Funding This study was supported by the Richland Memorial Hospital Research and Education Foundation.
Competing interests None
Provenance and peer review Not commissioned; externally peer reviewed.
Statistics from Altmetric.com
There is a paucity of research that addresses the physical activity patterns of attending physicians, resident and fellow physicians and medical students. However, there is some evidence that the level of physical activity of physicians may be correlated directly with physician counselling pattern about this behaviour.1 One study has indicated that general practitioners in the action or maintenance stage of changing their own behaviour as it relates to physical activity are three times more likely to encourage the same behaviour in their patients.2 Also, it has been shown that a primary-care physician can improve fitness and exercise confidence in patients using tailored exercise prescriptions, even in the older population.3 While there may be differences in the efficacy of different types of physical activity counselling with regard to gender, it is clear that physical activity counselling in the primary-care setting is beneficial.4
One of the most comprehensive studies that evaluated issues related to physician health and prevention is the Women Physicians' Health Study, which was a questionnaire-based study of a representative sample of 4501 US women physicians.5 The authors concluded that primary-care physicians who practiced a health behaviour were significantly more likely to report counselling and screening patients for healthful behaviours. A national population-based supplemental survey (year 2000) to the 1995 National Health Survey that included 17 317 respondents concluded that the rate of physician counselling about exercise is low.6 It noted that physicians are more likely to counsel for secondary prevention and are less likely to counsel patients at risk for obesity. Finally, it concluded that the failure to counsel younger, disease-free adults and those from lower socioeconomic groups may represent important missed opportunities for primary prevention.
The evaluation of resident physicians as exercise role models was done in a small cross-sectional study that used a self-administered survey, treadmill fitness testing and a 7-day physical activity recall to study 51 resident physicians.7 It concluded that internal medicine resident physicians may not be adequate role models for exercise adherence. Confidence in the knowledge of current guidelines, personal physical activity enjoyment and perceived success and self-efficacy in engaging in regular physical activity may be useful targets for enhancing resident physician physical activity counselling for their patients. If resident physicians are not exhibiting traits that will lead to promoting healthy behaviours in patients, as they become attending physicians, we may continue to see the rate of physical activity decrease throughout the country.
The predictors of US medical students' prevention counselling practices have been studied in more detail than those of the resident and fellow physicians.8 Frank et al surveyed 2316 medical students from 16 medical schools in the USA during their first year orientation, upon entrance to the wards, and in their senior year. Several of the variables shown to predict physician counselling also predict US medical students' reporting counselling (especially personal health practices and speciality type). In addition, the avidity with which medical schools encourage students to practice healthful behaviours significantly influences their reported patient counselling practices. These findings give an evidence-based direction to help create physicians who counsel patients about prevention.
The purpose of our study was to determine the physical activity level of attending physicians, resident and fellow physicians and medical students throughout the USA. We conducted a cross-sectional web-based survey to evaluate the physical activity level. Our main objective was to obtain data from a nationally representative sample to determine whether their physical activity meets the current recommendations of the Unites States Department of Health and Human Services (USDHHS). The current USDHHS guidelines indicate that for substantial health benefits, including lower risk for premature death, coronary heart disease, stroke, hypertension, type 2 diabetes and depression, adults should participate weekly in at least 150 min of moderate-intensity aerobic activity, 75 min of vigorous-intensity aerobic activity or a combination of intensities, with 1 min of vigorous intensity equalling 2 min of moderate intensity. Further, for increased and additional health benefits, including lower risk for colon and breast cancer and prevention of unhealthy weight gain, they should participate weekly in more than 300 min of moderate-intensity activity, 150 min of vigorous-intensity activity or an equivalent combination.9 10
Participants and data collection
We established an online exercise survey with the goal of surveying a total of 2100 individuals (700 attending physicians, 700 resident and fellow physicians and 700 medical students). Individuals in the aforementioned groups completed this 3–4-min survey at http://www.physicianhealthpractices.com. Initially, we advertised the survey through the American Medical Association online newsletter (AMA eVoice) with an aim to gather an adequate sample from each of the individual groups being studied. We also distributed fliers at AMA Annual Meeting held in June 2009. The initial yield of completed surveys with the aforementioned efforts was about 450.
