Insurance Discrimination Research Articles

Antidiscrimination Insurance Pricing: Regulations, Fairness Criteria, and Models

On the issue of insurance discrimination, a grey area in regulation has resulted from the growing use of big data analytics by insurance companies: direct discrimination is prohibited, but indirect discrimination using proxies or more complex and opaque algorithms is not clearly specified or assessed. This phenomenon has recently attracted the attention of insurance regulators all over the world. Meanwhile, various fairness criteria have been proposed and flourished in the machine learning literature with the rapid growth of artificial intelligence (AI) in the past decade and have mostly focused on classification decisions. In this article, we introduce some fairness criteria that are potentially applicable to insurance pricing as a regression problem to the actuarial field, match them with different levels of potential and existing antidiscrimination regulations, and implement them into a series of existing and newly proposed antidiscrimination insurance pricing models, using both generalized linear models (GLMs) and Extreme Gradient Boosting (XGBoost). Our empirical analysis compares the outcome of different models via the fairness–accuracy trade-off and shows their impact on adverse selection and solidarity.

Genetic discrimination by insurance companies in Aotearoa New Zealand: experiences and views of health professionals.

Genetic discrimination in insurance is a significant clinical, research and consumer issue. Recently, the Australian life insurance industry introduced a partial moratorium on the use of genetic test results. However, in Aotearoa New Zealand, both life and health insurers can still use genetic results legally to discriminate against applicants. We aimed to document experiences and concerns of New Zealand-based health professionals (HPs) around the potential misuse of genetic test results for insurance purposes. We administered an online survey to New Zealand HPs who discuss genetic testing with patients, their experiences regarding the use of genetic test results in insurance and views on regulation. Twenty-three New Zealand HPs responded, 15 of whom worked in genetics clinics, representing >60% of the total New Zealand clinical genetics workforce. Eleven respondents reported having patients who experienced adverse outcomes related to insurance based on genetic results. Respondents reported patients sometimes/often delayed (n=11) or refused (n=4) genetic testing due to insurance concerns. Over 80% of those who answered (n=17/21) believe insurers' use of genetic results should be legally regulated. New Zealand HPs have concerns about insurance companies using genetic test results in underwriting, including the effect on patients, and strongly believe government legislation is required.

Human Genetics Society of Australasia Position Statement: Genetic Testing and Personal Insurance Products in Australia.

The expansion of genetic and genomic testing in clinical practice and research, and the growing market for direct-to-consumer genomic testing has led to increased awareness about the impact of this form of testing on insurance. Genetic or genomic information can be requested by providers of mutually rated insurance products, who may then use it when setting premiums or determining eligibility for cover under a particular product. Australian insurers are subject to relevant legislation and an industry led standard that was updated in 2019 to introduce a moratorium on the use of genetic test results in life insurance underwriting for policies <AU$500K. The Human Genetics Society of Australasia has updated its position statement on genetic testing and life insurance to account for these changes and to increase the scope of the statement to include a wider range of personally-rated insurance products, such as life, critical care, and income protection products. Recommendations include that: providers of professional education involving genetics should include ethical, legal, and social aspects of insurance discrimination in their curricula; the Australian Government take a more active role in regulating use of genetic information in personal insurance; that information obtained in the course of a research project be excluded; insurers seek expert advice when making underwriting decisions regarding genetic testing; and engagement between the insurance industry, regulators, and the genetics profession be improved.

Perceived benefits and barriers to implementing precision preventive care: Results of a national physician survey.

Polygenic risk scores (PRS) may improve risk-stratification in preventive care. Their clinical implementation will depend on primary care physicians' (PCPs) uptake. We surveyed PCPs in a national physician database about the perceived clinical utility, benefits, and barriers to the use of PRS in preventive care. Among 367 respondents (participation rate 96.3%), mean (SD) age was 54.9 (12.9) years, 137 (37.3%) were female, and mean (SD) time since medical school graduation was 27.2 (13.3) years. Respondents reported greater perceived utility for more clinical action (e.g., earlier or more intensive screening, preventive medications, or lifestyle modification) for patients with high-risk PRS than for delayed or discontinued prevention actions for low-risk patients (p < 0.001). Respondents most often chose out-of-pocket costs (48%), lack of clinical guidelines (24%), and insurance discrimination concerns (22%) as extreme barriers. Latent class analysis identified 3 subclasses of respondents. Skeptics (n = 83, 22.6%) endorsed less agreement with individual clinical utilities, saw patient anxiety and insurance discrimination as significant barriers, and agreed less often that PRS could help patients make better health decisions. Learners (n = 134, 36.5%) and enthusiasts (n = 150, 40.9%) expressed similar levels of agreement that PRS had utility for preventive actions and that PRS could be useful for patient decision-making. Compared with enthusiasts, however, learners perceived greater barriers to the clinical use of PRS. Overall results suggest that PCPs generally endorse using PRS to guide medical decision-making about preventive care, and barriers identified suggest interventions to address their needs and concerns.

