Based on excerpts from Nosek, M.A., Howland, C.A., Hughes, R.B. (2001) The investigation of abuse and women with disabilities: Going beyond assumptions. Violence Against Women, 7 (4) 477- 499.
Our first investigation of violence against women with disabilities was delayed by almost insurmountable barriers we had never encountered in any of our previous studies. Problems with maintaining confidentiality, complying with mandatory reporting requirements, and ensuring the safety of both participants and staff were just a few of the challenges that honed our research skills. In an effort to advance the state of research on this topic, we offer our thoughts on the following precautions and special considerations:
Beware of problems in the research designs used in many Studies of Abuse and Disabilities
The few studies that have been published on abuse and disability can be criticized according to several principles of valid empirical research.
Essential constructs and variables important to statistical analysis were rarely defined operationally. There was a particular lack of distinction among emotional, physical, and sexual abuse. The problem of constructing definitions is confounded by overlap among the types; that is, sexual abuse is always accompanied by emotional and physical abuse.
The studies used nonstandardized measurement instruments and techniques.
Global references were made to the type of abuse, for example, emotional versus sexual; however, there was little attempt to document or categorize specific incidents by perpetrator.
Samples in these studies were generally quite heterogeneous in terms of disability type, gender, and age. Most of the studies that have been published examine samples of persons with developmental disabilities; however, they rarely distinguish those who have cognitive impairments or a combination of cognitive and physical impairments from those who have only physical impairments. For example, children with cerebral palsy or spina bifida are classified as having developmental disabilities, but often have no cognitive impairments.
Convenience sampling was predominant, such as examining clients of intervention programs or police reports, as opposed to representative, random, or population-based sampling.
Rarely were there comparison samples of persons without disabilities, or persons who have not received assistance from a particular program.
Statistical analyses rarely went beyond frequencies and measures of central tendency. Due to the heterogeneity of the samples, analyzing specific experiences of individuals with specific characteristics (such as sexual abuse among adult women with mental illness) would result in sub-samples too small to allow the use of more sophisticated analytic procedures.
Use literature-based definitions that distinguish emotional, physical, sexual, and disability-related abuse
Many different definitions exist for the various types of abuse. Virtually every battered women's program and every state human rights or social services law has its own set of definitions; however, there is no standard or consensus on which best suits the purposes of research. In its national study of women with physical disabilities, the Center for Research on Women with Disabilities used the following definitions. Emotional abuse was defined as being threatened, terrorized, or severely rejected, isolated, ignored, or verbally attacked (Finkelhor & Korbin, 1988; Claussen & Crittenden, 1991). Physical abuse was defined as any form of violence against her body, such as being hit, kicked, restrained, or deprived of food or water (Soeken, McFarlane, Parker & Campbell, 1997). Sexual abuse was defined as being forced, threatened, or deceived into sexual activities ranging from looking or touching to intercourse or rape (Soeken et al., 1997; Cole, 1984).
In addition to the types of abuse experienced by all women, women with physical disabilities are sometimes abused by withholding needed orthotic equipment (wheelchairs, braces), medications, transportation, or essential assistance with personal tasks, such as dressing or getting out of bed. The large majority of persons who have activity limitations depend on family for personal assistance. Whether or not the person providing the assistance is the perpetrator of abuse, the woman with a disability may perceive that this is her only option, that no one else would take care of her, and that abuse is the price she must pay for survival.
Many of the incidences of abuse reported to us during our national study occurred in medical settings and were committed by medical professionals. This confounds the definitional dilemma by introducing the concept of medical malpractice. Certain acts are clearly abusive and punishable by law, such as rape by an orderly. Others, however, are not so easily categorized, such as physical touch as part of psychotherapy, stripping for examination, or restraints for patients who are agitated. It is sometimes difficult to draw the line between actions that are sanctioned as medically necessary and actions that in any other context would be considered abusive.
Use population-based sampling methodologies
Identification of women who are in abusive situations or who have a history of abuse is highly problematic. By sampling only those women who have sought treatment, researchers are able to examine only a very small portion of the population. These are generally the women whose situation has reached a crisis stage and who have been able to garner the courage and resources to ask for help. Women who feel trapped in their abusive situation and do not feel safe asking for help, who are unaware of helping resources, or whose emotional attachment to the perpetrator keeps them from leaving, would not be included in samples constructed using this method. This bias in sampling is even more severe for women with disabilities, because they have far greater barriers to overcome in getting to battered women's programs and accessing their services.
