Smoking status, symptom significance and healthcare seeking with lung cancer symptoms in the Danish general population

We explored the associations between smoking status and perceived symptom significance among individuals with possible lung cancer symptoms in the general population and evaluated the influence of symptom significance on healthcare-seeking behaviour among individuals with different smoking status. Changes in a familiar cough and shortness of breath were the symptoms with the highest significance, as nearly one in ten individuals reported extreme symptom concern about these symptoms and considerable impact on their daily activities. For all symptoms, a substantial number of individuals reported no concern or influence at all. Overall, individuals who currently smoked, were more likely to report higher concern about both specific and non-specific lung cancer symptoms. The same tendencies were found for influence on daily activities, though mainly for shortness of breath, tiredness, and loss of appetite.

In general, reporting combinations of symptoms seemed to increase the likelihood of higher symptom significance. Yet, it appeared that reporting tiredness in combination with the specific lung cancer symptoms tended to decrease the odds of being influenced by tiredness. The likelihood of reporting GP contact with both specific and non-specific lung cancer symptoms increased with increasing symptom significance. Smoking status did not clearly modify this association, though the likelihood of healthcare seeking increased to a higher extent among individuals who never smoked.

A strength of this study is the large sample of individuals from the general population. The response rate of 34% is lower than in previous studies exploring symptoms and healthcare seeking in the general population^13,30, but similar to another population-based study conducted during the same period³¹. The response rate was sought enhanced by two reminders, opportunity of contacting the project group and by drawing lots among respondents for a gift certificate. By investigating symptom significance in the general population and not among individuals with a specific diagnosis, we are able to enhance the understanding of how the general population comprehends and manages potential lung cancer symptoms, without the answers being biased by knowledge of a malignant diagnosis.

The data used for this study was collected in 2022. Similar data were collected in 2012 and analysed (Supplementary materials, Tables S7–S12). They showed identical tendencies, even though e.g., the Corona Virus 19 pandemic could have changed the perception of respiratory symptoms in the general population.

A risk of selection bias cannot be eliminated even though the sample was randomly selected. Individuals experiencing several symptoms or who are extremely concerned or influenced by their symptoms might be more prone to answer a questionnaire about symptoms and healthcare seeking, possibly resulting in an overestimation of symptom prevalence, healthcare-seeking behaviour, or symptom significance. On the other hand, this group may not have the surplus energy to answer a comprehensive questionnaire, causing an underestimation. Furthermore, individuals who are concerned about their symptoms or overall health may avoid responding to a questionnaire about symptoms to prevent increasing their anxiety. On the other hand, they may be motivated to participate by the opportunity to contribute their experiences to research, addressing issues that affect their daily lives. The study design attempted to minimise selection bias by offering the opportunity of contacting the project group in case of challenges³², however, the respondents are likely more technology literate than the non-respondents and individuals with no digital mail were excluded, which may induce some bias. Inclusion of individuals without digital mail was explored in the field test prior to the final questionnaire distribution, by inviting this group by postal letter. Very few of the postal invitees responded (<5%), but we received several calls from invitees and their relatives apologising for not being able to respond due to severe cognitive issues or diseases and asking us not to contact them again. Hence, we considered it ethically sound to omit the postal letter to avoid causing unnecessary disturbance of this vulnerable group³³. The questionnaire was only available in Danish, excluding individuals who could not read in Danish. Although a higher percentage of the respondents were women, and the respondents were slightly older than the non-respondents, the respondents were found to be fairly representative of groups of general Danish population³³.

Development of the DaSC II questionnaire followed COSMIN guidelines. The questionnaire was pilot tested among both academics and representatives of the general population, and afterwards field-tested in the target population to ensure content validity, comprehensibility, and avoid misunderstandings²⁷.

The participants were asked to recall symptoms experienced within the preceding four weeks and indicate symptom significance and healthcare-seeking regarding each symptom. This time span was considered reasonable for adequate recall of symptoms and symptom-related factors. Self-reported smoking status is often underreported³⁴ due to e.g., stigma, blame or embarrassment^17,23. A misclassification of individuals currently smoking as having never smoked could underestimate the differences between e.g., individuals who currently and never smoked. Risk of misclassification was minimised by the survey being online and by emphasising that participation was anonymous.

