Factors influencing practice choices of early-career family physicians in Canada: a qualitative interview study

Abstract

Background

Comprehensiveness of primary care has been declining, and much of the blame has been placed on early-career family physicians and their practice choices. To better understand early-career family physicians’ practice choices in Canada, we sought to identify the factors that most influence their decisions about how to practice.

Methods

We conducted a qualitative study using framework analysis. Family physicians in their first 10 years of practice were recruited from three Canadian provinces: British Columbia, Ontario, and Nova Scotia. Interview data were coded inductively and then charted onto a matrix in which each participant’s data were summarized by code.

Results

Of the 63 participants that were interviewed, 24 worked solely in community-based practice, 7 worked solely in focused practice, and 32 worked in both settings. We identified four practice characteristics that were influenced (scope of practice, practice type and model, location of practice, and practice schedule and work volume) and three categories of influential factors (training, professional, and personal).

Conclusions

This study demonstrates the complex set of factors that influence practice choices by early-career physicians, some of which may be modifiable by policymakers (e.g., policies and regulations) while others are less so (e.g., family responsibilities). Participants described individual influences from family considerations to payment models to meeting community needs. These findings have implications for both educators and policymakers who seek to support and expand comprehensive care.

Background

Despite the importance of family physicians, many people in Canada and around the world do not have access to one [1, 2]. Concerns about declining access to comprehensive care [3,4,5,6,7] have generated speculation that the decline is a result of practice choices made by early-career physicians [8,9,10,11]. In Canada, where roughly 15% of Canadians report not having a family physician [2], lack of access to community-based primary care has garnered attention from the media and the public, especially since the start of the COVID-19 pandemic [12,13,14]. There is a need to deepen our understanding of the factors influencing this trend towards reduced access to primary care.

We know little about what influences early-career family physicians’ practice choices and what may be contributing to the move away from longitudinal community-based primary care. A recent study identified a decline in comprehensiveness (range of settings and service areas) across four Canadian provinces, but found that this decrease was not unique to new-to-practice physicians [15]. Evidence is only beginning to emerge about how individual factors may be influencing choices that could affect availability of comprehensive primary care—such as a desire for work-life balance leading to a narrowed scope [4] or a desire for a sustainable career encouraging choice of focused practice [16]. As the next generation of providers, early-career family physicians’ practice choices will play an important role in the availability of comprehensive primary care going forward. An in-depth understanding of the reasoning behind their practice choices can inform policies and system reforms that would support a return to greater comprehensiveness in primary care delivery. We define ‘practice choices’ as decisions about how a physician chooses to practice medicine, for example, their scope or location of practice. In this paper, we answer the question, “what factors have influenced the practice choices of early-career family physicians in Canada?”.

Methods

We conducted a qualitative interview study using framework analysis [17,18,19], to identify factors influencing practice choices of family physicians in their first 10 years of practice. This study is part of a larger, mixed-methods project examining the practice patterns of physicians completing family medicine residencies or within their first 10 years of practice in Canada [20]. We received research ethics board approval from the Simon Fraser University Office of Research Ethics with harmonized approval from partnering universities.

Data collection

We recruited participants in three Canadian provinces—British Columbia, Ontario, and Nova Scotia—who were in their first 10 years of family practice. A description of the Canadian context is provided in Box 1. We chose provinces that vary with respect to geographic context, population characteristics, and payment and practice models. Invitations to participate were sent through family medicine residency program alumni email lists, and via Twitter, Facebook, and posters at research conferences. Individuals interested in the study completed an online screening survey of demographic and practice-related characteristics. Maximum variation sampling of survey respondents was applied by selecting participants that varied by gender, years since residency, relationship status, dependents, practice setting, and specialized training. We aimed for approximately 20 participants from each province to ensure a sample size that allowed for wide-ranging practice experiences [20].

Box 1. Canadian family medicine context

• Funding and delivery of primary care falls under provincial/territorial jurisdiction.

• Primary care models vary across the country and within provinces and territories.

• Most family physicians are private providers and are paid fee-for service.

• Capitation, blended, and salary payment models also exist and vary by province.

• To become a family physician, medical school graduates need to complete a two-year family medicine residency.

• Family physicians play the role of gatekeeper to specialist physicians in the health care system.

