Delivery of Genomic Medicine for Common Chronic Adult Diseases: Title and subTitle BreakA Systematic Review FREE

Quiz Ref IDThe greatest public health benefit of advances in understanding the human genome will likely occur as genomic medicine expands its focus from rare genetic disorders to inclusion of more common chronic diseases, such as coronary heart disease, stroke, diabetes mellitus, and cancer. These diseases are generally due to complex interactions between variations in multiple genes and the environment and only rarely are due to single-gene forms of the disease. With genomics discoveries relating to common chronic diseases, numerous genetic tests may emerge that hold promise for significant changes in the delivery of health care, particularly in preventive medicine and in tailoring drug treatment.

Attempts to integrate genetic/genomic knowledge of common chronic conditions into clinical practice are in the early stages, and as a result, many questions surround the current state of this translation. These questions include, what are the outcomes of genomic medicine? What is the current level of consumer understanding about genomic medicine, and what information do consumers need before they seek services? How is genomic medicine best delivered? What are the challenges and barriers to integrating genomic medicine into clinical practice? By examining studies relevant to these questions, this systematic review attempts to synthesize available information on the delivery of genomic medicine for common adult-onset conditions.

A review of articles published in peer-reviewed journals between January 2000 and February 2008 was conducted using MEDLINE. Searches were conducted independently by 1 of us (M.T.S.) and a professional librarian using the following Medical Subject Headings: genetics/genomics AND, outcomes, patient information needs/willingness to participate, delivery of services, workforce manpower issues, role of primary care, privacy issues, including insurance and employment issues. Reference mining of retrieved articles from these searches was used to identify additional articles.

Only original research articles and systematic review articles were reviewed. Systematic reviews were defined by having described a search for literature using computerized databases (eg, MEDLINE) and the presentation of findings in a systematic way. If we identified an article as already summarized in an accepted systematic review, we excluded it to avoid double-counting data. Commentaries, essays, legal analyses, consensus statements, and editorials were excluded. Delivery of genetic services and the associated issues were different in developed as compared with developing countries; therefore, we excluded articles about developing countries.

Because we were interested in genetic services for adults who have or who are at risk for common chronic diseases, such as cancer, diabetes, and coronary heart disease, we only included articles concerning these types of conditions, including multifactorial forms of common chronic diseases, and single-gene disorders that typically present in adulthood and feature common chronic diseases, such as breast and ovarian cancer due to BRCA1 and BRCA2 gene mutations or iron overload due to HFE gene mutations. This means we excluded articles that dealt with children or adolescents; genetic disorders that generally manifest in childhood (eg, neurofibromatosis or sickle cell anemia); rare genetic disorders that manifest in adulthood (eg, Huntington disease); reproductive issues; animal studies; genetically modified foods; research that did not pertain to clinical genetic services (eg, genetic association studies or studies describing participation in basic genetics research); and genetics educational resources, materials, curricula, or educational interventions for health professionals or the public.

Data were extracted (M.T.S.) and checked (P.G.S.) with disagreement resolved by consensus. Variables assessed included study design and topics within 4 key areas based on the key questions. We encountered 2 instances of potential overlap between topic categories and elected to resolve them as follows: we considered studies describing consumer educational needs as information needs rather than a barrier to integration of genetic services, and articles assessing use of family history in practice were classified as barriers to integration of genetic services rather than a model of integration into primary care, since these articles uniformly described inadequacies of most physicians in collecting and assessing familial risk.

Frequencies of data were tabulated and articles were grouped into topic categories. Evidence tables were created for each key area, and a narrative synthesis was performed. Because of the heterogeneous nature of the studies, we could not justify statistical pooling.

The literature searches and reference mining yielded 10 866 titles. After removal of duplicates and clearly irrelevant titles, 3371 citations were reviewed, and from these, 68 articles were selected and retrieved (Figure).

The number of articles in each category and topic according to type of study are presented in Table 1. The most common study designs were cross-sectional surveys and qualitative or descriptive studies. There were 5 systematic reviews and 8 experimental studies.

