Adolescent health presentations to Victorian general practice: a descriptive study using electronic medical records

Data collection

A retrospective, quantitative descriptive study was undertaken using de-identified data from EMRs of adolescent patients from a sample of Australian general practices across a single year. EMR data (including reasons for presentation, diagnoses, prescriptions issued, pathology and imaging ordered, Medicare item billed, and patient demographic data) were extracted from the records of 29 consenting Victorian GP clinics participating in the ‘Primary Care Audit, Teaching and Research Open Network’ (Patron) program (Manski-Nankervis et al. 2024). Patron is a research initiative of the University of Melbourne’s Department of General Practice and Primary Care that aims to make better use of existing primary care data to improve knowledge, medical education, policy and the way medical care is delivered (University of Melbourne Department of General Practice and Primary Care 2024). At the time of the present study, data extracts from the three most commonly used EMR software systems in Australia (Best Practice, Medical Director and ZedMed) were not harmonised in the Patron dataset, and hence only practices using Best Practice were included for this analysis.

Data management

To prepare data for examination and explore key research questions, data management proceeded through three steps (see Fig. 1): (1) identifying which visits to participating GP clinics across the year 2019 were made by adolescent patients; (2) determining which encounters in the EMR from among this patient group were for actual clinical visits; and (3) assigning RFVs to each of these identified clinical visits and classifying them according to the International Classification for Primary Care-2 (ICPC-2; WHO 2005) classification system to calculate the relative frequency of RFVs.

Fig. 1.

Steps followed to identify, determine and classify adolescent clinical presentations to GP clinics across 2019.

Determining clinical visits

The central focus of this study was to examine the nature of clinical GP visits by young people to primary care. However, not all entries captured in the patient EMR relate to clinical visits (e.g. administrative and clinical management activities, such as file reviews, clerical tasks and documented failures of attendance may all be captured). Hence, it was necessary to identify which entries in patients’ data records reflected actual clinical visits (i.e. patient attended GP clinic seeking support for an identified clinical need).

The target encounters for this study were those conducted by a GP or medical student (overseen by a GP) and classified as clinical in nature. We followed a systematic approach to determine which visits in the dataset met these criteria (see Fig. 2). For example, visits that were undertaken by practice support staff or by providers other than GPs or medical students were dropped from the dataset, as were types of encounters considered non-clinical (e.g. email sent; patient did not attend; MBS or prescription fields indicated visit was not clinical; visit duration of 0 min).

Fig. 2.

Flowchart showing classification of adolescent patient clinical visits across 2019. Notes: ¹Patients may record more than one visit across the 12-month date period. MBS, Medicare Benefits Scheme.

Data analysis

The main indicators of interest in examining adolescent primary care visits were: (1) annual rates of clinical visit per patient (based on data gathered over a single year), and (2) frequencies of different RFVs. These were examined both for the overall cohort of adolescent patients, and by age band, gender and locality. Visit rates are expressed as mean annual visits per patient, whereas frequency of different RFVs are presented by ICPC-2 classifications, and were calculated as the number of RFVs per 100 patient visits. Patient visits for which no RFVs were able to be assigned were not included in the denominator when calculating rates of RFV per 100 visits.

Several cross-chapter groupings of interest were also examined (i.e. higher-order classifications that combined rates across ICPC-2 chapters), including reason visits relating to preventive vaccination and medication (e.g. sexual and reproductive health, vaccines, lifestyle counselling), and several key GP processes (discussion of results, follow-ups, medications and referrals). All analyses were conducted using Stata software (BE ver. 17.0; StataCorp).

Ethics approval

Ethics approval was granted by the Human Research Ethics Committee of the University of Melbourne on 08 April 2020 (Ethics ID # 13448). Authors had full access to all the data (including statistical reports and tables) in the study.

Sample characteristics

The final dataset consisted of 22,062 adolescent patients, who recorded a total of 63,706 visits to primary care across the 12-month period, while seeking medical support for 79,490 total RFVs (see Table 1). Adolescent patients were drawn from 29 participating general practice clinics across Victoria; 16 practices were located in a major city and 12 were classified as regional.

