The base data set for our study contains more than 3 million cases registered from 2008 to 2014. The data set used for our main analysis comprises N=2,007,513 accidents that required outpatient medical care but no hospital inpatient stay. The number of included and excluded cases as well as descriptive statistics of the analysis sample are reported in the supplemental information S 1.
Injuries suffered by patients without an inpatient stay are mostly of low to moderate severity, as can be indirectly inferred by the typically short incapacity duration and the average low case costs. Most of the accidents (58%) resulted in no work absence or in an incapacity duration of three working days or less. 32% of the cases resulted in 4 to 10 working days absence, 8% in two weeks to three months, and only 2% in more than three months. The median of reimbursed direct medical costs per accident was 352 CHF (mean: 806 CHF; 75th percentile: 804 CHF). Hence, the provided medical care consisted in half of the cases of just one or two visits to a healthcare provider performing minor treatments.
In 4.1% of all cases, more than one group of providers had their first patient contact on the same day, of which 71% were a combination of GP with ED outpatient care, 16% of GP with medical specialist, and 13% of medical specialist and ED outpatient. For these, we assumed a treatment sequence as outlined in the methods section to define an unequivocal initial care provider.
Of those patients receiving care from medical specialists, 35% were treated by orthopedic surgeons, 33% by ophthalmologists, 20% by non-orthopedic surgeons, and 17% by other specialists.
3.1 GPs’ involvement in accident care and initial care provision from 2008 to 2014
In 2014, GPs were involved in 70.2% of all injury cases requiring outpatient medical care but no inpatient stay and provided initial care in 56.4% (Figure 1). While GPs’ involvement remained almost constant between 2008 and 2014 (drop of -0.9 percentage points), there was a more pronounced decline by -4.3 percentage points of GPs providing initial care. For emergency departments (ED), we observe that both involvement (from 38.1% to 45.6%, 7.5 percentage points) and provision of initial care (from 30.4% to 35.4%, 5.0 percentage points) increased substantially. Medical specialists were involved in 16.4% of all cases in 2014 and provided initial care in 8.2% of cases. Over time, involvement increased slightly (+0.9 percentage points) while initial care provision declined somewhat (-0.8%).
[Figure 1 about here]
3.2 Variations in GPs’ probability of providing initial care by patients’ place of residence, gender, citizenship, age, and point in time of the accident
Whether GPs provide initial care varies substantially between patients residing in urban vs. rural regions (Figure 2). While in 2014, 54.4% of patients residing in urban regions were provided initial care by GPs, it was 58.1% of patients residing in intermediate, and 60.3% of patients residing in rural regions. EDs show a complementary pattern: the probability that EDs provided initial care was highest in urban regions with 37.0% and lowest in rural regions with 31.9%. Even when adjusting for differences in the patient population, these geographical differences remain.
Males (55.1%) had a somewhat lower probability of seeking initial care from GPs compared to females (61.3%), although the difference diminishes considerably when adjusting for other patient characteristics and injury type (adjusted: 56.1% males vs. 58.0% females). Non-Swiss citizens (49.6%) have a considerably lower probability of seeking initial care from GPs than Swiss citizens (59.0%). The difference decreases somewhat after adjustment but remains large (adjusted: 51.0% Non-Swiss vs. 58.7% Swiss). Finally, younger patients aged 20 to 35 had a lower probability of GPs figuring as initial care provider (about 54%) than elderly patients aged 55+ (above 61%) (Figure 3). Again, these differences get slightly smaller after covariate adjustment, but remain substantial.
Regarding ED outpatient use, we observe a complementary pattern. Groups with lower probability of receiving initial care from GPs show higher probabilities for ED use, which apparently serves as a substitute for GPs.
Concerning medical specialists, there are only minor differences in initial care provision: after adjustment, there is no substantial difference between males and females, a 1.8 percentage points lower probability for Non-Swiss vs. Swiss citizens, and a 2.1 percentage points lower probability for rural vs. urban regions.
[Figure 2 about here]
[Figure 3 about here]
Further analyses show that there is considerable variation in the provision of initial care depending on the time-of-day and day-of-week the accident occurred. As Figure 4 shows, accidents occurring during daytime/evening show a considerably higher probability (between 51% and 60%) of GPs providing initial care compared to those happening at night (49% at 11pm, 40% at 1am, 43% at 4am). The probability of GPs providing initial care is also higher for accidents occurring on Sunday (58%) and at the beginning of the week (Monday 59%, Tuesday 58%), compared to those occurring later in the week (Wednesday 56%, Thursday and Friday 55%, Saturday 54%, see figure in Supplementary Information S 4). The pattern for EDs is complementary, showing the substitution between GPs and EDs. After adjustment, the differences are less pronounced for time-of-day and slightly more pronounced for day-of-week, showing that the type of accidents and injuries varies substantially according to the point in time the accident occurred. Even after adjustment, considerable injury heterogeneity may remain, which prevents us assigning the observed variations in initial care provision unequivocally to the point in time of the accident.
[Figure 4 about here]
3.3 Changes in the role of GPs in the care pathway from 2008 to 2014
Who provides initial care after an accident (GPs, medical specialists or a hospital ED) determines strongly the role of GPs during the subsequent care pathway. In the following, we categorize care pathways by GPs’ role in the whole care pathway:
- GP sole care provider: GPs were the only care provider
- GPs è specialists/ED outpatient: initial care by GPs, subsequent care by medical specialists or emergency department (outpatient)
- GP follow-up only: GPs provided only follow-up care after a medical specialist or an ED provided initial care
- GP not involved: GPs were not involved at all
In 2014, GPs were the sole care provider in 44.4% of the cases, i.e. no other provider group than GPs was involved. This is a substantial drop by 7.0 percentage points compared to 2008 when this share was 51.4%, . Correspondingly, the share of cases where GPs provided follow-up-only care increased by 3.4 percentage points from 10.5% to 13.9%. Also, the share of cases where GPs provided initial care, whereafter patients saw an ED, increased from 6.0% to 8.2%, while cases where patients subsequently saw a medical specialist remained stable at a low level (2008: 3.2%, 2014: 3.8%). The share of cases GPs are not involved in remained almost unchanged (2008: 28.9%, 2014: 29.8%).
[Figure 5 about here]
Extending the above analyses of the involvement of the various providers, the initial care provider, and the role of GPs to subgroups of injuries shows that the observed pattern applies to a wide range of injuries (see Supplementing Information S 3). While the percentage of GPs providing initial and sole care is in general higher for less severe types of injuries such as bites or cuts compared to more severe ones such as fractures, the trend of an increasing substitution of EDs for GPs is observable for all injuries considered. The decline in GPs providing initial and/or sole care and the complementary increase in Eds’ involvement is, however, more pronounced for less severe injuries compared to more severe ones (see Supplementing Information S 3).