Identification of abused women through police reports has been a sampling method used quite often in general abuse research. This is problematic for studies of disability and abuse in that there is no mechanism to report disability status in the record keeping systems of law enforcement. Even if there were a section on police forms for reporting disability, it would put the burden of labeling on police officers. Only the most obvious disabilities would be correctly classified, such as blindness, deafness, amputation and wheelchair use. It would be very difficult to provide every police officer with enough training to be able to recognize the wide variety of disabilities, especially the less visible ones such as mental illness, mental retardation, learning disabilities, early multiple sclerosis, or milder forms of cerebral palsy. Although gross indications of disability in crime statistics would be better than no indication, it would take considerable effort to establish feasible reporting systems.
Population-based sampling is generally preferred in empirical research. Under ideal circumstances, all women who meet pre-established criteria within a given geographical area would be recruited for the study and then screened for past or current abuse. Strategies for recruiting women with disabilities include placing announcements about the study in newspapers (including neighborhood publications and foreign language print media), and on radio and television; posting signs in doctors' offices, clinics, offices of social service agencies, and community recreation centers; and sending announcements to organizations of and for people with disabilities. Because these methods of recruitment have the potential to reach women regardless of their abuse status, they allow for random assignment to intervention groups. By using these strategies, researchers can conduct comparisons of women who screen positive for abuse to those who screen negative for abuse.
Under some circumstances, in tightly controlled studies, it may be acceptable to ask physicians or counselors to refer women who are known to have a history of abuse for participation in a study. This technique is more appropriate for representative sampling used in qualitative studies than for random sampling used in quantitative studies. The burden of identification is placed on the referring physician or counselor; however confirmatory screening can be conducted after recruitment.
Secure informed consent
One of the key elements in receiving approval from an institutional review board for the use of human subjects in research is the guarantee of securing informed consent. For studies that will be recruiting women who may have cognitive impairments or whose primary language is not English, this step is particularly important.
When abuse is the topic of the research, confidentiality in securing informed consent is critical. In our studies involving face-to-face recruiting and interviewing in a public facility, such as a clinic or a recreation center, women who express interest in participating in the study are escorted to a private area by one or two project staff members. Anyone accompanying her is asked to remain in the waiting room. She is asked if she has any objection to audio taping the interview for data validation purposes only. If she objects, only written records are made. The woman is assured that complete confidentiality will be maintained and that she will be identified in the study only by means of a code number.
Each participant is provided with a name and number to call if additional information is required. Staff read her the informed consent form describing the study and indicating that the woman understands the purpose of the study, the explicit nature of the questions to be asked, that participation is totally voluntary, and that she has the right to discontinue participation at any time with no penalty to her and no loss of services that she may be receiving at one or more of the facilities that assisted in recruitment.
She is also told at this time that she will receive a cash payment in a specified amount for completing the interview. Cash payments eliminate the possibility of outside discovery of study participation through institutional records such as check requests. Indication of consent (oral via tape recording or written) must be secured from each participant before proceeding with the interview.
In states that have strict requirements for reporting abuse to law enforcement or social service authorities, securing informed consent is more difficult. In this case, women should be informed that if they are identified as being in an abusive situation, their names will be given to an agency and someone will contact them to determine their safety. These reporting requirements have the effect of screening out women who are unwilling to put themselves at such risk, thereby seriously biasing a sample.
We have had success in securing permission from our Institution Review Board for the Protection of Human Subjects to maintain strict anonymity in recruitment. In our most recent study we recruited 511 women with physical disabilities, primarily through public chronic care clinics, and screened them for abuse. We never asked their names, allowed them to sign the informed consent form with an "X", and paid them a participant honorarium in cash immediately after the interview. Since we did not have knowledge of their identity, we could not report their names if they screened positive for abuse. However, we were prepared to assist them in completing a report and offered each woman referral information to organizations that offer abuse counseling and other services.
Our project staff were trained to offer this information with the awareness that some women would decline for various reasons such as stigma associated with abuse or potential risk of discovery. Likewise, we offered them the project director's business card after concluding that the benefits outweighed the potential risks to the women. In one instance, a mental health professional contacted the project director to determine the reason for disseminating that particular type of information to her client, indicating a concern that the information was associated with uncovered abuse.
