Communication and risk presentation in genetic counseling

Communication and risk presentation in genetic counseling

Patient Education and Counseling 61 (2006) 126–133 www.elsevier.com/locate/pateducou Communication and risk presentation in genetic counseling Develo...

167KB Sizes 2 Downloads 32 Views

Patient Education and Counseling 61 (2006) 126–133 www.elsevier.com/locate/pateducou

Communication and risk presentation in genetic counseling Development of a checklist Mirjam Fransen a,1, Ree Meertens a,*, Connie Schrander-Stumpel b a

Department of Health Education and Promotion, Maastricht University, P.O. Box 616, NL 6200 MD Maastricht, The Netherlands b Department of Clinical Genetics, University Hospital Maastricht, Maastricht, The Netherlands Received 13 April 2004; received in revised form 23 February 2005; accepted 26 February 2005

Abstract Objective: Genetic counseling involves advising people about genetic disorders, genetic risks and preventive measures, and guiding them in the process of decision-making. It remains unclear what effect certain aspects of risk communication have on outcomes like risk perception and the decisions people take. In order to examine this relationship between process and outcomes, the present study aimed to develop a reliable checklist to assess aspects of risk communication in genetic counseling. Methods: A preliminary checklist was developed on base of literature and tested for manageability in a pilot study. The checklist was adapted and tested for inter-observer reliability in 14 and 56 genetic counseling sessions. Inter-observer reliability was measured by computing Kappa and proportions of agreement. Results: Most of the items of the last version of the checklist had Kappa values between 0.4 and 1, which means that the inter-observer reliability for most items was sufficient or good. On the majority of items, the observers showed more than 80% agreement. Conclusion: The checklist we have developed has adequate inter-observer reliability and may be applied in future studies to assess risk communication aspects in genetic counseling. Practice implications: This checklist could be a useful instrument to identify the relationship between aspects of risk communication in genetic counseling and outcomes. Showing which communication skills and risk presentations affect client’s feelings and decisions may help to improve genetic counseling. # 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Genetic counseling; Risk communication; Risk perception; Decision-making; Checklist

1. Introduction In recent years, knowledge about human genes and the genes predisposing people to various diseases has grown rapidly. This has resulted in an increased availability of genetic tests and a growth in genetic counseling. Increasingly, people can be informed about their risk of developing a genetic disorder, the consequences of this disorder, the probability of transmitting it and about ways in which this may be prevented [1]. Some of the tasks of the genetic counselor, usually a clinical geneticist (medical specialist) * Corresponding author. Tel.: +31 43 3882407; fax: +31 43 3671032. E-mail address: [email protected] (R. Meertens). 1 Present address: Department of Public Health and the Department of Obstetrics and Gynaecology, Erasmus MC, Rotterdam, The Netherlands.

or a clinical genetic nurse, include helping clients to understand their personal genetic risk and guiding clients in the process of decision-making. So far, little is known about the relationship between the characteristics and outcomes of genetic counseling [2–4]. What we do know is that people have difficulties understanding risks in general [5–7]. Genetic risks are often underestimated or overestimated, so people may base their decisions on incorrect assumptions. To enhance risk perception among clients and improve their ability to make well-founded decisions, it is important to gain insight into the relationship between aspects of risk communication in genetic counseling and outcomes like risk perception and decision-making [3,4,8–10]. In order to examine this relationship, different aspects of genetic counseling should be assessed first.

0738-3991/$ – see front matter # 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.pec.2005.02.018

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

Hence, the main goal of the present study was to develop a reliable checklist that can be used to assess aspects of risk communication in genetic counseling. This checklist should be applicable in future studies on risk communication and it should contribute to the improvement of risk communication in genetic counseling. Besides developing the checklist, this study also aimed to determine which risk communication skills were actually applied by the genetic counselors we observed. This should provide insight into how counselors present genetic risks and how they guide clients in decisionmaking. Before we describe the procedure used to develop the checklist, we first discuss various aspects of risk communication that were used as a basis for developing the checklist. In Section 3, we describe the development of the checklist, the pilot study where the checklist was tested for clarity and studies 1 and 2 where the checklist was tested for inter-observer reliability. The last sections of this article present and discuss the inter-observer reliability of the checklist, aspects of risk communication actually observed during the study and implications for future research (Note: The checklist can be requested from the authors).

