Ethical violations and discriminatory behavior in the MedPhys Match

Abstract Purpose The purpose of this survey study is to investigate behaviors in conflict with the ethical standards of the Medical Physics Residency (MedPhys) Match (MPM) process as stated in the MPM rules (a) and with the nondiscrimination regulations of the Equal Employment Opportunity Commission (EEOC) (b), in addition to other behaviors that may in other ways erode the fairness of the system. Methods A survey was sent to all applicants and program directors registered for the 2015 and 2016 MPM. Survey questions asked about application, interview, and postinterview experiences, match results, and overall satisfaction with the process. Results Thirteen percent of 2015 respondents and 20% of 2016 respondents were asked by at least one program how highly they planned to rank them or which program they would rank first. Thirty‐seven percent of 2015 and 40% of 2016 program directors indicated that candidates communicated to the program their rank intent, with 22.0% in 2015 and 12.5% in 2016 being told that their program would be ranked first. Twenty‐three percent of 2015 respondents indicated being asked by at least one program during the interview about children or plans to have children; including 19% of males and 33% of females. In 2016, these values were 28% overall, 22% male, and 36% female. Fifty‐seven percent of 2015 respondents who were asked this question indicated being uncomfortable or very uncomfortable answering, including 27.3% of males and 88.9% of females. In 2016, 42.9% of all respondents indicated being uncomfortable or very uncomfortable answering, including 10.0% of males and 80.0% of females. Conclusions In the first two years of the MPM, there were widespread instances of ethical violations and discriminatory questioning during the interview process. Educating both interviewers and candidates on the MPM rules and general EEOC guidelines should decrease these instances and increase the fairness of the residency selection process.


| INTRODUCTION
The National Resident Matching Program (NRMP) was created in 1952 in an effort to eliminate the chaos, pressure, and gamesmanship that existed when medical residency programs and graduating medical students competed to fill residency slots. 1,2 The purpose of a residency match system is to provide a fair system to address the problems resulting when many programs compete for the same top candidates and candidates compete for limited residency positions. Without such a system, the recruitment process can result in a race to make the earliest offer, thus pressuring applicants to commit to an early offer that might not be their first choice for fear of not receiving other better offers. Match systems require candidates and programs to submit numerically ranked lists by a single deadline; an algorithm then matches programs with applicants, assigning both their highest available placement rank. However, ethical violations can erode and diminish the potential of a fair match system. 3,4 Several publications exist in the literature discussing the ethical violations of residency match programs in other medical disciplines, reporting problems with prohibited questioning, postinterview communication, pressures to commit prior to the rank deadline, efforts to game the system, and dishonest communication. 3,5-7 "Gamesmanship" is the term used in the literature to describe manipulating or circumventing the rules and procedures for personal benefit. Some of these publications have a few suggestions for changes, however, the negative aspects of the culture of the residency match system for these disciplines largely did not improve. 8 Medical physics is a rewarding career for individuals interested in applying their physics background in the medical field. A limited number of CAMPEP-accredited residency training programs are creating similar levels of competition as experienced in medical residencies. 9   Responses to the questions regarding opinions were collected using a 5-point Likert scale. Study data were collected and managed using REDCap (Research Electronic Data Capture) hosted at the study institution and supported by an institutional grant. 10 The full survey instruments are included as Supplementary Files.

2.B | Data analysis
Summary statistics were used to describe the response rates, demographics, interview experiences, ranking and matching experiences, as well as opinions regarding the MPM process. All analyses were HENDRICKSON ET AL.

| RESULTS
Applicant surveys were returned by 111 of 402 emailed invitations in the initial survey year, yielding a response rate of 28%, consistent with typical AAPM response rates. One hundred nine surveys were completed and included in this analysis. In the second year survey, 101 of 331 applicants returned the survey, yielding a response rate of 30%. All surveys are included in the analysis.
The demographic distribution of the survey respondents is shown in Table 1. Eighty-four percent of the respondents submitted a rank list, including 81% of the male respondents and 94% of the female respondents. Forty-eight percent of the respondents were matched in the inaugural year, including 45% of the male respondents and 58% of the female respondents. This compares with 39% of all applicants who participated in the MPM and were matched (excludes those who withdrew or did not submit ranks). 11 In the second year, 83% of the respondents submitted a rank list, including 81% of the male respondents and 88% of the female respondents. Seventy percent of the respondents matched with a residency position in 2016, including 63% of the male respondents and 87% of the female respondents. This compares to 51% of all applicants participating in the MPM. 11 Sixty-eight percent of those respondents who did not match in 2016 indicated that they do intend to apply for residency positions again, and 24% officially withdrew from the MPM before the match deadline. Only 17% of all respondents had also applied to the MPM the previous year, and 65% of those had participated in at least one interview in the previous year.
In the second year survey, 41.6% of respondents indicated a final degree of MS, 53.5% indicated PhD, and 2% did not respond. This question was not included in the first year survey.
Program directors' surveys were returned by 42 out of 79 emailed invitations in the inaugural year, with a response rate of 53%. In the second year survey, 47 surveys were completed in response to 77 emailed invitations, with a response rate of 61%. In the inaugural year survey, none of the responding programs indicated that they interviewed only MS candidates. Forty-three percent interviewed PhD candidates only, and 57% interviewed both MS and PhD candidates. In the second year survey, one responding program indicated that they interviewed only MS candidates. Forty-two percent interviewed PhD candidates only, and 56% interviewed both MS and PhD candidates. In the inaugural year survey, program respondents indicated that they largely experienced an increase in applications over the previous year. This situation was reversed in the following year, where most programs reported a decrease in the number of applications (see Fig. 1). In the first and second year surveys only one and then two survey respondents, respectively, indicated that they were asked about uncomfortable or uncomfortable answering the question (see Fig. 4).

