In the course of performing an operation, a surgeon will occasionally discover an abnormality that is completely unrelated to the procedure for which patient consent was obtained.1,2 These ‘incidental findings’ (IFs) vary in their nature, severity and consequences. While treating an IF that poses a significant immediate threat to life or limb is justified, indeed required, treating an IF without consulting the patient in any other circumstances is considered contrary to autonomy. Yet in many instances of (non-life threatening) IF, there are potential clinical benefits in dealing with the IF immediately, including avoiding the need for a further operation at a later date. However, performing an additional surgical procedure that is beyond the scope of informed consent could be considered unlawful.6–8,9–10 Despite this, the alternative of leaving the IF to discuss its management with the patient may constitute a lost opportunity to benefit the patient. What should the surgeon do in these circumstances? What would the patient want? What is the right balance between seeking to benefit the patient and patient autonomy and legal requirements?3–5

In a recent survey of surgeons and surgeons-in-training we found a consensus within the surgical community regarding when treatment of IF is, and is not, believed to be appropriate.9,10 Significantly, and unsurprisingly, the surgeons preference to treat or wait varies according to the seriousness of the IF, the urgency of that treatment and the risks associated with performing an additional procedure. However, extending an operation without consent, except in emergency situations, may not comply with current legal and ethical standards. When this happens it can understandably cause significant disquiet for surgeons. Patients may similarly have concerns. Crucially, the viewpoint of patients on the issue of IF has not been systematically investigated. In order to provide some of that perspective, we surveyed samples of both New Zealand healthcare service users and those from the general public, comparing their perspectives regarding the treatment of IFs which may arise during a planned surgery and what they value in this context and what they would expect of their surgeon.

Methods

Study sample

Two samples were included in this study. The general public sample was recruited to identify differences between the opinion of patients and public in this healthcare issue. This was done using advertisements displayed in public places such as community notice boards, newspapers and on social media directed at citizens of Christchurch and Dunedin. The general public group were competent adults and potential healthcare users. They were recruited irrespective of demography and previous surgical history. General public participants completed the questionnaire online using a registered product from Survey Monkey. The patient sample was recruited from pre-admission clinics for elective surgery at Dunedin Hospital and Christchurch Hospital. All patients were approached with information about the study by a medical student or clinical nursing staff. The surgical disciplines included general surgery, urology, plastics and reconstructive, otolaryngology, cardiothoracic and vascular surgery. Those involved either completed the questionnaire while at the clinic or were given a form to return at a later date. The study was approved by the Otago University Ethics Committee (Health).

Questionnaire design

A questionnaire constructed with a mixed quantitative and qualitative methodology was administered to participants. The patients being pre-admitted for surgery were asked to consider the questions as if the situations posed were to arise during their planned surgery, whereas the public participants were asked to consider the questions as if the situations posed were to arise during an imagined surgery. Participants’ preferences were sought regarding the treatment of such a finding without prior discussion versus leaving it to treat at a later time to allow for patient consultation. The survey explored circumstances that might influence the participant’s preferences and included several key factors such as the level of risk to the participant’s health and the clinical consequence(s) of waiting, versus treating the IF immediately. These scenarios were aligned to those previously presented in our study of surgeons’ responses. Following each scenario, participants were asked whether:

I would rather:

Have treatment during the same operation

Wait for further discussion afterwards

The scenarios were presented in order as shown in Table 1. Additionally, participants were asked open-ended questions exploring how they would prefer this topic to be approached pre-operatively by their surgeon. Relevant demographic information was also collected: age, gender, ethnicity, education, occupational status and previous surgical history—including that of any previous incidental findings. For patients awaiting surgery only, information regarding their planned procedure, anxiety and reason(s) for the procedure were also collected.

Table 1: Scenarios presented in the questionnaire. They will be referred to by scenario number in subsequent analyses.

Number

Scenario description

0

General response to IF;

no factors specified

1

IF serious; prompt surgery required

2

IF not serious; surgery may be required

3

IF procedure safe; future surgery avoided

4

IF not serious; future surgery avoided

5

IF serious; surgery required; IF procedure may have long-term impact

6

IF serious; surgery required; IF procedure may have serious long-term consequences 

Statistical analysis

Univariate analyses were initially undertaken to compare various demographic groups and their preferences for treatment within the circumstances outlined using chi-square and t test. Thematic analysis was also undertaken for the qualitative data. This involved a combination of both template analysis and qualitative content analysis. A template of categories reflecting themes was generated inductively from the respondents’ text and the template was modified as themes emerged throughout the iterative categorisation process until a final template containing all themes represented within the dataset was produced. The data were then coded using this template and the categories were analysed.