2. Aspects of risk communication 2.1. Risk presentation A review of studies assessing risk communication aspects in health care shows there are many different ways of presenting genetic risks. The most important methods are outlined below. Risk presentation: First of all, genetic risks can be presented in percentages (25%) or proportions (1:4). These presentations can be assisted by comparing a genetic risk to other risks, such as risks in the general population, risks at different ages [11] or risks in other genetic conditions [12,13]. Besides using numbers, it is also possible to present genetic risks in words. A counselor can verbally present a risk by saying that the risk is low or that it is unlikely that a genetic disorder will develop [12]. Different studies show that risk presentation in percentages or proportions has pros and cons, depending on several factors [6,7,14–16]. Risk comparisons could be helpful in conveying the significance of genetic risks or provide a better perspective on risks [17]. However, comparing a risk with risks with (completely) different characteristics may clarify nothing, and may lead to the idea that the counselor is trying to trivialize the risk [18]. It also remains unclear if numerical risk presentations have more positive effects on outcomes like risk perception or trust in genetic counselors than verbal expressions. Therefore, further research is suggested [10,12,19]. Risk framing: Risks and preventive measures can be presented or framed in negative or positive terms. A risk is framed in negative terms if the probability of developing a particular disorder is given, and is framed in positive terms if

127

the client is informed about the probability that the genetic disorder will not develop [20]. Studies on framing show that framing risks effects the perception of genetic risks. The answer, how risks should be framed remains speculative, however. According to Shiloh and Sagi [12], the right way of framing risks must be tailored to the individual client and the specific counseling situation. 2.2. Guiding decision-making under uncertainty Decision analysis: Genetic counseling often involves uncertainties. Genetic risks are probabilistic and complete certainty is never achieved, while preventive measures or tests often involve uncertain consequences as well [21]. To guide clients in the process of making decisions under conditions of uncertainty, counselors can use a decision analysis. This is a mathematic tool, developed by Pauker and Pauker [22] that attempts to emulate the human decisionmaking process. The method can be summarized in four basic steps. First, the counselor has to identify and define exactly what the decision problem is. Pregnant women with a risk of at least 0.5% of having a child with Down’s syndrome could have trouble deciding between having an amniocentesis and facing the risk of miscarriage (0.5%) and having no amniocentesis and running the risk of a child with Down’s syndrome. In the second step, this problem is structured by dividing all possible options and consequences in a decision tree (Fig. 1). For each chance event in the tree structure, the probability of that event’s occurrence must be specified in step three. The likelihood of any of the possible outcomes of the decision is equal to the product of the probabilities along the path leading to that outcome. The attitudes or values clients assign to the different outcomes must be measured as well. To assess the women’s attitudes towards these outcomes, they must assign a relative value, representing the relative desirability or burden of each alternative. In the final step, these attitudes are combined with the likelihood of every outcome in order to select a preferred course of action. Directiveness: The client’s autonomy in decision-making is an important issue in genetic counseling and could be promoted by non-directive communication. However, directivity and non-directivity are very difficult to define [23]. In the checklist developed in the present study, directivity is specified by three different aspects: the level of neutrality [24], the level of reinforcement [25], and the level of coercion [26]. Neutrality literally means that the counselor is impartial and stays as close as possible to the client’s point of view [24]. Michie et al. [25] defines directivity as advice, expressed views about or selective reinforcement of clients’ behavior, thoughts or emotions. According to this view, a counselor is being directive, for example, when he or she states that the choice of having an amniocentesis is the best decision the client could make. Contrary to Michie and Van Zuuren, Kessler [26] states that reinforcement and deviations from neutrality can hardly be

128

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

Fig. 1. Example of a decision tree for pregnant women (and their partners) who are at increased risk of having a child with Down syndrome (abnormal karyotype) and experience difficulties in deciding between an amniocentesis and no amniocentesis.

