A total of 596,316 ambulatory visits to pediatric clinics from 2000 to 2011 were retrieved from the sampling data. Of these visits, 454,782 (76.3%) involved pediatric patients (age < 18 years) and 141,354 (23.7%) involved adult patients (age ≥ 18 years). The latter group of adult patient visits were included in this study. Stratifying the data by age group revealed that patients aged 18–39 years accounted for the highest proportion (52.2%, n = 73,900) of adult ambulatory visits to pediatric clinics, followed by patients aged 40–64 years (36.2%, n = 51,172); patients aged 65 years and older accounted for the lowest proportion (11.5%, n = 16,282).
The total number of ambulatory visits to pediatric clinics, regardless of age, has increased year by year. However, further analysis revealed that the proportion of children’s ambulatory visits to pediatric clinics decreased steadily year to year (from 84.8% in 2000 to 68.2% in 2011). On the contrary, the proportion of adult ambulatory visits to pediatric clinics, regardless of age, increased steadily year by year (Figure 1
). From the 2000–2003 period to the 2008–2011 period, there were 1.7-fold, 2.8-fold, and 2.7-fold increases in the proportions of adult patients utilizing pediatric ambulatory services in the 18–39, 40–64, and ≥65-year-old age groups, respectively (Table 1
In this study sample, physician clinics were the major ambulatory care providers for adult patients, accommodating 96.8% (n = 136,791) of adult ambulatory visits to pediatric clinics, followed by academic medical centers (1.3%), local community hospitals (1.0%), and metropolitan hospitals (0.9%). We assembled a top 10 list for the most common diagnoses of adult ambulatory care based on an analysis of the first diagnosis code in each medical record and categorized the data by hospital level. The most common diagnosis categories overall were acute inflammation and infection-related diseases. The top three most common diagnoses did not change across the three analyzed time periods. The top three diagnoses across the three periods were upper respiratory infection (45.5%), followed by acute bronchitis (9.3%) and acute or chronic tonsillitis (5.5%).
Differences among the most common diagnoses over time emerged when hospital level categories were considered (Table 2
, Table 3
and Table 4
). Ranking figures of top ten most common diagnoses of adult ambulatory care were assembled (Supplementary Materials Figures S1–S4
). For academic medical centers, epilepsy remained the most common diagnosis over the full study period. Although their relative rankings changed slightly over time, the following three diagnoses remained in the range of second to fifth most common from 2000 to 2011: congenital cardiac and circulatory anomalies; deficiency and other anemia; and jaundice. It is noteworthy that upper respiratory disease, which was the second most common diagnosis (n
= 45) during the period of 2000 to 2003 and the third most common diagnosis (n
= 38) during the period of 2004 to 2007, became the tenth most common diagnosis (n
= 22) during the period of 2008 to 2011. Meanwhile, the frequency of diabetes diagnoses increased over the 12-year study period, rising from the 28th most common diagnosis (n
= 1) in the period of 2000 to 2003, to the tenth most common diagnosis (n
= 19) in the period of 2004 to 2007, and to the fifth most common diagnosis (n
= 33) in the period of 2008 to 2011. Systemic lupus erythematosus remained the sixth (n
= 14, 2000–2003 and n
= 30, 2008–2011) or seventh (n
= 27, 2004–2007) most common diagnosis in the study period (Table 2
, Table 3
and Table 4
) (Figures S1–S4
At metropolitan hospitals and local community hospitals, the most common diagnoses during the study period were related mainly to inflammation, infection, and allergic diseases, especially airway infections. Although epilepsy was the sixth (n
= 11) and seventh (n
= 15) most common diagnosis at metropolitan hospitals in the periods of 2000 to 2003 and 2004 to 2007, respectively, it was no longer within the top 10 diagnoses in the period of 2008 to 2011. Conversely, jaundice was not a top ten diagnosis in 2000 to 2003, but in 2004 to 2012 it became the most common diagnosis at metropolitan hospitals and the second most common diagnosis at local community hospitals (Table 2
, Table 3
and Table 4
) (Figures S1–S4
Overall, airway infections/diseases and gastroenteritis were consistently common diagnoses at adult ambulatory visits to pediatrician clinics during the study period in this survey. Upper airway infections were the most common diagnosis during the study period of 2000 to 2011, followed by acute bronchitis and acute/chronic tonsillitis. Noninfectious gastroenteritis was the fourth (2000–2003) or fifth (2004–2011) most common diagnosis. It is noteworthy that the incidence of essential hypertension among adults seeing pediatricians rised. Essential hypertension represented the tenth most common diagnosis, accounting for 0.9% of ambulatory service care visits during the period of 2000–2003. Then, its proportion rose steadily, reaching 1.6% in the period of 2004–2007 and 2.9% in the period of 2008–2011. Essential hypertension was ranked the eighth most common diagnosis at physician clinics during the period of 2008–2011 (Table 2
