8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
1/14
ASSESSING VARIATION IN HEALTH STATUS IN THE
ANDES: A BIOCULTURAL MODEL
KATHRYN S. OTHS
Department of Anthropology, University of Alabama, P.O. Box 870210, Tuscaloosa, AL 35487-0210,U.S.A.
AbstractResearch on health status in the Andes highlands is synthesized to demonstrate that intracul-tural diversity in the region is greater than often appreciated. Understanding the range of diversity inhealth status requires a biocultural model of sickness processes. Ecological, sociocultural, political-econ-omic and historical factors, such as altitude, mode of production, labor relations and land reform, areshown to combine to produce varying levels of health as measured at the community level. The model
is illustrated both by comparing communities and by the in-depth examination of a single communityin the northern Peruvian highlands. A rapid assessment checklist is provided to aid health agents in bet-ter assessing the relative health of communities in a potential intervention area. # 1998 Elsevier ScienceLtd. All rights reserved
Key wordsAndes, biocultural, health status measurement, international health, medical anthropology,intracultural diversity
INTRODUCTION
Towards the collaborative aim of updating research
on Andean health, I propose a synthesis of some of
the diverse health status data that has been gener-
ated over the years by Andeanist biological and
social scientists. My aims here are twofold: (1) to
recognize and embrace the biocultural diversity
found in the high Andes, in order to begin to
understand its eects on health status of Andean
people and (2) to suggest that international aid
agencies must take heed of this intra-Andean diver-
sity if their programs are to be implemented suc-
cessfully.
Chugurpampa*, the site of my own research, is
part of an arid potato growing region centered at
3300 m in the western escarpment of the Northern
Peruvian Andes. My ndings regarding health indi-
cators there tend to diverge from those of other
areas, primarily the southern part of Bolivia andPeru, where the majority of anthropological work
has been done. This has led me to consider how we
can integrate cultural, ecological, political economic
and social factors to account for these intra-Andean
dierences in health status.
The need to take into account diversity within a
population is an oft-stated, if seldom heeded, axiom
in anthropology (Pelto and Pelto, 1975; Vayda,
1994). Likewise, the idea of microdierentiation in
the Andes is not a new idea. The 4700 mile-long
Andes mountain range contains within its borders a
tremendous amount of diversity, both ecological
and cultural. The original notion of Murra (1972)
of ecological zonation documented the ecological
diversity within an ayllu. Coming to recognize the
complexity and diversity of Andean ecology, Murra(1985) has recently updated his notion of ecological
zones by taking into account (1) horizontal as well
as vertical dierentiation (what he now calls ``eco-
logical complementarity'') and (2) the need to corre-
late ecological niches with the type of socio-cultural
systems that have developed within, and in turn
inuenced, them.
I suggest that Murra's notion of ecological and
sociocultural dierentiation can be protably
applied to the greater Andes region in general and
to other aspects of Andean life in particular, such
as health. Many years ago, Buck et al. (1968) high-
lighted the health status dierences between fourPeruvian highland and lowland communities. More
recently, Carey (1990) and Leatherman (1994), have
gone further by linking social structural factors
with ecological niches to determine health dier-
ences between three neighboring communities in the
Nun oa district. Their work shows signicant health
dierences by village, each of which obtains its live-
lihood under a dierent set of social, political and
economic arrangements.
In this paper, I hope to build upon these import-
ant earlier contributions by going a step further and
comparing villages from dierent regions of the
Andes. Too often one reads over-generalized asser-tions about Andean health, such as ` infant mor-
tality is high at high-altitude'' or ``malnutrition is
Soc. Sci. Med. Vol. 47, No. 8, pp. 10171030, 1998# 1998 Elsevier Science Ltd. All rights reserved
Printed in Great Britain0277-9536/98 $19.00 + 0.00
PII: S0277-9536(98)00161-0
*Chugurpampa is a communidad campesina (communal
peasant community) of 902 inhabitants located in theDepartment of La Libertad, in the Province of Otuzco,district of Julcan.
1017
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
2/14
endemic in the Peruvian highlands''. Assuming simi-
lar health conditions for all marginalized indigenous
groups throughout the Andes would be as ill-
advised as assuming the health of rural Blacks inAlabama were the same as that of rural Blacks in
Minnesota. This paper aims to correct any attempt
to generalize about health conditions in ` the
Andes''.
HEALTH STATUS COMPARISONS BETWEEN
COMMUNITIES
Following the methodological advice of de Meer
et al. (1993), the comparisons below will be based
exclusively upon the results of carefully conducted
empirical studies that include socio-cultural corre-
lates. Data from large-scale national censuses andsurveys are included as reference points only; as
Go mez admits (Go mez, 1988, p. 16), statistics such
as the Pan-American Health Organization collects
for Peru may be based on estimates and inferences.
In the service of making the main point about
health diversity in the highlands, from time to time
I will be introducing heretofore unpublished data
from my own research in northern Peru.
The Chugurpampa health data was collected
from March 1988 through April 1989. Initially, a
census and health survey were conducted of all 166
households of the agricultural hamlet, which are
dispersed over 1000 hectares. Birth and mortality
rates for 1987 were calculated on the basis of thehealth survey and, using the census as a base, all
births and deaths in the community were recorded
by the researcher as they occurred during 1988. All
data was cross-checked against district civil registry
records*. From July to December, 1988, case illness
histories were collected for all members of 32 ran-
domly selected households. Height and weight
measurements of infants and children and adult
CMI scores were also obtained during this period.The stimulus for and verication of the obser-
vations made here would not have been possible
without intensive participant observation for
3 months in 1987 prior to, and throughout, the
study period. This included eating meals with doz-
ens of families, staying in many households, work-
ing in the elds, attending Sunday market and
campesino community meetings, accompanying the
midwife and other healers to see patients, teaching
classes, enjoying estas and in general, immersing
myself in every aspect of daily life (for details, see
Oths, 1991).
The epidemiologic transition Merrick (1986) has
noted for Latin American in general is also occur-
ring in Chugurpampa. Serious problems of infec-
tious disease, infant diarrhea and malnutrition have
been largely alleviated through cultural interven-
tions in the form of immunization, pure water, and
adequate sanitation and diet (see McKeown,
1976){. However, these Andeans are hardly illness
free. Acute viral respiratory and chronic musculos-
keletal illnesses make up nearly 2/3 of the reported
illnesses in Chugurpampa. Gastrointestinal pro-
blems, while accounting for only 14% of all illness
complaints, represent the majority of grave illnesses,
most of which occur in adults.
Similarities as well as dierences in illness pat-terns are apparent throughout the Andes. Table 1
shows that the same three general illness types, res-
piratory, musculoskeletal and gastrointestinal, are
also the most highly ranked in two other highland
Andean communities for which comparable data
was available, Cuyo Cuyo in Southern Peru
(Larme, 1993) and Saraguro in Southern Ecuador
(Finerman, 1985){. A lowland Bolivian community
is included for contrast. The salience of these three
illnesses is widely documented across the Andes by
other researchers, including Little and Baker (1976,
p. 417), Dutt and Baker (1978, p. 33), Donahue
(1981, p. 225), Bastien (1987, p. 50), Leatherman
and Thomas (1987, p. 223), Carey (1990, p. 273)
and Mitchell (1991, p. 30). Comparisons show that
respiratory illnesses are the most common com-
plaint at the higher altitudes, especially where the
climate is dry. Gastrointestinal complaints, which
include infant diarrhea and parasites, tend to be
more frequent at the lower elevations. The amount
of musculoskeletal disorder varies with the mode of
production, which may be partially associated with
altitude and terrain.
