PUBLIC HEALTH BRIEFS
2. Davis JE: The future of major ambulatory surgery. Surg Clin North Am
3. Henderson JA: Surgery centers' success challenges hospitals. Modern
Healthcare 1989; 19(22):78-80.
4. Ambulatory care growth continues. Outreach 1988; 10(l):1-3.
5. Shannon K: Outpatient surgery up 77 percent: data. Hospitals 1985;
6. Davis JE: Preface. Surg Clin North Am 1987; 67:ix-viii.
7. Grimaldi PL, MichelettiJA: Prospective Payment: The Definitive Guide to
Reimbursement. Chicago: Pluribus Press, 1985; 232-234, 275-281.
8. Zeigenfuss JT: DRGs and Hospital Impact, An Organizational Systems
Analysis. New York: McGraw Hill, 1985; 109-111.
9. Leader S, Moon M: Medicare trends in ambulatory surgery. Health Affairs
10. Lagoe RJ, Milliren JW: A community-based analysis of ambulatory
surgery utilization. Am J Public Health 1986; 76:150-153.
11. Lagoe RJ, Bice SE, Abulencia PB: Ambulatory surgery utilization by age
level. Am J Public Health 1987; 77:33-37.
12. CentralNewYorkHealth SystemsAgency: Medical Facilities Plan 1986-1988:
Syracuse: Central New York Health Systems Agency, 1986; 73-84.
13. US Department of Health and Human Services: Health, United States,
1988. Hyattsville, MD: US Department of Health and Human Services,
1989; 1 18-119, 142-143.
14. Division of Health Care Statistics, National Center for Health Care
Statistics: Data from the National Hospital Discharge Survey, 1988.
Hip Fracture Incidence among the Old and Very Old:
A Population-Based Study of 745,435 Cases
STEVEN J. JACOBSEN, MS, PHD, JACK GOLDBERG, PHD, TONI P. MILES, MD, PHD, JACOB A. BRODY, MD,
WILLIAM STIERS, PHD, AND ALFRED A. RIMM, PHD
Abstract: Data were obtained from the Health Care Financing
Administration and the Department of Veterans Affairs (formerly
called Veterans Administration) on all hospital discharges among the
elderly population from 1984 through 1987 and combined with census
estimates to calculate incidence rates of hip fracture for the elderly
population of the United States. Rates for White women were the
highest, reaching 35.4 per 1,000 per year among 95 year-olds.
Comparably, White men, Black women, and Black men experienced
similar age-related increases in risk, although of less magnitude and
relatively less rate of change, respectively. (Am J Public Health
Hip fractures are an important cause of the morbidity
and mortality experienced by the elderly population in the
United States.' Despite this, there have been relatively few
national population-based studies of the incidence of hip
fracture in this country. Ofthose previous studies, most have
produced unstable estimates of incidence rates for non-
Whites (due to small sample sizes)2'3 or the oldest old (over
85 years of age).2-7 The present study uses newly available
datafrom the Health Care Financing Administration (HCFA)
and the Department of Veterans Affairs (VA) to provide
detailed estimates of the incidence of hip fracture in the
elderly by age, race, and sex.
Address reprint requests to Steven J. Jacobsen, PhD, Division of
Biostatistics/Clinical Epidemiology, Medical College of Wisconsin, P.O. Box
26509, Milwaukee, WI 53226. This work was done while Dr. Jacobsen
(currently a third-year medical student at Medical College of Wisconsin) was
completing the doctoral requirements in the Epidemiology Program, School of
Public Health, University of Illinois at Chicago. Drs. Goldberg and Miles are
both with the Epidemiology/Biostatistics Program there, and Dr. Brody is
Dean ofthe School ofPublic Health, University ofIllinois, Chicago; Dr. Stiers
is with Health Services Research and Development, VA Hospital, Hines, IL;
Dr. Rimm is Professor and Head, Division of Biostatistics/Clinical Epidemi-
ology, Medical College ofWisconsin, Milwaukee. This paper, submitted to the
Journal June 19, 1989, was revised and accepted for publication October 17,
© 1990 American Journal of Public Health 0090-0036/90$1.50
Data were obtained from HCFA on all 40 million
short-stay hospital discharges for the calendar years 1984
through 1987 and 4 million discharges from VA hospitals.
