Determinants of patient and health care services delays for tuberculosis diagnosis in Italy: a cross-sectional observational study

The delay in diagnosis of tuberculosis (TB) is one of the main factors contributing to the spread of Mycobacterium tuberculosis and preventing its elimination. Several studies reported a remarkable delay in TB diagnosis, often longer than 2 months. Both patients and healthcare-services can contribute to TB diagnostic delay, the former deferring their presentation to care and the latter missing the opportunity of a well-timed diagnosis. While most of the studies on this topic have been conducted in settings with limited resources and high TB incidence [1‐11], substantial delays in TB diagnosis have been described also in high-resource settings [12‐17]. A meta-analysis of the studies conducted between 1990 and 2008 showed that the mean diagnostic delay, including the proportion attributable to healthcare system, is not significantly different in constrained-resource and high-resource countries [18]. Determinants of TB diagnostic delay vary according to different studies and geographical contexts. Some studies reported that gender and age are associated with delayed TB diagnosis, with women and the elderly experiencing the highest delays [1, 4, 8, 11, 16, 17]. Social factors, such as unemployment, low income and low-grade education, have also been shown to influence time to TB diagnosis, particularly in low- and mid-resource settings [1, 7, 8, 11]. In addition, some studies from high-resource countries with low TB prevalence reported that TB diagnosis may require longer time in autochthonous patients than in immigrants [13, 15, 17]. In particular, in a study conducted in Norway, the delay attributable to healthcare services was significantly longer in patients born in Norway compared to patients born abroad. Of note, 86% of the 83 patients affected by TB included in the study were born abroad [13]. Similarly, an Italian survey from 2003 reported that a long delay attributable to healthcare services was significantly more frequent in Italians than in migrants while a long delay attributable to patients was significantly more frequent in recent migrants compared to Italians [15]. Also in UK, autochthonous patients had been demonstrated to have longer healthcare services delays than those born outside the UK [17].

Extra-pulmonary TB is often and universally reported as an important risk factor for diagnostic delay, because of its often non-specific signs and symptoms [14, 17]. This is true not only for high-incidence countries but also for low-incidence ones. Indeed, extra-pulmonary TB was associated with longer health care delay in Denmark, Norway and the UK [14, 15, 17].

With 4032 new cases diagnosed during the year 2016, incidence of TB in Italy has been estimated around 6.1 (5.3–7.1) per 100,000 population [19] with up to 50% of cases affecting migrants [20].

Country-specific assessment of TB diagnostic delay and identification of the determinants of such delay is essential to improve the access to care of patients with TB. Therefore, we performed a cross-sectional multicentre observational study, aimed at assessing patient- and healthcare-related delay in TB diagnosis in Italy and the risk factors associated with it.

Among 163 patients evaluated for the study, 137 were enrolled. Reasons of exclusion from enrolment were: inability to sign the informed consent (8 patients), language barrier (n = 5), patient declined to participate to the study (n = 8), loss to follow up or death after the first visit (n = 4 and 1, respectively). Six patients returned blank questionnaires, thus information on symptoms, healthcare seeking pathway, knowledge and awareness of TB was available for 131/137 patients.

Herein we report a cross-sectional observational study assessing the diagnostic delay in the diagnosis of tuberculosis in seven TB referral centres in Italy. A considerable latency between symptom onset, clinical presentation and TB diagnosis was observed. In most of the patients, the diagnostic delay was longer than 2 months and about a quarter of the population was diagnosed after > 5 months since the onset of the symptoms. Our results are consistent with those from previous studies [5, 8, 12‐18] and confirm that unacceptable diagnostic delays still exist, in high-resource as well as in resource-constrained settings. In our study, the diagnostic delay attributable to healthcare services was about two-times longer than that attributable to patients (in median, 31 versus 14 days, respectively). Although this finding is similar to a previous regional survey conducted in Italy [15], several other studies conducted in high-resource settings showed patients and healthcare service to be equally responsible for the overall diagnostic delay of TB [12‐14, 16, 22, 23]. Possible explanations for this apparent discrepancy might depend on the proportion of elderly and foreign-born patients in the general population and on differences between healthcare systems, in particular in their arrangements for migrants, translating into variable access to care.

We found that native and non-native patients had similar time of presentation to care. Nonetheless, when we distinguished patients according to the time of their first arrival in Italy, those who lived in Italy for < 5 years had considerable higher risk of late presentation to care than long-term and permanent residents. This suggests that recently immigrated subjects experience significant difficulties in accessing to Health Services, possibly because of bureaucratic, legal and economical barriers, which were, indeed, reported by 12–18% of the patients. By converse, foreign-born patients who had settled in Italy for ≥5 years behaved similarly to native Italians. Given the fact that recent immigrants are also those with the highest risk of TB reactivation/reinfection, specific screening program, targeted information campaigns and social inclusion programs could be beneficial for these patients. Symptoms underestimation was the most common reason leading patients not to seek early medical care and, in multivariate model, it was associated with a > 3-fold increased risk of high patient-related delay. In this respect, promoting an adequate knowledge about TB risk and its clinical presentation is advisable.

