Patient reported outcomes and patient empowerment in clinical genetics services

Evaluation of clinical genetics services (CGS), including genetic counseling and genetic testing, has been problematic. Patient mortality and morbidity are unlikely to be directly improved by interventions offered in CGS. Patient‐reported outcomes (PROs) are not routinely measured in CGS evaluation, but this may change as patient‐reported outcome measures (PROMs) become a key part of how healthcare services are managed and funded across the world. However, there is no clear consensus about which PROMs are most useful for CGS evaluation. This review summarizes the published research on how PROs from CGS have been measured and how patients may benefit from using those services, with a focus on patient empowerment. Many patient benefits (PROs) identified repeatedly in the research literature can be re‐interpreted within a patient empowerment framework. Other important PROs identified include family functioning, social functioning, altruism, sense of purpose, enabling development of future research and treatment/participating in research. Well‐validated measures are available to capture (dimensions of) patient empowerment. Although generic measures of family functioning are available, suitable measures capturing social functioning, development of future treatments, and altruism were not identified in this review. Patient empowerment provides one useful approach to measuring PROs from CGS.

Evaluation of clinical genetics services (CGS) has been problematic. Although patient mortality and morbidity may be improved downstream, particularly for patients at-risk for hereditary cancer (HC) syndromes who may be offered options for cancer screening and risk-reducing surgery, these health status outcomes cannot be directly attributed, in the short-term, to CGS. Are there positive changes (benefits) experienced by patients that could be attributed directly, and in the shorter term to the interventions offered in CGS? CGS in the UK National Health Service (NHS) are currently evaluated using process measures only e.g. patient waiting times, number of genetic tests, etc. Patient outcomes, including patient-reported outcomes (PROs) are not routinely used. This may change as healthcare evaluation using patient-reported outcome measures (PROMs) is set to become a key influence on how healthcare services are managed and funded across the world (1,2). PROs are patients' subjective outcomes from using healthcare, and PROMs are questionnaires used to capture, or measure PROs. However, there is no clear consensus about which PROs are directly attributable to CGS, or about what PROM(s) best capture or measure these to show positive change from the patient's perspective. The aim in this review is to summarize the published research on (i) how patients may benefit from using CGS, as a way of identifying potential PROs and (ii) how these PROs might best be measured. A focus on patient empowerment is taken throughout because (i) this is a broad construct comprising many subjective outcomes valued by CGS patients and (ii) this construct has recently been applied to CGS. Empowerment may be an overarching construct under which many other PROs from using CGS could be subsumed.

Do patients benefit from using CGS?
Services provided by CGS include diagnosis of genetic conditions, genetic risk assessment, genetic testing and genetic counseling. The first three of these focus on providing accurate information to patients; accuracy of information provided can be evaluated with objective measures e.g. diagnostic accuracy. However, patients may derive subjective benefits (PROs) from genetic information, and a different approach is required to assess these PROs, that focus on the patient's perspective. One approach is to investigate whether the goals of genetic counseling have been achieved, from the patients' perspective. Genetic counseling has been defined as integrating: (1) Interpretation of family and medical histories to assess the chance of disease occurrence or recurrence.
(2) Education about inheritance, testing, management, prevention, resources and research. Genetic counseling includes genetic risk assessment, but also emphasizes goals to educate patients and to support them to adapt to genetic information and make informed decisions. These are domains with potential for PRO measurement.

Education
Early studies explored whether genetic counseling can educate clients about genetic risks, testing and inheritance. Although some studies showed good evidence of information recall (4), other studies showed that patients do not always accurately recall a specific recurrence risk (5,6). Risk perception after genetic counseling/testing is problematic as an outcome domain because it is well known that people use heuristics, or mental 'short cuts' when processing information for decision-making (7,8). Indeed, a recent systematic review showed that educational interventions are not effective at improving subjective cancer risk perception, and furthermore, the strongest predictor of post-intervention risk perception across studies is baseline perceived risk (9).
Part of the explanation for this may be that the term 'risk' may be a dynamic multidimensional construct comprising the numerical probability of an event combined with other issues including severity, context (e.g. personal experience), and other factors (10,11). If risk is conceptualized in this way, genetic counseling patients may develop a composite sense of risk influenced by these factors as well as by the numeric probabilities provided in genetic counseling. However, only the numeric probability part of this is assessed in measures of subjective risk perception. Furthermore, contextual factors (family experiences) can change over time, so measuring risk perception as an outcome from genetic counseling may not be useful, because risk perception may be dynamic rather than static. A 2008 systematic review of risk communication in CGS found that the informational and educational elements of genetic counseling provided less benefit to patients than the counseling elements that addressed emotional issues such as loss and unresolved grief (12).

