Peters, Antoinette S. PhD; Schnaidt, Kathleen N.; Zivin, Kara PhD; Rifas-Shiman, Sheryl L. MPH; Katz, Harvey P. MD
Currently, almost all U.S. medical schools offer office-based primary care clerkships. Yet, primary care doctors are under such great pressure for productivity that teaching becomes a burden, if not a downright impossibility. Thus, there is an ongoing need to identify and recruit primary care physicians who can provide a sound educational experience for third-year medical students.
This tension between teaching and clinical productivity has led to changes in perspective on motivators and rewards for teaching, and to recommendations that schools pay “hard” money for teaching1 to relieve faculty from the burden of financially supporting themselves through clinical dollars.2 Although medical schools commonly reward primary care physicians who teach in their offices with academic appointments, plaques identifying their relationship to the medical school, and access to free CME programs, few have been able to pay enough for teaching to replace clinical income.3,4
Research has shown that teaching practices cost between 30% and 40% more to operate than nonteaching practices,4–6 patient encounters are approximately twice as long when a student is present,6,7 and either fewer patients are seen and/or the physician’s day is extended.4 Therefore, both the practices and the individual physicians absorb the costs of teaching. Despite this, because many physicians report that they teach for personal satisfaction and rate the value of money lower than the pleasure of teaching,3,7–11 we must ask whether payment for teaching is necessary.
The effect of motivators and rewards is complex. Psychologists have asked whether extrinsic reward is necessary at all, whether it undermines intrinsic rewards,12 and, if it is needed, how immediate, frequent, and direct it must be to be effective.13 In a seminal article on how to motivate employees, Herzberg14 explored the effectiveness of monetary incentives. He concluded that job satisfaction and motivation stem from needs “separate and distinct” from those leading to job dissatisfaction. Needs leading to motivation and satisfaction relate to an intrinsic and “unique human characteristic, the ability to achieve and, through achievement, to experience psychological growth.” Those leading to dissatisfaction are “extrinsic to the job”—negative “working conditions, salary, status and security.”
Today’s practice of primary care involves both sets of needs: a breadth of content and close relationships with patients and students (achievement and recognition), but also long work hours during which one must see many patients to meet the demands of insurers.15 It is important, then, to understand the complex nature of how primary care physicians balance the need for satisfaction of teaching with the need for compensation.
Harvard Medical School has offered a longitudinal, nine-month primary care clerkship for the past 10 years.16 From its inception in 1997, the clerkship paid preceptors a $600 stipend. However, for 27 four-hour sessions, this was more a token of appreciation than compensation for lost revenue. Because those of us in charge of the clerkship have always found it difficult to recruit enough preceptors to teach between 120 and 150 students each year, we hoped that raising the amount of the stipend would ease the financial burden for preceptors and, consequently, help us retain them as faculty. With cooperation from our deans, we raised the stipend to $900 in 2003 and then to $2,500 in 2004.
In this study, we examine the association between the rise in stipend and retention of faculty. Additionally, we explore faculty members’ perceptions of the relative value of (1) an external reward or motivator (the stipend) and internal rewards and motivators, and (2) receiving the stipend directly or indirectly.
Retention of preceptors
Between 1997 and 2006, a total of 404 primary care physicians (internists, pediatricians, and family physicians) taught in the Primary Care Clerkship (PCC) at Harvard Medical School. Each year, we needed between 120 and 150 preceptors, but some preceptors left the clerkship, and we sought replacements for them. According to our records, the reasons for their leaving varied: unfortunately, a few died, some moved to other medical schools, some took maternity leaves, some changed their days of practice and had no patients available on the day the clerkship met, and others simply no longer could or wanted to teach.
The preceptors worked in a variety of community and hospital-based practices, as well as a multispecialty health maintenance organization. When we recruited preceptors, we informed them of the amount of the stipend, which was paid at the end of the nine-month clerkship.
