PAPER 96-1 The Relative Importance of Undergraduate and Postgraduate Education to the Practice Decisions and Attitudes of Young Family Physicians Practising in Ontario C. Woodward, M. Cohen B.M. Ferrier, A.P. Williams "NOT FOR CITATION WITHOUT PERMISSION" 1 The Relative Importance of Undergraduate and Postgraduate Education to the Practice Decisions and Attitudes of Young Family Physicians Practising in Ontario * Christel A. Woodward, Ph.D. Professor, Clinical Epidemiology and Biostatistics McMaster University May Cohen, MD, CCFP, FCFP Associate Dean (Health Sciences) McMaster University Barbara M. Ferrier, Ph.D. Professor, Biochemistry McMaster University A. Paul Williams, Ph.D. Associate Professor, Health Administration University of Toronto * Corresponding author: Dr. C.A. Woodward, Room 3H4 Faculty of Health Sciences McMaster University 1200 Main Street West Hamilton, Ontario L8N 3Z5 telephone (905) 525-9140 ext. 22167 FAX (905) 546-5211 This study was funded through a Strategic Grant #822-92-0044 from the Social Sciences and Humanities Research Council of Canada. Cite as: Woodward C, Cohen M, Ferrier BM, Williams AP: "The Relative Importance of Undergraduate and Postgraduate Education to the Practice Decisions and Attitudes of Young Family Physicians Practising in Ontario", McMaster University Centre for Health Economics and Policy Analysis Working Paper #96-1, March 1996. 2 ABSTRACT Objectives: To examine the influence of medical education and residency education on decisions regarding type of practice, practice activities and attitudes of family physicians. Design: Secondary analysis of data from a mailed survey of a cohort of family physicians that had a 70% response rate. Setting: Ontario Participants: All physicians certified by the College of Family Physicians of Canada between 1989-1991 after completing a family medicine residency who resided in Ontario and had either graduated from an Ontario medical school or a residency program in family medicine offered by an Ontario medical school. Outcome measures: Selected questionnaire items regarding whether they extended their residency, practice location, primary source of remuneration, practice organization, services included in practice, satisfaction with their education and patient care attitudes. Results: Systematic differences were observed in professional decisions, behaviours and attitudes of the physicians studied that were associated with the medical or residency programs that they attended when other factors such as gender of the physician were taken into account. For example, provisions of shared obstetrical care was less frequently observed among those who extended their residency or took a residency in family medicine at the University of Toronto. Both these variables were associated with higher probability of providing intrapartum care. Solo practitioners were twice as likely as other practitioners to do shared obstetrical care. Rural practice was also associated with providing intrapartum care. Physicians involved in research were less likely to have a predominantly fee-for-service practice, but more likely to have attended McMaster medical school or extended their residency. 3 Conclusions: The differences observed are likely to reflect both self and program selection effects and the socialization/education provided by these educational programs. It is surprising that the differences by educational program were as striking as those we observed by gender, urban or rural location and type of practice despite the fact that all medical schools and family medicine residency programs exist in the same province. 4 Although there has been considerable conjecture about the “signatures” of different medical schools and postgraduate medical programs, their description in the Canadian literature is not extensive. One large-scale study followed the 1200 students who entered all Canadian Medical schools in 1965 and obtained information on careers entered four years after graduation <1,2>. In this study, academic inputs to a medical school (i.e., full time faculty and research activities) were found to have had more influence on career choices than clinical inputs (i.e., part-time faculty and residency programs), but the relationships observed were complex and not consistent. The authors concluded that this might mean that developments following medical school strongly affect physician career choices <1>. Ryten <3> has reported on differences in dispersion rates among graduates of Canadian medical schools and noted that schools differ in the proportion of graduates who migrate to other provinces or countries. Other studies have focused on comparing the outputs (in terms of location, field of medical practice chosen) for students graduating from all medical schools in a given province <4,5>. A series of studies from McMaster medical school have contrasted a single medical school’s output with a randomly selected national sample of peers <6> or with a randomly selected sample, matched to the index graduate by site of postgraduate training <7>. These studies have suggested that McMaster graduates, while not different from their peers in proportion choosing primary care and specialty practice, are more likely to be engaged in teaching, administration and research. McMaster graduates who become primary care physicians are more likely to have sought certification in family medicine and havea distinctive practice style. The practice patterns (as observed through administrative