Perceived discrimination in primary care: Does Payer mix matter?

IntroductionPrevious literature has explored patient perceptions of discrimination by race and insurance status, but little is known about whether the payer mix of the primary care clinic (i.e., that is majority public insurance vs. majority private insurance clinics) influences patient perceptions of race- or insurance-based discrimination. MethodsBetween 2015-2017, we assessed patient satisfaction and perceived race- and insurance-based discrimination using a brief, anonymous post-clinic visit survey. ResultsParticipants included 3,721 patients from seven primary care clinics—three public clinics and four private clinics. Results from unadjusted logistic regression models suggest higher overall reports of race- and insurance-based discrimination in public clinics compared with private clinics. In mulvariate analyses, increasing age, Black race, lower education and Medicaid insurance were associated with higher odds of reporting race- and insurance-based discrimination in both public and private settings. ConclusionReports of race and insurance discrimination are higher in public clinics than private clinics. Sociodemographic variables, such as age, Black race, education level, and type of insurance also influence reports of race- and insurance-based discrimination in primary care.

Ethical, legal, and sociocultural challenges of genomic research in Jordan: a mixed methods study in patients with breast cancer with Jordanian–Palestinian heritage

BackgroundBreast cancer is a leading cause of morbidity and mortality in women worldwide. At King Hussein Cancer Center (KHCC), Amman, Jordan, 38% of breast cancer diagnoses are in women younger than 45 years, which might reflect a high prevalence of hereditary breast cancer. The aim of this study was to assess the contribution of germline mutations in BRCA1 and BRCA2 to breast cancer treated at KHCC. We also describe our experience in dealing with this delicate cultural issue in a tribal-based country with Palestinian–Jordanian family origins. MethodsPatients treated at KHCC with stage I–IV breast cancer, with a high-risk profile according to US National Comprehensive Cancer Network guidelines, were invited to participate. Patients were interviewed for consent and had full autonomy to decide whether they wanted to know their test result, inform their treating physician, or have a copy of the result placed in their charts. Charts were reviewed for clinical data and tumour pathology. A 10 mL blood sample was obtained for DNA extraction. A detailed three-generation family history was obtained. BRCA sequencing was done at the Myriad Genetics laboratory using BRCAnalysis. BRCA mutations were classified as deleterious, suspected deleterious, variant of uncertain significance (VUS), or favour polymorphism (harmless). The report received for each mutation result contained a detailed description of level of risk associated with the involved variant and estimated lifetime risk of breast and ovarian cancer, risk of second breast cancer or subsequent ovarian cancer, and risk of cancer to first, second, and third-degree family members. Patients were invited to an ad-hoc clinic with the treating physician for disclosure of results and counselling about estimated cancer risk to patient and relatives. Interviews with patients were summarised as text and qualitiative analysis for common themes was done. Challenges encountered during patient consent and disclosure of results were documented, as well as proposed solutions. Ethical approval was granted by the KHCC Institutional Review Board (no 11 KHCC 63). FindingsThe first 102 patients to meet inclusion criteria were invited to participate in the study, of whom 100 enrolled and provided written informed consent. Two patients declined to participate (a response rate of 98%). Six participants were residents of the West Bank and treated at KHCC, and 66 were from families of Palestinian origin residing in Jordan. The remaining patients were Jordanians. Median age was 40 years (range 22–75). 20 patients had deleterious mutations and seven had suspected deleterious mutations in BRCA1 and BRCA2 genes. Seven VUS were detected. All patients with BRCA1 and BRCA2 mutations had a significant family history of breast or ovarian cancer. Nevertheless, many patients with a significant family history (62 of 100) tested negative for BRCA1 and BRCA2 mutations. Many ethical dilemmas were highlighted during patient encounters, mainly relating to the guarantee of confidentiality and privacy in a tribal-based culture, fears of false reassurance, autonomy versus beneficence, and the bounds of physician–patient confidentiality when relatives are at genetic risk of cancer. Potential insurance, employment, and social discrimination implications were also addressed with the patients before testing and in more detail after disclosure of positive mutation results, which flagged a legal void in local genetic testing and data protection regulations. InterpretationOur cohort showed an important incidence of deleterious and suspected deleterious BRCA1 and BRCA2 mutations, suggesting that genetic testing should be discussed and offered to patients with high-risk features. Many high-risk patients tested negative for BRCA1 and BRCA2 mutations, therefore future studies should aim to test for mutations in other breast cancer susceptibility genes (eg, CHEK2, PALB2, and BRIP1). Running culturally sensitive genetic research in a country with limited resources is a challenging exercise and highlights the urgent need for guiding regulations. We recommend establishing a clinical cancer genetics programme, through which unaffected family members would benefit from early breast cancer screening and appropriate risk-reduction measures. FundingSister Institution Network Fund of MD Anderson Cancer Center and KHCC.