We have a database with complete records of the years each preceptor taught, reasons for each former preceptor’s decision not to teach as stated at the moment of recruitment, and the manner in which the stipend was paid (directly to the preceptor or indirectly via the preceptor’s appointing department). We verified our records by checking the medical school faculty database to determine whether preceptors who had left the clerkship had maintained an appointment or resigned. We then categorized preceptors according to teaching status in each year of the program: 0 = available to teach but declined (not retained), 1 = currently teaching (retained), 2 = willing to teach but not matched this year (e.g., because all patients speak Russian and no Russian-speaking student is available; considered retained), or 3 = unavailable (e.g., resigned or died; considered not retained).
We developed a brief survey to elicit preceptors’ views of the rewards they feel are important when deciding to continue teaching and those they feel are sources of satisfaction. With IRB approval, we e-mailed the survey to 402 present and past living preceptors; after two e-mail reminders, two medical student research assistants telephoned nonrespondents to request an interview. In the survey, we gave the following instructions:
* Rank the following in order of their importance to you in deciding to continue to teach in the PCC from year to year. That is, without this you would be unwilling or unable to continue. Place a 1 next to the item of greatest importance and a 6 next to the item of least importance. Rank-order the rest. Please avoid ties as that will make it more difficult to evaluate.
* Rank-order the following in terms of your feeling a sense of satisfaction in teaching in the PCC. Place a 1 next to the item that provides you the greatest source of satisfaction and a 6 next to the item that provides you the least.
The six items—based on prior research3,6,10,17 relevant to the rewards we offer, which represent both intrinsic and extrinsic motivators—are listed in Table 1.
We also asked preceptors who had discontinued teaching to state why they had done so; however, although interested in their post hoc rationales, we based coding on reasons they gave at the point of recruitment, as described above. We also gave current preceptors an opportunity to comment in free text on incentives that might be important to them regarding their willingness to participate in future teaching.
Because we informed faculty of the stipend upon recruiting them, and because we paid them at the end of the clerkship, we assume that the stipend is both a motivator and a reward. Therefore, we computed the annual retention rates and the average retention rate for each stipend period (i.e., when the stipend was $600, $900, and then $2,500) as well as in the year after the rise in stipend (i.e., when the stipend changed from $600 to $900, and again when it changed from $900 to $2,500). We examined the relationship between change in stipend and the proportion of faculty we lost each year either because they were unavailable or they declined to teach versus the proportion of those who were willing to teach, whether or not they were matched to a student. Then, on the basis of physicians’ availability to teach, we conducted logistic regressions to estimate the odds of returning after each rise in stipend using general estimating equations to account for correlation within physician.18,19
Further, we examined the frequency of rankings as well as the mean rank-order of factors involved in decisions to continue teaching and sources of satisfaction with teaching; we conducted t tests to determine the relationship between these rankings to teaching status (declined versus retained), and directness of payment (direct versus indirect). All statistical tests were conducted with a significance level of P < .05.
Among the 404 primary care physicians who served as preceptors at any time between 1997 and 2006, 315 (78%) were available to teach when surveyed in 2006–2007. Over the years, 119 (29%) of the 404 faculty declined to teach. Moreover, only 33 (8%) of the faculty who began teaching in the clerkship in its first or second year continued until the present. Most preceptors received the stipend directly: 326 of the 404 preceptors (81%). A total of 170 (42%) of all past and present preceptors responded to the survey, which represented 54% of the 315 available faculty. However, after we discarded incomplete surveys, we had responses from 154 (49%) of the available faculty.
Table 2 depicts the rates of retention (all preceptors who were willing and available to teach) and the proportion of faculty who declined to teach or left the medical school (not retained). Annual retention rates ranged from a high of 91% in 2006 to a low of 69% in 2000; conversely, the percentage of preceptors declining to continue to teach ranged from a low of 3% in 2006 to a high of 17% in 2000. The mean retention rate during the period when the stipend was $600 was 78%, and the proportion of willing teachers with whom we were able to match a student was 72%. In the year after receipt of the $900 stipend, retention rose to 86%. In the years after the stipend rose to $2,500, the mean retention rate rose to 89%, and we were able to match 75% of the willing and available faculty. To understand the relationship between the stipend and declining to teach, we excluded unavailable faculty from our analysis; we then found that the odds of returning to teach (i.e., those willing to teach, coded “1” or “2”) in the highest pay period versus the odds of returning to teach in the lowest pay period were 2.66 (95% CI 1.96, 3.62, P < .0001). Moreover, when we compare the odds of retaining faculty whom we could match with a student (i.e., those coded “one” only) with those who declined to teach, the odds of retention in the latter period were even greater: OR = 4.70 (95% CI 2.54, 8.71, P < .0001).
