Introduction

Population-based reproductive genetic screening programmes offer individuals and couples without a specific family history or predisposition the opportunity to learn if they have an increased risk of having a child with a particular genetic condition or conditions in current or future pregnancies. Developments in genetic health care are constantly evolving, with screening technology becoming publically and privately available such as multiplex panels for carrier screening, whole exome/genome sequencing and noninvasive prenatal testing.

Genetic screening guidelines recommend that ethical and psychosocial aspects of screening programmes, such as informed choice,1, 2, 3 are evaluated in addition to technical aspects such as analytical and clinical validity.3 Screening decisions do not always have a clear medical imperative or benefit that outweighs the potential for psychological harms for the individual.4 Therefore, screening programmes should not directly encourage uptake of screening or influence reproductive decision making following a screen-positive result.5 The decision of whether to accept screening resides with the individual and it is therefore important that the decision is informed, based on correct knowledge and free of coercion from others.1, 3, 5 Thus, the success of population genetic screening programmes should not be based solely on the uptake of screening, and it is important to evaluate whether these programmes are in fact facilitating informed choice.2

Evaluation of genetic screening programmes requires accurate, validated measures of informed choice, which is a complex, multidimensional construct, defined in various ways throughout the literature.6, 7, 8, 9, 10, 11, 12 Informed choice may literally be measured through evaluation of knowledge and uptake of screening, to indicate whether the decision to accept or decline screening was ‘informed’. However, definitions of informed choice more commonly involve at least two elements: having a factual understanding and making a choice aligned with one’s values.12, 13 Some definitions distinguish between choice and decision making with the inclusion of deliberation and the behavioural implementation of the decision to reflect the cognitive processes involved in decision making.7, 10 However, underlying these various definitions is the same concept of an informed and autonomous choice.13, 14 The complexity of measuring informed choice, because of the different ways informed choice has been defined and measured in a range of screening settings, has been recognised.12, 14, 15, 16

Here we present the findings of a systematic review focussing on reproductive genetic screening in the general population, where screening was used to assess risk of a genetic condition in current or future pregnancies. A variety of conditions are included such as Down syndrome (DS) and single gene disorders. Screening choices in any of these scenarios can be complex, as individuals’ choices depend on the value they place on the benefits and harms of screening. Aspects of the condition also influence the way in which screening is offered, for example, screening a fetus for DS, a genetic but not usually inherited condition, is common and offered during each pregnancy in many countries. On the other hand, population genetic screening for carrier status, which provides an individual with reproductive risk, needs only to be performed once in a person’s lifetime and so can be offered anytime, including preconceptionally or during pregnancy.

Although there have been many studies evaluating population reproductive genetic screening, this review focusses on studies that specifically aimed to measure informed choice. This work will inform future research aiming to evaluate informed choice in genetic screening programmes. This systematic review investigates how informed choice has been measured in population reproductive genetic screening, what variables, methods and validation procedures have been used, and what rates of informed choice have been reported in these studies.

Materials and Methods

Databases indexing medical and psychosocial research were searched for studies from the past 30 years (1982 to September 2012), addressing informed choice in population-based genetic carrier or DS screening. The search strategy was based on: (1) genetic screening, (2) measurement, and (3) informed choice. Search terms were defined using key words and thesaurus terms specific to Medline, Embase, Cochrane Library, CINHAL, and PsycInfo databases that were searched using this strategy (see Supplementary Table 1).

Inclusion criteria

Studies were included if they met the following criteria: original research during the past 30 years; published in English; adults or pregnant women from an unselected, population-based sample; the study needed to be in the context of offering a genetic test to determine reproductive risk; the study needed to have reported on measures of informed choice obtained prospectively at the time of test offer, including uptake of screening and knowledge as a minimum requirement to evaluate informed choice. Genetic testing is defined here according to the Foundation for Genomics and Population Health as: ‘a test to detect the presence or absence of, or change in, a particular gene or chromosome’; the test did not need to have been a DNA or chromosomal assay, for example, it could be a full blood examination to identify thalassaemia carriers or ultrasound screening to identify markers for DS.17