All available predictor variables were explored by fixed effects stepwise regression. Population partition (either patient or public) and scenario number (0–6) were put into a logistic fixed effects model to estimate a mean for each level from the data. To model the correlations that exist within these data, including repeated-measures on the same individual grouped by choice framework, cross-classified multilevel mixed effects estimation was applied.11–14 The questionnaire included six ‘proceed or wait’ questions and one reference question. Participants were grouped and designated as random effects to allow analysis of the covariance structure introduced by the grouping of the data. Age (as a continuous variable) as well as gender, education, history of surgery, employment and population partition have fixed main effects on these data. Of the subject-level demographics obtained in the questionnaire, all but ethnicity were significant and were included in the model. Introduction of a random slope at the subject-level served to explain over half of the residual variation (55.4%). The partition-level variable anxiety, measured only in the patient sample as they were about to undergo elective surgery, did not significantly improve the fit (not shown). However, fit was improved when the outcome was cross-classified to the participants within the questions in a multilevel mixed model. The reference groups for the regression included male, no advanced education, no past history of surgery, general public and scenario 0 (overall; no factors considered). All analyses were performed using Stata/IC 13.0 for Mac (Stata Corporation, College Station, TX, USA).

Results

Sample characteristics

Data was collected from 368 patients undergoing elective surgery and 331 members of the public from the same regions. The analysis was not restricted to data from questionnaires with complete responses to all questions. A total of 31 respondents (9.4%) in the public sample and four respondents (1.2%) of the patient sample did not answer any demographic questions. Table 2 presents a summary of the characteristics of the participant groups. They were quite distinctive in that the public sample was younger (median: 31 years versus 65 years in the patient sample), had more males (56.8% versus 31.4%; p<0.01), fewer were retired (11.7% versus 48%; p<0.01), and more had university-level education (48% versus 14.7%; p<0.01). The Christchurch and Dunedin patient samples were compared and apart from a small, statistically significant difference in gender distribution with a greater proportion of male respondents in the Dunedin sample (58.4%) than the Christchurch sample (54.5%), there were no statistically significant differences between these two populations (not shown).

Table 2: Sample characteristics. Demographics and p-values for the two-sample t-tests.

Characteristic

 

Public

Patients

P-values

N

%

N

%

Age

Median

31

-

65

-

-

Range

16–82

-

19–94

-

-

Inter-quartile range

22–56

-

55–73

-

-

Unanswered

32

9.6%

10

2.7%

-

Gender

Male

104

31.4%

209

56.8%

p<0.01

Female

193

58.3%

149

40.5%

Unanswered

34

10.3%

10

2.7%

Ethnicity

NZ European

239

67.9%

329

87.0%

p<0.01

Māori

10

2.8%

18

4.8%

Pacific Is.

6

1.7%

3

0.8%

Asian

16

4.5%

2

0.5%

Other

50

14.2%

17

4.5%

Unanswered

31

8.8%

9

2.4%

Employment status

Paid employment (full or part time)

172

46.7%

151

39.3%

p<0.01

Retired

43

11.7%

171

44.5%

Providing care for home and family

16

4.3%

8

2.1%

Not working due to ill-health

12

3.3%

29

7.6%

Not working for other reasons

4

1.1%

9

2.3%

Studying

90

24.5%

6

1.6%

Unanswered

31

8.4%

10

2.6%

Highest education

None

4

1.2%

24

6.5%

p<0.01

Secondary school

85

25.7%

190

51.6%

University

159

48.0%

54

14.7%

Other tertiary

39

11.8%

56

15.2%

Vocational training

5

1.5%

18

4.9%

Other

8

2.4%

10

2.7%

Unanswered

31

9.4%

15

4.3%

History of surgery

Yes

189

57.1%

328

89.1%

p<0.01

No

113

34.1%

30

8.2%

Unanswered

29

8.8%

10

2.7%

 

Total

331

-

368

-

  

In addition to the baseline characteristics that may impact stated preferences to proceed or wait, there were 517 respondents with a past history of surgery (74%) of which 59 (11.4%, 15 patient, 44 public) indicated a past history of IFs (Table 2). Patients’ level of anxiety regarding their upcoming procedure and public respondents’ level of anxiety for a theoretical procedure were also measured on a 5-point Likert scale (Not anxious, a little, moderately, reasonably, extremely). Patients report less anxiety (median=”a little anxious”) than general public (median=”moderately anxious”).

