Discussion

In this work, we used a cohort of 164 344 individuals corresponding to the residential population of the Vercelli LHA located in the North-West of Italy. We found a prevalence of chronic diseases equal to 21.43% considering the subjects of all ages who appear in at least two clinic databases, except for osteoporosis. The number of subjects with at least one pathology increases with age, and a similar trend was observed among men and women.

Nineteen macro-categories were selected that represent the main chronic diseases diagnosed in Italy. In particular, the selection of pathologies was carried out in line with the previous literature. Goodman et al presented a selection of 20 chronic conditions identified according to the definition of chronicity, prevalence and potential interest for public health in terms of impact and possible prevention.10 This study differs in the choice of excluding hyperlipidaemia as we consider it a risk factor rather than a chronic condition. Furthermore, our work takes into consideration only rheumatoid arthritis, differentiating arthritis from arthrosis because in Italy, there are no exemptions, hospital discharge and specific drugs for this condition, unlike in America. In the literature, there are other studies comparable with ours, such as the ones conducted by Pefoyo et al and Moin et al, who have identified roughly the same pathologies.9 12

We decided to not include hypertension in the selected pathologies because we considered it more like a risk condition than a disease. Moreover, if this condition is properly treated, it does not have a substantial impact on life expectancy and quality of life of those affected. In fact, individuals who take indicated therapy and maintain correct blood pressure values have the same life expectancy as the population who is not affected by hypertension.

We have used three different databases to limit the chances that some subjects could escape our search, as choosing a single database would not have given sufficient guarantee to our results. For example, in the database of exemptions, not all subjects with chronic diseases actually request the exemption, either because they can be unaware of this procedure, or because the exemption itself does not lead to considerable advantages in economic terms or access to visits.

Moreover, the values may be underestimated as many individuals may have an income or disability exemption and therefore do not use the disease exemption. However, we decided not to include these two types of exemptions in the main analyses because we would have otherwise had an excessive overestimation due to the fact that not all of these exemptions are granted for chronic diseases. The hospital discharge database is reliable because any patient who has ever had hospitalisation also has a discharge card related to it. However, using data from the hospital discharge database has few downsides. First of all, the fast-paced work in the hospital often leads to inaccurate input by the medical staff, which leads to forgetfulness in reporting the chronic pathologies from which hospitalised patients are affected. Moreover, not all patients suffering from chronic diseases are hospitalised, as very often they can be safely managed from home; therefore, unless an acute event takes place that brings these patients to hospital, they are not identified. The drug database, on the other hand, may have the opposite problem: it is impossible to exactly match a drug to a disease, so we will find many patients who take drugs indicated for chronic diseases without actually having one; this involves a significant overestimation in the results obtained. Conversely, there are many types of drugs for which it is possible that some have not been included in this review, leading to the exclusion of patients who take them. It should also be considered that despite the existence of guidelines, every physician decides independently which drugs to administer based on individual evaluations; this means that among different doctors and different patients, the therapies assigned are very different, making it very difficult to assign the pathologies to different subjects.

A few focus on some pathologies can be done. Osteoporosis was identified only through the use of drugs. This is because there is no exemption for osteoporosis according to the Italian Ministry of Health. Furthermore, there are not even subjects identified by hospital discharge although the relative code exists. We suspect this may be because clinicians tend to enter only the codes that are most relevant to the reason for admission. From this, it follows that by using the intersection of two databases, no patient suffering from osteoporosis will be present, as can be seen in the analyses described above. Another point of interest is the strong discrepancy between subjects with kidney disease identified between the ATC codes compared with the other two databases, which could be explained by the fact that the drugs used for chronic renal failure are very often the same as those used for cardiovascular diseases. Therefore, the data are overestimated compared with the real one identified by overlapping with the other databases, which are more specific for kidney diseases.

Results obtained in terms of prevalence among age groups and gender are coherent with expectancy. Particularly, when we performed a graphical representation of prevalence of chronic diseases in the study population, we observed that the graph assumes a curvilinear trend with a peak of prevalence at the age of about 80 years; growth from birth up to 50 years is constant and slow, then after 50 years, the increase becomes more marked up to 80 years where the curve begins to decrease. We can suggest that this fact occurs due to a selection related to several factors: both genetic and behavioural. We suppose that a patient suffering from one or more chronic conditions has a lower life expectancy than those who are not affected. This fact leads to the number of subjects with chronic diseases who are older than 80 years to be a minority; this is also indirectly found by DuGoff et al who note that as the chronic diseases of an individual increase, his life expectancy decreases.27 Moreover, when we focus our attention on people suffering from multiple pathologies, we observed that after the age of 80 years, more than 13% of subjects have at least two chronic diseases. This should make us reflect because not only is there a high prevalence of sick people in the elderly, but also of complex multipathological clinical pictures that have a major economic and clinical impact on the health system.