We identified 16 articles that described outcomes resulting from genomic medicine. These studies were further categorized as relating to psychological, affective, and cognitive outcomes; behavioral outcomes; or clinical outcomes (Table 2).

Psychological, Affective, or Cognitive Outcomes. We identified 7 articles that assessed psychological, affective, or cognitive outcomes, including 2 systematic reviews, 2 randomized controlled trials (RCTs), 1 survey, and 2 pre/post studies. One systematic review addressed the perceived risks and psychological and behavioral impacts of genetic testing in adults with a family history of disease.¹ A total of 35 articles involving 30 studies met criteria for this review. Only 1 RCT was included, 2 were cross-sectional studies, and the remainder had prospective designs.

Sixteen studies described outcomes for hereditary breast and ovarian cancer, 11 for hereditary nonpolyposis colorectal cancer, 1 for both hereditary breast and ovarian cancer and hereditary nonpolyposis colorectal cancer, and 2 for Alzheimer disease. The majority of the studies reported negative effects on affective outcomes for individuals found to carry genetic susceptibility mutations—but these effects were generally short-lived, and there were generally no differences between carriers and noncarriers with respect to perceived risk 12 months after genetic testing. An increase in screening behavior was observed in carriers, but the change was less than expected.

We identified another systematic review that addressed the effects of genetic counseling and testing for breast cancer susceptibility.² This review identified 15 studies, which were mostly observational in design, and concluded that, in general, counseling and testing was not associated with increased psychological distress and was associated with some improvements in decreased worry, anxiety, and depression.

One RCT involving obese individuals found that consultation that included genetics content did not have negative effects on self-efficacy, self-control, or increase in body weight after 6 months. In addition, the individuals with a family history of obesity who received consultation with genetics content reported significant improvement of negative mood after 6 months.³ Another RCT reported that education and personalized prevention recommendations increased the accuracy of risk perception without increasing worry in adult nondiabetic offspring of diabetic parents.⁴

The importance of risk perception associated with family history was described in a survey of patients with and without a family history of colorectal cancer. Being screened appropriately was significantly associated with a high perception of risk for colorectal cancer among those with a family history.⁵ The remaining identified studies were pre/post in design, assessing psychological or affective outcomes in patients before and after genetic counseling, and reported improvements in satisfaction, knowledge, and reductions in worry.^{6 - 7}

Behavioral Outcomes. We identified 5 articles relating to patient or clinician behavioral outcomes resulting from genetic services. A systematic review of 11 studies assessing patients' behavioral responses to genetic risk information relating to cancer and heart disease concluded that results were mixed, with some studies showing no change in behavioral outcomes, whereas other studies reported increased motivation to participate in cancer screening.⁸

Results from a survey of US physicians about exposure to and attitudes toward advertisements for genetic tests for inherited cancer susceptibility found that 27% had received advertising for genetic susceptibility tests and 25% felt that such advertising would be important in their decision-making to recommend testing.⁹ Another survey of primary care clinicians in the United States assessed factors influencing whether these physicians would order a genetic test to individually tailor smoking-cessation treatment. The study found that the most important factors for ordering such a test included the ability to more precisely target smoking-cessation treatment and to encourage patients with the knowledge that their treatment was tailored. However, concerns about genetic discrimination were an important barrier.¹⁰

A survey of Massachusetts family physicians found that physicians expected their patients to be more likely to get screened for cancer and make lifestyle and behavioral changes if genetic testing identified them as at increased risk.¹¹ Survey data from 65 patients with history of depression found that depressed patients reported that they were willing to pay for improvements in treatment response to antidepressant medication, and their stated willingness to pay for a 5% improvement in treatment response exceeded the price of a pharmacogenetic test.¹²

Clinical Outcomes . We identified 4 studies that assessed clinical outcomes. A pre/post study assessed women with BRCA mutations and reported that after counseling, 15% and 50% underwent risk-reducing surgery for breast and ovarian cancer, respectively, and 4 early stage cancers were identified. Among women who did not undergo prophylactic surgery, recommended cancer screening increased.¹³