, Table 3
and Table 4
The incidence rate of ambulatory visits to pediatric clinics significantly increased from 2000 to 2011. The incidence rates of the top ten most common diagnoses all show steady increase from 2000 to 2011. When allergic reaction was set as reference level, all the top ten most common diagnoses, except essential hypertension, had significantly higher incidence rates than allergic reaction did. Special attention should be paid to essential hypertension, as the incidence rate ratio of essential hypertension over allergic reaction was 0.50 in 2000, and the ratio steadily increased to 0.97 in 2011 (Figure 2
and Table S1
In our study, we found that the proportion of adults among ambulatory patients visiting pediatric clinics rose steadily with statistically significantly differences observed over time. This increase might reflect the falling birth rate in Taiwan. Depopulation is a global trend and garnering substantial public attention in East Asia [9
]. In 2010, the total birth rate in Taiwan fell to 0.89, the lowest birth rate in the world [9
]; since then it has remained among the lowest three. This decreasing birth rate trend has not been reversed despite policy efforts by the Taiwanese government to encourage childbearing, including a childbirth subsidy, maternity pension, children’s education subsidy, and children’s healthcare subsidy [15
]. Due to depopulation, pediatricians in Taiwan are facing unprecedented levels of competition with other physicians. To compensate, pediatricians can elongate their care into adulthood, especially for long-term patients with chronic diseases they have been treating since childhood.
This practice is well justified by challenges that some young adult chronic disease patients face in the transition from pediatric to adult medical care, including the occurrence of undesirable health outcomes, such as a worsening of glycemic control for teenaged patients with diabetes [18
]. Deterioration of health status has also occurred for teens diagnosed with pediatric epilepsy and teens who received cardiovascular surgery or other interventions in childhood [20
]. Whereas some countries have a barrier between pediatric and adult care, which is managed by bureaucrats [24
], in the Taiwan NHI program, people are free to choose their physicians for ambulatory visits without strict regulations of referral requirements [1
]. In other words, as the population of children shrinks, pediatricians have some incentive to expand into adolescent and adult medical care to ensure the long-term quality of care that their patients receive and to maintain a full medical practice. Pediatricians might also explore a transboundary market, such as treating some common adult-onset diseases.
We found that the most common diagnoses made by pediatricians differed among hospital levels. At academic medical centers, chronic illnesses—such as epilepsy, congenital heart disease, anemia, jaundice, systemic lupus erythematosus, and diabetes—were the most common diagnoses for adult ambulatory visits to pediatric clinics. On the one hand, as the population is aging, pediatricians staffing in hospitals would have some incentive to provide medical services to young adults and/or patients with chronic conditions who are accustomed to the pediatric medical delivery system since childhood. On the other hand, improvements in medical treatments and surgical techniques have enabled patients with congenital diseases to survive longer [20
]. For example, most congenital heart disease mortality has been delayed into adulthood [2
]. Nevertheless, many children reach adolescence with little understanding of the implications of their cardiac condition [27
]. Patients who have grown up with congenital heart disease have particular medical service needs when transitioning from pediatric to adult care [26
]. Many guidelines and training programs for grown-up congenital heart disease have been set up, and pediatric cardiologists always play very important roles in that field [2
The principle of medical specialties indicates that pediatric care should transition to adult care when patients reach 18 years old [1
]. However, for patients with epilepsy, transferring from pediatric to adult care has shortcomings and adult neurologists often have insufficient information about new patients’ past and current medical conditions [3
]. Additionally, these patients may prefer to delay the need to develop new patient-physician relationships and move from a familiar to an unknown ward culture [6
]. Staying in the pediatric medical delivery system can enable them to avoid these transition difficulties.