billing data) of the primary care output of McMaster’s medical school (19721983) has been compared with a sample of other Ontario medical school graduates matched on sex, year of graduation, and location <8>. McMaster graduates provided fewer services, saw fewer patients and earned less than matched contemporaries. Particularly, McMaster graduates did less emergency room work and performed fewer minor assessments of patients whereas they did more psychotherapy and performed more prenatal care than their contemporaries. Some studies, which compared various groups of Canadian primary care physicians who followed different postgraduate routes to primary care, have shown that different types of postgraduate training appear to affect primary care physicians’ self-rated initial practice competencies, <9> quality of care, <10> practice patterns, <11> and attitudes toward comprehensive care <12>. To date, no attempt has been 5 made to examine whether residency programs within the same field of medicine (e.g. family medicine or internal medicine) are associated with differences in attitudes, practice decisions or practice patterns of their graduates. This paper examines the influence of medical school and residency program on the decisions regarding type of practice, practice activities and on the attitudes of a cohort of Ontario primary care physicians who had recently entered practice after graduating from one of the five medical schools in Ontario or attending a family medicine residency program offered by one of the five Ontario medical schools. 6 METHOD This analysis uses data from a survey of all 1989-1991 residency-eligible certificands ofthe College of Family Physicians of Canada who are licensed in and reside in Ontario. The survey, fielded in the autumn of 1993, closed data collection at the end of January, 1994. At that time, 395 of 564 eligible physicians had responded, yielding a response rate of 70%. No difference in response rate was observed by sex of respondent, medical school of respondent, or rural/urban location of practice. However, certificands in 1990 (62.9%) were less likely to respond than certificands in 1989 (77.8%) or 1991 (71.3%) (X22 = 8.44; p = 0.01). A small portion of otherwise eligible respondents who had neither graduated from a medical school in Ontario nor attended a family medicine residency in Ontario (N=35), were eliminated from this analysis. The early practice development of primary care physicians questionnaire sought information about (1) route to medicine and the career path of the physician; (2) sociodemographic and professional characteristics of the respondents; (3) responders’ attitudes toward patient care and professional issues; (4) responders’ professional job satisfaction using the sixteen item scale developed by McGlynn <13>; (5) responders’ practice arrangements, and (6) their practice profile. Physicians who were currently not in practice or were doing locums were asked to skip the practice profile section. Some physicians who had entered a different residency program since gaining certification in family medicine also skipped the practice profile section. Choices of questions used on the questionnaire were informed by focus groups and indepth interviews with new-to-practice physicians. The data were analyzed using SPSS-PC+. First, descriptive statistics were used to examine the characteristics of the groups and characteristics of respondents versus non-respondents. Bivariate techniques were used to examine simple associations between twovariables. We then constructed logistic regression equations to examine the effect of medical school and postgraduate training site on the likelihood of practising in a rural area, the likelihood of solo or group practice, the likelihood of reporting that 95% or more of your professional income comes from fee-forservice practice, and the likelihood of three categories of practice behaviours (i.e., types of antenatal care, surgical care and hospital care). We also used linear regression techniques to examine whether medical school or post-graduate site are associ7 ated with these physicians’ attitudes about patient care and their overall satisfaction with their current situation, their undergraduate education and their postgraduate education. Given the difference in the proportion of women graduates among medical schools, sex of physician was entered into each regression equation along with dummy variables for the five Ontario medical schools and the five residency programs. We also created dummy variables for the following practice choices: rural location, extended residency beyond two years, predominantly (95%+) fee-forservice income, solo practice and group practice. We examined the influence of medical school and residency program in conjunction with the influences of these other variables. We regarded as significant associations where the p value was 0.01 or less. Associations with a probability between p = 0.05 and p = 0.01 were regarded as interesting. Description of Cohort Studied The cohort selected for study comprised all those Ontario family physicians certified in 1989-1991 who responded that they graduated from an Ontario medical school (N=320; 81%) or attended a family medicine residency program offered by one of the five Ontario medical schools (N=321). Of the 320 physicians who were Ontario medical graduates, 10% graduatedform the University of Ottawa, 15% from Queens University, 20% from the University of Western Ontario, 21% from