[Skip to Navigation]
Sign In
June 2001

Impact of Problem-Based Learning on Residents' Self-directed Learning

Author Affiliations

From the Department of Pediatrics, Albert Einstein College of Medicine and the Children's Hospital at Montefiore, Bronx, NY.

Arch Pediatr Adolesc Med. 2001;155(6):669-672. doi:10.1001/archpedi.155.6.669

Objective  To examine the effect of a problem-based learning (PBL) curriculum on self-directed learning behaviors among a group of pediatric residents.

Methods  A controlled comparison study was conducted with 80 pediatric residents at a large urban academic medical center. Residents were observed over 3 distinct but consecutive periods. First, all residents participated in a 3-month-long daily lecture series (pre-exposure phase). Then, for another 3 months, 39 residents (PBL group) were exposed to twice-weekly PBL sessions while 41 residents continued with the daily lectures (lecture-based group) and served as controls. Problem-based learning was withdrawn after 3 months and all residents returned to the lecture series (follow-up phase). Residents' self-directed learning behaviors were assessed through self-administered questionnaires during the pre-exposure, exposure, and follow-up phases.

Results  There were no significant preexposure differences in self-directed learning behaviors between the groups. During the exposure phase, the PBL group had significantly higher self-directed learning: 5 or more hours of independent study (2% vs 7%) (P= .001); 5 or more hours of medical discussions (28% vs 4%) (P= .008); 2 or more computer literature searches (51% vs 30%) (P= .005); and total hours of self-study per week (6 vs 4 hours) (P<.05). At the 3-month follow-up, the PBL group had returned to baseline levels of self-directed learning and there were no significant differences between the groups.

Conclusion  Residents exposed to PBL engaged in significantly higher levels of self-directed learning than their counterparts.

PROBLEM-BASED learning (PBL) has emerged as a prevalent teaching technique in medical schools, especially during the preclinical years.1 Problem-based learning seeks to increase motivation for self-directed learning by presenting a relevant problem to a group of learners. The process of solving the problem requires that members of the group engage in independent reading and research. Proponents of PBL maintain that during such a process, the innate desire of a group of learners to solve a problem can be used as motivation for self-directed learning. To date, very little has been published to support this theory in the residency setting.

One of the goals of postgraduate medical education is to prepare physicians to be lifelong learners, self-directing their own learning after residency. Despite this, the use of PBL in postgraduate education has been sporadic and limited.2-5 The effect of PBL on self-directed learning among residents has not been studied. Thus, this study was designed to test the hypothesis that the introduction of a PBL curriculum would result in enhanced self-directed learning behaviors among a group of pediatric residents.

Materials and methods

Program development

As a first step, one of us (J.C.) was trained in the PBL teaching method by Howard Barrows, MD, at Southern Illinois University, Springfield. Subsequently, an 8-hour-long faculty development course was developed at our institution. All faculty members in our department were invited to volunteer to participate in the course. All volunteers were accepted. One hundred five faculty members took part in the course over a 4-year period. These trained facilitators served as the pool of faculty for the residency PBL experience.

Using the established core lecture series as a template, we designed a PBL curriculum by choosing topics that lent themselves to the PBL format. Examples included "The Diagnostic Approach to the Child in Shock" and "The Child With Fever and Rash." We searched our admission records and found actual cases that illustrated these topic areas. For the selected cases, the medical records were reviewed and summarized in a sequential written format that was designed for the PBL exercise, beginning with the chief complaint and followed by the history of present illness, medical history, family/social history, review of systems, and the physical examination, in that order.

Pbl sessions

We employed the 2-part closed-loop reiterative model of PBL that has been identified by Barrows6 as the best model for promoting self-directed learning. In this model, learners are presented with a clinical problem with no prior preparation. A faculty facilitator provides no factual information, but helps the group probe their current knowledge and stay on the task of solving the patient's problem.

The first PBL meeting occurred at the beginning of the week and began with the presentation of the chief complaint of the patient. This was followed by a discussion of the differential diagnoses. Next, a summary of the rest of the history was read. Then there was a group discussion aimed at generating a list of problems, defining learning issues, and ranking the differential diagnoses. After this, the findings on physical examination were then made known; this helped the group in refining the differential diagnosis. Finally, there was the self-assignment of the learning issues that had been generated during the session. These learning issues involved further information that the group felt it needed to have to solve the patient's problem. After a break of several days to allow for research and reading, the second meeting began with reports on the self-assigned learning issues. There was further discussion of the differential diagnosis based on the new information. The group generated an action plan of laboratory tests and patient treatment. Laboratory reports for the planned tests were then revealed and a final discussion ensued. After the group committed to a specific diagnosis, the patient's actual diagnosis was made known. The sessions ended with a review of the case and a discussion of any errors that the group might have made along the way.

