Dear Editor
The reduced capacity for routine surgery during the COVID-19 outbreak triggers severe
consequences on waiting lists, determining their impressive expansion with management
costs
1
. The problem immediately burdens patients with urgent issues and cancer, whose number
of avoidable deaths indirectly due to COVID-19 is estimated close to that of SARS-Cov-2
2
. Planning and scheduling of surgery becomes complex on clinical, ethical and technical
grounds. Although several authors and professional associations have proposed clinical
prioritization through urgency classifications
3
, pathways and data system models, specific tools are necessary actually to run priority-based
scheduling sustainably, in a usable and scalable fashion
4
. The Surgical Waiting List InfoSystem (SWALIS) has been proposed previously
5
with such aims. Here we report on the pilot adoption of a new (SWALIS-2020) model
to prioritize elective surgery during the COVID-19 pandemic (https://www.isrctn.com/ISRCTN11384058).
This was a 6-week (March to May 2020) feasibility pilot cohort study testing a bespoke
software-aided, interhospital, centralized, multidisciplinary pathway serving all
major elective urgent surgery from specialties in the Metropolitan area of Genoa with
840 000 inhabitants. The pathway is based on centralized and multidisciplinary team
triage of referrals, prioritized further by the SWALIS-2020 model (
Fig. 1):
Urgency categorization over maximum waiting time, defined by implicit clinical criteria:
A1, 15 days (certain rapid disease progression); A2, 21 days (probable progression);
A3, 30 days (potential progression); B, 60 days (no progression but severe symptoms);
C, 180 days (moderate symptoms); D, 360 days (mild symptoms)
Waiting list prioritization, real-time ordered by the SWALIS-2020 score (percentage
of waited-against-maximum time) computed by a proportional, time-based, linear cumulative
method (PAT-2020) (
Figs 1 and 2
)
Theatre capacity planning, based on prioritized demand
Flexible, service-based, priority-based scheduling
Fig. 1
The linear method of prioritization method (Patent (PAT)-2007, SWALIS-2009)
The referring surgeon declares patient’s clock start date (t
0) and clinical urgency category (U) based on the likelihood of quick deterioration
to the point where it may become an emergency, or on the level of symptoms, dysfunction
or disability. Clinical urgency (U) is then associated with maximum waiting time from
t
0. In the SWALIS-2020 model, U can assume six different values in days: U = {A1=15,
A2=21, A3=30, B=60, C=180, D=360}. Given U and t
0, and defining P(t
0 +U) = 1, the priority (P) at the time of prioritization P(t) is defined as follows:
P
t
=
1
U
t
-
t
0
t
1
0 = patient 1 clock start date; U
1 = patient 1 urgency category maximum allowed waiting time; t
2
0 = patient 2 clock start date; U
2 = patient 2 urgency category maximum allowed waiting time; P
1 = patient 1 priority at time of prioritization (t); P
2 = patient 2 priority at time of prioritization (t). See
Fig. 2
legend for explanation of colour coding.
Fig. 2
The cumulative linear method of prioritization (PAT-2020, SWALIS-2020)
Clinical conditions can change during the waiting time (t
0, t
1, t
2, … tn
), affecting the patient’s urgency (U
0, U
1, U
2, …Un
). Priority can be calculated as summation, based on urgency variations:
P
(
t
)
=
1
U
0
(
t
1
−
t
0
)
+
1
U
1
(
t
2
−
t
1
)
+
1
U
2
(
t
2
−
t
1
)
+
……
+
1
U
n
(
t
−
t
n
)
P
(
t
)
=
1
U
n
(
t
−
t
n
)
+
∑
1
n
1
U
n
−
1
(
t
n
−
t
n
−
1
)
t
0 = start waiting time; U
0 = urgency for patient at starting time t
0; tn
= updated urgency time; Un
= updated urgency for patient; t = time of prioritization. The SWALIS-2020 prioritization
method assumes four priority score stages: ‘ideal’ (0–50 per cent), colour code white;
‘optimal’ (51–75 per cent), colour code green; ‘due’ (76–100 per cent), colour code
yellow; ‘overdue’ (more than 100 per cent), colour code red.
