Using classification tree modelling to investigate drug prescription practices at health facilities in rural Tanzania

[Know-how and techniques of Family Health Strategy managers and professionals in the prevention of violence against adolescents]

The scope of this study was to identify empowering factors for the actions of the Family Health Strategy (FHS) in coping with violence among adolescents. The sample was random from 61 teams of the Family Health Strategy of three municipalities. Questionnaires tabulated by EpiINFO 7.0, with bivariate analyses were used. Chi-square, Fisher and the "classification tree" were then used to analyze the experiences in this theme and the following issues: bonding, time since training completed, qualification, knowledge of public policies among others. There is acknowledgement of adolescents' vulnerability to violence, however, only two teams have ongoing prevention experiences. The prevailing opinion is that such activities are not the responsibility of the FHS. The study identifies interdependent variables that increase the likelihood of preventing violence and the promoting of a culture of peace among adolescents.

How peoples’ ratings of dental implant treatment change over time?

Objectives

Dental implant treatment (DIT) improves peoples’ oral health-related quality of life (OHQoL). Assessment of longitudinal changes in OHRQoL may be undermined by response shift (RS). RS is the process by which quality of life changes, independent of health status as a result recalibration, reprioritization or reconceptualization. Thus, this study aimed to describe RS in the OHRQoL and perceived oral health of individuals receiving DIT and to compare the then-test, a self-anchored scale and the classification and regression trees (CRT) approaches for assessing RS.

Methods

OHRQoL was assessed in 100 patients receiving DIT using the OHIP-Edent (n = 100) and a self-anchored scale (n = 45) before placement of the final restoration and 3 to 6 months after treatment was completed. The OHIP-Edent was also used as a retrospective assessment at follow-up. CRT examined changes in the OHIP-Edent total score as a dependent variable with global changes in oral health and each OHIP-Edent subscale score as independent variables.

Results

OHRQoL and perceived oral health improved after treatment. The OHIP-Edent score decreased from 36.4 at baseline to 12.7 after treatment. On average, participants recalibrated their internal standard downwards (− 4.0 OHIP-Edent points). CRT detected downwards recalibration in 5% of participants and upwards in 15%. Reprioritization was observed in the social disability and psychological discomfort aspects of OHRQoL.

Conclusions

RS affects longitudinal assessments of OHRQoL in DIT, reducing the apparent magnitude of change. The then-test and CRT are valid and complementary methods to assess RS.

Role of patient descriptors in predicting antimicrobial resistance in urinary tract infections using a decision tree approach: A retrospective cohort study

BACKGROUND

In general practice, many infections are treated empirically prior to or without microbiological confirmation. Prediction of antimicrobial susceptibility could optimise prescribing thus improving patient outcomes. Decision tree models are a novel idea to predict AMR at the time of clinical presentation. This study aims to apply a prediction model using a decision tree approach to predict the antimicrobial resistance (AMR) of pathogens causing urinary tract infections (UTI) for patients over 65 years based on pre-existing routine laboratory data.

METHODS

Data were extracted from the database of the microbiological laboratory of the University Hospitals Galway (UHG). All urine results from patients over 65 years, their microbiological analysis and susceptibility (AST) results from January 2011 to December 2015 were included. The primary endpoint was culture result and resistance to antimicrobials (nitrofurantoin, trimethoprim, ciprofloxacin, co-amoxiclav, and amoxicillin) commonly used to treat UTI. A non-parametric regression tree analysis i.e. a decision tree model was generated with the 75% of the dataset (training set) and validated with the remaining 25% (test set). The model performance was evaluated measuring Area Under the Curve Receiver Operating Characteristic (AUC_ROC) curve.

RESULTS

A total of 99,101 urine samples of patients over 65 years were submitted for culture over the five years and 27% had significant bacteriuria (≥10⁴ cfu/ml) and AST. The most common identified causative organisms were E.coli, Klebsiella spp. and Proteus spp. E.coli was more often resistant to amoxicillin (66%) followed by Proteus spp. (41%). Klebsiella spp. and Proteus spp. were more often resistant to trimethoprim (78% and 54% respectively). E. coli resistance to nitrofurantoin is low (<10%). The decision tree model showed an AUC-ROC score of 0.68 for culture and in between 0.60 to 0.97 for antimicrobial resistance of the pathogens, with the inclusion of patient's descriptors only. Including the uropathogen in the model did not change model performance.

CONCLUSIONS

The decision tree models using patient descriptors available at the time of presentation showed fair to excellent performance in predicting culture and antimicrobial resistance. The presented models provide an alternative approach to decision making on antimicrobial prescribing for UTIs. Increasing more predictors in the model could improve the model performance. Prospective data collection, validation and feasibility testing of the model including data from other laboratories will progress the practical implementation of similar models.

The Reliability of Classification of Terminal Nodes in GUIDE Decision Tree to Predict the Nonalcoholic Fatty Liver Disease

Tree structured modeling is a data mining technique used to recursively partition a dataset into relatively homogeneous subgroups in order to make more accurate predictions on generated classes. One of the classification tree induction algorithms, GUIDE, is a nonparametric method with suitable accuracy and low bias selection, which is used for predicting binary classes based on many predictors. In this tree, evaluating the accuracy of predicted classes (terminal nodes) is clinically of special importance. For this purpose, we used GUIDE classification tree in two statuses of equal and unequal misclassification cost in order to predict nonalcoholic fatty liver disease (NAFLD), considering 30 predictors. Then, to evaluate the accuracy of predicted classes by using bootstrap method, first the classification reliability in which individuals are assigned to a unique class and next the prediction probability reliability as support for that are considered.

