Quantitative Results
All 296 patients who were scheduled for a PAC visit over the 4-week study period were included (Table 1); one patient did not attend their appointment, and therefore 295 PAC visit notes were analyzed (99.7%). The median age was 66 years (IQR 57-72 years) and 188 patients were female (63.7%). At the time of analysis, 283 surgeries were completed (95.9%), 10 were postponed or cancelled (3.4%), and 2 had not been scheduled (0.7%). Nine patients had no recommendations made (3.2%).
Table 1. Characteristics of patients seen in preadmission clinic
|
Characteristic
|
Full Cohort
(n=296) n(%)
|
All Recommendations Followed n(%)
|
Missed Recommendations n(%)
|
p-value*
|
Patients seen in clinic
|
295 (99.7)
|
142 (51.8)**
|
113 (41.2)**
|
|
Age (median, IQR)
|
66 (57-72)
|
64 (52-71)
|
68 (60-74)
|
0.007
|
Female
|
188 (63.7)
|
96 (67.6)
|
65 (57.5)
|
0.016
|
Surgeries completed
|
283 (95.9)
|
-
|
-
|
|
Surgeries postponed
|
6 (2.0)
|
-
|
-
|
|
Surgeries cancelled
|
4 (1.4)
|
-
|
-
|
|
Inpatient surgeries
|
215 (72.9)
|
102 (71.8)
|
86 (76.1)
|
0.44
|
Day surgeries
|
80 (27.1)
|
40 (28.2)
|
27 (23.9)
|
Surgical Discipline
|
|
|
|
General Surgery
|
90 (30.5)
|
45 (31.7)
|
34 (30.1)
|
|
Spine Surgery
|
56 (19.0)
|
23 (16.2)
|
27 (23.9)
|
|
Neurosurgery
|
36 (12.2)
|
13 (9.2)
|
19 (16.8)
|
|
General Gynecology
|
34 (11.5)
|
19 (13.4)
|
10 (8.8)
|
|
Gynecology Oncology
|
30 (10.2)
|
19 (13.4)
|
7 (6.2)
|
|
Otolaryngology
|
15 (5.1)
|
7 (4.9)
|
4 (3.5)
|
|
Plastic Surgery
|
13 (4.4)
|
7 (4.9)
|
3 (2.7)
|
|
Thoracic Surgery
|
10 (3.4)
|
1 (0.7)
|
5 (4.4)
|
|
Orthopedics
|
9 (3.1)
|
6 (4.2)
|
3 (2.7)
|
|
Cardiac Surgery
|
1 (0.3)
|
1 (0.7)
|
0
|
|
Dentistry
|
1 (0.3)
|
0
|
1 (0.9)
|
|
Urology
|
1 (0.3)
|
1 (0.7)
|
0
|
|
*Comparison of patients with all recommendations followed and patients with missed recommendations.
**Percentage is out of all completed surgeries that had at least one recommendation at the time of data analysis, n=274.
Overall, 51.8% of PAC visit notes (n=142) had all recommendations followed. Patients who had missed recommendations were older than patients who had all recommendations followed (median age 68 years (IQR 60-74 years) and 64 years (IQR 52-71 years), respectively; p=0.007) and were less likely to be female (57.5% and 67.6%; p=0.016) (Table 1). General surgery, spine, neurosurgical and gynecologic procedures had the greatest numbers of patients with missed recommendations (Table 1).
Characteristics of recommendations made by internists in the PAC visit note are presented in Table 2. Almost all PAC visit notes contained at least one recommendation (n=274, 96.8%). Recommendations that addressed perioperative management of any home medications were most common (n=258 of 272 eligible notes, 94.6%). There was no difference in the proportion of recommendations followed in the preoperative period compared to the postoperative period (55.4%, n=149; compared with 48.9%, n=109; p=0.15).