Patients in the oldest adolescent age bracket (20–24 years) accounted for the greatest proportion of patents (38.2%; 8428/22,062), as well as the greatest proportion of visits (40.4%; 25,741/63,706). By gender, females accounted for 56% of the patient cohort (3046 + 4223 + 5150/22,062) and >60% (8053 + 13,752 + 17,232/63,706) of the general practice visits observed; females aged 20–24 years accounted for the largest proportion of recorded patient visits (27.1%; 17,232/63,706). Patients whose gender was recorded as ‘other’ made up just 0.1% of total patients and attended 0.1% of recorded general practice visits (values not presented in tables due to very low frequencies). In terms of locality, patient numbers and visits were split roughly evenly between city and regional practices.

Rates of visit to general practice

The average rate of visits per patient across the cohort was 2.89 visits per patient (see Table 1). Visit rates were found to increase with increasing age, with females in the 20–24-year-old age band showing the highest average visit rate of 3.35 visits per patient. Within each age category, females visited general practice at slightly higher rates than males, and on average attended approximately 0.5 more general practice visits per year than males. Overall rates of visit to regional practices were slightly higher than those to city practices (3.06 vs 2.74, respectively).

Reasons for visit

Information (either partial or complete) was present in the ‘visit reason’ field of the EMR for a majority (almost 99%) of recorded clinical visits. For 14% of clinical visits (9132/63,708), we could not assign an ICPC-2 code to RFV, because there was insufficient detail in the EMR (across the three fields examined). Table 2 shows overall rates of RFV for each of the 17 ICPC-2 chapters (representing major body systems; presented in descending order of overall frequency) along with the 45 most frequently recorded RFVs per 100 visits (based on ICPC-2 classification) separately for males and females, and for each age band (10–14 years, 15–19 years and 20–24 years).

Overall, respiratory RFVs and psychological RFVs were most frequent (at a rate of 20.3 and 19.6 reasons per 100 visits, respectively), followed by skin-related RFVs (15.8), and general and unspecified RFVs (12.8). This same group of RFVs featured in the five most frequent reasons across each of the gender-by-age groups, with two exceptions: pregnancy, childbearing and family planning RFVs were common among 15–19-year-old females (15.4, 3rd highest RFV) and 20–24-year-old females (19.1, highest RFV); and female genital RFVs were common among 20–24-year-old females (14.9, 4th highest RFV). Overall, females were found to present more frequently for genital-related and urological-related RFVs than males across all age groups.

Rates of respiratory, skin and musculoskeletal RFVs were generally higher among males compared with females, with differences tending to increase with age. Conversely, females showed higher rates of psychological RFVs than males in the younger two age bands, with this trend reversing at 20–24 years. Psychological RFVs appeared to peak during mid-adolescence (aged 15–19 years) for females (RFV rate 21.6) and late adolescence (aged 20–24 years) for males (RFV rate 21.6).

Differences in adolescent RFVs were next examined by locality of GP clinic (city vs regional practices). These comparisons were based on higher-order groupings of ICPC-2 chapter (body system) only (see Table 3). Compared with city practices, visits to regional practices demonstrated higher rates of psychological and pregnancy/family planning RFVs. Conversely, city practices showed higher rates of respiratory and skin-related RFVs compared with regional practices. Otherwise, RFV rates for most other ICPC-2 categories were similar between different geographic locations.

Finally, rates of presentation for several broad categories of RFVs that combined findings across ICPC-2 chapters were examined (i.e. cross-chapter categories; see Table 4). These key groupings included RFVs relating to preventative vaccination and medication, and ‘process’ RFVs that represent common clinical activities that span multiple ICPC-2 chapters (viz results, follow-up, medications, referrals). Among these groupings, preventative RFVs were the most common, with 6.3 preventive vaccination/medication RFVs for every 100 visits.