McMaster University and 34% from the University of Toronto (see Table 1). Quite different patterns of retention for postgraduate training at the same institution were observed, which should be kept in mind as we look at subsequent analyses. Queens and Ottawa retained less than a third of their family medicine bound graduates who eventually chose to reside in Ontario and form part of our cohort. About half the family physicians in our cohort who graduated from University of Western Ontario and McMaster stayed to do their family medicine training there, while over three-quarters of Toronto graduates in this cohort also did their postgraduate training in Toronto. Finally, it should be noted that the proportion of women who graduated from the five medical schools differs significantly and ranges from 43% of Queens graduates in this cohort studied to 75% of McMaster graduates (X24 = 17.68 p <0.01). If we examine cohort characteristics by family medicine program, we observe that 40% of the group completed their residency at University of Toronto while the other four schools contribute 12-19% of cohort members. Nearly two-thirds of those in the Toronto and the University of Western Ontario residency programs were 8 from their own undergraduate medical school while the proportion of their own graduates in the family medicine residency programs at McMaster, Queens and Ottawa were 54%, 33% and 23% respectively. These associations should be noted because they introduce some collinearity into regression equations. (Toronto medical school and residency (r = 0.56); Western medical school and residency (r = 0.49); McMaster medical school and residency (r = 0.44).) Because our cohort contains physicians who either graduated from an Ontario medical school or attended an Ontario residency program, the final sample studied was 360. Influence of Medical School and Residency in Career Decisions. Extending Residency. Overall 81 (22.5%) of the physicians extended their family medicine residency program beyond two years. Of these, twenty-eight had additional training in obstetrics and gynecology (mean 4.7 months; sd = 2.30), four did additional training in anesthesia (mean 8.50 months; sd = 4.12), thirty-five chose to do additional training in emergency medicine (mean 9.71 months; sd = 4.11) and twenty-five reported training in other fields of medicine (mean 9.57 months; sd = 9.66). Most people (83%) extended their residency in only one area. Fourteen took two or more types of additional training. Table 3 shows that there was no significant difference in the proportion choosing to extend their residency by medical school. However, significant differences are observed in the likelihood of choosing to extend your residency by residency program. We entered dummy variables for each medical school and residency program, along with other potential confounding variables. Only three variables entering the equation explained extension of residency. Graduates of the University of Western Ontario (UWO) medical school were 2.80 times more likely than others to extend their residency. However, this odds ratio drops to near one (1.01) when they also attended the residency program at their home university which has a low odds ratio (0.36) for extending residency time. Thus, UWO graduates who went on to other residency programs were more likely to extend their residency. Graduates of the McMaster residency program are 2.64 times as likely to extend their residency as other physicians. Rural Location. Only forty-one (11.4%) of the 360 physicians had chosen to settle in a rurallocation 9 in Ontario. The distribution of rural physicians by medical school and residency program are shown in Table 4. Bivariate tests of association suggest that residency programs are more strongly associated with this location decision than medical school. Only two variables contributed significantly to explaining a rural location choice. Graduates of the University of Toronto’s residency program (OR = 0.21) and female physicians (OR = 0.50) were less likely than others to be in a rural location. Fee-for-service. This variable fee was created to describe physicians who derived 95% or more of their income from fee-for-service practice. Overall, 65.6% of the physicians studied said that 95% or more of their professional income in the past twelve months had been received from fee-for-service. The simple bivariate associations between this variable and medical school or residency program were equivocal. However, the logistic regression equation suggests that graduates of McMaster medical school are more likely than graduates of other schools to report that less than 95% of their professional income comes from fee-for-service. Just over 52% of McMaster graduates said 95% or more of their professional income came from fee-for-service compared with two-thirds or more of graduates from other medical schools. (See Table 5.) Type of Practice Organization. We divided physicians into three broad groups of practice organization: solo practice, group and other (including emergency medicine, locums, attending another residency program, and various institutional practices). As can be seen in Table 6, simple bivariate statistics were equivocal, with one in the interesting range. Thus, we again used logistic regression analysis. When examining independent variables associated with solo practice, thegroup practice dummy variable was not included in the equation. Three variables were important to predicting likelihood of solo practice: fee-for-service (OR = 2.70), Toronto medical school graduate (OR = 0.36), and female (OR = 0.48). Women and Toronto medical school graduates were least likely to have chosen solo practice, while deriving 95% or more of professional income from fee-for-service was associated with solo practice (OR = 2.70). We also examined predictors of being in a group practice. Having attended Toronto medical school (OR = 1.73) or being female was associated with greater odds (OR = 2.70) of being in a group. 