Promoting informed choice and adaptation
Patient adaptation or adjustment to genetic risk information has been explored by assessing the impact on psychological distress (anxiety, depression, worry), particularly in the context of genetic testing for Huntington disease (HD) and HC. Evidence indicates that distress amongst mutation carriers tends to increase shortly after test results are communicated, and return to pretesting levels over time; amongst non-carriers and those who obtain inconclusive results, distress has a tendency to decrease over time (13)(14)(15)(16)(17)(18)(19). Similar patterns have been reported for apolipoprotein E (APOE) genetic susceptibility testing for Alzheimers disease (20).
These findings are reassuring because they show that genetic testing does not cause significant long-term psychological harm to most people tested. A minority of people tested continue to experience distress, particularly those with a pre-existing history of anxiety or depression. However, usefulness of distress as a PRO is problematic in that positive change has not been showed for people identified to be mutation carriers. For mutation carriers, it appears that the best that can be achieved regarding psychological distress is that they will not experience an increase in distress in the long-term, which is not the same as achieving positive change. Psychological distress does not seem useful as a PRO that could be measured to capture positive change after predictive, pre-symptomatic or genetic susceptibility testing amongst those identified to be mutation carriers.
There have been calls for research to assess whether genetic counseling and/or testing promote informed choice (21). There is some limited evidence that genetic counseling/testing can reduce decisional conflict about specific decisions (22,23), but there is still little firm evidence that genetic counseling/testing promotes informed choice more generally. Most research in this area was conducted in the context of either prenatal screening, often not preceded by genetic counseling (24)(25)(26), or evaluation of decision aids as an adjunct to genetic counseling (27). This research may not be representative or applicable to general CGS.

PROMs used to evaluate CGS
A 2008 systematic review of outcome measures used to evaluate CGS, by author M. M. and others, identified 67 validated questionnaires (PROMs) and three 'objective' outcomes (accuracy of diagnosis, accuracy of tests, rate of terminated pregnancies) (28). Questionnaires identified captured 14 PRO domains, ranging from health status through knowledge to psychological distress, clearly showing discordance over what PROs should be measured. Furthermore, 46 measures identified had been reported in one study only, with only 21 measures used more than once.
Worryingly, psychometric validation of the 67 questionnaires identified was limited. Only 25 of the 67 measures identified were assessed for test re-test reliability and only five for responsiveness. However, both are key properties of any questionnaire to be used as a PROM; without demonstrable reliability and responsiveness, a PROM will not be fit for purpose (29,30).
A further concern is that only 2 of 67 questionnaires identified were assessed for interpretability (28), a key requirement for any questionnaire intended for use as a PROM. Interpretability describes the degree to which a clinical or commonly understood meaning can be assigned to quantitative scores or change scores and is usually assessed using minimal important change (MIC) or minimal important difference (MID) (29,30).
Not only is there lack of agreement about what we should be measuring to assess PROs from using CGS, and how we should be measuring these, but the quality of PROMs available in 2008 was limited (28).
In summary, there is some limited evidence that educational and adaptational goals of genetic counseling are being achieved. Furthermore, there are issues regarding many PROs and PROMs that have been used to evaluate CGS. There may be better ways to capture PROs.
The aims of the rest of this review are to identify what (potential) PROs of CGS have been reported in the literature that could be captured using PROMs, to examine these using the theoretical framework of patient empowerment, and to provide an update on measures available. A focus on patient empowerment is taken because firstly, patient empowerment has been identified as a key patient outcome goal of genetic counseling (31) and secondly, because the concept of patient empowerment may be an overarching construct that could represent many other PROs from using CGS. Patient empowerment has recently been applied to CGS as a PRO summarizing a range of patient benefits directly attributable to CGS, through development and validation of a new PROM (32)(33)(34).

