Factors Influencing Involvement in Research and Career Choice: A Survey of Graduating Physical Medicine and Rehabilitation Residents

Methods 

Results 

Electronic mail contact was attempted with 79 program directors; 19 directors confirmed forwarding the survey link to residents, and 9 required telephone contact because of failed contact by electronic mail. One hundred sixty (2004, n=84; 2005, n=76) residents representing 24% of their graduating classes (2004, n=331; 2005, n=326) (D. Morris, American Board of Physical Medicine and Rehabilitation, e-mail communication, June 2, 2004) responded. Twenty-five states, the District of Columbia, and Puerto Rico were represented. Data from the 2004 and 2005 surveys were pooled for analysis because there were no statistically significant differences between measured characteristics comparing respondents from the 2 years (table 1).

Table 1. Respondent Characteristics

	Characteristics	2004	2005	Combined
	No. of respondents	84	76	160
	Mean age ± SD (y)	33.4±4.5	33.0±4.7	33.2±4.6
	Sex (% female)	33	32	32
	Marital status (%)
	Married	70	71	70
	Single	27	26	27
	Divorced	3	3	3
	Medical degree (%)
	MD	76	83	79
	DO	23	17	21
	Other	1	0	<1
	Memberships (%)
	AAPM&R	60	70	66
	AAP	1	3	1
	Both	29	28	28
	None	7	0	3
	Other	2	0	1
	No. of residents in graduating class (%)
	≤3	19	29	24
	4−6	42	46	43
	>6	39	25	33

Abbreviations: AAP, Association of Academic Physiatrists; DO, doctor of osteopathy; MD, doctor of medicine.

A review of table 1 reveals that 79% of the respondents had a doctor of medicine degree with 69% and 31% reporting membership in the AAPM&R and the Association of Academic Physiatrists (AAP), respectively. Although not detailed in the table, a limited number also reported membership in other associations, such as the American Association of Electrodiagnostic Medicine (AAEM) (now American Association of Neuromuscular and Electrodiagnostic Medicine) and the American Medical Association.

Of the 160 respondents, 74% reported a residency requirement to conduct research, 85% reported involvement in research during residency, and 85% reported that their residency program provided financial support for presentation travel (table 2).

Table 2. PM&R Resident Research Involvement

	Primary Outcomes	Response	Values
	Research requirement (%)		74
	Involvement in research (%)		85
	Funding for travel (%)		85
	Primary focus of research (%)	Musculoskeletal	38
		Other⁎	35
		Spinal cord injury	12
		Brain injury	5
		Stroke	6
		Pediatrics	4
	No. of presentations at meetings during residency (%)	0	45
		1	27
		2	18
		3 or more	11
	Location of research presentations (%)	AAPM&R	22
		AAP	14
		ASIA	3
		AAEM	2
		Other	17
		None	43
	Practice plan after residency/fellowship (%)	Private practice	63
		Academic	12
		Academic affiliated	18
		Other†	7
	Median (IQR) anticipated percentage of postgraduate practice time spent on (%)	Research activities	5 (0–10)
		Clinical activities	80 (70–85)
		Administrative activities	10 (0–10)
		Teaching	5 (0–10)
	Degree of concern that the current rehabilitation research base is a problem (%)	Very concerned	21
		Somewhat concerned	42
		Neutral	28
		Somewhat unconcerned	3
		Very unconcerned	8
	Perceived barriers to research involvement (%)	Lack of time	47
		Lack of interest	21
		Insufficient research training	26
		Other (including lack of financial and faculty support)	4
		No significant barriers	4

NOTE. Because of rounding, some percentages may not equal 100%. Abbreviations: ASIA, American Spinal Injury Association; IQR, interquartile range.

⁎ Other responses were as follows: electrophysiology, amputee, disability evaluation, sports medicine, epidural injections, deep venous thrombosis, botulinum toxin, rheumatologic rehabilitation, brachial plexus avulsion, health of disabled women, lymphedema treatment, neuromuscular disease, osteoporosis, pain management, and racial disparities. † Other responses were as follows: military hospital, Veterans Affairs hospital, research position, and another residency.

Regional Factors 

There was no evidence of a significant association between region of the country and the presence of residency research requirement, support for presentation travel, or plans to enter an academic or private practice (all P>.10). However, there was a significant association between region and planned percentage of postgraduate practice time for research (P=.03), with pairwise comparisons revealing that respondents from programs in the Northeast and Midwest planned to devote a significantly greater amount of time to research (9.5%) than those from the West, the South, and Puerto Rico (4.6%) (P=.025). There was no apparent regional disparity in resident research submissions to the AAPM&R annual assembly (table 3), the most common society for respondent membership (66%) and presentation (67%). Additionally, there was no association between training program size (≤3, 4–6, >6 residents per year) and presence of a research requirement or support for presentation travel.

Table 3. Resident Submissions to AAPM&R’s 2004 Annual Assembly⁎

	Region	Resident Research Submissions,⁎ n (%)	Proportion of PM&R Residents in Region (%)
	Northeast	102(40)	34
	South	62(25)	31
	Midwest	54(21)	23
	West	35(14)	12

⁎ Submissions were not necessarily accepted; numbers indicative of all resident submissions regardless of postgraduate year.