During the study period, residents completed 4 PBL cases per month.


The study was based in a large tertiary academic medical center with 85 pediatric residents. Eighty pediatric residents on active rotations were enrolled in the study. We performed a time-series cohort analysis of an intervention group with a comparison group. A self-administered questionnaire was developed to measure the self-directed learning behaviors of each resident according to our operational definition described below. The study was designed to include 3 distinct periods of observation each of 3 months' duration: pre-exposure, exposure, and follow-up.

During the pre-exposure period (July-September 1998), all 80 residents participated in a lecture series. Baseline self-directed learning behaviors were assessed in September 1998. During the exposure phase (October-December 1998), 39 residents (PBL group) were exposed to twice-weekly PBL sessions. The PBL group was made up of residents on inpatient rotations at the central residency site. For this group, PBL was substituted for noon lectures and occurred twice weekly. This group had weekly self-assigned learning issues derived from the PBL meetings and includes residents with a minimum of 1 and a maximum of 3 months of PBL. At the end of each monthly rotation, self-directed learning was measured using the same questionnaire. Forty-one residents on other concurrent rotations served as a comparison group. These residents continued to receive daily noon didactic lectures and were designated as the lecture-based learning group. The lecture-based learning group consisted of residents on ambulatory rotations at the central site plus residents on rotations at another training facility affiliated with the residency program. The lecture-based learning group had no formal self-assigned research topics. Self-directed learning behaviors were assessed at the end of each month.

The PBL intervention was withdrawn in December 1998. During the follow-up period (January-March 1999), all residents returned to the daily lecture series. Self-directed learning behaviors were reassessed in March 1999.

In addition, we assessed residents' satisfaction and opinions about PBL and the lecture series using a 5-point Likert scale. Complete data were available on all 80 subjects.


Self-directed learning was operationally defined as the average time per week that a resident spent in (1) independent study of medical text and journals; (2) medical discussions with colleagues, outside of scheduled didactics (including PBL) and teaching rounds; and (3) performing computer literature searches. The period of recall assessed was the preceding month. For example, each resident was asked to answer the following question: "On average over the past 1 month, how many hours per week did you spend in performing computer literature searches?"

Statistical methods

Data were maintained in SPSS 9.0 statistical software (SPSS Inc, Chicago, Ill). Comparisons of percentages between PBL and lecture-based learning groups were done using χ2 or Fisher exact test as appropriate for dichotomous variables. Cut points for statistical analysis were determined prospectively based on pilot data on residents' self-directed learning while field-testing the questionnaire.


Compared with the lecture group, a significantly higher percentage of residents in the PBL group studied for 5 or more hours per week (26% vs 7%) (P = .001); engaged in 5 or more hours of medical discussions per week (28% vs 4%) (P = .008); and performed 2 or more computer literature searches per week during the exposure phase (51% vs 30%) (P = .005). At follow-up, self-directed learning in the PBL group had returned to baseline and there were no significant differences between the groups (Table 1).

Table 1. 
Comparison of Self-Directed Learning Behaviors Between the Groups*
Comparison of Self-Directed Learning Behaviors Between the Groups*

Table 2 shows that for the PBL group during exposure, there was a significant increase in the mean total weekly self-directed learning hours (6 vs 4 hours) (P<.05).

Table 2. 
Comparison of Mean Total Self-directed Learning Hours per Week*
Comparison of Mean Total Self-directed Learning Hours per Week*

On a 5-point satisfaction scale (with 5 as the best score), residents gave the PBL series a rating of 4.6, while the lecture series was rated as 3.3.


This is the first report to demonstrate that participation in PBL resulted in significant increases in self-directed learning behaviors in the postgraduate setting. We found that a group of residents who were exposed to PBL showed significant increases in the amount of time devoted to independent study, the amount of time spent in medical discussions, and the number of computer literature searches that were done. However, our findings also suggest that the gains in self-directed learning behaviors might be short lived. It is unclear whether residents who have prolonged or ongoing exposure to PBL would sustain the improvement in self-directed learning behaviors. In addition, it would be of interest to study whether the enhanced self-directed learning behaviors would resurface if another mechanism for self-assignment of research topics were instituted in the ambulatory setting.