Data were collected through semi-structured interviews addressing topics about current practice characteristics, personal and professional influences, training experiences, the policy environment, and considerations for possible future practice changes (see “Appendix” for interview guide). Interviews were conducted in English by phone or videoconference by a single female researcher in each province, audio recorded, and lasted 45–60 min. Interviewers (ER, LJ, MMcK) had graduate-level training in qualitative research and were supervised by PhD researchers (AG, EM, LJG) with extensive experience in qualitative health services and primary care research. Interviewees received an honorarium to compensate for their time. Interview recordings were transcribed by a professional transcriptionist.

Data analysis

Framework analysis is a comparative form of thematic analysis where inductive and deductive themes are organized to identify, describe, and interpret patterns across participants or cases. It requires the creation and application of an analytic framework in a systematic and comprehensive way in a matrix format that allows for a comparative analysis. There are five steps in framework analysis: (1) data familiarization (e.g., reading and rereading transcripts); (2) identifying a thematic framework; (3) indexing study data into the framework; (4) charting to summarize indexed data; and (5) mapping and interpretation of patterns [19]. We selected this method to organize our qualitative data and identify patterns in practice choices across participants. We developed an analytic framework consisting of a subset of codes from the larger project codebook relevant to our research questions (see Box 2 for codebook development). The codes included current practice characteristics, and influential training, professional, and personal factors. All interview transcripts were coded using this analytic framework (i.e., the “indexing” step of framework analysis). Each transcript was then summarized for each component of the analytic framework (i.e., the “charting” step of the framework analysis). To ensure a consistent approach to summarizing and populating individual cells in the framework matrix, the qualitative lead (AG) and three members of the research team developed an analytic dictionary.

Box 2. Codebook development process

• Step 1 Initial codes were developed inductively using transcripts from interviews conducted with family medicine residents (not reported here) from the same study [20].

• Step 2 The codebook was applied independently to one early-career family physician transcript by 11 team members. Team met to discuss refinements to the codebook.

• Step 3 Modified codebook was tested on additional family physician transcripts and discussed over multiple team meetings to inductively develop codes specific to physician data and finalize a codebook.

• Step 4 Research analysts used the final codebook to code interviews from their home province.

• Step 5 Research analysts coded one interview from another province. Double-coded interviews were discussed, with discrepancies resolved by consensus, and revisions to the codebook were used for subsequent solo coding.

The analysis resulted in four matrices containing cross-sectional summaries for all study participants. The first matrix was concerned with four key practice characteristics being influenced (scope of practice, practice type and model, location of practice, practice schedule and work volume). The remaining matrices addressed a specific category of influences (training, professional, personal). We then created interim matrices and tables that enabled us to explore the relationships between the characteristics and influences (i.e., final step of “mapping and interpretation”). We identified themes that summarized how each practice characteristic was influenced, together with illustrative quotes.

Results

We present participant and practice characteristics, followed by the factors that influence practice choices. We use the term ‘community-based practice’ to refer to primary care provided to a general population in a clinic setting. We use ‘focused practice’ to refer to one or more specific clinical areas that make up a part, or all, of a family physician’s practice outside a traditional community-based practice [21].

Participant characteristics

We interviewed 63 early-career family physicians between April and October 2019. Our sample contained slightly more women than men, with around half having dependent children, and most having trained in Canada (Table 1).

Table 1 Demographic characteristics for the full study sample (n = 63)

Practice characteristics

Participants worked in both community-based practices and focused practices, with almost a third working in locums (temporary replacement for another physician) (Table 2).

Table 2 Practice characteristics for the full study sample (n = 63)

Although our interview guide did not include structured questions about practice characteristics, 49 of the 56 physicians who worked in community-based practice described their practice characteristics (Table 3). Most participants worked in a group practice, at one site, and with other health professionals.

Table 3 Practice characteristics described by early-career family physicians in community-based practices (n = 49)

Factors influencing practice choices

Our analysis identified four key practice characteristics: (1) practice type and model, (2) scope of practice, (3) location of practice, and (4) practice schedule and work volume. These were influenced by three categories of influential factors: training, professional, and personal (see Table 4 for a visual overview of the data).

• 1. Practice type and model

  In our analysis, ‘practice type’ refers to community-based practice, focused practice, or locuming; ‘practice model’ refers to solo practice, group practice, or region-specific payment or practice model (e.g., Family Health Teams in Ontario). Illustrative quotes are presented in Table 5.