While the pattern of illness types varies across
contexts, the severity of illnesses does as well. When
it comes to health status indicators such as infant
mortality, crude death rate, maternal reproductiveloss or days missed from work due to illness,
Chugurpampa's rates stands apart from other sites.
*Relying on civil registry data alone would have overesti-mated the birth rate and underestimated the death ratefor Chugurpampa.
{Of the 25 cases of respiratory infection recorded in a20% random sample of Chugurpampa infants (N= 8)during a 24 week case collection period, four (16%)were grave in severity. Of 8 g.i. ailments, 1 case each
of diarrhea and empachado (intestinal blockage) weregrave (25%). No deaths occurred. These gures weresimilar to the rates for the surrounding district ofJulcan (population 37,000), including 390 infants. Thedistrict health post doctor recorded one infant death,from diarrhea, for the 6 month period that overlappedmy study period.
{Symptom patterns may change seasonally (see Carey,1988, p. 147); whether the eects of these changesmight be reected in the types of illness or simply thequantity of symptoms is not known. Data in Saraguroand Cuyo Cuyo were based on 12 months of data col-lection, compared to 6 months (JulyDecember) inChugurpampa. To achieve comparability across sites,respiratory illnesses included cold/u, sore or swollenthroat, asthma and chest pain. Musculoskeletal com-plaints also included accidents/injuries, rheumatism,
body aches and swollen legs. The category of g.i. pro-blems subsumed stomach aches, vomiting, diarrhea,parasites, liver problems and constipation.
K. S. Oths1018
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
3/14
Aware of established concerns about the design of
small area variation studies (see Parchman, 1995 for
caveats), I fully recognize the potential lack of uni-
formity in methods, sampling and interpretation of
symptoms among researchers whose data is used
herein. There is insucient data to compare all
samples statistically; nonetheless, some tentative
comparisons might be useful.
The infant mortality of 31 per 1000 live births in
Chugurpampa is one of the lowest reported for the
Andes above 2500 m (Table 2). At the other
extreme, Larme reported 242 deaths per 1000 births
in Southern Peru for the same year, while Cruz-
Coke found a rate of 333 in Chile 30 years ago.
Geographical, altitudinal and temporal trends are
discernible in Table 2. It appears that IMR
increases the more southern the location of the vil-
lage or district (r = 0.50), the higher the altitude
(r = 0.54) and the earlier the date, with the excep-
tion of Cuyo Cuyo (r = 0.43)*. The decrease in
rates over time reects the decline in mortality due
to infectious disease, or demographic transition
(Merrick, 1986).
Similar to its IMR, Chugurpampa's crude death
rate, at 6 per 1000 population, is less than half that
of the next lowest rate found recorded, as shown inTable 3. The high is 50 in Chile in 1965 and, for
more recent years, 27 in Cuyo Cuyo. Altitude, lo-
cation and year are again mildly correlated with the
death rate in degrees similar to that for IMR.
Maternal reproductive loss is measured as the
mean percentage of children who have died at any
time since birth during a mother's lifetime. While
not good, the rate of 18.5% in Chugurpampa is
somewhat lower than Carey found in any of the
three communities in the Nun oa district (p = n.s.)
and nearly 1/2 that of Larme's rate found in two
dierent communities in Cuyo Cuyo (Table 4), with
the dierence (combining Larme's two samples) sig-
nicant at the p < 0.01 level (t = 3.04). Dividing
Larme's and my samples into women older and
younger than 40 reveals an even smaller risk of
reproductive loss for Chugurpampa's younger
women compared to Cuyo Cuyo's, suggesting
greater improvement in Chugurpampa's health over
time.
Finally, the workdays lost due to illness in
Chugurpampa, at 0.44 days per 2 weeks, are a frac-
tion of those lost in Nun oa, as seen in Table 5. The
lowest Carey found was 2.05 days in Nun oa town
and the highest, 3.25 days in Chillihua. While there
were no signicant dierences between any of
Carey's three sites, each of which is engaged in a
distinct mode of production, the dierence between
Chugurpampa and each Nun oa community was sig-
nicant at p < 0.001 (Sincata, t = 5.7; Nun oa,
t = 3.8; Chillihua, t = 6.6). All rates were based on
a 2 week recall period during the same season.
MODELING ALTERNATIVE EXPLANATIONS OF
DIVERSITY
Carey alerts us ``to be sensitive to the possibility
that subtle dierences within sociocultural systems
may have important health consequences'' (Carey,
1990, p. 278). How do the political-economic sys-
tem, sociocultural meanings and biological pro-
cesses generate the dierent patterns of illness? A
biocultural approach is necessary to identify the
combination of factors that might be accounting for
the dierences in health status indicators that are
shown above.
At the outset, I acknowledge that any compari-
son between communities can suer from unad-
justed demographic dierences, such as age
distribution or sex ratio. Since health conditions
change over time, the period in which data werecollected can also inuence comparisons, as
Leonard (1989) found with nutritional status in
Table 1. The highest ranked illness types and their frequencies by Andean community
Community
Cuy o Cuyo ( 19 88 ) Chugu rpa mpa (1 98 8) Sar agu ro ( 19 84 ) Mont ero ( 19 77)
Illness typeRespiratory (%) 26 40 22 20Musculoskeletal (%) 28 18 10 NAGastrointestinal (%) 18 14 18 26Altitude (m) 3600 3300 2500 lowlandZone South Peru North Peru South Ecuador BoliviaSource Larme (1993) Oths (1991) Finerman (1985) Frerichs et al. (1980)
Larme: Convenience sample of 10 households in each of two communities ( N= 107 individuals) chosen to represent a range of economicstrategies, income levels, household composition and stages in the domestic cycle. Symptom rates were collected by researcher throughrepeat illness follow-up interviews with household heads for 1 year.
Oths: Random sample of 32 households (approximately 20% of population) drawn from censused community. Researcher collected ill-nesses cases at 2 week intervals for 6 months for all family members from all household members present at time of interview.
Finerman: Sample of 140 women (approximately 3% of population) drawn from volunteers of 4 Saraguro communities who had com-pleted one or more successful pregnancies. Researcher had each woman complete a monthly health questionnaire on family's symp-toms for 1 year.
Frerichs: Household health survey of 605 randomly selected households (3372 individuals) in Montero region carried out by 6 nursesaliated with health services delivery demonstration project.
NA = data not available.
*The rank order correlations reported take the commu-nities as the unit of analysis.
Assessing variation in health status in the Andes 1019
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
4/14
Table 2. Infant mortality rates for high-altitude andean peasant communities per 1000 live births
Site Rate Location Year Altitude Source
Saraguro 100 South Ecuador 1981 2500 Finerman (1985)Chugurpampa 31 North Peru 1988 3300 Oths (1991)Vicos 122 Central Peru 1952 2800 Alers (1965)Vicos 143 Central Peru 1963 2800 Alers (1965)Quinua district 124 South Peru 1955 3250 Mitchell (1991)Quinua district 54 South Peru 1985 3250 Mitchell (1991)Nun oa district 135 South Peru 195069 4000 Spector (1971)Nun oa district 129 South Peru 197584 4000 Carey (1988)Cuyo Cuyo 242 South Peru 198186 3600 Larme (1993)Belen 142 North Chile 1966 3200 Cruz-Coke et al.
(1966)Chapiquin a 130 North Chile 1966 3400 Cruz-Coke et al.