There are other national and international studies that have done research like ours.

Passi, for example, is a national surveillance system that continuously investigates aspects relating to the state of health of the Italian population through interviews and questionnaires, presenting data in the 3-year period.28 In Italy, according to the Passi survey (period 2017–2020), 18.4% (Piedmont 19.2%) of subjects between 18 and 69 years of age claim to have received a diagnosis of one or more of the following pathologies: renal failure, chronic bronchitis, emphysema, respiratory failure, bronchial asthma, stroke or cerebral ischaemia, diabetes, myocardial infarction, cardiac ischaemia or coronary artery disease, other heart diseases, tumours, liver disease or cirrhosis. When we restricted our analysis to this age group, we observed a prevalence of at least one chronic disease of 16.05% (95% CI 15.83%, 16.27%). A similar reasoning can be applied by comparing the data from the Passi d’Argento Surveillance (population over 65, period 2017–2020). Based on Passi d’Argento Surveillance, the prevalence of Italian subjects with at least one chronic disease is 60.8%: 49.5% in Piedmont and 47.08% (95% CI 46.62%, 47.54%) in Vercelli.29 We noted that our values are coherent with those proposed by Passi and Passi D’Argento. We assumed that little discrepancies between results could be related to different methodological approaches used: the data of the national surveillance must be self-reported by the population through telephone interviews and different pathologies included.

International comparisons were also done. Pefoyo et al, based on the data provided by the National Insurance System, calculated the prevalence of 16 chronic diseases in the population of Ontario in Canada (12 242 273): arthritis (excluding rheumatoid arthritis), hypertension, asthma, depression, diabetes, cancer, chronic coronary syndrome, cardiac arrhythmia, osteoporosis, COPD, congestive heart failure, renal failure, dementia, rheumatoid arthritis, stroke and acute myocardial infarction. They find that 43% of the population examined had at least one chronic disease, a prevalence significantly higher than the 21.64% estimated by our study. However, when we analysed the data defining chronic disease, a subject that was identified by at least one database (situation more similar to that presented in Pefoyo et al work), the prevalence was very similar as we obtained 45.28%. Interestingly, we observed that a similar age trend was observed: the prevalence increased as the age increased, but a little decrease was observed for subjects aged 90+ years. Moreover, the proportion of subjects with multimorbidity conditions is higher in older age. Finally, differences observed in estimates could be linked to different reasons including lifestyles, environmental conditions and different geographical areas.9 Another important study was conducted by Barnett et al using a national dataset held by the Primary Care Clinical Informatics Unit at the University of Aberdeen, UK, which contained data corresponding to one-third of the Scottish population. They extracted data on 40 morbidities and identified subjects with one or more chronic conditions. The prevalence was found to be 42.2%, and 23.2% were multimorbid.30 This is a higher value than that identified in our study probably due to a less restricted definition of chronic disease (in our analysis, we considered only 19 diseases).

In summary, our study presents certain limitations. First of all, the selection of the chronic diseases included was conducted without specific literature references; in fact, every cited author who conducted similar studies has chosen to include different pathologies. This fact happened because there is not a clear definition of ‘chronic pathology’, so each author gives his own definition. Another limitation may be the choice to require at least two databases (except osteoporosis). We decided to use this strategy because the number of subjects who had at least one chronic disease using this approach was roughly the same as the one in the Passi survey. Moreover, we decided to use one database only for osteoporosis, which is the unique pathology considered that does not have an exemption code; in fact, we can detect this disease only by using the ATC codes. However, we cannot be totally sure that this is the best choice. In addition, as we have explained before, even the three databases considered have some issues: for example, the database of exemptions underestimates the number of individuals with chronic conditions because not all of them make the exemption request. Considering the hospital discharge records, they have two important problems: first, they are often compiled approximately, and second, not all the individuals with chronic conditions are hospitalised. Finally, even the drug database has several problems, for example, many patients who take drugs indicated with a chronic condition do not really have it; in addition, every physician decides independently which type of drugs to prescribe to a single patient. Therefore, this leads to the impossibility to make completely accurate evaluations with this database.

Finally, considering the generalisability of our results, we can affirm that our conclusions could be extended to other Italian regions; in fact, we assume that the prevalence of chronic disease that we found in Vercelli’s area is quite similar to the prevalence of chronic disease in the entire Italian population; in fact, many conditions such as age, risk factors and healthcare system are roughly the same in all the Italian regions. Therefore, we can say that our results are generalisable to Italy. Conversely, it is more difficult to apply our results to the European or world population; in fact, the demographic, clinical and risk factor characteristics are totally different from the Italian ones.