Using PROMs during routine medical consultations: The perspectives of people with Parkinson’s disease and their health professionals

Abstract Background The use of patient‐reported outcomes measures (PROMs), such as quality of life or symptoms like pain or fatigue, is increasingly embraced within patient‐centred care and shared decision making. Objectives To investigate: (a) how patients and health professionals think about using PROMs during routine medical consultations; (b) for which purpose(s), patients and health professionals want to use PROMs during those consultations; and (c) how patients interpret PROMs information presented in various formats. People with Parkinson's disease and their health professionals served as case example. Methods We performed semi‐structured interviews with patients (N = 13) and professionals (N = 7 neurologists; N = 7 physiotherapists). We also used a survey in which patients (N = 115) were shown six figures displaying different information types. Presentation formats of this information varied (line/bar graphs). Interpretation by patients, perceived usefulness of information, attitude towards using information during routine medical consultations and (hypothetical) decisions were assessed. Findings Patients and professionals were generally positive about using PROMs during medical consultations. Professionals stressed the opportunity to monitor changes in individual PROMs over time. Patients were primarily positive about aggregated PROMs to make treatment decisions. This information was also most often interpreted correctly, especially when presented through a line graph (90.1% correct). Professionals thought patients should take the initiative in discussing PROMs, whereas patients thought professionals should do so. Conclusion/Discussion When used in routine medical consultations, PROMs seem to have potential to support shared decision making and facilitate patient‐professional communication. However, training seems needed for both patients and professionals to facilitate actual discussion and proper interpretation.


| INTRODUC TI ON
Where patient-reported outcome measures (PROMs) have traditionally been used to evaluate the effectiveness of medical interventions from the perspective of patients, 1 they are nowadays thought to have a more comprehensive role in health-care quality improvement. [2][3][4] PROMs are standardized questionnaires, completed by patients, measuring how they experience their health (ie symptoms, functional status). Well-known PROMs are quality of life questionnaires, such as the SF-36 and EQ-5D. Integrating PROMs within routine clinical practice might increase the motivation of patients to complete PROMs (which currently proves challenging), and such use has also been theorized to improve management and monitoring of patients' problems. 5,6 Apart from that, using PROMs in routine medical consultations could improve doctor-patient communication 7 and informed patient decisions 4,8 and as such improve quality of care. At an aggregated level, they can also be used to evaluate and compare treatment options and provider performance. 3,4,9 Using PROMs during routine medical consultations by health professionals and patients would suit current ideas on person-centred medicine and shared decision making. 4,10. Although shared decision making has become the standard in health care to enable patients in making informed decisions, [11][12][13] the use of PROMs in this context has remained understudied. For patients who must make choices between different treatment options and health-care providers, it is, however, not only important to have insight into clinical outcomes (eg survival rates), but also into how different treatment options may impact on their own health and quality of life, as viewed by patients themselves. Previously, patients have been positive about using information derived from PROMs, especially because it may prompt discussions about actual health issues with professionals. 14 However, patients do not find all PROMs equally relevant or even bothersome or unnecessarily confronting, and such perceptions may well hinder the adoption of PROMs into daily clinical practice. Several previous studies have therefore involved patients into developing PROMs. 10,15 Also, studies have shown that meaningful and comprehensible presentation of PROMs data is important to help patients with interpreting the data correctly and with using the data in important medical decisions, such as choosing a particular type of treatment or health-care provider. [16][17][18][19][20] Although a majority of patients seems to correctly interpret PROMs information (displayed through various formats), a substantial portion (10%-30% across studies) continues to experience trouble with interpretation, resulting in an inaccurate understanding of the information. [20][21][22][23] Health professionals are somewhat positive about using PROMs. 14 However, they often do not refer to PROMs during their routine consultations with patients or do this only on an ad hoc basis. [24][25][26] Barriers they experience include a lack of knowledge on how to use PROMs and lack of an adequate infrastructure for data collection and use. 6,27 The notion that PROMs such as a quality of life rating scales may be useful at an aggregated level as a benchmark to compare intuitions, but are perhaps less suitable (and were originally never designed to) guide decision making for individual patients, feed concerns or even resistance among health professionals towards using PROMs in consultations, thus creating important barriers for integration into daily clinical practice. 9,28 So, although there seems to be potential for using PROMs during