We monitored the safety and efficacy of the pathway by adverse events, drop-offs and
completions, auditing its performance weekly by the SWALIS cross-sectional and retrospective
waiting list indexes (dimensions and centrality), and by the SWALIS-2020 score at
admission. Applicability was tested over pathway deviation events, number of postponements
(before admission) and cancellations (on the day). Data were managed by live-running
interface, code-developed on MS VBA™ (Microsoft, Redmond, WA, USA). Statistical analysis
included use of Spearman’s rank test for correlation, the Mann–Whitney U test or one-way
ANOVA with the Kruskal–Wallis rank sum test, Dwass–Steel–Critchlow–Fligner or Loess
tests, performed with R software version 3.6.3 (The R Foundation for Statistical Computing,
Vienna, Austria).
After a 2-week feasibility phase (55 patients), 240 referrals were prioritized over
4 weeks with no major pathway-related critical events (M : F ratio 73 : 167; mean(s.d.)
age 68.7(14.0) years). Waiting lists were monitored, and theatres fully allocated
based on prioritized demand for the services. The mean(s.d.) SWALIS-2020 score at
admission was 88.7(45.2) in week 1, then persistently over 100 per cent (efficiency),
over a controlled variation (equity), with a difference between A3 compared with A1
(153.29(103.52) versus 97.24(107.93) respectively; P < 0.001), and A3 versus A2 (153.29(103.52)
versus 88.05(77.51); P < 0.001). A total of 222 patients eventually had surgery, with
no pathway-related complications or delayed/failed discharges.
Although different geographical areas are facing the COVID-19 outbreak asynchronously,
the waiting list backlog will continue for months, burdening hundreds of thousands
of patients, and prioritization will long remain a major issue. The SWALIS-2020 model
is designed for the broadest hospital acute care environment. It has smoothly selected
and prioritized the very few patients with the greatest need, scheduling their access
even with approximately 30 per cent capacity modifications weekly, managing active
and backlog waiting lists in the same process. The heterogeneity of established practices
in different services represents a challenge for waiting list pooling. However, the
SWALIS-2020 model has passed the test, allowing effectiveness, efficiency and equity.
These results encourage its wider adoption to prioritize surgery during the COVID-19
pandemic. We are looking for collaboration for further multicentre research.
Collaborators
E. Andorno, M. Filauro, G. Moscato, M. Rossi, S. Scabini and N. Solari (Department
of Surgery, Policlinico San Martino, Genoa, Italy); G. Buzzatti and P. Pronzato (Department
of Haematology and Oncology, Policlinico San Martino, Genoa, Italy); S. Campbell (Department
of Modern Languages and Cultures, University of Genoa, Genoa, Italy); W. Locatelli
(Regional Inter-Trust Surgical Departments, Regional Healthcare Trust, Liguria Region
Health Administration, Italy); M. Filauro and C. Introini (Department of Surgery,
Galliera Hospital, Genoa, Italy); M. Frascio, G. Peretti and C. Terrone (Department
of Surgery, Policlinico San Martino, and Department of Surgical Sciences and Integrated
Diagnostics, University of Genoa, Genoa, Italy); F. Martelli and G. Ucci (Hospital
Leadership Department, Policlinico San Martino, Genoa, Italy); G. Orsero (Department
of Emergency, Anaesthesia and Intensive Care, Policlinico San Martino, Genoa, Italy);
E. Raposio (Department of Surgical Sciences and Integrated Diagnostics, University
of Genoa, and Department of Haematology and Oncology, Policlinico San Martino, Genoa,
Italy), and L. Timossi (Department of Surgery, International Evangelical Hospital,
Genoa, Italy).