10 Services Included in Practice. Physicians who were currently in practice and not doing locums were asked to indicate whether or not they included a number of services in their practices. We examined three sets of interrelated services (1) services related to pregnant women (shared care, complete antenatal care, no antenatal care); (2) services related to hospitalized patients (direct care, concurrent care and supportive care) and (3) services related to surgery for their patients (minor office-based surgery and surgical assists). Antenatal Care. Forty-one physicians (14.1% of those answering this section) indicated that they were not involved in antenatal care. Most physicians (55.5% overall) said they shared care of their pregnant patients, while 37.8% said that they did complete antenatal care including uncomplicated deliveries. Some physicians indicated they did both shared and complete care. Bivariate analyses suggested no differences among medical school graduates in their approach to caring for pregnant women while an interesting difference was observed for shared care by residency program (See Table 7.) When logistic regression was used to explore correlates of thesethree practice activities, four variables entered the equation for shared care, all at the interesting level. Physicians who extended their residencies were less likely to be doing shared care (OR = 0.45) as were Toronto graduates (OR = 0.55). Physicians who were in solo practice (OR = 2 .08) or who graduated from the McMaster residency program (OR = 2.37) were more likely to do shared care. Physicians who had extended their residency (OR = 2.81) or who worked in rural areas (OR = 3.08) were significantly more likely to do deliveries. Being a Toronto medical school graduate was associated with the increased likelihood of doing deliveries (OR = 2.13). Physicians who earned 95%+ of their income from fee-for-service were also somewhat more likely to do deliveries. In contrast to shared care and doing deliveries, only one predictor was seen for not doing deliveries. Physicians in solo practice (OR = 0.30) were the most likely to be involved in any antenatal or obstetrical care, an interesting difference. Care of Hospitalized Patients. About sixty percent of physicians stated they provide direct care and concurrent care <14>. Supportive care1 in hospital was provided by over 75% of the cohort who were in group or solo practice. Interesting differences were noted among medical school graduates in the proportion doing di11 rect care which ranged from 50.5% of Toronto graduates to 73.9% of University of Western Ontario graduates. Significant differences were seen by residency program in the proportion providing direct hospital care, with University of WesternOntario again being the highest and Toronto the lowest. (See Table 8.) When we examined the independent influences of each variable taking its association with other variables into account, five variables explained an interesting or significant portion of the variance. Extending your residency (OR = 2.39), rural location (OR = 3.56) and having a predominantly fee-for-service practice (OR = 1.81) increased the likelihood that the physician provided direct care to patients in hospital. Being female (OR = 0.56) or a graduate of the Toronto residency program (OR = 0.35) decreased likelihood of being involved in direct hospital patient care. We also examined whether there were bivariate and multivariate associations with providing supportive care. Bivariate analyses were not interesting. However, being a Toronto medical school graduates (OR = 0.56) or a Ottawa residency program graduates (OR = 0.40) decreased the likelihood that the physician did supportive care. There were no significant bivariate or multivariate associations with provision of concurrent care. Office-based minor surgery and surgical assists. About half of the cohort of physicians studied indicated that they do surgical assists for their patients. No differences were observed across medical schools or residency programs in the likelihood of their graduates to report including surgical assists in their practices. (See Table 9.) The logistic regression equation suggested that female physicians (OR = 0.53) were about half as likely as males to include doing surgical assists in their practice while those who extended their residency or bill 95%+ fee-for-service were more likely to be involved in surgical assisting. Over eighty percent (83.2%) of physicians studied indicated that they include office-based minor surgery in their practices. Bivariate analyses suggested that medical school was unimportant to this 1 Concurrent care is defined in the Ontario Medical Association schedule of fees as applies when the family physician remains the most responsible physician but because of the seriousness or complexity of the condition, requests continued directive care by a consultant; supportive care is defined as care rendered in hospital by the referring physician who is not actively treating the case, to a patient under the care of another physician. 