HCFA requires all institutional providers to submit a uniform
billing form that contains the following data: age, race, sex,
a maximum of five ICD9-CM diagnostic categories, date of
admission and discharge, and a unique patient identifier. VA
maintains a similar database, keyed by the same patient
Cases of hip fracture were identified by scanning the 44
million discharge records fora discharge diagnosis offracture
of the hip (ICD9-CM 820.0 through 820.9) in any of the five
diagnostic fields. Patients were excluded if: 1) age at admis-
sion was less than 65 years; 2) the fracture could have been
attributed to a neoplastic process; 3) the patient had been
discharged previously for fracture of the hip in the four-year
period; 4) the first discharge diagnosis was for late effects of
fracture; or 5) the patient resided outside of Puerto Rico or
the United States proper. The remaining cases were consid-
ered to be first fractures of the hip and thus comprised the
numerator for the calculation of rates.
Denominator information was obtained from the Bureau
ofthe Census for 1985. Incidence rates were calculated as the
average annual age-, race-, and sex-specific number ofcases
divided by the population at risk. Rates were adjusted for
one-year age groups using the direct method, with the entire
United States population aged 65 years and older serving as
During the period of 1984 through 1987, 810,949 hospital
discharges for fracture of the hip were recorded by HCFA
and 11,435 were recorded by VA for persons over 65 years of
age. Ofthese, 54,055 were second fractures or rehospitaliza-
tions for a previous hip fracture. An additional 19,899 cases
were persons under the age of 65 from the HCFA database,
and 4,504 cases were discharged for late effects or attribut-
able to neoplastic processes. The final number of eligible
incident cases of first hip fracture was 745,435. (Some cases
were excluded for more than one reason.) Women accounted
AJPH July 1990, Vol. 80, No. 7
PUBLIC HEALTH BRIEFS
for 580,129 of these fractures (79 percent) whereas men
accounted for only 165,306 (21 percent) (Table 1). The great
majority ofthe cases were among Whites (93 percent), while
Blacks and those ofother orunknown racial background each
accounted for 3 percent.
The incidence of hip fracture among the elderly for this
time period was 6.63 fractures per 1,000 per year. Adjusting
incidence rates for age, White women experienced the
highest incidence rates at 8.07 per 1,000, followed by White
men at 4.28/1,000, Blackwomen at 3.06/1,000, and Black men
at 2.38/1,000. Examining incidence rates by one year age
groups (Figure 1) shows that rates for White women in-
creased exponentially from 1.63/1,000 in 65 year-olds, to
35.4/1,000 in women aged 95 years. A similar exponential
pattern ofincrease was observed for White males, with rates
increasing from 0.9/1,000 for 65-year-olds to 26.0/1,000 for
96-year-olds. The age-associated increase in incidence per-
sisted in Blacks, but was less than exponential in women and
nearly linear in men.
The risk offracture ofthe hip increased with age well into
the tenth decade of life for women and men, and Blacks and
Whites alike. Incidence rates for White women between the
ages of 85 and 95 were especially high, with more than 3
percent of this population suffering a fracture of the hip.
White men, Black women, and Black men also suffered from
high rates of injury, with rates reaching nearly 3, 2, and 1
percent of the nonagenarians, respectively.
The rates for White men and women calculated in this
study resemble those reported from the Scandinavian
countries,9-"1 except that rates in the oldest women appear to
be higher in Sweden."I Estimates from Rochester, Minnesota
appear to underestimate rates for both men and women over
TABLE 1-Demographic Characteristics of Elderly Hip Fracture Patients
Medicare and VA Population: 1984-87
Women N (%)
Men N (%)
Total N (%)
Percent of Race- and Sex-SpecHic Category.
Age at Fracture
FIGURE 1-Annual Age-Specific Incidence ofHip Fracture among the Elderly:
Population-based Rates by Race and Sex for 1984-87, United States
Lines represent a smoothed curve based on the calculated age-, race-, and
sex-specific incidence rates.
