Understanding Gender Sensitivity of the Health Care Workforce at the Veterans Health Administration

Abstract

Background:

Gender sensitivity of providers and staff has assumed increasing importance in closing historical gender disparities in health care quality and outcomes. The Department of Veterans Affairs (VA) has implemented several initiatives intended to improve gender sensitivity of its health care workforce. The current study examines practice- and individual-level characteristics associated with gender sensitivity of primary care providers (PCPs) and staff.

Methods:

We surveyed PCPs and staff (nurses, medical assistants, and clerks) at 12 VA medical centers (VAMCs) (n = 256 of 649; response rate, 39%). Gender sensitivity was measured using a 10-item scale adapted from the Gender Awareness Inventory-VA. We used weighted multivariate regression with maximum likelihood estimation to identify individual-and practice-level characteristics associated with gender sensitivity of PCPs and staff.

Results:

PCPs and staff had similar gender sensitivity but differed in most characteristics associated with that gender sensitivity. Among PCPs, women’s health training and positive communication with others in the clinic were associated with greater gender sensitivity. For staff, prior work experience caring for women, working in Women’s Health Patient-Aligned Care Teams, and rural location were associated with greater gender sensitivity, whereas more years of VA service was associated with lower gender sensitivity. Working at VA medical centers with a higher volume of women veteran patients was associated with greater gender sensitivity for both PCPs and staff.

Conclusions:

Women’s health training and experience in working with other women’s health professionals are strongly correlated with greater gender sensitivity in the clinical workforce.

Women veterans have traditionally accounted for a small percentage of patients receiving care within the Department of Veterans Affairs (VA) health care system (<10%) (Yano, Washington, Goldzweig, Caffrey, & Turner, 2003). However, as the number of women veterans seeking care at the VA has grown, their experiences have highlighted gender-based disparities in access to and experiences with VA care (Frayne et al., 2018; Klap et al., 2019). The VA has taken multiple steps to create a more equitable, high-quality care environment for women veterans, including policies recommending that all women veterans receive comprehensive primary care from a Women’s Health Patient-Aligned Care Team (WH-PACT) led by primary care providers (PCPs) with training and/or experience in women’s health (USVA, 2017). These changes have successfully reduced some gender disparities in care (Wright, Schaefer, Reyes-Harvey, & Francis, 2012); however, disparities in quality of care still persist, particularly for chronic disease management (e.g., diabetes and hypertension) (Whitehead, Czarnogorski, Wright, Hayes, & Haskell, 2014; Wright et al., 2012), continuity of care (Wright, Craig, Campbell, Schaefer, & Humble, 2006), inpatient services (Wright et al., 2012), and patient experience of care (Kehle-Forbes et al., 2017).

Persistent gender gaps have been attributed at least in part to the VA workforce’s readiness to care for women veterans (Yano, Haskell, & Hayes, 2014), and in particular to the fact that women veterans’ numerical minority status means that many providers and staff have little or no experience in providing care for women (Bergman, Frankel, Hamilton, & Yano, 2015; Chuang et al., 2017; Maisel et al., 2015; Yano et al., 2014). To maintain women’s health proficiency for providers, VA policy recommends that PCPs in WH-PACTs should have either 100 women veterans in their primary care panel, 3 years of prior experience working in women’s health, and/or recent training in women’s health (e.g., completion of VA’s women’s health mini-residency, women’s health fellowship, or preceptorship with an experienced women’s health provider) (USVA, 2017). VA guidelines also recommend that WH-PACT staff have “knowledge and skills to provide care to women veterans,” but provide little specificity as to how this will be achieved (USVA, 2017). A recent review of VA care for women veterans found that, in practice, adherence to these proficiency standards varies significantly across VA and that on average, only 44% of PCPs in WH-PACTs had documentation of proficiency in care of women (USVA Office of Inspector General, 2017).

Low gender sensitivity among VA employees can contribute not only to gender disparities in care quality and access (Washington, Bean-Mayberry, Riopelle, & Yano, 2011), but also to attrition of women veterans from VA care (Hamilton, Frayne, Cordasco, & Washington, 2013). Conversely, higher gender sensitivity has been found to be associated with increased provider and staff confidence in delivering gender-sensitive comprehensive primary care for women patients (Meredith, Wang, et al., 2017).

