The results of this study demonstrate that despite the high amount of unreported exacerbations, most COPD patients do recognize periods of symptom deteriorations and are willing to take certain self-management actions. To our knowledge, this is the first prospective study investigating patient's decisions and self-management behaviour during symptomatic days and exacerbation episodes.
It is well established that COPD exacerbations have clear negative impact on disease progression, morbidity, mortality, and HRQoL [3, 5, 20]. According to current guidelines, exacerbations should be treated with increased bronchodilator therapy, glucocorticosteroids and/or antibiotics . Patient recognition of exacerbation of symptoms and prompt intervention reduces the risk of hospitalisation and is associated with a better quality of life [11, 21]. However, in line with our findings different cohort studies have revealed that only a minority of exacerbations in fact are reported and subsequently treated [8–10]. Despite the fact that reported exacerbations have on average the worst outcomes, unreported exacerbations showed in a Canadian and a Chinese cohort to have important and non-negligible impact on annual change in health status [9, 10] Until now, only a few studies examined potential determinants of reporting exacerbations. Delay or failure to report exacerbations seems to be associated with a combination of disease, event and patient characteristics. Our study also confirmed that patients with more severe airflow limitation and higher hospitalisation rates in the previous year are more likely to report an exacerbation [9, 22]. In a Canadian cohort, a higher number of symptoms at onset and exacerbations with increased cough and sputum quantity were associated with reporting . In addition, other studies show that psychological characteristics may also be considered as potential predictors of seeking medical treatment [10, 23, 24].
In contrast to the lack of seeking medical attention, patients did seem to recognize consistent warning signs stimulating them to anticipate on these episodes. Rather than contacting a healthcare provider, the majority (70.7%) of patients showed a willingness to increase their attention on planning periods of rest, breathing techniques and/or sputum expectoration. A similar high proportion of patients increased their bronchodilation medication (62.7%). These findings are in line with two multinational interview-based studies investigating patients perceptions of retrospectively self-reported exacerbations [13, 19]. These studies also showed that the majority of COPD patients are able to recognize and respond to periods of symptom deterioration. Only one study reported on exacerbation-related actions focused on planning periods of rest (30% rest; 10% lie/sit down;, 10% stay calm/prepare) . Furthermore both studies found that 33% of the patients anticipated by increasing medication [13, 19]. It needs to be emphasized that comparisons partly hampers both because of differences in operational definition of an exacerbation and the methods used to assess these self-management actions.
In the current study, both type-A and type-B actions were conducted timely, mainly during prodrome and at exacerbation onset. The latter is consistent with findings of a previous study reporting that half of the exacerbations were related to self-initiated use of rescue medication, mainly during the week before reporting the exacerbation . Obviously, patients cannot predict whether prodromal days with symptom increase turn into actual exacerbations. This was reflected by examining actions during isolated symptomatic days that were not part of an exacerbation. Around 40% of these days were also associated with type-A actions and 25% with type-B actions. This suggests that a substantial amount of patients seem to respond immediately to symptom changes, even if this only concerns increase in minor symptoms.
In addition to reporting of exacerbations, we related patient characteristics for timely taken type-A and type-B action measures during exacerbations. Interestingly for both types of action, current smoking had an increased likelihood of been associated with less appropriate self-management actions. Similar defaults in self-management behaviour by current smokers were seen from previous studies in using appropriately meter dose inhaler technique , being adherent to long term nebulizer therapy  and attending outpatient pulmonary rehabilitation [27, 28]. Furthermore, correct type-A actions were more frequently taken in patients with longer exacerbations episodes. Finally, the type of symptoms showed to influence patients' self-management decisions during exacerbations. Patients perceiving increased dyspnea at exacerbation onset are more likely to take type-B actions that patients without increased dyspnea. Patients with increased sputum purulence are less likely than patients without increased sputum purulence to take action. We believe this study provides valuable new insights in how patient respond to symptom deterioriation. This knowledge can support the development of improved patient information and material to enhance appropriate self-management of exacerbations. However, a few words of caution are needed when interpreting the results. First, although this study was the first to examine self-management decisions prospectively, it comprised a relatively small group of 108 consecutive patients followed up for only 6 weeks resulting in 75 exacerbations. This is equal to an annual rate of 6.0 per patient-year This relatively high event-rate can be explained by the fact that all patients were simultaneously followed-up in the same 6-week winter period in which exacerbations have shown to be ~ 50% more likely than in other seasons [6, 29]. Also the relative high proportion of patients included immediately after hospitalisation might have contributed to a higher exacerbation rate.
Secondly, the relatively short period of follow-up only allowed for the evaluation of a single exacerbation per patient. Data from a well-known UK cohort indicates that exacerbations, although not validated in terms of aetiology, are not random events but clustered in time . The fact that we did not evaluate multiple exacerbation episodes per patient, could have resulted in biased and incomplete judgment of relapsed or recurrent exacerbations and subsequently the presence or absence of correct self-management behavior. A longer follow-up would have created the opportunity to judge within-patient variations over time, taking into account clustering of exacerbations. A stable run-in period of at least 4 weeks could have eliminated recurrent exacerbations .
Thirdly, another limiting factor is that after dichotomizing episodes by actions carried out, the number of events per predictor variable did not reach the rule of thumb to allow for multivariate logistic regression analysis . Therefore, the present study solely evaluated the association between single characteristics and actions taken. Adequately powered observational studies allowing multivariate analysis including adjustment for potential confounders are needed to validate current indicators of self-management behaviour.
This study was not designed and powered to examine the effects of different self-management decisions taken by patients on exacerbation related outcome (recovery time, severity etc). Nevertheless, indirectly from the literature, it can be expected that early and self-initiated response on these episodes including all three type of actions most likely affect outcome. Self-management measures during periods of symptom deterioration like changing activity, relaxation, and breathing pattern alteration, a proxy for type-A actions, have shown to be effective in faster symptom relief [32, 33]. Furthermore, there is sufficient evidence that prompt anticipation by increasing short-acting bronchodilators (type-B actions) is effective in reducing symptoms and improve airflow obstruction during exacerbations [1, 34]. Our study indicates that a substantial amount of exacerbations remain partly or completely unmanaged. As recommended, these results stress the importance of developing and investigating individualized therapies specifically aiming at early recognition and consequent actions to these episodes by patients . Within these interventions, our data suggests that specific attention concerning exacerbation-related health behavior should go out to current smokers.