Table 4 Practice characteristics and the factors that influence them

Table 5 Participant perspectives on influences on practice type and model

Training influences on practice type and model

Participants who reported enjoying their medical school rotations or seeing mentors demonstrate the breadth and depth of family practice said this influenced them to have a community-based practice of their own. Some noted that their medical school had an explicit focus on community-based practice and that this instilled a sense of responsibility to work in a community-based practice. Others learned from medical school mentors with a niche practice, and reflected this by offering focused services themselves (e.g., emergency, sports medicine, mental health services). Exposure to the challenges of the fee-for-service (FFS) payment model during medical school sparked a desire for alternative payment models (e.g., salary or capitation).

Some mentors and influential people in residency highlighted the benefits of community-based practice over working in more focused settings (e.g., emergency) and generally encouraged participants to explore a variety of settings. Numerous participants reported emulating their preceptor’s practice after residency. However, in some cases, preceptors’ large work volume actually discouraged participants from pursuing community-based practice. Participants perceived that urban placements offered fewer opportunities for hands-on training due to the availability of specialists, and this reinforced their interest in rural community-based practice. Those who had rural residency experience reported that they continued to take on rural locums after residency to maintain the skills needed for community-based practice. Rural residencies were also reported to include hospital-based practice (e.g., hospitalist care, inpatient care, obstetrical call), which some enjoyed; others found it stressful and chose to abandon that part of practice after training.

Professional influences on practice type and model

Early career experiences after residency, when participants tried different practices, were reported as strongly influential. Some participants used locuming to find a practice they wanted to join long-term; others built a roster of locum sites through which they enjoyed rotating. Participants reported that wanting to acquire and maintain their skills influenced their decision to locum, to have a community-based practice, or to complement this with work in emergency and obstetrics.

By contrast, policies and regulations influenced people away from community-based practice. Specifically, FFS, low compensation, and restrictions on joining certain practice or payment models were all reported as key factors in the decision to start a focused practice. Participants felt the FFS model was not conducive to providing quality services (e.g., it restricted appointment length times) and caused income instability/unpredictability. This was compounded by the perception of increasing costs relative to fixed fee codes and a resulting dissatisfaction with the level of compensation. The perceived amount of unpaid work in community-based practice further reduced physicians’ desire to open a new practice. Some participants sought alternative payment models that they felt facilitated patient connections and continuity; however, limited opportunities for positions in certain practice models (e.g., Community Health Centres in Ontario), restrictions on entering capitation-based models, and perceived constraints in contract positions caused some participants to turn to locuming or focused practice instead. Several participants were concerned about the future of family medicine (in particular due to the uncertain relationship between family physicians and provincial governments and changing health care policies) and said that this influenced their decision to avoid community-based practice.

The added workload associated with practice management—including administrative tasks, billing, and human resources—was reported by participants as a disincentive to work in a community-based practice. In particular, participants reported difficulties finding locums to cover their practice when needed, identifying fellow physicians to join their practice, and having inadequate business training.

A desire to have more control over practice management, schedule, and practice style influenced decisions on what type of practice to pursue. Others focused on meaningful and valued work such as providing diverse and intellectually stimulating services in acute care settings or as locums. Participants expressed a desire to spend more time with patients during appointments, have the time to build therapeutic relationships, and ensure patients have continuous access to services. These preferences caused some participants to select capitation-based practices or collaborative practices when they were available.

Personal influences on practice type and model

Some participants remarked that financial considerations such as uncertainty in their personal financial situation, debt, or need for additional income made them reluctant to start their own practice; instead, they pursued the most lucrative options (e.g., locums, focused practice). Participants shared the role that family and relationships played; being single and/or not having children gave them opportunities to take on various types of practices or locuming, while those with children reported feeling more limited in their choices. Having children reduced interest in emergency medicine, inpatient medicine, and evening walk-in practice, while increasing desire for salaried models that reduced take-home work. For others, difficulties finding locums for planned maternity leave and perceived judgement by patients about maternity leave were gender-related influences on decisions to focus practice. Personal discomfort with patients’ gender-based expectations about physician behavior, including expectations that women physicians would provide more emotional support than men, discouraged some from community-based practice.