Methods
This review is not a systematic literature review, as no formal meta-analysis or critical appraisal of included articles was conducted. This was because the aim was not to synthesize and critique the included studies, but to identify what (potential) PROs have been reported in the literature, to examine these using the theoretical framework of patient empowerment, and to provide a brief update on measures available. However, a structured approach was used to identify articles for inclusion. Two literature searches were conducted.
Search 1 aimed to identify articles reporting patient benefits or patient empowerment from using CGS. Medline, EMBASE, AMED and PsycINFO were searched from January 2006 to June 2014 (Box 1). The term 'benefit' was used instead of the term 'outcome' because the aim was to identify outcomes that provide benefits or positive change to patients. Grey literature was not searched and key-author and reference list searches were not conducted. Titles and abstracts were double-screened by M. M. and A. D. using the following exclusion criteria: (1) The words 'benefit' or 'empower*' do not appear in title or abstract. (2) Articles exclusively reporting clinical features, new genes/mutations, population screening, pharmacogenetics, nutrigenomics, pregnancy outcomes after assisted reproduction, tumor testing, clinical guidelines, conference abstracts, editorials, discursive articles. (3) No empirical data reported on (perceived) patient benefits or (perceived) patient empowerment from using CGS.
Disagreements were resolved by discussion. Articles were retained for full text screening if it was unclear from title and abstract that the article could be excluded. Full-text articles were included only if exclusion criteria were not met. Data were extracted regarding what patient benefits or aspects of patient empowerment were reported. Data on magnitude of these outcomes were not extracted because the aim was simply to identify what patient benefits of CGS have been reported.
Benefits identified in included studies were qualitatively re-interpreted using the theoretical framework of 'empowerment'. Patient empowerment is a contested term, with many definitions that all relate to patients as self-determining agents (35). Some definitions of patient empowerment focus on the process of empowering patients to self-manage long-term conditions. Other definitions focus on patient states, experiences and capacities such as perceived control over health and healthcare, ability to self-manage their condition, collaborate with healthcare providers, and use coping resources. (3) Behavioral control: feeling able to use the health and social care systems effectively to reduce harm/improve life for oneself and other relatives. (4) Emotional regulation: feeling able to effectively manage emotional consequences of genetic information. (5) Hope: for a fulfilling family life for oneself, relatives and future descendants.
Qualitative re-interpretation was conducted as follows: M. M. conducted the first round of re-interpretation by assigning one of the five dimensions of empowerment to each benefit shown in Appendix S1 if it seemed to fit, or by creating a new code (assigning an appropriate title) if the benefit could not be coded using one of the empowerment dimensions. Where the term 'empowerment' was identified as a benefit in the source article, then the generic code 'empowerment' was used. Coding was iterative and every benefit was coded to saturation (as many codes as seemed relevant were assigned to each benefit). A code-book was created with a description for each new code created.
M. M. and A. D. then independently coded the benefits shown in Appendix S1, using the code-book. Independent coding was done by either assigning one of the five dimensions of empowerment to each benefit if it seemed to fit, or by assigning another code from the code-book if the benefit could not be coded using one of the empowerment dimensions. Coding to saturation was not conducted in this second round; if a benefit could be re-interpreted clearly within the empowerment framework, no further codes were added to that benefit, unless there was uncertainty. Following independent re-interpretation, A. D. and M. M. compared their coding. This led to further iteration of the code book, because some codes were dropped in this second round. Most coding differences were easily resolved by discussion, and related to whether one or more of the empowerment dimensions was relevant. Other differences were more difficult to resolve, and these will be described below.
Search 2 was conducted to update the 2008 Payne et al. systematic review to identify any PROMs newly available or newly applied to evaluate CGS (28). Medline, EMBASE, AMED, Cochrane databases and PsycINFO were searched from February 2006 to June 2014 using Payne's search terms (28). Hapi was excluded as payment was required.