12 decision. However, there was an interestingdifference by residency programs. When factors associated with doing minor surgery were further explored using logistic regression, two interesting associations were seen. UWO residency program graduates (OR = 9.3) were more likely to do office-based minor surgery as were physicians working in groups (OR = 2.23). Female physicians were significantly less likely (OR = 0.40) to do so. Time spent in administration, research and teaching. Respondents were asked what percentage of their time they typically spend in various activities. We chose to examine how many recorded any percentage time on other administrative activities (outside of managing their practices) such as hospital committees, and on research and teaching activities. In each case we examined simply whether any time is spent to be able to compare these data with our previous study of all McMaster graduates 1972-1979. As seen in Table 10, the proportion spending any time in research was significantly different by medical school in the bivariate analysis while not significantly different by residency program. Logistic regression results suggested that deriving 95%+ of your income from fee-for-service decreased the odds of doing research (OR = 0.31) while graduates of McMaster medical school (OR = 2.89) were significantly more likely to do research than their colleagues in this cohort. The proportion doing teaching also varied (from 24% to 43.5%) by medical schools, an interesting difference. Among residency programs, a significant difference was observed in the proportion of graduates in this cohort who did teaching. Logistic regression identified two variables as contributing to the variance observed. Physicians who said they received 95%+ of their income from feefor-service were half as likely to spend some time teaching (OR = 0.49) and graduates of the Western Ontario residency program wereleast likely to report teaching activity (OR = 0.09). Using bivariate techniques, involvement in administration was associated more strongly with residency programs than undergraduate medical education. (See Table 10.) Logistic regression results suggest that physicians in group practices (OR = 4.34) and solo practices (3.62) are significantly more likely than the other group of physicians to be involved in administration beyond their practices. Ottawa residency graduates (OR = 2.70) are significantly more likely to be doing some administrative work beyond management of their own practices. McMaster medical graduates (OR = 1.82) were also more likely to be involved in some outside of prac13 tice administrative activities, at an interesting level. Satisfaction. Physicians were asked to complete a sixteen item satisfaction scale developed by McGlynn <13> at the Rand Corporation and used in several U.S. medical outcomes studies <15>. Coefficient alpha for the scale was 0.857 in this study. No differences were observed in overall scores on this scale by gender, medical school or residency program. The mean score of the cohort was 57.03 (out of a possible score of 80). Two factors, related to type of practice, were associated with overall satisfaction. Being in group practice (Beta = 0.20; t = 4.09 p = 0.0001) or solo practice (Beta = 0.17; t = 2.349; p = 0.019) were associated with overall satisfaction compared to those doing locums, emergency room work, and other types of practice. The Professional Satisfaction Scale also has several subscales. Of interest to us were the two items related to medical school or residency education which make up the training and education subscale. This subscale had an alpha of 0.472 for our sample. We thus chose to examine each item individually rather than together. The first item enquires about satisfactionwith the extent to which medical school has prepared you for practice and the second item queries satisfaction with the ‘quality of the teaching in your residency program’. As seen in Table 11, graduates of McMaster medical school were significantly more satisfied with the extent to which their medical school had prepared them to practice than others, while members of our cohort who graduated from the University of Ottawa reported the highest level of dissatisfaction. The responses from the graduates of other three schools were fairly similar. This pattern is also seen in the regression analysis. No significant difference is noted among residency programs in the satisfaction of their graduates using bivariate analytic techniques with 45.8% to 60.5% of responders indicating that they were satisfied. When both residency program and medical school are entered into the linear regression equation along with the six control variables, more graduates of the UWO residency are seen to be dissatisfied with the quality of teaching in their residency program, a difference at the interesting level. 