Preparedness of Internal Medicine and Family Practice Residents for Treating Common Conditions FREE

Residents in internal medicine (IM) and those in family practice (FP) are trained to provide primary care for adults, yet they have very different residency experiences. Internal medicine training was originally hospital based and emphasized acute care and pathophysiology of disease. More recently, IM training has included an increased emphasis on primary care.^1- 2 Family practice training, initiated in 1969, provides more experience in office-based ambulatory and long-term settings and focuses more on preventive and psychosocial aspects of care.^1,3- 4 In addition to primary care for adults, FP residency includes training in pediatrics, surgery, and obstetrical care. The different emphases in the 2 types of residencies may affect how internists and family physicians approach their adult patients. This may also influence their self-perceived preparedness to provide care for certain types of patients and treat their associated conditions.

Previous studies suggest that IM residents assume a more technical approach to patient care compared with FP residents, who emphasize preventive services and counseling.^5- 6 Surveys of IM residents have found that they believe they are underprepared for a variety of primary care tasks, including treating depression⁷ and performing pelvic examinations.⁸ Similarly, a study using case vignettes compared the self-perceived competencies of IM and FP residents and found that FP residents report greater competence in managing depression whereas IM residents report greater competence in treating acute myocardial infarction and its complications.⁹ These studies, however, are limited to single training sites. Little is known about differences in IM and FP training across programs or the extent to which these differences may affect resident preparedness to handle common adult conditions.

We hypothesized that the greater emphasis on outpatient care in FP residencies would be associated with greater self-perceived preparedness for outpatient conditions among FP residents than IM residents and that, similarly, the greater emphasis on inpatient care among IM residencies would be associated with greater self-perceived preparedness for inpatient conditions among IM residents than FP residents. We further sought to explore the relative importance of exposure to the settings in which the specific clinical conditions were found to IM and FP residents' sense of preparedness. To address these hypotheses, we used data from a national survey of resident preparedness.¹⁰

The sampling methods used in this study have been described previously.¹⁰ Briefly, we used a multistage process to assemble a sample of residents across 8 specialties (including IM and FP) to obtain a representative sample of residents training in their final year at US academic health centers. The final list of academic health centers, defined as medical schools and their closely affiliated or owned clinical facilities, contained 162 hospitals that were responsible for training 40 000 of the 98 000 residents in 1997.¹¹ Our final sample consisted of 578 IM residents in 25 programs (8% of accredited IM categorical programs) and 658 FP residents in 75 programs (16% of accredited FP programs).¹⁰ We conservatively estimated that our sample represents 53% of senior residents training in IM and 27% of senior residents training in FP, or 44% of all senior residents in both specialties. The lower percentage of FP residents reflects the greater number in nonacademic health centers that sponsor FP residencies.

Development of the survey instruments (1 for each specialty) was informed by literature reviews, focus groups, review of relevant Accreditation Council on Graduate Medical Education policies, and comments from experts in each of the respective specialties.¹⁰ The instruments were pretested using cognitive interviews and were designed to take 15 minutes to complete.

The survey was administered in the spring of 1998. Mailed surveys constituted 92% of the IM responses and 90% of the FP responses, with the remainder of the responses in each group obtained via telephone surveys. Response enhancement techniques included advance notification, multiple mailings, telephone follow-ups, and flexible scheduling. Respondents were eligible to receive cash prizes or (in some cases) payment for completed interviews. The initial (unadjusted) response rate was 48% for IM and 49% for FP, reflecting relatively high invalid identification rates (27%) of residents who had left their programs or were assigned the wrong specialty survey. The final survey response rate, adjusted for invalid sample, was 65%. Internal medicine and FP samples had adjusted response rates of 59% and 68%, respectively. The Massachusetts General Hospital (Boston) Institutional Review Board approved this protocol.

Clinical Preparedness. We asked residents how prepared they felt to diagnose and treat (1) inpatients, critically ill patients, and ambulatory patients in general; (2) four specific conditions typically associated with the inpatient setting (acute myocardial infarction, diabetic ketoacidosis, acute asthma, and acute renal failure); and (3) eight specific conditions associated with the outpatient setting (diabetes, hypertension, low back pain, vaginitis, headache, depression, upper respiratory tract infection, and hyperlipidemia). All of these adult conditions were listed together on both IM and FP surveys. The FP survey contained a separate section listing pediatric diagnoses. The response categories for each question were very unprepared, somewhat unprepared, somewhat prepared, and very prepared.

