Improving thromboprophylaxis in the medical inpatients: The role of the resident in an academic hospital

Introduction

Venous thromboembolism (VTE), which includes pulmonary embolism and deep vein thrombosis, represents a major healthcare concern, with an estimated worldwide incidence of 10 million cases per year. ¹ The annual healthcare costs associated with VTE are estimated to be €3 billion in Europe and US$10 billion in the USA. ² It is one of the main causes of preventable in-hospital death, accounting for almost 10% of mortality. ³

VTE incidence is higher in patients with hypercoagulable disorders (such as cancer, pregnancy, obesity, and autoimmune diseases), vascular damage, heart failure, hereditary factors, venostasis, immobilization and elderly people. ⁴ Medical inpatients often have multiple factors that increase VTE risk ⁵ and the majority of the fatal pulmonary embolisms occur in this population. ⁶

VTE risk assessment is reported to be low if no active interventions are implemented. In a study conducted in the United Kingdom, VTE risk assessment was present only in 5% of the patients admitted. ⁷ Only 39% of medical inpatients receive appropriate thromboprophylaxis (TP) compared with 58% of surgical patients as described in the multinational ENDORSE study. ⁸ Other studies report a TP guidelines adherence of 49.4–73.5% in medical patients.^9,10

There are several prediction models for VTE risk stratification in hospitalized patients. Two of the most used are the Padua prediction score (PPS) and the Caprini risk assessment model (RAM) score. The American College of Chest Physicians (ACCP) recommends the use of the PPS for in-hospital medical patients. ¹¹ However, some studies support the use of the Caprini RAM score as the first choice in a hospital environment since it has a higher sensitivity to detect patients at high risk.^12–14

Several strategies have been used to increase VTE risk assessment, including physician training, educational interventions, electronic notifications, weekly multidisciplinary meetings and continuing medical education programs.^15–17

It is recommended for hospitals to have an appropriate TP protocol to reduce morbidity and mortality associated with VTE. ⁵ The use of a risk assessment model (RAM) for VTE, along with other interventions, has demonstrated to reduce the incidence of VTE by half in hospitalized medical patients by increasing the TP guidelines adherence. ¹⁸

The Caprini RAM uses a thorough evaluation of risk factors, accounting for all of the following: age, planned or previous surgery, presence of varicose veins, history of inflammatory bowel disease, swollen legs, overweight or obesity, myocardial infarction, congestive heart failure, sepsis, severe lung disease, mobility restriction, cancer, central venous catheter, personal or family history of VTE, severe trauma, spinal cord injury, stroke, use of estrogen replacement therapy, abortion and recent pregnancy. ¹⁹

As a quality improvement strategy, the Caprini RAM was implemented by the internal medicine resident physicians to medical inpatients in our teaching hospital, giving feedback for treatment adjustment to staff physicians.

The primary objective of this study was to determine the effect of the resident physician feedback to the staff physician in TP guidelines adherence after the implementation of the Caprini RAM score. The secondary objective was to assess the feedback effect in individual VTE risk groups (low, moderate, high, and highest) according to the Caprini RAM score.

Methods

This was a retrospective, observational, comparative, cohort study conducted in a single teaching hospital in Northern Mexico. STROBE guidelines were followed accordingly.

This study was approved by our local ethics committee (REC number 04062019-b-MI-HM-CI). The audit was registered with the hospital quality improvement department. All procedures performed were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments. As this was a quality improvement project and only collected retrospective clinical data were used, it was considered not necessary to seek written consent from patients.

Records from medical patients age 18 years and older admitted to our hospital between May 2019 and June 2021 were included. Exclusion criteria were pregnancy and surgical patients. Data collected were age, gender, BMI, Caprini RAM score and Caprini RAM risk classification (low, moderate, high, and highest), adherence to TP guidelines and change to appropriate TP after feedback.

The audited intervention consisted in the implementation of the Caprini RAM by the resident to all medical inpatients within 24 h of their admission. Score was calculated and patients were categorized in risk stratification groups: low (0–1 points), moderate (2 points), high (3–4 points), or highest (≥5 points). If the indicated TP for a patient did not correspond to the assigned group, the resident would contact the staff physician in charge of the patient to give feedback on risk stratification and recommend the TP adjustment.

