Strengths and Limitations
To our knowledge, this is the first study to develop a predictive
nomogram for PTB of twin gestations and a web-based calculator to
improve the approachability of the prediction model. Our model could
synthesize the most relevant risk factors for PTB, such as nulliparity,
pre-pregnancy BMI, history of PTB or late abortion, chorionicity,
cervical funneling and cervical length, and generate a risk percentage
number for each patient, which had a predictive accuracy (AUC 0.856
[95% CI: 0.813-0.899]) significantly that was higher than that of
either variable alone. The strengths of the study include the detailed,
standardized data collection, high rate of follow-up and efficient
statistical analysis. Based on the model, we could provide reliable risk
estimation for clinical counselling, therapy decision-making, and
follow-up strategies, rather than complicating the clinicians’
lives with close monitoring and administration resulting from
an undefined or inherently subjective risk assessment. All data
pertaining to characteristics in our study can be easily obtained in the
obstetric units where the cervical assessment has been well
standardized, and these data were obtained by well-trained specialists
throughout the study period. Moreover, external validation and
restricted cubic splines supported the test performance.
On the other hand, the present study has some limitations. Most
importantly, our study is limited by its retrospective design. There is
a possibility of confounding bias: patients with unmeasured or
unobservable factors who were excluded may represent patients at higher
risk. Second, the sample number of marginal risk scores is relatively
insufficient, which limits generalizability because it might not fully
reflect the actual situations of patients at the highest risk and who
are may be the most clinically interesting population. Last, the study
population in the two centres is limited to our own population (Asian).
This potential limitation may also be considered as a strength. All
women included in the study were followed up and treated only in the two
tertiary medical centre, which limits confounding factors associated
with the heterogeneity in provider bias, such as clinicians’ experience,
and differences in process of monitoring and management for offering the
intervention.