Main findings

Initially when unprompted by the scenarios exploring different contexts, 95% of the respondents preferred to have an IF treated without additional consultation. At this scenario-naïve point, respondents were also asked in an open-ended format to state the factors that would encourage or discourage their preference to proceed (Figure 1), it emerged that most stated patient and public preferences were similar to the factors subsequently tested in the scenarios: ie, severity of the IF, the risk associated with the IF procedure, long-term consequences of the IF procedure and the likelihood that the IF would eventually require surgery. Several other factors, including the surgeon’s competence, cost-effectiveness and the inclusion of an IF clause in pre-operative discussions were referred to by both the patient and public respondents, but were not explored in the scenarios presented in the questionnaire.

Figure 1: Responses to open-ended questions “what would encourage or discourage participants’ choice to proceed during a planned procedure with treatment of a concurrent IF?”. 

c 

The relative prevalence of each theme showing the encouraging factors (dark grey, diverging left and discouraging factors (light grey, diverging right) for the patient sample upper bar (solid fill) and the public sample in the lower bar (hash fill).  

When respondents were subsequently asked to state their preference in the context of different scenarios, there was an overall preference to proceed in 75.1% (81% for patients, and 86.7% for the general public) in all scenarios. However, the proportion that chose to proceed was scenario-dependent for both the general public and patient samples (Figure 2A). When comparing the proportions for each scenario, a statistically significant greater proportion of patients preferred to proceed with immediate treatment than those in the public sample, except in scenario 2 (IF serious; prompt surgery required) where they were similar. Secondary analyses to examine an age bias, an age matched subset of 316 subjects showed a similar effect except patients and public were more alike when the scenarios associated seriousness with the IF and its outcome (Figure 2B). In these scenarios, if the seriousness of the IF required prompt surgery, the predominant expectation was for the surgeon to proceed (93.1–95.9%). However in the scenario where there was less urgency, but greater risk of longer-term consequences, there was a greater preference to wait and not to proceed (55.4% patients; 80.4% public), especially if these consequences might be serious. In contrast, if the IF was not serious or the surgery was safe and further surgery could be avoided, a large majority (95.7% of the patients, 83.7% of public) preferred the surgeon to proceed.

Figure 2: A comparison of the proportion of general public and patient respondents having a preference for the surgeon to proceed with treatment of an IF without prior patient consultation.  

c 

c 

Results are shown for each of the scenarios in the questionnaire. The p-values are for t-tests of the two-tailed null-hypothesis. Panel A describes all respondents (N=699). Panel B is an identical analysis on a subsample of age-matched public and patient samples (N=316; 158 respondents in each sample partition). 

Several scenario-independent factors were identified in addition to age and public/patient partition that may impact the preference to proceed or wait. These include gender, level of education, ethnicity, employment status, level of anxiety in patients and prior history of surgery. However; anxiety and ethnicity did not have a significant impact on the choice to proceed or wait. A past experience of an IF, as in 59 of 518 (11.4%) respondents with a past history of surgery, also did not appear to be associated with preference.

To better characterise the impact of these demographics on the outcome (proceed or wait), quantitative data were modelled as the choice to proceed with IF treatment introducing several fixed and random covariates. The final multilevel mixed effects model with the subject-level random effect was highly significant for these data (p<0.0001) compared to ordinary logistic regression. Neither the addition of interaction terms nor extending the random effects improved the model. Sub-analyses were undertaken to identify the fixed effects of factors, such as IF severity, consequence of treatment, necessity of surgical treatment of IF and convenience. These factors varied across the seven scenarios presented in the questionnaire but the individual contributions of each factor were not resolvable (not shown). Figure 3 shows the fixed main effects of these data in the model and shows that younger members of the public are the most protective of their autonomy while older patients with a history of previous surgery are the least, especially in scenarios where the IF is serious. Figure 3 also shows the relative level of scenario-dependence of public and patient preferences for differing combinations of factors tested in each scenario (severity of IF, procedural risk, convenience, consequences of treatment, consequences of non-treatment.)

