Disparities in health outcomes are one metric used for assessing progress toward health equity. One of the goals of Healthy People 2020 is for the nation to “achieve health equity, eliminate disparities and improve the health of the people.”1 The ideal environment necessary for health equity—in which everyone has a fair and just opportunity to achieve optimal health—requires the removal of a broad range of barriers in policies, laws, systems, environments, and practices that reduce patients' access to the opportunities and resources needed to be healthy.2 In an analysis over a 10-year period from 1991 to 2000, medical advances saved 176,633 lives but 886,202 deaths were attributed to racially associated health disparities.3
This article reports the results of a quality improvement intervention designed to highlight professional behaviors that are known to influence health disparities in the hope that participants would implement ideal behaviors more frequently. The quality improvement program covered five broad categories:
- Race and ethnicity
- Sexual orientation and gender identity
- Literacy (including numeracy)
- Physical, sensory, and cognitive disabilities
- Economic and geographic factors.
These five categories were chosen based on their prominence in the medical literature, availability of subject matter experts, and the ability for measurable interventions during a chart review.4,5
Race and ethnicity are two forms of social classifications used in the United States.6 The official racial and ethnic classifications used by federal government and applied to health disparities are American Indian or Alaska Native, Asian, black or African American, Hispanic or Latino, Native Hawaiian or Pacific Islander, and white.7 Health inequities associated with specific races and/or ethnicities exist across medical care, resulting from healthcare systems issues, legal and regulatory factors, and clinician bias, including stereotyping, prejudice, and discrimination, whether intentional or not.8-10 Studies show that health disparities associated with race and ethnicity persist even after controlling for socioeconomic factors.9,11,12
Health disparities can occur to some sexual and gender minority patients if healthcare providers make decisions based on assumptions that the patient is cisgender and heterosexual. Knowing about patients' sexual attraction, behavior, and identity, gender identity and assigned sex at birth, as well as having an adequate understanding of this terminology, is essential for providing optimal care to sexual and gender minority patients.13-17 The CDC found that a two-question system—asking patients their current gender identity and assigned sex at birth—resulted in a 64% increase in identifying transgender patients, compared with only asking one question.17
Health information literacy, or the ability to understand, process, and apply basic health information needed to make healthcare decisions, affects all aspects of healthcare. Health literacy refers to a constellation of reading and numeracy skills (for example, making calculations or understanding percentages) required to function in the healthcare environment.1,4,18 Optimal health literacy requires a combination of individual and system-level factors and environments including the individual, the media, government and education entities, and the healthcare system.19 Low health literacy can result in medication errors, hospital readmissions, lack of preventive care, increased use of EDs, and poorer health outcomes overall.20 Health literacy is not the same as functional literacy (the ability to read), and can affect patients with a wide range of education levels, backgrounds, and socioeconomic statuses, although patients of lower socioeconomic status, and particularly people of color, are disproportionately affected by low health literacy.21
Physical, sensory, cognitive, economic, and geographic barriers also can interfere with health outcomes. Patients who live in a “food desert” are unable to obtain healthful foods, increasing the incidence of obesity, hypertension, and diabetes.22,23 Physical disabilities inhibit patients' ability to obtain medical care, and cognitive impairments reduce the likelihood of them comprehending health management.22,23
Participants in this study used Outside the Box: Reducing Health Disparities (http://aanpa.org/outsidethebox.html), a quality improvement program offered to PAs nationwide that highlights health disparities in five broad domains. Participants chose an intervention to address the specific findings in their own practice. This article reports on the first 2 years of data collected.
METHODS
Procedure
The study was granted institutional review board (IRB) exemption from Chesapeake IRB. All participants completed a system examination survey about their practice. Five domains that commonly experience health disparities (Table 1) were grouped into a series of 11 preintervention assessment questions that were displayed electronically to the participants (Table 2). Participants answered this series of 11 questions for 10 randomly chosen charts of previously seen patients. Participants were instructed to “choose 10 charts of established patients at random and answer the following questions based on the data contained in the chart.” Random selection was defined by the participant without intervention from the researchers. There was no outside review of the blinded charts and no attempt to link chart data to individual patients.
;)
TABLE 1.: Results of system examination answered by respondents about their practiceEach domain contains multiple questions
TABLE 2.: Example questionnaire for one patient chart (ideal response is boldfaced)
Following completion of the preintervention assessment, participants were given feedback for each domain. When 50% or more charts were “ideal” for a given domain, participants saw a message stating “you're doing pretty good!” When fewer than 50% were “ideal” for a given domain, participants saw “this could be better!”
Participants were then given access to educational materials on each of the five domains. The Outside the Box toolkit included 2 to 4 pages of educational materials on each of the five domains, a chart review tool to assess and evaluate the participant's practice, a frequently asked questions (FAQ) sheet, and instructions on choosing interventions. Examples of potential interventions participants could use were noted in the FAQs, although participants were told they could develop their own intervention to improve their preintervention assessment scores.
