Discussion

Historically, microscopic PBS evaluation has been engrained throughout medical education as a reflexive screening tool to quickly evaluate a wide range of diseases including leukaemias, myeloproliferative disorders, haemolysis, hemoglobinopathy and parasitic infections. The additional benefit of being a rapidly available, easily accessible, and minimally invasive test accounts for why it is one of the most commonly ordered laboratory tests. With the introduction of highly sophisticated haematology analysers, the role of a manual PBS review by a pathologist is limited. The current day haematology analysers have improved the ability to detect morphological abnormalities on peripheral smears with great precision and accuracy to the point where manual review provides limited additional information.7 Correspondingly, a CAP Q-probe study reported that more than one third (35.7%) of participants believed that PBS review did not provide any additional clinical information.8 However, there are few instances where PBS reviews screened by analysers may miss significant findings that impacts patient care, which necessitate the expertise of a second review by a pathologist. These criteria are highlighted in box 1 and table 1, based on our institution’s experience. Reassuringly, abnormal CBC values at our institution are flagged automatically by our analysers, and are manually reviewed by a CLS. This prompts a laboratory requested pathologist PBS review based on the guidelines set by the individual lab (box 1). Furthermore, laboratory requested PBS review flagged by analysers have been shown to have higher sensitivity of detecting novel and clinically relevant information compared with clinician requested pathologist review.9

To understand the clinical need for PBS review in our patient population and the impact of our new policy implemented on 1 September 2016, we reviewed the indications for all PBS orders within our inpatient facilities and outpatient clinics. Finalised reports from 1 April 2016 to 31 January 2017 were grouped into diagnostic categories and generally paralleled clinical indication patterns. Clinical indications for RBC, platelet, and white blood cell morphological evaluation were the most common request for PBS review in descending order, with anaemia and thrombocytopaenia listed as the most common indications. In most cases, pathologist PBS review did not provide additional clinical information beyond that inferred by the CBC and differential count parameters. For example, in cases of anaemia, basic red cell indices provided by the CBC such as haemoglobin, mean corpuscular volume and reticulocyte count could easily be used to make a preliminary assessment of the type of anaemia without a PBS review. Similarly, the CBC and differential count parameters could be used in clinical indications of ‘cytopaenia’ or ‘cytosis’ without a peripheral smear review. With the understanding of clinical need and limitation of PBS reviews, pathology trainees at our institution were educated on PBS approval criteria in order to instruct clinicians, recommend appropriate ancillary testing and prevent duplicate orders (box 1 and table 1). Our new policy reduced unremarkable PBS reviews by screening for appropriate indications and educating clinicians on best testing practices. We believe that our institutional policy change did not impact patient care adversely, since there were no complaints from clinicians after the policy was implemented. Correspondingly, appropriate ancillary testing was recommended in lieu of cases that did not meet indication for a pathologist PBS review. Thus, we believe that the recommendations improved resource utilisation without compromising patient care.

Interestingly, Beckman et al reported that the added clinical value of PBS review did not align with clinician ordering practices and noted that orders for pathologist review for white blood cell morphology had the most added clinical value, yet assessment for RBC morphology was the most commonly requested order. To evaluate the added value of PBS, Sandhaus et al retrospectively interviewed ordering physicians within 48 hours of 64 signed out cases.10 Surprisingly, 51% of ordering physicians replied that they had not seen the results, 30% replied that they had seen the results, 6% did not remember the results, and 13% were contacted directly regarding urgent findings. Possible explanations for poor follow-up within the study included ordering physicians going off service, designating staff to check results, following up results after 2 days, orders from the emergency department when the patient was already discharged, or denial of knowledge of the order.10 Due to lack of urgency and inadequate follow-up of results from ordering clinicians, these studies question the clinical utility of PBS review.

Time for pathologist PBS review is abstractly underestimated. On average, morphologic review of PBS by a pathologist can take approximately 3 min per PBS, and longer when there are abnormal findings. The pathologist then issues a report that includes the clinical history, detailed characterisation of each haematopoietic cell line, summary of findings, final interpretation with differential diagnosis and recommendation of additional tests if needed. Overall, it may take an average of 10–15 min per PBS review to analyse morphological findings, review the patient’s clinical history, and generate a written report. These findings have also been documented in previous published papers.9 11

Education has been shown as an inexpensive effective way of changing practice habits, with increased promising effects when combined with feedback.1 2 In a study to optimise CBC with differential ordering practices within the ICU, Shen et al implemented Choosing Wisely, a 2-week education intervention that recommended replacement of the CBC with differential test with a CBC test only during the first 48 hours of inpatient care in the ICU. This educational session was implemented after initial analysis that indicated that the majority of the frequently repeated CBC with differential testing was being used in the ICU. Educational sessions reminded clinicians of the 2 day half-life of granulocytes and provided ordering algorithms, which were reinforced throughout the unit. This intervention decreased the total number of CBC with differential test orders and increased the total quantity of CBC tests without differential orders. Although educational interventions may only have temporary effects, Shen et al noticed that ordering practices continued well beyond the 2-week intervention and were also unexpectedly adopted by other units.2 While our institution does not provide educational seminars or handouts for PBS review indications, we rely on educating pathology trainees to provide constant verbal feedback and rapport with our outpatient providers, along with automated EMR popup boxes of PBS request guidelines for in-patient providers to promote ongoing lasting change in practice habits.

There were some limitations to our study. Clinical indications and diagnostic categories were obtained retrospectively from review of our EMR. We also did not document the actual diagnostic time saved by ordering recommended tests in lieu of PBS review; however, the pathologists interviewed at our institution have reported notable time savings and it is assumed that the recommendations improves resources for definitive diagnosis. Ordering sets are periodically reviewed during weekly clinical pathology lab meetings when notable trends deviate. From these reviews, PBS smears were removed from linked ER ordering sets, where patients are often discharged before results are released, and all CBC orders. In these scenarios, PBS smears must be requested separately. However, ordering sets containing PBS reviews are not routinely reviewed due our overwhelming volume. This is an intervention that should be addressed in future quality improvement projects to proactively catch misutilisation of PBS orders.12

The request for PBS review reflects an unmet need for clinical consultation to explain unusual laboratory findings and their clinical relevance. Accurate diagnosis of PBS abnormalities requires the integration of CBC values, additional laboratory findings, clinical history and morphological findings that can either lead to diagnosis or guide cost-effective workup. One interesting approach to quickly communicate clinical correlation of abnormal PBS findings was explored by Jaso et al by the use of a web-based synoptic reporting system that generates a detailed templated report generated by selection of PBS findings, which resulted in decreased turnaround time and reduced typing errors.13 This need opens the possibility of developing a venue for a tailored billable pathology consultation report to promote effective communication and improve communication clinical outcomes.

Clinicians may also want reassurance from secondary PBS review of abnormal CBC findings. Educating clinicians about the low rate of missed information provided by advanced analysers and the sensitivity of laboratory-initiated pathology PBS review may provide comfort through the awareness of a built-in mechanism to screen for haematological abnormalities. While our implemented policy provides a venue for clinicians to inquire about best testing modalities for abnormal PBS findings, it may be valuable to extend this practice to advise on best testing modalities for other pathology services. Future studies may look into optimising visibility of results in the EMRs and re-evaluation of ordering sets. In conclusion, adopting polices to request PBS approval reduces the number of noncontributory tests, serves as a beneficial educational exercise of resource utilisation for both trainees and clinicians, and reduces hospital costs.