The cost of being a woman:

Article Outline

• Abstract
• Methods
  • Data
  • Panel Survey Design
  • Measurement of Female-Specific Conditions
  • Covariates
  • Analyses
• Results
• Discussion
• References
• Biography
• Copyright

Purpose

An important component of women’s health care is for conditions that are exclusive to women, yet little research has addressed the economic impact of health care for these conditions. The purpose of this study was to describe health care utilization for female-specific conditions, the incremental expenditures attributable to these conditions, and the overall incremental expenditures across the lifespan.

Methods

We analyzed 3 years of a nationally representative survey of the US noninstitutionalized population, the 2000–2002 National Medical Expenditure Panel Survey, which included 25,361 females aged ≥14, representing 38,170 person-years.

Results

More than one fifth of women (21.2%) reported having a female-specific condition during a 1-year period, the most common of which were gynecologic disorders (7.4%); pregnancy-related conditions (6.4%); and menopausal symptoms (5.3%). The mean increment in annual total expenditures attributable to female-specific conditions ranged from $483 for menopausal disorders to $3,896 for female cancers. The annual total health care expenditures of women with female-specific conditions were estimated to be $108 billion, of which >40% ($43.3 billion) was attributable to female-specific conditions. Women with female-specific conditions who had no health insurance were less likely to have visited a doctor (p = .0002), filled a prescription (p = .001), and been hospitalized (p = .0001) for these conditions, but more likely to have visited an emergency department (p = .02) seeking treatment for these conditions.

Conclusions

In this nationally representative sample of American women aged ≥14, female-specific conditions were common and substantially increased costs of health care.

Although it is recognized that women’s health care needs are different from men’s in part because of the complexity of reproduction and the female reproductive system (Clancy & Massion, 1992), knowledge of how reproduction and disorders of the reproductive system affect health care utilization and expenditures is quite limited. This is an important topic because the vast majority of women use health care services for female-specific conditions at some point in their lives, some many times. A better understanding of how women use the health care system for care of female-specific conditions could inform health planning efforts and improve the design of health care services for women across the life span.

Research on the topic of health care utilization for female-specific conditions is limited. A study of hospitalizations among women of reproductive age (15–44 years old) using the 1988–1990 National Hospital Discharge Summary found that 10% of all hospitalizations among women in this age group were for nonpregnancy reproductive tract disorders, in particular pelvic inflammatory disease, benign ovarian cysts, endometriosis, menstrual disorders, and uterine leiomyomas (Velebil, Wingo, Xia, Wilcox, & Peterson, 1995). A study of the prevalence of gynecologic disorders among women aged 18–50 using the 1984–1992 National Health Interview Survey found that menstrual disorders were the most common gynecologic condition in the general population, followed by adnexal conditions, uterine leiomyoma, and inflammatory disorders (Kjerulff, Erickson, & Langenberg, 1996). Women with inflammatory disorders were most likely to have seen a doctor during the prior year for this condition (95.6%), whereas women with menstrual disorders were the least likely (67.0%). Gynecologic conditions accounted for 6.3% of emergency department visits among women aged 15–44 in 1992–1994 based on analyses of the National Hospital Ambulatory Medical Care Survey (Curtis et al., 1998). Pelvic inflammatory disease and other lower genital tract infections accounted for nearly half of emergency department visits for gynecologic conditions. Analyses of ambulatory care for gynecologic disorders among women aged ≥18 using the 1995–1996 National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey found that symptoms of menopause accounted for the largest number of ambulatory visits for gynecologic disorders, followed by menstrual disorders and pelvic inflammatory disease (Nicholson, Ellison, Grason, & Powe, 2001).

Based on the 1995 National Survey of Family Growth, 88% of women aged 40–44 in the United States have been pregnant at least once and 83% have given birth (Abma, Chandra, Mosher, Peterson, & Piccinino, 1997). There are >4 million births in the United States each year, and obstetric delivery is the most common reason for hospital admission (Kozak, Owings, & Hall, 2005). Nondelivery pregnancy hospitalizations are also common and occur most frequently as treatment for preterm labor (Bacak, Callaghan, Dietz, & Crouse, 2005). In addition, cesarean section is the most common major operative procedure performed in the United States (Kozak et al., 2005). Although prenatal care and delivery are clearly important aspects of women’s health care, we are not aware of any studies that have measured the extent to which pregnancy-related health care utilization contributes to the cost of health care for female-specific conditions.