Studies were excluded if they: were reviews, opinion or commentary, letters to the editor, guidelines, or used only qualitative methodology; used a hypothetical scenario; sampled selected subgroups of the population, such as those identified in high-risk categories through first-trimester screening for DS and trisomy 18, a group affected by, or with family history of, a genetic condition, or those undergoing preimplantation genetic diagnosis as part of in vitro fertilisation treatment; reported fetal anomaly scans, or other tests for structural anomalies without specifically reporting on risk of DS for the pregnancy. The cited references of the included papers were searched for additional relevant titles. Corresponding authors of the included papers and experts in the field were contacted for recommendations of any additional relevant papers.

Results

The review of titles and abstracts was conducted by AGA, and JE confirmed the application of the exclusion criteria in 10% of papers (see Figure 1). During data extraction, eight papers were excluded either because they measured informed choice many months after test offer18, 19 or did not report uptake of screening.20, 21, 22, 23, 24, 25 In total, 34 articles met the inclusion criteria; data were extracted for 33 articles as two studies reported the same data on informed choice.26, 27 The final inclusion of papers was confirmed by the authors JE and SAM.

Figure 1
figure 1

Flow diagram of systematic review search results. *All corresponding authors (30) and 3 experts were contacted. Of the 21 studies identified, 19 were from corresponding authors and 2 from experts. The four studies meeting inclusion criteria were identified by corresponding authors (two) and by experts (two).

Summary of studies

Of the 33 studies included in the review, less than half (n=11) of the studies were randomised control trials (RCTs, Tables 1, 2, 3, 4).11, 13, 26, 28, 29, 30, 31, 32, 33, 34, 35 Twenty-two studies were prospective,12, 28, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55 including cross-sectional,41, 42, 49 quasi-experimental,53 pilot,50 and longitudinal study designs.52 Of the RCTs, five involved an information or educational intervention,28, 29, 32, 33, 34 five randomised the screening offer,11, 13, 26, 30, 35 and there was one decision aid trial.31 The 33 studies were published between 1991,54 and 2012,36 with 25 studies published between 2002 and 2012. Three studies used the same cohort as in previous studies;13, 39, 46 in subsequent studies the sample was stratified differently,40 different measures were applied,11 and additional validation data were reported.45 Eight studies were conducted in continental Europe,11, 13, 38, 43, 44, 45, 46, 53 and 14 were from the United Kingdom.12, 26, 29, 30, 32, 33, 34, 35, 37, 39, 40, 41, 51, 54 Seven studies were from Australia,31, 36, 48, 49, 50, 52, 55 three from the United States,28, 42, 56 and one from Taiwan.47 Prenatal screening for DS was the most common screening programme evaluated (27 studies).11, 12, 13, 28, 29, 31, 32, 33, 34, 35, 37, 38, 39, 40, 42, 43, 44, 45, 46, 47, 48, 49, 51, 52, 53, 55, 56 Three studies offering population carrier screening in prenatal settings included screening single gene disorders: sickle cell disease and thalassaemia (haemoglobinopathies (Hb))26, 41 and cystic fibrosis (CF).30 Of the 33 studies reviewed, only one included partners of the pregnant women,30 and one offered CF carrier screening to men as well as women.54 Only two studies offered carrier screening to a nonpregnant population, for CF54 and fragile X syndrome (FXS).50

Table 1 Down syndrome screening studies measuring informed choice with the MMIC
Table 2 Down syndrome screening studies using measures of informed choice (other than the MMIC)
Table 3 Population carrier screening studies measuring informed choice with the MMIC
Table 4 Population carrier screening studies using other measures of informed choice

