ABP Neonatal-Perinatal Medicine common mistakes
ABP Neonatal-Perinatal Medicine · Understand the traps and how to avoid them.
Classifying an intrapartum tracing as either reassuring or an emergency: documenting "normal" for a strip with recurrent late decelerations and moderate variability, or calling for an immediate cesarean for a category II strip with a single variable deceleration and otherwise normal features.
Candidates memorize the category I and category III definitions and then have no algorithm for the large category II middle ground, so they default to whichever label feels safer under time pressure.
What to do instead
Build a category II triage algorithm: name the specific abnormal component (recurrent late or variable decelerations, minimal variability, tachycardia, loss of accelerations), list the intrauterine resuscitation steps for each, and define an explicit time point at which the pattern's persistence triggers delivery. Practice with tracings that must be sorted into "resuscitate and observe" versus "proceed to delivery."
Running the term-infant delivery room script on a 26-week infant: 100% oxygen to start, immediate vigorous sustained inflations, no plastic wrap or thermal mattress, and a 3:1 compression-to-ventilation ratio applied without asking whether the bradycardia is of cardiac or respiratory origin.
The resuscitation algorithm is taught and rehearsed as one linear sequence, and the gestational-age branches (initial FiO2 blended up from 21-30%, gentler ventilation, aggressive thermoregulation, and the fact that most bradycardia in a preterm infant is respiratory) are learned separately and not integrated.
What to do instead
Rehearse resuscitation as at least three parallel scripts keyed to gestational age and to the likely cause of bradycardia, and state the branch point out loud in practice: "preterm under 32 weeks, so I start at 21-30% and titrate to target, apply the wrap, and correct ventilation before compressions."
Anchoring on transient tachypnea of the newborn in a term infant with tachypnea and mild retractions, never obtaining pre- and post-ductal saturations, never re-examining after 4-6 hours, and never considering delayed transition or pulmonary hypertension when the oxygen requirement persists.
TTN is the most statistically likely diagnosis in a term infant with tachypnea, so the first plausible label is held and the trajectory of the illness is not used to revise it.
What to do instead
Attach a diagnostic checkpoint to every term respiratory distress admission: paired pre- and post-ductal saturations, a repeat examination and blood gas at a defined interval, and an explicit differential that includes delayed transition/PPHN, pneumonia, pneumothorax, and congenital heart disease. Compare pre- and post-ductal values and escalate if the gradient is abnormal.
Palpating only a brachial pulse and concluding that perfusion and pulses are normal, then discharging a newborn with a systolic murmur and mild lower-extremity coolness without four-extremity blood pressures or pre-/post-ductal saturations.
The newborn cardiovascular examination is taught as "murmur, cyanosis, pulses," and femoral pulses are the step most often skipped when the infant looks well and the brachial pulse is brisk.
What to do instead
Adopt a fixed, non-negotiable cardiovascular discharge screen: both brachial and both femoral pulses, four-extremity blood pressures, and pre-/post-ductal saturations, documented as a set. Any abnormal screening result converts to an echocardiogram before discharge rather than a repeat examination the next day.
Reassuring the team with a normal cranial ultrasound obtained in the first 24 hours in an infant with an abnormal neurologic examination and a low cord pH, then deferring further imaging or therapeutic hypothermia decisions on that basis.
Candidates conflate "imaging was normal" with "no injury," and early cranial ultrasound is indeed frequently normal in the first day of hypoxic-ischemic injury.
What to do instead
Use magnetic resonance imaging (conventionally days 4-7 of life, with spectroscopy) as the injury-defining study, and treat the early cranial ultrasound only as a screen for hemorrhage, hydrocephalus, or a major malformation. Base hypothermia eligibility on gestational age, examination, and biochemical criteria, not on an early ultrasound.
Ordering a contrast enema or upper gastrointestinal contrast series in a preterm infant with bloody stools, abdominal distension, and pneumatosis intestinalis on plain film.
The reflex "abdominal distension plus blood equals a need for a contrast study" overrides the recognition that the working diagnosis is necrotizing enterocolitis, in which contrast studies risk perforation and are not diagnostic.
What to do instead
Use a staged NEC pathway: plain abdominal films at defined intervals, serial abdominal examinations and measurements, nothing by mouth with decompression, antibiotics, and surgical consultation. Reserve contrast studies for the specific problem of differentiating atresia or obstruction in a stable infant without evidence of NEC.
Initiating a full sepsis evaluation and a 48-72 hour antibiotic course in a well-appearing 38-week infant whose only risk factor is maternal group B streptococcal colonization with adequate intrapartum prophylaxis given more than 4 hours before delivery, or conversely discharging without evaluation an ill-appearing infant because the mother was culture-negative.
Risk-factor counting is learned as a list rather than as a categorical decision tree, so each factor is added up independently instead of being read as a branch that yields "observe" versus "evaluate."
What to do instead
Memorize the categorical algorithm by branch: maternal chorioamnionitis, gestational age, duration of membrane rupture, GBS status, adequacy of prophylaxis, and clinical appearance. For each branch rehearse the specific actions (no culture and observation versus limited evaluation versus full evaluation with lumbar puncture) and the duration of antibiotics.
Interpreting a day-10 infant with poor feeding, vomiting, weight loss, hyponatremia, and hyperkalemia as sepsis with a metabolic component, without measuring 17-hydroxyprogesterone, renin, and aldosterone, and without recognizing that a male infant can have a salt-wasting crisis with normal-appearing genitalia.
Neonatal sepsis is the default explanation for lethargy and poor feeding, so electrolyte derangements are treated as consequences of sepsis rather than pursued as a second, coexisting diagnosis.
What to do instead
Trigger a fixed metabolic review whenever sodium is low and potassium is high in a neonate: repeat the electrolytes and check glucose, 17-hydroxyprogesterone, renin, aldosterone, and a cortisol if indicated, and start stress-dose hydrocortisone plus sodium chloride replacement once the diagnosis is suspected rather than waiting for the screening result.
Selecting a standard G-banded karyotype as the first-line genetic test for a dysmorphic newborn with a suspected microdeletion, or ordering targeted FISH only for a syndrome the examiner has already committed to clinically.
The tests are taught as a hierarchy of "chromosomes, then genes," and the resolution limits of karyotype (roughly 5-10 megabases) and the broader diagnostic yield of microarray are not part of the working knowledge base.
What to do instead
Use chromosomal microarray as the first-tier test for a dysmorphic infant with an unexplained multiple-malformation pattern or developmental concern, and reserve karyotype for suspected aneuploidy in a rapidly evolving situation, for a specific rearrangement known in the family, or for culture-based confirmation of a balanced rearrangement.
Declaring a quality improvement project successful after three consecutive data points above the baseline mean, or proposing an EHR order-set change alone as the complete intervention for a persistent safety problem such as unintended extubation or central line-associated bloodstream infection.
Candidates read any upward run of points as improvement and view documentation or order-set changes as equivalent to a full intervention, whereas reliable improvement requires measurement over time with control rules and attention to process reliability and balancing measures.
What to do instead
Use statistical process control explicitly: apply the standard run and shift rules before calling special-cause variation, report centerlines and control limits, and design the intervention as a bundle with a named owner, a process measure with a reliability target, an outcome measure, and at least one balancing measure, tested over multiple plan-do-study-act cycles.