Of special interest in this study is the effect of lack of health insurance on health care utilization for female-specific conditions. In particular, we wanted to see if women lacking health insurance tended to forgo care for female-specific conditions. Prior research indicates that women without health insurance are less likely to receive preventive care, such as Pap tests and mammograms (Ayanian, Weissman, Schneider, Ginsburg, & Saslavsky, 2000), and receive fewer prenatal services (Institute of Medicine, 2004), which results in poorer cancer and birth outcomes for women lacking health insurance (Institute of Medicine, 2004). However, we are not aware of any research that has addressed whether lack of insurance affects health care utilization for female-specific conditions overall, nor have studies addressed the effect of lack of health insurance among women with female-specific conditions on multiple types of health care utilization (such as ambulatory visits and emergency room visits) within the same population.

Utilizing the 2000–2002 Medical Expenditure Panel Survey (MEPS) we sought to 1) describe the prevalence of female-specific conditions in a population-based representative sample of American women aged ≥14; 2) provide estimates of the health care utilization for these conditions in terms of ambulatory visits, prescription medications filled, hospitalizations, emergency department visits, and the effect of health insurance coverage on utilization; 3) describe the out-of-pocket and total expenditures of women with and without these conditions and the incremental expenditures attributable to each of these conditions; and 4) describe the overall incremental expenditures owing to female-specific conditions across the life span.

Back to Article Outline

Methods 

Data 

MEPS is conducted by the Agency for Healthcare Research and Quality (AHRQ) and obtains insurance costs and out-of-pocket spending for medical services, including ambulatory care, prescription drugs, hospitalizations, emergency department visits, home health care, dental care, and medical devices. The expenditures data are derived from the MEPS household interviews in combination with information obtained from participants’ providers and are defined as the actual medical care expenditures, regardless of the source of payment. Each MEPS panel is a sample of the previous year’s participants in the National Health Interview Survey, which is conducted by the National Center for Health Statistics. The National Center for Health Statistics uses a 2-stage cluster sampling design to obtain a nationally representative sample of the civilian noninstitutionalized population across the United States, with an oversampling of African Americans and Hispanics. The MEPS also obtains data on the sociodemographic characteristics of study participants and self-reported medical conditions, which are classified using the International Classification of Diseases, Ninth Revision, Clinical Modifications (ICD-9-CM) codes (Health Care Financing Administration, 1980). Each calendar year, a new panel of MEPS participants are enrolled. Participants are subsequently interviewed 5 times over the course of several years, covering 2 years of health care utilization and medical expenditures.

The MEPS is a representative sample of the US population and is designed to produce national estimates of health care utilization and expenditures. The sampling weights take into account the cluster sampling design, household nonresponse, attrition over time among respondents, and oversampling among specific ethnic groups. The response rate was 65.8% in 2000, 66.3% in 2001, and 64.7% in 2002. More detailed information about the survey design and sampling methods can be seen in several articles (Cohen 1997, Cohen 1997, Cohen et al 1997, Cohen 2003) and on the MEPS web site (www.meps.ahrq.gov).

Panel Survey Design 

Because some of the conditions of interest in this study occur relatively rarely in a population-based sample, such as female-specific cancers, it was necessary to combine multiple years of data to have sample sizes adequate to make national estimates. The MEPS uses an overlapping panel design, such that in each year some of the participants are in their first year of participation and others are in their second year. To adjust for the correlation between years caused by this overlapping panel design, we used files provided by the AHRQ that allow researchers to specify a common complex survey design structure across multiple years and with the use of SUDAAN (SUDAAN User’s Manual, 2001) or STATA (STATA, release 9, 2005), develop estimates that account for the complex survey design as well as the correlation in the data sets caused by multiple records per participant (MEPS HC-036, 2005). The estimates calculated from multiyear combined data sets can be interpreted as annual averages for the time period covered, in this case the 3-year period of 2000–2002 (Machlin, Zodet, & Nixon, 2003).