SOURCES: Health Care Financing Administration, Department of Veterans
Affairs, US Bureau of Census.
the age of 80 years,57 but this may be due to small sample
sizes. It is noteworthy that our results substantiate prior
national estimates using data from the National Hospital
Discharge Survey for all race and age groups,2.4.'2 despite the
previously noted problems with similar data.13
There are several limitations in these data. The assump-
tion that all fractures of the hip necessarily result in hospi-
talization is probably valid, because hip fractures are asso-
ciated with a great deal of pain and incapacity; but persons
who did not survive long enough to be hospitalized would be
missed. A second assumption is that persons hospitalized in
this age range are covered by the Medicare insurance
program or cared for in VA hospitals. According to the
Health Care Financing Administration,'4 97 percent of the
United States population ages 65 and older is enrolled in the
Medicare program. The number of cases treated in VA
hospitals appears to account for the vast majority of the
remaining 3 percent, suggesting these rates reflect true
These data may over-estimate the risk ofan incident first
fracture since data are only available for up to four years of
follow-up. The rates presented herein may represent up to a
6 percent over-estimate of the risk of a first fracture.
Likewise, the use ofthese data to estimate the prevalence of
fracture would result in an underestimate for the same
The apparent decline in fracture incidence after age 95
among white women is intriguing. Possibly, women who are
at risk of fracture have been removed from the population
through fracture-related mortality or competing causes. It is
noteworthy that this decline persists when all fractures during
the four years of follow-up are included in the numerator.
Alternatively, this result may reflect some degree of non-
sampling error inherent to the data. Only more detailed
surveillance of an elderly population will provide the oppor-
tunity to ascertain the meaning of the decline.
The data were presented at the Steenbock Symposium on Osteoporosis in
Madison, WI, June 1989 and the 117th Annual Meeting ofthe American Public
Health Association, in Chicago, Illinois, October 1989. This work was
completed with the cooperation ofthe Health Care Financing Administration.
AJPH July 1990, Vol. 80, No. 7
PUBLIC HEALTH BRIEFS Download full-text
1. Cummings SR, Kelsey JL, Nevitt MC, O'Dowd KJ: Epidemiology of
osteoporosis and osteoporotic fractures. Epidemiol Rev 1985; 7:178-207.
2. Farmer ME, White LR, Brody JA, Bailey KR: Race and sex differences
in hip fracture incidence. Am J Public Health 1984; 74:1374-1380.
3. Silverman SL, Madison RE: Decreased incidence of hip fracture in
Hispanics, Asians, and Blacks: California hospital discharge data. Am J
Public Health 1988; 78:1482-1483.
4. Bacon WE, Smith GS, Baker SP: Geographic variation in the occurrence
of hip fractures among the elderly population of the United States. Am J
Public Health 1989; 79:1556-1558.
5. Garraway WM, Stauffer RN, Kurland LT, O'Fallon WM: Limb fractures
in a defined population. I: Frequency and distribution. Mayo Clin Proc
6. Gallagher JC, Melton LJ, III, Riggs BL, Bergstrath E: Epidemiology of
fractures of the proximal femur in Rochester, Minnesota. Clin Orthop
7. Melton LJ, III, O'Fallon WM, Riggs BL: Secular trends in the incidence
of hip fractures. Calcif Tissue Int 1987; 41:57-64.
8. Fleiss JL: Statistical Methods for Rates and Proportions. 2nd ed. New
York: John Wiley and Sons, 1981.
9. Alffram PA: An epidemiologic study ofcervical and trochanteric fractures
of the femur in an urban population. Acta Orthop Scand 1964 (Suppl 65),
10. Jensen JS: Incidence of hip fractures. Acta Orthop Scand 1980; 51:511-
11. Elabdien BS, Olerud S, Karlstrom G, Smedby B: Rising incidence of hip
fracture in Uppsala, 1965-1980. Acta Orthop Scand 1984; 55:284-289.
12. National Center for Health Statistics, Hospital Care Statistics Branch:
1987 Summary: National Hospital Discharge Survey. Advance Data From
Vital and Health Statistics. No. 159. DHHS Pub. No. (PHS) 88-1250.
Hyattsville, MD: Public Health Service, 1988.