Personality and personal interests were likewise reported as influential. Those who enjoyed travel, an active lifestyle, and were comfortable with change and uncertainty reported interest in locums or work across many settings. By contrast those participants who were not comfortable with uncertainty were drawn to a focused practice.

Participants highlighted that work-life balance was an important consideration in their choice of practice model; however, there was no consensus on how to achieve it. Locuming was perceived as a way to gain schedule flexibility and a manageable work commitment for some, but others feared burnout from locuming and thus chose to start their own practice. A desire for flexibility generally influenced participants away from a community-based practice. Solo practice was perceived as a model associated with high work volume, need for around-the-clock availability, and difficulty finding coverage for time off. This led some participants to instead choose team-based care models. Hospitalist practice was either perceived as allowing for flexibility and better hours, or as requiring overtime and potentially leading to burnout if combined with community-based practice.

• 2. Scope of practice

  ‘Scope of practice’ refers to current areas of clinical practice, populations served, services delivered (in community-based or focused practice), and extra-clinical activities (e.g., research, teaching). Illustrative quotes are presented in Table 6.

Table 6 Participant perspectives on influences on scope of practice

Training influences on scope of practice

Residency experiences were reported as being associated with either an increased scope of practice—when preceptors provided a diverse range of services and increased participants’ competencies in these services—or a decreased scope of practice when training was lacking. Additional training enabled participants to provide certain services (e.g., emergency medicine) or care for populations with specific needs (e.g., people with substance use disorders). Some participants described not being able to obtain additional training, which limited the services they could provide in their practice (e.g., obstetrics, palliative care). However, some found workarounds to gaps in training, such as working with other physicians to gain experience or learning on the job.

Professional influences on scope of practice

Early career experiences were reflected in some participants’ practice choices. For example, previous non-medical degrees or career experiences increased participants’ interests in mental health services, social determinants of health, or specific patient populations (e.g., equity-deserving groups). Some participants reported that their early career experiences in rural settings increased their desire to offer a broader range of services.

Participants reported that their scope of practice was influenced by a desire to engage in meaningful and valued work. Their desire to have a positive impact on their patients’ lives influenced their decision to offer specific services such as sexual health care, palliative care, or obstetrical care. Similarly, they valued meeting patient and population needs, which influenced participants’ scope of practice as they worked to fill service gaps in their community (e.g., medical assistance in dying, opioid agonist treatment) or support particular patient populations (e.g., women facing barriers to care, transgendered people). Teaching was incorporated by some to ‘give back’ to the next generation of physicians.

Reduced confidence in the delivery of specific services or care for specific populations led some to limit their scope of practice. Confidence was influenced by time away from providing certain services (e.g., minor surgical procedures) or a lack of volume (e.g., IUD insertions, obstetrics). When participants felt confident with specific populations or services, they were more inclined to include those in their practice.

Personal influences on scope of practice

Family and relationships were reported as very influential. Participants with children chose to not offer certain types of services, such as hospitalist or obstetrics care, due to unpredictable schedules or night shifts. For some, having a spouse with more lucrative employment gave them the freedom to pursue a scope of practice that was in line with their interests rather than with higher income. For others, however, having a working spouse meant they had more family responsibilities, which restricted their ability to provide full-time community-based care or obstetrics services.

Participants also reported that their gender was sometimes influential, particularly when faced with gender-based expectations about physician behavior. The perception that patients prefer women physicians for women’s health issues influenced some men physicians to not provide those services. Participants reported that taking maternity leave reduced their confidence to provide certain services due to time away from active practice. Lastly, participants spoke about their personality and personal interests; some said that their empathy and patience led them to work with certain populations (e.g., geriatric populations).

• 3. Location of practice

  We define ‘location’ as the geographic location of participants’ practice or as the distinction between rural or urban practice. Illustrative quotes are presented in Table 7.

Table 7 Participant perspectives on influences on location of practice

Training influences on location of practice

Participants reported that having medical school electives or residency placements in rural, small town, and northern practices resulted in a greater desire to work in those locations, in particular when mentors modeled community-based rural practice or encouraged the participant to join their rural practice after graduation. Participants noted that their residency placements either allowed them to get familiar with a location and want to stay, or influenced the choice to not work in the area.