(1966)Hullatire 333 North Chile 1966 4300 Cruz-Coke et al.
(1966)
Ecuador 66 national 1987 all USDC (1987)Peru 90 national 1987 all USDC (1987)Bolivia 127 national 1987 all USDC (1987)
Chile 19 national 1987 all USDC (1987)
Finerman: Infant mortality rate was calculated in 1981 by researcher from birth histories of a subsample of 76 of the 140 women volun-teers selected for the symptom survey (see Table 1). The recall covered the years from 1931 to 1981 and deaths are counted only ifoccurring in the rst 6 months of life, in contrast to the standard method of using the rst 12 months. This results in an underestima-tion of the rate compared to the others in the table. On the other hand, due to the epidemiologic transition, the rate probably smoothout a dierence between a higher IMR in 1931 and a lower rate by 1981.
Oths: Calculated on health survey and census data of entire community conducted by researcher, which included current pregnancies.Pregnancies and births followed-up for 1 year through collaboration with village midwives, illness research, community surveillance,and monitoring of cemetery. Data was cross-checked through review of District birth and death records. Method helped avoid biasessuch as the ocial underreporting of births or deaths, especially of preterm or very young infants, as well as the misreporting ofbirths and deaths for communities in which they did not occur.
Alers: Rates based on records kept by Cornell-Peru Project and registry of Marcara District. The Project's two community censuses, in1952 and 1963, serve as baseline. This population study was the rst of its kind for the Andes.
Mitchell: Longitudinal data on births and deaths obtained by researcher for district of Quinua from municipal records, with total popu-lation derived from national census for closest years.
Spector: Rate based on civil registry birth and death records and national census data reviewed by various researchers from years 1950through 1969 for Nunoa district. Mean for period presented, with range of 83 to 237.
Carey: Mortality rate generated from longitudinal data on births and deaths obtained by research team for district of Nun oa from mu-
nicipal records, with total population derived from national census. Mean for period presented, with range of 104 to 186.Larme: Infant mortality was calculated by researcher from Cuyo Cuyo birth and death records, Ministry of Health epidemiologic data in
Puno, and review of local health post records. Average for 19811986 period presented.Cruz-Coke: Demographic surveys of communities' inhabitants (sampling method not specied) and review of civil registry records by
researchers and health team.US Department of Commerce: Large-scale, systematically sampled government survey.
Table 3. Crude death rates for high altitude andean communities per 1000 population
Site Rate Location Year Altitude Source
Chugurpampa 6 North Peru 1988 3300 Oths (1991)Vicos 15 Central Peru 1952 2800 Alers (1965)Vicos 25 Central Peru 1963 2800 Alers (1965)Quinua district 18 South Peru 1960 3250 Mitchell (1991)
Quinua district 15 South Peru 1980 3250 Mitchell (1991)Nun oa district 20 South Peru 195069 4000 Spector (1971)Nun oa district 15 South Peru 197584 4000 Carey (1988)Cuyo Cuyo 27 South Peru 1981 3600 Larme (1993)
Belen district21 North Chile 1965 3200 Cruz-Coke et al.
(1966)
Lauca district50 North Chile 1965 4300 Cruz-Coke et al.
(1966)
Ecuador 8 national 1987 all USDC (1987)Peru 9 national 1987 all USDC (1987)Bolivia 14 national 1987 all USDC (1987)Chile 6 national 1987 all USDC (1987)
Oths: See Table 2. Sex and age adjusted rate does not dier from crude rate.Alers: See Table 2.Mitchell: See Table 2.
Spector: See Table 2. Rate is sex and age adjusted. Mean for period presented, with range of 14 to 33.Carey: See Table 2. Rate is sex and age adjusted. Mean for period presented, with range of 11 to 18.Larme: See Table 2.Cruz-Coke: Rate for each community based on data extracted from civil registry records by researchers and assisting health team.
K. S. Oths1020
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
5/14
Nun oa. This acknowledged, I now turn my atten-tion towards other factors of anthropological inter-
est that might be systematically generating the
observed health status dierences between commu-
nities.
The model depicted in Fig. 1 is divided into 2
major parts, microrisk and macrorisk factors, fol-
lowing the distinction Brown (1987) makes between
micro- and macroparasites. Though the distinction
is somewhat articial, it helps to dierentiate
between stresses produced at the local level and
those produced in exogenous arenas of power. Of
course, there can be interactions between any two
or more factors regardless of their levels. The
model takes into account aggregate-level factorsthat can distinguish between one community and
another. These factors are hypothesized to aect
community health either directly or indirectly.
The microrisk portion of the model is subdivided
into ecological considerations and socio-cultural
factors. Ecological risks result from environmental
factors that, through their action on the human
organism, produce physiological conditions that are
potentially stressful. Andean zones are customarily
demarcated longitudinally as being north, central or
south and latitudinally as the east or west escarp-
ment or central puna (high plains). Each zone has
its own general climatic and cultural characteristics,
with further distinctions within each zone (micro-niches and ethnicities). There are two ways of
thinking about altitude. One is in absolute meters
above sea level, which gives an indication of oxygen
pressure and UV radiation. Another is the altitude
of a community relative to other communities sur-
rounding it. Climate, measured in terms of humid-
ity, temperature and wind, may vary diurnally,
seasonally and by ecological microniche between, as
well as within, places in the Andes. UV radiation
increases directly with altitude and well as nearness
to the Equator. Barometric pressure drops with
increasing altitude; the lower the pressure, the
higher the risk of mountain sickness and other
severe physiologic problems characterized byedema, fatigue and dizziness.
The sociocultural microrisk factors merit some
explanation as well. Settlement patterns when dis-
persed tend to be more sanitary, whereas nucleated
villages tend to put dwellers in close contact with
human and animal waste. While the general consen-
sus is that health problems in a family increase with
the density of household occupation, the other
extreme, living alone, also puts one at high risk
in a labor intensive agriculture-based economy.
Demographic distributions, such as age (for depen-
dency ratio), sex (for gender ratio) and education
levels should also be considered. The primary mode
Table 5. Adult workdays lost due to illness by Andean community
Site
Chugurpampa Nun oa district
Sincata Nun oa town Chillihua
N 90 18 47 18Mean (SD) 0.44 (0.54) 3.22 (4.52) 2.05 (3.99) 3.25 (3.90)Year 1988 1984Source Oths Carey (1990)
Oths: See Table 1. My workdays lost are comparable to those of Carey's third phase of research in that our denitions are similar andour data is gathered from the same late year season (July to mid-December and August through October, respectively). As little vari-ation exists in workdays lost across the study period, I averaged my data collected over 10 two-week recall periods so as to be consist-ent with Carey's two-week recall period.
Carey: See Table 4. Sample sizes, as they are for the 3rd phase of Carey's research, exhibit some loss to follow up. Data obtained, how-ever, are comparable to the initial phase, in which the three household sample sizes were 90% (of 117) 8% (of 483) and 100% (of127), respectively.
Statistical signicance (2-tailed): Chug-Sincata p < 0.001, Ch ug-Nun oa p < 0.001, Chug-Chillihua p < 0.001.