The percentage of residents from both groups performing computer literature searches dropped during the follow-up period. We are not sure why this happened. A speculation might be that the fatigue and increased clinical demands associated with the winter months were contributing factors. We did not ask residents to identify the location from which they performed their literature searches. However, in addition to the medical center's library resources, we have computer terminals with Internet access located on the wards and in the ambulatory setting.

Outcome variables that have been used to assess PBL in the past have included student satisfaction, student performance on standardized tests or problem-solving paradigms, student motivation to learn as demonstrated by topics covered, and self-directed learning as demonstrated by use of library and other materials.4,5,7-10 In the residency setting, Schwartz et al4 showed a high level of satisfaction among surgical residents engaged in PBL. In another study, attendance at PBL was highly correlated with performance on the American Board of Surgery In-Service In Training Exam.5 Itani et al5 speculated that this finding might have been the result of enhanced self-study on the part of residents engaged in PBL, but they did not establish this link. Our findings strongly support the association between PBL and enhanced self-directed study.

A limitation of our study is that we relied on self-reports by residents and could not independently verify the behaviors that were being reported. However, we believe that this limitation is partially controlled for by the fact that we administered the same questionnaire to the same group of residents at several intervals both preintervention and postintervention. Moreover, the residents were unaware of the study hypothesis. We found that residents were very willing to report low levels of self-directed learning behaviors. As an example, approximately 30% of residents in both groups consistently reported studying for less than 1 hour per week. The fact that residents were quite willing to report such low levels of self-directed learning activity gives us some confidence in the honesty of the self-reports. We believe that this level of reporting was encouraged by the fact that the questionnaires were anonymous.

This was not a randomized study, but rather a convenience sample of residents based on rotation assignments. However, we were unaware of any bias in the way that residents were assigned to various rotations. We do not believe that the changes observed were a result of rotation assignments because during the follow-up period, the group rotation assignments were essentially reversed. Despite this fact, we found no significant differences in self-directed learning behaviors during this period.

Overall, we found very low levels of baseline self-directed learning behaviors among residents. This lends some validity to the general consensus among residency educators about the need for finding ways to enhance self-directed learning. However, the demonstrated effectiveness of PBL in improving self-directed learning among residents in this report challenges the presumption that the enormous time demands of residency may preclude the attainment of this goal. This study demonstrates that it is in fact possible to increase residents' self-directed learning in spite of heavy clinical demands. Based on our findings, further use and study of the PBL technique in residency is warranted.

Accepted for publication February 5, 2001.

Presented in part at the annual meetings of the Pediatric Academic Society, San Francisco, Calif, May 3, 1999, and Boston, Mass, May 16, 2000.

Corresponding author: Philip O. Ozuah, MD, MSEd, Montefiore Medical Center, 3544 Jerome Ave, Bronx, NY 10467 (e-mail: pozuah@pol.net).

Barrows  HSTamblyn  RM Problem-Based Learning: An Approach to Medical Education.  New York, NY Springer Publishing1980;
Foley  RPPolson  ALVance  JM Review of the literature on PBL in the clinical setting.  Teach Learn Med. 1997;94- 9Google ScholarCrossref
Schwartz  RWDonnelly  MBMayo  WPStrodel  WE Problem-based learning: a formal curriculum for postgraduate surgical education.  Curr Surg. 1993;50285- 290Google Scholar
Schwartz  RWDonnelly  MBSloan  DAStrodel  WE Residents' evaluation of a problem-based learning curriculum in a general surgery residency program.  Am J Surg. 1997;173338- 341Google ScholarCrossref
Itani  KMMiller  CCChurch  HMMcCollum  CH Impact of a problem-based learning conference on surgery residents in training exam (ABSITE) scores: American Board of Surgery in Training Exam.  J Surg Res. 1997;7066- 68Google ScholarCrossref
Barrows  HS A taxonomy of problem-based learning methods.  Med Educ. 1986;20481- 486Google ScholarCrossref
Thomas  RE Problem-based learning: measurable outcomes.  Med Educ. 1997;31320- 329Google ScholarCrossref
Albanese  MAMitchell  S Problem-based learning: a review of literature on its outcome and implementation issues.  Acad Med. 1993;6852- 81Google ScholarCrossref
Vernon  DTBlakc  RL Does problem-based learning work? a meta-analysis of evaluative research.  Acad Med. 1993;68550- 563Google ScholarCrossref
Saunders  KNorthup  DEMennin  SP The library in a problem-based curriculum. Kaufman  Aed. Implementing Problem-Based Medical Education: Lessons From Successful Innovations New York, NY Springer Publishing1985;Google Scholar