problems, high blood pressure, heart problems, obesity,
muscle strains, or back problems; (2) psychotherapy for depression or anxiety;
and (3) self-help groups for depression or anxiety. The category "alternative
medicine" was thus delimited to exclude those practices that are already part
of standard medical care and recommendations such as exercise to treat hypertension
or psychotherapy to treat depression. (The category "lifestyle diet" could
include more standard or conventional dietary recommendations such as a low-fat
or low-salt regimen for treating cardiovascular disease or hypertension.)
Analyses were repeated using a second dependent variable, primary reliance
on alternative medicine, a dichotomous measure defined by those respondents
who reported using primarily alternative therapies to treat health-related
Table 1 lists the independent
variables considered possible predictors of alternative health care use. Since
constructs like "satisfaction with conventional medicine" are highly generalized,
multiple measures of these variables were used. Using principal components
analysis with varimax rotation of selected questionnaire items, 4 multi-item
factors were identified: satisfaction with conventional practitioners; health
status; belief in the power of religious faith to heal; and belief in the
efficacy of conventional medicine.
Since the dependent variable was dichotomous, logistic regression analyses
were carried out. Demographic variables were entered in a first block with
the remaining variables entered in a second block. These variables were entered
together in the second block because their precise causal ordering was not
readily apparent (ie, there was no clear theoretical rationale for entering
them in separate blocks). The variables that then remained significant (P<.05) in the logistic regression analyses constituted
the final multivariate model.
The following hypothesized relationships were tested in the multiple
Users of alternative health care will be distinguished
from nonusers in that they will (a) report less satisfaction
with conventional medicine; (b) demonstrate a greater
desire to exercise personal control over health-related matters; and (c) subscribe to a holistic philosophical orientation to
Since the majority of health care alternatives
are not covered by insurers, having access to more financial resources will
predict use of alternative medicine.25,26
As suggested by previous research,1,26- 28
higher levels of education will be predictive of alternative medical use.
Users of alternative health care will be more likely
to be part of a cultural group, described by Ray24,29
as "cultural creatives," and identifiable by the following values: commitment
to environmentalism; commitment to feminism; involvement with esoteric forms
of spirituality and personal growth psychology, self-actualization, and self-expression;
and love of the foreign and exotic. These individuals tend to be at the leading
edge of cultural change and innovation, coming up with the most new ideas
in the society, and are therefore hypothesized to be more inclined to use
alternative health care.
(Ray developed his value classifications,
what he termed "value subcultures," empirically using factor analysis and
multidimensional scaling to create orthogonal value dimensions. K-means clustering
was then used to cluster respondents into the different value groupings. According
to Ray,24 the cultural creative group has been
steadily growing in the culture at large since the late 1960s and now represents
approximately 44 million Americans [23.6% of the adult population]. While
there is likely some crossover in terms of values and orientation with those
identified by the popular media as New Agers, the latter term has no operational
definition while the categorization of cultural creative is based on empirical
research examining specific values held by individuals in the culture at large.)
Those who report relying primarily on alternative
forms of health care will be more likely to subscribe to a holistic philosophy
of health (their greater commitment to these health practices being reflected
in a set of health beliefs that are more congruent with many forms of alternative
Survey respondents were comparable to census data from the same time
period with the exception of a slight underrepresentation of younger, less
educated, and poor persons (Table 2).
Respondents were asked whether they had experienced any of a list of
26 health-related problems within the past year (Table 1). They were then asked to list the 3 most "bothersome" or
"serious" ones. The top 5 problems listed were (1) back problems (19.7%);
(2) allergies (16.6%); (3) sprains/muscle strains (15.7%); (4) digestive problems
(14.5%); and (5) lung problems, pneumonia, or respiratory infections (13%).
Forty percent of respondents reported using some form of alternative
health care during the past year. The top 4 treatment categories were chiropractic
(15.7%); lifestyle diet (8.0%); exercise/movement (7.2%); and relaxation (6.9%).
The most frequently cited health problems treated with alternative therapies
were chronic pain (37%); anxiety, chronic fatigue syndrome, and "other health
condition" (31% each); sprains/muscle strains (26%); addictive problems and
arthritis (both 25%); and headaches (24%).
Analyses were also carried out to determine which specific treatments
were being used for which therapeutic modalities. Table 3 lists the top 10 health problems (in terms of percentage
who treated them with alternative medicine) and the most frequently used alternative
therapies for each.
Although certain alternative therapies tended to be used more frequently,
a broad range of alternatives were, in fact, being used for the majority of
health problems. For example, although chiropractic care represented close
to 50% of all alternative treatments used for headaches, individuals also
reported using acupuncture, homeopathy, megavitamins, spiritual healing, lifestyle
diets, relaxation, massage, folk medicine, exercise, psychotherapy, and art/music
therapy to treat this health problem. A similar pattern is evident across
many of the health problems listed on the survey; ie, although particular
alternative treatments may predominate, use is by no means confined to any
particular therapy or even a few therapies.