Table 4. Maternal reproductive loss among Andean communities
Site
Cuyo Cuyo Nun oa district
Chugurpampa Ura Ayllu Puna Aylllu Sincata Nun oa town Chillihua
N (individuals) 180 10 10 21 47 18Mean (SD) 18.5% (19.4) 30.1% (32.1) 35.4% (17.3) 20.9% (19.5) 20.9% (24.8) 22.8% (27.3)Women 40+ 26.5% 36.3% 46.5% NAWomen
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
6/14
of production in Andean settlements is usually one
or a combination of the following: agriculture,herding, wage labor and mining. Each brings its
own host of related health problems, the latter two
modes more likely to produce the worst health
rates. Wealth is a variable that may also be viewed
in absolute or relative terms, depending on whether
the comparison is with other local communities,
other peasant regions or the nation as a whole. The
degree of equity in distribution of the land base is a
critical factor determining the health of peasant
communities. Community cohesion may be measured
by, among other things, the degree of stability of
residence and the size of kinship networks. Some
communities have abandoned traditional work
arrangements based on social networks, such as
labor exchange, while in others these arrangements
are quite viable. Social stratication varies from one
Andean community to the next. For instance, a
higher ratio of mestizos to indigenous people is
often related to inequalities for the latter, such as
restricted access to health care (see Crandon, 1983).
Wide variation in gender relations is reported
throughout the Andes, with male dominance and/or
preference found in some places while not in others.
Finally, certain regions possess greater traditional
medical resources than others, thereby increasing
options with which to maintain or restore highlan-
ders' health.On the macrorisk side, political-economic pro-
cesses and historical antecedents may also have a
bearing on modern day health status. One political-
economic dimension, market participation, refers tothe extent to which peasants production is oriented
to the satisfaction of needs exogenous to the local
community, in contrast to production for subsis-
tence. The more a community participates in the
former, the more economically vulnerable it is to
the vagaries of world markets. A corollary of mar-
ket participation is market penetration, or the
quantity of externally produced goods purchased by
peasants. These include foods, pesticides, clothing,
medicines and luxury items, among other things
goods which may increase dependency or lead to
unwise investment of limited cash resources. Labor
relations refers to the work arrangements between
kin groups; producer, middlemen and buyer; or
employer and employee. The degree of exploitative
relations, e.g. sharecropping, in a community is a
good indicator of its degree of social stability and
stratication. The political stability in a region
depends upon the integrity of local political insti-
tutions and the integration of the area into national
politics, as well as the amount of competition for
power in an area (e.g. guerrilla insurgency). The
availability (in terms of distance and cost) of state-
provided and private biomedical services may aect
morbidity and mortality levels, especially for serious
conditions such as acute respiratory illness.
The degree of historical as well as present daySpanish inuence in an area can determine among
other things its economy, language, political organ-
Fig. 1. Model for systematizing diversity among Andean communities.
K. S. Oths1022
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
7/14
ization, religious inuence and other customs. Land
reform has not been carried out evenly throughout
the Andes. As a result several types of community
organization currently exist in the Andes, such asthe hacienda, free peasant community, private prop-
erty, state-run cooperative and mining camp.
Hacienda workers and miners characteristically
exhibit the worst health, while cooperative members
(Leatherman, 1994) appear to fare the best.
Agricultural communities that have not benetted
from land reform are also at increased risk for ill
health (Leatherman, 1994; Luerssen, 1994).
In the center of the model are listed some key
risk mediators of special interest to international
health workers which may modify health outcomes.
These modiers, including water quality, sanitation,
diet and medical treatment, result from the conu-ence of the antecedent micro- and macrorisk fac-
tors. Diet, for instance, might be poor at the
community level under the following conditions:
land is scarce and distributed poorly, men migrate
to the coast for wage labor, and women are left to
shoulder typical men's roles, such as plowing.
Health outcomes, then, can be viewed as a pro-
duct of, or dependent on, the micro and macrorisk
factors and key risk mediators which obtain in any
given community. While this model cannot yet be
tested completely with data from Andean commu-
nities, it may prove helpful in framing appropriate
questions to ask in the search for explanationsabout health diversity.
Initial comparisons across communities could be
made eciently with an ecological model; besides
making comparative research easier, the model
could help national and international health
agencies identify those communities with the great-
est need for health intervention programs. Later,
large-scale research could t a model for multi-level
analyses using data from both a large number of
communities as well as a random sample of individ-
uals within each of those communities. Such a con-
textual model could then look at the eect of an
aggregate measure, say community cohesion, on thehealth of individuals, of the entire community or
the interaction between variables at both levels
(Sampson, 1991).
APPLYING THE MODEL TO A COMMUNITY EXAMPLE
Infant mortality
The utility of this model can be illustrated usingthe indicators of health status already presented.
For instance, one health outcome that can be exam-
ined with this model is infant mortality and some
of its known risk mediators. Infant mortality is a
sensitive measure of the overall health of the com-
munity. In Chugurpampa, potable water is available
from capped and free owing springs, in contrast to
that from rivers, gutters and irrigation ditches
reported in many places. Ecologically,
Chugurpampa is located on a hilltop at an altitude
high not only in absolute terms but relative to
neighboring villages, i.e. they are not downstream
from anyone. Politically, community leaders' asser-
tiveness has aided in getting PROJECT CARE as-
sistance to cap their reservoir and provide piping to
part of the community. Community cohesion has
been essential to initially build and then maintain
the reservoir and piped water system through the
collection of user fees and communal labor.
Infant diarrhea, often the result of fecal contami-
nation, is a leading cause of infant mortality
worldwide. Adequate sanitary conditions in
Chugurpampa follow from its dispersed and low
density community settlement pattern, which allows
for elimination and garbage disposal at a distance
from the home. This contrasts sharply with that of
nucleated, low density villages or nucleated highdensity towns where, for privacy, elimination
usually takes place near or within the house com-
pound, increasing chances of human contamination.
Childhood infectious diseases often lead to early
mortality. With few exceptions, low rates of infec-
tious bacterial and parasitic diseases are reported
throughout the high Andes. Buck et al. (1968) meti-
culously demonstrated this with their Andean
research of nearly 30 years ago. This is due in part
to UV radiation, cold temperatures and low air
pressure which inhibit the growth of infectious dis-
ease agents (Heath and Williams, 1989; Ward,
1989).
Immunization coverage of children ve and
under in Chugurpampa reached 60% complete cov-
erage for measles and polio and DPT series (an ad-
ditional 16% lacked only one of the total required
doses)*. Peruvian national averages for full cover-
age for the same time period were 43% for urban
areas and 10% for rural areas (Instituto Nacional
de Estadistica, 1986, p. 64). The hamlet relies heav-
ily on the active representation and commitment by
the Ministry of Health, the international aid project
and community leaders all of whom organize and
promote the vaccination campaigns at the commu-
nity level.
In reference to diet, Chugurpampans, in a fertilenorthern zone, appear to produce and consume
adequate calories to sustain healthy life. Children's
*For several weeks before beginning eldwork inChugurpampa, I assisted with the vaccination drivethroughout the district, visiting dozens of communities.This allowed me to familiarize myself with local vacci-nation routines and record keeping procedures. TheChugurpampa rate is drawn from my review of chil-dren's vaccination cards as part of the census. Thesecards are well-guarded by parents, as school admissionis dependent upon proof of full coverage. Incidentally,local health ocials overestimated the actual rate of
coverage by about 20%. For more detail on vaccinesgiven and the percent covered at each year of age, seeOths (1991, p. 187).