Comfort with Pharmacogenetic Testing Amongst Pediatric Oncology Providers and Their Patients.

Background: Pharmacogenetic (PGx) testing, a component of personalized medicine, aims to ensure treatment efficacy while reducing side effects and symptoms. Before this testing becomes routine in the pediatric oncology population, nurses need to understand the knowledge and concerns of providers, patients, and family members with regard to the timing, extent, interpretation, and incorporation of PGx testing. Methods: As part of a comprehensive PGx study (larger study) for children diagnosed with cancer, we surveyed providers and caregivers of children with cancer about their knowledge of and comfort with PGx testing. Caregivers who declined to participate in the larger PGx study were also asked to participate in the survey. Chi-square tests and a two-sample t-test were used to compare variables. Results: One hundred and two participants from the larger PGx study and 12 families who refused (response rate of 77% and 54%, respectively) as well as 29 providers (88%) completed surveys. Families not on the study were less interested in and comfortable with PGx results. Both groups were concerned about health or life insurance discrimination and payment. Providers would like support in ordering PGx testing and interpreting PGx. Discussion: Providers remain wary of most PGx testing, uncomfortable with interpreting and applying the results. Families are interested in the possibilities of personalized prescribing while worried about who has access to their child's genetic information. Further education on relevant tests for providers, including nurses, and the testing process for families, including details on privacy and sharing of genetic information, appear necessary.

EP507: Patients’ reasons for genetic testing refusal in cancer evaluations: An analysis of the literature and opportunities to educate

eP507: Patients’ reasons for genetic testing refusal in cancer evaluations: An analysis of the literature and opportunities to educate

Oral health inequality among Indonesian workers in South Korea: role of health insurance and discrimination factors

BackgroundThe health of migrant workers is becoming an important public health issue. Although there are an increasing number of migrant workers in Korea, the health status in migrant populations remains unknown. The aims of this study were (1) to evaluate the association between income and self-rated oral health (SROH), and (2) to assess the role of health insurance and self-perceived discrimination in the association between income and SROH among Indonesian migrant workers in Korea.MethodsInformation about self-reported income, SROH, coverage/utilization of health insurance (HI), living difficulties related to oral health (LDROH), oral health literacy (OHL), and discrimination were obtained from Indonesian migrant workers in Korea (n = 248). The main explanatory variable was income, and SROH was an outcome variable. Logistic regression analyses were performed controlling for age, gender, HI, LDROH, OHL, and discrimination. The paths from income to SROH were analyzed using the Partial Least Square-Structural Equation Model (PLS-SEM).ResultsAmong Indonesian migrant workers, the lower income group had the highest probability of a poor SROH compared to the higher income group. The variables showing a high explanatory power were discrimination among the low income group and HI among the middle income group. In PLS-SEM, the variables such as HI, LDROH, OHL, and discrimination contributed 11% to explaining the association between income and SROH.ConclusionA monotonic gradient was revealed among migrant workers according to the association between income and SROH. Discrimination and HI contributed to oral health inequalities.

