As is well-known, there has been a substantial shift in the sex mix of dentists and dental students. Women accounted for fewer than 3 percent of practicing dentists in 1982, 12 percent in 1990 and 13 percent in 1997, and it is projected for 2010 and 2020 that women will account for 22 percent and 28 percent, respectively, of all practicing dentists.^2–4 Furthermore, the proportion of women dental students increased from 24 percent in 1985 to 36 percent in 1993, 39 percent in 2001, and 40 percent in 2002, suggesting there may be even greater proportions of female dentists in future years.⁵

Studies from the late 1980s assessed the impact of sex trends in the work force through surveys asking about sociodemographic characteristics and work patterns, income and practice organization.^6,7 In 1987, Dolan and Lewis⁶ studied recent graduates (a relatively defined age cohort) and reported that women were more likely to work part time (< 30 hours per week) than men (12 percent versus 4 percent), and that, overall, women worked about 2.5 fewer hours per week than did men (41.3 hours versus 43.8 hours).

Surveys conducted by the American Dental Association have been the major source of data on the distribution, practices, careers and education of the U.S. dentist work force.³ These data have allowed for analyses of the practice characteristics and work patterns (including hours and weeks worked), patient visits, staffing, career patterns, practice finances and sociodemographic characteristics of the overall work force; they also have facilitated supply projections.^4,8–13

Several studies have used ADA data to assess differences in work patterns according to sex and age, often presenting data for various dentist subgroups (such as recent graduates, full-time owner dentists, private practice dentists) and for a single year or selected years. A 1995 study reported that among dentists in private practice, women worked about four fewer hours per week in their primary practice than did men (34.3 versus 37.4 hours).¹² A 1999 survey of career patterns across age and sex found that about 34 percent of women dentists reported working part time (defined as less than 32 hours per week), and that this was consistent across age groups for women. Overall, fewer men worked part time; however, the proportion increased with age (only 9 percent for men younger than 40 years of age and 21 percent for those older than 40 years of age).¹³

We conducted a study to assess dentists’ work patterns using 21 years of cross-sectional national survey data from the U.S. Bureau of Labor Statistics, which has provided repeated snapshots of the national dentist work force through the years since 1979. It characterizes dentists’ participation in the work force, "participation" encompassing the proportion of dentists working and the hours they worked per week. This report assessed the relationship of hours worked across different periods and across dentists’ sociodemographic characteristics, with a special focus on sex, age, marital status and having children. It also suggested ways in which these data can allow for work pattern comparisons with other professions that have witnessed a growing number of women entrants (such as medicine, law and pharmacy).

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
We used data from the Current Population Survey, or CPS, a national cross-sectional survey conducted by U.S. Bureau of Labor Statistics, or BLS, each month with 50,000 households based on a rotating sample. We conducted detailed analyses of hours worked by dentists using data from the CPS’s Outgoing Rotation Group, or ORG, covering 1979 through 1999.¹⁴ The analyses included descriptive statistics of work force participation and hours worked, as well as standard bivariate statistical tests (for example, t tests for differences in hours by sex) regarding differences in participation and hours worked across sex, age and survey period. In addition, we used linear regression models to measure multivariate relationships between hours worked and survey period, sex, age, marital status, having children and other observable demographic characteristics. We also examined differences across sex of the impact of age, marriage, having children and survey period on hours worked.

Data. The ORG data are compiled from a subset of the CPS households each month in which respondents complete an additional questionnaire regarding their current occupation, industry, labor force participation and earnings, as well as standard demographic questions such as education, age, race, sex and marital status. The sample size of the ORG data during the 1979–1999 period was about 300,000 people per year, yielding about 250 people each year who reported their occupation as "dentist."

We used three subsets of this ORG data set in our study. Our first analysis of work force participation used all years (1979–1999) and encompassed 5,296 respondents who identified themselves as dentists, after we excluded those who did not fit inclusion criteria (376 excluded because their age was not between 20 and 76 years, or because they had less than 16 years of education). Our second analysis of work hours used all respondents included in the first subset, all years, and it involved 4,563 respondents after we excluded those who did not report any hours worked (412) or had missing values for hours worked (321). Our third analysis focused on whether respondents had children and how having children affected the number of hours they worked; for it, we used only working dentists (from the second subset) surveyed in the years when these questions were asked (1984–1993), resulting in a sample size of 2,348.