Assessing variation in health status in the Andes 1023
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
8/14
height and weight, infant growth curves, daily ob-
servation of household consumption patterns and
the highlander's own perceptions (see vignette
below) corroborate this. The children ve andunder in my study sample (N= 39) compared
favorably with the Peruvian National averages for
the northern rural sierra, with 40% of
Chugurpampans height-for-age less than 2 standard
deviations (SD) below the median compared to
67% for the national sample and 24% of their
weight-for-age below 2SDs compared to 29% over-
all (Instituto Nacional de Estadistica, 1986). Infant
growth curves, while remaining in the lowest per-
centiles, showed steady progress for all 8 children
under one year old in the study*. The quality of
diet cannot be divorced from a general discussion
of labor relations, access to land and fair marketprices (Leonard et al., 1990). Leonard et al. (1990)
note that the stature of Nun oan children is among
the lowest in the Andes, which they attribute to
poor nutrition, again conrming the notion of inter-
regional diversity. For a Nun oan town, peasant
community and cooperative alike, Leatherman
(1994) found rates of at-risk children based on
height-for-age that ranged from 54 to 73%, all
exceeding Chugurpampa's rate{.
Furthermore, environmentally, in the Northern
Peruvian Andes the cold temperature, while con-
stant, is less severe than in the Central and
Southern portions. In Chugurpampa, diurnal tem-
perature uctuations are less drastic, frost is rare
and snow unknown. Frigid temperatures in other
parts of the Andes are associated with higher infant
mortality rates.
Cultural practices which can contribute to high
infant mortality are found in some areas of the
Andes. de Meer et al. (1993) in Peru and McKee
(1984) and Scrimshaw (1978) in Ecuador, found
male gender preference, early weaning of females
and female infanticide. There was no evidence of
these practices in Chugurpampa. While Scrimshaw
reports a sex-ratio imbalance of 68 males to 32
females for the rst child born (Scrimshaw, 1978, p.
389), implying infant neglect of rst-born females,rst-born females slightly outnumbered males in
Chugurpampa.
Loss of function and morbidity
Measures of function and morbidity are more
subjective than mortality rates. Still, striking dier-
ences occur in the total number of workdays lost
between Chugurpampa and Nun oa. Here, not
simply morbidity, but the incentive to work or not
work must be considered in arriving at an adequate
explanation. Are households in a particular commu-
nity employed in wage labor, in a cooperative, or
independently in cooperation within kin networks?
What is the marginal utility of a day of work for a
peasant who is sick and might otherwise be recuper-
ating? Labor relations and land ownership are im-
portant here. In Carey's study, the peasant villagers
of Sincata experienced 40% more symptoms on
average than in Chillihua, a cooperative, yet their
``workdays missed'' relative to their symptoms wasnearly 1/3 lower. In Sincata, as in Chugurpampa,
independent peasants receive no secondary gains
from missing work. Every day missed just means
that much more work to do when they return
(Leatherman (1992) concurs on this point).
Women's health appears to be quite directly re-
lated to the parity of gender roles, which in turn
are highly conditioned by the mode of production.
In Cuyo Cuyo, where women were left alone to run
the household due to men's migration to the mines,
they suered a noticeable health burden (Larme,
1993; Luerssen, 1994). Inevitably, the health of the
entire family suers when the head female of thehousehold suers.
The drier, desert conditions of the northern
Peruvian Andes may be implicated in the high num-
ber of respiratory conditions in Chugurpampa (see
Table 1). The aridity and low absolute humidity in
this region cause rapid dehydration in humans,
which contributes to respiratory and g.i. ailments.
Ventilation and sweating with exertion compounds
the eect (Heath and Williams, 1989; Ward, 1989).
Diurnal temperature uctuations contribute to res-
piratory, musculoskeletal and g.i. illness (Ward,
1989). In Chugurpampa, highs average around 708F
and lows around freezing.
In Chugurpampa, the association between muscu-loskeletal problems, economic strategy and labor re-
lations is clear. It is characteristic of peasant
populations that food production be directed by the
politically dominant, coastal elite. In the northern
Peruvian Andes, agricultural production of potatoes
for the market nds the highlanders caught in a
vicious cycle of spiraling costs of petrochemical pro-
ducts and deating market prices, thus necessitating
ever-increasing production to stay ahead (also see
Mitchell, 1991){.
Potato agriculture requires constant stooping and
the lifting, carrying and swinging of sharp tools and
heavy loads, usually 100150 lb at a time. Duringharvest in Chugurpampa, for example, old and
young, men, women and children alike, may typi-
*WHO (1983) measurements of medians and standard de-viations were used for assessing height and weight.One infant's growth curve was at for 1 month due toillness, but recouped the growth in the followingmonths. Some of the repeat measures of infant datawere lost as a consequence of political unrest.Therefore, my recall does not permit more precisedetails.
{The percentage of children 2SDs below the medianfor height- and weight-for-age in the southern ruralsierra are 67 and 22%, respectively (Instituto Nacional
de Estadistica, 1986).{Ination on government agricultural loans to peasants
was 70% for 1988.
K. S. Oths1024
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
9/14
cally transport up to one ton of sacked potatoes per
day. The result of such arduous labor with little
respite under harsh climatic conditions is a high
incidence of physical degenerative illness such asmusculoskeletal aches, pains, strains, injuries,
lacerations, arthritis and kidney problems.
What cannot yet be determined, but could be
articulated by using the proposed model of diver-
sity, is the extent to which musculoskeletal com-
plaints are associated with (1) the degree of
exploitative labor relations found in dierent com-
munities and (2) the degree to which the eects of
exploitation are modied by other resources within
the community.
Historically speaking, the CajamarcaLa
LibertadSan Martin corridor in which
Chugurpampa resides was more heavily occupiedby the early colonizers than other areas of the
Andes. This early Spanish inuence has resulted in
the virtual loss of the Quechua language, traditional
dress, and native species of camelid, a point which
Brush (1977) corroborates*. Thus, lacking llamas
and with burros being expensive, people have
become accustomed to carrying heavy weight on
their own backs, with obvious musculoskeletal
health eects. Chugurpampans own fewer pack ani-
mals than peasants in certain other Andean areas:
with 14% of households owning horses or mules
(0.2 per household) and 56% owning burros (1.0
per household), Chugurpampa fares more favorably
than Vicos (Alers, 1965, p. 445) or Ura Ayllu(Bullard, 1990), but less so than Quinoa (Mitchell,
1991, p. 60) or Puna Ayllu (Bullard, 1990). Data of
Figueroa (1982, p. 135) show the great variability in
animal ownership among eight communities in the
southern sierra of Peru.
The heavy use of pesticides should not be over-
looked as another potential agriculture-related con-
tributor to the categories of respiratory (through
inhalation) and gastrointestinal (through consump-
tion) illness (see Bull, 1982, p. 37). Pesticides are
commonly used, despite their high cost, due to their
promotion by agricultural agents and the eventual
dependency on the product after several years ofuse (Mitchell, 1991, p. 108, 212; Hamilton, 1994).
Additionally, musculoskeletal and respiratory
problems are a near inevitable outcome of mine
labor. Nash (1979) among others has documented
the abysmal conditions of industrial mining in the
Andes. The amount of mining activity in each com-
munity correlates with musculoskeletal illness.
Referencing Table 1 again, Cuyo Cuyo has the
highest percentage of musculoskeletal complaints at
28%, Chugurpampa is next at 18% and Saraguro
has only 10%. The communities of Cuyo Cuyo
viewed their principal work to be migrant gold
mining and, probably not coincidentally, had corre-spondingly poorer health status than most commu-
nities. Larme showed this as due not only to the
physical cost of mining, but also to the social disor-
ganization and increased workload on women pro-
duced by the absence of males from the
community. The four ex-silver miners in the
Chugurpampa sample (11% of adult men) were
also the sickest persons. Finerman relates no mining
activity in Saraguro.
