Impact of a machine learning based decision support for Urinary Tract Infections: Prospective observational study in 36 primary care practices (Preprint)

2021 ◽  
Author(s):  
Willem Ernst Herter ◽  
Janine Khuc ◽  
Giovanni Cinà ◽  
Bart Knottnerus ◽  
Mattijs E. Numans ◽  
...  
Keyword(s):  
Machine Learning ◽  
Primary Care ◽  
Treatment Success ◽  
Treatment Practices ◽  
Significant Difference ◽  
Implementation Period ◽  
Matching Procedure ◽  
Primary Care Practices ◽  
Tract Infections ◽  
Prescription Behavior

BACKGROUND There is increasing attention for machine learning based clinical decision support systems (CDSS), but their added value and pitfalls are very rarely evaluated at the frontline of clinical practice. We have implemented a CDSS to aid general practitioners (GP) with the treatment of patients with urinary tract infections (UTI). UTIs are a large health burden worldwide and the scientific evidence for clinically effective treatments with increased risk of a complicated UTI is limited. OBJECTIVE In this study, we prospectively assess the impact of this CDSS on treatment success and change in antibiotic prescription behavior of the physician. Doing so, we hope to identify drivers and obstacles for positively impacting the quality of healthcare practice with machine learning. METHODS The CDSS was developed as a joint effort by Pacmed, Nivel and LUMC. The CDSS presents the expected outcomes of treatments together with context information needed to assess the expected outcomes well. Treatment success was defined as a subsequent period of 28 days where no new antibiotic treatment for the UTI was needed. In this prospective observational study, 36 primary care practices used the software for a period of four months, starting in November 2017. Twenty-nine control practices were identified through a propensity score matching procedure. All analyses have been done on electronic health records from the Nivel Primary Care database. Patients for which the software has been used have been identified in the Nivel database through a sequential matching procedure using the CDSS usage data. To evaluate treatment success, we have compared the proportion of successful treatments prior and during the study within the treatment arm. The same analysis has been done for the control practices and for the subgroup of patients we were sure of the software has been used for. All analyses were statistically tested by two-sided z-tests with an alpha level of .05. In assessing the difference of treatment success for several patient subgroups, Bonferroni corrections were applied. Lastly, the antibiotic prescription behavior of the physicians was analyzed through the same z-tests. RESULTS In the treatment practices, 4998 patients were included in the period of time before the implementation study, 3422 patients were included during the implementation period. In the control practices, 5044 patients were included before the implementation period, 3360 patients were included during. The proportion of successful treatments increased significantly from 75% to 80% on average in the treatment practices (z=5.47, p<.001). In the control practices, no significant difference was detected (76% before and 76% during the pilot, z=0.02, p=0.98). We have been able to identify 734 out of 1200 patients in the CDSS usage database in the Nivel database. For these patients, of whom we are certain the software has been used for, the proportion of successful treatments during the study was 83%. This is a statistically significant difference with the 75% of successful treatments prior to the study in the treatment practices (z=4.95, p<.001). CONCLUSIONS The introduction of the CDSS as intervention in the 36 treatment practices was associated with statistically significant improved treatment success. We have excluded temporal effects and validated the result with the subgroup analysis in patients for whom we are certain the software was used. The study shows important strengths and points of attention for the development and implementation of a machine learning based CDSS in clinical practice. CLINICALTRIAL The trial was registered on ClinicalTrials.gov under the Identifier NCT04408976.

OP08 Using Real World Data To Support National Postmarketing Surveillance

2018 ◽  
Vol 34 (S1) ◽  
pp. 3-3
Author(s):  
Yeqin Zuo ◽  
Bernie Mullen ◽  
Rachel Hayhurst ◽  
Karen Kaye ◽  
Renee Granger ◽  
...  
Keyword(s):  
Primary Care ◽  
Real World ◽  
National Level ◽  
Clinical Information ◽  
Health Priorities ◽  
Care Practice ◽  
Postmarketing Surveillance ◽  
Care Practices ◽  
Primary Care Practices ◽  
Tract Infections