Data analysis. Variables. The dependent variable in our analysis was the total number of hours worked in the week before the survey (again, where a value of one or more was used to indicate or select those who participated in the work force). We used two indicators for part-time work (less than 20 hours and less than 32 hours) and one for longer work hours (more than 42 hours), where we selected the total of 42 hours to differentiate overtime dentists in the sample from a large group clustered at 40 hours per week.

The independent variables included in the analyses were sex, age, marital status, having children younger than 18 years of age, survey year, and indicator variables for the geographic (U.S. Census) region and living in a metropolitan area. We examined age as reported and by categories (younger than 35 years, 35 to 55 years, and older than 55 years). We examined survey year using three subgroups of the annual cross-sectional samples that evenly divided the 21 survey years (1979–1985, 1986–1992 and 1993–1999).

Statistical methods. We calculated mean values of selected demographic and labor market variables for all dentists, and across sex and age (Table 1). Various descriptive statistics examined the percentage of dentists working (work force participation), as well as hours worked by sex, by three subgroups of survey years (Table 2) and by age groups (Table 3, page 641). We conducted linear regression analyses to further examine the multivariate relationship between hours worked and the independent variables. We ran each regression for all dentists and separately for men and women (Table 4, page 642). In addition, because of changes in the variables collected in the data, we ran the analysis of the impact of children on the subsample of years that included information on children (Table 4). We used Stata Statistical Software (Release 5.0, StataCorp, College Station, Texas) for all of the analyses, and all of the analyses incorporated survey weights (provided in the data by the BLS based on the sampling design) to make the statistics consistent with U.S. averages.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

Demographic characteristics and work patterns. Male dentists who were working were, on average, older than their female counterparts, more likely to be married and more likely to have younger children. Furthermore, while both men and women were more likely to live in metropolitan areas, men had a somewhat higher rate of living in nonmetropolitan areas (Table 1). Overall, the average age of dentists increased by about a year across the three periods (data not shown). Average reported hours worked per week by all dentists was about 40 hours, with women reporting about four fewer hours than men (36.2 versus 40.2 hours). Consistent with these data, more women reported working part time and fewer reported working more than 42 hours.

Hours worked by sex across time and age categories. Table 2 offers comparisons of hours worked among working dentists by three groups of survey periods and by sex. The average number of hours worked by all working dentists varied little by period and was close to the average of about 40 hours per week. In the two periods covering 1986 through 1999, women reported working about five hours less per week than did men. Notably, there was no significant difference between sexes in work hours in the earliest period (1979–1985), although women in those years reported higher numbers of hours worked than men. In more recent years (1986–1999), roughly 30 percent of women and 14 percent of men worked part time (< 32 hours), and about 30 percent of men and 16 percent of women worked more than 42 hours per week.

Comparisons of hours worked by age showed that dentists older than 55 years of age worked about five hours less than their younger colleagues (Table 3). For each age group, women worked fewer hours than men. Further analysis showed that men between 56 and 65 years of age worked 37.6 hours, and men older than 65 years of age worked 30.4 hours. About 25 percent of female dentists and 10 percent of male dentists 55 years of age or younger worked part time, and this increased to 30 percent for men older than 55 years of age. About 20 percent of women and 32 percent of men in the younger age groups worked more than 42 hours per week.

For women, we found that being married had no effect on number of hours worked. However, having children had a strong negative effect.

Multivariate analyses. Results from the two regression analyses examine the impact of each independent variable on mean hours worked (dependent variable) (Table 4). The first regression data column includes all working dentists and all survey years. The second includes an indicator for having children as an independent variable, and hence used only survey years that included information on children. We ran each regression on the total dentist sample (including an indicator for women), and then separately for men and women.

A robust finding in each regression with the total sample is that when we controlled for various factors, women worked significantly fewer hours than men. Controlling for other factors, and hence comparing women and men who were similar in all characteristics other than sex, indicated that women worked 4.6 and 4.8 fewer hours in the respective study periods (1984–1993 and 1979–1999) per week than men, which is slightly higher than the simple difference in mean hours (4.0 hours) shown in Table 1. Age also had a significant impact on hours worked. Those in the younger age groups worked about five to six hours more per week than did those older than 55 years of age. The only notable survey period finding was that women in the earliest survey period worked about five to six hours more than in more recent times (Table 2).