Measurement of Female-Specific Conditions 

All medical conditions reported by participants were recorded verbatim and coded into fully specified ICD-9 codes by professional coders. These ICD-9 codes were further organized by AHRQ into 3-digit clinical classification codes (MEPS HC-006, 1999), which often combined groups of ICD-9 codes. We used these clinical classification codes to identify women with female-specific conditions and disorders. The clinical classification codes that we used were as follows. For gynecologic disorders the codes were: 175, “Other female genital disorders”; 171, “Menstrual disorders”; 168, “Inflammatory diseases of female pelvic organs”; 172, “Ovarian cyst”; 169, “Endometriosis”; 046, “Benign neoplasm of uterus”; and 170, “Prolapse of female genital organs.” For pregnancy-related conditions the codes were: 196, “Normal pregnancy and/or delivery”; 177, “Spontaneous abortion”; 181, “Other complications of pregnancy”; 195, “Other complications of birth, puerperium affecting management of mother”; 174, “Female infertility”; 186, “Diabetes or abnormal glucose tolerance complicating pregnancy, childbirth, or the puerperium”; 184, “Early or threatened labor”; 178, “Induced abortion”; 218, “Liveborn”; 180, “Ectopic pregnancy”; 183, “Hypertension complicating pregnancy, childbirth and the puerperium”; 182, “Hemorrhage during pregnancy, abruptio placenta, placenta previa”; 191, “Polyhydramnios and other problems of amniotic cavity”; and 187, “Malposition, malpresentation.” There was only one code for menopausal disorders (173) and for nonmalignant breast disease (167). Women with female cancers had the following codes: 024, “Cancer of breast”; 026, “Cancer of cervix”; 025, “Cancer of uterus”; 027, “Cancer of ovary”; and 028, “Cancer of other female genital organs.”

Covariates 

In addition to age, in several analyses we controlled for the number of chronic non–female-specific conditions. We used the clinical classification codes used in previous research to define chronic conditions and measured the number of chronic conditions per person (Hwang, Weller, Ireys, & Anderson, 2001).

We measured the extent of health insurance coverage by counting the number of months in each year that each participant reported having health insurance coverage. Women who reported having health insurance in every month were categorized as being fully insured, regardless of whether the insurance coverage was private or public. Women who were uninsured for 1–11 months of the year were categorized as “uninsured some of the year” and women who were uninsured in all 12 months were categorized as “uninsured all year.”

Analyses 

Expenditures in the MEPS data are not linked to specific conditions. To estimate the contribution of female-specific conditions to total health expenditures, we used a series of separate regression equations for women with each female-specific condition and those without, following the method suggested by Yelin, Herrndorf, Trupin, & Sonneborn (2001). This involved fitting the expenditure data into a 2-stage hurdle model. In the first stage, a logistic regression model was fit to predict the probability of having zero expenditures. In the second stage, log-transformed expenditures were fit to a linear model for patients with nonzero expenditures, controlling for age, race, income, education, and number of chronic conditions. Expected expenditures were exponentiated to their original scale and multiplied by a smearing correction (Duan, 1983). Finally, expected expenditures were multiplied by the predicted probability of having nonzero expenditures. This approach was used to account for both the presence of patients with no expenditures and the typical skewness in the expenditure data. We substituted the characteristics of those with each condition into the regression equations used to estimate the expected level of expenditures of those women if they did not have a female-specific condition. Each increment was calculated as the difference between the estimated expenditures for women with each female-specific condition compared with the estimated expenditures for women who were similar in terms of age, race, income, education, and number of chronic conditions, but did not have a female-specific condition. We also calculated the incremental expenditure for having any female-specific condition versus having no female-specific condition. The same methods were used to calculate the estimated incremental expenditures for the age categories of 14–29, 30–44, 45–64, and ≥65, with a separate regression model for each age category, controlling for age (within age category), race, income, education, and number of chronic conditions.

We also calculated the odds of having ≥1 ambulatory visits for female-specific conditions, as well as prescriptions filled, hospitalizations, and emergency department visits for female-specific conditions comparing women who were insured all or some of the year versus not insured all year, adjusting for age, race, and number of female-specific conditions via logistic regression equations. We calculated these adjusted odds ratios among women with ≥1 female-specific conditions overall because, for the most part, there were too few women who were uninsured within each of the female-specific conditions to look at this issue within conditions. However, there were enough women among the 2 most common condition categories—gynecologic disorders and pregnancy-related conditions—who were not covered by health insurance all year to allow us to measure the effects of lack of insurance on ambulatory visits for gynecologic disorders and pregnancy-related conditions.