| ME THODS
This study used a cross-sectional design, with Parkinson's disease as case example. For various reasons, Parkinson's disease can be regarded as exemplary for many other chronic progressive illnesses.
First, Parkinson's disease is a chronic condition characterized by a wide range of both motor and non-motor symptoms. Second, optimal management requires a multi-disciplinary approach with involvement of a wide range of professional disciplines, involving all traditional echelons of care. The disease course is gradually progressive, with a lengthy time span of typically multiple decades. In the first part of our study, we conducted a series of semi-structured interviews with people diagnosed with Parkinson's disease and with neurologists and physiotherapists providing care to people with Parkinson's disease. Based on these interviews, we identified key themes from the perspectives of both patients and professionals.
This qualitative design allowed us to gain a better in-depth understanding of the views of patients and to compare it in an efficient way to the professionals' perspective. In the second part, we used these themes to build a survey. The survey was self-administered by people with Parkinson's disease to assess their interpretation and perceived usefulness of different types of PROMs information. We chose to conduct this survey as a confirmatory study, allowing us to collect some quantitative evidence for the qualitative findings, especially concerning the interpretation of information in different formats. A patient representative from the Dutch Parkinson's Disease Association actively collaborated with us throughout the project. She commented on the interview guide, helped us recruiting patients for the interviews and was present during some of the interviews with patients. We also discussed the initial analyses of the interviews with the patient representative.
Professionals were recruited through ParkinsonNet, a Dutch national network of professionals specialized in treatment and care of people with Parkinson's disease. 29 Professionals within this network were invited to participate through email, followed by a reminder one week later to non-responders. The researcher (MV) contacted professionals who responded. One physiotherapist was recruited through the network of one of the researchers in ParkinsonNet.

| Procedures
Interviews with patients were conducted by the research assistant (SV). The interviewer first again explained the study as well as anonymity and confidentiality, subsequently asked permission for audio-recording, and then obtained informed consent. Next, the interviewer conducted the interview. The interviews had a total duration from about 45-60 minutes. After the interviews, patients completed a short questionnaire about their socio-demographic characteristics, health literacy 30 and numeracy. 31  Out of the 500 inhabitants that are not in the choir, 300 are men.
What is the probability that a randomly drawn man is a member of the choir? Please indicate the probability in a percentage." The correct answer is 25% and this was considered adequate numeracy; all other responses were considered to indicate inadequate numeracy.
Semi-structured interviews with health professionals were conducted at different locations in several Dutch hospitals and physiotherapy practices. Five professionals were interviewed by telephone for logistic/pragmatic reasons. A researcher (MV) conducted the interviews with professionals. Professionals were provided with brief information about the study and were asked permission for audio-recording. Next, the actual interviews were conducted (see measures). All interviews with professionals lasted between about 45 and 60 minutes.

| Interview topics
For all interviews, an interview protocol was used to assess partici- Professionals were asked to reflect on these examples, for example to what extent they thought they would be useful and usable, and how they would explain the information to patients.
Both interview protocols were developed in an iterative way by the research team; OD and SV developed the draft version of the interview protocol for patients, and MV and MF developed the draft version of the interview protocol for professionals, addressing the main research questions. We did not use a specific conceptual or theoretical model, but rather used an explorative approach using broad concepts thought to be related to information needs and use (patients) and attitudes and perceived barriers/facilitators (professionals). These protocols were then exchanged between the researchers and refined further. Next, neurologists involved in Parkinson care and the patient representative involved in the study were asked to give feedback on the draft versions, both with respect to content and the way we formulated interview questions. After this feedback, the interview protocols were finalized.

| Data analyses
All interviews were transcribed literally. We employed qualitative thematic analyses with an inductive character. All transcripts were read and re-read (14 interviews with professionals by MV and 13 interviews with patients by SV (patients)). Passages were selected, coded and related to our main topics in interview protocols. This resulted in two sets of codes (one for the professional interviews and one for the patient interviews), which were grouped into overarching themes. A subset of interviews (three patient interviews and six professional interviews) was coded by a second researcher (OD and MF) to ensure reliability of analysis. Disagreements between codes and/or identification of themes were resolved in separate consensus meetings between SV and OD and between MV and MF. We did not perform a member check nor a review by an independent analyst.
However, we did involve the patient representative in the identifica-  Patients who were interested in participating after viewing the website announcements could directly access the survey through a web link. In total, 125 patients completed the survey. We excluded 12 cases because of poor data quality (ie no or hardly any questions were answered) resulting in a data set of 113 cases.