For the postintervention assessment, 30 days after implementation of the intervention, participants were sent a link with the same instructions as provided for the preintervention assessment and were asked to answer the same 11 questions from 10 randomly selected charts that were different from the preintervention assessment charts. Following electronic submission of the postintervention assessment, feedback from preintervention scores compared with postintervention scores was displayed to the participant.
For the qualitative aspect of the program, participants were asked to evaluate their experience with the Outside the Box program and were given an open-ended response prompt to report how their thinking about health disparities in relation to patient care was altered as a result of the activity. Participants also identified the intervention(s) used to improve their practice from preintervention to postintervention.
Participants
The 102 participants in this quality improvement intervention were identified from a convenience sample of PAs who voluntarily selected this activity as a practice improvement CME. Participants were recruited online and at professional meetings including the American Academy of PAs (AAPA) annual conference.
Analysis
Descriptive and inferential statistics were performed using SPSS 24. Descriptive statistics for the system examination were calculated based on percentages across the sample. Participant responses to questions in each domain were coded as ideal or not ideal and percentages of ideal behavior were calculated for each domain at the preintervention and postintervention assessments.
Paired-samples analyses were conducted on preintervention assessment and postintervention scores for each of the five domains. In addition to significance testing, Cohen d was calculated for each effect as a measure of the magnitude of the effect in pooled standard deviations: although an effect may reach statistical significance, it may be trivial and not meaningful (for example, d < 0.2), so effect size must be considered in an assessment. For this analysis, we considered d < 0.2 as trivial or nonmeaningful, d = 0.2 as small, d = 0.5 as medium, d = 0.8 as large, and d = 1.2 as very large, in line with previous interpretations of Cohen d.24,25
RESULTS
Participants
The average experience in clinical practice for the cohort was 8.85 years (median, 6 years). The average number of patients seen per week by the cohort was 67.99 (median, 60). Participants were from 26 different specialties as defined by AAPA, and from 27 states. The largest concentration represented among the cohort was primary care (31.1%) followed by surgical specialties (29.1%), internal medicine subspecialties (13.6%), emergency medicine (6.9%), and others (15.3%). Four percent did not specify practice type. Within surgical specialties, orthopedics was the most common. Most participants were from the South and the West.
System examination scores
To ascertain present practice in regard to the five domains evaluated, PAs were asked to report systemwide procedures at their practice on patient intake, assessment, and education before any preassessment, intervention, or postassessment (Table 1).
Pre- and postintervention assessment
In each of the five domains, statistically significant improvements in the percentage of ideal behaviors were reported from preintervention assessment to postintervention assessment, and effect sizes were medium to large in each domain. The exception was the economic and geographic domain, which showed a smaller, although still statistically significant improvement (Table 3).
TABLE 3.: Pre- and postintervention scores for each domain
Specifically, for sexual orientation and gender identity, before participation in the Outside the Box intervention, participants reported ideal behavior 11.77% of the time, compared with 29.81% of the time postintervention, t(102) = 8.40, P < .001, d = 0.84. For race and ethnicity, a preintervention score of 9.56% improved to 30.93% at postintervention, t(102) = 7.43, P < .001, d = 0.81. For physical, cognitive, and sensory disabilities, a preintervention score of 45.15% improved to 64.22% at postintervention, t(102) = 5.47, P < .001, d = 0.55. For literacy and numeracy, a preintervention score of 15.59% improved to 40.59% at postintervention, t(102) = 6.91, P < .001, d = 0.72. Finally, the economic and geographic domain improved from 49.26% at preintervention to 55.10% at postintervention, t(102) = 4.50, P < .001, d = 0.19.
Participant open-ended responses were collected, but many responses lacked specificity regarding any one domain, preventing a systematic thematic analysis.
DISCUSSION
Regardless of the specific participant's intervention, improvements were documented in all five domains. The large effect sizes of the cohort completing the program took the developers of this program by surprise. When queried before the chart review, 37% of participants stated that their practice included asking about gender identity and assigned sex at birth separately, and 54% stated their practice had visual education materials available, with 39% reporting that patient materials were at or below a 5th-grade reading level and 81% stating that handouts were available in multiple languages. Based on these numbers, it would appear that practices had already implemented methods to reduce health disparities, but Outside the Box still identified gaps in participants' practices.
Ideal behavior in relationship to sexual orientation and gender identity and race/ethnicity showed a substantial improvement after the program (11.77% and 9.56% preintervention compared with 29.81% and 30.93% postintervention, respectively). Although most participants reported that their practices had low-literacy or multilanguage handouts, the improvement for literacy and numeracy issues was still significant after implementing the program (15.59% preimprovement compared with 40.59% postimprovement). The improvement for economic and geographic issues compared with physical, cognitive, and sensory issues was smaller, but still statistically significant (45.15% and 49.26% preintervention compared with 64.22% and 55.10% postintervention, respectively).