TP recommendation adherence was assessed according to the Caprini RAM stratification group. The low-risk group had no need for TP. The moderate-risk group was warranted either mechanical TP with elastic compression stockings, intermittent pneumatic compression device or pharmacological TP with either low molecular weight heparin (LMWH) or unfractionated heparin (UFH). The high-risk group recommendation was pharmacological TP with LMWH or UFH and with or without mechanical TP. The highest-risk group was recommended to have pharmacological TP with LMWH, UFH, and direct oral anticoagulants (DOAC) or adjusted-dose vitamin K antagonists (warfarin or acenocoumarol) and with or without mechanical TP.

Contraindications for pharmacological TP were recent head trauma, active or high risk of bleeding, recent cardiovascular surgery, recent stroke and presence of epidural catheter. Mechanical TP was recommended for these cases in the moderate-to-highest groups.

Statistical analysis

Patient data was deidentified for statistical analysis. We used the Kolmogorov–Smirnov test for normality. We described variables with a normal distribution using mean and standard deviation (SD), whereas median and the 25th and 75th percentiles (p25–p75) were used for the non-normal distribution variables. We used the McNemar test to compare pre- and post-intervention adherence to TP guidelines. A p-value ≤0.05 was considered significant. There were no missing values. Data were analyzed using SPSS version 25.

Sample size was calculated according to proportions inequality, two dependant groups McNemar test. With a two-tailed, odds ratio 5.4 effect size at an alpha of 5%, 80% power and 5.3% of the population changing as a result of the intervention, 265 participants were needed to detect the effect. Simple random sampling was used for the selection of the medical records.

Results

A total of 265 medical records were included. Men represented 54.3% of the cohort. Median age was 50 (36–66) years. Mean (SD) BMI was 27.2 (5.2) Kg/m². Median Caprini RAM score was 2 (1–4). Of these patients, VTE risk was low in 74 (27.9%), moderate in 63 (23.8%), high in 67 (25.3%) and highest in 61 (23%) according to the Caprini RAM. The demographic characteristics, Caprini RAM score and risk stratification are shown in Table 1.

OPEN IN VIEWER

Table 1.

Demographic characteristics and venous thromboembolism risk.

Variable	Result
Total patients, n (%)	265 (100)
Male gender, n (%)	144 (54.3)
Age (years), median (p25–p75)	50 (36–66)
BMI (Kg/m2), mean (SD)	27.2 (5.2)
Caprini RAM score, median (p25–p75)	2 (1–4)
Risk group, n (%)
Low	74 (27.9)
Moderate	63 (23.8)
High	67 (25.3)
Highest	61 (23)

Before intervention, 193 (72.8%) of all patients had appropriate TP; post-intervention, 207 (78.1%) patients received adequate TP, indicating a difference of 5.3% after feedback (OR 5.14 CI 3.21–8.23) (p < .001). When dividing patients by VTE risk stratification groups, none of the 74 patients had inappropriate TP; therefore, required no intervention. In the moderate-risk group, 33 (52.4%) patients had proper TP, with an increase to 37 (58.7%) after feedback (OR 7.5 CI 3.21–8.23). In the highest-risk group, 46 (75.4%) had adequate TP, increasing to 49 (80.3%) after intervention (OR 2.03–7.29). These differences were not statistically significant. However, in the high VTE risk group, appropriate TP was present in 40 (59.7%) patients and in 47 (70.1%) post-intervention (OR 1.81–13.75) (p = .016) (Table 2).

OPEN IN VIEWER

Table 2.

Adequate TP.

Risk group	Pre-intervention	Post-intervention	OR	CI	p
All, n (%)	193 (72.8)	207 (78.1)	5.15	3.21–8.23	<0.001
Low, n (%)	74 (100)	74 (100)	—	—	—
Moderate, n (%)	33 (52.4)	37 (58.7)	7.5	3.01–18.67	0.125
High, n (%)	40 (59.7)	47 (70.1)	3.85	2.03–7.29	0.016
Highest, n (%)	46 (75.4)	49 (80.3)	5	1.81–13.75	0.25
					McNemar

Discussion

This study demonstrated that resident physician feedback significantly increases adequate TP in 5.3%, from 72.8% to 78.3%. In the subgroup analysis, a significant difference was observed only in the high-risk group, with an increase of 10.4%. The role of the resident physicians is relevant to improve VTE risk awareness among internal medicine staff physicians in our academic hospital, successfully improving TP, and their collaboration may be helpful in the prevention of VTE and its associated complications.

