In a mixed-methods design the investigators will study decision making processes and experiences regarding a diagnostic trajectory before and after the introduction of a patient decision aid for people with memory complaints, their significant others and their GPs. The 'before group' will receive diagnostics as usual from their GPs. The 'after group' will use the patient decision aid. The investigators expect the patient decision aid to increase the level of SDM and to contribute to a timely and personalized diagnostic trajectory. Data will be collected using semi-structured interviews, questionnaires and information retrieved from people with memory complaints' medical records.

Study design This study will use a before-after mixed-methods design. In the before group (BG) GPs will execute care as usual in the decision-making process for a diagnostic trajectory for memory complaints. In the after group (AG), a patient decision aid will be added to the decision-making process provided by the GP. The same GPs will participate in both the BG and the AG, which maximizes the comparability between the BG and the AG. The investigators aim to use the mixed-methods approach to gain a broad and in-depth understanding of the issues important for people with memory complaints, their significant others, and the GPs involved in the decision-making process. This approach is taken to fully unravel the complexity of the decision-making process for a diagnostic trajectory for memory complaints and of the SDM intervention.

Setting The study was developed by Maastricht University Medical Centre (MUMC+) in close collaboration with Radboud University Medical Center Nijmegen (Radboudumc Nijmegen). The participating GPs, people with memory complaints and their significant others will be recruited from general practices located in the south-eastern part of the Netherlands.

Participants Patient involvement Before the start of the study, a client panel of the Alzheimer Centre Limburg (i.e. a panel consisting of people with dementia and their significant others) was asked to reflect on the research design. During the study, people with memory complaints and their significant others will be closely involved in the development and implementation of the SDM-intervention.

Recruitment and study participants First, the researchers will approach GPs working in the south-eastern part of the Netherlands with an information letter about the study. No specific inclusion and exclusion criteria have been formulated for the GPs. GPs who want to participate will be asked to return an application form to the research team. Subsequently, the research team will then provide participating GPs with study material (patient application forms, information letters, and informed consent forms). GPs will be asked to recruit patients with memory complaints during an inclusion period of 6 months in both the before and after group. In the before group, the GPs will inform potential participants about the study after a decision about a diagnostic trajectory (i.e. to wait and see, undergo diagnostic testing by the GP, or referral to a medical centre) has been made. Patients can be included up to a maximum of 3 months after a decision has been made. Significant others of people with memory complaints will be asked to participate as well; their participation is preferred, but not required. The inclusion and exclusion criteria of the study are shown in Table 1. If a potential participant is interested in the study, the GP will send his/her contact details to the researchers. The researchers will contact the potential participant to provide detailed information about the study. Eligible participants will sign informed consent forms before participation. The inclusion/consent procedure in the after group will be finalized based on the developed intervention in a later stage of the study.

Sample size Given the dependence between patients and their GPs in our sample, clustering of patients per GP was taken into account in sample size calculations. Clusters were expected to include at least two people with memory complaints and intraclass correlation to be 0.05 or lower. To achieve power of 0.8 and 2-sided testing at 0.05, 35 GPs are required.

The investigators therefore aim to include between 30 and 40 GPs in our study which seems to be a realistic number based on a previous (comparable) study.

For the qualitative component of the study, it is expected that interviews with 15-30 patients, with or without their significant other will be sufficient to reach data-saturation (i.e. the point at which interviews reveal no new information or insights). Participants will be asked to participate in an interview based on the outcome of the decision-making process (i.e. to wait and see, undergo diagnostic testing at the GP, be referred to a medical center for advanced diagnostic testing). In this way, the investigators can ensure that each 'choice-category' is represented in the selective sample. From those patients not selected for the interviews, only qualitative data from their medical journals (e.g. free text notes) will be analysed. This procedure will be the same in the before and after group. For the interviews with the GPs, it is anticipated that participation of 20-25% of the included GPs will be enough to reach data-saturation.