In order to increase the sample size, we decided to use other efforts to gather survey responses. Using the American College of Graduate Medical Education (ACGME) website (http://www.acgme.org/adspublic/), we sent individual emails to each designated institutional official for all residency and fellowship programmes in the USA (see online only example). In addition to the ACGME-directed emails, we also sent individual emails to the leaders of the Aerospace Medical Association, the American Academy of Pediatrics, the American College of Physicians and the American Association of Public Health Physicians. Many individuals that received the survey decided to forward the survey to their colleagues throughout the country. The period of data collection was about 8 months (June 2009–January 2010). As the distribution of the survey was through numerous sources, there was no way to evaluate the number of surveys actually received by medical students, resident and fellow physicians, or attending physicians. There were no incentives for participation in the study, and the study was considered exempt by the institutional review board of Palmetto Health Richland Hospital.
We developed an anonymous cross-sectional survey to obtain self-reported information on physical activity level, body mass index (BMI), age, gender, ethnicity, speciality, practice setting, frequency with which they encourage physical activity to their patients, comfort level with providing physical activity counselling and assessment of whether or not the practitioner felt as if he/she participated in an adequate amount of physical activity. We used the short form of the International Physical Activity Questionnaire (IPAQ) to quantify the level of physical activity of the respondents (see appendices). Each survey was concise such that all participants were able to complete the survey in about a 3-min time frame.
The IPAQ has undergone a rigorous testing in numerous countries, and this research indicates that it is a valid and reliable questionnaire for assessing physical activity and inactivity.11,–,13 We decided to use the short form of the IPAQ to encourage the completion of the survey in its entirety. The short form of the IPAQ, in which one recalls their activity over the last 7 days, has been indicated for national monitoring while the long form is useful for more detailed assessments. The IPAQ has been evaluated in numerous populations and in many disease states such as fibromyalgia and schizophrenia.14 15
In addition to the short form of the IPAQ, the final survey asked for information on age, stage of training (medical student, resident physician, fellow physician or attending physician), gender, ethnic background, speciality of training, confidence with regard to physical activity counselling, belief of adequacy of personal level of physical activity, state (where practitioner practices or attends medical school), setting (urban, rural, suburban), practice type (academic, hospital-based, non-clinical, private practice, retired), average number of work hours per week, weight, height, marital status and the number of children in the household. We obtained the demographical data to characterise whether or not the population responding to the survey fit the profile of medical students, resident and fellow physicians and attending physicians throughout the USA. By obtaining the aforementioned demographics, we were able to address whether or not our sample is a representative sample of medical students, resident and fellow physicians and attending physicians throughout the USA.
Statistical data analysis
Data from Survey Monkey were downloaded into a Microsoft Excel spreadsheet, where survey responses were converted into variables for analysis in the R statistical programme.16 The responses for the number of hours and days per week for vigorous, moderate and walking exercise were used to create a dichotomous variable, indicating whether or not the USDHHS guidelines were met (outcome for this study). Descriptive tables and graphics were developed. Because of sparse data, some levels of categorical predictor variables were collapsed. Thirteen candidate predictor variables were chosen for the logistic regression model-building process. Seventy-seven observations had missing values for one or more predictor variables. Therefore, the model building was carried out on 1872 individuals. The final main effects model contained seven predictors. The importance of two-factor interactions with the major predictor variable (training level) and the remaining predictors was explored.
There were a total of 1949 respondents to the survey representing each of the four levels of training: medical student, resident physician, fellow physician and attending physician. The largest group represented in the survey participants was resident physicians with 897 (46%) respondents from this subset of population surveyed. There were slightly more female respondents 1026 (52.6%) in the survey population. The distribution of the respondents by the number of work hours was fairly symmetrical, perhaps slightly skewed to the right. Most of the survey participants worked between 60 and 69 h with 453 (25.2%) indicating their average number of work hours in this range (table 1). There was a wide ethnic variation represented in the survey respondents, but most of the survey respondents were Caucasian with 1368 (70.2%) respondents from this ethnic group. Survey respondents' age ranged from 20 to 79 years, but the largest age group represented were survey participants aged 20–29 years (37.9%) (table 1).