| Measures
Based on the interviews with patients, we developed a survey to assess patients' interpretation and use of PROMs in different presentation formats. We provided patients with six figures showing fictitious PROMs information (Figures 1-6). We used the PROM "level of discomfort" as an example instead of "Quality of life" in the titles of the figures, to ensure alignment with the higher = worse directionality of the scores. All six figures were accompanied by explanations of the y-as (0-100 score, 0 = no discomfort and 100 = great discomfort) and x-as (0-10 years, 0 = time of diagnosis and 10 = 10 years after diagnosis).
After showing each figure, we posed several questions to assess the way patients interpreted and used the presented information.

| Interviews with health professionals
We identified nine themes across the three interview topics (Table 2), with no distinct differences between neurologists and physiotherapists.

| Perceived pros and cons of discussing PROMs and types of PROMs considered useful
Health professionals had a strong preference for individual PROMs Although professionals generally agreed that aggregated PROMs scores for provider options could be useful for internal quality improvement, only few said they would use this information in consultations with their patients (Theme 3). This was largely due to a lack of trust in data quality as well as in patients' capacity to judge the data at its true value, but also to their perception that patients do not want to use this information. Some professionals also perceived individual PROMs scores with comparative data of similar patients useful (Theme 4). According to these professionals, this information could contribute to improve disease knowledge and acceptance, and Level of discomfort

Number of years after diagnosis
Level of discomfort Mr. Jansen

| Factors influencing (non) use of PROMS data in the medical encounter
Professionals perceived the availability of an online portal, in which patients could fill out PROMs and receive the results, as the most important factor enabling the use of PROMs data during routine consultation (Theme 7). In addition to such a "within-encounter" tool, professionals also saw possibilities for patients to review PROMs data in a "pre-encounter" tool. Professionals felt the need for training in PROMs in general and more specifically in how to use PROMs during consultations with patients (Theme 8). Finally, appearing more indirectly from the data was the need to have a positive attitude towards using PROMs in consultations (Theme 9). Several professionals appeared to be sceptical towards the quality and reliability of PROMs as health-care quality measurements. In this respect, several professionals also emphasized the need for role models and pioneers working in the field of Parkinson's disease.

| Interview study patients
We identified 12 themes across the four interview topics. These are listed in Table 3, illustrated with patients' quotes.

| Comprehension of different types of PROMs data
Overall, patients adequately comprehended the PROMs information

| Preferred ways to receive PROMs data
There was a strong preference to receive the information as soon as possible (ie immediately after diagnosis; Theme 10) and from their health professional (Theme 11). Neurologists and nurses were the preferred health professionals to receive information from (Theme 12 TA B L E 3 Qualitative themes and subthemes derived from the interviews with patients (N = 13)

Interview topic Themes and quotes
Current use of PROMs Theme 1: Medical encounter is concentrated on medication regimen, and not on treatment decisions in general "Well, how far should such a neurologist go? It's only 15 min. (..) The neurologist is more like a medication person." (female, 43 y, highly educate, disease duration of 6 y) "I walk and do exercises with that hand, and then she comes in and advises a pill or another pill." (Male, 70 y, highly educated, disease duration of 6 y) Theme 2: Patients do fill out PROMs questionnaires, but they do not remember the data being discussed by professionals Patient: "Indeed, I filled out such a questionnaire. (..) On my PC at home. Yes, He < neurologist>sent it to me." Interviewer: "And did your doctor discuss the information afterwards with you?" Patient: "No." (Female, 52 y, medium educated, disease duration of 10 y) "No I don't believe that we discussed it." (Female, 77 y, highly educated, disease duration of 3 y) "No, not with the neurologist (..). I did so with my oncologist but that was 7 y ago." (Female, 66 y, medium educated, disease duration of 3 y) Theme 3: Mixed preferences as to discussing PROMs data with professionals "Because it offers an entrance to talk. I really think that I would start on that myself, at that moment." (Female, 52 y, medium educated, disease duration of 10 y) "About quality and life and yes.. Well, … we're far from that. That's why I say: I am busy with the things I'm still able to do." (Male, 71 y old, medium educated, disease duration of 3 y)

