Prediction of Next Glucose Measurement in Hospitalized Patients by Comparing Various Regression Methods: Retrospective Cohort Study

BACKGROUND

Continuous glucose monitors have shown great promise in improving outpatient blood glucose (BG) control; however, continuous glucose monitors are not routinely used in hospitals, and glucose management is driven by point-of-care (finger stick) and serum glucose measurements in most patients.

OBJECTIVE

This study aimed to evaluate times series approaches for prediction of inpatient BG using only point-of-care and serum glucose observations.

METHODS

Our data set included electronic health record data from 184,320 admissions, from patients who received at least one unit of subcutaneous insulin, had at least 4 BG measurements, and were discharged between January 1, 2015, and May 31, 2019, from 5 Johns Hopkins Health System hospitals. A total of 2,436,228 BG observations were included after excluding measurements obtained in quick succession, from patients who received intravenous insulin, or from critically ill patients. After exclusion criteria, 2.85% (3253/113,976), 32.5% (37,045/113,976), and 1.06% (1207/113,976) of admissions had a coded diagnosis of type 1, type 2, and other diabetes, respectively. The outcome of interest was the predicted value of the next BG measurement (mg/dL). Multiple time series predictors were created and analyzed by comparing those predictors and the index BG measurement (sample-and-hold technique) with next BG measurement. The population was classified by glycemic variability based on the coefficient of variation. To compare the performance of different time series predictors among one another, R2, root mean squared error, and Clarke Error Grid were calculated and compared with the next BG measurement. All these time series predictors were then used together in Cubist, linear, random forest, partial least squares, and k-nearest neighbor methods.

RESULTS

The median number of BG measurements from 113,976 admissions was 12 (IQR 5-24). The R2 values for the sample-and-hold, 2-hour, 4-hour, 16-hour, and 24-hour moving average were 0.529, 0.504, 0.481, 0.467, and 0.459, respectively. The R2 values for 4-hour moving average based on glycemic variability were 0.680, 0.480, 0.290, and 0.205 for low, medium, high, and very high glucose variability, respectively. The proportion of BG predictions in zone A of the Clarke Error Grid analysis was 61%, 59%, 27%, and 53% for 4-hour moving average, 24-hour moving average, 3 observation rolling regression, and recursive regression predictors, respectively. In a fully adjusted Cubist, linear, random forest, partial least squares, and k-nearest neighbor model, the R2 values were 0.563, 0.526, 0.538, and 0.472, respectively.

CONCLUSIONS

When analyzing time series predictors independently, increasing variability in a patient's BG decreased predictive accuracy. Similarly, inclusion of older BG measurements decreased predictive accuracy. These relationships become weaker as glucose variability increases. Machine learning techniques marginally augmented the performance of time series predictors for predicting a patient's next BG measurement. Further studies should determine the potential of using time series analyses for prediction of inpatient dysglycemia.

Insulin Dosing and Glycemic Outcomes among Steroid-treated Hospitalized Patients

OBJECTIVE

To determine the optimal insulin-to-steroid dose ratio for attainment of glycemic control in hospitalized patients.

METHODS

We retrospectively studied data collected from the electronic health record within an academic medical center from 18,599 patient-days where patients were treated concurrently with insulin and steroids. Multivariate logistic regression analyses, which included demographic and clinical variables, assessed the relationships between the exposures of total and basal insulin-to-steroid ratios and the outcomes of glycemic control (all blood glucose readings on the following patient-day >70 and ≤180 mg/dl) and hypoglycemia within three subgroups of steroid dosing: low (≤10 mg prednisone equivalent dose [PED]), medium (>10 to ≤40 mg PED), and high (>40 mg PED).

RESULTS

Increased insulin-to-steroid ratio was associated with increased odds of both glycemic control and hypoglycemia. The optimal total insulin-to-steroid ratio for attaining glycemic control was 0.294 units/kg/10 mg PED in the low-dose subgroup, 0.257 units/kg/10 mg PED in the medium-dose subgroup, and 0.085 units/kg/10 mg PED in the high-dose subgroup. The optimal basal insulin-to-steroid ratio was 0.215 units/kg/10 mg PED in the low-dose subgroup, 0.126 units/kg/10 mg PED in the medium-dose subgroup, and 0.036 units/kg/10 mg PED in the high-dose subgroup.

CONCLUSIONS

Increasing insulin-to-steroid ratios are positively associated with glycemic control and hypoglycemia. Our study suggests that ∼0.3 units/kg/10 mg PED is an optimal dose for low and medium dose steroids, while ∼0.1 units/kg/10 mg PED is optimal for high dose steroids. Further prospective studies are needed to identify insulin regimens that will optimize glycemic control in steroid treated patients while minimizing hypoglycemia risk.