considered directive. In his view, directiveness must be seen, as a form of persuasive communication in which there is a deliberate attempt (through coercion for example) to undermine the individual’s autonomy and compromise his or her ability to make an autonomous decision. This does not mean that a counselor is non-directive when he or she is not persuasive. Non-directiveness, according to Kessler [23], has a goal and direction and is a strategy to assist clients to achieve a personal decision, discussing all relevant options. 2.3. Best way of risk presentation The question what the best way is to present risks is very complex, because one has to define first what the desired outcome of the process is; and defining desired outcomes may conflict with being non-directive. According to Kessler [23] no definition of non-directiveness will apply to every single case. He emphasizes that the attempt to help clients to reach personal and emotionally charged decisions requires human experience and counseling skills. So, we suggest that the most appropriative answer is that the best way of risk communication depends on the individual client and the aims of genetic counseling.

risk presentation in proportions, percentages or words; risk comparisons; risk framing; decision analysis; directivity. All aspects were transformed into measurable items and arranged in a logical order. The checklist was supplemented by items relating to general communication skills such as starting and ending the conversation, using technical jargon and dealing with emotions and non-verbal communication [27–30]. These aspects play an important role in the communication between counselor and client and could influence outcomes of genetic counseling as well. Because the content of these aspects is considered well known, this is not discussed further in this paper. Every item in the checklist had two answering options: yes and no. Since some skills, like the commitment and expertise of the genetic counselor, are difficult to measure objectively, a separate part was added to the checklist, in which observers could give their subjective impression on a scale from 1 to 7. In addition, a manual for observers was developed, providing explanations of the items and instructions for completing the checklist. The checklist was then tested for manageability in a pilot study (n = 12), adjusted and tested for inter-observer reliability in study 1 (n = 14), and again adjusted and tested for inter-observer reliability in study 2 (n = 56).

3. Methods

3.2. Pilot study

3.1. Development of the checklist

The first version of the checklist was tested for clarity and manageability in a pilot study at the Department of Clinical Genetics, University Hospital Maastricht in the Netherlands. Four different genetic counselors (clinical geneticists) were

The concepts measured in the checklist were the various aspects of risk communication mentioned earlier in this paper:

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

observed in 12 genetic counseling sessions by the first author of this paper. For practical reasons, it was decided to exclude oncogenetic counseling sessions. The subjects of the observed counseling sessions were risks for various genetic disorders concerning offspring. In six sessions, the risk for Down’s syndrome on the basis of maternal age was discussed. On base of the findings in this pilot study, the checklist was adjusted. The yes/no answering option was extended with ‘not applicable’ and several items were more extensively explained in the manual. 3.3. Study 1 After the pilot study, the checklist was completed by two observers in 14 genetic counseling sessions and tested for inter-observer reliability. Six different genetic nurses and clinical geneticists were observed in this stage. The observer pair was formed by the first author of this paper and a genetic nurse who did not belong to the research sample. Inter-observer reliability was measured by computing Kappa. Kappa is a measure that presents the actual agreement between observers as a proportion of the potential agreement, after correction for the agreement based on coincidence [31]. During the observations, the observers were found not to have had enough time to complete the checklist correctly, especially in the shorter sessions. Hence, the checklist needed to be shortened. Mainly the general communication aspects were merged or omitted, as they were not the main theme of the checklist. The remaining items were improved on the basis of their inter-observer reliability. More than half of the items had a low or moderate Kappa and were adjusted. Items on risk presentation in percentages or proportions were divided over two columns; a distinction was made between framing the risk of the disorder and framing the risk associated with preventive measurements; the items concerning the decision analysis were improved by clarifying them and providing more examples and/or better instructions. Besides the adjustments to items, the answering options were also adjusted. The answer ‘not applicable’ was found to be useless in many items. Only the items that had proved to be really not applicable in one or more of the 26 sessions retained this answering option, while the other items reverted to the dichotomous ‘yes/no’ options. 3.4. Study 2 After these adjustments, the final version of the checklist was once more tested for inter-observer reliability. To determine how many sessions should be observed to establish a reliable measure of agreement, a power calculation was used. This power calculation was based on the standard error of Kappa in the 14 counseling sessions that already had been observed. Most of these Kappas had a standard error of 0.20. This means that Kappa itself was not reliable enough, because the reliability interval was simply

