What is risk?

A risk is the likelihood (or probability) that harm will occur from a particular hazard.4 Patients may be concerned with broader qualitative attributes, such as the origin of the risk (natural or technological), whether a risk is imposed or temporary, and the power that they have to control the risk.5 Many times in reproductive healthcare there are hazards associated with a particular decision. For example, there are risks of pelvic inflammatory disease (PID) with an IUD, a probability of up to 9 in 100 of pregnancy with typical use of combined oral contraception,6 or a consequence that well over half of progestogen implant users experience a change in menstrual bleeding pattern.7 Patients will generally take risks if they perceive that there is an advantage or benefit. Normally, the benefits should outweigh the risks by a significant margin. There is no such thing as a zero risk8 ,9 and patients should be made aware that there is no medicine or drug without a possible risk or side effect, including medicines used for contraception.

Risk perception is subjective.10 Patients tend to act on the basis of their perception of the risk rather than the on ‘actual’ risk presented to them.11 Their estimates of risk may differ from those of the healthcare professional who counsels them.2 ,12 ,13 Other factors may play a role. For example, uptake of genetic testing in prenatal care was strongly predicted by women's general attitude to genetic testing, independent of risk perception.14 Risks may be perceived as far less tolerable if the treatment or intervention is voluntary, the technology is unfamiliar or has no benefit for the patient.

Importance of good risk communication

Good communication helps to build trusting relationships between patients and professionals, leads to greater satisfaction on both sides,4 with less fear of treatments,15 helps people to take more responsibility for their own health, reduces medical errors and mishaps and even malpractice claims.10 ,16 Healthcare professionals tend to consider risk as the mathematical probability of something happening within a whole population group (e.g. the chance of an ectopic pregnancy in women using an IUD), which is supported by statistical information.1 This way of assessing risk is important, but it is objective, impersonal and deals with populations rather than individuals. Conversely, patients use an everyday concept of risk that is bound up in the individual's concerns, anxieties and fears about the present and the future (e.g. “What is my risk of having an ectopic pregnancy while using an IUD?”). Understanding the uncertainty about a treatment is thought to be a critical element of an informed decision and should be communicated effectively to patients.17 ,18 At the same time, the healthcare professional should make an assessment of the patient's mental capacity to understand the information given, considering this in the light of the Mental Capacity Act 2005. Healthcare professionals have traditionally communicated risk poorly,19 but developments such as patient decision aids can improve risk communication.20

Legal aspects of risk pertaining to patient consent

The legal position on whether or not a doctor or other healthcare professional is negligent in failing to mention a risk to a surgical patient was decided in the UK in the case of Sidaway.21 The more recent Montgomery v Lanarkshire Health Board judgment22 highlights the importance of an individual approach to advising a patient about a risk. When discussing the benefits and risks of various treatment options with patients, the new ruling requires doctors to consider whether “a reasonable person in the patient's position would be likely to attach significance to the risk, or the doctor is or should reasonably be aware that the particular patient would be likely to attach significance to it”.22 Rather than a radical change in medical practice, the judgement has been considered to bring the law in line with current ethical guidance for UK doctors, as stated in the General Medical Council guidance on consent.23 It is therefore advisable to inform the patient of any ‘material’ or ‘significant’ risks of the proposed treatment, any alternatives to it, and the risks incurred by doing nothing. A Court of Appeal judgment stated that it will normally be the responsibility of the doctor to inform a patient of “a significant risk which would affect the judgment of a reasonable patient” [Pearce v United Bristol Healthcare NHS Trust (1998) 48 BMLR 118 (CA)].24 It is appropriate to warn of a relatively rare risk for an elective procedure (e.g. pregnancy following laparoscopic sterilisation) if this might have a huge impact on the patient if it occurs.

How to communicate risks

Risks in SRH are personal and rich in uncertainty and there are no clear best practices for presenting information about uncertainty.17 Risk communication can be improved, however, by employing certain strategies as described below.

1. Personalised risk communication tailors information to an individual's risk status. For example, current evidence suggests a risk of weight gain with Depo-Provera^® injection use compared to the oral contraception or no hormonal contraceptive method, and more so if the user is already overweight. This risk can be personalised, for example, by telling a woman with a body mass index of 32 that the average increase in weight over 18 months after starting Depo-Provera is 9.4 kg compared to 0.2 kg if she took oral contraception and 3.1 kg if she did not use hormonal contraception.25 Personalising risk communication enhances informed decision-making.26 A Cochrane Systematic Review found that for individuals undergoing screening procedures, interventions with a ‘personalised risk communication element’ enhance informed decision-making.26 Overall 45.2% (592/1309) of participants who received personalised risk information made informed choices, compared to 20.2% (229/1135) of participants who received generic risk information [odds ratio for random effects 3.65, 95% confidence interval 2.13–6.23]. Personalised risk communication does not significantly increase the uptake of such screening tests.