14 Patient Care Attitudes. A number of items were written or adapted from other sources <16-19> to capture physicians’ attitudes toward patient care. These items were subjected to principal components factor analysis with varimax rotation. Four factors (scales) were derived that have acceptable eigen values. After examining the item content of each scale, we labelled the scales empathy (coefficient alpha = 0.74); directive, symptom oriented approach (coefficient alpha = 0.55); interest in counselling (alpha = 0.61); and disinterest in psychosocial issues (alpha = 0.61). Empathy. The five items that comprise the empathy scale are given in Table 13. The mean for the group was 20 out of a possible 25 points, suggesting a high level of empathy. Interesting differences among medical schools were noted in scores on this scale but no difference was observed at the postgraduate training level. Linear regression indicated that as a group UWO graduates scored significantly lower on this scale, while female physicians and physicians in solo practice scored significantly higher. Directive, symptom-oriented approach scale consists of five items. (See Table 14.) The scoring of one item, ‘working with patients who require major life-style changes to improve their health status is enjoyable’ is reversed. Again, greater differences were noted among participants in different medical school programs than in different residency programs. Two variables entered the linear regression equation. Graduates of the University of Toronto and McMaster medical schools scored significantly lower (were more likely to disagree) on items on this scale. Interest in counselling scale contains four items. (See Table 15.) Overall, there was high interest in counselling among respondents. Again, an interesting difference was observed among medical schools but not residency training programs in how their graduates responded. The linear regression equation suggests that two factors can account significantly for the explained variance. Female physicians score higher on this scale as do graduates from the University of Ottawa. Disinterest in psycho-social aspects of medicine is a four item scale (where scores can range from 4 to 20). Most respondents scored low on this scale, suggesting interest in psycho-social aspectsof care. (See Table 16.) Again, a significant difference was observed in the bivariate analysis at the medical school level but not at the residency program level. Lower scores on this scale were seen among female 15 family physicians, McMaster medical school graduates and University of Toronto medical school graduates when compared to the remainder of the group. 16 DISCUSSION The observed influence of gender on a range of practice decisions and behaviours examined is not surprising. Although gender did not affect satisfaction with medical school or residency training, women were less likely than men to have chosen to be in solo practice or to locate in rural areas (seen many times in previous studies) <20-23>. They were less likely to include minor office surgery or surgical assists as part of their practice, which fits with previous observations that women may be less interested in performing technical procedures and being involved in surgery than men <24-25>. As well, female physicians’ attitudes toward patient care were often different from men. Women were higher in empathy and interest in counselling as measured in this study and more interested in the psycho-social aspects of medicine, an observation made previously by Maheux <26>. Many of the differences noted are congruent with differences between men and women physicians observed by other investigators <20-28>. Further analysis of this rich data set will allow greater explication of the reasons for these differences in choices between men and women. Many of the correlates of solo practice, a style of practice chosen less often by female physicians and Toronto graduates in this study, have also been previously observed. Physicians in solo practice reported more often that they were involved in shared antenatal care but it wasunlikely that they did not provide any care to pregnant women. Physicians in solo practice were most likely to report 95% or more of their income comes from fee-for-service. Solo practitioners scored higher on the empathy scale. Another important explanatory variable was whether or not the physician had extended his/her residency. Those who extended their residency were more likely to do their own antenatal care, provide direct care to hospital patients and do surgical assists. Yet, many of those who extended their residency did so to focus on one area of practice, most often emergency medicine and obstetrics. Interestingly, graduates of the McMaster residency program, although more likely to extend their residency, were more likely as a group to do shared care rather than complete antenatal care. Extending residency also links to later doing research. Rural practice was not associated with extending your residency. 17 Rural location was associated with greater likelihood of reporting that the physician did complete obstetrical care and was likely to be directly responsible physician for his/her hospitalized patients, a finding consistent with other reports <29>. Toronto residency graduates were the least likely to be in rural practice or to be doing direct care of hospitalized patients. Systematic differences were repeatedly observed in the professional decisions, behaviours and attitudes of family physicians studied that are associated with the medical school or residency program which they attended. Toronto medical school graduates have a fairly distinctive, and likely more urban, pattern of practice. (There was some collinearity between the Toronto undergraduate and residency variable given that two thirds of Toronto medical school graduates stayed in Toronto for their residency.) They are less likely to provide supportive care to their hospitalized patients but are more