Uptake of screening varied in the DS studies from 13%44 to 91.6%.38 Six DS studies reported 100% uptake;42, 45, 46, 47, 48, 49 however, these studies only included women who had accepted screening as they were recruited at their screening consultation. High rates of screening uptake were reported for prenatal Hb screening in two studies (100%41 and 90.7%26). Similar uptake rates were reported in CF carrier screening in a prenatal setting (89–91%30) and a preconception setting (89%, family planning clinic site).54 However, the two preconception screening studies (CF54 and FXS)50 reported that uptake varied with recruitment method and site.50, 54 Studies reporting 100% uptake, however, should not be considered representative as they either only included individuals already having screening,47 or recruited from a site where women were presenting for prenatal screening.42, 45, 46, 48, 49

Measures of informed choice

Knowledge

As determined by the inclusion criteria, all of the 33 studies measured and reported on knowledge, a core component of informed choice. The knowledge items used in these studies measured knowledge recall, although some studies also measured perceived knowledge47, 56 and risk accuracy.56 Throughout these studies the type and number of knowledge items varied, and knowledge levels were not generally comparable across studies. Two validated DS knowledge scales were used in more than one study; six studies used the scale developed by Marteau et al12, 31, 39, 40, 49, 52, 55 and two studies used the scale by Dormandy et al,34, 44, 45 whereas the other DS studies each used their own specific knowledge scale. Studies screening for single gene disorders developed condition-specific knowledge scales for FXS,50 Hb,26, 41 and CF.30, 54 Knowledge results were reported in different ways using means,11, 12, 29, 32, 34, 36, 38, 40, 41, 51, 52, 53, 55, 56 medians,26, 28, 35, 36 dichotomised cutoffs of good knowledge,11, 13, 26, 31, 36, 39, 43, 44, 45, 46, 49, 55 categories,32, 37, 45, 46, 48, 50 and percentage of items answered correctly.28, 30, 42, 47, 54

Other measures of informed choice

Measures of informed choice other than knowledge were heterogeneous, involving a variety of different aspects such as perceptions of knowledge and information,32, 37, 43, 49, 56 the multidimensional measure of informed choice (MMIC),11, 12, 13, 26, 31, 35, 36, 39, 40, 41, 43, 44, 45, 46, 49, 52 perceptions of choice,32, 43, 49, 51, 56 attitudes and feelings towards screening,30, 43, 47, 50, 56 decisional conflict,11, 28, 31, 38, 56 satisfaction and decisional regret,11, 13, 28, 29, 32 and systematic and informed decision making.29, 42, 51

Perceptions of knowledge and information

Two studies measured perception of knowledge,47, 56 and five studies measured perception of the information provided.32, 37, 43, 49, 53 Measuring participants’ perception of knowledge and information is another way of assessing the quality and level of information provided in a screening study. These measures directly asked how women felt about the information provided,53 and if they were provided with enough information,49 with the purpose to inform how to present information in a meaningful way to aid participants’ understanding and facilitate informed choice. One study used a single item of perceived adequacy of information as their primary outcome measure, and showed that 60% of participants in both trial arms reported that there was adequate information to make a choice about screening.32 Another study showed informed choice was dependent on a measure of the quality of the information provided (P<0.001)43 and that less satisfactory information was related to the perception of the screening being compulsory rather than optional.43

Multidimensional measure of informed choice

The MMIC was the most widely used measure, and was used in 17 studies,11, 12, 13, 26, 31, 35, 36, 39, 40, 41, 43, 44, 45, 46, 49, 52, 55 with the majority (n=14) assessing informed choice in prenatal screening for DS in the United Kingdom, continental Europe, and Australia (Table 1).11, 12, 13, 31, 35, 39, 40, 43, 44, 45, 46, 49, 52, 55 The MMIC was adapted for use in population carrier screening for Hb in pregnancy by using a specific knowledge scale for Hb,26 and for low literacy populations,41 as well as for FXS carrier screening in a preconception setting (Table 3).36, 50

The MMIC is a theory-based measure, modelling the attitudes–behaviour domain on the Theory of Planned Behaviour.57 The MMIC is based on the definition of informed choice being ‘one that is based on relevant knowledge, consistent with the decision-maker’s values and behaviourally implemented’.12 Some of the studies in this review were published before the MMIC, and hence they did not use the same terminology for defining informed choice as the authors of the MMIC.29, 30, 33, 34, 51

The MMIC combines knowledge with ‘value-consistency’ comprising attitudes and screening behaviour into a single measure. A choice is value consistent when an individual has positive attitudes and accepts screening or has negative attitudes and declines screening. A value consistent choice becomes an informed choice when made with good knowledge. All other combinations are classified as uninformed choices.