In the postintervention data collection, an open-ended written response was collected from the study cohort. Responses ranged from “I have become more aware and did not realize how many charts were missing this information” to “I assumed my patients knew more about their treatment plans and medications” to “affected my practice minimally, due to the fact that in orthopedics many of the health disparities do not apply to fracture treatment.” A substantial number of PAs reported that asking more questions of patients elicited more information, and this was true across specialties. One PA in pediatrics reported “I found out that my pediatric patients are more willing to answer questions about their sexual history when given a questionnaire. They were also more open with me when I asked for more information.” Another PA reported that colleagues realized they needed to “revise and reword our adolescent questionnaire to include specific questions about gender (for the teens to self-identify, rather than what parents have filled out for them) and sexual practices.” A PA in internal medicine reported “I was impressed with how many times the information I received after asking such questions from well-known patients was new and changed the direction of the care and health plan the patient received.”
Differential diagnoses were noted to be more inclusive. One PA in urgent care reported diagnosing a patient with “inclusion (chlamydia) conjunctivitis and another patient with gonorrhea pharyngitis. In both cases, their sexual history alerted me to getting a culture vs. making a presumptive diagnosis of antibiotic resistance and treating with a second-line antibiotic.” Another PA noted that patients whose diabetes was difficult to control actually had issues with understanding the titration directions given at the office visit. Weekly phone interventions by the nurse manager helped these patients bring their diabetes under control, reducing the incidence of end-organ damage and healthcare costs. Still another PA used the social worker for hospital discharged patients resulting in increased medication compliance, home care inclusion, and accessing the local Healthcare for the Homeless program, as needed. Without the push to ask patients for financial information, this PA reported that discharge paperwork was given to everyone without knowing whether the patient could afford to fill prescriptions or was able to arrange transportation to future appointments. The qualitative and quantitative data suggest that awareness is the first step, and PAs can make a difference in the health outcomes of the populations they serve.
This study demonstrates that Outside the Box can improve practice performance among PAs in the domains that are known to experience health disparities. Other studies also have shown that targeted quality improvement strategies can improve diagnostics and outcomes in populations that experience health disparities.26,27 We recommend further research in the use of quality improvement strategies for PAs and health disparities. Other recommendations for PA practice behavior that may address health disparities and bring improvement to care include education on health disparities and opportunities to mitigate health inequities, exploring personal unconscious or implicit associations through tools such as the implicit association test (https:/implicit.harvard.edu/implicit), and pursuing continuing medical education (CME) to enhance provider communication skills for health literacy, improve cultural competence, and improve the frequency of ideal practice behaviors.28 PAs should begin by thinking about each patient as a uniquely complex person; then take a good history including the patient's ethnicity or ancestral background, family history, sexual history, literacy level, and socioeconomic status to guide appropriate screenings; search for data on that population group and the disease of interest; and ask “What does this mean for my patient?”
LIMITATIONS
Limitations to this study included cohort size as well as potential selection bias. Participants who self-selected into the program may be more aware of biases and/or disparities or were more likely to change explicit behaviors. The data collected pre- and postintervention were not independently reviewed, and participant bias in reporting of results cannot be excluded. Without a control group of clinicians who did not participate in Outside the Box, we cannot exclude the possibility that changes in practice patterns were due to outside influences (articles, CME, or such) although due to the short timeline of the intervention, that is unlikely.
Strengths include moderate to large-sized statistical results, as well as the representativeness of the cohort in terms of geography and specialty practice. Improvements in scores were not associated with a specific type of intervention, such that scores improved on average although all cohort members did not target every domain for improvement.
Even with statistically significant large improvements in each domain, the postintervention ideal numbers were not reached across the full cohort. On the positive side, published data have shown that even a short-term quality improvement project can have significant long-term change in professional behavior.29 As health disparities continue to be a major factor in health outcomes, the success of simple quality improvement interventions such as Outside the Box show promise as a catalyst for change in PA practices.
REFERENCES
1. US Department of Health and Human Services. Healthy People 2020.
www.healthypeople.gov/2020/about/foundation-health-measures/Disparities. Accessed October 15, 2019.
2. Braveman P, Arkin E, Orleans T, et al. What is health equity? And what difference does a definition make?
www.rwjf.org/en/library/research/2017/05/what-is-health-equity-.html. Accessed October 15, 2019.
3. Woolf SH, Johnson RE, Fryer GE Jr, et al. The health impact of resolving racial disparities: an analysis of US mortality data.
Am J Public Health. 2004;94(12):2078–2081.