129

too wide. To determine the number of observations needed to calculate reliable Kappas in the second part of the study, the standard error was set at 0.10. Halving this standard error meant that the number of observations in the second part of the study had to be 4  14 = 56 (S.E. = 1/HN). Therefore, nine different genetic nurses and clinical geneticists were observed in 56 counseling sessions. All observations took place at the Department of Clinical Genetics, University Hospital Maastricht in the Netherlands. To allow generalization to other observers, the checklist had to be completed by different pairs of observers instead of one fixed pair. Six different pairs each observed about 8–12 counseling sessions. For logistic and practical reasons, the first author of this paper was a constant observer in all sessions, while six other observers (students in health sciences) took turns. 3.5. Analysis As mentioned before, inter-observer reliability was measured by computing Kappa. In addition to Kappa, the proportion of answers that the observers agreed about was computed to estimate the inter-observer reliability of each item in the checklist. This was done because the Kappa calculations in the first phase of the study had shown that a few items yielded an overabundance of one particular combination of scores (e.g. yes–yes). Since Kappa corrects for agreement based on coincidence, the accuracy of Kappa is actually based on the smallest group of score combinations. These combinations, however, do not provide a good impression of the agreement between observers on this item, because there are simply not enough combinations left. If most observers choose ‘yes’, for example, this definitely indicates that they almost always agree about a ‘yes’ score. But because they have hardly used the other scores, it is unclear whether they would also agree about the ‘no’ or ‘not applicable’ scores. We, therefore, also calculated the agreement proportions to give a more qualified impression of the agreement between observers in the items with skewed scores. The proportion of agreement was calculated by dividing the number of agreed-upon scores by the total number of scores. Since the scores of the subjective impressions are ordinal, these agreements were determined by the (quadratic) weighted Kappa. In this approach, the weight of a nondiagonal cell is smaller, the less the cell deviates from the diagonal in the table.

4. Results 4.1. Genetic counseling sessions A substantial number of the sessions observed in this study involved genetic counselors communicating with pregnant women who were at increased risk for fetal Down’s

130

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

Table 1 Kappa, agreement proportions and application of risk communication skills Item

Kappa

Agree

Yes

No

N.A.

Opening the conversation Asking for or referring to the client’s request for help Announcing what will be discussed in this session

0.82 0.21

0.91 0.89

0.49 0.07

0.51 0.93

– –

Giving information Disorder Explicitly mentioning the disorder Explaining the genetic background of the disorder