Understanding and improving gender sensitivity of the VA workforce is critical to efforts to improve women veterans’ experiences with VA care. However, the few prior studies that have examined gender sensitivity occurred before the implementation of recent VA policy initiatives intended to improve gender sensitivity of its workforce (Fox et al., 2016; Vogt, Barry, & King, 2008), such as training existing providers in women’s health and hiring new providers who already have this expertise (Cordasco et al., 2015).

The goal of the present study was to identify individual- and practice-level characteristics associated with gender sensitivity of VA PCPs and staff. Informed by prior conceptual framework on gender-sensitive care (de Kleijn, Lagro-Janssen, Canelo, & Yano, 2015) and research on gender sensitivity (Vogt et al., 2008; Vogt et al., 2001) we hypothesized that individual-level characteristics such as female gender, role (e.g., PCP vs. staff), years of service in VA, training, and prior experience in women’s health would be associated with higher gender sensitivity (Vogt et al., 2008; Vogt et al., 2001). We also hypothesized that practice-level characteristics, such as being part of a primary care team designated specifically for care of women veterans (WH-PACT) versus PCPACT, quality of communication within the clinic, proportion of women veterans seen locally, and rurality of the clinic would be associated with gender sensitivity.

Methods

Sample

We surveyed PCPs and staff at 12 VA Medical Centers (VAMCs) implementing WH-PACTs (Yano et al., 2016). Primary care settings included both general primary care clinics with care for women integrated with care for men, and women’s health clinics for primary care. PCPs included physicians, nurse practitioners, and physician assistants. Staff included nurse care managers (RN), medical assistants/medical technicians (LPN/ LVN), and clerks. We excluded PCPs and staff in PACTs for geriatrics, infectious disease, home based care, homelessness, post-deployment health, renal or dialysis, serious mental illness, and spinal cord injuries and disorders (VA Handbook 1101.10). We surveyed 280 PCPs and 369 staff, who were identified using VA’s primary care panel management databases (e.g., VA Corporate Data Warehouse and VA Support Service Center). We contracted with RAND for survey administration, which was carried out online and via mail between September 8, 2014, and June 18, 2015. Survey development has been described elsewhere (Meredith, Azhar, et al., 2017; Meredith, Wang, et al., 2017). We received 256 survey responses (94 PCPs and 162 staff). Our analysis included 91 PCPs (33% response rate) and 151 staff (41% response rate) who answered the gender sensitivity questions. There were no differences between respondents and non-respondents by gender or type of profession. However, we found that staff with no experience caring for women patients did not answer all the gender-sensitivity questions. The VA Greater Los Angeles Healthcare System and RAND Santa Monica Institutional Review Boards approved the study.

Gender Sensitivity Measure

Gender sensitivity was assessed using 10 items adapted from the previously validated Gender Awareness Inventory-VA (Salgado, Vogt, King, & King, 2002; Vogt et al., 2001). Survey items were modified and pretested with eight PCPs using cognitive interviewing techniques. Their feedback was incorporated in the final surveys. Respondents were asked to respond on a 5-point Likert scale, from 1 (strongly disagree) to 5 (strongly agree), to the following items: 1) The VA should not be expected to provide special health services for women, 2) It would bother me to see a woman breastfeed in the clinic, 3) Having a special room for women to breastfeed would be a good clinic policy, 4) Sometimes I wish VA primary care clinics had only male patients, 5) It is nice to have female patients at VA primary clinics, 6) Special women’s clinics should be at all VA health facilities, 7) Having female patients at VA primary care clinics makes things too difficult, 8) Compared with men, women expect too much courtesy from clinic staff, 9) Female patients care too much about the way the clinic looks, and 10) Having female patients makes this a better clinic. We reverse coded the six negatively worded items (items 1, 2, 4, 7, 8, and 9) and then created a single composite score based on respondents’ average responses to all 10 items (Cronbach’s alpha = 0.78). The composite score ranged from 1 to 5, with higher scores reflecting greater gender sensitivity. The score was a continuous variable slightly skewed to the left. Because transformations of the score did not improve the data distribution, we analyzed using the untransformed composite score for ease of interpretation.