Association of influenza vaccine uptake with health, access to health care, and medical mistreatment among adults from low-income neighborhoods in New Haven, CT: a classification tree analysis

OBJECTIVE

The aim of this study is to identify population subgroups under-vaccinated for influenza through classification tree analysis to inform interventions aimed at improving future vaccine uptake.

METHOD

A cross-sectional community health needs assessment was conducted from 09/2012 through 11/2012 among randomly selected households in six low-income neighborhoods in New Haven, CT (N=1300 adults, aged 18-65). Known correlates of influenza vaccine uptake plus experience of medical mistreatment were used to develop a classification tree to identify under-vaccinated population subgroups

RESULTS

Forty-five percent of respondents reported receiving the influenza vaccine. The classification tree identified healthy adults and uninsured adults at increased risk of influenza complications as subgroups with low vaccine uptake (40% and 30%, respectively). The subgroup representing insured, high-risk adults who reported experience of medical mistreatment had moderate vaccine uptake (45%). Sensitivity of the classification tree was high (83%, 95% CI=80% to 86%), indicating a strong true positive rate using these subgroups.

CONCLUSION

Results highlight the need for renewed attention to promoting the influenza vaccination recommendation for all adults, particularly among healthy adults, uninsured, high-risk adults and insured, high-risk adults who have experienced medical mistreatment. Further research is needed to better understand how to reach these population subgroups.

Supervising community health workers in low-income countries--a review of impact and implementation issues

Background

Community health workers (CHWs) are an increasingly important component of health systems and programs. Despite the recognized role of supervision in ensuring CHWs are effective, supervision is often weak and under-supported. Little is known about what constitutes adequate supervision and how different supervision strategies influence performance, motivation, and retention.

Objective

To determine the impact of supervision strategies used in low- and middle-income countries and discuss implementation and feasibility issues with a focus on CHWs.

Design

A search of peer-reviewed, English language articles evaluating health provider supervision strategies was conducted through November 2013. Included articles evaluated the impact of supervision in low- or middle-income countries using a controlled, pre-/post- or observational design. Implementation and feasibility literature included both peer-reviewed and gray literature.

Results

A total of 22 impact papers were identified. Papers were from a range of low- and middle-income countries addressing the supervision of a variety of health care providers. We classified interventions as testing supervision frequency, the supportive/facilitative supervision package, supervision mode (peer, group, and community), tools (self-assessment and checklists), focus (quality assurance/problem solving), and training. Outcomes included coverage, performance, and perception of quality but were not uniform across studies. Evidence suggests that improving supervision quality has a greater impact than increasing frequency of supervision alone. Supportive supervision packages, community monitoring, and quality improvement/problem-solving approaches show the most promise; however, evaluation of all strategies was weak.

Conclusion

Few supervision strategies have been rigorously tested and data on CHW supervision is particularly sparse. This review highlights the diversity of supervision approaches that policy makers have to choose from and, while choices should be context specific, our findings suggest that high-quality supervision that focuses on supportive approaches, community monitoring, and/or quality assurance/problem solving may be most effective.

Correlates of inappropriate prescribing of antibiotics to patients with malaria in Uganda

Background

In many rural areas of Uganda, febrile patients presenting to health facilities are prescribed both antimalarials and antibiotics, contributing to the overuse of antibiotics. We identified the prevalence and correlates of inappropriate antibiotic management of patients with confirmed malaria.

Methods

We utilized individual outpatient data from 36 health centers from January to September 2011. We identified patients who were prescribed antibiotics without an appropriate clinical indication, as well as patients who were not prescribed antibiotics when treatment was clinically indicated. Multivariate logistic regression models were used to identify clinical and operational factors associated with inappropriate case management.

Findings

Of the 45,591 patients with parasitological diagnosis of malaria, 40,870 (90%) did not have a clinical indication for antibiotic treatment. Within this group, 17,152 (42%) were inappropriately prescribed antibiotics. The odds of inappropriate prescribing were higher if the patient was less than five years old (aOR 1.96, 95% CI 1.75–2.19) and if the health provider had the fewest years of training (aOR 1.86, 95% CI 1.05–3.29). The odds of inappropriate prescribing were lower if patients had emergency triage status (aOR 0.75, 95% CI 0.59–0.96) or were HIV positive (aOR 0.31, 95% CI 0.20–0.45). Of the 4,721 (10%) patients with clinical indications for antibiotic treatment, 521 (11%) were inappropriately not prescribed antibiotics. Clinical officers were less likely than medical officers to inappropriately withhold antibiotics (aOR 0.54, 95% CI 0.29–0.98).

Conclusion

Over 40% of the antibiotic treatment in malaria positive patients is prescribed despite a lack of documented clinical indication. In addition, over 10% of patients with malaria and a clinical indication for antibiotics do not receive them. These findings should inform facility-level trainings and interventions to optimize patient care and slow trends of rising antibiotic resistance.