0.30 0.76

0.87 0.87

0.90 0.42

0.10 0.54

– 0.04

0.66

0.98

0.97

0.03



0.86 0.84 0.59

0.95 0.93 0.77

0.26 0.65 0.09

0.74 0.35 0.35

– – 0.56

0.57

0.77

0.06

0.37

0.57

0.82 0.89

0.91 0.93 0.93 0.89 1.00 0.94

0.06 0.46 0.01 0.04 0.04 0.03

0.60 0.23 0.67 0.64 0.96 0.97

0.33 0.31 0.32 0.32 – –

0.86 0.91 0.85 0.78 0.65

0.91 0.95 0.93 0.91 0.80

0.36 0.20 0.09 0.04 0.18

0.42 0.57 0.70 0.74 0.60

0.22 0.23 0.21 0.22 0.22

0.74

0.89

0.05

0.72

0.23

0.53 0.23

0.76 0.89

0.55 0.07

0.45 0.93

– –

0.44 0.60

0.82 0.80

0.20 0.54

0.80 0.46

– –

0.45 0.55 0.74 0.82 0.46 0.37

0.80 0.82 0.91 0.91 0.93 0.95

0.78 0.27 0.22 0.54 0.07 0.04

0.22 0.73 0.78 0.46 0.93 0.96

– – – – – –

0.35 0.77 0.58 0.42 0.68 0.28

0.66 0.91 0.84 0.71 0.84 0.77

0.49 0.74 0.74 0.44 0.53 0.19

0.51 0.26 0.26 0.56 0.47 0.81

– – – – – –

0.09 0.13

0.78 0.64 0.98

0.12 0.29 0.02

0.88 0.71 0.98

– – –

0.66

0.68

0.32



Medical jargon Adapting medical jargon to the language of the client Risk presentation Risk of disorder Presenting the risk of the disorder in percentages Presenting the risk of the disorder in proportions Comparing the personal risk with the risk without genetic transmission in percentages Comparing the personal risk with the risk without genetic transmission in proportions Presenting the risk in percentages as an age related risk Presenting the risk in proportions as an age related risk Comparing risks in percentages at different ages Comparing risks in proportions at different ages Comparing the risk with other risks in percentages Comparing the risk with other risks in proportions Risks associated with preventive measures Presenting the risk associated with preventive measures in percentages Presenting the risk associated with preventive measures in proportions Comparing the risk with the risk of the disorder in percentages Comparing the risk with the risk of the disorder in proportions Comparing the risk with the risk associated with other preventive measures in percentages Comparing the risk with the risk associated with other preventive measures in proportions Presenting the risk in words Using visual methods Uncertainties Presenting the uncertainty associated with the risk Presenting the uncertainty associated with preventive measures Framing Framing the risk of the disorder in negative terms Framing the risk of the disorder in positive terms Framing the risk of the disorder in both negative and positive terms Framing the risk associated with preventive measures in negative terms Framing the risk associated with preventive measures in positive terms Framing the risk associated with preventive measures in both negative and positive terms Decision analysis Discussing problems of decision-making Discussing possible decisions Structuring possible outcomes of decisions Presenting probabilities of possible outcomes Asking the clients value towards possible outcomes Giving advice based on probabilities and values Directivity Expressing own opinion Reinforcing choices positively or negatively Urging the client to make a certain decision Dealing with emotions Discussing emotions

a

0.78 1.00 a

a

0.35

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

131

Table 1 (Continued ) Item Looking for eye contact Ending the session Verifying whether the client has understood the information given Offering printed information Offering psychological counseling

Kappa

Agree

Yes

No

N.A.

a

0.98

0.98

0.02



0.32 0.62

0.66 0.91 1.00

0.56 0.13 0.00

0.44 0.87 1.00

– – –

a

Agree: agreement between observers (in proportions); yes: risk communication aspect is applied (in proportions); no: risk communication aspect is not applied (in proportions); N.A.: risk communication aspect is not applicable (in proportions). a Kappa could not be computed because of an asymmetric table.

syndrome because they were 36 years or older. The subjects discussed in the remaining sessions were variable. Clients who had certain disorders themselves, such as coronary or muscle disorders or had relatives with such disorders, wanted to know the risk of developing the same condition for themselves, their children or future children. In a few sessions, the genetic counselors talked with pregnant women about the possible risk of using medication during pregnancy. Couples, who had a child affected by a neural tube defect (e.g. spina bifida), wanted information about the recurrence risk and to discuss their options, including prenatal diagnosis. Also the autosomal recessively inherited disorder cystic fibrosis (CF) was the subject in a few sessions. 4.2. Inter-observer reliability of the checklist Table 1 presents the Kappa values and proportions of agreement for the various items in the final version of the checklist. Most of the Kappa values were between 0.4 and 1, which means that the inter-observer reliability for most items was sufficient or good [31]. The proportions of agreement ranged from 0.64 to 1.00. On the majority of items, the observers showed more than 80% agreement. Comparing the Kappas per item with the agreement proportions shows that many of the items with a low Kappa (<0.4) had rather high agreement proportions. All of these items showed a skewed distribution of answers. A comparison of the Kappa values in study 2 with those obtained in study 1 reveals a general improvement. It was mainly the items on risk presentation in study 2 that had a better inter-observer reliability than in study 1. Kappas of the items on directivity could not be compared, because they could not be measured in study 1 and/or study 2 due to asymmetric tables. The agreement proportions for these items

