Physician and Surgeon Communication Assessed via the Pathology Requisition in a Regional Laboratory Over Ten Years

Background Ineffective communication between healthcare providers is a known risk factor for adverse events. Objective The aim of this study was to retrospectively assess the communication with pathology via an analysis of the information provided on the pathology requisitions over ten years. Methods All in-house surgical specimens and all non-gynecologic cytopathology specimens accessioned from 2011 to 2020 were retrieved at a regional laboratory. Cases with any clinical information were deemed to have a clinical history present (CHP). CHP was tabulated by submitting physicians/surgeons (SPS), hospital site, year, and tissue group. Results The study period contained 554,817 relevant pathology reports, of which 553,966 could be extracted. The overall CHP rate was 74% and varied from 76% to 67% over the study period. SPSes submitting ≥200 cases (n=314) had a mean/median/standard deviation/max/min CHP rate of 81%/92%/23%/100%/5%. The CHP varied between four hospital sites, from 53% to 97%. CHP varied from 61% to 99% by tissue group. Conclusions CHP is associated with several factors and appears to depend on the hospital culture, specialty, and individual physician/surgeon. The pathology requisition is a way to measure and track the communication that is clinically relevant. Improving communication with pathologists/the pathology department will likely require process changes and mandates. Hospital and laboratory accreditation bodies should consider effective communication with pathology a marker of quality and an accreditation issue.


Introduction
A large mature body of evidence demonstrates that effective communication is a key factor in managing complex environments, such as aviation and space travel [1][2][3]. Thus, it is not surprising that communication failures between healthcare providers (in complex care environments) are a welldocumented cause of adverse events [4]. Accordingly, communication with pathology is also recognized as important and considered a quality metric [5].
A large multi-institution study demonstrated that pathologists need information from the clinic to arrive at an accurate diagnosis [6]. In the legal context, submitting physicians and surgeons may be required to provide a relevant history (as in Ontario [7]), and it may even be specified what information should be supplied within a consultation request (as in Poland [8]). 1 2, 3 1 1 4

A local perspective and assessment
A prior local project examined clinical history provision in the context of prostate core biopsies and found significant variation among submitting physicians/surgeons [9]. An earlier version of this work was presented in poster form at the European Congress of Pathology in 2019 in Nice, France. There appears to be a significant difference between actual practice and the known fact that suboptimal communication is associated with suboptimal care.

Objective of study
The goal of this study is to objectively assess communication with pathology in several hospitals served by a regional pathology laboratory over several years. Additional goals are to investigate whether hospital site, department (as assessed by the surrogate of 'tissue group'), and individual physician/surgeon are predictors of clinical history present.

Materials And Methods
Approval from the Hamilton Integrated Research Ethics Board (HiREB; approval no. 4879) was obtained to extract all pathology reports over a period of ten years and examine the provided clinical history in relation to the healthcare providers. The data was fully anonymized with respect to patient identifiers and healthcare providers. Consent was not obtained as the data were completely anonymized with regard to patient identifiers and healthcare providers.
The anonymized data was fed into a custom program written in Python (https://www.python.org/) that classified cases as clinical history present (CHP) and categorized them into one or more tissue groups, based on a dictionary that was created for that task (Appendix A). A further custom program in Python outputs the anonymized data in a coded format readable by R (https://cran.r-project.org; RStudio, Boston, MA). Subsequently, the tabulations and plots were made with R.
The clinical history present (CHP) status was determined by searching for in-house standardized phrases ('Not available on requisition' for surgical cases and 'Clinical history not provided' for cytology cases) and (if applicable) examining the length of clinical history. If one of the standardized phrases was found the case was marked as 'NAOR' (not available on requisition). As some reports did not have a clinical history section at all: cases that were not 'NAOR' were further examined to determine the length of the clinical history section (number of characters). All cases where the clinical history had a zero-character length were labelled 'no information available' (NIA). All cases that were 'NAOR' or 'NIA' were labelled 'incomplete'. All cases not 'incomplete' were deemed clinical history present (CHP).
The tissue group (TG) was meant to be a rough surrogate for the originating (clinical) department; thus, each case was only assigned to one TG in the post-processing in R. In this context, cases were arranged in a hierarchy (Appendix B). 'Cytology' was set very high in the hierarchy to separate these from the surgical cases.
Case complexity was captured using the number of tissue blocks and the workload units (L4E 2018), as described in a prior study [10]. Each specimen was categorized into one of four mutually exclusive categories ('biopsy', 'resection', 'ambiguous', 'unknown') using a string-matching algorithm.
The regional laboratory (Hamilton Regional Laboratory Medicine Program) consists of multiple hospital sites. Pathology cases were identified by the hospital site in which the specimen was obtained/accessioned Submitting physicians/surgeons were assigned to a hospital site on the basis of the hospital site where the largest number of specimens they submitted came from/were accessioned. It should be noted that the term 'submitting physician/surgeon' in this study corresponds to the field 'requesting physician' on the pathology report; invariably, this field on the pathology report is synonymous with the term 'most responsible physician [or surgeon]'. The hospital sites were anonymized in the results.

