Introduction
Second medical opinions (SOs) can assist patients in making treatment decisions and increase the understanding of their diagnosis [
1,
2]. In addition, they reduce the number of surgeries [
3‐
7]. It has been found that they can lead to changes in diagnosis or treatment, even though there is some variation with regard to the rate of disagreements between first opinions and SOs [
8,
9]. SO programs were originally introduced in the USA in the 1970s [
4]. Currently, SOs are implemented in many countries [
3,
10‐
13]. A survey of the German population found that 24% had already obtained one or more SOs [
10].
In Germany, the right to obtain an independent SO free of charge was initially implemented by the SO Directive in 2019. This applies to individuals with statutory health insurance, making up a vast majority in the German population. The first medical indications included in the SO Directive were hysterectomy and tonsillectomy or tonsillotomy. The list of indications included in the SO Directive was later extended to shoulder arthroscopy, amputation of the lower extremities, knee endoprosthesis, and spine surgery. The list of indications will probably be extended further [
14]. The SO Directives sets the framework within which SOs must be offered to individuals with statutory health insurance. The SO Directive is set up by a committee that includes representatives of service providers (physicians, dentists, hospitals) as well as statutory health insurers. According to the SO Directive, SOs for the included indications need to be provided verbally, thus allowing for telemedical SOs. Statutory health insurers may also provide SO programs for indications that are not part of the SO Directive. Many statutory health insurers provide (telemedical) SO programs [
15]. These SO programs are often provided based on documents only and often exclude direct and personal contact between patients and physicians [
16]. Even though high patient satisfaction was found for customers of an online portal providing SOs based on documents only [
17], a survey of the general population in Germany found that 90% preferred personally delivered SOs [
10]. This indicates that there are discrepancies between the provision of SOs by health insurers and the patients’ needs.
We aimed to survey the general population in Germany on their views of SO programs, and to identify diverging perspectives for people living in areas with diverse degrees of urbanization. We asked participants to compare SOs based on documents only with personally delivered SOs, and to name potential situations in which they would seek SOs based on documents only. This was done to characterize patient collectives and situations in which either type of SO provision would be appropriate, and to assess advantages and disadvantages. We further aimed to explore how SOs should be provided to best fulfil the patients’ needs. This included, but was not restricted to, qualification criteria for the physician providing the SO and acceptable waiting and travel times to obtain an SO.
Methods
The survey was conducted as part of the ‘ZWEIT’ project funded by the ‘Innovation Fund’ of the Federal Joint Committee. More details on ZWEIT can be found in the study protocol [
18]. Our survey was approved by the ethics committee of Witten/Herdecke University (case number: 162/2019). We followed the guidelines on conducting and reporting survey research [
19]. The questionnaire was developed in multistage meetings between researchers. Its structure is shown in Table
1. Parts of the questionnaire were adapted from our previous survey of customers of a telemedical SO provider. This relates mainly to questions on sociodemographic characteristics and a few questions on patients’ experience with SOs [
20]. Health literacy was assessed using the 16-item European Health Literacy Survey [
21]. The answer category ‘do not know’ was added to the 4 answer options (very easy, fairly easy, fairly difficult, and very difficult). Furthermore, our questionnaire contained 4 items adapted (at least partially) from a previous survey on SOs [
10]. Our questionnaire consisted of a mixture of open-ended and close-ended questions. For some of the close-ended questions, we provided dichotomous answers, whereas for others, we provided a list of possible answers. Some questions allowed for multiple answers, and others required a single answer. We piloted the questionnaire with a sample of 18 persons (heterogeneous in terms of sociodemographic characteristics).
Table 1
Structure of the questionnaire
1 | 2 | health-related items |
2 | 5 | local health care situation |
3 | 7 | patients’ needs concerning SOs |
4 | 8 | patients’ experiences with SOs |
5 | 8 | design of the SO procedure |
6 | 3 | experiences with and knowledge of SO programs by health insurers |
7 | 3 | experiences with and knowledge of the offer of SO providers |
8 | 11 | sociodemographic characteristics |
In the first step, all local registration offices in Berlin and Brandenburg were listed and sorted by settlement pattern (cities, towns and suburbs, rural areas). For each settlement pattern, 10 local registration offices were randomly selected. Because there are only five local registration offices for the settlement pattern cities in Berlin and Brandenburg, we selected all of them. We asked the local registration offices to provide postal addresses of randomly selected adult residents (n = 333 each for the 10 local registration offices of towns and suburbs as well as of rural areas and n = 666 for local registration offices of cities). This process led to a 1:1:1 disproportionate stratified sampling based on settlement patterns (cities, towns and suburbs, rural areas). This sampling method was chosen to detect potential differences in the attitude toward SOs based on the settlement pattern. Finally, through the 25 local registration offices, we identified a stratified sample of 9990 adults (≥18 years) living in the federal states of Berlin and Brandenburg, Germany.