remained sufficient or good, however. The items in the decision analysis generally were found to have improved as well. The five weighted Kappas of the subjective impressions presented in Table 2 were considerably lower than the weighted Kappas of the earlier version of the checklist. The agreement proportions of these items were not very high either, between 0.18 and 0.45. A closer examination of the data shows that the differences between the scores per subjective item were not very large, usually 1 or 2 points on the scale from 1 to 7. Some observers often chose the more extreme scores, like 1 or 7, while others often chose 2 or 6. When the scale was divided into three parts, new calculations with the same data show, as could be expected that the agreement between observers is higher in this reduced scale. Apart from directivity (agreement 0.43), all items had an agreement proportion between 0.64 and 0.73 this time. 4.3. Aspects of risk communication Table 1 also shows which risk communication aspects were most frequently addressed by the genetic counselors in study 2 and which were addressed least. The table shows that risks of genetic disorders were mainly presented in terms of proportions and were rarely compared with other risks or risks at different ages. When risks associated with preventive measures were mentioned by genetic counselors, percentages were used slightly more often than proportions. Genetic counselors did not use only numbers to express risks. In fact, more than half of the risks were presented in words. In the majority of these cases, however, words were used in addition to numbers. It is clear that genetic disorders were mainly framed in negative terms by the counselors. Risks associated with preventive measures, however, were framed in both negative and positive terms.

Table 2 Weighted Kappa and subjective observations Item

Kappa (W)

1

2

3

4

5

6

7

Commitment Threat Expertise Directivity Adaptation

0.42 0.17 0.33 0.20 0.39

0.00 0.23 0.00 0.07 0.00

0.05 0.46 0.00 0.39 0.04

0.12 0.21 0.02 0.14 0.06

0.30 0.07 0.21 0.18 0.36

0.32 0.02 0.43 0.14 0.31

0.18 0.01 0.31 0.08 0.17

0.03 0.00 0.03 0.00 0.06

For every subjective impression the proportion per score is given, starting with score 1 (not at all committed/threatening/skilled/directive/no adaptation at all) and ending with score 7 (very committed/threatening/skilled/directive/a lot of adaptation).

132

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

Although the decision analysis was not applied sequentially as described in the literature, counselors usually did discuss possible decisions and outcomes, presented probabilities of outcomes and asked the clients about their attitude towards certain outcomes. The data further show that counselors did not urge clients to make particular decisions that they rarely expressed their own opinion, and that they did not reinforce choices. General communication skills were usually not practiced by the genetic counselors. Only in half of the counseling sessions the counselor asked the client to state his or her request for help or referred to the request at the beginning of the session or verified whether the client had understood the information at the end of the session. The subjective impressions that the observers had about the counselors in study 2 are listed in Table 2. According to the observers, the degree of commitment, the perceived expertise of the counselors and the way the counselors adapted their language to that of the clients ranged from moderate to high. Counselors did not make a threatening impression and did not seem to be very directive either.

5. Discussion and conclusion 5.1. Discussion As we experienced in the development of the checklist, risk communication is difficult to measure. Items concerning decision analysis, for example, cannot be formulated too specifically, because the checklist must be applicable in various types of genetic counseling sessions. The items should not be too general either, because subjective interpretations make the checklist less reliable. Other items, for example, ‘framing the risk in positive terms’, are difficult to measure properly because they require close attention on the part of the observers. We did not explicitly check whether all elements of risk communication observed in the sessions could be coded in the checklist. The subjective impression of the observers, however, was that there were hardly any subjects discussed in most of the observed sessions that could not be addressed in the checklist. Besides these difficulties of the checklist itself, it is important to note that the inter-observer reliability of some items of the checklist was hard to interpret because of a skewed distribution of scores. Another disadvantage of these skewed distributions is that they also make it more difficult to draw conclusions about the effects of some aspects of risk communication. When counsellors, for example, are always looking for eye contact, it is hard to relate outcomes to this aspect. As the results show, the second version of the checklist generally had a better inter-observer reliability than the first version, except for the five subjective impressions. It seems that the seven categories in the scale were too narrow and