Of the patients seen, 38.5% (n=109) were on home anticoagulation or antiplatelet medications. Preoperative recommendations related to home anticoagulation or antiplatelets were followed more often than postoperative recommendations (61.5%, n=64; and 45.2%, n=28; respectively; p=0.03). Nearly one-quarter of patients were on medications for diabetes and most of these patients had recommendations related to the management of diabetes in the perioperative period (n=60 of 67 eligible consult notes, 89.6%). Only 70.0% of recommendations related to perioperative diabetes medications were followed (n=42). Recommendations on management of these medications after surgery were made for less than half of patients (n=32, 47.8%), though these postoperative recommendations were commonly followed (n=29, 90.6%). Nearly three-quarters of patients were eligible for cardiac biomarker screening (referring to preoperative brain natriuretic peptides and/or postoperative troponin measurement) as per the
Canadian Cardiovascular Society guidelines20 (73.5%, n=208). Of these, only 63.0% (n=131) had a recommendation for patients to undergo cardiac biomarker testing, while 64.1% of patients with a recommendation had it completed (n=84).
Table 2. Categories of perioperative recommendations made in PAC visit notes, with proportion of recommendations followed, proportion followed preoperatively, and proportion followed postoperatively.
Recommendation Type
|
Eligible consult notes* n (%)
|
Consult notes with any recommendation n (%)
|
All recommendations followed n (%)
|
Consult notes with preoperative recommendations n (%)
|
Preoperative recommendations followed n (%)
|
Consult notes with postoperative recommendations n (%)
|
Postoperative Recommendations followed n (%)
|
All types
|
283
|
274 (96.8)
|
142 (51.8)
|
269 (95.1)
|
149 (55.4)
|
223 (78.8)
|
109 (48.9)
|
Any home medication
|
272 (96.1)
|
258 (94.6)
|
145 (53.3)
|
258 (94.6)
|
150 (58.1)
|
131 (48.2)
|
67 (51.1)
|
Antiplatelet & Anticoagulation
|
Home medication
|
109 (38.5)
|
104 (95.4)
|
59 (56.7)
|
103 (94.5)
|
64 (61.5)
|
62 (56.9)
|
28 (45.2)
|
DVT prophylaxis
|
209 (73.9)
|
151 (72.2)
|
105 (69.5)
|
n/a
|
n/a
|
151 (72.2)
|
105 (69.5)
|
Opioids
|
Home medication
|
28 (9.9)
|
22 (78.6)
|
18 (81.8)
|
22 (78.6)
|
18 (81.8)
|
4 (7.1)
|
0
|
Withdrawal management
|
28 (9.9)
|
1 (3.6)
|
0
|
n/a
|
n/a
|
1 (3.6)
|
0
|
Diabetes
|
Home medication
|
67 (23.7)
|
60 (89.6)
|
42 (70.0)
|
60 (89.6)
|
44 (73.3)
|
32 (47.8)
|
29 (90.6)
|
Cardiac
|
Cardiac biomarkers
|
208 (73.5)
|
131 (63.0)
|
84 (64.1)
|
n/r
|
n/r
|
131 (63.0)
|
84 (64.1)
|
Home medications
|
174 (61.5)
|
152 (87.4)
|
115 (75.6)
|
152 (87.4)
|
115 (75.6)
|
55 (31.6)
|
36 (65.5)
|
Delirium management
|
49 (17.3)
|
5 (10.2)
|
2 (40.0)
|
n/a
|
n/a
|
5 (10.2)
|
2 (40.0)
|
*Eligible consult notes refer to the number of PAC visit notes that could have included a recommendation in that category, based on the patient's past medical history or home medications. This serves as the denominator for the proportions reported.
n/a = not applicable; n/r = not reported.
Based on these data, the interview guide was adapted to: (1) characterize the process by which PAC visit notes are reviewed and recommendations are entered to identify steps in the process where recommendations may be missed; and (2) understand why inpatient surgical teams may not follow recommendations, especially those related to perioperative cardiac biomarker surveillance, diabetes, and anticoagulation. These specific questions were intended to address the overall aim of this study, which was to understand the number of and reasons for missed recommendations in PAC visit notes. This will allow us to better inform interventions to improve the quality of perioperative care in our centre.