Women’s Health Training, Experience, and Individual Employee Characteristics

Women’s health training was coded as a binary (yes or no) variable and defined differently for PCPs and staff. For PCPs, women’s health training was defined as self-reported completion of or attendance at any of the following: 1) VA women’s health mini-residency, 2) Veterans Integrated Service Network–sponsored mini-residency, 3) VA women’s health conferences or trainings in person or audio/video-presentations, such as VA’s eHealth University, 4) women’s health or gynecology Specialty Care Access Network-Extension for Community (SCAN-ECHO) (Arora et al., 2011), a virtual program that trains and supports PCPs with specialist consultation on patient cases, 5) non-VA women’s health conferences, 6) preceptorship with experienced WH-PCPs on a regularly scheduled basis, or 7) family practice or internal medical residency, or women’s health fellowship within the past 3 years.

Among staff, women’s health training was defined as completion of or attendance at one of the following: 1) shadowed an experienced women’s health provider/staff on a regularly scheduled basis, 2) VA women’s health conferences or trainings in person or via audio/video-presentations, 3) women’s health or gynecology SCAN-ECHO, 4) non-VA women’s health conferences, or 5) other relevant trainings in women’s health.

We constructed dichotomous variables for individual gender, and for prior women’s health experience. We defined prior women’s health experience as having cared for at least 50% women patients in panel for at least 3 years (yes vs. no) because information on prior panel size was not available and current percent of women patients in panel was highly correlated with working in WH-PACT. We measured an individual’s length of service at VA in years.

Practice Characteristics

We asked individuals to indicate whether they were working in WH-PACTs at the time of the survey and included it as a binary variable (1 = WH-PACTs and 0 = PC-PACTs). We measured communication with other clinical care professionals in clinics using five items responded on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree). Sample items included, “Our staff and clinicians have constructive work relationships,” and “In this clinic, co-workers from different clinical or administrative backgrounds frequently interact to solve quality of care problems.” We formed a composite score for communication by taking the average of individual responses to the five items (Cronbach’s alpha = 0.80), where higher scores reflected more positive communication. We used the VA Support Service Center database to identify the rurality of clinic location (urban vs. rural) and the percentage of women veterans enrolled at each participating VAMC.

Data Analysis

We evaluated sample characteristics combined and separately by PCPs and staff. To compare gender sensitivity by the number of women’s health trainings, we recoded women’s health training as the incremental number of the trainings individuals completed and their gender sensitivity score at each level. We also conducted sensitivity analyses in which we examined the relationship between each individual type of women’s health training and gender sensitivity (Appendix).

To identify the association between individual- and practice-level characteristics and gender sensitivity, we used multivariate linear regression with full information maximum likelihood estimation (FIMLE) (Allison, 2012). We checked for multicollinearity and verified that variance inflation factors for independent variables were less than three. Because we had a relatively small sample size and about 10% cases with observed values for the dependent variable had missing data on one or more predictors, FIMLE allowed us to retain all cases with observed values (i.e., cases with missing data on some of the independent variables were not deleted). We used FIMLE for our analysis, instead of alternate linear regression approaches, because 1) multiple imputation resulted in imputation of only 3 to 4 cases for about 25 missing cases of predictors, and 2) linear regression limited to complete cases showed similar coefficient estimates as FIMLE.

We conducted a sensitivity analysis including other providers and staff in the broader medical neighborhood for WH-PACTs and PC-PACTs. They included social workers, dietitians, psychiatrists, psychologists, pharmacists, and nutritionists. These providers and staff support several PACTs at the same time and are less likely to engage in daily communication occurring between PCPs and core PACT teamlet staff. The surveys included 126 individuals from the broader medical neighborhood. Because there were 29 respondents (23% response rate), we conducted sensitivity analyses in combination with the 151 core PACT teamlet staff respondents and reported our findings descriptively.