Weights were applied to all estimates in order to adjust for multiple records per participant and the complex survey design.

Back to Article Outline

Results 

In 2000 there were 9,729 females aged ≥14 in the MEPS; in 2001 there were 13,129 and in 2002 there were 15,312, for a total of 38,170 woman-years, provided by 25,361 women. Because the MEPS is a longitudinal panel survey, participants stay in the study for a 2-year period. There was a 4.1% attrition rate among females aged ≥14 in panel 5 from 2000–2001, and a 3.6% attrition rate from 2001–2002 in panel 6.

Women with pregnancy-related conditions were the youngest female-specific condition grouping, with a mean age of 28.5, whereas women with female cancers were the oldest, with a mean age of 58.4 (Table 1). The women with pregnancy-related conditions were also most likely to be minority (Black non-Hispanic or Hispanic), to be poor/near poor or low income, and to be uninsured some or all of the year (Table 1). Overall, more than one fifth of females aged ≥14 (21.2%) reported having a female-specific condition or disorder annually, 17% of whom had ≥2 female-specific conditions. Having a female-specific condition was common across the life span (Table 2), although some conditions were more common in younger women, such as pregnancy-related conditions, whereas others were more common among women in middle age, such as menopausal disorders, or at older ages, such as female cancers. Overall, slightly ≥60% of those with female-specific conditions had ≥1 ambulatory visits for their condition in the previous year and more than half had ≥1 prescriptions filled (Table 3). Women with female-specific conditions were generally not hospitalized for these conditions, with the exception of women with pregnancy-related conditions, 40.8% of whom were hospitalized over the course of a 1-year period, and women with gynecologic disorders, 5.8% of whom were hospitalized.

Table 1. Description of study population (weighted): Females aged ≥14, Medical Expenditure Panel Survey, 2000–2002

Sample size	Gynecologic disorders	Pregnancy-related conditions	Menopausal symptoms	Benign breast disease	Female cancers	One or more female-specific conditions	No female-specific conditions
	(n = 2736)	(n = 2472)	(n = 1918)	(n = 893)	(n = 536)	(n = 7828)	(n = 30342)
Age (mean, SE)	37.46(0.40)	28.48(0.20)	56.42(0.34)	50.21(0.70)	58.37(1.13)	41.98(0.29)	44.68(0.25)
Race (%)
White, non-Hispanic and other	85.4	79.5	91.2	89.5	93.4	86.0	84.3
Black, non-Hispanic and Hispanic	14.6	20.5	8.8	—	—	14.0	15.7
Income (%)
Poor/near poor/low income	25.4	43.2	18.9	24.3	31.6	29.0	30.6
Middle/high Income	74.6	56.8	81.1	75.7	68.4	71.0	69.4
Education (%)
No degree	20.2	20.0	11.2	11.2	—	17.3	25.8
HS degree	49.2	46.1	52.1	55.6	50.0	49.6	48.0
Beyond HS	30.6	34.0	36.6	33.2	32.2	33.2	26.2
Insurance (%)
Insured all year	79.0	70.5	89.6	87.7	90.2	80.7	78.1
Uninsured some of year	12.5	21.9	5.7	—	—	12.3	10.1
Uninsured all year	8.5	7.6	—	—	—	7.0	11.8
Employment (%)
Consistently	57.7	45.1	57.0	53.5	38.5	52.3	49.0
Inconsistently	21.2	30.0	9.2	—	—	19.0	19.2
Not employed	21.1	24.9	33.8	35.4	52.0	28.7	31.8
No. of chronic conditions (%)
0	49.1	70.7	26.6	35.1	—	46.9	49.5
1	24.2	19.6	25.4	25.0	23.7	23.2	23.5
≥2	26.7	9.7	48.0	39.9	58.8	29.9	27.0
Health status (%)
Excellent/very good	57.6	68.8	52.7	52.4	42.9	58.6	58.9
Good	29.6	24.3	30.5	31.6	32.7	28.6	27.6
Fair/poor	12.8	7.0	16.7	16.0	24.5	12.9	13.5
No. female-specific conditions (%)
1	76.4	67.7	85.8	70.6	77.7	82.5	NA
≥2	23.6	32.3	14.2	29.4	22.3	17.5	NA