Qualitative Data
We interviewed 18 participants, including 2 nurse practitioners (11.1%), 8 surgical residents (44.4%), 6 surgeons (33.3%) and 2 internists (11.1%) (Table 3). Participants represented neurosurgery (n=4, 22.2%), gynecology (n=4, 22.2%), general surgery (n=3, 16.7%), spine surgery (n=2, 11.1%), orthopedic surgery (n=1, 5.6%), otolaryngology (n=1, 5.6%) and thoracic surgery (n=1, 5.6%). One internist and one surgeon declined to participate due to schedule conflicts.
Table 3. Characteristics of Interview Participants.
Characteristic
|
Number (%)
|
Faculty Status
|
Nurse Practitioners
|
2 (11.1)
|
Residents
|
8 (44.4)
|
Staff Physicians
|
8 (44.4)
|
Discipline of Practice
|
General Surgery
|
3 (16.7)
|
Orthopedic Surgery
|
1 (5.6)
|
Neurosurgery
|
4 (22.2)
|
ENT Surgery
|
1(5.6)
|
Gynecology
|
4 (22.2)
|
Spine Surgery
|
2 (11.1)
|
Thoracic Surgery
|
1 (5.6)
|
Internal Medicine
|
2 (11.1)
|
ENT = otolaryngology
We used 31 themes in the final codebook (Appendix 4). Twenty-five codes were based on constructs from CFIR and TDF and 6 were developed through inductive analysis. We separated social context and networks into two subconstructs (peer pressure and social influences, and external networks and internal networks, respectively) to discriminate between similar but different ideas. The six codes developed inductively captured specific medical topics (for example: anticoagulation, diabetes), critiques or suggestions for improvement of the PAC visit note, and differing ideas of the intention of the PAC visit note. Relevant constructs are listed in brackets in the reporting of results.
We categorized reasons for missed recommendations into a conceptual framework which delineates between different kinds of unintentionally and intentionally missed recommendations (Figure 1). Our qualitative data shows that surgical teams may unintentionally miss recommendations due to both individual-level and systems-level factors. Intentionally missed recommendations can be divided into appropriate modifications (recommendations which surgical teams did not follow because they were not safe or reasonable for their patients) and user error (recommendations which were not followed due to misinterpretation or lack of knowledge of evidence and best practices). Additional supportive quotes are provided in Table 4.
Unintentional Missed Recommendations
Unintentional missed recommendations occurred when participants were unaware of PAC visit recommendations because they did not read the PAC visit note. Many participants expressed surprise at the proportion of recommendations that were not followed by surgical teams.
Individual-level drivers of missed recommendations
Participants from all groups reported that the most junior person on the surgical team was responsible for reading the PAC visit note and entering relevant orders after the end of the surgical case; this was the resident in most instances (Social Identity and Role).
We found that surgical team members were often unaware that patients may have had a PAC visit note (Knowledge) (Figure 1). One neurosurgery resident stated "It wasn’t right away in my training [that I was told about the PAC visit note], it wasn’t something that was explicitly told to us, that, to look for when you started putting post-op orders... I wouldn’t look, and I didn’t know they were there [until] probably mid-way through first year" (P04).
There was variability between residents regarding whether they had an established routine or process for reading the PAC visit note (Behaviour Regulation) (Figure 1). The absence of an established routine acted as a barrier to implementing recommendations while an established routine was a facilitator of implementing recommendations. The process of reviewing all surgical patients for a PAC visit note was able to overcome a lack of knowledge of which patients had been in PAC. For example, one surgical resident stated "Pretty much for all of our [elective] patients, I’ll ... see if they’ve been seen [by PAC internists], and if so then I typically try and implement most of the suggestions that they’ve recommended... I actually usually look at them either the night before or the morning of" (P10). Some surgical team members reviewed the PAC visit note only if they had time, while other residents always reviewed the PAC visit note (Relative Priority). Overall, this suggests that a routine of always reviewing the PAC visit note was the most important individual-level driver of missed PAC recommendations.