Interpreting the impact of the variables of marital status and having children required examining the regressions for men and women separately and using the data that included information on children. For women, after we controlled for children and all other variables, we found that being married had no effect on number of hours worked. However, having children had a strong negative effect (reducing the number of hours worked per week by about seven hours). For men, being married added about one hour of work per week, and having children had no effect on number of hours worked.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
This study found high overall work force participation among both male (93 percent) and female (91 percent) dentists. Participation rates declined with age among men and could not be assessed among older women dentists because of a small sample size. Among the cross-sectional samples of working dentists, on average, women worked about five fewer hours per week than in the period of 1986 through 1999, though not in the period from 1979 through 1985. These findings, drawn from a repeated national survey, add support to patterns seen across time from previous separate single-year surveys, as well as to patterns reported using ADA survey data.^3,6,12

Age also had a significant impact on hours worked per week, and being older than 55 years of age had a larger impact among male dentists than among female dentists in terms of hours worked. For men, the findings again were consistent with those of other studies showing that male dentists reduce the number of hours they work per week after the age of 55 years, with further reduction in hours after the age of 65 years.^2,13 However, a small number of older female dentists has precluded study of sex differences in the impact of age on hours worked per week among dentists older than 55 years of age.

Further analyses demonstrated that a significant factor affecting the reduced hours for women was having children younger than 18 years of age, and that marital status was not significant once we controlled for children and other factors. Though our data did not allow us to follow individual career patterns, the cross-sectional data indicated that differences in hours worked between men and women were less for older dentists and for those without young children. This finding is consistent with the few previous studies of dentists examining the impact of having children on hours worked. For example, a 1992 study found that among dentists with children, men work more hours and women fewer.¹⁵

Differences in hours worked by sex also have been found among other health professions with increased populations of women. About 45 percent of active pharmacists are women, and an analysis using this same BLS data set found that female pharmacists worked almost seven hours less per week (37.2 hours) than men (44.1).¹⁶ In addition, an earlier study of physicians found that, on average, women worked 52 hours per week while men worked 59 hours per week.¹⁷ The stability of findings over the survey years for dentists, combined with similar findings from other health professions, suggest that the differences by sex among dentists may be expected to persist. These studies show the importance of examining number of hours worked and other outcome variables by sex in professions that will see continued sex shifts. However, more research is needed to better understand underlying reasons for differences in labor force behavior by sex.

The cross-sectional data indicated that differences in hours worked between men and women were less for older dentists and for those without young children.

Study limitations. Our data set had a limited amount of information available for analysis. For example, hours worked per week is a less comprehensive collection of data than hours per year. In addition, productivity within each hour (for instance, how many patients seen, how many services were billed) and what is produced (that is, patient care or administrative work) during an hour may vary, and this is not captured in the data. Similarly, income can vary, and this can in turn relate to decisions regarding how much to work.¹¹ Our study did not incorporate income because the income measures in the CPS are censored at high values and are not reported for all cases. In addition, the location of practice and patient selection affect the overall impact of hours worked on the population’s access to dental care and the population’s dental health. Unfortunately, these questions also were beyond the scope of the available data.

The data had three other inherent limitations. First, the data set was cross-sectional, which limited the ability of the analysis to characterize work patterns across a career. Second, there may be a reporting bias in the hours variable in that responses tended to cluster at 40 hours for men and 30 for women, though to the extent it is "even" for those actually working above and below those numbers, it should not affect the averages. Third, the data set used respondents’ self-reports of occupation. However, there was no reason to suspect that people would systematically incorrectly identify themselves as dentists, these data commonly are used for work force analyses, and the findings relate well to dentist surveys conducted by the ADA. Furthermore, as part of the data selection, we dropped responses with demographic information inconsistent with being a U.S. dentist.

	CONCLUSION

TOP ABSTRACT METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Growing numbers of women entering dentistry have added a vital human resource to the profession, particularly during the past 20 years when the number of men entering dentistry declined. However, their work patterns relative to men’s will affect measurement of the overall dental work force capacity. On the basis of the consistency of our findings for dentists with those of previous studies, as well as consistency among those previous studies about dentists, and the similarities of sex work patterns of other professions, the differences by sex in hours worked can be expected to continue for some time. Our analysis also found significant differences in hours worked according to age. Consequently, consideration of these and other factors that affect work capacity could have an impact on dentist supply projections.