The intra-Andean diversity of general community
level factors such as social stratication and wealth
can also be examined. It might be instructive not
only to gauge the overall degree of social stratica-
tion within a community, and the relative wealth of
households, but also the wealth of one community
vis-a-vis another. Ethnicity and wealth may be
highly correlated within a region but not always
across regions, as Crandon-Malamud (1991) has
shown for the highland Bolivian mestizos and
Aymara after the 1952 revolution. Although in my
unstratied community of indigenous peasants
(there were no mestizos) there was dierentiation in
socio-economic standing, it appears from the sket-
chy absolute indicators of wealth available (such as
hectares of land, number of animals, cash income,
etc.) that the poor in Chugurpampa are less poor
than those of other regions. For instance, a mini-mum of 3.5 hectares is considered necessary for ade-
quate household production in the highlands (Deere
and de Janvry, 1979, p. 604). The mean of 5.8 hec-
tares of arable land per Chugurpampan household
exceeds other published gures for peasant commu-
nities: 4.5 in Cajamarca (Deere and de Janvry,
1979), 2.9 in the Latacunga-Ambato region of
Ecuador (Hamilton, 1994), 1.6 in Uchumarca, San
Martin (Brush, 1977, p. 86), 0.8 in Soqa, Puno
(Lewellen, 1978, p. 187), no more than 0.5 in
Quinoa, Ayacucho (Mitchell, 1991, p. 56) and from
0.4 to 1.8 among various communities in the district
of Moho, Puno (Collins, 1988, p. 100). This fact, of
course, is tied directly to several micro- and macro-
risk factors, such as land reform, soil depletion,
mode of subsistence, labor relations, market entry,
etc. Regarding landlessness, in Chugurpampa, while
16% held land rights to 1 (or, equally, 0.5) hectare
or less, only 3% were without access to land by
other means (tenancy, rental, etc.). This degree of
landlessness is similar to what Deere and de Janvry
(1979) has recorded for Cajamarca, where 13%
have 0.25 or fewer hectares and much less than
cited for other communities in southern Peru: 29
and 54% have less than 1 hectare in the Tambopata
Valley, Puno, and Andarapa, Ayacucho, respect-
ively (Collins, 1988, p. 161; Sanchez, 1982, p. 163)and 59% have less than 0.5 in Soqa, Puno
(Lewellen, 1978, p. 42). The relationship within a
*Alers (1965) notes that in 1951, of those persons seven orolder in Vicos, Ancash, the department that bordersLa Libertad to the south, 98% spoke no Spanish. No
Chugurpampans could recollect anyone ever havingspoken Quechua, though remnant words remain intheir Spanish.
Assessing variation in health status in the Andes 1025
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
10/14
community between lower wealth, based on indi-
cators such as land, and poorer health, as measured
by nutritional status and workdays lost, has been
quantied by Leonard (1989) and Luerssen (1994).While this discussion has been speculative, the
observed health dierences among Andean commu-
nities appear to be generated out of the unique con-
uence of ecological factors, local meaning and
political-economic and historical processes which
combine to shape systematic intra-Andean diversity.
Because of the tremendous diversity that exists in
the Andes, it is clear that health research must rest
upon a solid ethnographic foundation.
IMPLICATIONS FOR HEALTH PROGRAMS
How does the foregoing apply to national and in-ternational health programs? I relate the following
anecdote to illustrate the necessity of understanding
the extent to which there is local variation in health
and illness throughout the Andean regions:
The setting is Chugurpampa, a large peasant
community in the northern highlands of Peru, at a
meeting of the newly formed Club de Madres
(mother's club). A festive luncheon awaits the repre-
sentatives of the internationally funded health and
nutrition project that sponsors the club: the fare
consists of soup, potatoes, beans, guinea pig, corn,
hot pepper sauce, salad and beverages. The under-
lying emphasis of the health project is child survi-
val; the goal is to eliminate infant mortality and
childhood malnutrition by re-educating mothers.
The rst comments to be heard from the arriving
nurses and nutritionists are ``me muero de hambre''
(I'm dying of hunger). After the lling lunch, the
mothers are invited to gather in a vacant classroom
to listen to lectures on nutrition, infant feeding and
the basic food groups; in short, to teach Andean
mothers how to cook and feed their families. Later
that week, several women express that they are
quitting the Club de Madres.
Applied health care personnel are not interested
in nor capable of solving the root causes of exploi-
tation and hunger, such as an inequitable distri-bution of land or exploitative market prices. But
they need to know the type and extent of actual
health problems that exist before they can begin to
redress them. It would be benecial for inter-
national health agents to perform a rapid assess-
ment and scoring of the micro- and macrorisk
factors of several potential recipient communities in
the area in which they wish to work. In addition to
identifying the greatest perceived health problems
of a community, the model aords a reasonable in-
dication of the general health of the community,
which could help them target those communitieswith the greatest health decits.
Scrimshaw and Hurtado (1987) have pioneered
the rapid assessment procedure and others have
adapted the idea to specic illnesses or ethnographic
uses. A rst attempt to develop a usable instrument,
based on the foregoing evidence, appears in Table 6.
While the checklist is by no means all-inclusive and
will hopefully undergo renements if eld testing
proves it useful, the operationalization of many key
community factors contributing to health that it
provides is designed to be applicable by health and
social science professionals. The rst section can be
completed with a brief visit to the community, with
the second part (or any portion thereof) obtainable
within a 1 to 2 day stay in the community. For an
even quicker assessment, in lieu of assigning absol-
ute values on items, an alternative coding scheme
would be to rank all villages under consideration
for each item; again one would represent the best
and the highest number the worst, condition.
Disregarding the diversity in the Andes and fail-
ing to perceive the relevant local cultural, material
and environmental conditions can lead to inap-
propriate interventions, the likes of which I wit-
nessed during my stay in Chugurpampa. Now, the
problem of program inappropriateness has plagued
international aid eorts since their inception (Paul,1955). This Andean case illustrates just one of the
``generic problems'' Jordan (1993, p. 175) identied
that can cause health intervention eorts to fail: the
inappropriateness of contentin this case writ
large*. My critique regards the assumption that a
model paradigm of child survival, popular since the
1980's, is inherently applicable everywhere there are
poor people. With this paradigm, topics such as
infant diarrhea, oral rehydration therapy, infant
mortality, nutrition and breast-feeding receive
major attention regardless of actual local conditions
(for review, see WHO, 1986; Scheper-Hughes, 1987;
UNICEF, 1989; Mosley et al., 1990; Nightingale et
al., 1990; Nichter and Kendall, 1991).
Health care workers made assumptions about
Chugurpampa's overall health based on national
and regional statistics, gearing their eorts toward
children and ignoring adult morbidity and mor-
tality. Unfortunately, this led to the failure of their
program, depicted in the vignette. They deigned to
teach their mother-hosts how to cook for and rear
their already healthy hamlet children. Interventions
were geared toward particular illnesses (diarrhea,
malnutrition) that were assumed to be causing high
infant and childhood mortality. That was not the
case. The result was much eort wasted on introdu-
cing well-meaning programs that were meaning-lessto the people they were intended to serve. There
were two other such examples of child survival
*Jordan also notes the ``multiple agendas'' of aid pro-grams which can include the political aim of imposingpower and authority on a marginalized people throughthe guise of superior knowledge (Jordan, 1993, p. 169).
While this paper seeks to make a dierent point, herobservation certainly applies to the Andean context aswell.