Retention was also significantly related to directness of payment: preceptors paid directly were more likely to continue teaching (P < .0001).
Because 16 respondents failed to rank-order all six factors or ranked more than one factor as number one, we analyzed the data twice, using first all responses and second only complete and correct rankings. Relative ranks did not change. We report the complete rankings here.
Important factors in deciding to continue teaching.
The majority—105 (73%)—of respondents ranked having a good student as the most important factor in their continuing to teach in the PCC (mean rank-order = 1.5 from among six factors—see Table 1). A total of 11 (about 8%) of the preceptors ranked receiving the stipend as the most important factor in their continuing to teach (mean rank-order = 3.7). Respondents who were teaching or willing to teach at the time of the survey—when the stipend was highest—ranked the stipend significantly higher than did those who had previously declined to teach (3.5 versus 4.1, P < .03).
Wondering whether the data were tainted by responses from the 33 preceptors who had taught 9 or 10 years (and for whom, perhaps, no stipend was too little), we compared their responses with those from respondents with shorter service. There were no significant differences between the groups.
Sources of satisfaction with current teaching.
Most respondents—127 (82%)—ranked having a good student first as a source of satisfaction with teaching (mean rank-order = 1.3). A total of 18 (about 12%) ranked having a student interested in a primary care career first (mean rank-order = 3.0). No one ranked receiving the stipend first as a source of satisfaction (mean rank-order = 4.2) (Table 1).
The stipend was significantly more important as an incentive to teach in the future (i.e., something faculty felt they needed to continue to teach) than as a source of satisfaction with teaching (mean rank-order 3.7 versus 4.2, P = .02).
Preceptors who received the stipend directly were more likely than those whose departments received the stipend to rank it high as a source of satisfaction with teaching (P = .03). But, in contrast with their actual behavior as evidenced by retention rates, preceptors who received the stipend directly were no more likely than those whose departments received the stipend to rank it high as an important factor in their decision to continue teaching (P > .05).
Stated importance of the stipend.
Seventy of the 170 survey respondents (41%) provided open-ended comments. When we examined these comments, we found that few respondents offered alternative incentives or sources of satisfaction, but, instead, most elaborated on their rankings of the six factors, especially remarking on the value of the stipend. The following comments sum up the matter:
The stipend is important for two reasons. I see fewer patients when I precept and I get penalized if I do not meet certain productivity goals. And a stipend is a concrete indication that one’s time and energy are valued.
Money ain’t everything, but it’s way ahead of whether [the students are] going into primary care.
Reasons for declining to teach.
We found no discrepancy between our records and respondents’ stated reasons for not teaching. However, their written comments provided richer detail because, although they explained their reasons for leaving in terms of time and money, they put their reasons in a broader professional perspective. For instance, some said that teaching conflicted with other professional activities, such as committee or volunteer work, as well as with clinical productivity; some felt that they could not meet patients’ demands for short waits and visits with a student in the office and that the flow of the office’s work was interrupted with a student present. While some directly mentioned lengthened workdays, others’ responses indirectly spoke to the idea that there was not enough time to teach and meet other obligations. One ex-preceptor’s comment describes his concern about the conflict between good teaching and clinical productivity:
The stipend did not cover the time needed to teach the student…. I did not feel I could do a good job without this compensated time.
Retention of preceptors in our PCC significantly improved after we raised the stipend for teaching. Moreover, receiving payment directly was associated with higher retention. In contrast, preceptors’ stated preferences for rewards and motivators for teaching suggested that, in the context of six possible intrinsic and extrinsic factors, a stipend was important but not primary. However, receiving the stipend was ranked higher as an incentive to continue teaching than as a source of satisfaction.