—, <100 subjects; NA, not applicable

Table 2. Estimated annual prevalence of female-specific conditions among females aged ≥14 in the United States by age categories (weighted): Medical Expenditure Panel Survey, 2000–2002

	Estimated Number and % in US by Age Categories				Number and % Overall
	14–29	30–44	45–64	≥65
Gynecologic disorders	3,005,138(9.8)	3,042,922(9.3)	2,082,828(6.2)	516,923(2.6)	8,647,810(7.4)
Pregnancy related	4,110,303(13.4)	3,274,061(10.0)	—	—	7,422,247(6.4)
Menopausal symptoms	—	441,135(1.4)	4,549,779(13.6)	1,172,529(5.9)	6,174,624(5.3)
Benign breast disease	—	—	1,145,428(3.4)	617,186(3.1)	2,907,759(2.5)
Female cancers	—	265,571(0.8)	652,399(2.0)	767,470(3.9)	1,832,680(1.6)
Overall—women with ≥1 female-specific conditions	6,923,465(22.6)	7,303,216(22.4)	7,711,299(23.1)	2,793,533(14.0)	24,731,512(21.2)
Women with no female specific conditions	23,785,282(77.4)	25,370,930(77.6)	25,721,786(76.9)	17,148,735(86.0)	92,026,733(78.8)

—, <100 subjects.

Table 3. Annual health care utilization for female-specific conditions among females aged ≥14 who have these conditions(weighted): Medical Expenditure Panel Survey, 2000–2002

Groups	Ambulatory visits for female-specific conditions		Prescriptions for female-specific conditions		Hospitalizations for female-specific conditions		ER visits for female-specific conditions
	% who had ambulatory visits	No. of visits among those who had ambulatory visits	% who had prescriptions filled	No. of prescriptions filled among those who had prescriptions filled	% who had hospitalizations	No. of hospitalizations among those who had hospitalizations	% who had ER visits	No. of ER visits among those who had ER visits
		Mean (SE)		Mean (SE)		Mean (SE)		Mean (SE)
Gynecologic disorders	54.6	2.5(0.08)	49.4	1.6(0.03)	5.8	1.1(0.04)	4.7	1.2(0.07)
Pregnancy related	78.5	7.2(0.18)	46.7	1.7(0.06)	40.8	1.1(0.02)	11.4	1.4(0.07)
Menopausal symptoms	23.3	2.1(0.13)	90.1	2.4(0.04)	—	—	—	—
Benign breast disease	83.9	2.6(0.12)	13.2	1.4(0.10)	—	—	—	—
Female cancers	81.0	9.3(0.99)	38.1	2.6(0.11)	—	—	—	—
Overall—women with ≥1 female-specific conditions	60.7	5.1(0.14)	56.2	2.1(0.03)	15.1	1.1(0.02)	5.2	1.4(0.05)

—, <100 subjects.

Overall, women with female-specific conditions generated $108 billion in total health expenditures in the United States annually, of which >40% ($43.3 billion dollars) were attributable to their female-specific conditions (Table 4). The incremental expenditures due to female-specific conditions were highest for women aged 30–44 (Figure 1).

Table 4. Estimated annual health expenditures among females aged ≥14 in the United States by condition grouping (weighted): Medical Expenditure Panel Survey; 2000–2002^a

NA, not applicable.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Estimated annual incremental expenditures due to female-specific conditions for females aged ≥14 in the United States by age categories (weighted): Medical Expenditure Panel Survey, 2000–2002. Zero expenditures were included, less than 2% had zero incremental health expenditures. Incremental expenditures calculated for women with 1 or more female-specific conditions in comparison to women with no female specific conditions, controlling for number of chronic conditions, race, income and education. The numbers graphed (in billions) are as follows: age 14–29, 14.189; age 30–44, 16.769; age 45–64, 9.839; and age 65+, 2.486.

The adjusted odds of having ≥1 ambulatory visits for female-specific conditions for women who were insured all or some of the year versus not insured was 1.58 (95% confidence interval [CI] = 1.25–1.99; p = .0002); for having ≥1 prescriptions for female-specific conditions was 1.40 (95% CI = 1.14–1.71; p = .001); for having ≥1 hospitalizations for female-specific conditions was 1.73 (95% CI = 1.32–2.27; p = .0001); and for having ≥1 emergency department visits for female-specific conditions was 0.61 (95% CI = 0.40–0.94; p = .02). For women with gynecologic disorders, the adjusted odds of having ≥1 ambulatory visits for this condition for women who were insured all or some of the year versus not insured was 1.50 (95% CI = 1.08–2.06; p = .01) and for women with a pregnancy-related condition this was 2.05 (95% CI = 1.44–2.90; p = .0001).