2. Using a range of presentation styles and risk communication tools is the most effective way to give information about risks to patients.27 A randomised controlled trial of women's understanding of the effectiveness of contraceptive methods found that tables with categories such as ‘more’ or ‘less effective’ communicated relative effectiveness better than numeric tables. However, women grossly overestimated the risk of pregnancy unless they were shown tables with numbers, so a combination of both methods was best.28 Graphically presenting risk information in addition helped general practitioners to build a relationship with women, and allowed them to convey absolute and relative risks quickly without having to explain these concepts.29 A specifically designed tool, the Paling palette (Risk Communication Institute, http://www.riskcomm.com/paling_palettes.htm) can display most medical risks with a probability of higher than 1 in 1000. The chart shows 1000 people, with the number experiencing a particular effect coloured in (Figure 1). An alternative system uses icon arrays (http://www.iconarray.com), a matrix of usually 100–1000 icons to represent an at-risk population, simultaneously displaying both the number of expected events and the number of expected non-events.

3. Framing health information as negative or positive can affect patient understanding of risk.30 ,31 As an example, comparing the chances of combined oral contraception failing in a good user (<1 in 100, negative framing) versus the chances of successful contraception (<99 in 100, positive framing). People receiving information framed as a positive or a gain have more confidence in a treatment than those receiving information presented as a negative or loss. Since framing of information can have such a profound effect on patient perception of risk, framing information in a variety of ways is recommended to avoid misunderstanding.32 For example, a patient may be told that there is a 1 in 10 chance that she will have a pregnancy in the first year of using a diaphragm, so this is a 9 in 10 chance that she will not conceive in the first year (typical risk of contraceptive failure with a diaphragm is 16%).

4. The duration of the risk should also be discussed with the patient at the same time as discussing the amount of risk.33 For example, the risk of PID after IUD insertion among 22 908 IUD insertions and during 51 399 woman-years of follow-up was 1.6 cases per 1000 woman-years of use on average, but after adjustment for confounding factors, PID risk was more than six times higher during the 20 days after insertion than during later times (unadjusted rates, 9.7 vs 1.4 per 1000 woman-years, respectively); the risk was low and constant for up to 8 years of follow-up.34

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

A graphic display to compare the effectiveness of different types of emergency contraception (developed using data from references 56, 57). CI, confidence interval; EC, emergency contraception; IUD, intrauterine device; OR, odds ratio; UPSI, unprotected sexual intercourse.

Aiding decision making

Decision aids such as pamphlets or videos prepare patients to participate in decisions by providing information about treatment or screening options and their associated outcomes. When compared with usual care, decision aids performed better in terms of greater knowledge, more realistic expectations, lower decisional conflict, and an increased proportion of people participating in decision-making.20 More detailed decision aids such as booklets or computer-based programmes are better than simpler ones and significantly improve knowledge, create more realistic expectations and greater agreement between values and choice. A decision aid for choosing pregnancy termination methods, for example, improved knowledge and reduced risk-perception scores about both methods compared to usual counselling methods.35 In general, structured, tailored and/or interactive communication tools are more effective at increasing a patient's understanding of evidence.36 For contraceptive methods, a recent systematic review found that a range of educational interventions can increase knowledge.37 Further research is needed, however, to determine what aspects of educational interventions are most effective, whether an educator is needed, and how education may impact behaviours.

Presenting numerical risk information for patients

The format in which written and verbal information on probability is presented affects patient understanding of risk. Patients have a more accurate perception of risk if the information is presented as numbers rather than words.36 Presenting natural frequencies, which is a joint frequency of two events, such as the number of patients with a disease and who have a positive test result,38 may also be beneficial.

A change in risk is better understood if absolute risk reduction, or a relative risk reduction with the baseline risk format for the disease, is used.36 For example, there is up to a 30% (30 in 100) chance of pregnancy when a woman has unprotected sex during her fertile period. Typical failure of condoms to prevent pregnancy is 10–15% (10–15 per 100 women). By using the IUD for emergency contraception this risk of pregnancy is reduced to <1% (<1 in 100 women will get pregnant). The absolute pregnancy risk reduction is 29% with IUD use, greater than with condom use. Further worked examples for calculating risk reduction are available.39 Descriptive terms such as ‘very low risk’ or ‘high risk’ may be used, but they should be qualified by giving a figure, such as a frequency, since they can reflect the speaker's perspective and not the patient's view.40