All analyses were weighted for nonresponse weights, the inverse predicted probabilities of response by the type of clinic (primary care vs. women’s health), position (e.g., physicians, nurse practitioner, physician assistant, registered nurse, medical assistant, and clerk), and gender, so the estimates were representative of PCPs and staff in PC-PACTs and WH-PACTs at the 12 participating VAMCs. We used Stata version 13.1 for all analyses (StataCorp, College Station, TX).

Results

Overall, PCPs and staff (nurses, medical assistants, and clerks) were predominantly female (74.0%) and had an average of 14.9 years of service at VA (Table 1). Similar proportions of PCPs and staff worked in WH-PACTs (41.0% vs. 39.5%), and PCPs and staff reported similar average communication scores. Among the PCPs in WH-PACTs, 81.2% reported having had at least one women’s health training, and the remaining reported prior women’s health experience for at least 3 years (results not reported). Among the staff in WH-PACTs, 43.0% reported completion of at least one women’s health training, 19.6% reported no training activity but had prior women’s health experiences for at least 3 years, and 37.4% reported neither a women’s health training nor women’s health experience (results not reported).

Table 1.

Characteristics of Primary Care Providers and Staff

Characteristics	Primary Care
	All (N = 242)†	Providers (n = 91)	Staff (n = 151)
Women’s health training/experience
Had ≥1 women’s health training	36.7%	50.4%	27.5%
Had experience in practice with ≥50% women patients in past 3 years	32.1%	38.4%	27.9%
Individual characteristics
Years of service at VA	14.9 ± 11.7	13.2 ± 12.1	17.3 ± 10.7
Female	74.0%	58.8%	84.2%
Practice characteristics
WH PACT member	40.1%	41.0%	39.5%
Communication across discipline in clinic*	3.4 ± 0.8	3.4 ± 0.8	3.4 ± 0.8
Clinic location in rural area	11.3%	10.3%	12.0%
% women veterans at VAMC	7.4 ± 2.4	7.9 ± 2.4	7.1 ± 2.4

Abbreviations: SD, standard deviation; VA, Department of Veterans Affairs; VAMC, VA medical center; WH-PACT, Women’s Health Patient-Aligned Care Team.

^*Communication score ranges from 1 to 5 where higher score indicates more positive communication with others within clinic.

^†Fourteen respondents (3 primary care providers and 11 staff) who did not respond to all the gender sensitivity questions were excluded. Primary care providers included physicians, nurse practitioners, and physician assistants. Staff included care managers, medical assistants, medical technicians, and clerks.

Responses to gender sensitivity items revealed mostly positive views (Table 2). The overall mean gender sensitivity score was high, 4.04 out of 5 (standard deviation, 0.6), with the mean scores of 4.10 (standard deviation, 0.50) for PCPs, and 4.00 (standard deviation, 0.66) for staff (Table 3). On average, gender sensitivity scores were higher with a greater number of women’s health trainings and with women’s health experience.

Table 2.

Gender Sensitivity Item-Level Responses by Primary Care Providers and Staff

Gender Sensitivity Survey Items	Primary Care (%)
	All (N = 242)*	Providers (n = 91)	Staff (n = 151)
1. The VA should not be expected to provide special health services for women. (disagree/strongly disagree)	85	91	81
2. It would bother me to see a woman breastfeed in the clinic. (disagree/strongly disagree)	80	90	73
3. Having a special room for women to breastfeed would be a good clinic policy. (agree/strongly agree)	79	82	77
4. Sometimes I wish VA primary care clinics had only male patients. (disagree/strongly disagree)	75	77	74
5. It is nice to have female patients at VA primary clinic. (agree/strongly agree)	73	75	72
6. Special women’s clinics should be at all VA health facilities. (agree/strongly agree)	74	72	76
7. Having female patients at VA primary care clinics makes things too difficult. (disagree/strongly disagree)	69	60	75
8. Compared to men, women expect too much courtesy from clinic staff. (disagree/strongly disagree)	69	77	63
9. Female patients care too much about the way the clinic looks. (disagree/strongly disagree)	64	76	56
10. Having female patients makes this a better clinic. (agree/strongly agree)	53	59	48

Abbreviation: VA, Department of Veterans Affairs.