We initially contacted eligible persons by post in April 2020 and sent a reminder in May 2020. Because responders were asked to return the completed questionnaire, consent for participation was implicitly provided. Postage was paid by Witten/Herdecke University, and participants could win one of 125 Amazon vouchers amounting to 50€. The declaration of consent required to participate in the lottery was returned in a separate envelope. The questionnaire and consent form were separated immediately upon arrival. Therefore, no assignment to individuals was possible. One researcher extracted data from each questionnaire received into an Excel spreadsheet. Another researcher checked a random sample of 10% for accuracy. For open-ended questions, categorization by one author was verified by another author. We resolved discrepancies by discussion and, if necessary, by consulting another author. To analyze the extracted data, we used Microsoft Excel. We compared settlement patterns between all persons invited for participation and the participants. We categorized postal codes according to the ‘degree of urbanisation’ of Eurostat [
22] and regional statistics by the Federal Statistical Office [
23]. The Comparative Analysis of Social Mobility in Industrial Nations (CASMIN) categorization [
24] was used to classify the participants’ education level. CASMIN contains three main categories (primary, secondary and tertiary education) and several subcategories. Primary education (inadequately completed general education) was not applicable to our sample. Therefore, we only considered the categories tertiary education (academic degree, independent of school education and type of degree), higher secondary education (no or any vocational education with at least intermediate school education) and lower secondary education (no or any vocational education with at least general school education).
We planned to perform a subgroup analysis comparing participants living in rural areas and participants living in cities. For continuous data, we planned to use a t-test to test for significant differences. In case of dichotomous data, we planned to use a chi-squared test.
Discussion
SOs are generally viewed positively and utilized by the German general population. The willingness to seek an SO is high. Most participants preferred SOs delivered within direct patient-physician contact, either in a practice or in hospital. SO programs offered by health insurers or dedicated SO providers seem to play only a minor role, as very few respondents are aware of them.
A medical record analysis and a telephone survey with a representative sample from Israel found that 14.9 and 17.2% of participants, respectively, had obtained an SO during a 1.5-year period [
25]. In our sample, we found a higher proportion of people obtaining one or more SOs (47%). However, because we did not limit our question to a specific time period, the results are not comparable. A systematic review investigating oncological SOs found high heterogeneity in the rate of those who had already obtained an SO [
8]. The high heterogeneity might reflect differences in focus and research methods of the included studies making a direct comparison with our results difficult. On a national level, we found a high degree of acceptance of SOs, particularly in comparison to an earlier survey of the German general population [
10].
The preferred way of SO delivery was via direct and personal patient-physician contact. This is in accordance with the SO Directive, which mandates that the SO is provided verbally. Furthermore, this is comparable to another survey [
10]. However, a subgroup of 15% stated that there are situations in which they would obtain an SO based on documents only instead of a personally delivered SO.
We were unable to find considerable deviations in sociodemographic characteristics between our participants and two previous surveys of patients who had obtained SOs based on documents only [
20,
26]. This would have allowed us to further characterize the subgroup of patients for whom a telemedical SO might be more attractive. We were also unable to characterize patient subgroups in a related, published survey, because patient characteristics were not reported. The sample consisted of US patients who obtained an SO based on documents only, which was subsequently discussed with the patients [
27]. Given the high satisfaction of the customers of the online portal [
20], the SO Directive might exclude the needs of a patient subgroup by limiting SOs to those delivered verbally. However, we are currently unable to specify to which patients this would apply.
We found that participants were willing to wait about a month and travel for an hour to obtain an SO. This high willingness was documented in the context of the hypothetical knee pain described in the questionnaire. It remains unclear to what extent participants would have different views in the case of other complaints, or even in the presence of real complaints.