Individuals from a wide variety of medical subspecialities completed the survey. Internal medicine respondents represented the largest single subspecialty group represented with 338 responses (17.3%), but the other primary-care subspecialities including paediatrics and family medicine also had several respondents with 295 (15.1%) and 220 (11.3%) responses from these groups, respectively. The other subspecialities represented in the survey in which 408 respondents were distributed included the following: aerospace medicine, allergy and immunology, colon and rectal surgery, dermatology, general preventive medicine, medical genetics, neurological surgery, neurology, nuclear medicine, occupational medicine, ophthalmology, otolaryngology, pathology, physical medicine and rehabilitation, plastic surgery, public health, radiation oncology, radiology-diagnostic, thoracic surgery, urology, vascular surgery and many combined subspecialities such as internal medicine/paediatrics.
Survey participants represented 48 states, 1 district, Puerto Rico and Federated States of Micronesia and there were 8 states/1 district with >100 respondents including Michigan, South Carolina, Massachusetts, New York, District of Columbia, Maryland, Texas and California (online supplementars table). Other key demographics that were obtained from the survey were marital status, number of children in the household and self-reported height/weight for BMI calculation. Of the 1949 respondents, 1213 (62.6%) indicated that they were married, 1300 (66.9%) of all respondents had no children and the BMI for most participants fell in normal or overweight ranges with BMIs in the 18.5–24.9 range (normal weight) or 25–29.9 range (overweight). Also, participants responded to two questions to gauge their personal opinions about physical activity. A majority of respondents (72.1%) felt that they did not get enough physical activity. However, most (83.2%) felt confident about counselling patients about physical activity as indicated by their response to the question ‘do you feel confident about counselling patients about physical activity?’
Survey respondents and the US Department of Health and Human Services Guidelines
Our primary objective was to determine the percentage of medical students, resident and fellow physicians and attending physicians who meet the current US Department of Health and Human Services Guidelines (hereafter ‘USDHHS guidelines’). There was a variation in this response across levels of training, number of work hours, gender, ethnicity, age and BMI. Attending physicians (84.8%) and medical students (84%) were more likely than resident (73.2%) and fellow (67.9%) physicians to meet the guidelines for physical activity (table 1). There was an indirect relationship with number of work hours per week and meeting the guidelines with 85% of individuals working less than 40 h a week meeting the guidelines and only 70% of those working >80 h a week meeting the guidelines (table 1). For BMI, 80% of those persons in the normal weight (BMI of 18.5–24.9) and 78% persons in the overweight (BMI of 25–29.9) range were more likely to meet the guidelines than 67% of underweight persons (BMI <18.5) or 68% of obese persons (BMI >30) (table 1).
Logistic regression analysis
Using the attending physicians as the reference group with the highest likelihood of meeting the guidelines for physical activity, the odds of medical students meeting the USDHHS guidelines were 0.81 the same odds of attending physicians meeting the guidelines (95% CI 0.50 to 1.34), but this was not significant at the 0.05 level. The OR for resident (0.56, 95% CI 0.38 to 0.82) and fellow (0.50, 95% CI 0.31 to 0.80) physicians indicated that they were significantly less likely to meet the guidelines (table 2). Men were more likely to meet the USDHHS guidelines with an OR of 1.39 (95% CI 1.10 to 1.76) compared with women who were the reference group. Persons aged 30–39 years were less likely to meet USDHHS guidelines with an OR of 0.69 (95% CI 0.52 to 0.95) compared with the reference group of persons aged 20–29 years, and persons who were 40 years and older were more likely than both groups to meet the guidelines with an OR of 1.07, but these data did not show statistical significance (95% CI 0.69 to 1.66). In the general US population, older adults are less likely to engage in leisure time physical activity which differs from the findings illustrated in our study.17 Non-Caucasian survey respondents were less likely (OR 0.58, 95% CI 0.46 to 0.74) than the reference group of Caucasian-respondent to meet the guidelines. Persons who worked less than 70 h per week were more likely (OR 1.57 ,95% CI 1.22 to 2.02) than those who worked more than 70 h per week to meet USDHHS guidelines. Persons without children were more likely to meet the USDHHS guidelines compared with those with children (OR 0.70, 95% CI 0.54 to 0.91).