A woman participating in studies about abuse should be assured that complete confidentiality will be maintained and that she will be identified in the study only by means of a code number. She should be offered the option of having her responses recorded only in writing by a staff person and not recorded on audio tape. All written and taped evidence of informed consent containing any identifying information about participants should be stored separately from responses to the interview in a locked file cabinet.
Staff should record participants' answers on a computerized version of the interview protocol backed up by an audio tape recording. Both the recordings and diskettes containing responses to the questionnaires should be kept in a locked cabinet separate from the consent tapes. Staff should be trained and carefully monitored to ensure that no references to participants' real names are used within the office or in any communication outside of the office.
Install safety measures to protect both the study participants and the project staff from retaliation by the perpetrator
Asking questions about abuse may trigger very strong emotions in some women. Referral to counseling should always be a feasible option for women who are disturbed by the interview process. Collaboration with a battered women's program or other type of counseling resource is valuable for making such referrals. A mental health professional with violence-related clinical experience and expertise should be available for consultation, referral, or crisis intervention at all times that the interviews are being conducted.
For many women, responding to questions about abuse may be more than emotionally disturbing; it may put them at risk for further abuse. Some abusive spouses go to excessive lengths to monitor their wives' activities and conversations. They may install tapes and caller identification devices on phones in the home. They forbid their wives from contacting friends and service providers, or insist on being present during every interaction. It is common for them to search their wives' purses and belongings. Finding a name card, a brochure, or any other identifying information about an abuse program could trigger a battering incident. Enforcing isolation is one of abusers' main tactics for protecting themselves.
Participants should be asked about the safest means for making contact. It may be by phone at certain times of the day, at certain locations, or through a relative or friend. If phone is the preferred means of contact, project staff should establish a system of code words that can be used if the abuser enters the room, if children are within listening range, or if the situation becomes dangerous in any other way. Mailing materials, including descriptions of the study or questionnaires, to a participant's home should generally be avoided.
Project staff may also be at risk for violence. If an abuser finds a name card or traces a call to the project office, he or she may stalk or commit violent acts in retaliation. It is not safe to think that a study focusing on disability is exempt from this type of risk. Staff who visit women in their homes should travel in pairs and be trained in how to recognize and escape potentially violent situations. We experienced a situation in which a perpetrator answered the door and project staff had to implement risk management procedures. Langford (2000) recommends that investigators do not leave the interview site with any of the participants and that the interviews are conducted in public places. It is advisable for all project staff, from principal investigators to receptionists, to receive training from a local battered women's program on safety measures that should be taken when interacting with women in abusive situations.
Take special efforts to ensure the inclusion of women with disabilities from minority backgrounds
Applications for funds from the Public Health Service, including the National Institutes of Health and the Centers for Disease Control, as well as the Department of Education, now require a section on how minorities and women will be involved in all proposed research. A good foundation for involving underrepresented groups is to take steps to ensure that members from these groups are included among your project investigators, staff, consultants, and advisors. In recruiting applicants for any vacant positions, outreach should be conducted to women, persons with disabilities, and minorities.
Special efforts should be taken in all projects to ensure the inclusion of women from minority backgrounds as participants. Collaboration with programs that serve a high percentage of minority women can be helpful in implementing recruitment strategies. The likelihood of minority women agreeing to participate in and complete a study is determined in large part by the method used for data collection. It is documented in the literature and confirmed by our research experience that women from minority backgrounds are less willing to complete lengthy written questionnaires. In order to eliminate this selection bias against minority participation, studies should be designed using methods that are more successful with minority populations, such as the personal interview.
It is helpful to use interviewers who share characteristics with the participants, including having a disability and being of minority background. To the extent possible, interview protocols and project materials should be translated into Spanish for studies being conducted in areas in which large numbers of first-generation Hispanics reside.
The Office of Research on Women's Health has issued recommendations on the recruitment and retention of women in clinical studies (National Institutes of Health, 1995) by involving representatives of the study population in planning the study, and designing culturally sensitive data collection materials, information materials, consent forms, and research methods, including qualitative interview methodology. Materials should be concise and at an appropriate reading level. Efforts should be taken to ensure that participants understand the purpose of the study and the benefit to them. Investigators should work toward developing a mutual relationship of trust, respect, and honesty, reinforcing their importance to the study and their personal worth. The project should provide follow up to participants and their community once the study is completed by sharing information about the results and benefits. Written authorization should be obtained from the women before initiating follow-up contacts or mailing study results to the home.