Premium Based on ‘Like, Share and Post’: Use of Social Media Data in Life Insurance and Proxy Discrimination

Premium Based on ‘Like, Share and Post’: Use of Social Media Data in Life Insurance and Proxy Discrimination

Attitudes among parents of persons with autism spectrum disorder towards information about genetic risk and future health

Clinical relevance of genetic testing is increasing in autism spectrum disorder (ASD). Information about genetic risk may contribute to improved diagnostics, treatment and family planning, but may also be perceived as a burden. Knowledge about the families’ preferences with regard to genetic risk information is important for both health care professionals and policy makers. We investigated attitudes towards sharing information about genetic risk of ASD and knowledge about future health among parent members of the Norwegian Autism Association (N = 1455) using a questionnaire, and the relationships with parent and child characteristics, such as age, gender and ASD severity. Most preferred autonomy in deciding whom to inform about genetic risk of ASD (74.4%) and a minority supported extensive intra-familial disclosure of the genetic risk (41.1%). The majority agreed that it is an obligation to know as much as possible relevant for future health (58.0%) and only 51.7% agreed to a principle of a ‘right not to know’. In regression models, the attitudes were associated with opinions about benefits and harms of genetic testing (e.g., treatment, family planning, understanding of ASD pathology, insurance discrimination and family conflict). In sum, the findings show that most parents want to know as much as possible relevant for their children’s future health and keep their autonomy and intra-familial confidentiality about genetic risk information. Nearly half of the parents were not concerned with a “right not to know”. These attitudes can inform development of guidelines and bioethics in the age of genomic precision medicine.

Universal screening and trauma informed care: Current concerns and future directions.

There is a growing awareness of the prevalence and negative health effects associated with traumatic events, including childhood abuse and adversity, intimate partner violence, adult sexual assault, and exposure to combat and community violence. Health care systems have attempted to address this link by becoming trauma informed through universal trauma precautions and screening protocols. We review several clinical and methodological concerns associated with universal trauma screening in adult health care settings including: deciding which traumas to assess in which populations, integrating retrospective recall with current functioning to facilitate referrals, and guarding against adverse patient reactions and insurance discrimination. We outline potential implications for program development and future research including: adapting and refining screening tools, integrating patient preferences and privacy concerns into screening protocols, assessing resource limitations, and integrating public health advocacy into screening programs. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Informed choice and attitudes regarding a genomic test to predict risk of colorectal cancer in general practice

ObjectiveA genomic test to predict personal risk of colorectal cancer (CRC) that targets screening and could be feasibly implemented in primary care. We explored informed decision-making and attitudes towards genomic testing in this setting. MethodsA CRC genomic test was offered to 150 general practice patients with brief discussion of its implications. We measured informed choice about the test, consisting knowledge, attitudes and test uptake. Sixteen purposively-sampled participants were interviewed. ResultsOf 150, 142 (95%) completed the informed choice measure and of 27 invited, 16 (59%) completed an interview. 73% made an informed choice about the test. Interviews revealed that participants with inadequate knowledge on the informed choice scale still understood the gist of the test. While positive attitudes were most prevalent, some had concerns, and many were indifferent to the test. Positive attitudes included: that risk information could facilitate risk reduction; negative attitudes included: that risk results could cause worry and be used for insurance discrimination; indifferent attitudes included: that the test seemed benign and it was easy to do. ConclusionsOur study adds to the evidence that genomic tests for CRC risk do not pose significant concern to patients in community settings. Practice implicationsAs genomic tests become more prevalent, this study’s findings can be used to facilitate informed decision-making and ensure equitable access.

The Discriminating (Pricing) Actuary

The insurance industry is built on risk classification, grouping insureds into homogeneous classes. Through actions such as underwriting, pricing, and so forth, it differentiates, or discriminates, among insureds. Actuaries have responsibility for pricing insurance risk transfers and are intimately involved in other aspects of company actions and so have a keen interest in whether or not discrimination is appropriate from both company and societal viewpoints. This article reviews social and economic principles that can be used to assess the appropriateness of insurance discrimination. Discrimination issues vary by the line of insurance business and by the country and legal jurisdiction. This article examines social and economic principles from the vantage of a specific line of business and jurisdiction; these vantage points provide insights into principles. To sharpen understanding of the social and economic principles, this article also describes discrimination considerations for prohibitions based on diagnosis of COVID-19, the pandemic that swept the globe in 2020. Insurance discrimination issues have been an important topic for the insurance industry for decades and are evolving in part due to insurers’ extensive use of Big Data; that is, the increasing capacity and computational abilities of computers, availability of new and innovative sources of data, and advanced algorithms that can detect patterns in insurance activities that were previously unknown. On the one hand, the fundamental issues of insurance discrimination have not changed with Big Data; one can think of credit-based insurance scoring and price optimization as simply forerunners of this movement. On the other hand, issues regarding privacy and use of algorithmic proxies take on increased importance as insurers’ extensive use of data and computational abilities evolve.