13. Rees JL: Accuracy of hospital activity analysis data in estimating the
incidence of proximal femoral fracture. Br Med J 1982; 284:1856-1857.
14. Health Care Financing Administration, Bureau of Data Management and
Strategy. Medicare Program Statistics: Health Care Financing Adminis-
tration: Medicare Enrollment, Reimbursement, and Utilization, 1983.
HCFA Pub. No. 03234. Baltimore, MD: HCFA, 1987.
Swaddling and Acute Respiratory Infections
KADRIYE YURDAKOK, MD, TUNA YAVUZ, MD, AND CARL E. TAYLOR, MD, DRPH
Abstract: In Turkey and China the ancient practice ofswaddling
is still commonly practiced. Both countries have extremely high rates
of pneumonia, especially during the neonatal period. Preliminary
evidence on the possibility that swaddling may interfere with normal
respiratory function and thereby predispose to pneumonia was
gathered in a teaching health center in Ankara. Babies who had been
swaddled for at least three months were four times more likely to
have developed pneumonia (confirmed radiologically) and upper
respiratory infections than babies who were unswaddled. These
preliminary findings were highly significant and are being followed up
by further studies. (Am J Public Health 1990; 80:873-875.)
The ancient practice of swaddling has almost disap-
peared in most countries of the world. A national sample
survey in Turkey, however, showed that 93 percent of
mothers swaddle their children.' In China also, most babies
are tightly swaddled from birth through the first several
months of life.2 Since these two countries include over
one-fifth ofthe children ofthe world, the number ofswaddled
babies is substantial.
In both countries pneumonia is the first cause of death
among children, with particularly high incidence among
neonates. In Turkey about 50,000 infant deaths occur annu-
ally due to pneumonia.3 In China over 300,000 child deaths
per year are attributed to pneumonia with a child mortality
rate twice as high as the second highest cause ofdeath.4 Forty
percent of these deaths are under one month of age.
Address reprint requests to Kadriye Yurdakok, MD, Director of Child
Health, Ministry of Health, Cankaya, Kuloglu Sokak, 6/12, 06680, Ankara,
Turkey. Dr. Yavuz is with the Gulveren Health Center in Ankara; Dr. Taylor
is professor emeritus, Johns Hopkins School of Hygiene and Public Health,
Institute for International Programs, Baltimore. This paper, submitted to the
Journal July 5, 1989, was revised and accepted for publication December 7,
© 1990 American Journal of Public Health 0090-0036/90$1.50
In trying to explain this extremely high incidence of
pneumonia we focused on the possibility that swaddling
might interfere with normal respiratory function and lung
expansion. No studies have been published of possible
relationships between swaddling and acute respiratory infec-
tions or pneumonia.
The common cultural practice is that immediately after
birth babies are tightly bound in layers of cloth. Complete
swaddling immobilizes the baby from the neck to the feet.
The legs are pressed firmly together with the knees straight
and the arms are bound to the sides or slightly to the front of
the body. The layers of cloth are not only pulled tightly but
they are also securely tied to minimize body movement. In
partial swaddling cloth is wrapped around the legs and torso
up to the armpits, but the arms are free. In both types of
swaddling the child may also be covered with netting or a
blanket to prevent exposure to flies, drafts or cold air. Babies
are almost always laid on their backs and kept in a dark room
to induce sleep.5 Swaddled babies seldom cry and respira-
tions seem shallow to an observer, raising the question of
whether full expansion of the lungs occurs. A variety of
devices are used to dispose of excreta in Turkey but in the
urban area where this study was done families use cloth
The records from Gulveren Health Center in a suburb of
Ankara were examined to look for associations between
swaddling and selected health problems. This teaching health
center of Hacettepe Medical School has high standards of
follow-up and care ofall the children in the health center area.
Records were available on 186 infants, ofwhom 94 had been
unswaddled, 29 had been partially swaddled, and 63 had been
completely swaddled. The rate ofswaddling is lower than the
national figure because this study included only those babies
who were swaddled for at least three months.
All infants were examined and detailed histories were
taken. The ages of the children at the time of this cross-
sectional study ranged from three to 12 months, with a mean
of6.8 months. Sex ratios were similar in all groups as shown
AJPH July 1990, Vol. 80, No. 7