3.A | The candidate survey: Interview experience
Twenty-four percent were neutral in their response, while the remaining respondents were comfortable answering this question. In the second year survey, this value increased to 79% of all survey respondents indicating that they were asked at least once where else they were interviewing; 38% were very uncomfortable or uncomfortable answering this question.
Thirteen percent of all respondents in the first year survey and 16% in the second year survey indicated that they were offered incentives (future faculty position, etc.) during their interview.
Twenty-nine percent in the first year and 27% in the second year surveys indicated that they were told by at least one program that they were "ranked to match" or told their rank number prior to the match deadline. If they had knowledge of their rank position, 78% in the first year and 67% in the second year surveys indicated that the information did not affect how they ranked the program. Thirteen percent of all survey respondents in the inaugural year and 20% in the second year surveys indicated that they were asked by at least one program how highly they intended to rank that program or asked which program the interviewee would rank number one. If they were asked about program rank, 64% and then 56% indicated that they were very uncomfortable or uncomfortable answering the question in the 2015 and 2016 surveys, respectively ( Fig. 4).
Furthermore, one applicant respondent to the 2015 survey stated, "My only opposition to the process was regarding schools that sent out emails or called multiple people, essentially inferring they were a top pick, only to find out that they had told several other applicants the same thing. This situation happened to several friends going through the process and one friend and I even received nearly identical emails from an institution that was only offering one position." This respondent continued to complain that this behavior by programs was unfair and misleading to candidates, to the point that applicants altered their rank lists based on the information. This sentiment was echoed in other applicant comments.
Ten percent of all respondents in both years of the survey indicated that they were told by at least one program that they would not match to their program. Seventy-eight percent of the positive respondents in the inaugural year and 75% in the second year surveys said this knowledge did affect how they ranked programs.
In the first year survey, 23% of all survey respondents indicated that they were asked about having children or their plans to have children during their interview. Of these respondents, 57% were very uncomfortable or uncomfortable answering the question. In the second year survey, 28% of all survey respondents were asked this question. Of the 2016 respondents, 42.9% were uncomfortable or very uncomfortable answering this question, including 10% of the males and 80% of the females.
Sixteen percent of survey respondents in 2015 and then 15% in 2016 indicated that they were offered a residency position outside the MedPhys Match Program.

3.B | The candidate survey: Postinterview experience
Fifty percent of all survey respondents in the inaugural year survey indicated that they were contacted via a phone call, email, or letter by a program director, faculty or staff member, or a resident after their interview, which was not in direct response to a contact or question initiated by the candidate. In the second year survey, the positive response rate was 53%. Of all survey respondents, 21% in 2015 and 25% in 2016 said they felt very or moderately pressured by the program to offer assurances (see Fig. 4).
Some candidates initiated communication with the program after their interviews. Figure 5 summarizes their responses, where 93% and 85% in the inaugural and second year surveys reported that they did not say that they would rank a program highly in their thank you note.

3.C | Ethical statements
Applicants were asked about their beliefs in the honesty and motivation of participants in the MPM process. In response to "Applicants often make dishonest or misleading assurances or statements to programs about their level of interest," 37.9% of candidates in the 2015 survey strongly agreed or agreed, while the majority (62.1%) were neutral to strongly disagreed (see Fig. 6). These positive response rates increased in the 2016 survey with 43.1% of candidates indicating that they strongly agree or agree with this statement, while 56.9% were neutral to strongly disagreed.
Candidates were also asked how they felt about the following statement regarding possible results of dishonesty, "Applicants who mislead programs about how strongly they plan to rank them improve their position in the match." Responses were fairly even among agree, neutral, and disagree in both survey years, as shown in As an added value judgment, candidates were asked how they feel about the statement "Applicants may be justified in making dishonest or misleading assurances or statements to programs." Responses again were fairly even among agree, neutral, and disagree in both surveys, as shown in Fig. 6.
Candidates were asked whether they believed their rank position could be improved if senior faculty from their current institution intervened on their behalf. In the first year survey, the majority (54.6%) responded positively to this question (see Fig. 6). Fewer candidate respondents (45.3%) felt the same way in the second year survey.