The following variables predicted use of alternative medicine in the
multiple logistic regression (criterion for entering was P<.05): (1) being more educated; (2) being classified in the value
subculture of cultural creatives; (3) having a transformational experience
that changed the person's worldview; (4) having poorer overall health; (5)
believing in the importance of body, mind, and spirit in treating health problems
(holistic health philosophy); and (6) reporting any of the following health
problems: anxiety, back problems, chronic pain, or urinary tract problems. Table 4 presents the intercorrelations
of all hypothesized predictors and use of alternative medicine. Table 5 presents the adjusted odds ratios and 95% confidence intervals
for the independent variables that emerged as significant predictors.
Contrary to a number of previous findings6- 13,27
and the present study's hypothesis, negative attitudes toward or experiences
with conventional medicine were not predictive of alternative health care
use. Among those who reported being highly satisfied with their conventional
practitioners (54%), 39% used alternative therapies, while 40% of those reporting
high levels of dissatisfaction (9% of respondents) were users of alternative
Although there was a trend in the direction of those desiring to keep
control in their own hands being more likely to use alternative medicine,
this variable was also not a significant predictor of alternative medicine
use as hypothesized.
Racial/ethnic differences also did not predict use of alternative medicine.
Use was found across all groups (eg, whites, 41%; blacks, 29%; Hispanics,
40%). (Percentages of Asian and Native American respondents who used alternative
medicine are not reported here as their overall numbers in the sample are
too small. Also, the fact that certain ethnic groups had relatively low representation
in the sample may explain why they did not emerge as predictors in the regression.)
No significant differences were found with respect to sex with 41% of women
and 39% of men reporting use of alternative health care. Finally, neither
income nor age predicted use of alternative medicine in the regression.
The results do, however, provide strong support for the philosophical/value
congruence theory in several ways. First, as hypothesized, having a holistic
philosophy of health ("The health of my body, mind, and spirit are related,
and whoever cares for my health should take that into account") was predictive
of alternative health care use. Among those subscribing to this philosophy,
46% reported being users of alternative medicine, while only 33% of those
not endorsing the item were users. This finding suggests that use of alternative
medicine may, in part, reflect shifting cultural paradigms, particularly with
respect to recognizing the importance of spiritual factors in health. Second,
the statement, "I've had a transformational experience that causes me to see
the world differently than before," also emerged as a significant predictor.
Of those who answered "yes" (18.3%), 53% reported use of alternative health
care compared with 37% of those who responded "no" or "not sure." Third, those
categorized as cultural creatives were significantly more likely to use alternative
health care. Among this subcultural group, 55% reported using alternative
health care compared with only 35% of those not in this group.
Education emerged as the 1 sociodemographic variable that predicted
use of alternative medicine; individuals with higher educational attainment
were more likely to use alternative forms of health care (eg, 31% of those
with high school education or less reported use compared with 50% of those
with graduate degrees).
The 3-item factor, health status, also emerged as a significant predictor
of alternative health care use, with use increasing as health status declined.
A number of specific health problems (ie, back problems, chronic pain, anxiety,
and urinary tract problems) were also predictive of alternative health care
use. These results suggest that experiencing certain health problems increases
the likelihood that one will be a user of alternative medicine in a general
sense (ie, not simply to treat that particular disorder). For example, those
individuals citing anxiety as 1 of their 3 most serious health problems were
almost twice as likely as nonanxiety sufferers (67% vs 39%) to be users of
alternative health care.
To test the validity of the logistic regression model, 2 techniques
were used. First, predicted values from the multivariate equation were divided
into quintiles. The percentage of respondents within each quintile who used
alternative medicine was then calculated. This analysis is typically used
to assess the extent to which there is any clinical or policy relevance to
the predictor variables beyond their being statistically significant.30 Within the quintile of lowest predicted value scores,
17% used alternative medicine; within the highest quintile, 68% were users.
These results suggest that the model is fairly strong and has practical (not
merely statistical) significance.
To further examine the model's validity, the sample was randomly split
into 2 even subsamples, and separate logistic regressions were run for each.
These multivariate models were then compared, and there were no significant
differences observed in the coefficients of each model. Finally, predicted
values were again divided into quintiles in each subsample, and the spread
of probabilities across each group was quite consistent between each model
and in comparison with the overall regression model.