Back to Article Outline

Discussion 

This is the first study to provide a comprehensive picture of health care utilization for conditions unique to women across the life span, measuring ambulatory visits, prescription usage, hospitalizations, and emergency department visits. This is also the first study to investigate the extent to which women forgo care for female-specific conditions and the degree to which this is driven by lack of health insurance. This study provides important information of relevance to understanding women’s unique health care needs.

Our analysis of a representative population-based sample of American women indicates that female-specific conditions and disorders affect more than one fifth of women annually and increase the per-capita medical expenditures of women with these conditions by an average of $1,764, or $43.3 billion for the US overall. Women aged 30–44 had the highest incremental expenditures ($16.8 billion for this age category), primarily because women in this age category had high rates of both gynecologic disorders and pregnancy, as did the women aged 14–29, and some were also beginning to experience menopausal symptoms and female cancers.

As expected, pregnancy-related conditions generated the highest incremental expenditures among female-specific conditions for the US overall and accounted for more than half of the incremental expenditures attributable to female-specific conditions. Pregnancy and related conditions are an important reason why women seek health care. This was the most common female-specific condition category among women aged 14–29 and the second most common among women aged 30–44. Women with pregnancy-related conditions (most were coded as “normal pregnancy or delivery”) were the female-specific condition most likely to be uninsured some of the year. More than one fifth of the pregnant women were uninsured some of the year and nearly 8% had no health insurance coverage at all. Women who were pregnant but not covered by health insurance were twice as likely as women who had health insurance all or part of the year to have not visited a doctor about this condition. This is concerning and supports the findings of other studies that indicate that pregnant women without health insurance may forgo prenatal care, potentially increasing their risk of adverse pregnancy outcomes (Institute of Medicine, 2004).

This study provides the first analysis of health care utilization and expenditures on behalf of women with female-specific conditions and disorders across the life span and demonstrates the continuing presence of female-specific conditions, even among women well past menopause. We were surprised to see that even among women aged ≥65, 14% reported a female-specific condition, most commonly menopausal symptoms. The MEPS condition files only provide the 3-digit ICD-9 code for menopausal and postmenopausal disorders (627), which includes a range of conditions under this heading. To help us understand this variable better, AHRQ was kind enough to provided us with an analysis of the more specific codes in this category, which indicated that 99% of women who reported having menopausal symptoms were coded as 627.2 “Menopausal or female climacteric states, symptoms such as flushing, sleeplessness, headache, lack of concentration associated with menopause.”

Female cancers occurred most commonly among women aged ≥65 (41.9% of the female cancers occurred in this age category) and were the condition grouping that was most devastating financially. Women with female cancers incurred the highest out-of-pocket expenditures, the highest total medical expenditures, and the highest incremental expenditures of the female-specific conditions. Women with female cancers were the most likely to have other chronic conditions (>80%), indicating that these women are already particularly vulnerable and have other health burdens as well.

An analysis of incident cases of self-reported benign breast disease among participants in the Nurses’ Health Study reported 5,012 cases among 165,141 person-years of follow-up, for an annual incident rate of 3.0% (Baer et al., 2003). In this study there, were 894 self-reported cases of benign breast disease among 38,170 participants, for an annual incident rate of 2.3%. If we just include the women aged ≥45, which would make it more similar to the ages of the participants in the Nurses’ Health Study, our rate is slightly over 3.0%, which is comparable to the rate obtained in the Nurses’ Health Study. In this study, nearly 84% of the women with benign breast disease had ≥1 ambulatory visit for this condition annually, perhaps indicating a high degree of concern that this condition could lead to breast cancer.