We did not find any significant association between other socioeconomic determinants (such as level of education, employment status, annual income or family condition) and delayed presentation to care or health care-related delay. Despite the fact that the association between social disadvantage and TB has been well documented, this finding is not completely surprising because no single measure of socioeconomic status can capture all the aspects of social vulnerabilities and health inequalities that determine delayed diagnosis.

Healthcare related delay was, in our study, the main contributor to total TB diagnostic delay. Consistently with other studies [13, 14, 16, 17], extra-pulmonary TB involvement and lack of respiratory symptoms were associated with longer diagnostic delays. Unspecific symptoms, unusual clinical presentations and difficulties in obtaining specimens for the diagnosis are all likely explanation for this finding and reflect the challenges faced by clinicians during TB diagnosis. It is unrealistic to believe that, particularly in low-incidence countries, all TB can get to a prompt and rapid diagnosis and, from an epidemiological point of view, delays in extra-pulmonary TB diagnosis can look less important than delaying the diagnosis of other communicable forms of TB. Nonetheless, extra-pulmonary TB is often clinically severe and life-threatening and diagnostic delays can be associated with disease progression and a bad prognosis. Before a definite diagnosis is determined, a certain index of suspicion for TB disease should be maintained in patients with compatible symptoms and/or epidemiological risk factors, even when other diagnoses are deemed more likely.

Health-seeking pathways had high influence on diagnostic delay, with a major role played by the first health care provider consulted by patients. Patients initially assessed by the general practitioner experienced longer healthcare-related delay and needed a higher number of contacts with the healthcare system before diagnosis, than those who had direct access to the hospital or consulted a specialist. Various reasons may explain this difference. First, Italy is a low-incidence country and TB rates have steadily reduced in the last 50 years. In such setting, the ability of general practitioner to recognize and diagnose a disease that they rarely see is likely to have reduced over time. As a matter of fact, TB diagnostic delay in healthcare structures has been previously demonstrated to be inversely proportional to the number of cases annually diagnosed by the health-care structures themselves [24]. It is not surprising, therefore, that specialist physicians play an important role in case detection, nowadays. Nonetheless, more than 40% of the patients, in our study, referred to the general practitioner, in the first instance. Informative campaigns and continuous education programs on TB for general practitioners are therefore highly warranted and likely to have an impact on diagnostic delay. Another possible explanation for the health-seeking pathway impact on healthcare-related delay is that symptoms of patients presenting to the emergency department could be more severe or specific than those of patients who contacted their general practitioner, thus facilitating a more prompt diagnosis. Lastly, patients presenting to the emergency department or specialist consultation could also be those more likely to be perceived at risk for TB, because undocumented migrants and other marginalized patients could have easy and free access to hospital care and to TB facilities, but not to primary care.

Results from the present survey add new and important information about TB diagnostic delay in a low-incidence country. Some limitations of the study should be, however, acknowledged. In particular, the small sample size (about 10% of the TB cases notified every year in Italy), the exclusion from enrolment of patients due to language barriers and inability to sign informed consent and the heterogeneous geographical distribution of the TB referral centres included in the study may limit its generalizability. Nonetheless, it should be noted that the preponderance of centres located in large urban areas of Northern Italy included in our study reflects the higher burden of TB in these areas. Moreover, two important reference centres, one in Central and one in Southern Italy, were also involved. Another limitation of the study is the lack of previous validation of the questionnaire on knowledge and beliefs of the patients, which was indeed derived and adapted from a questionnaire previously validated in a different context [21], but not further validated on our population. On the other hand, despite the previously discussed recall bias, the use of specifically designed questionnaires is an invaluable strength of this research, because it allowed to collect precise data on patients’ socio-economic conditions, path of access to the healthcare services and patients’ beliefs and opinions. Finally, consensus about what duration between onset of TB symptoms and diagnosis is considered as delay is lacking in the literature and to this extent quantiles were used in our analysis. Nonetheless, the principle “the later, the worse” is certainly reasonable at least for highly contagious forms of TB.

In conclusion, our study showed that TB diagnostic delay is common. Recently immigrated patients are those most at risk for a delayed presentation to care, whereas those with extra-pulmonary disease have lower chances of being diagnosed early after presentation for care. In addition, patient evaluated, in the first instance, by primary care physicians are more likely to delay the diagnosis and to need repeat consultations before TB is suspected. Since delaying TB diagnosis can have severe implications for patients health as well as for TB transmission, informative campaigns for patients and continuous education programs for physicians are highly warranted.