After dividing patients by VTE risk, only the high-risk group had a significant difference in TP appropriateness. This may be explained by a higher compliance to TP in patients with the highest risk due to the awareness of a possible occurrence of VTE in a patient with many predisposing factors. An explanation for the absence of a significant difference in the moderate-risk group may be the unwillingness of the staff physicians to give pharmacological TP due to the concern of bleeding complications and the discomfort caused by elastic compression stockings in these patients. ²⁰ An alternative explanation was the small sample size by group.

Many other options for improving TP have been described. In a single-center, retrospective, cohort study, appropriate TP increased in medical patients from the respiratory department with physician training and educational interventions (13% and 27%, respectively). ¹⁶ In another study, the use of e-mail notifications and weekly multidisciplinary meetings improved proper TP in internal medicine patients from 63% to 92.8% after 15 weeks. ¹⁷ A continuing medical education program conducted in Saudi Arabia showed an increase of 35% on proper TP in medical patients. ¹⁵

However, not every single-faceted intervention may increase TP guidelines adherence. In a randomized control study conducted in Switzerland, the implementation of an automated electronic alert system did not improve TP in medical inpatients. ²¹ The possible explanation for this result was that the automated alerts were ignored by physicians, as they were frequent and non-specific. This represents an advantage for human alerts since they can be focused individually for each patient. In our study, the fact that the recommendation is provided by a resident physician instructed in VTE risk assessment may be considered an additional strength.

The extent of the effect in TP guidelines adherence may be further increased when combining several intervention strategies, as described in a review by Kahn et al. Multifaceted interventions that include an alert component (either human or electronic) are more effective in improving TP guidelines adherence than those without it. This strategy also demonstrates a reduction in the rate of VTE after 3 months. ²² Multifaceted interventions implemented in a multicentre cohort study conducted in the USA consisted in the use of automatic electronic alerts, monthly audit-and-feedback assessments, and a continuing medical education activity. An increase of 18% in TP guidelines adherence was reported using multiple interventions compared to electronic alerts only. ²³

National strategies may be implemented to increase VTE risk assessment. In the United Kingdom, a financial incentive was used to increase TP awareness, and a quality improvement project was run in a teaching hospital in Plymouth. Appropriate VTE risk assessment increased from 5% to 95% after 6 years. This led to a relative risk reduction of 14% in VTE related to inadequate TP. ⁷

When comparing feedback from resident physicians to other methods, the difference in TP appropriateness after the intervention is small. This may be explained by the reluctance of the staff physicians to switch to pharmacological TP guided by a RAM originally validated for surgical patients,^11,24,25 although there is evidence that supports its use in medical patients.^12–14 Moreover, it is worth to mention that the PPS, which is recommended for medical inpatients by the ACCP, ¹¹ is based in a single cohort study. ¹² It is possible that multidisciplinary meetings and a continuing medical education program which included the information gathered from these studies may be of use to increase the acceptance rate of the Caprini RAM for medical inpatients among staff physicians in our hospital.

Some strengths of this study should be highlighted. To our knowledge, it is the first study to evaluate the role of the resident physicians in TP quality improvement strategies in an academic hospital. It used the Caprini RAM in medical inpatients, which has a high sensitivity for the detection of VTE risk.

This study has some limitations that must be noted. It is a single-center study and results may not be generalizable. The power after fractioning the sample in VTE risk subgroups decreased by the reduction in the size. Moreover, it did not focus on long-term events, development of VTE nor major bleeding complications.

Conclusions

In conclusion, our study demonstrated that resident physician feedback significantly increases appropriate TP in medical inpatients. Further investigation and prospective studies are necessary to address the relevance of the intervention in the outcomes of this population.