Systems-level drivers of missed recommendations
Similar to individual processes to view PAC visit notes and enter recommendations, surgical team members reported inconsistent systems-level mechanisms to ensure that the PAC visit note had been read and recommendations had been followed (External and Internal Networks) (Figure 1). Team-based processes where orders were consistently reviewed to ensure all recommendations had been ordered were considered a facilitator of implementing PAC recommendations, and when this process was undefined, it was considered a barrier.
Some surgical teams relied on allied healthcare team members, including charge nurses and pharmacists, to notice omissions in the orders. As one neurosurgery resident stated, "No [one is checking orders], not from the surgical team perspective... the orders [are] put in after surgery... and most charge nurses will review and say GIM said to start this medication after 48 hours" (P04). Other surgical teams had a defined chain of responsibility that involved only physicians, where the more senior residents were double-checking the work of junior residents who entered orders.
While some staff surgeons reported reviewing the PAC visit note, none endorsed checking the postoperative orders to ensure that PAC recommendations had been followed and many stated that they never checked postoperative orders entered by residents. One staff neurosurgeon stated, "If I had to double check every drug [or] order a resident put in, I wouldn't get anything else done" (P18). Contrary to resident self-report, staff surgeons believed that the residents were consistently using the PAC visit note to guide postoperative order entry. This suggests a disconnect between the reported behaviours of surgical residents and the perceptions of staff surgeons, which may account for some unintentionally missed PAC visit note recommendations.
Drivers of Intentional Missed Recommendations
We classified intentional missed recommendations into user error, referring to the decision to not follow a recommendation based on a mistake in judgement, or appropriate modifications, when surgical teams chose not to follow a recommendation because of changing patient status (Figure 1). In addition, we observed that themes that explained the missed recommendation clustered by the medical topic of the recommendation.
User Error
Participants did not follow postoperative diabetes recommendations because of lack of skills to manage and titrate insulin, even when recommendations were specific about dosages and regimens (Skills) (Figure 1). Residents felt that "we’re so not used [to prescribing] insulin, we’re worried [about] giving little or too much, and then we don’t really have the time or the capacity to make some fine adjustments... and patients are further more complicated, their oral intake is all messed up, so [diabetes] is... the thing that we’re struggling the most with" (P01, general surgery resident). Participants relied on sliding scale insulin to get patients through the postoperative period due to increased comfort with this regimen compared to basal bolus insulin therapy, even when there were PAC visit note recommendations to use basal bolus insulin therapy.
Members of the surgical team did not always follow recommendations for postoperative troponin surveillance because they did not understand the rationale of testing or benefits to patients (Knowledge) (Figure 1). For example, "Some people don’t feel that troponin monitoring and the interventions actually result in changes to patient’s [management]... so even when we monitor them and [the troponin] go[es up] by five points, it’s... not like it’s a real problem" (P06, nurse practitioner).
Appropriate modifications of recommendations
Internists reported difficulty providing postoperative recommendations for patients because of anticipated physiologic changes after surgery and the difficulties in anticipating all the possible complications (Adaptability) (Figure 1). One internist stated "It’s not realistic [to provide postoperative recommendations]... if they need help after the surgery, they should call for help... but to guess all the pathways that the patient could go down from what the problems may or may not be, what the surgeons may or may not be comfortable with... there are too many variables" (P15, internist). Surgical residents agreed, stating "from our standpoint, our postop patients are not... the same person they were three days before surgery" (P12, gynecologic oncology resident).
Unclassifiable
Intentionally missed recommendations related to perioperative anticoagulation management were intentionally not followed when surgical team members did not agree with the evidence-base for the recommendations (Evidence Strength and Quality) (Figure 1). This occurred when surgical team members reported that the guidelines used by internists to make recommendation did not apply to their specific surgical discipline based on their evaluation of the supporting evidence. In these cases, it is not possible to make a statement about which course of action is safest for patients.