Professional influences on location of practice

Participants reported that their choice of practice location was influenced by policies and regulations that impacted work volume and pay. For example, opportunities to join a payment model that aligned with their preferred style of practice, provided higher income, or required no overhead payments influenced the decision to work in certain locations (e.g., rural locums). Provincial licensing regulations were reported as costly and deterred participants from working in more than one province. Those who did get a second provincial license described low remuneration in their home province as a factor in their decision.

Some participants shared that they moved to an urban setting because of the challenges they faced working in low-resource, rural communities due to high work demands. Similarly, others felt uncomfortable working in rural settings because they lacked confidence to offer a broad scope of services or to be the only physician in a region. Those who were uncomfortable working in emergency departments noted it as a limitation to their ability to locum in smaller communities. Others reported that rural locations provided an opportunity for meaningful and valued work, including gaining and maintaining a broad range of skills, pursuing intellectually stimulating work, having variety in practice, and finding meaning working in under-resourced settings.

Personal influences on location of practice

Participants shared that they worked in or near their childhood community because of family and relationships. This was emphasized by participants with children, who described wanting to be near family and in a community with a strong education system. The desire to live close to aging parents and concerns about future caregiving responsibilities were other considerations in choosing a practice location. Many participants also noted that their spouse influenced their decision about where to work, such as when a partner did not want to live in a rural area. In other instances, participants worked where their spouse had their residency, practice, or other work site. Participants reported that rural communities reduced opportunities for friendships and relationships. However, others found that being familiar with a community and feeling appreciated influenced their decision to stay there. Participants also reported being drawn to communities that aligned with their personality and personal interests (e.g., outdoor recreation).

• 4. Practice schedule and work volume

  We define ‘practice schedule’ as both timetable and time allotted to practice. ‘Work volume’ is defined as the workload associated with participants’ practice. Illustrative quotes are presented in Table 8.

Table 8 Participant perspectives on influences on practice schedule and work volume

Professional influences on practice schedule and work volume

Participants shared that wanting to offer patients better access to care influenced their clinic hours and work schedule (e.g., open in evenings, working more). Remuneration considerations also influenced how participants allocated work time. For example, the ability to earn a higher income as a hospitalist influenced some participants to consider reducing their hours in community-based practice and spending more time working in the hospital. Working in a FFS model was cited by some as a reason to not work evenings and weekends, since there were more no-shows for which they would not be paid. Some participants also shared that wanting to acquire and maintain skills either influenced them to work more days per week (i.e., to practice the skills) or to reduce their work volume until they felt more comfortable providing a specific service.

Personal influences on practice schedule and work volume

Participants noted that work-life balance considerations influenced their schedule and work volume: some adjusted their schedules around personal interests or family while others reduced work volume, built in rest time, opted out of night shifts, or kept their schedules flexible to avoid burnout. Financial considerations led some participants to increase their work volume. Family considerations, such as parenting responsibilities and limited childcare availability influenced some participants to decrease work hours and/or prevented them from working weekends and night shifts.

Discussion

This qualitative study of 63 early-career family physicians identified how training, professional, and personal influences were reported to influence choice of practice type or model, scope, location, and practice schedule and work volume. Multiple factors were reported by each participant, resulting in a complex and nuanced set of often interacting factors that influenced individual practice choices. Training was reported as influential for all these practice characteristics, except schedule and volume. Participants described the influence of rural training experiences and the important role played by mentors and preceptors in medical school and residency. Professional factors influenced all practice characteristics; policies and regulations and a desire for professional satisfaction were commonly mentioned. Personal factors, in particular family responsibilities, work-life balance, and personality or lifestyle were reported to influence all practice characteristics.

While many studies have examined factors that influence medical students’ choice of family medicine [22, 23], few have examined choice of practice characteristics among active family physicians. Those that have often focus on a single practice characteristic, such as scope [4, 24,25,26] or practice type and model [27], and few examine early-career family physicians [24, 27]. These studies confirm our findings of the influential nature of certain individual and regulatory factors on practice choices—including training and mentorship [4, 25,26,27], early-career experiences [25, 26], work-life balance [4, 26, 27], professional satisfaction [4, 24,25,26,27], family [4, 26], policies and regulations [4, 26], and remuneration [24,25,26,27].