Impact of medical mistrust (MM) on perceptions of tumor genomic profiling (TGP) among African American (AA) cancer patients: Application of perceptual mapping (PM).

e22527 Background: TGP is widely used to identify targetable mutations for precision cancer treatment and clinical trials. Many patients have poor understanding of TGP and are unaware of possible secondary hereditary risks. Lack of clarity regarding the relevance of informed consent and genetic counseling further magnify risks for patients. AA patients have lower genetic knowledge and health literacy and higher MM than Caucasian patients, making them especially vulnerable in the clinical setting. Perceptions of TGP in AA cancer patients have not been well-characterized. Methods: 120 AA pts from 1 suburban and 1 urban site (Fox Chase Cancer Center[FCCC] and Temple University Hospital[TUH]) were surveyed. A k-means cluster analysis using a modified MM scale was conducted; chi-square analysis assessed demographic differences. Perceptual mapping (PM)/multidimensional scaling and vector modeling was used to create 3-dimensional maps to study how TGP barriers/facilitators differed by MM group and how message strategies for communicating about TGP may also differ. Results: Data from 112 analyzable patients from FCCC (55%) and TUH (45%) were parsed into less MM (MM-L, n = 42, 37.5%) and more MM (MM-H, n = 70, 72.5%) clusters. MM-L and MM-H clusters were demographically indistinct with no significant associations by sex (p = 0.49), education (p = 0.3), income (p = 0.65), or location (p = 0.43); only age was significant (older = higher MM, p = 0.006). Patients in the MM-H cluster reported higher concerns about TGP, including cost (p &lt; 0.001), insurance discrimination (p &lt; 0.001), privacy breaches (p = 0.001), test performance/accuracy (p = 0.001), secondary gain by providers (p &lt; 0.001) and provider ability to explain results (p = 0.04). Perceptual mapping identified both shared and contrasting barriers between MM clusters (Table). Conclusions: More than 2/3 of AA patients comprised a MM-H cluster. Communication strategies should focus on concerns about family and how to discuss TGP with an oncologist. PM can identify distinct and shared information needs of vulnerable populations undergoing TGP. [Table: see text]

Opinions of adults affected with later-onset lysosomal storage diseases regarding newborn screening: A qualitative study.

Lysosomal storage diseases (LSDs) are a heterogeneous group of conditions causing substrate accumulation leading to progressive organ damage. Newborn screening (NBS) for several LSDs has become available in recent years due to advances in technology and treatment availability. While early initiation of treatment is lifesaving for those with infantile presentations, controversy continues regarding diagnosis of milder, later-onset diseases in infancy, including creation of pre-symptomatic populations of 'patients-in-waiting', the potential for medicalization, stigmatization, and/or discrimination. In-depth interviews were conducted with 36 adults [11 with Fabry disease (FD), 8 with Gaucher disease (GD), and 17 with late-onset Pompe disease (LOPD)], to determine their perspectives on NBS for their respective conditions. Thirty-four of 36 participants were in favor of NBS; both participants not in favor had GD1. Emergent themes influencing participants favorably toward NBS included earlier age of onset, a long diagnostic odyssey, less efficacious treatment, and the desire to have made different life decisions (e.g., relationships, career, or lifestyle) with the knowledge of their diagnosis. Concerns about insurance discrimination and psychological or physical burdens were associated with less favorable opinions of NBS. The ability for parents to make future reproductive decisions based their child's NBS result was considered favorably by some participants and unfavorably by others. Participants' specific condition (GD1, FD, or LOPD) contributed to these experiences differently. Participants with LOPD and FD favored NBS to initiate earlier treatment and prevent irreversible organ damage, whereas fewer patients with GD1 mentioned this benefit. Participants with LOPD had the longest diagnostic odyssey, while those with FD were more likely to report feeling misunderstood and experiencing accusations of malingering, both contributing to favorable views of NBS. Results expand prior quantitative findings by illuminating how participants' lived experiences can shape opinions about NBS. By understanding how currently affected individuals perceive the lifelong impact of a NBS result, genetic counselors can provide better anticipatory guidance to the parents of individuals diagnosed with a later-onset LSD by NBS.

Choosing how to discriminate: navigating ethical trade-offs in fair algorithmic design for the insurance sector

Here, we provide an ethical analysis of discrimination in private insurance to guide the application of non-discriminatory algorithms for risk prediction in the insurance context. This addresses the need for ethical guidance of data-science experts, business managers, and regulators, proposing a framework of moral reasoning behind the choice of fairness goals for prediction-based decisions in the insurance domain. The reference to private insurance as a business practice is essential in our approach, because the consequences of discrimination and predictive inaccuracy in underwriting are different from those of using predictive algorithms in other sectors (e.g., medical diagnosis, sentencing). Here we focus on the trade-off in the extent to which one can pursue indirect non-discrimination versus predictive accuracy. The moral assessment of this trade-off is related to the context of application—to the consequences of inaccurate risk predictions in the insurance domain.