Each continuous knowledge and attitude scale is dichotomised into ‘good’/‘poor’ knowledge and positive/negative attitudes for which there is no commonly agreed upon acceptable standard. Good knowledge was defined in various ways using the median12 and a guess-corrected midpoint.13 Validation of the ‘good’ knowledge cutoff was generally lacking, with only two studies validating the cutoff using interview data.12, 41

Informed choice varied for DS screening from 9.7%55 to 74.5%.49 In prenatal Hb carrier screening, 30.4–50% were classified as having made an informed choice.26, 41 Low rates of informed choice were generally because of poor knowledge. In preconception carrier screening for FXS, only 27% of participants were classified as having made an informed choice.36 In this setting, uninformed choices were due to low value consistency, largely because screening required people to return for a separate appointment to have blood taken, resulting in women having positive attitudes towards screening but not attending the return appointment to provide the blood sample for screening.50

Perception of choice

Five studies incorporated measures of participants’ perception of their choice about the screening test.32, 43, 49, 51, 56 When screening is offered as part of ‘routine’ care, it can be interpreted by participants as being compulsory. However, ideally, screening should be offered in a way that promotes autonomy; being free from coercion. Assessing the subjective perception of choice is one way of assessing the degree of autonomy that individuals exercised when making their choice.

The items measuring perception of choice varied across studies; four studies used a single item,32, 43, 49, 56 and one study included a single item within a systematic decision-making scale.51 All items asked women, in different ways, whether they thought the screening was optional, or who should make their choice about screening.

In prenatal screening for DS, almost half (46%) of women thought screening was routine,48 and most (82%) thought it was compulsory.43 However, most women reported believing that the choice should lie with the individual offered screening.56 Perceptions of choice are likely to be context specific; for example, in countries where screening is funded by the government or is socially accepted, testing might be seen as part of ‘routine’ care.38, 46

Attitudes and feelings towards screening

Of the 33 studies, 5 used an attitudes scale different from that used in the MMIC,30, 43, 47, 56 or the same scale but used independently of the MMIC.50 These studies used a variety of different items within the attitudes measures; three studies used various items regarding participants’ general feelings towards screening,30, 50 and whether or not the participant was in favour of DS screening.43 One study measured a broad range of attitudes and values towards screening such as ‘faith or fatalism’.28 Two studies also measured participants’ acceptance and values relating to screening, termination of pregnancy, and having a child with the condition being screened.30, 56 One study did not report the items or results of the attitudes measure.47

Attitudes regarding screening are particularly important to measure as it has been argued that screening choices ought to be aligned with the morals and values of the individual.3 These measures however need to be more consistent and validated within the sample population.

Decisional conflict

The decisional conflict scale (DCS) was used to evaluate participants’ feelings regarding the choice outcome in five studies; four RCTs13, 28, 31, 38 and an adapted version was used in one prospective study.56 The DCS was designed as a general scale, not for a particular condition, and validated with health-care consumers deciding about hypothetical scenarios of influenza vaccination and breast cancer screening.10 The scale comprises five subscales (information, values clarity, support, uncertainty, and effective decision) measuring the outcome of how an individual feels about a health-related decision, rather than evaluating whether the decision was an informed one. Decisional conflict is defined as ‘a state of uncertainty about the course of action to take’ that can occur when decisions are difficult, resulting in decisions that are not effective, and where an individual may feel uncertain about their decision. Rather, an effective decision is defined as one that is ‘informed, consistent with personal values and acted upon’.10 Decisional conflict is important in informed choice evaluation because it directly measures participants’ perspectives of how certain they feel about the decision and their perceived autonomy in making the decision.