3.D | Match experience
Respondents to the candidate surveys were asked how satisfied they felt with the MPM experience. As shown in Fig. 7, the responses were influenced by whether the individual was matched to a residency, with matched applicants significantly more satisfied with the process overall than applicants left without a residency position.
In 2015, 69% of all candidate respondents were asked where else they were interviewing. Of those candidates who were asked, 32% indicated that they were very uncomfortable or uncomfortable answering the question. In 2016, 79% of all respondents were asked this question. In 2015, 13% of all candidate respondents indicated that they were asked at least once how highly they were going to rank the program or asked which program that they would rank number one. In 2016, this value was 20%. The majority of candidates are not comfortable answering this question. Information regarding program ranking may be disclosed during the interview or in postinterview communications. A majority of candidates report not feeling pressured at all to offer assurances to programs regarding ranking.
3.E | The program directors' survey: Interview experience Seventy-seven percent of the program directors who responded to the survey indicated that they did give instructions to resident interview participants on the rules, ethics, and guidelines for Match participation. This positive response rate is consistent with the second year value of 75%.
Thirty-one percent indicated that they did initiate some form of communication (phone call, email, or letter) to at least one candidate after their interview that was not in direct response to contact initiated by the candidate. This value is the same in the second year survey. Thirty-six percent in 2015 indicated that they contacted all candidates after the interview, while 15% indicated that they only contacted those candidates that the program was interested in ranking. In 2016, these values were 31% and 25% respectively. and second year surveys, respectively. Of these 37% in 2015, none of the programs agreed with the statement that knowledge of the candidate's intent to rank influenced their own ranking of candidates. In 2016, two program directors responded that the intent to rank information did influence the program's ranking of candidates.
Again of these 37% in 2015, 60% said interviewees indicated that they would rank that program first and 20% were asked by the interviewee how the program would rank them; the values changed to 32% and 0% in the second year survey. Forty-seven percent of the program directors in both surveys indicated that they felt applicants were (always, frequently, or sometimes) dishonest about their intent to rank the program when the ranking statements were made.
Additionally, 13% (two in 2015) and 11% (two in 2016) of the programs failed to match with a candidate that had made a commitment to rank that program number one.
Program directors were asked about the importance or role of postinterview communication by an applicant's mentor or other advocate. In the inaugural year survey, 20% of program directors responded that at least one contact was made on behalf of an applicant. This value decreased to 13% in 2106. Figure 9 shows the relative influence of these communications.  Any expression of interest that may be made during the free discussion between a program and an applicant is subject to change based on further considerations by either party.

3.F | Match experience
The intent is for both applicants and programs to rank their selections in order of their true preference, without regard to how they believe they will be ranked by the other party. That is, to eliminate any "gaming of the system" by pressuring the other party in an effort to ensure a match by making false assurances of rank intent. tion such as advocacy calls is that it has been used as a means by which to subtly or overtly convey rank intentions in order to "prearrange" the match ahead of the deadline. 16 It can take the form of a program promising to rank a candidate highly or "to match" or a candidate claiming that they will rank that program (or every program at which they interview) first in order to convince the other party to do likewise. This behavior is inconsistent with the letter or spirit of residency match rules. Three of four candidates reported sending thank you letters to programs, with 21% (2015) and 34% (2016) only sending to programs that they were particularly interested in ranking. Seven percent (2015) and 15% (2016) admitted that they indicated in their thank you letter their intention to rank that program highly. The use of postinterview communication is widespread and clearly some of this communication is used to convey some ranking intention information. In Fig. 4, 79.1% (2015) and 75.3% (2016) of candidates report not feeling pressured at all by programs to offer assurances regarding ranking, which suggests that the use of thank you letters is relatively benign.