Figure 3: Forest plot showing the factors that influence the public and patient respondents’ preference to proceed or wait.  

c 

The plot shows the fixed main effects (age, gender, education, sample partition (patient/public), history of surgery and scenario) of the model expressed as the odds ratio and 95% confidence intervals. The random effects are not shown but were statistically significant (p=0.0001). 

The majority of scenario-primed respondents (in contrast to the earlier scenario-naïve viewpoint) preferred to have the possibility of IFs discussed preoperatively, while a minority but not insignificant number 80/699 (11.4%) who preferred to be informed after surgery. Following on from this, the respondents’ opinion of the appropriate IF-related content in pre-operative discussions was then explored in an open-ended format.

The main theme that emerged from this scenario-primed thematic analysis was the desire to be informed beforehand of common IFs and the associated procedures to treat them as well as a cursory discussion of their relative risks. Some respondents indicated a desire to be informed with greater details about risks, impact on quality of life, and outcome in terms of recovery and impact on the original procedure while others desired less information. Secondary themes included the surgeon’s competence, cost-effectiveness, inclusion of an IF clause in pre-operative discussions, and the preference to be told nothing about IFs. When respondents offered their reason for wanting to be told nothing or as little as possible, they reported that the additional information on IFs would lead to undue anxiety, information overload, or confusion. Overall, the public were more responsive in the open questions and had broader expectations of a preoperative process regarding IF.

This survey of patients and public, follows a previous survey of surgeons’ preferences to proceed or wait when an IF is found given varying levels of procedural risk and IF severity. The same reticence was identified in surgeon respondents in circumstances where the IF was severe or there were long-term disability implications of IF treatment.15 While the surgeons’ questionnaire and the public/patient questionnaire were designed to be delivered to different groups, the results of both surveys illustrate similar trends across surgeon and public/patient respondents. The tandem questionnaires allow direct comparison of preferences in the cases of differing severity of IFs and risk of IF procedures. The results show that in addition to agreement between public and patient healthcare users, there was also strong agreement between the surgeons and healthcare users with no statistically significant difference between them (Figure 4).

Figure 4: The frequency that patient and public respondents chose to proceed with IF treatment compared to the frequency that surgeon respondents chose to proceed in scenarios that were similarly presented across the tandem surveys and which varied procedural risk (moderate or high risk) while holding IF severity constant (high severity). 

c 

Discussion

This is the first study of New Zealand healthcare service users (patients and the general public) stated preferences regarding IFs. The findings illustrate that a high number of respondents generally prefer surgeons to immediately treat an IF when there has been no explicit prior consent for the unforeseen event, especially when the immediate risks of the IF are high and surgery is required. There is also a similar preference when the risk is low and the need for another surgery can be avoided. However, when the unwanted consequences of concurrent IF treatment could be severe or long-term, the contrary preference was expressed: not to treat immediately. Both patient and public samples expressed preferences to proceed based on the severity of the IF and risks of proceeding with treatment. These findings are important because it suggests that most surveyed New Zealand healthcare users prefer concurrent IF treatment when there is a favourable risk-benefit ratio. It is also consistent with the preference for some pre-operative discussion of IF.

This study also sought to obtain an understanding of the extent to which participants value their autonomy and what they would expect of their surgeon in the case of an IF. When it came to IFs, patients were willing to forgo a degree of autonomy when the perceived risk-benefit ratio was low, but communicated a preference to have the possibility of IF included in the surgical consent. Even so, considerable trust in the surgeon was implied in regard to recognising an IF and quantifying the relevant risks, as well as having some cognisance of what might be the patient’s preference had a discussion been possible. A small proportion of respondents expressed the desire for reassurance that the surgeon was well trained and competent to deal with the IF.

Furthermore, in qualitative analyses, the themes that emerged from the surveyed New Zealand healthcare users were consistent with themes that emerged from the survey of New Zealand surgeons.15 The surgeons stated similar values and very closely mirrored the patient and public wishes in their clinical management of IFs. As such, patient, public and surgeon respondents in New Zealand have similar values and rationales in approaching the management of IFs.