Review 1: patient benefits
In total, Search 1 identified 2424 records (Appendix S2). Title and abstract screening excluded 2379 records and the remaining 45 full-text articles were retrieved and assessed for eligibility (28,32,33,. This led to inclusion of 39 studies that used both quantitative and qualitative study designs. Benefits identified are shown in Appendix S1. Qualitative re-interpretation of benefits using the empowerment framework is indicated using superscripts in Appendix S1. Within each study, benefits are superscripted in column 2, and cross-refer to numbered themes in column 3 to illustrate how benefits were re-interpreted, where possible, using the empowerment framework. Benefits that could not be re-interpreted within the empowerment framework are identified separately under the sub-heading 'Other' in column 3. Following both rounds of re-interpretation, 27 benefits were identified, with 21 of these not interpretable within the empowerment framework. However, empowerment, and/or one or more of its dimensions could be identified in 38 of 39 studies; one identified post-traumatic growth (PTG) only (61). Only 10 articles explicitly identified empowerment or one of its dimensions, including perceived personal control (cognitive, decisional and behavioral control). Knowledge was identified in 19 studies, but in all of these this was re-interpreted as cognitive control. Reduction in psychological distress was reported in nine studies, although in all cases this was re-interpreted as emotional regulation. Intention or motivation to change health behavior was reported in 11 studies, but none of these studies reported evidence of actual behavior change, apart from use of interventions to reduce breast cancer risk (e.g. ref. (60), and was therefore re-interpreted as behavioral control. Ability to cope was identified in three studies, and was re-interpreted as behavioral control and emotional regulation. Of the remaining benefits, family functioning was identified in six studies, social functioning/support in six studies, risk perception in five studies, altruism in five studies, sense of purpose in four studies, and enabling development of future treatments and research/research participation in four studies. All other benefits were identified in just one or two studies.