A nonrandomized comparison of Greek patients who had at least 2 unsuccessful attempts at weight loss compared conventional diet and nutrition counseling with nutrition counseling informed by a nutrigenomics test result. For the first 300 days, weight loss between groups was not statistically different, but among the 50% of patients with more than 300 days of follow-up, there was a statistically significant difference of about 5 kg, with the test group weighing less on average.¹⁴

We identified only 1 RCT of a genetic testing intervention for a common condition that measured a clinical outcome.¹⁵ In this trial, 200 adults requiring warfarin anticoagulation were randomized to standard dosing vs a dosing schedule determined from a formula that takes into account CYP2CP and VKORC1 genotypes. During the time to achieve a stable maintenance dose, there was no difference between groups in the primary outcome of the percentage of international normalized ratios outside the therapeutic range (31% in the pharmacogenetic group and 33% in the standard dosing group). Patients in the pharmacogenetic group did have significantly fewer dose adjustments and fewer blood samples drawn for testing international normalized ratios.

One study reported possible harms from genetic testing. Among 24 patients who had BRCA testing and were found to have an unclassified variant (meaning the result of the test was uninformative), semistructured interviews revealed that 19 interpreted this information as “pathogenic” and 10 went on to undergo preventive surgery. None of the 5 patients who interpreted this test result as uninformative had preventive surgery.¹⁶

We identified 16 articles describing consumer information needs, including 1 systematic review, 1 RCT, 10 surveys, 2 pre/post studies, and 2 qualitative studies. Five articles addressed consumers' knowledge, attitudes, and beliefs about genetics and the other 11 addressed the intention to seek or participate in genetic services, including uptake of genetic testing (Table 3).

Five surveys from the United States and the Netherlands assessed consumers' knowledge, attitudes, and beliefs, and generally they all found that genetic knowledge was low, but attitudes about genetics were positive. However, there were concerns about genetic discrimination.^{17 - 21} Two surveys from the United States investigated knowledge and attitudes of Hispanic participants regarding cancer genetic services, found interest in such services,²⁰ and found that acculturation factors related to language may affect awareness.²¹

Six articles focused on the intention to seek genetic services, including motivations for genetic testing. An RCT randomized 2165 primary care patients into 2 groups—one offered a DNA-based test for hemochromatosis and the other a phenotypic test—and then assessed the willingness to have the testing. The authors found that acceptance of testing was about the same for both groups (56% for genotypic and 58% for phenotypic testing) and that reasons for refusal included a need to talk with a physician (44%), concern about privacy (32%), and dislike of blood drawing (29%).²²

Two surveys of patients with colorectal cancer assessed factors contributing to decision making about genetic testing for hereditary nonpolyposis colorectal cancer, including germline testing²³ and tumor microsatellite instability testing.²⁴ Among 314 patients, motivations for germline testing included learning of increased risk to children and finding out whether additional screening was needed. For 125 patients eligible for tumor microsatellite instability testing, the majority had positive attitudes about the potential benefits and perceived few barriers to undergoing the test, despite having little knowledge about the test. A pre/post study used interviews and questionnaires to assess barriers to participating in genetic counseling and BRCA gene testing for patients newly diagnosed with breast cancer.²⁵ Participation was not influenced by distress, knowledge about hereditary breast cancer, previous genetic testing in relatives, or perceived risk and barriers.

A survey of 187 mothers undergoing BRCA gene testing who had children aged 8 to 21 years found that most of these women were motivated to have testing for the information it provided to their children, and they believed that having such information could be used to prevent or control cancer in their children. The 2 most common informational resource needs of these mothers for discussing their BRCA test results with their children included reading educational literature (93%) and speaking with a family counselor (86%).²⁶

Another survey of 139 women previously treated for early stage breast cancer found that the majority (76%) would have been interested in genetic testing to determine risk of recurrence with most (84%) wanting to include information from test results in decision making about treatment.²⁷