Results
The pathology reports were successfully extracted for the study period; however, it was noted that the default report format did not contain any history for gynecologic cytopathology cases. This discovery led to a change in the process such that the clinical history is now provided; however, the data were not re-extracted for the purpose of this study. Thus, the gynecologic cytopathology cases were excluded from the analysis.
The study period (2011-2020), after excluding gynecologic cytopathology cases, contained 554,817 relevant pathology reports, and 553,996 of these could be extracted. The no information available (NIA) cases (cases without a 'clinical history' section) numbered as high as 3,815 per year. To draw attention to the absent clinical history, a change within the pathology department was instituted in 2017; the phrases 'not available on requisition' (surgical cases) and 'clinical history not provided' (cytology cases) were used as a space holder if no history was provided. The 'NIA' cases subsequently decreased to 44 out of 46,652 in the year 2020 (trend in Appendix-Supplemental Figures).
In the study group, 553,960 reports could be classified. Over the study period, 410,147 cases (74%) were CHP. CHP varied significantly by year ( Figure 1) and decreased at the end of the study period.

FIGURE 1: Overall clinical history present rate by year
The hospital sites 1, 2, 3, and 4 had 242,563, 123,041, 137,720, and 50,636 cases over the study period, respectively. A sub-analysis by hospital site ( Figure 2) showed that significant differences existed between the different hospital sites and that there was a significant decrease in the overall CHP rate. The data presented in Figure 2 suggest that the overall CHP rate decrease can be attributed primarily to changes at 'Site 1'. 2022      CHP: clinical history present, CVS: cardiovascular system, Unknown: case could not be classified but 'source of specimen' retrieved, Unclassified: 'source of specimen' not found (n=35) or unclassified in categories above (e.g., 'c_mass').

Discussion
The provision of clinical history can be assessed via the pathology requisition and it is apparent that it varies due to multiple factors. The individual submitting physician/surgeon, hospital site, and tissue type are all strong predictors. It should be noted that the SPS and hospital sites are not independent, as most SPS predominantly work at one of the institutions.
A limitation was that the threshold for CHP was very low in this study. Further analysis work could be done on the basis of the clinical history length. A partial analysis was done on the content of the clinical history using a large dictionary of terms. This analysis was put aside, as the approach was computationally expensive. More work remains on assessing the quality of the provided information within the context of the case, in order to assess whether it is relevant, important, and correct, based on what should be known to optimally manage the case. A further limitation is that the tissue group is a crude approximation of the department. A more complex analysis might attempt to (1) classify cases by string-matching, (2) assign each SPS to a tissue group, (3) re-construct the departments from the SPS' primary tissue group, and (4) calculate the CHP by the department from the sum of the SPS assigned to the tissue group. The individual that added the clinical history information (if present) cannot be ascertained in this study; the information was not recorded on the pathology report. This also represents a limitation in the analysis. It is possible that the communication between the submitting physician/surgeon and their delegate(s) is suboptimal. Further analysis on the communication with the submitting physician/surgeon and their delegate(s), using other information sources, may be informative.
The CHP has decreased with time, despite some efforts to raise awareness about the issue and the issue rising to the level of the hospital's central quality committee ('Medical Advisory Committee'). The time period also encompassed a major transition in 'Site 1's' hospital's electronic patient records. In the context of the transition, the proposal to make the provision of clinical history mandatory was not adopted. This was done despite a process change that would potentially change the hospital's liability: all health professionals (e.g., nurses) delegated to submit a pathology case would henceforth be documented in the patient record, thus better establishing the pathology case chain of custody.
It is said that there are three big questions in pathology: (1) What is it (biopsies)? (2) Did I get it all (resections)? (3) Did I get the right thing (everything else)? The lower CHP rate in biopsies (69%) compared to excisions (77%) is a concern; when the clinical history is most likely to be needed, it is more commonly absent.
Lack of a relevant clinical history can make the triage of cases difficult in pathology. In busy practice environments with a minimal reserve service capacity, it is more likely that in busy times, urgent cases are inappropriately triaged and may be delayed. Lack of a relevant clinical history can affect how specimens are handled within the laboratory. For example, if the clinical history is a 'mass lesion', deeper cuts are prepared immediately. In liver biopsies, if the indication is suspected medical liver disease, special stains are done.
The CHP rate found (74%) is low in relation to a prior large-scale multi-institution study that included 1,004,115 specimens; the prior study found that the clinical history was missing in 2.4% of cases [11].