Note: The response options for gender sensitivity questions are: strongly agree, agree, neither agree nor disagree, disagree, and strongly disagree. The numbers reported are total percentages of respondents who selected the options in the parenthesis.

^*Fourteen respondents (3 primary care providers and 11 staff) who did not respond to all the gender sensitivity questions were excluded. Primary care providers included physicians, nurse practitioners, and physician assistants. Staff included care managers, medical assistants, medical technicians, and clerks.

Table 3.

Gender Sensitivity Score by the Number of Women’s Health Trainings and Years of Experience

Women’s Health Trainings and Years of Experience	Primary Care
	All (N = 242)*		Providers (n = 91)		Staff (n = 151)
	N	Mean ± SD	N	Mean ± SD	N	Mean ± SD
Overall score	242	4.04 ± 0.60	91	4.10 ± 0.50	151	4.00 ± 0.66
Number of women’s health trainings
0	152	3.94 ± 0.60	44	3.93 ± 0.50	108	3.94 ± 0.65
1	47	4.17 ± 0.59	19	4.24 ± 0.44	28	4.11 ± 0.70
2	23	4.24 ± 0.56	15	4.15 ± 0.52	8	4.40 ± 0.61
3	12	4.11 ± 0.35	7	4.23 ± 0.31	5	3.92 ± 0.31
4	6	4.46 ± 0.37	5	4.53 ± 0.36	1	4.1 ± 0
5	2	4.95 ± 0.05	1	4.90 ± 0	1	5.0 ± 0
Had experience ≥3 years in practice with ≥50% women patients
0	164	3.97 ± 0.61	56	4.05 ± 0.54	108	3.93 ± 0.64
1	78	4.18 ± 0.55	35	4.17 ± 0.43	43	4.20 ± 0.64

Note: Means and standard deviations (SDs) present gender sensitivity composite score, ranging from 1 to 5 where higher scores indicate greater gender sensitivity.

^*Fourteen respondents (3 primary care providers and 11 staff) who did not respond to all the gender sensitivity questions were excluded. Primary care providers included physicians, nurse practitioners, and physician assistants. Staff included care managers, medical assistants, medical technicians, and clerks.

In multivariate regression analyses, gender sensitivity did not vary by gender for either PCPs or staff (Table 4). In analyses of PCPs and staff combined, prior experience in caring for women patients, working in WH-PACTs (vs. PC-PACTs), more positive communication within clinics, and working at VAMCs with higher volume of women veterans were significantly associated with higher gender sensitivity. However, individuals with more years at VA had lower gender sensitivity. When analyzed separately, the factors associated with gender sensitivity were different for PCPs and staff, except for working at VAMCs with a higher volume of women veterans. Among PCPs, having had at least one women’s health training and more positive communication within clinic were associated with higher gender sensitivity. Among staff, prior experience in caring for women patients, working in WH-PACTs, and working in a rural area were significantly associated with greater gender sensitivity. However, staff with more years in VA had significantly lower gender sensitivity. To test whether the association of tenure in VA on gender sensitivity was moderated by working in WH-PACTs, we evaluated the interaction between years of service in VA and working in WH-PACT while controlling for all other variables. The interaction coefficient was negative and significant for staff, but not significant for PCPs (results not reported). A scatterplot confirmed that the negative association between years of service and gender sensitivity was more apparent among staff in WH-PACTs than staff in PC-PACTs. When we examined staff with or without at least 3 years of prior experience in women’s health or women’s health training, we found that the negative interaction term between years in VA and WH-PACT was significant only for staff without at least 3 years of prior experience in women’s health or women’s health training.

Table 4.