Resident Exposure by Setting. Because we considered exposure to patients a potentially important mediator of differences between IM and FP preparedness, we attempted to assess the influence of exposure to patients in different settings on IM and FP reports of preparedness. Exposure to patients is a function of both the time spent in a setting and the volume of patients seen in that setting. We hypothesized that both components of exposure would be important because programs vary in the amount of time residents are scheduled in inpatient and outpatient settings, and medical centers and practices vary in the volume of patients seen in each setting. We had previously tested questions related to resident reports of the percentage of time spent and the number of patients seen in inpatient and outpatient settings, both alone and in combination. We found that a combined measure provided, on average, the greatest explanatory power and the highest statistical significance in our models, and, therefore, used those in our analyses. Inpatient exposure was calculated by multiplying the average number of patients admitted during a shift as an intern by the percentage of residency spent in the inpatient setting (excluding the operating room and emergency department). Outpatient exposure was calculated by multiplying the average number of patients typically seen in a 4-hour outpatient clinic by the percentage of residency spent in ambulatory settings.

Other Variables. Respondents were asked about their sex, whether they attended medical school in the United States, and their race/ethnicity (white non-Hispanic, black non-Hispanic, Hispanic, Asian, or Alaska native/American Indian). In separate analyses, we found that white non-Hispanic and Asian respondents were not significantly different in demographics and responses, so we combined their responses and compared these with the other groups constituting underrepresented minorities. We also asked residents about their intention to subspecialize after residency (yes, no, or undecided). Because all respondents surveyed were in their final year of residency, "undecided" responses were combined with "no" responses since these both reflected an absence of definite plans for subspecialty training.

Questions with multiple response categories were collapsed to form dichotomous variables. For our main outcome variable, preparedness, we compared very prepared with somewhat prepared/somewhat unprepared/very unprepared because we were interested in factors that determine superior preparation for diagnosis and treatment of the conditions surveyed. Because differences between very prepared and somewhat prepared may not be clinically meaningful, we compared all conditions using very prepared and somewhat prepared with somewhat unprepared and very unprepared in a secondary analysis. The patient exposure variables were entered into our models as continuous variables.

We tested differences between IM and FP characteristics using the χ² statistic (dichotomous variables) and the t test (continuous variables). We tested for differences between IM and FP in responses to questions regarding site-specific preparedness (inpatient, outpatient, or critically ill setting) as well as differences in reported preparedness for each of the 4 inpatient and 8 outpatient conditions. All dichotomous comparisons of preparedness between IM and FP were initially tested using the χ² test. In addition, we conducted multivariate analyses of comparisons of preparedness between IM and FP residents controlling for exposure to patients, sex, US medical school graduate status, race/ethnicity, and plans to pursue a subspecialty or fellowship. Odds ratios (ORs) from multivariate analyses were converted to adjusted relative risks following the method described by Zhang and Yu.¹² Analyses were weighted to adjust for differences in sampling and response rates among the strata. Multivariate analyses included adjustment for the multistage sampling design. Analyses were computed using SUDAAN software version 7.5.1(Research Triangle Institute, Research Triangle Park, NC). P≤.05 was considered statistically significant.

Internal medicine and FP residents were similar in race/ethnicity and the proportion who graduated from a US medical school. Family practice respondents were more likely to be women and were less likely to be planning subspecialty training (Table 1). Differences between IM and FP exposure to patients differed by clinical setting. Internal medicine residents reported greater exposure to inpatients, and FP residents reported greater exposure to outpatients (P<.001 for both comparisons).

We tested differences in resident-reported preparedness to treat 3 different types of patients: inpatients, critically ill patients, and outpatients. Internal medicine residents were more likely than FP residents to report being very prepared to treat inpatients (252/277 [91%] vs 179/320 [55%]; P<.001) and critically ill patients (183/278 [68%] vs 57/321 [17%]; P<.001). Family practice residents were more likely than IM residents to report being very prepared to treat outpatients (262/320 [82%] vs 134/277 [48%], P<.001).

Internal medicine residents were significantly more likely than FP residents to report being very prepared to care for patients with acute myocardial infarction, diabetic ketoacidosis, acute asthma, and acute renal failure (Table 2). These findings persisted in multivariate analyses. Resident characteristics were not significantly associated with preparedness for any of the 4 inpatient conditions (data not shown). Our measure of resident exposure to inpatients was generally not associated with reports of preparedness for inpatient conditions. One exception was diabetic ketoacidosis, which residents reporting increased inpatient exposure were more likely to report being very prepared to manage (OR, 1.34; 95% confidence interval [CI], 1.05-1.72; P = .02).

The results were more variable for the 8 outpatient conditions (Table 2). More IM residents than FP residents reported being very prepared to treat diabetes. In contrast, FP residents were significantly more likely than IM residents to report being very prepared to treat depression, headache, low back pain, upper respiratory tract infection, and vaginitis. Differences between IM and FP residents were not statistically significant for hyperlipidemia and hypertension. The direction of the associations between specialty and preparedness did not change when we recategorized the dependent variable to very prepared and somewhat prepared vs somewhat unprepared and very unprepared, but most differences were no longer statistically significant.