We chose an ordered logistic regression model, as it is possible to include all categories for the outcome variable in an ordered manner, under the assumption that the ‘distance’ between the categories is equal across the range, i.e., fulfilling the proportional odds assumption. Some analyses, e.g., the stratified logistic regression models, were limited by power issues due to e.g., low number of individuals who formerly smoked and reported high symptom significance. Further, some of the differences implied in the results were small with overlapping confidence intervals. The regression models were adjusted for possible confounders, yet other variables such as type of occupation could have been important as well. For instance, some individuals may have jobs that require physical activity, while others may have jobs that heighten the risk of respiratory symptoms possibly influence the perception of these symptoms. This information was not available in this study. Other factors as alcohol consumption, body mass index and physical activity could act as cofounders as well but have previously showed ambiguous or low impact on healthcare seeking with lung cancer symptom²⁴. Both are correlated with smoking status, yet in the context of perception of respiratory symptom we consider smoking history as essential even though some residual confounding cannot be eliminated.

The percentages of individuals reporting extreme symptom concern for both the specific and non-specific lung cancer symptoms in the present study were similar or higher than the findings in studies exploring symptom concern for gynaecological cancer symptoms³⁵ and for studies exploring both symptom significance in general and for urological symptoms among men^36,37. Both lung-, and gynaecological cancers may present with specific as well as non-specific symptoms, which are also common in the general population. This may explain why the perceived significance of these symptoms are comparable to the significance of more benign and common symptoms as headaches, backpain and urological symptoms³⁶. This emphasises the challenge for both citizens and healthcare providers with interpretation of symptoms and balancing that the symptoms on one hand may be signs of cancer, and on the other may be benign and for some even self-limiting^6,7,8.

In the present study, individuals who currently smoked, reported higher symptom significance than individuals who never smoked. Hence, it seems like the a priori hypothesis, that individuals who currently smoke normalise their symptoms may not be correct regarding symptom concern and influence. Yet, far from all with high symptom significance had contacted their GP. McLachlan et al. describe that individuals may try to explain their symptoms as caused by alternative benign conditions, and therefore do not act on them even though perceiving the symptoms as concerning³⁸. Lung cancer screening has proven to contribute to early-stage diagnosis of lung cancer^4,5. However, not all individuals are included in the risk group invited to screening, and far from all invitees will participate in screening³⁹, hence timely diagnostics based on symptoms remains essential. This is particularly the case among individuals in high risk of developing lung cancer but low likelihood of perceiving their symptoms as significant, and probably also less likely to participate in the screening programms^39,40 Hence, awareness of non-participants in screening programmes who may not interpret and act upon their symptoms are important for both GPs and in planning of screening programs. Furthermore, the influence of a screening without signs of disease may lead to a lower concern about symptom in the following period. Diagnostic imaging does not guarantee a risk-free period⁴¹, therefore, it should be emphasised to screening participants that acting upon symptoms is recommended, regardless of the screening result.

In line with Molassiotis et al.⁴², the present study implied that the experience of more than one lung cancer symptom seemed to increase the symptom significance. Molassiotis et al. also proposed that coughing plays a significant role in the interpretation and explanation of other symptoms such as shortness of breath and tiredness by suggesting that coughing may be used to normalise or explain other symptoms, lowering the perceived significance. In our study this was mainly seen for the influence of tiredness, which seemed to be lowered when reported together with prolonged coughing or shortness of breath.

A notable and gratifying result in the present study is that the combination of prolonged coughing and changes in a familiar cough increases the significance among individuals who currently smoke. Individuals with chronic disease may attribute their symptoms to existing disease, and consequently, postpone care seeking⁴³. Therefore, it is of crucial for individuals with chronic obstructive pulmonary disease or other respiratory conditions to act when a familiar symptom as cough changes to ensure timely diagnosis lung cancer.