Comprehension of PROMs data
Theme 4: Gist of PROMs data is adequately comprehended, but not the exact details <this means> "that hospital two has a better score than hospital one." <patient responding to comparative hospital information, only focusing on comparison but not on dimension on which hospitals are compared> (Male, 71 y old, medium educated, disease duration of 3 y) <this means> "well, that I was quite ill.. that's what I think. Or is that incorrect?"<patient responding to individual PROMs scores on 'pain', focusing on the general concept of 'illness' and not on the Quality of Life dimension 'pain'> (Female, 66 y, medium educated, disease duration of 3 y) "I would choose the red line." Interviewer: "the combination of medication and occupational therapy?" Patient: "Yes" <patient responding to choice-information about treatment options, only focusing on one option and on general concept of Quality of Life (Female, 72 y, highly educated, disease duration of 6 y). that they did not feel equipped to start this conversation themselves. Most patients also preferred the possibility of taking the PROMs information home so they could reconsider the information after having discussed it with their neurologist or nurse. Parkinson's disease who participated in the survey (N = 113). A substantial number of participants did not complete the full survey and dropout seemed to be a function of the survey length; response rates for the initial questions of the survey were relatively high (only 7.1% missing for the first presentation format), whereas response rates increased towards the end of the survey (25.7% missing for sixth presentation format). Table 4 displays descriptive findings relating to our main variables, only for those participants who filled out the corresponding questions.

| Correct interpretation of PROMs information
Overall, most patients interpreted the PROMs information reason- Individual PROMs scores over time were interpreted more often correctly when presented in a bar graph (Figure 2; 87.8% correct) compared to a line graph (Figure 1; 74.3% correct). Bar graphs were also preferred by patients (57.2%) compared to a line graph (42.3%). providers was perceived as least useful (only 43% found this information certainly useful; Figure 6).

| Attitude towards use of PROMs information in routine clinical consultations
Patients were most positive towards using the aggregated PROMs scores comparing two treatment options in medical encounters with their professionals (94.5% of participants said they wanted to discuss Figure 4 and 92.1% to discuss Figure 5). The main reasons why patients wanted to discuss these scores were that it gave them in-

| Hypothetical decisions
We

| Strength and limitations
A limitation is that we used convenience samples of patients and professionals, which may limit the generalizability of the findings.
Qualitative interviews were performed with 13 patients, and the question remains whether their perspective also reflects the more general perspective of Dutch people with Parkinson's disease. We were however able to include a diverse group of patients, both with respect to disease duration and educational level, which decreases the chance of systematic selection bias. Among the health professionals, there were no distinct differences in the views of neurologists vs physiotherapists, which also supports the notion that that there was no great selection bias. Of concern is that a quarter of our participants did not complete the full survey, which is also a more general concern when using web-based questionnaires 33 and which may impact the external validity of our findings.   Whether shared decision making based on PROMs will really occur will, however, likely depend on the role that professionals take on in consultations. Our interviewed patients relied on the initiative of professionals to actually discuss PROMs data, which aligns with previous study findings in various clinical settings. [35][36][37][38] Patients are interested in information about treatment options and in shared decision making, 39,40 but also tend to leave the initiative in the consultation largely to health professionals when they are not explicitly invited to collaborate. 41,42 Because professionals in this study indicated that patients should take the initiative in discussing PROMs, it seems somewhat unlikely that PROMs information will actually be used for shared decision making, unless professionals will be explicitly instructed and trained in taking the initiative. In addition, initiatives to stimulate an active role of patients may be needed. 13 Correct interpretation of the PROMs information in our survey study varied from 74% to 90% across presentation formats.

| Discussion of findings
This finding is similar to previous findings in this field. For example, McNair et al 22 found accuracy rates ranging from 85% to 98% across six formats in a sample of 192 patients. Others found similar percentages. 23,43,44 Interpretation in our study seemed to be most accurate when using a bar chart compared to a line graph, although it should be said that one particular figure using a line graph (ie aggregated PROMs scores over time with results of two treatment options) was also accurately interpreted. Therefore, it seems to depend partly on the type of information presented, which format was associated with better comprehension. The finding that especially bar charts support user understanding seems to be somewhat in con-

| CON CLUS ION
Overall, the findings of our study confirm that both patients and professionals are positive towards the use of PROMs information in routine medical consultations. However, we found conflicting expectations and preferences of patients and professional as to who should initiate the conversation about PROMs information and which types of PROMs should be prioritized in those conversations. Overall, patients interpreted the PROMs information reasonably well, especially when presented through bar or line graphs.

ACK N OWLED G EM ENTS
This project was funded by the Dutch National Health Care Institute and The Netherlands Federation of University Medical Centres (Nederlandse Federatie van Universitair Medische Centra) (NFU)).
We would like to thank Loeki Feringa and Tanja Berk, for their valuable contribution as patient representatives of the Dutch Parkinson's Disease Association throughout the project. We also acknowledge the assistance of several researchers and from ParkinsonNet, particularly Tessa Geltink (intern) and Sandra Boots, in conducting the survey study.

E TH I C A L CO N S I D ER ATI O N S
The study protocol was approved by the Medical Ethics Review