were not satisfactorily specified, because the level of agreement rose when the scale was divided into three parts. A three- or four-point scale to measure subjective impressions in the checklist would have been preferable in study 2, because it would probably yield enough information on the subjective impressions of the different observers and would be more reliable at the same time. 5.2. Conclusion The main goal of the present study was to develop a reliable checklist that can be used to assess aspects of risk communication in genetic counseling. This checklist should be applicable in future studies on risk communication and it should contribute to the improvement of risk communication in genetic counseling. Besides developing the checklist, this study also aimed to determine which risk communication skills were actually applied by the genetic counselors we observed. This should provide insight into how counselors present genetic risks and how they guide clients in decisionmaking. According to the results of this study, this goal is achieved. The checklist we have developed has adequate inter-observer reliability and can be applied in future studies to assess risk communication aspects in genetic counseling. Other important conclusions can be derived from our observations on the risk communication habits of genetic counselors in this study. These observations show that these counselors often presented a one-sided view of risks, in the sense that risks of genetic disorders were often presented only in proportions and were merely framed in negative terms. Since they rarely expressed their own opinions and did not reinforce choices or urge clients to make particular decisions, it seems that the counselors were not directive towards the clients. They guided clients in coming to a decision by discussing possible decisions and outcomes, presenting the probabilities of these outcomes and discovering the importance the client attached to these outcomes. According to the subjective impressions of the observers, the genetic counselors were skillful and committed and took the clients’ level of knowledge and attitudes into account. They were not threatening or directive at all. However, our findings on the modes of communication used by the counselors taking part in this study may not necessarily be valid for all genetic counselors, not even for all Dutch ones. 5.3. Practice implications Since the checklist we developed has adequate interobserver reliability, it is suitable to assess aspects of risk communication in genetic counseling. In addition, the final version of the checklist was found to be easy to use. It was applied in various counseling sessions and was completed by different observers, who had no specific education or background. They prepared themselves by reading the instructions beforehand and filling in some trial checklists.

M. Fransen et al. / Patient Education and Counseling 61 (2006) 126–133

We acknowledge that there are many other factors besides risk communication aspects that play an important role in genetic counseling. Our emphasis, however, was on general risk communication aspects and less on psychological and social aspects, although very important as well. Although we give insight into the aspects of risk communication addressed in the sessions observed, it would also have been interesting to study the underlying reasons for genetic counselors to apply or not apply certain aspects of risk communication. This might also be an interesting topic for future research. This checklist could be seen as a starting point and a helpful instrument to identify relations between aspects of risk communication and the outcomes of genetic counseling, such as the decision to apply genetic tests or preventive measures. Showing which communication skills and risk presentations affect client’s feelings and decisions may help to improve genetic counseling.

References [1] Harper PS. Practical genetic counseling. Oxford: Butterworth Heinemann; 1998. [2] Edwards A, Hood K, Matthews E, Russell D, Russell I, Barker J, Bloor M, Burnard P, Covey J, Pill R, Wilkinson C, Stott N. The effectiveness of one-to-one risk-communication inverventions in health care: a systematic review. Med Decis Making 2000;20:290–7. [3] Marteau TM, Saidi G, Goodburn S, Lawton J, Michie S, Bobrow M. Numbers or words? A randomized controlled trial of presenting screen negative results to pregnant women. Prenat Diagn 2000;20:714–8. [4] Pilnick A, Dingwall R. Research directions in genetic counselling: a review of the literature. Patient Educ Couns 2001;44:95–105. [5] Kahneman D, Tversky A. On the psychology of prediction. Psychol Rev 1973;80:237–51. [6] Grimes DA, Snively GR. Patients’ understanding of medical risks: implications for genetic counseling. Obstet Gynocol 1999;93:910–4. [7] Kessler S, Levine EK. Psychological aspects of genetic counseling. IV. The subjective assessment of probability. Am J Med Genet 1987;28: 361–70. [8] Eiser JR. Communication and interpretation of risk. Brit Med Bull 1998;54:779–90. [9] Bottorff JL, Ratner PA, Johnson JL, Lovato CY, Joab SA. Communicating cancer risk information: the challenges of uncertainty. Patient Educ Couns 1998;33:67–81. [10] Welkenhuysen M, Evers-Kiebooms G, d’Ydewalle G. The language of uncertainty in genetic risk communication: framing and verbal versus numerical information. Patient Educ Couns 2001;43:179–87. [11] Kelly PT. Informational needs for individuals and families with hereditary cancers. Semin Oncol Nurs 1992;8:288–92.