K. S. Oths1026
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
11/14
Table 6. Checklist for rapid assessment of relative need for intervention in Andean peasant communities. For rating two or more villages.The highest score indicates the greatest potential for health problems. Community name: ( F F F)
Table 6continued overleaf
Assessing variation in health status in the Andes 1027
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
12/14
strategies the agency employed concurrently with
the Mother's Club in Chugurpampa: a Community
Garden (Huerto Comunal) and a Height and
Weight Monitoring of Children (Control de Peso de
Nin os Infantiles). I will briey detail the case of the
garden.
On the instigation of the international aid agency
working in Chugurpampa, a community garden was
planted near the school to provide for hot school
lunches. Mothers were to come on a rotating basis
to prepare the food. The implication was that chil-dren were malnourished in their own home and
thus inattentive during the school day. Ironically,
community mothers ended up supplementing the
meager garden produce from their own larders.
A common belief of the international aid commu-
nity, that children's inattentiveness in primary
school is a result of poor nutrition at home, served
as the rationale for this project. On the contrary, by
the reckoning of Chugurpampa mothers, it is pre-
cisely a young child's presence in school that can
cause sickliness, debility and weight loss. Day-long
school attendance curtails the young ones' habit of
constant snacking between meals to fulll high-alti-tude high-carbohydrate energy needs (Ward, 1989).
Consistent with the mother's belief, the school aged
children in my sample accounted for nearly two-
thirds of those who exhibited nutritional stress.
The aid programs met with reactions ranging
from mild amusement or consternation to suspicion,
defensiveness and anger. Locals simply viewed the
program aims as irrelevant. Instead of a participa-
tory spirit, a distrust of aid workers was instilled in
the community members, certain to disadvantage
future intervention eorts.
Involvement of Chugurpampans in the identi-
cation and denition of health problems which needto be addressed would result in a list of problems
very distinct from that produced by the Ministry of
Health and foreign aid agencies. People really want
advice and therapy for their most prevalent and
intractable problems such as respiratory, musculos-
keletal, stomach and dental ailments (see Oths,
1996). Infant and childhood illnesses certainly are
not isolated as the only, or even the major, health
risks which face these highlanders on a daily basis.
It should be noted that in other contexts, there
might be dierent but equally compelling reasons
that the poor resist well-intended child health pro-
grams. Elsewhere, where high morbidity and mor-tality are viewed as `normal', the ` apathetic
response to programs attempting to improve the
6Use WHO (1983) standards.
Table 6continued
K. S. Oths1028
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
13/14
health of children may have a valid underlying (but
unconscious) purpose'' as a strategy to control
family size where resources are scarce (Scrimshaw,
1978, p. 393).
CONCLUSION
The international health community alone should
not be faulted for assuming a homogeneity of
health problems within developing countries. Social
science in general, and anthropology in particular,
has not paid enough attention to intracultural or
intrasocietal diversity. The explanatory sketch pre-
sented in this paper has been intended to suggest
the importance of going beyond models that, at
best, examine intracommunity dierences to the
development of models that incorporate intercom-munity dierences as well.
In epidemiology, there is a perspective developing
for ``contextual'' or ``multi-level'' models (Von
Kor et al., 1992). Here, in essence, the community
has been the unit of analysis. Because I have largely
worked from published sources I have not been
able to examine how intracommunity dierences in,
for example, wealth might have dierent impli-
cations for health status depending on the overall
level of wealth of that community (i.e. what the
health implications are of being poor in a poor
community versus poor in a wealthy community).
Multi-level analyses can address these issues. The
model and rapid assessment checklist are oered asan on-the-ground tool for health care deliverers. A
systematic program of research designed to examine
diversity in health status in the Andes, using a
multi-level methodological strategy, would help to
answer many of the questions raised in this paper,
and to point the way to the more rational allo-
cation of helping resources. Ultimately, it is up to
us, as social scientists, to appreciate, describe and
explain the diversity in health found within the
Andean landscape.
AcknowledgementsThis work was funded by grants from
the National Science Foundation (No. 8813774) and theInter-American Foundation (No. F1-135-A1). I am grate-ful to William W. Dressler and several anonymousreviewers for invaluable comments on earlier drafts. Also,I wish to acknowledge the aid and support of my eld as-sistants, Amable Burgos and Genaro Aguilar and theTrujillo-based PRAS project, without whom this workwould not have been possible.
REFERENCES
Alers, J. O. (1965) Population and development in aPeruvian community. Journal of Inter-American Studies7, 423448.
Bastien, J. (1987) Healers of the Andes: KallawayaHerbalists and Their Medicinal Plants. University of
Utah Press, Salt Lake City.Brown, P. J. (1987) Microparasites and macroparasites.
Cultural Anthropology 2(1), 155171.
Brush, S. B. (1977) Mountain, Field, and Family: TheEconomy and Human Ecology of an Andean Valley.University of Pennsylvania Press, Philadelphia.
Buck, A. A., Sasaki, T. T. and Anderson, R.I. (1968)
Health and Disease in Four Peruvian Villages. JohnsHopkins Press, Baltimore, Maryland.
Bull, D. (1982) A Growing Problem: Pesticides and theThird World Poor. OXFAM, Oxford.
Bullard, J. R. (1990) Value and economic culture amongthe peasant gold miners of the Cuyo Cuyo district(Northern Puno, Peru). Working Paper No. 3.Production Storage and Exchange Project. Departmentof Anthropology, University of North Carolina, ChapelHill, NC.
Carey, J. (1988) Health, social support and social net-works in a rural Andean community of Southern Peru.Ph.D. dissertation, University of Massachusetts,Amherst.
Carey, J. (1990) Social system eects on local level mor-bidity and adaptation in the rural Peruvian Andes.
Medical Anthropology Quarterly 4(3), 266295.Collins, J. L. (1988) Unseasonal Migrations: The Eects of
Rural Labor Scarcity in Peru. Princeton UniversityPress, Princeton, NJ.
Crandon, L. (1983) Grass roots, herbs, promotors andpreventions: a re-evaluation of contemporary inter-national health care planning. The Bolivian case. SocialScience and Medicine 17, 12811289.
Crandon-Malamud, L. (1991) From the Fat of Our Souls:Social Change, Political Process, and Medical Pluralism
in Bolivia. University of California Press, Berkeley.
Cruz-Coke, R., Cristoanini, A. P., Aspillaga, M. andBiancani, F. (1966) Evolutionary forces in human popu-lations in an environmental gradient in Arica, Chile.Human Biology 8, 421438.
Deere, C. D. and de Janvry, A. (1979) A conceptual
framework for the empirical analysis of peasants.American Journal of Agricultural Economics 61, 601611.
de Meer, K., Bergman, R. and Kusner, J. S. (1993) Socio-cultural determinants of child mortality in southernPeru: Including some methodological considerations.Social Science and Medicine 36(3), 317332.
Donahue, J. (1981) Conclusion. In Health in the Andes,eds. J. W. Bastien and J. M. Donahue, pp. 224230.American Anthropological Association, Washington,D.C.
Dutt, J. S. and Baker, P. T. (1978) Environment, mi-gration and health in southern Peru. Social Science andMedicine 12, 2938.
Figueroa, A. (1982) Production and market exchange inpeasant economies: the case of the southern highlands
in Peru. In Ecology and Exchange in the Andes, ed.David Lehmann, pp. 123156. Cambridge UniversityPress, Cambridge.