Use appropriate, validated, disability-sensitive screening instruments
Screening for abuse in all clinical and counseling settings is the best strategy for offering women with disabilities a way to get help to resolve violence in their lives. CROWD has developed a simple, brief, and useful tool for opening the door to conversations about abuse and violence. It consists of four items, the first two were developed and tested previously in a general population of women, the second two address abuse that is specifically related to having a disability. A copy of this screening instrument is available.
Understand the legal requirements for reporting abusive incidents
Whenever questions about abuse are asked, issues of reporting requirements arise (Hyman, Schillinger & Lo, 1995). Some states have very strict requirements placed on all persons to report observed abuse to the local criminal justice agency or the state social service agency. Other states place this requirement only on certain professions, such as counseling, medicine, and law enforcement.
Researchers who have the dual status as clinician and investigator face ethical dilemmas when the mandatory reporting introduces a risk for the woman who is in potential danger of retaliatory violence (Langford, 2000). Some statutes have separate requirements for persons with disabilities and older persons; however, these are often restricted to institutional settings.
Although reporting abuse of adults is optional in many states, reporting of children who are also being abused by the perpetrator who is battering their mother is always mandatory. While obtaining informed consent from the women, the project staff should emphasize that any information related to the abuse of children will be reported.
There is no uniform federal standard. Researchers should be knowledgeable about state mandatory reporting requirements, consult with the appropriate protective authorities to resolve dilemmas, and address such issues in their application for approval from institutional review boards for the use of human subjects.
Implement abuse studies in clinical settings
We recommend recruitment through general or disability-related clinics as an effective strategy for recruiting representative or population-based samples. The following are some observations from our experience recruiting and conducting abuse screening interviews with 511 women with physical disabilities at public and other specialty clinics. We have observed several challenges to effective and efficient recruitment.
1. There are time delays related to busy medical staff who may or may not have the interest and willingness to refer women to an abuse study or who may fail to identify a woman as having a physical disability.
2. There is generally a lack of private and accessible space arrangements that invite participant comfort and foster a sense of safety in discussing sensitive issues, such as when the examining rooms offered for the interviews are not easily accessible for women using wheelchairs, especially when both participant and project staff use wheelchairs.
3. Raising unfulfilled expectations for women who are referred by medical staff but are ineligible for the study can be minimized by asking staff to use a pre-enrollment eligibility checklist.
4. Disability-related fatigue discourages women from participating in an interview after completing a lengthy physician consultation, therefore recruitment is more successful when initiated while women are waiting for their appointments.
5. Some recruitment strategies can be experienced as coercive and could re-victimize a woman who has experienced violence.
6. The women's first priority is to obtain a physician consultation, which is their reason for coming to the clinic. Even though they may express interest in the study, they may end the interview before it is finished when they are called in to see the physician.
7. Cultural differences related to the use of the word "disability" appear to affect recruitment as some women meeting the disability eligibility requirements do not necessarily identify themselves as having a disability but prefer to use the word "condition."
In addition to the recruitment issues, we have identified several pitfalls related to conducting interviews in clinics.
1. Time limitations inherent in a brief screening interview may reduce essential rapport, inhibiting the full disclosure of abuse-related information, especially in cases where the abuse is being disclosed for the first time.
2. Cultural differences demand that the project staff respect the woman's refusal to reveal certain personal information that may be construed as a violation of family loyalty.
3. Repeated use of the word "abuse" can introduce trauma and feelings of dis-empowerment, whereas substituting words such as "the incident" or "the event" appears to be less invasive.
4. Dilemmas related to the participants' perception of the interviewer as "expert" or "peer" must be resolved before beginning the study as women may expect that an interviewer in the clinic setting will provide expert informational or support assistance. Project staff must be trained to adhere to their assigned roles, to recognize their limitations, and to offer appropriate referrals for psychological support and other needs. Whether or not the woman conducting the interviews has a disability may create a different dynamic in her relationship with the participant. Able-bodied interviewers can create psychological distance when interviewing a woman with a disability. Interviewers with disabilities, on the other hand, may establish a peer relationship more quickly, but may digress by discussing problems they have in common.