133

[12] Shiloh S, Sagi M. Effect of framing on the perception of genetic recurrence risk. Am J Med Genet 1989;33:130–5. [13] Shiloh S, Saxe L. Perception of risk in genetic counseling. Psychol Health 1989;3:45–66. [14] Watson M, Duvivier V, Wade Walsh M, Ashley S, Davidson J, Papaikonomou M, Murday V, Sacks N, Eeles R. Family history of breast cancer: what do women understand and recall about their genetic risk? J Med Genet 1998;35:731–8. [15] Van Vliet HAAM, Grimes DA, Popkin B, Smith U. Lay persons’ understanding of the risk of Down’s syndrome in genetic counselling. Br J Obstet Gynaecol 2001;108:649–50. [16] Hopwood P, Howell A, Lalloo F, Evans G. Do women understand the odds? Risk perceptions and recall of risk information in women with a family history of breast cancer. Commun Genet 2003;6: 214–23. [17] Williams PR. Health risk communication using comparative risk analyses. J Expo Anal Environ Epidemiol 2004;14:498–515. [18] National Research Council. Improving risk communication. Washington: National Academy Press; 1989. [19] Gurmankin AD, Baron J, Armstrong K. The effect of numerical statements of risk on trust and comfort with hypothetical physician risk communication. Med Decis Making 2004;24:265–71. [20] Kahneman D, Tversky A. Prospect theory: an analysis of decisions under risk. Econometrica 1979;47:263–91. [21] Van Zuuren FJ, Van Schie ECM, Van Baaren NK. Uncertainty in the information provided during genetic counseling. Patient Educ Couns 1997;32:129–39. [22] Pauker SP, Pauker SG. The amniocentesis decision: an explicit guide for parents. Birth Defects Orig Art Ser 1979;15:289–324. [23] Kessler S. Psychological aspects of genetic counseling. XIV. Nondirectiveness and counseling skills. Genet Test 2001;5:187–91. [24] Van Zuuren FJ. The standard of neutrality during genetic counselling: an empirical investigation. Patient Educ Couns 1997;32:69–79. [25] Michie S, Bron F, Bobrow M, Marteau TM. Nondirectiveness in genetic counseling: an empirical study. Am J Med Genet 1997;60: 40–7. [26] Kessler S. Genetic counseling is directive? Look again. Am J Med Genet 1997;61:466–7. [27] De Haes JCJM, Hoos AM, Van Everdingen JJE. Communiceren met patie¨nten/communicating with patients. Maarssen: Elsevier/Bunge; 1999. [28] Silverman J, Kurtz S, Draper J. Skills for communicating with patients. Abingdon: Radcliffe Medical Press; 1998. [29] Watson M, Lloyd S, Davidson J, Meyer L, Eeles R, Ebbs S, Murday V. The impact of genetic counselling on risk perception and mental health in women with a family history of breast cancer. Br J Cancer 1999; 79:868–74. [30] Wouda J, Van der Wiel H, Van Vliet K. Medische communicatie: gespreksvaardigheden voor de arts/gedical communication: conversation skills for the doctor. Maarssen: Elsevier/De Tijdstroom; 1998. [31] Dunn G. Design and analysis of reliability studies: the statistical evaluation of measurement errors. New York: Oxford University Press; 1989.