Finerman, R. D. (1985) Health care decisions in anAndean indian community: getting the best of bothworlds. Ph.D. dissertation, University of California, LosAngeles.
Frerichs, R. R., Becht, J. N. and Foxman, B. (1980)Prevalence and cost of illness episodes in rural Bolivia.International Journal of Epidemiology 9(3), 233238.
Go mez, L. C. (1988) Health status of the Peruvian popu-lation. In Health Care in Peru: Resources and Policy, ed.D. K. Zschock, pp. 1552. Westview Press, Boulder.
Hamilton, S. (1994) Gender and agrochemicals: Linkingproductivity, environment and health in an Andean eco-system. Paper presented at the American Anthropological
Association Annual Meetings. Atlanta, December 3.Heath, D. and Williams, D. (1989) High-Altitude Medicine
and Pathology. Butterworth and Company, London.
Assessing variation in health status in the Andes 1029
8/2/2019 Oths Assessing Variation in Health in the Andes- A Bio Cultural Model (SSM, 1998) Oths
14/14
Instituto Nacional de Estadistica (1986) Encuesta Nacionalde Nutricion y Salud 1984. Ministerio de Salud, Lima,Peru.
Jordan, B. (1993) Birth in Four Cultures, 4th edn.
Waveland Press, Prospect Heights, Illinois.Larme, A. (1993) Work, reproduction and health in two
Andean communities. Working Paper No. 5. Production,Storage and Exchange Project. Department ofAnthropology, University of North Carolina, ChapelHill, NC.
Leatherman, T. L. (1992) Illness as lifestyle change.MASCA Research Papers in Science and Archaeology 9,8389.
Leatherman, T. L. (1994) Health implications of changingagrarian economies in the southern Andes. HumanOrganization 53, 371379.
Leatherman, T. L. and Thomas, R. B. (1987) Patterns ofillness and work disruption in a rural Andean popu-lation. Abstracts of the Annual Meeting, AmericanAssociation of Physical Anthropologists. American
Journal of Physical Anthropology 72, 223.Leonard, W. (1989) Nutritional determinants of high-alti-tude growth in Nun oa, Peru. American Journal ofPhysical Anthropology 80, 341352.
Leonard, W., Leatherman, T., Carey, J. and Thomas, R.B. (1990) Contributions of nutrition versus hypoxia togrowth in rural Andean populations. American Journalof Human Biology 2, 613626.
Lewellen, T. (1978) Peasants in Transition: The ChangingEconomy of the Peruvian Aymara: A General SystemsApproach. Westview Press, Boulder.
Little, M. A. and Baker, P. T. (1976) Environmental adap-tations and perspectives. In Man in the Andes, eds. P. T.Baker and M. A. Little, pp. 405428. Dowden,Hutchinson and Ross, Stroudsburg, Pennsylvania.
Luerssen, J. S. (1994) Landlessness, health and the failuresof reform in the Peruvian highlands. Human
Organization 53, 380387.McKee, L. (1984) Sex dierentials in survivorship and the
customary treatment of infants and children. MedicalAnthropology 8(2), 91108.
McKeown, T. (1976) The Modern Rise of Population.Academic Press, New York.
Merrick, T. W. (1986) Population pressure in LatinAmerica. Population Bulletin, Vol. 41, No. 3. PopulationReference Bureau, Washington, D.C.
Mitchell, W. P. (1991) Peasants on the Edge: Crop, Cultand Crisis in the Andes. University of Austin Press,Austin.
Mosley, H., Jamison, D. T. and Henderson, D. A. (1990)The health sector in developing countries: Problems forthe 1990s and beyond. Annual Review of Public Health11, 335358.
Murra, J. V. (1972) El control ``vertical'' de un ma ximo depisos ecolo gicos en la economia de las sociedadesAndinas. In Visita a la provincia de Leon de 1562, ed. J.V. Murra, Vol. 2, pp. 429476. Universidad NacionalHermilio Valdiza n, Hua nuco.
Murra, J. V. (1985) ``El archipie lago vertical'' revisited. InAndean Ecology and Civilization, eds. Masuda, Shozo,Izumi Shimada and Craig Morris, pp. 313. Universityof Tokyo Press, Japan.
Nash, J. (1979) We Eat the Mines and the Mines Eat Us.Columbia University Press, New York.
Nichter, M. and Kendall, C. (1991) Beyond child survival:Anthropology and international health in the 1990s.Medical Anthropology Quarterly 5(3), 195203.
Nightingale, E. O., Hamburg, D. A. and Mortimer, A. M.(1990) International scientic cooperation for maternaland child health. In Issues in Contemporary InternationalHealth, eds. T. A. Lambo and S. B. Day, pp. 113133.
Plenum Medical Book Company, New York.Oths, K. S. (1991) Medical treatment choice and health
outcomes in the northern Peruvian Andes. Ph.D. disser-tation, Case Western Reserve University, Cleveland.
Oths, K. S. (1996) Ecological and macrolevel inuences onillness in northern Peru: beyond the international healthparadigm. In Society, Health and Disease: TransculturalPerspectives, eds. Subedi, Janardan and EugeneGallagher, pp. 107129. Prentice Hall, Upper SaddleRiver, NJ.
Parchman, M. L. (1995) Small area variation analysis: Atool for primary care research. Family Medicine 27,272276.
Paul, B. D. (1955) Health, Culture, and Community: CaseStudies of Public Reactions to Health Programs. RussellSage Foundation, New York.
Pelto, P. J. and Pelto, G. H. (1975) Intracultural diversity:Some theoretical issues. American Ethnologist 2, 118.
Sampson, R. J. (1991) Linking the micro- and macroleveldimensions of community social organization. SocialForces 70, 4364.
Sanchez, R. (1982) The Andean economic system andcapitalism. In Ecology and Exchange in the Andes, ed.David Lehmann, pp. 157190. Cambridge UniversityPress, Cambridge.
Scheper-Hughes, N. (ed.) (1987) Child Survival:Anthropological Perspectives on the Treatment andMaltreatment of Children. Kluwer Academic Publishers,Dordrecht.
Scrimshaw, S. (1978) Infant mortality and behavior in theregulation of family size. Population and DevelopmentReview 4(3), 383403.
Scrimshaw, S. and Hurtado, E. (1987) Rapid AssessmentProcedures for Nutrition and Primary Health Care.UCLA Latin American Center Publications, LosAngeles.
Spector, R. M. (1971) Mortality characteristics of a highaltitude Peruvian population. M.A. thesis, PennsylvaniaState University, University Park, PA.
United Nations International Children's Emergency Fund(1989) The State of the World's Children. OxfordUniversity Press, New York.
United States Department of Commerce, Bureau of theCensus (1987) World Population Prole. U.S.Government Printing Oce, Washington, D.C.
Vayda, A. P. (1994) Actions, variations and change: Theemerging anti-essentialist view in anthropology. InAssessing Cultural Anthropology, ed. Robert Borofsky,
pp. 320330. McGraw-Hill, New York.Von Kor, M., Koepsell, T., Curry, S. and Diehr,
P. (1992) Multi-level analysis in epidemiologic researchon health behaviors and outcomes. American Journal ofEpidemiology 135, 10771082.
Ward, M. P. (1989) High-Altitude Medicine andPhysiology. University of Pennsylvania Press,Philadelphia.
World Health Organization (1983) Measuring Change inNutritional Status. WHO, Geneva.
World Health Organization (1986) Health ResearchStrategy for Health for All by the Year 2000. AdvisoryCommittee on Health Research, WHO, Geneva.
K. S. Oths1030