Multivariate Regression for Predictors of PCP and Staff Gender Sensitivity

Independent Variables	Primary Care
	All (N = 242)*	Providers (n = 91)	Staff (n = 151)
	Beta (SE)	Beta (SE)	Beta (SE)
Intercept	3.31 (0.24)†	2.94 (0.36)†	3.52 (0.31)†
PCPs (vs. staff)	0.01 (0.08)
Women’s health training/experience
Had ≥1 women’s health training	0.16 (0.08)	0.28 (0.12)‡	0.17 (0.11)
Had experience in practice with ≥50% women patients in past 3 years	0.19 (0.08)‡	0.11 (0.10)	0.26 (0.11)‡
Practice/individual characteristics
Female	−0.08 (0.10)	0.12 (0.13)	−0.26 (0.17)
Years of service at VA	−0.008 (0.003)§	−0.002 (0.005)	−0.009 (0.004)‡
WH-PACT (vs. PC-PACT)	0.24 (0.08)§	0.04 (0.12)	0.33 (0.10)§
Communication across discipline within clinic	0.09 (0.04)‡	0.11 (0.05)‡	0.08 (0.07)
Clinic location in rural area	0.17 (0.11)	0.16 (0.21)	0.21 (0.11)‡
% women veterans at VAMC	0.05 (0.01†	0.07 (0.02)†	0.04 (0.02)‡

Abbreviations: PCP, primary care providers; SE, standard error; VA, Department of Veterans Affairs; VAMC, VA medical center; WH-PACT, women’s health Patient-Aligned Care Team.

Note: The regressions used maximum likelihood estimation.

^*Fourteen respondents (3 PCPs and 11 staff) who did not respond to all the gender sensitivity questions were excluded. PCPs included physicians, nurse practitioners, and physician assistants. Staff included care managers, medical assistants, medical technicians, and clerks.

^†p < .001.

^‡p < .05.

^§p < .01.

When examined by the specific type of women’s health training received, PCPs and staff who completed the training had generally higher gender sensitivity (Appendix Table 1). For PCPs, the association between the specific training and gender sensitivity was not significant after adjusting for covariates (Appendix Table 2). For staff, participation in the women’s health or gynecology SCAN-ECHO sessions was significantly associated with gender sensitivity after adjusting for covariates (Appendix Table 2).

Sensitivity analysis combining 29 staff from the broader medical neighborhood with the 151 core PACT teamlet staff showed similar associations in the regression model compared to the model limited to the 151 PACT teamlet staff. One exception was that the volume of women veterans at VAMCs no longer predicted gender sensitivity (results not reported).

Discussion

The VA has implemented policies with aims to improve overall care experiences for women veterans, but no studies have evaluated correlates of gender sensitivity since these policies were enacted. Using a representative sample of the primary care workforce at 12 VA facilities, we examined factors within the current care and policy environments that were associated with gender sensitivity of PCPs and staff. We found that the volume of women veterans seen locally, individual practice experience with caring for women patients, working in WH-PACTs, communication quality within clinics, and years worked at the VA were significantly associated with gender sensitivity. However, the magnitude and statistical significance of each variable associated with gender sensitivity varied between PCPs and staff.

For PCPs, the positive association between women’s health training and gender sensitivity highlights the importance of VA’s investment in provider training for creating a culture of care that better meets the needs of women veterans (Cordasco et al., 2015; Zuchowski et al., 2017). The finding that the local volume of women veterans was associated with gender sensitivity underscores the importance of having a sufficient number of women veteran patients in the VA care environment independent of WH-PACT or women’s health training and experience. Smaller sites of care, including community-based outpatient clinics, may require alternative care arrangements to offset small patient volumes (Cordasco, Mengeling, Yano, & Washington, 2016). These measures may include telehealth arrangements with more experienced women’s health providers, more active preceptorship at a distance, and/or engagement with non-VA community providers with ample exposure to women patients, or other novel approaches (Moreau et al., 2018).