Input and output mismatch
Pathologists (via organizations like the College of American Pathologists) have worked on creating a workflow conducive to higher quality and more complete reports [12]. Pathologists have done much on the 'output' side of the lab; these have increased completeness [13] and standardized communication [14]. Communication from pathology, in relation to reporting, is regulated (in California) to allow the systematic rapid analysis of health data [15,16].
Improvements on the 'input' side of the laboratory are overdue and ideally should be standardized into a structured form with elements relevant to the particular submitted specimen, so it mirrors the 'output' side of the laboratory, that is, increasingly synoptic type pathology reports.

Objective continued measurement
The computerization of medical records presents ongoing challenges and opportunities in communication with pathology [17]. Measuring communication is becoming easier with the further computerization of medical records. Adding more measurements and refining them will be essential for improving communication. In the absence of continuous measurements and the assessment of communication, there can be no evidence-based dialogue about how to improve communication and what is effective to bring about sustained improvement.
How the crude measurement of clinical history completeness herein is related to hard outcomes (number of medical legal actions, patient harm) remains to be established. Prior work does show that a deficient clinical history leads to delayed reporting and increased work in pathology [6].

Clinical History is an Externality
An externality is a cost (or benefit) that is borne by a third party. In our practise environment, submitting physicians/surgeons are compensated by a single (government) payer for the services rendered. Pathologists in this context are third parties, and the clinical history is an externality; the presence or absence of a provided clinical history does not affect the SPS compensation.
Since providing a clinical history requires time, there is a significant disincentive to doing so. Costs associated with a deficient history are primarily borne by the pathologist. An adverse event that is associated with a deficient history is, primarily, a cost to the patient. Additional costs may be incurred by physicians/surgeons that treat the sequelae of an adverse event.
Indirectly, all medico-legally insured physicians/surgeons pay for deficient histories that lead to medicolegal insurance payouts. Thus, submitting physicians/surgeons who judiciously and consistently supply a clinical history are paying for individuals who do not-through higher medico-legal insurance rates.

Power Dynamic and Standards
The state of communication with pathology is likely a reflection of referral patterns and the power dynamic in physician/surgeon and pathologist interactions. Submitting physicians/surgeons often decide where to send the specimens they generate. Thus, pathologists (and pathology practices) may face repercussions when asking for a complete relevant history. In the context of competition between pathology practices, the provision of clinical history may be a casualty of catering to the customer. Accreditation standards that require a clinical history (with audits to ensure the provided information from the submitting physician/surgeon is relevant and sufficient) would break a dynamic that likely leads to a lower standard of care with adverse effects for patients.
Ideally, submission of a specimen to pathology without a clinical history should not be possible (in the context of electronic physician/surgeon pathology order entry), as is the case in one of the hospital sites for requested radiology tests.

Conclusions
The provision of clinical history can be measured from the pathology requisition. Whether information is provided can be predicted by the hospital site, individual submitting physician/surgeon, and type of specimen.
Improving communication with pathology will likely require process changes and mandates. Hospital and laboratory accreditation bodies should consider communication with pathology a quality metric and make accreditation contingent on a minimum standard of communication from submitting physicians/surgeons.