Five articles addressed patients' information needs related to participation in genetic services. Two articles underscored the importance of eliciting the patient's perspective when discussing risk of chronic disease, particularly in the context of the family history. A systematic literature review that included 11 qualitative research studies found that for individuals with a family history of cancer, coronary heart disease, or diabetes, the salience of their family history was determined by acknowledging that a disease runs in the family, and this was strongly influenced by the personal experiences of that illness in relatives. Some of these factors were the same as the medical factors used to assess familial risk (ie, the number of affected relatives and age at onset), but others were more personal, such as emotional or physical closeness of a family member.²⁸ A qualitative study published the following year delved further into patients' perceived familial risk and found that a sense of vulnerability to common chronic disease depended on the emotional impact of experiencing the illness in the family, especially if the illness was sudden, premature, or fatal, and the nature of personal relationships within the family.²⁹

A pre/post study from the United States of women eligible for referral to cancer genetics consultation found that after 6 months those who had attended the consultation were significantly more knowledgeable, but clinical history, perceived risk, family history, psychological state, and anxiety were not associated with attendance.³⁰

Two articles addressed the genetic counselor-patient interaction. A cross-sectional study of observed behavior from the Netherlands found that adults referred for cancer genetics consultation had a strong psychosocial focus, which was important to their agenda for the consult; however, these needs only minimally influenced the interaction during the visit because patients did not communicate their previsit needs.³¹ Quiz Ref IDA qualitative study from the United States that interviewed genetic counselors and their patients found that the goals of genetic counseling were often unclear to both. Although counselors objected to stating goals because doing so implied a preset agenda, patients were uncertain about the role of the counselor and the boundaries of what would be discussed. Most patients appreciated the time spent by genetic counselors and believed that their physicians could not provide the service.³²

We identified 16 articles that described the delivery of genomic medicine. These studies were further classified as relating to existing genetic services and the genetics workforce, integration of genetics into primary care, and new models of genetic services delivery (Table 4).

Five articles described genetic services and the current genetics workforce: 4 surveys of the genetics workforce in the United States and 1 descriptive study of regional genetic health centers in the United Kingdom. Two surveys used data from a mailed questionnaire to the 1576 active physician and PhD members of the American Board of Medical Genetics and found that 60% of respondents perceived that there were inadequate numbers of geneticists³³ and that geneticists providing services for adult-onset diseases represented the smallest proportion of geneticists (7%).³⁴ Another survey obtained data from active members of the American Academy of Pediatrics Section of Genetics and Birth Defects and physicians certified by the American Board of Medical Genetics and found that 5% had training in internal medicine, and adult patients seen for reasons other than prenatal services represented only 12% of patients seen by geneticists.³⁵ The fourth survey described characteristics and professional roles of 211 nurse genetics specialists in the United States and found that most (57%) provided direct patient care and worked in genetics (26%) or oncology (22%) settings.³⁶ The descriptive study from the United Kingdom found that half of referrals come from primary care, genetic counselors were key to providing genetic services, and genetic testing was primarily for single-gene disorders.³⁷

Six articles described the integration of genetics into primary care. This included 1 qualitative study from Canada that found family physicians anticipate providing more genetic services, but they were uncertain and had concerns about their abilities to do so.³⁸ The remaining 5 articles were from the United Kingdom and described a range of approaches to integrating clinical genetic services into primary care, including 2 pre/post studies, 1 survey, 1 qualitative study, and 1 cluster RCT.

Three studies (the RCT, the survey, and a pre/post study) addressed the use of a nurse genetics specialist within the primary care setting. Both the RCT and the survey found that patient satisfaction with the service provided by the nurse genetics specialist was high and costs with the new model were less.^{39 - 40} The pre/post study found that physicians believed that contact with a nurse genetics specialist helped in identifying patients in need of a genetics referral and that nurse genetics specialists provided current genetics information to them.⁴¹

The other 2 studies examined the effect of referral guidelines for cancer genetics services among general practitioners. The qualitative study found that physicians appreciated that referral guidelines would improve their referral patterns,⁴² and the pre/post study found that guidelines were associated with improved referral patterns.⁴³