In addition, providers’ perceptions of positive communication within clinics were associated with higher gender sensitivity. The relationship between communication and attitudes has been documented in social studies more broadly (Conway & Schaller, 2007). The more individuals in a clinic come together to solve problems and have constructive work relationships, the more likely that individuals will share their opinions through conversation. In the context of WH-PACT, caring for women veterans in a clinic may have created opportunities for repeated communication about women’s health and women veterans that contributes to increasingly consensual opinions about and approaches to care delivery for women veterans (Celik, Lagro-Janssen, Widdershoven, & Abma, 2011). The VA has already created an environment for enhanced communication through team-based care, where PCPs in all PACTs are expected to communicate with staff in the teamlet and from the broader medical neighborhood (e.g., pharmacy, social work) for care coordination and have discussions regarding the care needed for patients (Schectman, & Stark, 2014; Rodriguez et al., 2014). Studies of PCPs’ experiences with PACT implementation found that having a good communication relationship with other providers and staff can influence whether delivery of women’s health comprehensive care succeeds (Bergman et al., 2015; Chuang et al., 2017). Efforts to support such communication in the context of team-building are likely to be positive adjuncts to other efforts to enhance gender sensitivity and women’s primary care delivery.

Among staff, at least 3 prior years of experience caring for women patients and working in WH-PACTs were associated with higher gender sensitivity. Participation in women’s health or gynecology SCAN-ECHO sessions was also associated with higher gender sensitivity. SCAN-ECHO sessions are the only training modality that combines consultation with specialists regarding specific clinical cases with didactic women’s health education (Cordasco et al., 2015). These results highlight the importance of hands-on experiences for staff whose job positions required different types of training. One in four staff were clerks and medical assistants who generally had fewer women’s health trainings than nurses. Perhaps unsurprisingly, the local volume of women veterans was also related to enhanced gender sensitivity among staff. Because many VA facilities have relatively low volumes of women veterans (5 of 12 participating VAMCs had <7% women veterans enrolled), staff may have limited opportunities to gain hands-on experience working with women veterans, particularly if not part of a WH-PACT.

In contrast, staff with more years of service at VA had lower gender sensitivity, indicating greater gender sensitivity among newer staff compared to staff with longer tenures. The differences may be due to different expectations about serving women veterans. Longer-term staff were hired when even fewer women received care in the VA, whereas newer staff were likely to receive information about women veterans and their use of VA services during orientation. We verified that staff with more years of service at VA had similar women’s health training and experience compared to staff with fewer years of service. Additional analyses showed a positive association between working in WH-PACTs and gender sensitivity diminished for staff with longer tenure in VA if they had not had a women’s health training or experience. The findings highlight the fact that simply working in a women’s health environment does not automatically imbue providers and staff with greater gender sensitivity and that having a training or experience in caring for women patients carries greater importance for enhancing gender sensitivity. Trainings that are specifically geared toward increasing provider and staff gender sensitivity should be evaluated (Vogt et.al., 2008).

The finding that working in rural VA facilities was associated with greater gender sensitivity is noteworthy. Rural VA facilities had smaller volumes of women veterans seen than urban VA facilities. PCPs and staff in rural and urban facilities had similar results for reported women’s health training, women’s health experience, and other characteristics, except that communication scores were higher among PCPs and staff in rural than urban facilities. Differences between rural and urban facilities may present in other attributes, such as staffing, team effectiveness, or leadership commitment to enhance access to women’s health services, and/or other local resources not captured in the survey (Bergman et al., 2015; Chuang et al., 2017; Cordasco et al., 2016). In general, rural practices are less likely to have a sufficient volume of women veteran patients to warrant the establishment of women’s health clinics (Bean-Mayberry, Yano, Caffrey, Altman, & Washington, 2007), but rural providers and staff are likely to spend more time with patients as they have smaller caseloads (Cordasco et al., 2016). Because smaller rural sites may have fewer resources in general (e.g., less specialty care), their clinical workforce may take on multiple roles and work more closely with each other (Crump et al., 2019; Fischer et al., 2016; Yano, Goldzweig, Canelo, & Washington, 2006). The affinity among the rural workforce and their patients may have reinforced a more positive communication environment and shared positive gender attitudes toward women patients. Future investigation on facilitators of or barriers to promoting gender sensitivity and WH-PACTs should explore other possible drivers of rural and urban differences in VA facilities or clinics.