A unique aspect of this study is that we were able to measure nonutilization of health services among women with female-specific conditions. The female-specific conditions varied greatly in the extent to which they were associated with ambulatory visits. Around 80% of the women with benign breast disease, female cancers, and pregnancy-related conditions had ≥1 ambulatory visits over a 1-year period for these conditions, whereas much lower percentages of women with gynecologic disorders and menopausal symptoms did. With the exception of pregnancy, female-specific conditions did not lead to hospitalizations or emergency department visits very often. Although one might expect higher levels of hospitalization for the women with a pregnancy-related condition than 40.8%, this relatively low level of hospitalization is primarily because in any given calendar year women are in various stages of pregnancy, some would conceive in 1 year and deliver in the next. In addition, it is <50% because some pregnancies do not end in hospitalization, for various reasons.

Being insured all or part of the year increased the likelihood that a woman with a female-specific condition would see a doctor about this condition by 60%, fill a prescription for this condition by 40%, experience a hospitalization for this condition by 73%, and decreased the likelihood that she would go to the emergency department as treatment for this condition by 39%. Previous research found that 6.3% of visits made to emergency departments by women aged 15–44 were for gynecologic disorders, most commonly pelvic inflammatory disease (Curtis et al., 1998). It was also found that younger women (aged 15–24) and Black women were more likely to visit emergency departments for gynecologic disorders than women who were older and White (Curtis et al., 1998). Although that study did not investigate the association between lack of health insurance coverage and visits to emergency departments for gynecologic disorders, they noted that the higher rates of visits to emergency departments among young, Black women suggested problems with access to more appropriate primary health care services for treatment for gynecologic disorders. Our study also suggests that limited access to primary care among women who do not have health insurance could be an important reason why women may seek care for female-specific conditions at emergency departments rather than more appropriate venues.

We must note several important limitations of this study. In this study. we were not able to look at provider specialty of the physicians that women with female-specific conditions went to or the effect of type of provider on health care utilization or related issues. The MEPS did not begin to collect information on provider specialty until 2002; therefore, we only had information on type of provider seen for a subset of our study population. Another limitation of this study is that due to small sample sizes for many specific conditions, we combined women with similar types of conditions, such as female cancers, into categories. Although this allowed us to describe a broad picture of health care utilization and expenditures for female-specific conditions overall, and for types of conditions, we were not able to investigate specific conditions individually.

This study has important implications for health care for women. First, using a nationally representative sample of American women aged ≥14, we found that female-specific conditions and disorders were common among women across the life span, affecting >20% of women annually. Health care utilization for these conditions was primarily outpatient, often requiring multiple outpatient visits. This underscores the importance of ambulatory care for women’s health, as well as access to providers with the expertise to provide care to women with these conditions and disorders. Second, we found that >20% of the women in the United States had no health insurance coverage for some or all of the year and nearly 30% of pregnant women did not. This indicates that there are significant gaps in health care coverage for women that need to be addressed. Third, lack of health insurance coverage increased the likelihood that women with female-specific conditions would forgo care for these conditions in terms of the usual venues of health care, such as physician’s offices, but would instead seek care for these conditions at emergency departments. Women who do not have health insurance may have limited access to private physician’s offices and may find emergency departments to be their only available source of health care. This is not an optimal situation for the treatment of female-specific conditions; emergency departments are often not staffed with providers who have expertise in the treatment of female-specific conditions, nor can emergency departments provide the continuity of care that appropriate treatment for these conditions requires. In summary, the design of health care services for women should take into account the high prevalence of female-specific conditions among women across the life span and seek ways to improve access to care for women who have these conditions but lack health insurance coverage.