Results can be compared across the studies as the DCS is a standardised measure on a scale of 0–100 with validated cutoff scores for low decisional conflict (scores <25) and increased decisional conflict (scores >37.5).10 Decisional conflict was reported within the normal range in two studies (mean ranging from 1.65 to 23.9)28, 31 with another study also demonstrating a relationship between low decisional conflict and high levels of knowledge (OR=1.31, 95% CI, 1.26–1.37, P<0.001).38

Satisfaction with decision and decisional regret

Five studies measured satisfaction with decision or decisional regret.13, 28, 29, 32, 50 Satisfaction and decision regret scales are explicit measures of how people feel about their choices,4 and can be used to measure the outcome of decision making.4, 58 An informed decision may lead to better outcomes59 such as increased satisfaction, or low regret with the decision.4, 10, 60 Decision regret can be defined as ‘remorse or distress over a decision’ that can stem from ‘knowledge that the choice made was nonoptimal’.4 Decision satisfaction and regret, however, are not the same as making an informed choice; someone may have low satisfaction but have made an informed choice.59

Two studies used scales of decision satisfaction,13, 29 one study used the decision regret scale,28 whereas two others used individual items to measure satisfaction with how the choices were made.32, 50 In studies where satisfaction with screening choices was measured, most women were very satisfied with their choices,29 and did not regret their decisions.28, 50 One study showed significantly higher satisfaction scores in people making informed decisions compared with uninformed decisions (t=4.9, P<0.001, t=2.1, P<0.05).13

Measures of systematic decision making

Three studies used measures of systematic, informed decision making.11, 29, 51 According to a rational decision-making perspective, an informed decision needs to involve a systematic process of weighing up the advantages and disadvantages of the choice, that is, deliberating, as well as the decision being informed in order to have the best outcome.7 Informed decision making is differentiated from an informed choice, where the former involves the process of deliberation. Decisions that are not systematic are often made using mental short cuts or heuristics that can lead to biases in decision making and poorer quality decisions.61 To account for this aspect of decision making, an 18-item systematic decision-making scale was developed,29, 51 and a 6-item deliberation scale was used with the existing MMIC.11 These measures have been well validated,51 indicating that each accurately measures the systematic decision-making construct. However, these scales have not been widely used to date.

Discussion

This systematic review evaluated how informed choice has been measured in population reproductive genetic screening. Studies offering population genetic screening should endeavour to measure informed choice as well as uptake of screening.2 The studies identified in this review demonstrate the array of methods used to measure informed choice in population reproductive genetic screening. Our findings highlight well-validated measures, and also some limitations and inconsistencies in evaluating informed choice in this setting.

The concept of informed choice is a multidimensional construct and has been defined in a numbers of ways.11, 12, 32, 33, 48 Measuring knowledge is integral for informed choice evaluation, yet there has been little development and validation of knowledge measures. A heterogeneous range of knowledge items were used in the studies reviewed. Even the reporting of validated knowledge scales varied within DS screening where there has been much development of informed choice measurement.12, 31, 39, 40, 46, 49

There are no agreed thresholds or minimum standards for the knowledge an individual is required to have to make an ‘informed’ choice. As such, a variety of methods were used to classify participants as having ‘good enough’ knowledge for an informed choice. This is problematic because when good knowledge is defined differently, it is not possible to compare the proportion of informed choice across studies. Some measures, such as the MMIC, required the scales to be dichotomised. Dichotomising continuous scales into binary or categorical data has been criticised and is not recommended by some editorial guidelines.62

The authors of the MMIC drew on work by O’Brien-Pallas, integrating the definition of informed choice with the theoretical perspective of rational behaviour.57, 63 The MMIC has been validated and applied widely in DS screening,11, 12, 13, 31, 35, 39, 40, 43, 44, 49, 52, 55 and now has been used in Hb26, 41 and FXS carrier screening.36 In nonpregnancy screening, the construct validity of the value-consistency domain needs to be validated because in nonpregnancy carrier screening individuals do not need to make a decision when offered screening, rather they may defer the decision to another time in their life.36, 64 Therefore, the MMIC needs to be further explored and undergo validation in screening settings other than prenatal screening.