Limitations

Our study has several limitations. First, our analyses are based on only a single point in time, so we cannot make causal inferences. Second, our sample had a low response rate from VA PCPs and staff. Although our response rate is consistent with other studies of the VA primary care workforce (Nelson et al., 2014), views and experiences of the larger primary care workforce may be under-represented. Although we weighted our analyses to represent PCPs and staff at a geographically diverse group of 12 VA facilities, results may not generalize to VA facilities serving different proportions of women veterans or with different types of care arrangements, such as providers and staff in other specialty PACTs such as geriatrics or homeless PACTs, or those working in VAs with different care structures. Future research should explore other organizational factors related to gender sensitivity among these providers and staff.

Finally, gender sensitivity was generally quite high in our sample. Although our findings were generally consistent with those from an earlier 2001 study by Vogt et al. (2001), gender sensitivity may be higher in our sample than in the general population of VA PCPs and staff. Providers and staff who like to care for women patients may be more likely to participate in women’s health trainings and/or work in WH-PACTs, which could have biased the findings toward greater gender sensitivity. In addition, staff who did not have any prior women’s health experience did not respond to all of the gender sensitivity questions, which could also have influenced our results.

Implications for Policy and/or Practice

Strengthening gender sensitivity and a culture of care that is more accepting and understanding of the health needs of women veterans is important in reducing barriers to care related to how women veterans are treated and respected in the VA and improving their overall care experiences (Women’s Health Services, 2015; Yano et al., 2014). Our findings illustrate that women’s health training, opportunities to work with women veterans, and communication with other health professionals in the clinic are associated with higher levels of gender sensitivity. VA has already taken proactive steps in addressing some of these elements. Efforts to effectively recruit and train the primary care and women’s health workforce are also underway as part of building capacity for comprehensive women’s health services (Cordasco et al., 2015). However, availability and completion of training may vary by location and between PCPs and staff. Unlike the provisions for PCPs in WH-PACTs based on training and experience, staff working in WH-PACTs are not currently required to have specific women’s health related experience or training. Ensuring protected time to attend relevant trainings and leadership support and participation in trainings can promote training uptake (Fox et al., 2016). In addition, opportunities to work with other women’s health professionals and participate in clinical case discussions may be beneficial in increasing sensitivity for staff. Both trainings and hands-on experiences should be evaluated for their effectiveness on enhancing the workforce gender sensitivity.

Last, gender sensitivity is not just an issue in the VA (Celik et al., 2011). Lessons learned from the VA about the value of training and fostering more positive and supportive professional communication within and across care teams may help decrease gender disparities in other care settings as well.

Author Descriptions

Claire Than, PhD, MPH, is a research health scientist at the VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy. Her research focuses on health care access, quality, and outcomes for women.

Emmeline Chuang, PhD, is a health services researcher and Associate Professor, Health Policy and Management at the UCLA Fielding School of Public Health. Her research focuses on organizational factors affecting service access and quality of care for low-income women and children.

Donna L. Washington, MD, MPH, is Women’s Health Focused Research Area Lead, VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy, and Professor of Medicine at UCLA. Her research examines health care access, quality, and equity for women/vulnerable populations.

Jack Needleman, PhD, FAAN, is Fred W. and Pamela K. Wasserman Professor, Chair of the Department of Health Policy and Management, UCLA Fielding School of Public Health. He focuses on the impact of changing markets/public policy on quality and access to care.

Ismelda Canelo, MPA, is a Health Science Specialist, VA HSR&D Center for the Study of Healthcare Innovation, Implementation and Policy. She has 20 years of experience managing VA primary care and women’s health organizational and implementation research.

Lisa S. Meredith, PhD, is Senior Behavioral Scientist, RAND; Professor, Pardee RAND Graduate School; and Research Scientist, VA HSR&D Center of Excellence for Innovation, Implementation and Policy. Her research focuses on mental health services, health care provider behavior, and quality improvement.

Elizabeth M. Yano, PhD, MSPH, is Director, VA HSR&D Center for the Study of Innovation, Implementation and Policy, and Adjunct Professor, Health Policy and Management, UCLA Fielding School of Public Health. Her research focuses on implementation of research into evidence-based practice and policy.