To test whether individuals who report relying primarily on alternative
forms of health care show a different profile from those who use alternative
medicine more in conjunction with conventional means, separate logistic analyses
were carried out. This exploratory analysis suggests that primary reliance
on alternative forms of medicine is explained by a considerably different
set of variables. The following independent variables were significant predictors
in the multiple logistic regression: (1) distrust of conventional physicians
and hospitals; (2) desire for control over health matters; (3) dissatisfaction
with conventional practitioners; and (4) belief in the importance and value
of one's inner life and experiences. The fact that only 4.4% (n=45) of the
sample was categorized as relying primarily on alternative forms of health
care is consistent with previous findings 1
suggesting that the vast majority of individuals appear to use alternative
therapies in conjunction with, rather than instead of, more conventional treatment.
In contrast to individuals who use alternative therapies in conjunction
with conventional medicine, for whom dissatisfaction with conventional medicine
was not a significant predictor of alternative health care use, 2 of the 4
predictors of primary reliance on alternative medicine reflect a general lack
of trust in and satisfaction with conventional medical care. It is also individuals
who report a desire to keep control in their own hands who are more likely
to report relying primarily on unconventional forms of health care.
Education and health status did not predict primary reliance on alternative
medicine. Neither being a cultural creative nor holding a holistic philosophy
of health was a significant predictor in this model. These findings suggest
that, contrary to my hypothesis, those evidencing a greater commitment to
or reliance on alternative health care may be doing so primarily as a result
of their dissatisfaction with conventional medicine rather than on ideological
or philosophical grounds.
Because of the small sample size in the above analyses and the relatively
imprecise measure of the dependent variable (ie, one can only infer that respondents
who report relying primarily on these alternatives tend to use them more as
a replacement than as a complement to conventional approaches), one must interpret
these findings with caution.
Perceived benefits of alternative therapies were considered as potential
determinants of use (eg, if someone reports receiving some benefit from a
given treatment, this could in turn serve as an important determining factor
in future health care decisions). The 2 most frequently endorsed benefits
were, "I get relief for my symptoms, the pain or discomfort is less or goes
away, I feel better," and "The treatment works better for my particular health
problem than standard medicine's." These responses suggest that the most influential
or salient factor in people's decision to use alternative health care may
be its perceived efficacy. The response, "The treatment promotes health rather
than just focusing on illness," was the third most frequently reported benefit
and offers further support for the philosophical congruence theory.
The present study was designed to provide a comprehensive analysis of
factors influencing the decision to use various forms of alternative health
care. Based on the results from the multiple logistic regression, users of
alternative medicine (40% of those surveyed) can generally be characterized
as follows: Users tend to be better educated and to hold a philosophical orientation
toward health that can be described as holistic (ie, they believe in the importance
of body, mind, and spirit in health). They are more likely to have had some
type of transformational experience that has changed their worldview in some
significant way, and they tend to be classified in a value subculture as cultural
creatives. Users of alternative health care are also more likely to report
poorer health status than nonusers.
Relief of symptoms is the main benefit reported (the perceived efficacy
of alternative medicine being cited nearly twice as often as other reported
benefits). A central finding is that users of alternative health care are
no more dissatisfied with or distrustful of conventional care than nonusers
Among those categorized as primarily reliant on alternative health care—fewer
than 5% of the surveyed population—a different pattern emerged. Unlike
those who used alternative therapies in conjunction with or as a supplement
to conventional forms of medical care, these individuals were more likely
to be dissatisfied with and distrustful of standard care as well as desirous
of maintaining exclusive control over their health care decisions. They were
also more likely to report being interested in their inner life and experiences,
suggesting some crossover with the set of spiritually relevant variables that
predicted nonexclusive use of alternative health care. These results suggest
that future studies examining predictors of alternative health care use need
to more carefully measure this phenomenon so that individuals who use these
therapies in conjunction with or as a supplement to conventional means can
be clearly distinguished from those who use them predominantly or more exclusively.
Several possible interpretations can be offered for certain variables
that emerged as predictors of alternative health care use. Education, for
example, may increase the likelihood that people will (1) be exposed to various
nontraditional forms of health care through their own reading of popular or
academic books on the subject; (2) educate themselves about their illnesses
and the variety of treatments available to them; and/or (3) question the authority
of conventional practitioners (ie, be less inclined simply to accept unquestionably
the physician's knowledge and expertise).
There are also at least 2 possible explanations for the finding that
poorer health status predicts alternative medical use. First, since those
who are in poor health have, by definition, had less success in treating their
health problems, their continued suffering may have prompted them to seek
out alternatives. Second, a significant number of individuals who report poor
health, more pain, disability, and physical symptoms may be somatizers. Somatization
has been defined as "the propensity to experience and report somatic symptoms
that have no pathophysiological explanation, to misattribute them to disease,
and to seek medical attention for them."31
Since research suggests that somatizers are disproportionately high users
of medical services, get more medical tests, and tend to experiment with (shop
around for) different health care practitioners, it seems reasonable that
they would be more likely to seek out various health care alternatives.32 It would be useful to design future studies examining
predictors of alternative health care use in such a way that somatizers and
nonsomatizers can be differentiated more clearly.