4.D | How does MedPhys Match compare with
other medical residency match systems?
Dermatology is an example of a competitive medical field that participates in a national residency matching program with similar rules for ethical conduct. In a 2009 survey study that included Stanford dermatology applicants, current US dermatology residents, and US dermatology program directors, residents felt pressured to reveal rank intent to the programs. 3 Thirty-one percent of Stanford applicants and 19% of US dermatology residents responded that they felt pressured to reveal their rank-ordered lists before the Match deadline.
Twenty-seven percent of program directors reported that they told applicants they were "ranked to match." Twenty-one percent of Stanford applicants and 17% of US dermatology residents reported that they were told they were "ranked to match" by the program at which they matched. Fifteen percent of Stanford applicants and 26% of US dermatology residents said they changed their rank-order lists based on knowledge they received about their rank order from programs. No program director reported promising an applicant an incentive to rank their program highly; however, 5% of Stanford applicants and 3% of US dermatology residents reported that they were promised an incentive to rank a program highly.
Similarly, 21-25% of MPM applicants felt pressured to offer assurances regarding ranking. Twenty-nine percent (2015) and 27% (2016) reported being told by at least one program that they were "ranked to match" or their rank number prior to the MPM deadline.
Furthermore, 13% (2015) and 20% (2016) reported being asked by at least one program how they intended to rank that or other programs.
A follow-up survey in dermatology was launched the subsequent year, to investigate whether raising awareness of the violations and issues in the initial publication had led to improved behavior. 8 The conclusions were that the behaviors had not changed.

Jena et al. surveyed senior medical students who had applied
to firstand second-year residency positions in the NRMP in a variety of subspecialties. 6 In this survey, the majority (86.4%) of respondents reported being contacted by at least one residency program after the interview. These communications included feedback suggesting that they would "fit well" in the program (76.2%), that they would be "ranked highly" (52.8%), or that they would be "ranked to match" (34.6%). These types of statements are allowed by NRMP. Five percent of the respondents reported that a program had asked where it would be ranked on the applicant's list; this is a violation of NRMP rules. Sixty-three percent of the applicants reported that they informed a single program that they would rank it first; while only 1.1% indicated that they informed more than one program that it would be ranked first. Both of those behaviors are violations of the NRMP rules. Almost onequarter (23.4%) of respondents reported altering their rank list based on postinterview communications, thereby admitting that they are influenced by the postinterview communication. While only 1.2% of the applicants reported failing to match at a program that had told them that they were "ranked to match" and therefore ranked the program first, 18.6% reported not matching with a program despite feeling assured by postinterview communications that they would match there and therefore ranked the program first.
While the nonspecific feedback from programs is allowed by the NRMP rules, it is nonetheless a form of pressure on applicants to alter rankings. In

4.E | Study limitations
This study is limited by the overall response rate, and the potential differences in results with a higher response rate cannot be esti- Other medical specialties tout that their residency positions were filled by their top ranked choices as a measure of their program's value or desirability. 7 This low "number needed to fill" seems a hollow metric for a relatively small field such as medical physics and therefore is unlikely to be adopted by the medical physics community.
Applicants are under pressure to secure a match given limited Rank according to quality of the candidate and not on how highly it is believed that they will rank the programs.

2.
Do not directly or indirectly divulge ranking intent; and do not solicit this information from candidates.

3.
Eliminate postinterview communications to further minimize the risk of divulging or soliciting rank intent information. Questions about the program can be addressed to a neutral and knowledgeable program administrator.

4.
Determine rankings immediately following interviews.
Postinterview communication via a thank you letter or by an applicant's mentor has overall little effect on the program's ranking of the candidate. Therefore adopting a system of determining rankings immediately following interviews, such as recommended by Wu,16 would have little to no effect on final rankings and makes a clear statement to applicants that any postinterview communication is unnecessary, with or without potentially false promises of ranking intent. This also makes a thank you letter truly a simple, courteous expression of thanks and respect, without the unnecessary and unfortunate need to consider an ulterior motive by the applicant.
Program directors can inform interviewees as a group, "Please feel free to contact us about any questions that you have about our program. We will not ask you how you intend to rank us. We will not share with you where you are positioned in our ranking." If it is true that the program will determine its rankings immediately following interviews, then it is pertinent and helpful to tell this to all inter- viewees.
An example template that programs could choose to adapt and adopt in an effort to educate all interview participants on EEOC guidelines and Match rules is included in Fig. 10.

University of Washington Medical Physics Residency Program Interview Code of Conduct
To promote the highest ethical standards during the interviews, ranking, and any matching processes for the University of Washington Medical Physics Residency Program, interviewers and involved others must commit to: • Ensuring that all candidate interviews are conducted in a friendly atmosphere that is safe, respectful, and nonjudgmental. • Respecting an applicant's right to privacy and confidentiality by not asking candidates to disclose their preferences, intentions, or locations of other programs to which they are applying. • Never asking illegal or coercive questions about age, gender, religion, sexual orientation, health and family status (including marriage, children or pregnancy Please sign to indicate that you understand and pledge to follow the letter and intent of these rules.
Sign Name Date F I G . 1 0 . A sample code of conduct agreement statement for all participants in the interviewing process of resident selection.