Discussion
Patient benefits (PROs) reported in 38 of 39 studies could be re-interpreted within the empowerment framework (33). These findings support the reciprocal-engagement model of genetic counseling practice, which includes empowering patients to feel more in control of their situation as a patient outcome goal (31). However, re-interpretation within the empowerment framework was not always clear-cut. For example, in the first re-interpretation (conducted by M. M. only), 'ability to cope' was coded separately, outside of the empowerment framework. However, when the second round of independent coding (M. M. and A. D.) was done, and discussed, it was agreed to uniquely code within the empowerment framework because coping is an activity that can be emotion-focused or problem-focused, and it was agreed to re-interpret as a combination of behavioral control and emotional regulation (77). A second example of a difficult concept to re-interpret was PTG. This led to considerable discussion, because both authors felt that PTG does overlap with empowerment, but there was uncertainty about whether the empowerment construct encompassed all elements of PTG. This may be because the challenges posed by having a genetic condition in the family may sometimes, but not always be so traumatic as to have potential to lead to the kind of dramatic 'life-changing' psychological shifts in thinking and relating to the world that are associated with PTG (61). Perhaps empowerment can occur even where such dramatic shifts do not take place, with the potential for PTG where the change is most extreme.
Other important PROs identified in this review, that could potentially be useful as PROs in CGS include family functioning, social functioning, altruism, sense of purpose, enabling development of future research and treatment/participating in research. Although the theoretical framework of empowerment used in this study does not include these four dimensions, it is plausible to suggest that patients who achieve positive change in these areas following their use of CGS may well consider themselves to be more empowered. In particular, identification of family functioning as a patient benefit and potential PRO from CGS is particularly interesting because genetic conditions are whole family concerns. The presence of a genetic condition in the family may pose a considerable challenge to the functioning of the whole family, as members struggle to nurture and care for multiple affected family members, whilst worrying about risks to themselves and their children, and try to find ways to communicate about genetic risks to relatives. There is much that can be drawn from areas such as family therapy that can usefully be applied to genetic counseling practice, for example the Family System Genetic Illness Model (78,79).
Although measures of family functioning are available and have been applied to CGS (28), suitable measures to capture social functioning, altruism, sense of purpose, enabling development of future research and treatment/participating in research were not identified in this review. The 'relating to others' sub-scale of the Post-Traumatic Growth Inventory (61) may capture some aspects of social functioning. Change in risk perception was identified as a patient benefit in four studies, but this is problematic because of heuristic biases and the cognitive difficulty of interpreting and making sense of risk information (7)(8)(9)(10)(11).
Findings in Review 2 identified three questionnaires capturing outcome domains that were not captured by the measures identified by Payne et al. (2008): GCOS-24, IPQ, and PDDS. All other outcome domains captured by the additional questionnaires identified in this review (health status, anxiety, depression, satisfaction and health behavior) are captured by the measures identified by Payne et al. (28).
Of the novel domains captured by questionnaires identified in this review, only GCOS-24 captures patient empowerment, a construct that could encompass many PROs identified in studies included in Review 1. As it was developed specifically for use in CGS, GCOS-24 has potential for a high degree of clinical utility. IPQ is a generic measure and captures a construct not dissimilar to cognitive and behavioral control, dimensions of empowerment (80,81). IPQ captures patient ideas about illness identity, cause, timeline, consequences, and cure-control, which all relate to patients' knowledge and understanding about the condition. These include dimensions which could be interpreted as dimensions of empowerment, including knowledge and understanding (cognitive control), and their beliefs about whether they can effectively reduce harm/improve life for themselves and other relatives (behavioral control). Both IPQ and GCOS-24 have potential as generic (IPQ) and condition-specific (GCOS-24) PROMs for CGS that may capture important patient benefits.
PDDS measures (reduction in) stigma experienced. However, this benefit was identified in only one of 39 studies identified in Review 1. As PDDS is not widely used as an outcome measure in CGS, and (reduction in) stigma is not widely reported as a PRO from using CGS, it is unclear how useful PDDS is for CGS evaluation, although it may be useful in some areas of CGS e.g. genetic counseling for conditions affecting mental health.
Review 1 showed that apart from use of interventions to reduce high genetic risk for cancer, firm evidence of actual behavior change in response to genetic risk information is lacking. The concept of perceived personal control (PPC) may help to identify and measure benefits from CGS despite lack of demonstrated health behavior change in response to genetic information (82). PPC may also be a useful alternative to measuring knowledge, or recall of genetic risk information. Feelings of control, including feeling that you have enough knowledge and understanding about the condition, including risks to self and relatives, may be valued as outcomes from using CGS, even if that control is not or cannot be translated by the patient into health behavior change with downstream measurable health status benefits. Measures available to capture PPC defined as cognitive, decisional, and behavioral control include the PPC scale (82,83), GCOS-24 (34) and, arguably, IPQ (80,81).
A combination of generic and condition-specific measures is recommended for use in healthcare evaluation (1). The generic IPQ, and the CGS-specific PPC and GCOS-24 are all well-validated for reliability and internal consistency, and both PPC and GCOS-24 have been shown to be responsive to change before-after genetic counseling/testing (34,(80)(81)(82)(83)(84). However, none of these three measures has yet been assessed for interpretability (MID or MIC). Despite this, all three measures have been used in studies evaluating CGS. For example, GCOS-24, the most recently developed of these, was used to evaluate a specialist psychiatric genetic counseling clinic in Canada (84).
A limitation of this review is the lack of any formal critical appraisal of included studies. A further limitation is the subjective qualitative re-interpretation of patient benefits within the empowerment framework because other interpretations are also possible. However this review is the first attempt to identify what PROs of CGS have been reported in the literature, and what validated measures (PROMs) are available to capture these.

Conclusion
Patient empowerment provides one useful approach to conceptualizing and measuring PROs from CGS. However, patient empowerment as operationalized in the GCOS-24 may not capture all the important PROs from CGS such as family functioning, social functioning, and altruism. It may be that a range of generic and condition-specific PROMs, including GCOS-24, PPC and IPQ and others will be required to capture all the important PROs from using CGS. We have yet to establish consensus about what PROs we aim to promote in CGS, and the best PROMs to measure these. Until we have consensus about these issues, it will be difficult to compare services and interventions on the basis of how effectively they deliver patient benefits. Findings in this study can inform these discussions.

Supporting Information
Additional supporting information may be found in the online version of this article at the publisher's web-site.