The principle aim of the present study was to determine the 1-year incidence and persistence of neck pain among undergraduate students. The 1-year incidence of neck pain among undergraduate students in this study was high, at 46% among whom 33% reported persistent neck pain. Grimby-Ekman et al , on the other hand, reported that the annual incidence of neck or upper back pain in Swedish undergraduate students was 15%. The discrepancy between our and previous studies may be due to difference in frequency of data collection during follow-up and the definition of symptomatic case. Grimby-Ekman et al  followed their subjects on a yearly basis, whereas in this study subjects were followed every 3 months for ≤1 year. Data collection every 3 months would reduce the influence of recall bias. Also, in the previous study a symptomatic case was defined as an individual who experienced pain for > 7 days whereas in this study it was > 24 hours. Consequently, it is likely that a far greater number of subjects were identified as symptomatic cases in this study.
Neck pain is regarded as a chronic episodic condition characterized by persistent, recurrent, or fluctuating pain and disability . Earlier studies showed that persistent musculoskeletal symptoms were common among young population [[8, 14, 39]]. Slightly more than one quarter of our study sample who reported new onset of neck pain experienced persistent neck pain. In this study, persistent neck pain was defined as that reported for ≥ 2 consecutive follow-ups. There is a lack of consensus over the operational definition of persistent musculoskeletal pain. Stanton et al  defined recurrence as persistence of pain reported both at baseline and follow-up assessment with no recovery, whereas Hill et al  proposed that persistent pain could reflect chronic, recurrent, or continuous pain. In the present study, we did not ask subjects whether they had experienced any recovery period of pain since they previously reported symptoms. Thus it is not possible to differentiate those who had recurrent pain from those who had continuous pain. However, within this limitation, our findings suggest that students with neck pain may become symptomatic adults, highlighting an urgent need for stakeholders to pay more attention to the problem of neck pain in the young population to reduce the impact of neck pain later in life.
A secondary aim of this study was to identify risk factors for onset and persistence of neck pain. Computer use-related factors contributed significantly to these. However, risk factors for onset differed from those for persistence of neck pain, as was found in previous studies [8, 42]. Persistent pain can have broad and profound effects on well-being with significant impairment of physical and psychological health . Thus information about risk factors for persistent neck pain is of considerable importance.
Onset of neck pain was predicted by the computer screen position not being level with the eyes. High computer screen height results in the neck being more erect ; simultaneous increase in muscle activity of the neck extensor and sternocleidomastoid muscle was reported in this posture, and therefore prolonged computer use in this posture may be harmful . The effect of low computer screen is inconclusive. Studies showed that low computer screen height increases neck flexion, neck extensor activities and compressive loading of neck ligaments, joint capsules, and other structures of the cervical spine, thus possibly increasing musculoskeletal strain in the upper body [[46–49]]. However, Fostervold et al  showed that working with low monitor screen height improved oculomotor status with significant reductions in musculoskeletal symptoms in the upper body. In the present study respondents were asked only whether the top of the computer screen was positioned at a level horizontal with the eyes when they sat and looked straight ahead. Further study is required to evaluate the relation between high/low computer screen height and neck pain in undergraduate students.
A mouse position self-rated as too low decreased the risk of onset of neck pain. This finding is contrary to the common concept of "good computer posture" often described as a position in which the upper arm is perpendicular to the floor, the elbow at a right angle, and forearm parallel to the floor . However, Marcus et al  showed that elbow angles between 137° and 148° while using the mouse were correlated to lower risk of developing neck and shoulder pain in newly hired computer workers. Although the exact elbow angle while using a mouse was not measured in the present study, it is plausible that mouse position self-reported as too low may correlate with elbow angles > 90°. This hypothesis warrants further investigation.
Factors that predicted persistence of neck pain were students being in the second year of their studies, use of computer for entertainment < 70% of total computer usage time, and position of the keyboard being too high. The risk of persistent neck pain was 1.9-fold higher for second year students in comparison with first year students. Ndetan et al  found that chiropractic students are predominantly exposed to injury risk factors during the first, third, and sixth academic trimesters. However, the finding may not necessarily reflect other student groups. To our knowledge, this study is the first to demonstrate that year of undergraduate study significantly correlated with neck pain. This information may be of importance for developing viable prevention strategies of neck pain in young population. Further study is needed to focus on identifying year of study factors so as better to understand how year of study interacts with persistent neck pain.
A higher percent time of computer use for entertainment reduced the risk of persistent neck symptoms. Computer use for entertainment included activities such as chatting, playing games, listening to music, and watching movies. Hakala et al  suggested that the basic mechanism of computer game playing mostly required repetitive hand motion in sitting position and dynamic action where players change postures freely, thereby minimizing loading of the upper extremities. Thus computer use for entertainment may not require the user to be in static postures for prolonged periods. It is also plausible that computer use for entertainment reduces mental stress . Bongers et al  suggested that psychosocial demands can exceed an individual's coping capabilities, resulting in a stress response that in turn could produce muscle tension, static loading of muscles, or generate other physiological responses that may result in musculoskeletal symptoms.
A keyboard position self-rated as too high increased the risk of persistent neck pain. Faucett et al  found that higher keyboard height (with respect to elbow height) was associated with increased risk of neck, upper back, and upper extremity discomfort. Mekhora et al  found that neck and shoulder discomfort significantly declined when keyboard level was adjusted to suit the individual's comfort and proposed that adjustment of keyboard level would reduce the need to reach the hand forward and backward or to elevate the shoulder. Therefore, less muscle activity would be present after the adjustment.
Interestingly, no psychosocial and clinical domains remained in the final model. Although it is currently accepted that a large index number within psychosocial dimensions contributes to onset and persistence of musculoskeletal symptoms in adults , evidence for such effect in the young population is far from conclusive. Grimby-Ekman et al  found that perceived stress was a risk factor for developing neck or upper back pain and for persistent neck or upper back pain in Swedish undergraduate students. On the other hand, Hanvold et al  reported no association between stress level and incidence of neck pain during 3-year follow-up among technical school students. In the present study we only examined a selected group of psychosocial and clinical factors. Other important psychosocial and clinical factors may be identified in future work.
The major strength of this study is its prospective design and the evaluation of broad biopsychosocial factors for their contribution to neck pain. However, the study has several weak points. First, the nature of several biopsychosocial factors and the diagnosis of neck pain were subjective, which may have led to inaccuracy. Future studies should consider inclusion of objective information to increase accuracy. Second, information regarding frequency and severity of pain was not collected in this study. Further study should gather this information to enhance understanding regarding relations between risk factors and musculoskeletal symptoms in undergraduate students. Third, subjects in this study were recruited only from one university. Thus generalization of the results to other undergraduate student populations may be limited.