Five articles described new models of genetic services. One experimental study evaluated telephone disclosure of BRCA genetic test results compared with traditional genetic services and found there were no significant differences in anxiety and general well-being between the 2 groups, and both groups reported similar rates of satisfaction with the services.⁴⁴ A survey of rural physician practices in northern Maine and public health nurses statewide found that health professionals receiving and providing educational and clinical genetic services via videoconferencing generally found it to be a positive and satisfying experience.⁴⁵ One experimental study randomized general practice teams in the United Kingdom to the use of computerized decision support software that aided in familial risk assessment of breast, ovarian, and colorectal cancer or to a comparison group that received education and guidelines; it found that with decision support, there were significantly more genetics referrals, and these referrals were more likely to be consistent with referral guidelines.⁴⁶

Quiz Ref IDRegarding direct-to-consumer marketing of genetic services, an observational study that assessed nutrigenomics tests found that the genetic test results provided little influence on the assessment and recommendations provided by the test companies, and test results frequently included recommendations for the consumer to purchase costly dietary supplements.⁴⁷ A survey of primary care physicians and consumers in the United States found that a small percentage (14%) of consumers were aware of nutrigenomic testing and of those who had a nutrigenomics test, only 10% had discussed the results with their physicians.⁴⁸

We identified 20 articles that described barriers to the clinical integration of genomic medicine. These articles were further classified as relating to health professionals' knowledge, attitudes, beliefs, and abilities; lack of oversight of genetic testing; and privacy, confidentiality, and genetic discrimination concerns (Table 5).

Fifteen articles described health professionals' knowledge, attitudes, beliefs, and abilities pertaining to medical genetics. This included 1 systematic review, 10 surveys, 2 observational studies, 1 descriptive study, and 1 qualitative study.

The systematic review included 18 studies that dealt with primary care physicians' perceived barriers concerning the provision of genetic services. The studies were published between 1993 and 2001 and included 7 surveys, 4 pre/post tests, 4 qualitative studies, 1 observational study, and 1 cluster RCT. These studies found that physicians reported that they lack sufficient knowledge and confidence relating to provision of genetic services and they have limited time to obtain family history.⁴⁹ These themes were also identified in the other 14 articles.

The 10 surveys and the qualitative study we found assessed genetics knowledge and attitudes of a variety of health professionals from the United States, Australia, and the Netherlands, and these studies found similar results to the systematic review,⁴⁹ with respondents feeling underprepared for assessing and managing genetic issues in their practice and lacking basic genetic knowledge.^{50 - 60}

The 2 observational studies and the descriptive study we retrieved described health professionals' abilities to collect, document, and interpret family history. These studies found that a majority of patients with a strong familial risk for a common disease lacked documentation of key family history elements in their medical record, as well as documentation of genetic risk assessment or referral for genetics evaluation.^{61 - 63} A survey of family physicians also found that most (87%) believed that family history screening is important; however, only 62% were confident in the ability to recognize patients with hereditary risk.⁵³

One article described oversight of genetic testing. A survey of 190 genetic testing laboratory directors found that most would support formal registration under a genetic testing specialty regulated by the Clinical Laboratory Improvement Amendments and that proficiency testing could improve the quality of genetic testing.⁶⁴

Four articles described concerns about privacy and genetic discrimination. Two articles used data from the same qualitative comparative analysis of states with and without laws prohibiting genetic discrimination by health insurers. One article found that there are no well-documented cases of health insurers either asking for or using presymptomatic genetic test results for underwriting purposes.⁶⁵ The other study found that patients' and clinicians' fears of genetic discrimination in health insurance exceeds reality, and this fear may deter genetic testing, especially for high-cost tests and adult-onset diseases.⁶⁶

A cross-sectional survey of 86 859 patients in primary care settings participating in a hemochromatosis screening study in North America found that African American and Asian individuals were less likely to have concerns about genetic discrimination, and younger participants, US residents, and those with less than a high school education were more likely to have these concerns.⁶⁷ Quiz Ref IDA follow-up survey of 832 participants from this cohort who were screened for iron phenotypes and HFE genotypes associated with hemochromatosis found that after 1 year, none reported problems with health insurance or employment and only 3 (0.4%) reported problems with life insurance or long-term care insurance.⁶⁸

Genomic advances hold the promise to improve care and prevention of common chronic diseases. Although that promise has yet to be fully realized, new genetic discoveries make the reality of that promise seem closer than ever before. However, if past experience with other health care advances is any guide, patients will not realize the full benefit of genomic advances without a thorough understanding of the organization, clinician, and patient needs that are required to translate these advances into improved clinical outcomes. Our systematic review reveals a large gap between what knowledge is available and what health systems still need to know about the outcomes, consumer needs, organization of health services, and barriers, to ensure appropriate and effective clinical integration of genomic information and technologies for common chronic disease.