European Union (EU) guidelines recommend verbal descriptors of risk.9 From a regulatory perspective for patient information leaflets, the emphasis is to express risk as a statistical probability, for example, a verbal descriptor such as ‘very rare’ corresponds to ‘up to 0.01%’ (<1/10 000). However, research into what individuals understand by terms such as ‘very rare’ or ‘common’ suggests that the current EU guidelines on verbal descriptors are not correctly matched with statistical probabilities.41 ,42 The public equate the verbal descriptors (very rare, common, and so on) to risks that are substantially higher than those defined in regulatory documents. The current EU guidelines are provided in Table 1.9

Percentages are an abstract way of portraying risk (e.g. 5%), whereas the actual number of people who could be affected is more vivid to some patients (e.g. 5 in every 100 people). Consequently, when quantifying risks it is recommended to use frequencies rather than percentages to portray risk,41 and to help compare between risks, use a consistent denominator.42 Example 1 below uses a consistent denominator to compare venous thromboembolism (VTE) risks. Another example using a constant denominator considers contraceptive efficacy, stating that in 1 year of use, 99/100 women will not get pregnant with an IUD, whereas only 85/100 women will not get pregnant with typical use of a male condom.

Using relative risks alone to compare different options can be misleading41 as was demonstrated in the 1995 ‘pill scare’. Epidemiologists reported that use of third-generation oral contraceptives doubled the risk of VTE compared with use of second-generation oral contraceptives.43 The initial reports did not mention that the absolute risks were low: 1/3000 compared with 1/6000, that the background risk in non-users was 1/20 000, or that the increased risk of death was around 1/1 000 000. The risk was not put into context and media and public panic ensued.44 Evidence suggests that presenting the risk of thrombosis complications in terms of incidence rather than relative risk may improve communication of side effects during counselling for combined hormonal contraception initiation.45

Over the last decade there has, however, been little improvement in the reporting of absolute risks.46 For many observational studies and meta-analyses, absolute risks cannot be reported easily or meaningfully. Wherever possible absolute risks should be stated, since using absolute data allows the consumer to more easily compare different risks.41

Examples of communicating risks in SRH are provided below.

Example 1: A woman asks you what her risk is of having VTE while using the combined pill

The risk of VTE (deep venous thrombosis and pulmonary embolism) in women who are not using the combined contraceptive pill is 2–5/10 000 woman-years.47 This means that in 10 000 women not using the pill over a 1-year period up to five may have VTE (hence it is ‘rare’). This doubles to 5–12/10 000 woman-years, which is still rare to uncommon, in women who are using combined hormonal contraception.48 ,49 In pregnancy and the immediate postpartum, the risk of VTE is much higher: 29/10 000 woman-years and over 10 times higher postpartum– 300–400/10 000 woman-years. VTE is therefore rare in women who use the combined pill. Women are at a far greater risk of VTE postpartum or during pregnancy than while taking the combined pill as shown in Table 2.

Example 2: A woman is interested in having intrauterine contraception but has heard IUDs are associated with pelvic infection. She asks what her risk is of having a pelvic infection if she chooses an IUD.

The risk of a pelvic infection while using intrauterine contraception is generally low,50 but is increased where there is a sexually transmitted infection (STI) and in the first 3 weeks after IUD insertion. The most comprehensive review to date of insertions and IUD use suggests a risk of up to 16/10 000 woman-years,34 hence pelvic infection after IUD insertion or with IUD use is uncommon. Also, pelvic infection has not been found to be higher in women using IUDs compared to women who are using other methods of contraception such as progestogen-only injectables51 or no contraception.52 By preventing STIs, the risk of pelvic infection with IUD is further reduced. The patient should be offered an STI screen before her IUD insertion and should practise safer sex to reduce her risk of pelvic infection with the IUD.53

Example 3: You offer a woman with postcoital bleeding a chlamydia test but she declines. Her last STI screen, 6 months ago, was negative and she has not changed partner since. Her partner has never been screened for STIs before, but she believes he has no infection as her own results have been negative. What will your advice be?

The woman's own negative test results may suggest, but cannot guarantee, that her partner does not have an infection. Most currently used STI test kits have a sensitivity and specificity of over 97%, so the test will miss 3 in 100 (or 1 in 30) infections. The efficiency of a test depends on the test's ability to give a positive result when there is an infection (sensitivity) and its ability to give a negative result when there is no infection (specificity). Contaminants and hormonal factors may affect a test result; hence no test is 100% accurate.54 It is therefore good practice to repeat some tests when indicated, even if the patient may have done the test previously. This patient has postcoital bleeding and a chlamydia test is clinically indicated. You would therefore recommend a repeat chlamydia test, and an STI screen for her partner.