Back to Article Outline

References 

1. Abma JC, Chandra A, Mosher WD, Peterson LS, Piccinino LJ. Fertility, family planning, and women’s health: New data from the 1995 National Survey of Family Growth. Vital Health Statistics. 1997;23:1–114
   • View In Article
2. Ayanian JZ, Weissman JS, Schneider EC, Ginsburg JA, Saslavsky AM. Unmet health needs of uninsured adults in the United States. Journal of the American Medical Association. 2000;284:2061–2069
   • View In Article
3. Bacak SJ, Callaghan WM, Dietz PM, Crouse C. Pregnancy-associated hospitalizations in the United States, 1999–2000. American Journal of Obstetrics and Gynecology. 2005;192:592–597
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (128 KB)
   • CrossRef
4. Baer HJ, Schnitt SJ, Connolly JL, Byrne C, Cho E, Willett WC, et al. Adolescent diet and incidence of proliferative benign breast disease. Cancer Epidemiology Biomarkers and Prevention. 2003;12:1159–1167
   • View In Article
5. Clancy CM, Massion CT. American women’s health care: A patchwork quilt with gaps. Journal of the American Medical Association. 1992;268:1918
   • View In Article
6. Cohen JW. A design and methods of the Medical Expenditure Panel Survey household component. MEPS Methodology Report No. 1. Rockville, MD: Agency for Health Care Policy and Research; 1997;[AHCPR Publication No. 97-0026].
   • View In Article
7. Cohen JW, Monheit AC, Beauregard KM, Cohen SB, Lefkowitz DC, Potter DE, et al. The Medical Expenditure Panel Survey: A national health information resource. Inquiry. 1997;33:373–389
   • View In Article
   • MEDLINE
8. Cohen SB. A sample design of the 1996 Medical Expenditure Panel Survey Household Component (MEPS methodology report No 2 [AHCPR Publication No. 97-0027]). Rockville, MD: Agency for Health Care Policy and Research; 1997;
   • View In Article
9. Cohen SB. Design strategies and innovations in the Medical Expenditure Panel Survey. Medical Care. 2003;41:5–12
   • View In Article
10. Curtis KM, Hillis SD, Kiefe BA, Brett KM, Marchbanks PA, Peterson HB. Visits to emergency departments for gynecologic disorders in the United States, 1992–1994. Obstetrics and Gynecology. 1998;91:1007–1012
    • View In Article
    • MEDLINE
    • CrossRef
11. Duan N. Smearing estimate: A nonparametric retransformation method. Journal of the American Statistical Association. 1983;78:605–610
    • View In Article
12. Health Care Financing Administration. International classification of diseases, clinical modification(9th rev.; vol. 1). Washington, DC: DHHS, US Public Health Services; 1980;[DHHS Pub No (PHS) 80-1260]
    • View In Article
13. Hwang W, Weller W, Ireys H, Anderson G. Out of pocket medical spending for care of chronic conditions. Health Affairs. 2001;20:267–278
    • View In Article
14. Institute of Medicine. Insuring America’s health: Principles and recommendations. Washington, DC: Author; 2004;
    • View In Article
15. Kjerulff KH, Erickson BA, Langenberg PW. Chronic gynecological conditions reported by US women: Findings from the National Health Interview Survey, 1984–1992. American Journal of Public Health. 1996;86:195–199
    • View In Article
    • MEDLINE
    • CrossRef
16. Kozak LJ, Owings MF, Hall MJ. National Hospital Discharge Survey: 2002 annual summary with detailed diagnosis and procedure data. Vital Health Statistics. 2005;13(158):1–199
    • View In Article
17. Machlin SR, Zodet MW, Nixon JAA. Estimates of medical expenditures from the Medical Expenditure Panel Survey: Gains in precision from combining consecutive years of data. Alexandria, VA: American Statistical Association; 2003;Proceedings of the Section on Survey Research Methods (pp. 2615–2622)
    • View In Article
18. MEPS HC-006. 1996 Medical conditions [AHRQ Publication No. 99-DP06]. Rockville, MD: Agency for Healthcare Research and Quality; 1999;
    • View In Article
19. MEPS HC-036. 1996–2003 Pooled estimation file (Center for Financing, access and cost trends). Rockville, MD: Agency for Healthcare Research and Quality; 2005;
    • View In Article
20. Nicholson WK, Ellison SA, Grason H, Powe NR. Patterns of ambulatory care for gynecologic conditions: A national study. American Journal of Obstetrics and Gynecology. 2001;184:523–530
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (83 KB)
    • CrossRef
21. STATA Release 9. (2005). College Station, TX: StataCorp.
    • View In Article
22. SUDAAN user’s manual (Release 8.0). (2001). Research Triangle Park, NC: Research Triangle Institute.
    • View In Article
23. Velebil P, Wingo PA, Xia Z, Wilcox LS, Peterson HB. Rate of hospitalization for gynecologic disorders among reproductive-age women in the United States. Obstetrics and Gynecology. 1995;86:764–769
    • View In Article
    • MEDLINE
    • CrossRef
24. Yelin E, Herrndorf A, Trupin L, Sonneborn D. A national study of medical care expenditures for musculoskeletal conditions: The impact of health insurance and managed care. Arthritis and Rheumatism. 2001;44:1160–1169
    • View In Article
    • MEDLINE
    • CrossRef