Additional logistic models containing two-way interactions with training status were run (data not shown). We found that the race modified the association of training status in meeting the USDHHS guidelines. There was no difference in the OR for attending physicians between the races, but for all other levels of training the odds of meeting USDHHS guidelines were lower in non-Caucasians than for Caucasians, after adjusting for the other factors in the model. Similarly, there was no difference in the OR for attending physicians with or without children, but for all other levels of training the odds of meeting USDHHS guidelines were lower in those with children than those without children, after adjusting for other factors. We also ran a multiple linear regression model to determine the impact of greater number of work hours on BMI. We determined that more work hours corresponded to an increase in BMI, adjusting for age and race. The BMI of an individual who works less than 40 h per week, is 20–29 years old, and is Caucasian is estimated to be 22.3. If, for example, an individual worked 40–49 h/week (instead of <40—same age and race), the predicted BMI would be 23.42.
In this cross-sectional web-based survey, physicians and medical students engage in more physical activity and tend to have a lower BMI than the general population.9 18,–,20 In 2008, 43.5% of the US adults were aerobically active, 28.4% were highly active, 21.9% met the muscle-strengthening guideline and 18.2% met the muscle-strengthening guideline and were aerobically active.9 In 2007–2008, the age-adjusted prevalence of obesity was 33.8% (95% CI 31.6 to 36.0%) overall, 32.2% (95% CI 29.5 to 35.0%) among men and 35.5% (95% CI 33.2 to 37.7%) among women.19 Resident and fellow physicians engage in less physical activity than attending physicians and medical students, and this association was strongly associated with the number of work hours per week. The work hour demands of physicians in training limit their ability to engage in regular physical activity which places them at risk for deleterious health effects associated with inactivity. Also, normal BMI survey participants were most likely to meet USDHHS guidelines regarding physical activity when compared with respondents in other weight categories. The survey respondents reflected similar demographic of medical students, resident and fellow physicians and attending physicians as it relates to gender, ethnic background and subspeciality.21,–,23
Our study had several strengths. First, our survey respondents represented a diverse demographical group in terms of age, gender, ethnicity, subspeciality and the level of training. Second, our study used the well-established IPAQ for assessing the physical activity levels.11 24 Third, to our knowledge, this is the largest study addressing the physical activity levels of resident and fellow physicians.7 Finally, no current studies address the physical activity of our study population as it relates to the current USDHHS guidelines.
There are several weaknesses in our study. First, while we obtained a diverse sample, we were unable to determine the response rate as there was no mechanism to monitor the number of persons who accessed the survey who went on to complete the survey in its entirety. Second as with any self-reported survey, there are likely misclassifications of activity level and BMI. Third, while several states had a very strong representation in the study population, many states had very few numbers. Finally, while we did assess the confidence level of the physical activity of survey respondents, we did not assess familiarity with the current USDHHS recommendations for physical activity.
Physical activity has a wide variety of health benefits. Chaput et al address the importance of physical activity to prevent weight gain and maintain weight loss over time, and they propose that legislative measures should be taken to promote physical activity to reduce the incidence of chronic disease.25 Blair et al demonstrate that while physical activity benefits have been purported since Hippocrates, physical inactivity remains a major public health issue that requires attention at the national, state and local levels.26 While physical activity affects all persons within our society, there is a dire need for development of programmes for the medically underserved population.27 Physical activity reduces chronic disease, and improves the overall quality of life.28
What is already known on this topic?
▶ The level of physical activity of physicians may be correlated directly with physician counselling pattern about this behaviour.
▶ Physician counselling about the importance of physical activity is low.
What this study adds
▶ Physicians and medical students engage in more physical activity and tend to have a lower body mass index than the general population.
▶ Resident and fellow physicians are less likely to meet the US Department of Health and Human Services Guidelines for physical activity than attending physicians and medical students.
▶ An increase in the number of work hours corresponds to an increase in BMI and a decrease in physical activity.
Many studies have evaluated the impact that physicians have on supporting health behaviours such as physical activity, especially in the primary-care environment. A recent study conducted by Grandes et al as a part of the PEPAF (Evaluation of Family Physician's Effectiveness for Physical Activity Promotion) showed that family physicians were effective in increasing physical activity in their patients.29 While most respondents to the survey indicated that they felt competent in prescribing physical activity to patients, a few studies that have evaluated medical students' competence with prescribing physical activity have noted a low confidence level with counselling about physical activity. This suggests that medical students may not be prepared to address this issue.30 31 As medical students are the future healthcare providers for patients, efforts should be made to encourage physical activity in this population as it will likely have a positive impact on the health of their patients.32 33
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Funding This study was supported by the Richland Memorial Hospital Research and Education Foundation.
Competing interests None
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.