Development and validation of a machine learning model for classification of next glucose measurement in hospitalized patients

BACKGROUND

Inpatient glucose management can be challenging due to evolving factors that influence a patient's blood glucose (BG) throughout hospital admission. The purpose of our study was to predict the category of a patient's next BG measurement based on electronic medical record (EMR) data.

METHODS

EMR data from 184,361 admissions containing 4,538,418 BG measurements from five hospitals in the Johns Hopkins Health System were collected from patients who were discharged between January 1, 2015 and May 31, 2019. Index BGs used for prediction included the 5th to penultimate BG measurements (N = 2,740,539). The outcome was category of next BG measurement: hypoglycemic (BG  ≤  70 mg/dl), controlled (BG 71-180 mg/dl), or hyperglycemic (BG > 180 mg/dl). A random forest algorithm that included a broad range of clinical covariates predicted the outcome and was validated internally and externally.

FINDINGS

In our internal validation test set, 72·8%, 25·7%, and 1·5% of BG measurements occurring after the index BG were controlled, hyperglycemic, and hypoglycemic respectively. The sensitivity/specificity for prediction of controlled, hyperglycemic, and hypoglycemic were 0·77/0·81, 0·77/0·89, and 0·73/0·91, respectively. On external validation in four hospitals, the ranges of sensitivity/specificity for prediction of controlled, hyperglycemic, and hypoglycemic were 0·64-0·70/0·80-0·87, 0·75-0·80/0·82-0·84, and 0·76-0·78/0·87-0·90, respectively.

INTERPRETATION

A machine learning algorithm using EMR data can accurately predict the category of a hospitalized patient's next BG measurement. Further studies should determine the effectiveness of integration of this model into the EMR in reducing rates of hypoglycemia and hyperglycemia.

A Qualitative Study of the Barriers and Benefits to Resident Education in Ambulatory Surgical Centers

OBJECTIVE

As Ambulatory Surgical Centers (ASCs) become more common in academic medical centers, large hospital systems must determine how to shift resident education from inpatient to outpatient surgical centers. This study aims to define stakeholders' views regarding the integration of surgical residents into ASCs.

DESIGN

Long-form interviews lasting 30 to 60 minutes were conducted. Interviews were hand-transcribed and analyzed by qualitative analysis to determine benefits of learning in ASCs for residents, challenges that arise from integrating residents, and recommendations to improve resident incorporation.

SETTING

Interviews were conducted using a video conferencing platform.

PARTICIPANTS

Residency program directors, attending surgeons, graduate medical learners, and a nursing manager were interviewed. Twenty-one total interviews were conducted, representing ten different departments.

RESULTS

Stakeholders agreed that residents benefit from being placed in ASCs because the fast, surgical pace allows the residents to engage in more cases. However, different stakeholders highlighted different challenges, all centered around the notion of inter-stakeholder conflict due to conflicting priorities among residents, attending physicians, and administration. Likewise, recommendations differed by stakeholder group-faculty members sought more defined learning objectives and enhanced communication, whereas residents desired that ambulatory surgical time be more structured.

CONCLUSIONS

Despite the pressures of rapid case turnover, stakeholders agreed that there are many benefits to resident education in ASCs. Findings related to challenges and recommendations support the need to strengthen communication between stakeholder groups and better plan for resident integration into ASCs.

Patient Navigator and Community Health Worker Attitudes Toward End-of-Life Care

Background: There are racial/ethnic disparities in hospice use and end-of-life (EOL) care outcomes in the United States. Although the use of community health workers (CHWs) and patient navigators (PNs) has been suggested as a means of reducing them, CHW/PNs' attitudes toward a palliative care philosophy remain unknown. The purpose of this study was to examine how personal attributes affect a CHW/PN's attitude toward EOL care. Methods: CHWs/PNs were recruited from two state-wide organizations and invited to complete an online survey. We collected information on demographics, attitudes toward the palliative care philosophy, and comfort with caring for patients at the EOL. Results: Of the 70 CHWs/PNs who responded to the survey, 82.5% identified as female, 56.4% identified as black, and 56.2% had a four-year college degree or higher. The mean score on a validated scale to assess attitudes toward EOL care was 33.5 (SD = 4.9; possible range, 8-40). Eighty percent strongly agreed or agreed with being open to discussing death with a dying patient. Higher self-efficacy scores were associated with more favorable attitudes toward hospice (r = 0.306, p = 0.016). Conclusions: CHWs/PNs have an overall favorable attitude toward the palliative care philosophy and may be inclined to providing EOL care.