Discussion 

Nearly two thirds of residents surveyed in this study responded that they were somewhat to very concerned about the current base of evidence in PM&R. It is interesting that although a requirement to perform research during training was associated with an increased likelihood of a resident performing research, the results of this survey did not support the assumption that this participation necessarily leads to an increased likelihood of selecting an academic or research-oriented practice.

It is not surprising that residents in programs that supported travel for presentation reported increased involvement in research. However, despite an association with increased research participation, presentation support did not appear to influence postgraduate research plans.

Although it is hoped that requiring residents to be involved in research may increase the probability of selecting an academic career,5 this study did not support such a relation. This result differs from a report involving a single internal medicine residency program that the presence of a research requirement affected career choice.7 This difference may stem from a difference in traditions of the 2 specialties or from the incorporation of programs both with and without research requirements in the current study. Departmental factors, such as faculty interest and training, availability of adequate mentors, research milieu, financial ability to support resident projects, and dedicated research time, may contribute. Personal factors including degree of interest in research before entering residency; perceived program support for research involvement; and ability to manage PM&R study, clinical, research, and personal time are also probable contributors to resident research experiences, attitudes, and future career choices.

The selection of a career path is undeniably multifactorial and likely contains elements of serendipity. Currently, there are many clinical opportunities for graduating physiatrists. For many, the work hours, compensation, lifestyle, and content of clinical work are attractive. In contrast, research can appear tedious and frustrating, with added burdens and stresses of needing to seek funding to sustain one’s career. A lack of role models and mentors to interest PM&R residents in research and guide them along the challenging path are significant barriers to the selection of a research career.

The lack of a correlation between financial support for scientific presentations and selection of an academic practice also differs from results reported by an emergency medicine residency survey.8 Reasons for this discrepancy may again be multifactorial, but differences in survey design may be important because the emergency medicine study surveyed program directors who may have provided different responses than residents.

Although research requirement and funding for travel were not associated with the selection of an academic practice, it is encouraging that nearly one third of the respondents intend to pursue a practice with an academic component. It is interesting that the residents reported that their greatest research interest involves the musculoskeletal system (see table 2). Such research may be of greatest benefit if directed at reducing disability in the aging population.

Overall, graduating residents are concerned about the state of rehabilitation research but perceive lack of time, their own lack of interest, and deficient training as barriers to their involvement in research. The selection of the AAPM&R annual assembly as the most common venue for resident research presentation is consistent with the pattern of respondent membership. This also may relate to multiple other factors, such as the job fair, resident assembly, Physiatric Association of Spine, Sports and Occupational Rehabilitation courses, or other attractions.

There are several potential limitations of this study. Most significantly was the 24% response rate. Following the recommendation of Baruch,9 we assessed the response rate acceptability by comparing it with the response rates of similar surveys in the years 1995 to 2004. Baruch recommended that a response rate is acceptable if it falls in a range no lower than 1 SD below the mean computed from the population of response rates of similar surveys. The mean and SD of 10 comparable surveys8, 10, 11, 12, 13, 14, 15, 16, 17, 18 were 45.4% and 23.2%, respectively. Thus, the pooled response rate of 24% achieved in this work was considered to be acceptable. Additionally, in comparison with the full population, response-frequency distributions appeared representative of graduating PM&R residents with regard to region, age, sex, and size of residency programs (D. Morris, American Board of Physical Medicine and Rehabilitation, e-mail communication, June 2, 2004). The data collected in this preliminary survey will provide estimates of parameters needed for the power analysis of subsequent surveys.

This rate, although low, was likely representative of graduating PM&R residents with regards to region, age, sex, and size of residency programs (D. Morris, American Board of Physical Medicine and Rehabilitation, email communication, June 2, 2004). Hindsight shows that an earlier survey date (perhaps May) as well as a more developed set of questions dealing with research experience before residency training, mentorship programs, and financial support might have been beneficial.

In interpreting the results of this study, it is important to realize that a primary outcome measure was postgraduate career choice. It is possible that research requirement, training, and involvement may have predicted outcome measures not tested, such as ability to comprehend medical literature, propensity to use clinical evidence, or ability to interpret data in clinical practice. In follow-up studies, we believe that it may be useful to assess some of these other outcome measures.

Interpretation of the implications of our results is dependent on the existence of a relation between graduating residents’ perceptions of their postgraduate practice and reality. Although one would not expect a 1 to 1 correlation between perception and reality, a correlation is likely because if residents in their final month of residency have already selected a practice without capacity for research, it would be more challenging to build this component later. Although this survey was sensitive to those who selected practices in which research would not be a component, it would be insensitive to detecting situations in which a graduating resident selected academic practice and later discontinued this. Thus, the results were more likely to have been biased toward increased research involvement. Additionally, if there were a selection bias in residents who volunteered for this study, it would most likely have been increased inclusion of residents with greater research interest. Despite these biases toward representation of those with increased research interest, the results did not suggest a correlation between residency research requirement and practice selection.

The repetition of all questions used in 2004 in the 2005 survey served to validate the results, and the addition of questions in 2005 enabled clearer interpretation of data. The nearly identical results for demographic, personal, primary, and secondary outcome measures comparing years suggests that the addition of questions did not significantly alter results and may have strengthened interpretation.

Conclusions 

This study suggests there is a relation between a research requirement and financial support for presentations during residency training and the participation of PM&R residents in research. However, neither the existence of a research requirement nor financial support was found to be associated with resident selection of an academically oriented career.