Table 4. Variability in the reported and perceived processes for reviewing the PAC visit note by residents, nurse practitioners, and staff surgeons.
Driver of Missed Recommendations
|
Exemplar Quotation
|
Unintentionally missed recommendations
|
Individual-level drivers
|
Knowledge of the PAC visit note
(Barrier)
|
"It's hard to know who was seen by internal medicine before and who was not" (P06, nurse practitioner)
|
Behaviour Regulation
(Facilitator)
|
"Certainly before we’re operating on any patient I look to see if there’s [a PAC visit] note on [our EHR], and so if there is, I read it... it took me a while to make it into a habit where I like will look at it before every patient, to make sure there’s anything I need to do"
(P08, surgical resident, ENT)
|
Behaviour Regulation
(Barrier)
|
"It’s just not in [the resident's] routine practice [to look at the PAC visit note], cause not everyone has preop consults… they just have their plan of postop orders, and they just put them in without looking, even for patients that go to the OR from the unit, they just do the same thing all the time"
(P06, nurse practitioner)
|
Relative Priority
(Barrier)
|
"When we finish a surgery [and] we put in a post-op order, [that's] typically the first sense we’ll get to as to how medically complex they are. Having said that, when you have time in the evenings, like at the Sunday before your week it is nice to be able to sit down and look through the cases that you have coming for the week, but that takes time and is not always something that’s realistic"
(P04, surgical resident, neurosurgery)
|
Systems-level drivers
|
External and Internal Networks
(Barrier)
|
"No, I mean not any more than any other service where there’s fellows and senior residents and juniors residents and everyone’s is part of the team, and everyone’s trying to do what’s best for the patient, so we’re all looking at things, we’re all trying to you know help each other out, but there’s no, I don’t think there’s a formal process unless the chief resident has made it a priority for him or herself to go through and say I’m gonna look at every preop assessment and make a point of it, there’s no like safety check"
(P12, obstetrics and gynecology)
"Not terribly often, no... if I had to double check every drug [or] order a resident put in I wouldn’t get anything else done,"
(P18, neurosurgery)
|
External and Internal Networks
(Facilitator)
|
"The junior resident puts the orders, but I look up the medicine consult and still review it, ok they asked for troponins, they did this and that, so don’t forget about this and that...then [the fellow] double [checks], he looks at it too"
(P01, general surgery resident)
"No, there’s not [a mechanism to check if orders are entered], if it’s a medication we have a pharmacist on our unit... I would say she’s like a safety net"
(P17, gynecologic oncology staff)
|
Intentionally missed recommendations
|
User error
|
Skills (Barrier)
|
"I know you’re supposed to do basal bolus insulin... but the concern would be like how do you titrate it, how do you get them off of it to go home, all those things and so, I just use the sliding scale to like bring them down if they’re high and then let them ride it out"
(P06, nurse practitioner)
|
Knowledge of guidelines and
evidence (Barrier)
|
"We are not understanding whether all those [troponin] suggestions apply to all patients and then who should act [on abnormal results]... [implementation] initially was a little bit vague... I can’t do this so I’m gonna ignore it, and that’s maybe not the best thing for people overall"
(P14, staff spine surgery)
|
Appropriate modifications
|
Adaptability (Barrier)
|
"A lot of times their recommendations tend to be to prevent that from happening, so by that point the cat’s kind of out of the bag in terms of what they thought when the patient was well two weeks prior to surgery... maybe [their recommendation] doesn’t apply anymore, we need a more updated plan"
(P04, neurosurgery resident).
|
Unclassifiable
|
Evidence Strength & Quality
|
"Things like Xarelto, these other blood thinners, I tell patients different instructions than internal medicine simply based on experience, not evidence"
(P14, neurosurgery staff)
|