5. Inexperienced interviewers must be trained to refrain from overreacting to abuse involving bizarre or otherwise inflammatory circumstances, asking leading or suggestive questions (those beginning with "what" or "when" are examples of non-leading questions), physically touching a woman who has not first extended the gesture, and assuming gender issues such as the woman's sexual orientation or the gender of the perpetrator.
Training programs are needed in primary care settings to improve the identification and management of patients in violent situations (O'Leary, 1999), including programs specifically targeting disability-related violence. Given the evidence that psychological abuse usually precedes physical abuse and may exert more adverse consequences (Thompson et al., 1998), these training programs should include the assessment of psychological abuse as well as physical and sexual abuse and violence.
Agonito, R. (1993). No more "nice girl": Power, sexuality, and success in the workplace. Holbrook, MA: Bob Adams Publishing.
Andrews, A. B., & Veronen, L. J. (1993). Sexual assault and people with disabilities. Special issue: Sexuality and disabilities: A guide for human service practitioners. Journal of Social Work and Human Sexuality, 8(2), 137-159.
Ansello, E. F. (1996). Causes and theories. In L. A. Baumhover & S. C. Beall (Eds.), Abuse, neglect, and exploitation of older persons: Strategies for assessment and intervention. (pp. 9-29). Baltimore, MD: Health Professions Press.
Belsky, J. (1980). Child maltreatment: An ecological integration. American Psychologist, 35(4), 320-335.
Bowe, F. (1984). Women with Disabilities in America: A Statistical Report Drawn from Census Bureau Data. Washington, D.C.: President's Committee on Employment of the Handicapped.
Briere, J., & Runtz, M. (1988). Symptomatology associated with childhood sexual victimization in a nonclinical adult sample. Child Abuse & Neglect, 12, 51-59.
Briere, J., & Zaidi, L. Y. (1989). Sexual abuse histories and sequelae in female psychiatric emergency room patients. American Journal of Psychiatry, 146, 1602-1606.
Browne, A., & Finkelhor, D. (1986). Impact of child sexual abuse: A review of the research. Psychological Bulletin, 99, 66-77.
Claussen, A. H., & Crittenden, P. M. (1991). Physical and psychological maltreatment: Relations among types of maltreatment. Child Abuse & Neglect, 15, 5-18.
Cole, S.S. (1984). Facing the challenges of sexual abuse in persons with disabilities. Sexuality and Disability, 7(3/4), 71-88.
Courtois, C. A., & Watts, D. C. (1982). Counseling adult women who experienced incest in childhood or adolescence. The Personnel and Guidance Journal, 60, 275-279.
Cunningham, J., Pearce, T., & Pearce, P. (1988). Childhood sexual abuse and medical complaints in adult women. Journal of Interpersonal Violence, 3, 131-144.
Danek, M. M. (1992). The status of women with disabilities revisited. Journal of Applied Rehabilitation Counseling, 23(4), 7-13.
Faria, G., & Belohlavek, N. (1984). Treating female adult survivors of childhood incest. Social Casework, 65, 465-471.
Farmer, A., & Tiefenthaler, J. (1997). An economic analysis of domestic violence. The Review of Social Economy,55(3), 337-358.
Fine, M ., & Asch, A. (Eds.). (1988). Women with disabilities: Essays in psychology, culture, and politics. Philadelphia, PA: Temple University Press.
Finkelhor, D., Araji, S., Baron, L., Browne, A., Peters, S. D., & Wyatt, G. E. (1986). Sourcebook on child sexual abuse. Newbury Park, CA: Sage.
Finkelhor, D., & Korbin, J. (1988). Child abuse as an international issue. Child Abuse & Neglect, 12, 3-23.
Gordy, P. L. (1983). Group work that supports adult victims of childhood incest. Social Casework, 64, 300-307.
Grunfeld, A. F., Larsson, D. M., MacKay, K., & Hotch, D. (1996). Domestic violence against elderly women. Canadian Family Physician, 42(1), 1485-1493.
Healing the shame that binds you (Cassette Recording No. 1-55874-043-0). (1989). Bradshaw, J. S. Deerfield Beach, FL: Health Communications.
Hyman, A., Schillinger, D., & Lo, B. (1995). Laws mandating reporting of domestic violence: Do they promote patient well-being? Journal of the American Medical Association, 273(22), 1781-1787.