Patient-Reported Outcomes and Experiences with Population Genetic Testing Offered Through a Primary Care Network.

Aims: To explore patient experiences in a large-scale primary care-based, preemptive genetic testing program.Methods: Patients who received genetic results from the initiative were invited to participate in an online survey 3 weeks postresult disclosure. A 6-month follow-up survey was sent to assess changes over time.Results: The initial survey was completed by 1646 patients, with 544 completing the 6-month follow-up survey. The following outcomes were high overall: patient-reported understanding of results (cancer: 87%; cardiac: 86%); perceived utility (75%); positive emotions (relieved: 66.8%; happy: 62.0%); family result sharing (67.6%); and satisfaction (87%), although analysis by demographic factors identified groups who may benefit from additional education and emotional support. Results-related health behaviors and discussions with providers increased over time (screening procedures 6.1% to 14.2% p < 0.001; provider discussion 10.3% to 25.3%, p < 0.001), and were more likely to take place for patients with positive cancer and/or cardiac results (39.8% vs. 7.6%, p < 0.001). Forty-seven percent of patients reported insurance discrimination concerns, and most (79.4%) were not familiar with privacy and nondiscrimination laws. Concerns regarding discrimination and negative emotions decreased between the two survey time points (privacy issues 44.6% to 35.1% p < 0.001; life insurance discrimination concerns 35.5% to 29.6%, p = 0.001; anxiety 8.1% to 3.3%, p < 0.001; and uncertainty 19.8% to 12.8%, p < 0.001). These findings led to the development and integration of additional patient resources to improve program implementation.Conclusion: Our findings highlight patient experiences with and areas of need in a community-based genomic screening pilot initiative using a mixed primary care/genetics provider model to deliver precision medicine.

Gender-Affirming Health Insurance Reform in the United States.

In May 2014, the US Department of Health and Human Services prohibited insurance discrimination of transgender individuals. Despite this, insurance plans often lack explicit guidelines on gender transition-related care and coverage of surgical procedures is extremely varied. We evaluated the evolution of insurance coverage of gender-affirming care following the 2014 legislative change. Insurance providers were selected based on company market share. We conducted a Web-based search and telephone interviews to identify the corresponding policies related to gender-affirming health care. We compared policy changes made before and after the 2014 US Department of Health and Human Services decision. Of the 92 insurers surveyed, 7% did not have a policy, and 315 policy revisions were documented. After the legislation, a significantly higher proportion of policy revisions were related to coverage of services (36% vs 11%, P < 0.0001), removal of existing criteria significantly decreased (23% vs 49%, P = 0.0044), and addition of criteria unrelated to international standards sharply increased (32% vs 2%, P = 0.0002). This resulted in reduced coverage of facial feminization, hair transplantation, laryngochondroplasty, and voice modification surgery. However, nipple reconstruction experienced increased coverage. The percentage of revisions to add preauthorization criteria to meet international standards (49% vs 45%, P = 0.6714) or to change terminology (37% vs 27%, P = 0.1055) were similar before and after the legislation. After the transformative legislation in 2014, an increasing number of insurance companies established gender transition-related policies. As more patients seek gender-affirming care, insurers deviate from international guidelines and create additional benchmarks that may act as barriers to care.

Anti-Discrimination Insurance Pricing: Regulations, Fairness Criteria, and Models

On the issue of insurance discrimination, a grey area in regulation has resulted from the growing use of big data analytics by insurance companies – direct discrimination is prohibited, but indirect discrimination using proxies or more complex and opaque algorithms can be tolerated without restrictions. Meanwhile, various fairness criteria have been proposed and flourish in the machine learning literature with the rapid growth of artificial intelligence (AI) in the past decade, which generally focus on a classification decision. However, there is little research on insurance applications, particularly on insurance pricing as a regression problem. In this paper, we summarise the fairness criteria that are potentially applicable to insurance pricing, match them with different levels of anti-discrimination regulations, and implement them into a series of existing and newly proposed anti-discrimination insurance pricing models. Our empirical analysis compares the outcome of different models and shows the potential of indirect discrimination.