With the increasing prevalence of type 1 and type 2 diabetes mellitus in children worldwide [4
], the demand for medical care for diabetic children is increasing. As these diabetic children grow up, they are expected to transition from pediatric to adult medical care [7
]. Nonetheless, similar to the difficulties faced by epileptic patients, diabetic patients may prefer to continue being treated by their pediatricians. Such a preference could explain, at least in part, why diabetes rose from a relatively uncommon diagnosis (28th) to the fifth most common diagnosis for adult ambulatory visits to pediatric clinics in the period of time analyzed in this study. In summary, for those patients with congenital diseases or chronic diseases diagnosed in their childhood or infancy, these patients prefer to stay in the pediatric medical delivery system for continuity. On the supply side, faced with a declining pediatric population, pediatricians have incentive to provide medical services to these patients. As the effect of the demand side and supply side combined, the proportion of adult ambulatory visits to pediatricians staffing in hospitals has steadily increased.
Although 74.3% of children’s ambulatory visits to pediatricians occurred at physician clinics, notably, 96.8% of adult ambulatory visits to pediatricians occurred at physician clinics, a rate vastly higher than at the other three site types [1
], and also much higher than the proportion of ambulatory visits to physician clinics with other types of specialists [11
]. Indeed, it is reasonable to suppose that the 12.4% growth in the number of physician clinics in Taiwan from 2004 to 2011 may have played a role in motivating pediatricians to expand their medical service offerings to adults due to intensifying competition.
Although chronic diseases occupy a large proportion of medical services delivered at adult ambulatory visits to pediatricians in academic medical centers, most adult patients appeared to seek medical care for airway infections/diseases and gastroenteritis in physician clinics. Similar observations were reported in a previous study in which upper respiratory tract infection was found to be the most common diagnosis made by pediatricians at children’s ambulatory visits [1
]. In our study, we found that the diagnosis of essential hypertension increased steadily over the examined years. This increase might be explained in part by population aging, which raises the prevalence of chronic diseases, such as hypertension, and associated healthcare needs [30
In a society with an aging population and a decreasing birth rate, we postulated that pediatricians may provide more medical services to adults. The present results are consistent with that supposition. Although pediatricians may have the opportunity to care for some adult patients with chronic illnesses diagnosed in childhood during their residencies, the current residency training program for pediatricians does not include an adult ward/ambulatory rotation or the distinct opportunity to evaluate and treat sick adults, especially geriatric patients. Nevertheless, the addition of adult or geriatric medical training per se to a pediatric residency would be on its face irrational. Therefore, adding formal adult and geriatric medicine training before doctors-in-training begin a pediatric residency, such as in a postgraduate year, may improve adult ambulatory care quality. The effects of the declining birth rate on the population are approaching irreversibility. Therefore, policy makers should re-consider what role pediatricians could and should play in the healthcare arena after completing their residency training. Although a theoretical system dynamics model for the Taiwanese pediatric workforce has been reported [31
], it has not yet been determined empirically how many pediatricians should be trained. This question merits further discussion and study.
This study had some limitations. First, we did not obtain data on quality of care or cost effectiveness, such as adequacy of medication for different diagnoses, medication dosage, or medical costs per visit. Therefore, we could not evaluate the quality of care provided by pediatricians to adult patients. Second, by analyzing a sampling of claim data, we could not know the exact final diagnosis associated with every ambulatory visit. Third, there is sometimes a time gap between births and formal birth registrations in Taiwan. In this time gap, neonates might utilize medical services via their parents’ NHI identity in so-called dependent visits. Therefore, some of the NHI claims data-associated neonatal diagnoses may be linked with parents’ NHI identities. This phenomenon could potentially have introduced some bias into our analyzed datasets. In our clinical experience in Taiwan, this phenomenon is most prevalent in inpatient departments shortly after neonates are born. Generally, families register a newborn child’s birth very soon after the neonate has been discharged. It is illegal to delay birth registration and doing so would prevent the child from accessing government funded social welfare benefits. Consequently, dependent visits for newborns brought to pediatric clinics are not typical. However, because we used the systematic sampling CD dataset, we cannot trace all ambulatory and inpatient claim data for each patient. Hence, we could not clarify this bias absolutely. Nonetheless, based on the aforementioned circumstances, we believe this possible caveat should not affect our conclusions.