Kearney-Cooke, A. (1988). Group treatment of sexual abuse among women with eating disorders. Women and Therapy, 7(1), 5-21.
Kellerman, A., & Mercy, J. (1992). Men, women, and murder: Gender-specific differences in rates of fatal violence and victimization. Journal of Trauma, 33(1), 1-5.
Kreigsman, K. H., & Bregman, S. (1985). Women with disabilities at midlife: [Special issue: Transition and disability over the life span]. Rehabilitation Counseling Bulletin, 29(2), 112-122.
Langford, D. R. (2000). Pearls, pith, and provocation: Developing a safety protocol in qualitative research involving battered women. Qualitative Health Research, 10(1), 133-142.
McFarlane, J., Parker, B., Soeken, K., & Bullock, L. (1992). Assessing for abuse during pregnancy: Severity and frequency of injuries and associated entry into prenatal care. Journal of the American Medical Association, 267(23), 3176-3178.
Mullen, P., Romans-Clarkson, S. E., Walton, V. A., & Herbison, G. P. (1988). Impact of sexual and physical abuse on women's mental health. Lancet, 1, 841-845.
Murphy, P. A. (1992). Taking an abuse history in the initial evaluation. National Association of Rehabilitation Professionals in the Private Sector, 7(5), 187-190.
Murphy, P. A. (1993). Making the connections: Women, work, and abuse. Orlando, FL: Paul M. Deutsch Press, Inc.
National Institutes of Health. (1995). Recruitment and retention of women in clinical studies: A report of the workshop sponsored by the Office of Research on Women's Health. Washington, DC: National Institutes of Health.
Nosek, M. A. (1996). Sexual abuse of women with physical disabilities. In D. M. Krotoski, M. A. Nosek, & M. A. Turk (Eds.), Women with physical disabilities: Achieving and maintaining health and well-being. (pp. 153-173). Baltimore, MD: Paul H. Brookes.
Nosek, M. A., Howland, C. A., & Young, M. E. (1997). Abuse of women with disabilities: Policy implications. Journal of Disability Policy Studies, 8(1-2), 157-176.
O'Leary, K. D. (1999). Psychological abuse: A variable deserving critical attention in domestic violence. Violence and Victims, 14(1), 3-23.
Ratican, K. L. (1992). Sexual abuse survivors: Identifying symptoms and special treatment considerations. Journal of Counseling and Development, 71(1), 33-38.
Rothstein, M. A. (1997). Family violence and the health care system in Texas. Houston, TX: Health Law and Policy Institute, University of Houston Law Center.
Schaller, J., & DeLaGarza, D. (1995). Issues of gender in vocational testing and counseling. Journal of Job Placement, 11(1), 6-14.
Soeken, K., McFarlane, J., Parker, B., & Campbell, J. C. (1997). The abuse assessment screen: A clinical instrument to measure frequency, severity and perpetrator of abuse against women. In Soeken, K., McFarlane, J., Parker, B., & Campbell, J. C. (Eds.), Beyond diagnosis: Health care advocacy for battered women and their children. Thousand Oaks, CA: Sage.
Soeken, K., McFarlane, J., Parker, B., & Campbell, J. C. (Eds.). (1997). Beyond diagnosis: Health care advocacy for battered women and their children. Thousand Oaks, CA: Sage.
Thompson, R. S., Meyer, B. A., Smith-DiJulio, K., Caplow, M. P., Maiuro, R.D., Thompson, D.C., Sugg, N.K., Rivara, F.P. (1998). A training program to improve domestic violence identification and management in primary care: Preliminary results. Violence and Victims, 13(4), 395-410.
U.S. Bureau of the Census. (1989). Labor force status and other characteristics of persons with a work disability: 1981 to 1988 (Current Population Reports Series P-23, No. 160). Washington, DC: U.S. Government Printing Office.
Womendez, C., & Schneiderman, K. (1991). Escaping from abuse: Unique issues for women with disabilities. Sexuality and Disability, 9(3), 273-280.
Young, M. E., Nosek, M. A., Howland, C., Chanpong, G., & Rintala, D. H. (1997). Prevalence of abuse of women with physical disabilities. Arch Phys Med Rehabil, 78(12 Suppl 5), S-34-S-38.