Rates of attendance

Overall, young people’s rates of visit and their patterns of primary care presentations showed noticeable variation by age band, gender and locality. Rates of attendance to primary care were found to be highest among older patients, females and those in regional localities; females aged 20–24 years accounted for the largest proportion of recorded patient visits. The reasons that young people present to general practice were found to vary by age, gender and practice locality. Overall, respiratory RFVs and psychological RFVs were most frequent, followed by skin-related RFVs, and general and unspecified RFVs. Rates of respiratory, skin and musculoskeletal RFVs were generally higher among males compared with females, whereas rates of psychological, genital-related and urological-related RFVs tended to be higher among females.

The overall mean attendance rate of young people in our sample was 2.89 visits/patient per year, comparable to the value of 3.00 observed in Booth et al.’s (2008) Australia-wide study of BEACH data from 1999 to 2005. Given that the National Prescribing Service General Practice Insights Report found a similar value of 3.3 visits per patient among Australian young people in 2017–2018 (NPS MedicineWise 2021), this suggests that the rate of general practice encounters for this group has remained relatively stable over time. Although comparison between these datasets may be complicated by the fact that our data reflect Victorian and not national activity, overall rates of Victorian GP encounters have been shown to closely approximate national rates, allowing for meaningful comparison (NPS MedicineWise 2021). These relatively low rates complement recent trends in national health service use data that show young Australians (aged 15–24 years) remain the group least likely to attend general practice relative to all other adult age bands (Australian Bureau of Statistics 2022–2023), a pattern also noted two decades ago in the BEACH data (Booth et al. 2008). It has been argued that low rates of primary care presentation among young people may simply be due to their relative good health (Booth et al. 2008) and consequent reduced healthcare needs. However, a growing body of research has identified a range of barriers to accessing primary care among young people, including confidentiality concerns, cost, lack of youth-friendly services and lack of service awareness (Booth et al. 2004; Tylee et al. 2007), all of which may impact access rates.

Reasons for visit

Young people presented to general practice most frequently for physical (biomedical) concerns, particularly respiratory, skin and general/unspecified physical complaints. This finding is consistent with existing literature on adolescent primary care presentations both from Australia and overseas that show young people present most commonly for biomedical concerns (Booth et al. 2008; Frese et al. 2011).

A key point of divergence in our findings compared with these studies, however, is the high frequency of psychological (mental health) reasons for encounter (the second most frequent RFV by ICPC-2 chapter in our sample). This marks a notable upward shift since the Booth et al. (2008) examination of encounters in Australian general practice between 1998 and 2004, in which mental health presentations were found to be infrequent among adolescent patients. The apparent increase in youth mental health consults to primary care corresponds with recent Australian trends showing a rapid decline in mental health for 15–24-year-olds across the past decade (Wilkins et al. 2022). It may also be related to the introduction of the Better Access scheme by the Australian government in 2006 (Department of Health 2023), which led to a considerably expanded range of mental health treatment services covered under Medicare and provided by GPs and mental health practitioners. Youth access to mental health support has increased dramatically since the inception of this scheme (Jorm and Kitchener 2021).

Increased frequency in adolescent mental health presentations may also be in part due to plausible reduction in stigma associated with mental illness (a known barrier to help-seeking among young people with mental health problems (Brown et al. (2015)) following well-publicised campaigns by national mental health organisations, and through various initiatives that aim to reduce stigma and discrimination towards people with mental illness, although research evidence of program effectiveness is limited (Morgan et al. 2021).

In light of previous research showing stark discrepancies between adolescent mental health needs and rates of general practice management (Hickie et al. 2007), our findings provide encouraging signs of greater adolescent engagement with general practice for psychological issues, and emphasise the continued need for proficiency among GPs regarding youth mental health.

