Monitoring the indirect impact of Covid-19 on hospital care pathways in Italy: results from the Mimico-19 network

The COVID-19 pandemic has a non-negligible indirect impact on hospital care pathways, which is important to estimate. To this aim, we set up the Mimico-19 network of seven Italian regions (62% of the Italian population) representing different socio-demographic areas of the country with also a different burden of the epidemic. We retrospectively analysed regional hospital discharges data, computing twelve indicators of volumes and performance in three clinical areas: cardiology, oncology and orthopaedics, including time-dependent pathways and elective surgery. Weekly indicators for the period January-July 2020 were compared with the average of the corresponding indicators in 2018 and 2019; comparisons were performed within 3 sub-periods: pre-lockdown, lockdown and post-lockdown. The weekly trend of hospitalizations for ST-segment elevation myocardial infarction (STEMI) showed a 40% reduction, but the proportion of STEMI patients with a primary PTCA did not significantly change from previous years. Volumes of malignant neoplasms surgeries differed substantially by site, with a limited reduction for lung cancer (<20%) and greater declines (30-40%) for breast and prostate cancers. Hospitalizations for femoral neck fracture in the elderly decreased by 20%, but the percentage of timely interventions remained constantly higher than the previous years. General trends did not show important differences across regions, regardless of the different Covid-19 burden. Hospitalizations have generally decreased, but the capacity of a timely and effective response in time-dependent pathways of care was not jeopardized throughout the period. The drop in the care demand for cardiovascular diseases and cancers needs to be further investigated and monitored more thoroughly.


Introduction
The Covid-19 pandemic in Italy has caused more than 4,2 million cases of infection and over 130,000 deaths to date [1]. Following the first epidemic outbreak in Lombardy, which rapidly spread to other regions, the government issued the first national lockdown starting on 9 March 2020, and since then, evidence had accrued on its possible effects on people's health [2][3]. Services whose benefits could be lower than both the patient's risk of infection and the organizational difficulties were postponed, and citizens were recommended to avoid unnecessary access to health services. This indication, however, may have causedfor fear of contagion or misinterpretation of the norm -further delays in recognizing symptoms and timely accessing diagnosis and treatment, even for non-deferrable conditions. Furthermore, several specialist departments had their hospital beds reduced to be used to face the huge flow of COVID-19 patients.
Earlier in 2020, seven Italian regions (Piedmont, Lombardy, Emilia-Romagna, Tuscany, Lazio, Puglia and Sicily) have therefore set up the Mimico-19 network to monitor the side effects of the restrictive measures against Covid-19 on the quality of care. The seven regions total about 37 million inhabitants (62% of the Italian population) and are representative of different geographical areas of the country, with a different epidemic burden [4].
The objective of this report is to describe the indirect impact of the pandemic and the lockdown on hospital activities through indicators of volume and performance in three clinical areas: cardiology, oncology, and orthopaedics, characterized by the high volumes of activity and the severity of the conditions.

Methods
Using data from the regional hospital discharge databases, we defined four indicators for each of the three clinical areas (the ICD-9-CM codes included in the indicators are detailed in the Online Resource 1):

Results
Here we report the results for the pooled indicators, highlighting only important regional variations (detailed regional results are presented in the Online Resource 2).

Cardiology
The weekly trend of hospitalizations for STEMI ( Figure 1a) exhibited a gradual reduction in volumes, which began in late February and peaked in mid-April (% variation about -40%). Trends revealed a highly significant difference during lockdown, although STEMI hospitalization showed a significant 4% reduction even before the pandemic. In the post-lockdown, volumes went slowly back to the starting values. The hospitalization for NSTEMI ( Figure 1b) showed a similar trend, but the average percent reduction was greater (over 55% at the end of March) and the post-lockdown recovery was slower. The proportion of STEMI patients with a primary PTCA (Figure 1c), which could be estimated only in four regions, remained stable over time without significant changes from previous years. On the other hand, in-hospital mortality significantly increased by 26% on average during lockdown and by 15% in the subsequent period ( Figure   1d). This unfavourable outcome was almost entirely driven by Lombardy, the region with the earliest and greatest burden of Covid-19; all the other regions showed non-significant differences in mortality, although with large weekly fluctuations ( Figure S1d, Online Resource 2).