Introduction:While medicines and medical tests are developed in a controlled clinical trial environment, postmarketing surveillance in the real world can be challenging. MedicineInsight—a database of longitudinal patient-level clinical information from primary care practices in Australia—is a novel program that collects primary care data to improve postmarketing surveillance at a national level.Methods:MedicineInsight collects de-identified clinical information from primary care practice information systems using data extraction tools. MedicineInsight currently includes 3.6 million regular patients of 3,300 family physicians (general practitioners) from 650 primary care practices across Australia. MedicineInsight data include longitudinal clinical information on diagnosis and medicines (dose, strength, route of administration, medication switches over time, adverse events, and allergies), and pathology testing data. A series of observational studies was developed for postmarketing surveillance of management of a range of health priorities including type 2 diabetes mellitus (T2DM), chronic obstructive pulmonary disease (COPD), depression, and antibiotics use.Results:Forty-four percent of patients with T2DM in the MedicineInsight database did not have a recorded hemoglobin A1c result and thirty-one percent did not have a recorded blood pressure reading in the previous 6 months. While guidelines recommend a stepwise approach to the initiation of COPD therapy, forty-nine percent of patients with COPD (with or without asthma) were prescribed dual therapy at initiation and a small number (4.5 percent) were prescribed triple therapy. Between 2011 and 2015, the annual rate of antidepressant prescribing per 1,000 family physician encounters increased by eight percent. High volumes of antibiotics were prescribed for respiratory tract infections in Australian primary care, notwithstanding guideline recommendations that antibiotics are not recommended in most cases.Conclusions:Large scale, real-world clinical data from primary care practices can play an important role in postmarketing surveillance at a national level.

scholarly journals Cost-effectiveness of depression case management in small practices

2013 ◽  
Vol 202 (6) ◽  
pp. 441-446 ◽  
Author(s):  
Jochen Gensichen ◽  
Juliana J. Petersen ◽  
Michael Von Korff ◽  
Dirk Heider ◽  
Steffen Baron ◽  
...  
Keyword(s):  
Primary Care ◽  
Cost Effectiveness ◽  
Case Management ◽  
Controlled Trial ◽  
Cost Effective ◽  
Control Group ◽  
Significant Difference ◽  
Small Primary ◽  
Care Practices ◽  
Primary Care Practices

BackgroundCase management undertaken by healthcare assistants in small primary care practices is effective in improving depression symptoms and adherence in patients with major depression.AimsTo evaluate the cost-effectiveness of depression case management by healthcare assistants in small primary care practices.MethodCost-effectiveness analysis on the basis of a pragmatic randomised controlled trial (2005-2008): practice-based healthcare assistants in 74 practices provided case management to 562 patients with major depression over 1 year. Our primary outcome was the incremental costeffectiveness ratio (ICER) calculated as the ratio of differences in mean costs and mean number of qualityadjusted life-years (QALYs). Our secondary outcome was the mean depression-free days (DFDs) between the intervention and control group at 24-month follow-up. The study was registered at the International Standard Randomised Controlled Trial Number Registry: ISRCTN66386086.ResultsIntervention v. control group: no significant difference in QALYs; significantly more DFDs (mean: 373 v. 311, P<0.01); no significant difference in mean direct healthcare costs (€4495 v. €3506, P = 0.16); considerably lower mean indirect costs (€5228 v. €7539, P = 0.06), resulting in lower total costs (€9723 v. €11 045, P = 0.41). The point estimate for the cost-utility ratio was €38 429 per QALY gained if only direct costs were considered, and ‘dominance’ of the intervention if total costs were considered. Yet, regardless of decision makers' willingness to pay per QALY, the probability of the intervention being cost-effective was never above 90%.ConclusionsIn small primary care practices, 1 year of case management did not increase the number of QALYs but it did increase the number of DFDs. The intervention was likely to be cost-effective.

scholarly journals Haemoglobin concentration and erythrocyte sedimentation rate in primary care patients

1997 ◽  
Vol 90 (1) ◽  
pp. 16-18 ◽  
Author(s):  
E J Kanfer ◽  
B A Nicol
Keyword(s):  
Primary Care ◽  
Erythrocyte Sedimentation Rate ◽  
Sedimentation Rate ◽  
Haemoglobin Concentration ◽  
Inverse Correlation ◽  
Clinical Analysis ◽  
Plasma Viscosity ◽  
Erythrocyte Sedimentation ◽  
Significant Difference ◽  
Primary Care Practices

The erythrocyte sedimentation rate (ESR) remains a commonly measured indicator of disease, but is subject to several non-disease influences. The haemoglobin concentration (Hb) and ESR were measured in 1249 consecutive patients (492 men, 757 women) from primary care practices. An inverse correlation was found between Hb and ESR throughout the range of measured Hb, and in particular there was a significant difference in the median ESR of patients in the highest and lowest quartile for non-anaemic Hb (P<0.001). These results indicate that correct clinical analysis of an ESR result should take into account the Hb, both in anaemic and in non-anaemic patients. Interpretative difficulties due to external influences on the measured ESR could be resolved by replacement of this test with plasma viscosity estimation.