In the multivariate analysis controlling for exposure to patients, IM residents reported greater preparedness to treat diabetes, hyperlipidemia, and hypertension. Reports of greater preparedness by FP residents remained significant for vaginitis and marginally significant for depression (Table 2). We found that increased exposure to patients in the outpatient setting was associated with greater preparedness to treat outpatient conditions for 6 of the 8 conditions studied (Table 3). The effects of exposure to outpatients on resident reports of preparedness were most marked for headache and depression and somewhat less marked for low back pain.

Some resident characteristics other than specialty were significantly related to preparedness to diagnosis and treat the outpatient conditions. Women were significantly more likely than men to report being very prepared to diagnose and treat depression (OR, 1.70; 95% CI, 0.99-3.06; P = .02) and vaginitis (OR, 2.53; 95% CI, 1.63-3.94; P<.001). Also, residents who did not intend to subspecialize were more likely to report being very prepared to treat vaginitis (OR, 2.04; 95% CI, 1.36-3.05; P = .005).

This report presents the results of a national survey of IM and FP residents in their final year of training in 162 US academic health center hospitals. We found significant differences between IM and FP residents' self-reported preparedness to diagnose and treat common conditions in inpatient and outpatient settings. Internal medicine residents reported greater preparedness than FP residents to diagnose and treat all 4 inpatient conditions, while FP residents reported greater preparedness to diagnosis and treat most of the 8 outpatient conditions included in our survey.

In this study, we measured residents' perceptions of their preparedness. The technique of self-assessment has been widely used,^{5- 6,13- 16} although how well residents' perceptions match some objective standard of preparedness is not well known.¹⁷ Some authors have found little correlation between physician self-assessment and objectively measured competency,^18- 20 but others have found that physicians are able to predict their performance reliably.^21- 24 Residents, however, may underrate themselves compared with the ratings of their supervisors.^25- 26 Therefore, our results must be interpreted within the limitations of self-assessment.

There are several possible explanations for the differences in IM and FP residents' reports of preparedness found in this study. Internal medicine and FP programs differ in their training emphases and their residency review committee requirements for exposure to inpatients and outpatients. Internal medicine residents spend at least 12 months in inpatient teaching services compared with 6 months of adult inpatient care for FP residents. The greater time spent on inpatient care for IM residents and outpatient care for FP residents was reflected in our findings. The IM residency review committee has noted internists are distinguished by their diagnostic skills² whereas the FP residency review committee highlights abilities in providing continuous and comprehensive care.⁴ In addition, differences in self-reported preparedness may reflect differences in the interests and career choices of the residents in each specialty.²⁷ Internal medicine is a gateway for many subspecialties, whereas FP is predicated on providing comprehensive primary care for a broad range of patients.

Exposure to patients in inpatient and outpatient settings was significantly associated with preparedness for several conditions included in the study. It is unlikely that the variable we used for patient exposure reflects the complexity of patient-resident interactions; nonetheless, our results do reflect the importance of patient exposure on resident reports of preparedness for certain inpatient and outpatient conditions. Of note, the 3 outpatient conditions for which IM residents rated themselves higher than did FP residents (diabetes, hypertension, and hyperlipidemia) were conditions also commonly found in inpatient settings. Additional condition and training program characteristics influencing the relationship between exposure and preparedness may include diagnostic and therapeutic complexity, evidence base, frequency of comorbidities, likelihood of complications, resident workload, and quality of resident instruction. In addition, there may be a lower limit (threshold) of necessary exposure. Internal medicine and FP residents may all have sufficient exposure to several conditions, which may explain the lack of association between exposure and preparedness for several of the inpatient and outpatient conditions.

Sex of the resident was the most significant factor for care of patients with vaginitis. This is consistent with a previous study of IM and FP physicians that found that both types of physicians rated their skill and comfort with sex-specific examinations higher with patients of the same sex. Because female patients often prefer female physicians,^28- 31 it is possible that female residents encountered more cases of vaginitis in their training.

Although we found that IM and FP residents reported differences in preparedness, there are several limitations to our conclusions. There may be systematic biases in the way IM and FP residents respond to questions about preparedness. Individuals with a particular approach to self-assessment may be drawn more to one specialty than the other, or residents within each specialty may become acculturated to rate themselves a certain way. Although we separated adult and pediatric diagnoses in our survey, FP residents may have factored their feelings of preparedness with pediatric care into their responses for preparedness on the adult conditions where these overlap. Additionally, IM and FP practices may not be directly comparable because FP adult outpatient populations have been found, on average, to be younger and to have fewer chronic conditions.³² The generalizability of our results may be limited by our sample, which targeted residents at academic health centers.

In summary, in a national survey of residents, we found differences in IM and FP resident reports of preparedness to diagnose and treat common inpatient and outpatient conditions. This study found IM and FP reports of preparedness to be generally consistent with the differing emphases of these distinct specialty training programs. These differences raise an important policy question. If different training regimens result in differing levels of preparedness, can training programs be designed to optimize preparedness for the anticipated practice setting? To improve residency education and the care physicians provide for adult patients, we must continue to investigate the factors promoting better resident preparedness as well as the best means to measure them.