4. Centers for Disease Control and Prevention. Office of Minority Health and Health Equity. Strategies for reducing health disparities.
www.cdc.gov/minorityhealth/strategies2016/index.html. Accessed October 15, 2019.
5. Pomeranz H, Horvath T. Promoting diversity to eliminate health disparities.
J Physician Assist Educ. 2017;28(suppl 1):S43–S48.
6. Foster MW, Sharp RR. Race, ethnicity, and genomics: social classifications as proxies of biological heterogeneity.
Genome Res. 2002;12(6):844–850.
7. Office of Management and Budget. Office of Information and Regulatory Affairs. Revisions to the standards for the classification of federal data on race and ethnicity.
www.federalregister.gov/documents/1997/10/30/97-28653/revisions-to-the-standards-for-the-classification-of-federal-data-on-race-and-ethnicity. Accessed October 15, 2019.
8. FitzGerald C, Hurst S. Implicit bias in healthcare professionals: a systematic review.
BMC Med Ethics. 2017;18(1):19.
9. Smedley BD, Stith AY, Nelson AR. Institute of Medicine. Committee on Understanding and Eliminating Racial and Ethnic Disparities in Health Care.
Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: National Academies Press; 2003.
10. Elo IT, Beltrán-Sánchez H, Macinko J. The contribution of health care and other interventions to black-white disparities in life expectancy, 1980-2007.
Popul Res Policy Rev. 2014;33(1):97–126.
11. LeLacheur S, Straker H. Culture, diversity, race, and the standards: assessing and addressing the hidden curricula.
J Physician Assist Educ. 2011;22(2):34–37.
12. Braveman PA, Cubbin C, Egerter S, et al. Socioeconomic disparities in health in the United States: what the patterns tell us.
Am J Public Health. 2010;100(suppl 1):S186–S196.
13. Bruessow D, Poteat T. Primary care providers' role in transgender healthcare.
JAAPA. 2018;31(2):8–11.
14. Sobecki JN, Curlin FA, Rasinski KA, Lindau ST. What we don't talk about when we don't talk about sex: results of a national survey of U.S. obstetrician/gynecologists.
J Sex Med. 2012;9(5):1285–1294.
15. Althof SE, Rosen RC, Perelman MA, Rubio-Aurioles E. Standard operating procedures for taking a sexual history.
J Sex Med. 2013;10(1):26–35.
16. Bruessow D. Keeping up with LGBT health: why it matters to your patients.
JAAPA. 2011;24(3):14.
17. Institute of Medicine.
Collecting Sexual Orientation and Gender Identity Data in Electronic Health Records: Workshop Summary. Washington, DC: National Academies Press; 2013.
18. US Agency Healthcare Research Quality. AHRQ health literacy tools for use in pharmacies.
www.ahrq.gov/health-literacy/pharmhealthlit/tools.html. Accessed October 29, 2019.
19. Nielsen-Bohlman L, Panzer AM, Kindig DA Institute of Medicine Committee on Health Literacy.
Health Literacy: A Prescription to End Confusion. Washington, DC: National Academies Press; 2004.
20. Berkman ND, Sheridan SL, Donahue KE, et al. Low health literacy and health outcomes: an updated systematic review.
Ann Intern Med. 2011;155(2):97–107.
21. Scott TL, Gazmararian JA, Williams MV, Baker DW. Health literacy among Medicare enrollees in a managed care organization.
Med Care. 2002;40(5):395–404.
22. US Department of Health and Human Services. Healthy People 2020. Social determinants of health.
www.healthypeople.gov/2020/leading-health-indicators/2020-lhi-topics/Social-Determinants. Accessed October 29, 2019.
23. Norton JM, Moxey-Mims MM, Eggers PW, et al. Social determinants of racial disparities in CKD.
J Am Soc Nephrol. 2016;27(9):2576–2595.
24. Sawilowsky S. New effect size rules of thumb.
J Mod Appl Methods. 2009;8(2):467–474.
25. Cohen J.
Statistical Power Analysis for the Behavioral Sciences. 2nd ed. New York, NY: Lawrence Erlbaum Associates; 1977.
26. Anarella JP, Wagner VL, McCauley SG, et al. Eliminating disparities in asthma care: identifying broad challenges in quality improvement.
Am J Med Qual. 2017;32(6):598–604.
27. Nietert PJ, Ornstein SM, Jenkins RG, et al. The effect of ethnicity on outcomes in a practice-based trial to improve cardiovascular disease prevention.
Int J Equity Health. 2004;3(1):12.
28. Straker H. Reducing health disparities: every PA's obligation.
JAAPA. 2008;21(2):13.
29. Shaw-Gallagher M, Boyle R, Zuber K. Longitudinal survey of clinician behavior change in CKD management.
JAAPA. 2019;32(4):39–43.