Demographic factors

Between males and females, the largest differences with respect to RFVs were seen for pregnancy/family planning, genital and urological RFVs, all of which were observed at notably higher rates for females across all age groups. This concentration of urogenital and reproductive issues among female adolescents in general practice has been observed in previous studies of adolescent general practice encounters (Booth et al. 2008), and is a likely contributor to the higher overall attendance rate of females observed in our investigation. Moreover, the initiation of sexual activity common during this period of adolescence (Power et al. 2022) may account for the sharp increase in rates of pregnancy/family planning and genital RFVs observed among females in the 15–19 years age group.

Our study used the Modified Monash Model to classify visits as occurring in either city or regional GP settings, and found the rate of youth visits to rural practices to be slightly higher than that for city practices. This is despite a body of research indicating greater barriers to healthcare access among adolescents in regional and remote settings (Quine et al. 2003; Booth et al. 2004). This finding may be due to the relative lack of alternative services in rural communities, necessitating presentation to general practice in cases where adolescents in cities might seek help from other healthcare providers.

Our findings also indicate that regional general practice visits more frequently involve psychological RFVs compared with metropolitan practices, where physical problems (e.g. respiratory, skin) tend to be more common. Despite the prevalence of mental illness being similar between those living in major cities and those in regional areas (Australian Bureau of Statistics 2020–2022), suicide-related contacts in general practice are known to increase with increasing remoteness from urban areas (Harrison et al. 2013). As such, more frequent presentations relating to suicidal ideation or attempt may contribute to the observed higher rate of psychological RFVs in regional practices compared with urban areas. A more detailed analysis of mental health RFVs is required to determine which complaints are responsible for this difference. Also notable is that RFVs related to pregnancy/family planning were almost twice as frequent among regional practices. This may be due, in part, to relatively higher rates of teen pregnancy in regional, rural and remote areas (Family Planning Australia 2022), as well as increased barriers faced by women accessing specialist sexual and reproductive health services in these areas (Wood et al. 2024), and their greater reliance on GPs to deliver reproductive healthcare services in Australian rural communities (The Royal Australian College of General Practitioners 2018).

Strengths and limitations

To the best of our knowledge, our study is the first to use electronic medical records to comprehensively describe RFVs of adolescent patients in Australian primary care. The use of a large repository of secondary EMR data allowed us to study a sizeable sample of patients and visits across geographically diverse locations over a full calendar year. Our use of EMR data collected at the point-of-care as part of routine clinical practice meant that results were unaffected by recall bias. The records obtained provide rich data on patient encounters across multiple practices, and allow detailed insights to be gleaned of adolescent presentations in Victoria. Although data incompleteness and quality are known limitations of secondary research using EMRs (Canaway et al. 2019), our study sought to address this concern by following a systematic approach to determine which visits in the dataset were clinical in nature, and by utilising multiple data fields (such as MBS item recorded and script provided) to infer clinical activity where RFV was incomplete or not recorded. However, despite the benefits of our adopted multi-step approach, our RFV identification process may have included inaccuracies, as EMRs are imperfect data repositories and their use may lead to under-reporting of diagnoses (Canaway et al. 2024). Although other EMR fields could plausibly have been added to our approach, the combination of the three fields examined was considered to offer the best balance of accuracy, reliability and efficiency.

Several key limitations are also noteworthy. Although our focus on Victorian GP clinics only may limit the generalisability of findings, previous data suggest that overall rates of Victorian GP encounters closely approximate national rates (NPS MedicineWise 2021). Additionally, our decision to maximise consistency in EMR data structure between clinics by including only those sites using the common software package, Best Practice, resulted in the exclusion of >60 Victorian clinics with available data, representing a further potential limit on the generalisability of findings. Likewise, our reliance on data from clinics enrolled in the Patron program (Manski-Nankervis et al. 2024) may impact generalisability. Finally, our reporting of findings based on patient gender may be incomplete, as experiences of primary care among gender-diverse youth may vary from those identifying as male or female (Strauss et al. 2022). These potential limitations may be addressed in future studies by expanding program participation to practices in other Australian states and territories, incorporating additional software vendors, and seeking broader representation of patients from diverse backgrounds.