There are also several possible explanations for the finding that alternative
medicine users are more likely to subscribe to a holistic philosophy of health.
People who hold this philosophical orientation may be attracted to alternative
forms of health care because they see in these therapeutic systems a greater
acknowledgment of the role of nonphysical (mind/spirit) factors in creating
health and illness. An alternate explanation (which would reverse the direction
of causation) is that people who have been involved with alternative medicine
have had their belief systems influenced by these therapeutic modalities and
the philosophies underlying them.
That users of alternative health care are more likely to report having
had a transformational experience that changed the way they saw the world
lends partial support to the hypothesis that involvement with alternative
medicine may be reflective of shifting cultural paradigms regarding beliefs
about the nature of life, spirituality, and the world in general. As suggested
by Charlton,20 a subset of individuals may
be attracted to these nontraditional therapies because they find in them an
acknowledgment of the importance of treating illness within a larger context
of spirituality and life meaning.
The apparent effect of one's spiritual/philosophical orientation on
involvement with alternative health practices is further supported by the
finding that being a cultural creative is a significant predictor of use.
This suggests that the growing interest in alternative medicine may not simply
represent a shift in individual beliefs about the nature of health and illness,
but is rather a phenomenon that is transmitted through and influenced by the
culture. This interpretation is supported by the finding that the effect of
membership in this value subculture is not accounted for simply by holding
a holistic philosophy of health; that is, both of these variables contributed
independently in the logistic regression equation.
As with other studies that attempt to explain complex human behavioral
phenomena, a significant amount of variance is not explained by the regression
equation. There are obviously unaccounted for variables such as a general
openness to novelty and experimentation, or curiosity, that need to be examined
in future studies. Another possibility is measurement error associated with
the independent variables; for example, the variable "belief in the importance
of body, mind, and spirit" might be interpreted differently depending on one's
religious background. Moreover, the decision to use alternative medicine is
sufficiently context or situation dependent (eg, influence of significant
others who have used or not used various alternatives) to make prediction
Another limitation to this study is its cross-sectional nature, which
precludes drawing any definitive conclusions regarding cause-and-effect relationships.
For example, it is unclear whether holding a holistic philosophical orientation
has led certain individuals to seek out alternative therapies, whether exposure
to these therapies has somehow influenced the way they view health and illness,
or whether both effects occur. Moreover, the reliance on self-report may weaken
the internal validity of the study as retrospective accounts of one's health
status, health practices, and reasons for making certain health care decisions
may be subject to distortion and inaccuracy.
Since the sample underrepresented the poorer, less educated, and non–English-speaking
segments of the population, it is unclear if (and how) the results would be
different had these groups been better represented. It is possible that the
modest overrepresentation of more educated respondents in the study sample
may have slightly inflated the estimates of use of alternative therapies.
Finally, since information could not be obtained on nonrespondents, there
remains the possibility of some self-selection bias in the study sample.
Despite these limitations, the study results make several contributions
to our understanding of alternative health care use. First, the results provide
useful information to conventional practitioners about the health beliefs
and practices of many of their patients and may suggest areas where practitioners
and the present health care system may be failing to meet peoples' health
care needs adequately. This seems particularly important given research suggesting
that the vast majority of medical symptoms are self-diagnosed and self-treated33 and that a significant portion of alternative medical
use (eg, use of herbal therapies and nutritional supplements) falls into the
realm of self-care.1 Subsequently, if health
care professionals are to effectively support individuals in making informed,
safe, and appropriate choices, it is critical that they develop greater awareness
of the nature of, potential efficacy of, and reasons for patients' use of
unconventional self-care approaches.
Second, the results can help identify and clarify prevailing cultural
conceptions about and attitudes toward health and illness and examine the
degree to which the growing interest in alternative medicine may represent
a type of cultural (Kuhnian34) paradigm shift
regarding health beliefs and practices. Results from the present study lend
support to the notion that for many individuals, the use of alternative health
care is part of a broader value orientation and set of cultural beliefs, one
that embraces a holistic, spiritual orientation to life.
Third, the information derived from this and similar studies can serve
as a useful adjunct to data derived from controlled studies of the clinical
efficacy of alternative therapies. These combined research efforts not only
have the potential to change some of the ways conventional biomedicine is
practiced, but can also serve to stimulate further dialogue among the biomedical
community, governmental agencies, insurance companies, and managed care organizations
regarding the potential value of alternative treatments.
Finally, as policymakers and health care professionals continue to debate
reforms of the present health care system, it seems important to understand
why a significant portion of the population is going outside mainstream biomedicine
to treat a variety of illnesses and to maintain their general health and well-being.