likely to work in groups (than solo practice) and to deliverbabies (rather than offer shared care arrangements to pregnant women) than other medical graduates. Both Toronto and McMaster medical graduates are less likely to take a directive, symptom-oriented approach to patients and are more likely to be interested in the psycho-social aspects of care than other medical school graduates. McMasters’ medical graduates were less likely to earn their professional income predominantly from fee-for-service activities, an observation made previously <8>, were the most satisfied with the preparation for practice offered by their medical school education, and were more likely to be involved in research and administrative duties beyond office management than graduates of the other Ontario medical schools. This latter finding is similar to that observed in an earlier cohort of all McMaster graduates (1972-79) who were compared with a national sample of other medical school graduates <6>. Fewer correlates were found among the decisions and attitudes examined to distinctively characterize the graduates of the three other Ontario medical schools. Ottawa graduates were most likely to express an interest in the counselling role of the physician and were the least satisfied with their medical education. Western graduates were lower in empathy as measured in this study. Queens graduates did not display a distinctive pattern of behaviours or attitudes among those we examined. Residency programs contributed to the type of practice activity their graduates undertake. Particularly, Ottawa residency program graduates were more likely to report administrative activities while UWO residency graduates were least likely 18 to be involved in teaching. Toronto residency graduates were more unlikely to settle in rural areas or to be involved in direct care of hospital patients. While residents in the McMaster program were the most likely to choose to extend their residency, they were more likely to choose to do shared antenatal care. UWOresidency graduates were least likely to extend their residency, least satisfied with their residency, and most likely to engage in office-based surgery. Some caveats must be added about the cohort studied. We did not attempt to study every graduate of these medical schools and residency programs or even every graduate of these programs who had entered family medicine. Our sampling frame began with selection of those certified in family medicine during a three year period following a family medicine residency and was limited to physicians who had current addresses in Ontario. We recognize that different medical schools have historically had somewhat different in province retention rates. Particularly, the Univer sity of Ottawa, which is the only medical school studied where students are expected to be bilingual, has had a higher outflow of graduates than the other medical schools <3,30>. However, there was not a differential response rate by graduates from some medical schools; cohort members who resided in Ontario were equally likely to respond irrespective of medical school of graduation. We were unable to obtain information on the residency programs undertaken by non-respondents and thus are less confident that no bias in response occurred by residency program. Yet, the numbers in the cohort who graduated from the different programs also look proportional to the intake of the programs. Finally, two medical schools retained a sizeable proportion of their graduates who entered family medicine in their own residency program. This caused some collinearity in the medical school and residency variables for these schools which may account for the more unstable regression coefficients (i.e., things attributed to one level of education may be true of the other). These study limitations are acknowledged. We recognize that we must be cautious about attributing the differences observed to the socialization/education which occurred in these educational settings. Physicians apply both tomedical schools and residency programs. Self-selection into educational programs occurs. Educational programs admit their students after a selection process; differential selection by programs may also be operating.The question becomes - which, if any of these factors are salient? Yet, fascinating differences were observed which beg for further explication. Differences by educational 19 program, although all were medical schools and family medicine programs in the same province, were often as striking or more conspicuous than the differences that we observed by gender, urban/rural practice location or type of practice. These latter associations are easier to accept because there is already a larger body of supporting evidence in the literature which is consistent with our findings. Our study suggests that medical schools and residency programs also have ‘signatures’ although the factors which help create those signatures are a long way from being understood. Developing such understandings, while a daunting task, should be part of our research agenda. Acknowledgements: The authors would like to thank the physicians who participated in this study. 20 REFERENCES 1. Roos NP, Roos LL. Medical school impact on student choice: A longitudinal study. Evaluation and the Health Professions, 198; 3:3-19. 2. Roos NP, Fish DG. Career and training patterns of students entering Canadian medical schools in 1965. Canadian Medical Association Journal, 1976; 112:65-70. 