Oncology
The total volume of malignant cancers surgeries (Figure 2a) showed a decline, which was moderate during the first two weeks of lockdown, reached -25% by the end of March, and remained at these levels even in the post-lockdown. Trends differed by cancer site: the reduction in volumes was limited (less than 20%) for lung cancers (Figure 2b), whereas it fell by about 30-40% for breast and prostate surgery (Figures 2c-d), earlier for the latter and in the post-lockdown for breast cancer.

Orthopaedics
Hospitalizations for femoral neck fracture in the elderly (Figure 3a) decreased by about 20% during lockdown, remaining significantly below the volumes of the previous two years in the post-lockdown too. In contrast, the percentage of timely interventions (Figure 3b) was constantly higher than the previous years throughout the whole period, with the difference increasing in the post-lockdown. Hip and knee replacements plunged dramatically following national indication to suspend scheduled operations. In both cases, however, there was an inversion of the trends by the end of July 2020, with volumes significantly exceeding those in 2018-2019.

Discussion
In The reduction in STEMI hospitalizations may be linked to a real lower incidence of heart attacks (e.g., related to better air quality or reduced mobility), but we cannot exclude delays in diagnosis or self-limitation of the demand by patients for fear of contagion. The secular trend of reduced hospitalization for STEMI may have also played a role, though marginal. Indeed, the lack of timely access to hospital care would also be supported by the reported increase in the number of out-of-hospital cardiac arrest deaths [5,6]. Moreover, we observed an increase in STEMI in-hospital mortality in Lombardy, where the high prevalence of Covid-19 was associated with worse STEMI outcomes [7]. This could also be associated with the delay in seeking hospital care by patients suffering a heart attack, which would result in a reduction in the efficacy of a primary PTCA.
The overall volume of cancer surgery hides profound differences across sites: in each of them, in fact, the decision to postpone the surgery depends on the balance between the benefit of an immediate operation and the risk of hospital infection from Covid-19 [8]. As expected, the decline in volumes of lung surgery was negligible, mainly due to its non-deferrable nature, while reductions for breast and prostate surgery were greater. Notably, especially for breast cancer, the persistence of the reduction in the post-lockdown is likely an indirect consequence of the suspension of screening activities [9], whose real impact on women's health will only be visible in the future. Indeed, a four-week delay in breast cancer surgery has been estimated to be associated with an 8% increased risk of mortality [10].
Hospitalizations for femoral neck fracture in the elderly decreased, probably due to the limited mobility imposed by the confinement measures, while hip and knee replacements were affected by the interruption of non-urgent surgery and the need to adapt surgical and outpatient activities to the new safety regulations [3].
This suspension has likely left room for an improvement in the timeliness management of emergencies, as in the case of the femoral neck fracture. Interestingly, the offer of orthopaedic surgery during the 2020 summer exceeded the volume of operations usually performed in this period. This reflects the timely and effective reorganisation of the sector after the lockdown in order to recover the accrued delays.

Conclusion
Rescheduling less urgent interventions may not affect survival in the short term but may have amplified the severity of unresolved health problems. Furthermore, the deferral involves a lengthening of waiting lists that will require a re-scheduling attentive to priorities, equity and efficiency. At this stage of the epidemic, it is necessary to strengthen primary care services so that they can adopt a pro-active approach and move towards the identification of risk conditions that were neglected during the pandemic and timely address patients to the secondary care system.
Finally, the drop in cardiovascular or oncological care needs to be monitored more thoroughly to counteract a possible reduced access to early diagnosis (as in the case of breast cancer screening) and follow-up of severe chronic diseases.