Data analysis Quantitative analysis Descriptive analyses will be conducted on the sociodemographic characteristics of people with memory complaints, their significant others, and GPs. To assess the decision-making process, descriptive analyses will be conducted on: number of consultations before final decision, final decision (i.e. number of decisions in each 'choice category'), level of SDM, and level of self-efficacy regarding SDM as indicated by GPs. Data will be reported as percentages, numeric counts, and means/standard deviations (SD). Furthermore, multiple regression analysis will be performed to investigate: (1) whether sociodemographic characteristics or characteristics of the decision-making process are related to the outcome of the decision-making process or to the level of SDM and (2) whether the level of SDM is associated with the outcome of the decision-making process. A random effect multilevel analysis will be used to assess changes in the decision-making process before and after implementing the SDM intervention while taking clustering at the GP level into account. Baseline characteristics will be added as confounders in case of observed differences in the before and after group. Integration of these analyses with qualitative measures will help explain these results. All analyses will be conducted using SPSS.

Qualitative analysis Qualitative methods will be used to analyse data regarding experiences in the decision-making process, SDM, and decisional conflict. Interviews will be audio-taped and transcribed verbatim. Two independent researchers will analyse the transcripts using content analysis. The same researchers will agree on codes, combine codes to create categories (axial coding), and combine categories and refine themes (selective coding). Atlas.ti will be used to support the analyses.

Mixed model analysis Integration of quantitative and qualitative research methods will be applied during data collection as well as during analyses. During data collection, attention will be paid to 'building' and 'merging'. Building will be applied by selecting patients for semi-structured interviews based on quantitative data such as age, gender, and the final decision outcome of the decision-making process. GPs will be selected based on their scores on the quantitative measures SDM-Q-Doc and the self-efficacy VAS as well as on the same criteria of age and gender. Merging will be applied through tailoring topic lists for the semi-structured interviews to the personal situation of patients and their significant others, a task which will be accomplished with the help of information from their medical record. During data analyses, 'merging' and 'embedding' will be applied (i.e. qualitative data will be used to explain identified patterns in quantitative outcome measures which will provide insight into the working mechanisms of the SDM intervention).

Ethics and Dissemination Care as usual will most likely not be influenced in the before group of the S-DeciDeD project. Only a small-time investment will be required from GPs and patients: 5 minutes for completing the questionnaires and twenty minutes for participating in the interview. The protocol of the before group has been approved by the Medical Review Ethics Committee (MEC) of the Maastricht University Medical Centre (MUMC+) in the Netherlands, number 2018-0333. Data collection in the before group will start in the beginning of 2021.

The protocol for the after group will be submitted to the MEC after the development of the patient decision aid as judgement should include full information on this novel intervention. The S-DeciDeD project is part of the national Memorabel programme funded by ZonMw.

Furthermore, as Maastricht University aims to become a FAIR (Findable, Accessible, Interoperable and Re-usable) university, the data will also be published on DataverseNL (https://dataverse.nl/). This is a website where researchers can upload their anonymous data with restricted access, which allows other researchers to use the data or replicate the findings after receiving permission from the S-DeciDeD project. Finally, the findings will be published in peer-reviewed international journals and presented at conferences and in one dissertation.

Inclusion Criteria:

• Person has visited the general practitioner because of memory complaints no longer than 3 months ago (noticed by the person, a significant other and/or the general practitioner)
• Person is above the age of 60
• Person is able to complete baseline assessments
• Person must be mentally competent
• Person must give written informed consent prior to participation

Exclusion Criteria:

• Person has communication/language/comprehension/literacy or (severe) hearing problems
• Person has had a severe mental illness in the last 12 months, such as schizophrenia, depression or a bipolar disorder not otherwise specified
• Person has a life-threatening comorbid illness

As Maastricht University aims to become a FAIR (Findable, Accessible, Interoperable and Re-usable) university, the data will be published on DataverseNL (https://dataverse.nl/). This is a website where researchers can upload their anonymous data with restricted access, which allows other researchers to use the data or replicate the findings after receiving permission from the S-DeciDeD project.