Unfortunately, no studies were found to date regarding healthcare users’ preferences of IF management in other locations or populations to offer further guidance. Much of the published literature on the topic of IFs and surgical consent (SC) focuses on prevalence of specific IFs,16–18 complications of treating Ifs,19 delays in treatment, and pre-empting IFs through more thorough pre-surgical clinical assessments.20,21 There are also numbers of case reports of non-consensual IF treatment leading to legal action.22–26 The implications of IF in other areas of healthcare including radiology, oncology and genetic counselling have been explored, but have limited relevance to the immediacy of the surgical context and the unconscious patient.26–28 Importantly, this study suggests that in general, patient, public and surgeon views are similar regarding the factors relevant to IF and IF decision-making. Since the findings in this study are among the first reporting public and patient preferences regarding this ethically, legally and clinically important issue, the results suggest several ways forward. Firstly, these findings provide an indication of what factors influence a patient’s preferences in the decision-making and disclosure processes. This study shows that preferences are influenced by the immediate clinical surgical context of IF severity and procedural risk. Clearly, if there is immediate risk to life or limb, the preference is for immediate treatment. If there is a high likelihood of an adverse outcome associated with the IF or long-term adverse implications from the treatment, patients prefer to participate directly with the surgeon to make the decision; a surgeon who proceeds in such circumstances without consent risks complaint and/or sanctions. However, it also suggests that if the immediate treatment of IF is low risk and another operation is avoided, the majority of patients are more likely to accept concurrent treatment. Generally, in these circumstances, people are happy for their surgeon to take an action by which they would avoid the burden of returning for further surgery. This was evident in the qualitative analysis with reference to avoidance of the added risks and inconvenience of further surgery as well as added costs or waiting times. This is a reflection also of the knowledge of the health system constraints, limitations on access and long waits that exist for some surgical procedures.

Demographic characteristics of the respondents may also influence an individual’s preference to proceed or wait. For example, increased age is associated with a greater preference for a surgeon to proceed, while tertiary education is associated with a reduced preference to proceed. For the surgeon, these findings offer some guidance when confronted with an IF as to the likelihood of a specific patient’s preference. However, there is a self-evident note of caution that a patient’s demographic does not mean that they will definitely support the expected preference for further immediate IF surgical intervention or not and therefore the surgeon should not rely solely on this. It is unwise to make assumptions based on some demographic characteristic of the patient. While most times the patient will agree with a surgeon’s actions, they also may not.

Secondly, the study confirms that the majority of patients and public expect IF to be included in surgical consent. This should be done carefully to avoid overload in the surgical consent process and unnecessary anxiety. This was also the preferred approach of New Zealand surgeons surveyed. In the survey of surgeons, when such consent was obtained, the surgeon’s decisions regarding IF were significantly more informed and lead to more interventions for patient benefit.15 However, the level of detail regarding IF expected by the respondents varied widely from the very detailed and extensive to the much more cursory. Numerous studies of surgical consent have reported that being informed of all risks, both common and rare, were among the top priorities for patients and ranked much higher than being informed of legal rights.29 Some studies suggest that patients want explicit information about their operation and post-operation recovery as well as direct involvement in all decision-making along the way30,31 regardless of their ability to comprehend the complexities and despite reported anxiety. Care should be taken not to make pre-operative discussions too lengthy, dense and difficult for the patient to digest.32–34 Surgeons are often reported to prefer including less information to make the discussions more comprehensible, but a reasonable middle ground should be achievable.33–35 Some limitations to this study have been acknowledged and taken into account above. Given the recruitment strategy,there are several characteristic differences between the patient and general public samples. For the general public, there was a larger social media component while the patient sample recruitment was based on contact in the pre-admission clinics. The age disparity between the two populations offers the most likely explanation for corresponding differences in employment status (11.7% retired in the public sample; 44.5% in the patient sample) and education level (48.0% university educated in the public sample and 14.7% in the patient sample). The gender distribution was significantly different with more male respondents from the patient sample (56.8%) than from the general public sample (31.4%). However, the sub-analyses show these are accounted for. This sampling difference has usefully pointed to some differences between the opinion of patients and sampled public in issues of healthcare. In particular, a patient awaiting surgery appears to be more open to additional surgery for an IF, more accepting of greater risk, and more trusting of their surgeon.

This study provides evidence that there is general consistency of opinion among patients and public with that of the surgical community regarding concurrent IF treatments. It confirms that treatment of IF need not just be based on the threat to life and that many times patients are willing to forgo autonomy for the sake of other benefits. The safest way to ensure that the values of the surgeon match with those of the patient with regard to IFs is to include a discussion about this possibility prior to surgery.