Recently stress during medical training is increasingly being reported in published literature. Previous studies have shown fairly high levels of distress, such as symptoms of depression [1, 2] and even suicide thoughts [3, 4] among medical undergraduates. The potential negative effects of emotional distress on medical students include impairment of functioning in class-room performance and clinical practice, stress-induced disorders and deteriorating performance [5, 6]. Perceived medical stress has also been linked to current mental distress [7] and to forthcoming health problems [8]. However, there is very little information about the effect of stress on academic performance during medical training [9].

In addition to stress the students' social, emotional and physical as well as family problems may influence their learning ability and academic performance [10, 11]. Higher levels of stress may have a negative impact on the students learning ability. Excessive stress may result in mental and physical problems and may diminish a student's sense of worth and might affect his/her academic achievement [12, 13]. Earlier studies have classified stressors into three main categories: academic pressures, social issues and financial problems [14]. Studies from developing countries like Thailand, India, Malaysia and even Pakistan have reported stress among medical students and have also underscored the role of academics as a potential stressor [15–18]. But these studies either did not use the perceived stress scale or did not study the correlation between stress and academic performance [19]. Studies have reported that academic constraints, [20] and also factors like age, gender, ethnicity and marital status may also influence students' severity of stress and hence academic performance [21, 22].

In our study, we evaluated perceived stress among medical students including its sources and severity along with its correlation with academic performance, which may be of importance to both medical teachers and psychologists. To our knowledge such a detailed study has not been reported from Pakistani medical schools. In our study, medical students reported a higher level of perceived stress, which was significantly higher among female students, but it's correlation with self-reported academic performance was not significant. Academic and psychosocial stressors were reportedly more common both in terms of frequency of occurrence and severity. By logistic regression analyses stressed cases were associated with female gender, psychosocial and academic related stressors.

Life of a medical student or a health care professional can be very stressful. Mild, moderate, and high levels of stress and even burnout have been reported amongst medical students and health care professionals from other countries [25–29]. Previous studies have shown that medical school stress is a good predictor of nervous symptoms even when psychosocial variables such as marital status or cohabitation status, confident and other general self-esteem are taken into consideration [6]. Apart from all this, end-of-the-year first year medical students have shown to be worse off psychosocially than they were when they entered [30].

The amount and severity of stress experienced by medical students may vary according the settings of the medical school, the curricula, evaluation (examination) system etc. Previous studies from medical schools in different countries have reported varying levels of stress [1, 5, 6, 9, 13, 15, 16, 18]. A study from Agha Khan University, Pakistan has reported that more than 90% of its students experienced stressed at one time or the other during their course [18]. A similar study from India reported that 73% of the students had perceived stress at some point or the other during their medical schooling [15]. Saipanish reported that 61.4% of students in a Thai Medical School had come across some degree of stress as calculated by the Thai Stress Test [16]. These studies have used different instruments to measure stress. This limits the comparability among these studies. We chose the perceived stress scale since this instrument has been documented for its reliability and validity [19, 31–42]. The advantage of PSS is that it can be applied to a wide range of settings, to different subject types and includes items measuring reactions to stressful situations as well as measures of stress [19, 31–42]. An important limitation of other reviewed stress scales for health professions students is that it focuses only on academic stressors, and lack of inclusion of personal issues or reactions to stressful situations (psychosocial issues), and poor applicability to broader settings.

In our study, sample proportion (67.08%) of the female students was higher than their male (32.92%) counterparts. Mean PSS scores among female students was significantly higher than that of male students. A plausible reason may be the conservative nature of society in Pakistan, where women lack freedom to participate in extra-curricular activities owing to the restrictions imposed on them by the society. However, Cohen has reported that there was no significant difference in stress using PSS between male and female students. In addition, private life may conflict with professional life and this may cause stress. We lacked sufficient information, which could assist us in carrying out further analysis about this. The overall average stress score in our sample was significantly higher (p < 0.05) than that reported by Cohen [31].

In our study the correlation between perceived stress and academic performance was negative and not statistically significant. This finding is contrary to existing literature which states that acute stress is a predictor of reduced scholastic performance, especially in examinations [9, 43]. Plausible explanations for such contradictory reports could be that only acute stress but not longstanding stress may affect academic grades during examinations. In addition, the students who are striving to perform well in examinations may make themselves stressed. Individual coping styles and skills along with their access to different forms of social support may have played a role in negating the effect of stress on academic performance; this could not be explored in our study.

Most students had experienced either academic and/or psychosocial stressors. Among academic stressors, 'tests/exams' were the chief sources of stress. Despite this, test/exams are important in the medical training as a standard for evaluation/assessment. Also examinations encourage students' learning and also provide feedback to the teachers. Those students who perceive tests/exams as a burden may experience stressful situations while for others, who consider exams useful, they may assist in their learning. Previous studies have also reported that academics/exams are common sources of stress among medical students [9, 15–19, 44–47]. There may be a need to revisit the evaluation/examination system to make it less stressful to the students. An important source of stress was related to psychosocial factors. This may be due to time constraints for self, family, friends and entertainment due to the demanding medical curriculum. Another reason could be inadequate recreational facilities provided by the college as reported by the students. The students, who were stressed, reported that psychosocial and academic related stressor groups had occurred more frequently. This was revealed by the association of these factors with stressed cases by logistic regression analysis. This may suggest that students had a global response to a wide range of potential stressors rather than being limited to a few specific items. However, there was a high odds ratio on both univariate and multivariate analysis suggesting high collinearity. This suggests that students having academic stressors were more likely to have psychosocial stressors as well or vice versa. A study from USA has recommended that teaching stress management and self-care skills to medical students may prove to be beneficial [48]. There is a need to look at the applicability of such measures which are feasible in our medical school setting.

Lack of generalisation of our results to other medical schools in Pakistan is an important limitation of this study. Since the information was collected on self-administered questionnaires/instruments we cannot rule out information bias. The use of self-reported academic scores is another limitation and future studies should use their academic records. Cross-sectional design of our study is yet another limitation since associations presented lack temporality. Prospective studies are necessary to study the associations between occurrence stressors and incidence of stress. Though response rates (80%) were fairly good there could be some selection (non-response) bias. Though participation was anonymous, authors had the opportunity to meet the non-respondents and ask them reasons for not participating in the survey. The main reasons that surfaced were the sensitive and personal nature of the study as well as the length of questionnaire. The other reasons were absence from college due to sickness or travel on the day of the survey.

The students reported higher levels of stress. The most frequently occurring stressors among the students were related to academic and psychosocial domains. The associations between stressed cases and female gender, occurrence of academic and psychosocial stressors needs to be further tested by prospective studies.