Like demonstrated in prior studies^35,36,37, a rise in the significance of symptoms is associated with an increased likelihood of healthcare seeking when it comes to symptoms related to lung cancer. Nevertheless, both heightened symptom concern and influence had a lesser impact on the probability of contacting the GP for individuals who currently smoked, as opposed to those who never smoked, yet with overlapping confidence intervals. This suggests that, even within the population at the highest risk of developing lung cancer, symptoms perceived as significant do not invariably result in healthcare seeking. This aligns with the findings of a qualitative study, which suggests that individuals who smoke are less aware, for instance of, the benefits of timely lung cancer diagnosis⁴⁴. Other possible explanations include that individuals who currently smoke encounter more barriers to healthcare seeking^16,17,45, and that their symptom appraisal may be influenced by pre-existing symptoms associated with smoking and existing chronic diseases^46,47. Kaushal et al. showed that potential lung cancer symptoms are rarely identified as cancer symptoms and that the anticipated healthcare-seeking behaviour was not altered by co-morbidity⁴⁶. Including chronic disease as potential confounder in the logistic regression models in the present study did not alter the associations with neither symptom significance nor GP contact. Additionally, fear of stigma^23,48 and the concern about potential findings can serve a barrier to healthcare seeking^16,17.

Symptom appraisal and the decision to seek healthcare is a multifaceted process influenced by various factors. Several theories address consultation behaviour and symptom appraisal⁴⁹. In the Health Belief Model health behaviour decisions such as seeking care are argued to be triggered by the ‘cue to act’ such as of symptom experiences. Symptoms are interpreted and evaluated in an interplay of perceived severity, susceptibility, benefits and barriers. All of these depend on the individual’s self-efficacy defined as the confidence and ability to comprehend and act^50,51. Symptom significance is embedded in more of the abovementioned factors, for instance, the perceived severity and susceptibility may depend on the impact of the symptoms on daily activity and perceived risk of illness, which in the case of respiratory symptoms is likely to be related to the individuals’ smoking history. This is supported by the findings by Black et al. who analysed healthcare-seeking behaviour among patients with lung cancer based on the common sense of illness self-regulation model. They also found that the emotional response to symptoms depend on the severity and perceived risk^52,53. As such, we argue that symptom significance and smoking status could act as both barriers and facilitators of care seeking.

Self-efficacy, among others, may be expressed as the individual’s health literacy^51,54. A recent study showed that health literacy challenges are associated with lower healthcare seeking with symptoms of lung cancer regardless of smoking status⁵⁵. Nevertheless, current smoking has also been associated with lower levels of health literacy, which may influence the perception of the symptom significance as implied by Samoil et al.⁵⁶. This remains to be investigated in population-based studies. Further, social relations, societal trends, and previous experiences with the healthcare system are also theorised to influence symptom appraisal and care-seeking behaviour, as emphasised by both Hay and Leventhal et al.^53,57,58.

The symptoms examined in the present study may serve as indicators of lung cancer. The absolute risk of lung cancer when presenting with the symptoms, is though very low^59,60, and the symptoms may more likely be indicators of benign conditions. Thus, it seems reasonable that not everyone is concerned or influenced by the symptoms.

Since a history of tobacco use is the most potent risk factor for developing lung cancer, knowledge about how smoking status influences perceived symptom significance and modulates healthcare-seeking behaviour is important for future healthcare planning, education and public awareness campaigns. GPs, and other healthcare professionals may play a crucial role in disseminating information on this topic to both the general population and specific patients. The increased symptom significance among individuals who currently smoke, as found in the present study, emphasise that this group may perceive their symptoms as severe, and that they may be aware of their susceptibility, but not necessarily prompt healthcare seeking. The GPs could play an important role in articulating the benefits of acting on symptoms, especially when concerned or bothered, for instance in planned consultations regarding existing disease. A challenge lies in raising the awareness and vigilance of lung cancer symptoms among individuals at risk while avoiding alarming the general population unnecessary.

The perception of symptom significance could be considered in future implementation of lung cancer screening, where a screening test without signs of disease may diminish the level of concern for symptoms, and potentially hamper the decision of seeking care. Thus, participants, who comprise of high-risk individuals, should be informed that acting on symptoms is important regardless of the results of recent screening.

Overall, individuals who currently smoked were more likely to perceive their lung cancer symptoms as significant, and individuals who reported high symptom significance were more likely to seek healthcare with both specific and non-specific lung cancer symptoms. The significance of symptoms appeared to have a less pronounced influence on healthcare seeking among individuals with a history of current smoking. This suggests that these specific groups may derive greater benefits from receiving support and encouragement to seek medical attention when experiencing concerns or discomfort related to symptoms that could be indicative of lung cancer. Consequently, such interventions could potentially reduce the likelihood of patient delay and improve the chance of timely diagnosis for lung cancer.