Our results confirm findings by Geraedts et al. on the importance of seeking an SO for different diagnoses. Participants rated obtaining an SO following a cancer diagnosis to be more important compared to diseases of bones, joints, and muscles. Although we found a wide range of indications for which participants had obtained an SO in the past, most of those were obtained for orthopedic indications (30%) and significantly fewer for oncologic indications (8%). At first glance, this contradicts the finding that oncological SOs are rated to be more important than orthopedic SOs. The high proportion of orthopedic SOs may be related to the fact that the quality of medical indications for orthopedic surgery has been discussed controversially in the media [
28,
29]. This may have raised awareness and created a certain degree of public mistrust. Perhaps as a result, orthopedic SOs are common [
10,
16,
17]. Furthermore, the numbers may reflect the prevalence of the indications. We found that SOs for conservative and surgical treatment recommendations were equally distributed. This shows that the patients’ need for SOs goes beyond surgical treatment recommendations and includes nonsurgical treatments. This is particularly interesting because the SO Directive only refers to elective surgical treatment recommendations and explicitly excludes oncological indications. The exclusion seems to contradict patients’ needs. Systematic review data show that internationally, SOs are often obtained for indications not currently covered by the German SO Directive. Many SOs are provided for oncological indications and elective surgery, but also for general medical concerns [
8,
9]. Overall, indications rated as important by patients have not been included in the SO Directive to date, while some indications included in the SO Directive are less important from the patient perspective. The list of indications has been extended gradually in the past and will probably be expanded further in the future. The patients’ perspective should be considered when selecting further indications to include in future versions of the SO Directive .
In the context of hypothetical knee pain, participants rated experience with the recommended diagnosis/treatment, knowledge of the current state of research, and independence to be (very) important qualifying factors for the physician providing an SO by 93, 86, and 72%, respectively. These criteria are part of the SO Directive. However, 80% of participants rated the access a physician providing the SO has to a network of experts as (very) important. This is only part of the SO Directive for a subset of indications but it is part of some SO programs provided by health insurers [
16]. Around 70% of participants would consider it (very) important that the second opinion provider can perform further treatment. This is contrary not only to many SO programs offered by health insurers [
16] and to the SO Directive but also to the fact that independence was rated as (very) important by many participants. Perhaps the participants’ understanding of independence was different to ours. We referred to financial aspects such as (conscious or unconscious) incentives to recommend a particular treatment which could be lower when further treatment is not permitted.
According to the SO Directive, the physician providing the first opinion must inform patients about their right to obtain an SO. Given that 93% of participants want to be informed about the possibility of obtaining an SO by a physician (followed by 56% who want to be informed by the health insurer), this is mostly in accordance with the patients’ needs. Furthermore, most participants want to be informed about the possibility of obtaining an SO in a personal and direct way (89%). This can best be achieved during the consultation with the physician providing the indication. The SO Directive requires that physicians who provide the first opinion inform patients about information available online on treatment options and about a list of physicians offering SOs. This is in accordance with patients’ wishes for information on treatment options and for lists of potential SO providers. Currently, it is unknown how many physicians fulfil the obligations set out in the SO Directive.
Finally, health insurers could certainly better promote their offer of SO programs and the SO Directive. This would be indicated particularly in view of the low awareness level of SO programs in the population. The low awareness might also explain why 43% of the participants considered seeking an SO via their health insurer, but only 4% of participants who obtained an SO in the past had done so.
Strengths and limitations
The response rate to our survey was low, despite various strategies to increase response, such as sending a reminder and offering participation in a lottery. The number of items included in our questionnaire is a possible explanation for this (47 items). We used only one validated instrument in our questionnaire. As in a previous survey [
20], we found a high proportion of people for whom no overall score for health literacy could be calculated (17%). This is probably due to the provision of a ‘do not know’ category. Because we used data provided by local registration offices, we were unable to systematically compare participants with the basic population (except for the residence classification). Therefore, we cannot exclude selection bias. Furthermore, the impact of the COVID-19 pandemic on the results of our survey remains unclear. For example, the pandemic may have influenced the comparison between different ways of SO delivery and the willingness to wait or travel for an SO.
This is a comprehensive survey on SOs, and its results help to better understand patient needs, especially in the context of other parts from our extensive project ZWEIT.
Implications for research and practice
It is important to tailor SO programs to the patients’ needs. SO programs should be evaluated with the goal of defining specific groups and circumstances for each type of SO program. Some patients’ needs identified in this project appear contradictory. Future research should examine to what extent these needs can be reconciled. Because many participants are interested in obtaining an SO but only few are aware that SO programs are offered by health insurers or SO providers, the level of awareness of these programs should be increased.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.