3. Ryten E. The changing demographics of physician supply in Canada: How did we get where we are and were are we going? Does it matter? In: M. Watanabe (Ed.) Physician Manpower in Canada, Calgary: University of Calgary Printing Services, 1988. 4. Rothman AI, Cox JA, Kraemer J. Ontario’s 1972 medical graduates: Roles and locations in 1982. Ontario Medical Review, June, 1983; 302-305. 5. Jennett PA, Hunter KL. Career and practice profiles of Alberta medical graduates 1973-85: Implications for Manpower planning and decision making. Canadian Medical Association Journal, 1988; 139:345-349. 6. Ferrier BM, Woodward CA. A comparison of the career choices of McMaster medical graduates and contemporary Canadian graduates: A secondary analysis of physician manpower data collected by the Canadian Medical Association. Canadian Medical Association Journal, 1987; 136:39-44. 7. Woodward CA, McAuley RG, Ridge H. Unravelling the meaning of global comparative ratings of interns. Research in Medical Education. Proceedings of the Twentieth Annual Conference, Washington, D.C., November, 1981; 149-154. 8. Woodward CA, Ferrier BM, Cohen M, Goldsmith C. A comparison of the practice patterns of general practitioners and family physicians graduating from McMaster and other Ontario medical schools. Teaching and Learning 21 in Medicine, 1990; 2:79-88. 9. Curry L, Woodward CA. A survey of postgraduate training for family practice. Canadian Medical Association Journal, 1985; 132:345-349. 10. Borgiel AEM, Williams IJ, Bass MJ, Dunn EV, et al. Quality of care in family practice: Does residency training make a difference? Canadian Medical Association Journal, 1989; 140:1035-1043. 11. Woodward CA, Cohen M, Ferrier BM, Goldsmith CH, Keane D. Correlates of certification in family medicine in the billing patterns of Ontario general practitioners. Canadian Medical Association Journal. 1989; 147:897-904. 12. Beaudoin C, Maheux B, Beland F. Influence of training in family medicine residency on physicians’ attitudes toward comprehensive care. Canadian Family Physician, 1989; 35:2413-2416. 13. McGlynn E. Physician job satisfactions: Its measurement and use as an indicator of system performance. Unpublished doctor dissertation, RAND Graduate School, Santa Monica, California, 1988. 14. Ontario Medical Association Schedule of Fees. April 1, 1989. Toronto: Ontario Medical Association, 1989. 15. DiMatteo MR, Sherbourne CD, Hays RD, Ordway L, et al. Physicians’ characteristics influence patients’ adherence to medical treatment: Results from the medical outcomes study. Health Psychology, 1993; 12:93-102. 16. Ashworth CD, Williamsom P, Montarno D. A scale to measure physician beliefs about psychosocial aspects of care. Social Science and Medicine, 1984; 19:1235-1238. 17. Wolf MH, Putnam SM, Stiles WB, James SA. The medical interview satisfaction scale: Development of a scale to measure patient perception of physician behaviour. Journal of Behavioural Medicine, 1978; 1:391-401. 22 18. Ware JE, Snyder MK, Wright WR. Development and validation of scales to measure patient satisfaction with health services. Volume 1 of a Final Report; Part A: A review of the literature, overview of methods and results regarding construction of scales. Carbondale, Illinois: Southern Illinois University, School of Medicine, 1976. 19. McGaghie WC, Boehlecke B, DeVillis BM., Contreras A, Beckei M. Development of a measure of attitude toward pulmonary disease prevention. Evaluation and the Health Professions, 1993; 16:106-118. 20. Williams AP, Domnick-Pierre K, Vayda E, et al. Women in medicine: Practice patterns and attitudes. Canadian Medical Association Journal, 1990; 143:194-201. 21. Jennett P, Hunter KL. Career and practice profiles of Alberta medical graduates (1973-1985). Canadian Medical Association Journal, 1988; 143:625628. 22. Carter RG. The relation between personal characteristics of physicians and practice location. Canadian Medical Association Journal, 1987; 136:360368. 23. Maheux B, Dufort F, Lambert J, et al. Do female general practitioners have a distinctive type of medical practice? Canadian Medical Association Journal, 1988; 139:737-740. 24. Railton RH, Tholl W, Sanmartin CA. The case of Canadian general surgeons: Getting the message heard. Canadian Journal of Surgery, 1993; 36:111-113. 25. Martin SC, Arnold RM, Parker RM. Gender and medical specialization. Journal of Health and Social Behaviour. 1988; 29:333-343. 26. Maheux B, Dufort F, Beland F, et al. Female medical practitioners: More preventive and patient oriented? Medical Care, 1990; 28:87-92. 23 27. Keane D, Woodward CA, Ferrier BM, Cohen M, Goldsmith CH. Female and male physicians: Different practice profiles. Canadian Family Physician, 1991; 37:72-81. 28. Cohen M, Ferrier BM, Woodward CA, Goldsmith C. Gender differences in the practice patterns of Ontario family physicians (McMaster University graduates). Journal of the American Medical Womens Association, 1991; 46(2):49-54. 29. Rouke J. Perspectives on rural medical care in Ontario. Canadian Family Physician, 1991; 37:1581, 1583-1584, 1647. 30. Thurber D. A snapshot from Caper. ACMC Forum, 1994; 27:6-9. 24 25 26 27 28 29 Logistic Regression Resluts Solo: OR Group Fee 2.70 # Toronoto MS 0.36 # Toronto MS Female 0.48 # Female OR 1.73 * 2.70 # + locums, emergency medicine, new residency, and other institutional practices * 0.05 > p > 0.01 # p < 0.01 30 31 32 * 0.05 > p > 0.01 + p > 0.01 33 Solo 3.62 + * 0.05 > P > 0.01 + P < 0.01 34 35 36 37 38 39 40
© Copyright 2026 Paperzz