According to our respondents, the most important motivator, as well as the most important source of satisfaction with teaching, was an intrinsic factor, “having a good student.” Similarly, having a student interested in a primary care career was also a source of satisfaction. Moreover, a few faculty members reported that they not only enjoyed but continued teaching only as long as they had a good student; they discontinued after having a difficult student. This finding is in accordance with Herzberg’s14 report that intrinsic factors are associated with job satisfaction.
Other researchers have surveyed primary care physicians to ask why they teach medical students and what influences their decision to continue to teach. None has examined the relationship between expressed attitudes and actual behavior. In those studies, intrinsic factors (e.g., belonging to a group of teachers,10 being celebrated publicly or receiving tokens of appreciation,20 and personal satisfaction11) were often mentioned as primary reasons for teaching. Physicians’ ratings of external rewards (financial compensation, services, and free CME credits) varied. Both Single3 and Kumar11 and their associates reported that physicians rated most highly receiving CME credit as a reward. Yet, Single et al’s respondents rated financial compensation lowest among all factors influencing their intention to continue teaching. When Kumar et al examined both ratings (extent to which respondents “appreciated” the reward) and rank-order of rewards, they found that physicians rated financial payment relatively high (3.94 on a 5-point scale) but ranked it, along with gifts and services, at the bottom of the rank-order.
We examined the effects of different levels of stipend and directness of payment on actual retention, at the point of recruitment when physicians decided to accept or to decline to teach. Others have, instead, asked their faculty to rate the value of payment and to state their intention to teach again. Two studies reported that physicians who were paid for teaching rated the value of payment higher than did those who were not paid.11,21 The latter (Christiansen et al21) also reported that 94% of salaried faculty said they intended to teach again, whereas only 65% of the unsalaried planned to teach again.21
It is important to note that a small proportion (12%) of our primary care faculty have served as PCC preceptors since the clerkship’s inception in 1997. Like the 65% of unsalaried physicians in the Christiansen et al21 study who intended to teach again, these physicians seemed to derive intrinsic rewards from teaching. This is not to say that payment is neither desired nor helpful. As some indicated in open-text remarks, receiving the stipend, especially directly, serves as a token of appreciation and a statement of one’s value to the medical school. When the stipend level was raised, faculty members expressed the belief that administrators understood their sacrifices and cared enough to try to compensate them for lost revenue.
This study has several limitations. First, the response rate to the survey was very low, and the analyses were compromised by a few respondents’ failure to provide complete and discrete rankings for all six factors. Although their responses conform to those of primary care physicians previously studied, our ability to generalize the findings is limited. Second, we have no comparison group. Everyone, whether paid directly or indirectly, received the same compensation. Last, we did not retain records about the number of faculty we attempted to recruit each year. We base our study on those who, once recruited and having taught, either continued or declined to continue teaching.
Naturally, the extent to which we can understand why primary care physicians teach is limited by our research methods. First, it is important to differentiate between attitudes, intentions, and actual behaviors. We are fortunate to have complete data on all 404 preceptors’ tenures as teachers in the PCC as a measure of actual behavior. Their rank-ordering of six relevant rewards let us observe the relative importance of a stipend and to compare those attitudes with their willingness to teach. Second, we need to explore more fully the extent to which those factors we define as intrinsic or extrinsic rewards or motivators are, in fact, perceived to be such by faculty respondents. We allowed our respondents to define “good student” in their own terms. Is a good student one whose intelligence sparks excitement for the preceptor (an intrinsic reward) or one who can lend a hand and lighten the work load (an extrinsic reward)? Is a stipend money that buys one’s time or a token of esteem? Each factor might have an element of both intrinsic and extrinsic reward.
Third, we must consider how truthful respondents are likely to be or, indeed, are able to be, when asked about the importance of money—especially when those respondents are perceived as well-paid members of society. Some of the faculty who declined to teach but responded to our survey couched their admissions of the importance of the stipend in terms of its allowing them to have the time to teach well—possibly a more socially acceptable response. Moreover, the reasons one gives post hoc for ceasing to teach may differ from those one gives (or is willing to give) when called in the midst of a busy practice to make a commitment to teaching.