Quiz Ref IDThe most important and consistent finding from our literature review is that the primary care workforce, which will be required to be on the front lines of the integration of genomics into the regular practice of medicine, feels woefully underprepared to do so. Remediation of this deficiency should be a top priority, and more studies are needed to test models for how this can be feasibly accomplished.

A second theme we identified is that consumers have unclear and dissonant notions about the value of genetic testing for common chronic disease. In general, consumers knew little about genetics/genomics but were interested in the prospect of this technology helping to better identify diseases for which they and their family members were at increased risk. Consumers were also worried about the possible adverse consequences of genetic testing, particularly the privacy issues and discrimination in health insurance and employment. Therefore, another research need is the development and testing of interventions to promote greater understanding of genomic medicine among consumers. Such interventions may need to be tailored for the needs of specific populations.

A third theme we identified is that there is a great need to better understand the outcomes of genomic medicine interventions for common chronic disease. To date most studies have assessed these outcomes using weak pre/post study designs, and in general these studies have assessed changes in psychological, affective, and cognitive outcomes of patients receiving genetic counseling and testing for single-gene disorders. More research describing clinical outcomes is needed: do patients who receive counseling and testing have better clinical outcomes in terms of mortality, decreases in incidence of disease, and better clinical responses to pharmaceuticals? And at what cost?

We identified other barriers to the clinical integration of genomic medicine for common chronic disease, in addition to the perceived inadequacy of the primary care workforce to participate in this. The most prominent of these include health professionals' lack of basic knowledge about genetics and their lack of confidence in interpreting familial patterns of disease, which limits their ability to appropriately counsel their patients, order and accurately interpret genetic tests, and refer their patients for genetics consultation. In addition, there may not be sufficient numbers of genetics professionals available to meet the demand for genetic services.

Our review has several limitations. First was the difficulty in identifying health services studies of genomic medicine for common chronic disease. We know of no empirical assessments identifying an optimal search strategy for such studies as, for example, there exist optimal search strategies to identify RCTs and studies of risk and prognosis. We searched MEDLINE using reasonable search terms, but we acknowledge it is likely that we have failed to identify some studies of relevance. However, given the consistency of the studies we did identify, we believe it is unlikely that missing literature would substantially change our main conclusions.

Second, we did not address the literature dealing with prenatal genetics, pediatric genetics, genetic conditions with pediatric onset experienced by adults, or rare single-gene disorders presenting in adulthood. Our focus was on the opportunities and challenges of genomics interventions for common chronic diseases of adulthood. Even then we found that the evidence regarding these conditions is largely limited to single-gene disorders that manifest as common diseases in adulthood, particularly cancer.

Third, because this literature was a heterogeneous combination of study questions, designs, populations, and outcomes, synthesizing the study findings involved making a certain number of judgments on our part. We tried to be explicit and describe these judgments as fully as possible so readers can understand what we did.

In summary, this review of health services research studies revealed many gaps in the organization, clinician, and patient needs that must be filled to realize the full benefits of genomics. These studies reported that consumers need more and better information to understand the value of genomic medicine for common chronic diseases; that the primary care workforce is not prepared to meet the challenge of integrating common disease genetics/genomics into the regular practice of medicine; that the genetics workforce appears to have inadequate numbers to respond to genetic service needs for adult-onset conditions; that few models of integration into primary care have been successfully implemented on a wide scale; and that a host of barriers exist, such as lack of oversight for genetic testing technologies and concerns about privacy and discrimination. All of these issues need attention. It will be a lost opportunity if the health services research components of genomic medicine fail to keep pace with the rapid basic science advances and clinical discoveries.