Our findings, however, did not always align directly with those in the literature. Likely due to differences between health care systems in Canada and the United States, Russell et al. identified more factors affecting practice choices among American physicians at the institutional level (e.g., choices made by organizational leadership about service provision) [4], while we identified more factors related to provincial policy (e.g., choosing not to work in a CHC model because of provincial government funding decisions). A study by Gosden et al. in England found that new family physicians were sometimes averse to practices in areas of high deprivation [28]. Some of our participants instead highlighted the professional satisfaction they gained from meeting community needs and working with equity-deserving populations (e.g., choosing to meet the needs of vulnerable populations by offering addiction care).

Family physicians face tensions between caring for a patient population with increasing complexity and the growing administrative burdens and costs of community-based practice [29, 30]. For some, including some focused practice services helped create a more sustainable career [16]. Given the complexity of often interacting factors in our study’s findings, governments that want to address workforce concerns and increase the availability of comprehensive primary care must look at modifying multiple factors that create barriers to community-based practice. This includes policies and regulations, but also factors such as physicians’ desire for meaningful and valued work, work-life balance, and meeting the needs of patients. Lastly, factors such as family responsibilities, friendships and relationships, and personal interests should not be ignored.

A strength of this study is its breadth and open-ended nature. The use of qualitative enquiry allowed participants to describe the most relevant factors that shaped how they set up and maintained their practices. Given the broad nature of our research question, our interviews did not allow for an in-depth examination into any particular practice characteristic or influence. While study participants worked across different provincial systems, the results may not directly apply to other health care systems in Canada and internationally.

Conclusion

These rich qualitative data demonstrate the complexity of factors which shape the way early-career physicians’ practice. Policies should take this complexity into account; overly simplified policies that target a single factor such as training are unlikely to reverse the trend away from comprehensive community-based practice. Meaningful system reform will be better supported with a nuanced and comprehensive understanding of the variety of influences informing how family physicians choose to practice.

Appendix: interview guide

Current and future practice description

• 1. How did your career unfold out of residency?

• 2. Tell me briefly about your current practice.

  • How is your practice organized?

  • Who else do you work with?

  • How are you (and your team) compensated?

  • What does your work week look like?

• 3. Tell me about any particular clinical interests that you have as a family physician.

  • Which patient populations are you interested in?

  • How have you incorporated these interests into your practice?

  • Do you see this changing over time?

• 4. In what ways, if any, does your current practice differ from your ideal type of practice?

• 5. How do you see your practice changing over time?

  • If IMG with a return of service: how will your practice change after your return of service is satisfied?

• 6. [If participant mentions “comprehensive” probe: what does comprehensive mean for you?].

Priorities

• 7. When you think about your career, what is most important to you?

• 8. In what ways, if any, did your personal priorities or goals influence your career?

  • How did your personal relationships influence your career?

  • How did parenthood or caregiving influence your career?

  • How did financial considerations influence your career?

  • How did your gender influence your career?

  • How did your other personal characteristics influence your career plans?

• 9. if no exogenous factors emerge in Q 4, 5, 6) What kinds of other influences have you experienced or anticipate that may influence your practice changes over time?

  • i. E.g. Community, professional, regulatory influences

Past experiences

• 10. How did your medical school experience influence your career plans as a family physician?

  (Re-direct away from responses about why family medicine was chosen as a specialty)

  • Positive or negative experiences?

  • Did you have any experiences stand out in primary care during training?

  • Did any key people influence your plans?

• 11. How did your residency influence your plans for family practice?

  • Positive or negative experiences?

  • Did you have any experiences in primary care stand out during training?

  • Did any key people influence your plans?

• 12. Tell me about your CaRMs experience

  • Was family medicine your first choice?

  • Did you have to make trade-offs between specialty and location of residency?

• 13. Tell me about any other life experiences you’ve had that influence your career plans as a family physician?

Why: policy environment and practice opportunities

• 14. At the start of the interview, you told me about how you practice now and what you would like your clinical practice to look like. Do you expect you will be able to achieve this ideal type of practice? Why / why not?

  • Are opportunities available for your preferred type of practice?

  • Are there restrictions or barriers to you having this ideal practice?

  • How will you populate your practice?

  • How do your gender or other personal characteristics impact your ability to achieve this type of practice?

Wrap up

• 15. If you were mentoring a new family medicine resident, what advice would you give them about planning their career in family medicine?

• 16. Anything else that you think is important for me to know?