Last, it is important to note that recruitment and retention of primary care faculty is a complex issue. We need good teachers whose practices can accommodate students at the time when the students are scheduled to take the course. It is every clerkship director’s dream to have a surplus of faculty from whom she or he may select the best teachers and the best sites. For us, this has meant being able to send a Spanish-speaking student to the Spanish clinic or a student interested in serving the indigent to a homeless clinic. Being able to recruit and retain more faculty has allowed us to match students and preceptors carefully, which may, in turn, add to preceptors’ satisfaction with teaching.
As one researcher stated, “higher income, although extremely important, will not alone permit medical school faculties to attract and retain talented teachers…. Other rewards are important to professionals, especially academics.”20 However, we conclude that raising stipends beyond a mere token of thanks eases the burden of teaching in a busy practice such that the intrinsic pleasure of teaching is given play.
The authors acknowledge with gratitude the assistance of Ken Kleinman, PhD, and Fang Zhang, PhD, with the general estimating equations analyses. Dr. Kleinman is associate professor, and Dr. Zhang is instructor, Department of Ambulatory Care & Prevention, Harvard Medical School and Harvard Pilgrim Health Care, Boston, Massachusetts. The authors also thank Morgan Chessia and Sarah Kimball for conducting telephone interviews with initial nonrespondents; both are students at Harvard Medical School.
1 Ebert RH, Ginzberg E. The reform of medical education. Health Aff (Millwood). 1988;7(2 suppl):5–38.
2 Educating medical students. Assessing change in medical education—the road to implementation (ACME-TRI report). Acad Med. 1993;68(6 suppl):S1–S46.
3 Single PB, Jaffe A, Schwartz R. Evaluating programs for recruiting and retaining community faculty. Fam Med. 1999;31:114–121.
4 Boex JR, Boll AA, Franzini L, et al. Measuring the costs of primary care education in the ambulatory setting. Acad Med. 2000;75:419–425.
5 Hogan AJ, Franzini L, Boex JR. Estimating the cost of primary care training in ambulatory settings. Health Econ. 2000;9:715–726.
6 Bowen JL, Irby DM. Assessing quality and costs of education in the ambulatory setting: A review of the literature. Acad Med. 2002;77:621–680.
7 Usatine RP, Nguyen K, Randall J, Irby DM. Four exemplary preceptors’ strategies for efficient teaching in managed care settings. Acad Med. 1997;72:766–769.
8 Jason H, Westberg J. Teachers and Teaching in U.S. Medical Schools. Norwalk, Conn: Appleton-Century-Crofts; 1982.
9 Schindler N, Winchester DP, Sherman H. Recognizing clinical faculty’s contributions in education. Acad Med. 2002;77:940–941.
10 Starr S, Ferguson WJ, Haley HL, Quirk M. Community preceptors’ views of their identities as teachers. Acad Med. 2003;78:820–825.
11 Kumar A, Kallen DJ, Mathew T. Volunteer faculty: What rewards or incentives do they prefer? Teach Learn Med. 2002;14:119–123.
12 Tang S, Hall VC. The overjustification effect: A meta-analysis. Appl Cogn Psychol. 1995;9:365–404.
13 Bandura A. Social Foundations of Thought and Action. Englewood Cliffs, NJ: Prentice-Hall; 1986.
14 Herzberg F. One more time: How do you motivate employees? Harv Bus Rev. 2003;81:87–96.
15 Bodenheimer T. Primary care—will it survive? N Engl J Med. 2006;355:861–864.
16 Peters AS, Feins A, Rubin R, Seward S, Schnaidt K, Fletcher RH. The longitudinal primary care clerkship at Harvard Medical School. Acad Med. 2001;76:478–482.
17 Kumar A, Loomba D, Rahangdale RY, Kallen DJ. Rewards and incentives for non-salaried clinical faculty who teach medical students. J Gen Intern Med. 1999;14:370–372.
18 Liang K, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73:13–22.
19 Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42:121–130.
20 Greer DS. Faculty rewards for the generalist clinician–teacher. J Gen Intern Med. 1990;5:S53–S58.
21 Christiansen RG, Wark K, Levenstein JP. Attitudes of part-time community internal medicine faculty about their teaching. Acad Med. 1992;67:863–864.