Another approach to encapsulate the informed choice construct is to use the DCS with a knowledge measure. The DCS is a well validated measure10 used in five studies measuring informed choice.13, 28, 31, 38, 56 Incorporating a knowledge scale as well, perhaps using ‘good’ knowledge, combined with a low decisional conflict could be one way of defining informed decision making. This way, participants’ perspectives of whether they made an informed and free choice, and whether they were supported, are taken into consideration. These aspects are similar to other concepts such as the perception of choice and satisfaction, yet more research and agreement is needed to determine how these measures should be used and how they relate to evaluating informed choice.

When using the MMIC, or any measure out of its original context, the limitations of the measure need to be recognised until the measure has been validated within the new setting. The MMIC has been criticised for oversimplifying informed choice into a binary variable, potentially incorrectly classifying choices that are informed.52, 65 The MMIC also does not take into account the cognitive processes involved in making a decision, such as deliberation.7, 11 Although there are limitations to the MMIC, as for any measure, the MMIC remains the only single, multidimensional measure of informed choice for genetic screening. The MMIC is a relatively short instrument to administer and easy to use, with many published studies detailing how to apply and interpret the measure. However, the MMIC does require a knowledge measure that is specific to the condition being tested.

In the future, a knowledge measure may need to address the general phenotype (multiple congenital anomalies, developmental delay, reduced life expectancy, and so on),66 rather than be condition specific. With multiplex genetic technologies now enabling screening for multiple conditions,67 measuring informed choice with condition-specific knowledge scales is problematic. With this technology becoming available directly to consumers, there is a need for more research about informed choice in the general population, and outside the prenatal context. There is also a need for ongoing evaluation and publication of informed choice in existing programmes, with implementation of changes where needed to facilitate informed choice. Moving forward, qualitative and mixed methods studies could be used to help understand informed choice in genetic screening and further inform how best to measure informed choice in the future.

Although we have applied rigorous systematic review methods, there are some limitations to this study. As informed choice is defined in a number of ways, it was difficult to develop a search strategy to locate all studies for this review. The search terms may not have included every definition and possible way that informed choice could be indexed in the databases. To account for this, we used citation tracking and contacted corresponding authors and experts in the field, and this yielded 13 additional relevant papers. The studies identified in the review came from a limited number of countries; almost half of the studies (14/33) were from the United Kingdom where informed choice is a key component of prenatal screening guidelines,68 and where the widely used measure, MMIC, was developed. Informed choice may be a component of guidelines and the evaluation of screening programmes in other countries, but the evaluations may not be published and hence not included in this review. Because of the heterogeneity of included studies, we were only able to conduct a narrative synthesis. However, our review has highlighted important challenges for this field of research and suggests the need for further work in the area especially with the expansion of carrier screening and noninvasive prenatal screening.69, 70

The focus of this review was on the measurement tools used in population genetic screening for reproductive risk. For this reason the review did not include guidelines or qualitative studies. However, qualitative studies are important for exploring how individuals offered screening perceive informed choice. In fact, qualitative research could contribute to understanding issues of informed choice evaluation such as what is good knowledge, and whether individual’s perceptions of informed choice reflect how informed choice is measured. This approach could be especially useful for new screening programmes that have not been evaluated.

This review has highlighted that there is no gold standard for evaluating good knowledge or informed choice in population reproductive genetic screening. Knowledge and autonomy are key aspects of informed choice; it is important to define what is important for participants to understand when accepting or declining screening and for individuals to understand that screening is optional and their own personal choice. A dearth of studies involving future fathers and partners suggests more research is needed to better understand their role, and women’s perspectives about this in making choices about screening. There are a variety of informed choice measures with variable quality that would benefit from further development and evaluation. Results of informed choice evaluations must be presented in a meaningful way to indicate the programme’s efficacy to facilitate informed choice. Particular attention ought to be directed towards addressing how ‘good’ knowledge and informed choice are defined and measured from the outset of a screening programme.