scholarly journals Early prediction of COVID-19 patient survival by targeted plasma multi-omics and machine learning

2021 ◽  
Author(s):  
Christoph Borchers ◽  
Vincent Richard ◽  
Claudia Gaither ◽  
Robert Popp ◽  
Daria Chaplygina ◽  
...  
Keyword(s):  
Machine Learning ◽  
Mass Spectrometry ◽  
Patient Survival ◽  
Treatment Success ◽  
Healthcare Systems ◽  
Early Prediction ◽  
Triple Quadrupole ◽  
Clinical Laboratories ◽  
Significant Difference ◽  
Length Of Hospitalization

Abstract The recent surge of COVID-19 hospitalizations severely challenges healthcare systems around the globe and demands for reliable tests predictive of disease severity and mortality. Using multiplexed targeted mass spectrometry assays on a robust triple quadrupole MS setup which is available in many clinical laboratories, we determined the precise concentrations of 100s of proteins and metabolites in plasma from hospitalized COVID-19 patients. We observed a clear distinction between COVID-19 patients and controls and, strikingly, a significant difference between survivors and non-survivors. With increasing length of hospitalization, the survivors’ samples showed a trend towards normal concentrations, indicating a potential sensitive readout of treatment success. Building a machine learning multi-omic model that considers the concentrations of ten proteins and five metabolites we could predict patient survival with 92% accuracy (AUC 0.97) at the day of hospitalization. Hence, our standardized assays represent a unique opportunity for the early stratification of hospitalized COVID-19 patients.

Smoking screening and management in primary care practices

1997 ◽  
Vol 6 (2) ◽  
pp. 165-172 ◽  
Author(s):  
P. E. McBride
Keyword(s):  
Primary Care ◽  
Care Practices ◽  
Primary Care Practices

Risk Factors in Diabetes and Pregnant Women with Urinary Tract Infections Compared to Younger Aged Female.

2018 ◽  
Vol 1 (3) ◽  
pp. 26-38
Author(s):  
Abdulghani Mohamed Alsamarai ◽  
Shler Ali Khorshed
Keyword(s):  
Risk Factors ◽  
Urinary Tract Infection ◽  
Urinary Tract ◽  
Tract Infection ◽  
Pregnant Women ◽  
Female Student ◽  
Mean Value ◽  
Infected Female ◽  
Significant Difference ◽  
Tract Infections

Background: Urinary tract infection is common with health impact in women and characterised by failure to treatment and recurrent episodes. Aim: This study was conducted to determine the risk factors for the development of urinary tract infection in diabetic and pregnant women in comparison to student female. Materials and methods: A prospective cross-sectional study conducted during the period from 1st of June 2015 to the end of January 2016. The population included in the study are 563 women, of them 425 were outpatients, and 138 were inpatients. Their age range between 18 and 80 years, with a mean age of 33.59±15.29 years. Urine samples collected and cultured on blood agar and MacConkey agar by spread plate technique. Bacterial colonies with different morphology were selected, purified and identified according to their biochemical characteristics using conventional standard methods. Results: In diabetic women, there were no significant difference in mean age and BMI values between culture positive and culture negative groups. However, pus cell mean scale was significantly higher [P=0.000] in women with urinary tract infection [1.76±1.25] than in those with negative culture [0.69±1.00]. In pregnant women, BMI mean value was significantly [P=0.013] lower in pregnant women with UTI [26.14] as compared to those without infection [26.99]. Pus cell scale mean value was significantly [P=0.000] higher in pregnant women with UTI [1.55] than women with negative UTI [0.85]. While there was no significant difference in mean age between UTI positive and negative pregnant women. In female student, there was a significant difference between UTI infected and non-infected in mean age [P=0.041] and pus cell scale [P=0.000]. However, BMI was not significantly different between infected and non-infected female student. Other risk factors association are variables in the 3 groups when analysed using X2, while AUC and OR show different trends of association between risk factors and UTI. Conclusion: BMI, pus cell scale, child number, delivery method, operation history and hospital setting were significantly associated with culture positivity in the 3 studied groups as determined by AUC. While OR confirmed association with pus sale scale in the 3 groups.

Sign in / Sign up

Export Citation Format

Share Document