Using machine learning models to predict falls in hospitalised adults

BACKGROUND

Identifying patients at high risk of falling is crucial in implementing effective fall prevention programs. While the integration of information systems is becoming more widespread in the healthcare industry, it poses a significant challenge in analysing vast amounts of data to identify factors that could enhance patient safety.

OBJECTIVE

To determine fall-associated factors and develop high-performance prediction tools for at-risk patients in acute and sub-acute care services in Australia.

METHODS

A retrospective study of 672,400 patients admitted to acute and sub-acute care services within a large metropolitan tertiary health service in Victoria, Australia, between January 1, 2019, and December 31, 2021. Data were obtained from four sources: the Department of Health Victorian Admitted Episodes Dataset, RiskManTM, electronic health records, and the health workforce dataset. Machine learning techniques, including Random Forest and Deep Neural Network models, were used to analyse the data, predict patient falls, and identify the most important risk factors for falls in this population. Model performance was evaluated using accuracy, F1-score, precision, recall, specificity, Matthew's correlation coefficient, and the area under the receiver operating characteristic curve (AUC).

RESULTS

The deep neural network and random forest models were highly accurate in predicting hospital patient falls. The deep neural network model achieved an accuracy of 0.988 and a specificity of 0.999, while the RF achieved an accuracy of 0.989 and a specificity of 1.000. The top 20 variables impacting falls were compared across both models, and 12 common factors were identified. These factors can be broadly classified into three categories: patient-related factors, staffing-related factors, and admission-related factors. Although not all factors are modifiable, they must be considered when planning fall prevention interventions.

CONCLUSION

The study demonstrated machine learning's potential to predict falls and identify key risk factors. Further validation across diverse populations and settings is essential for broader applicability.

Clinical risk prediction model for 30-day all-cause re-hospitalisation or mortality in patients hospitalised with heart failure

BACKGROUND

This study aimed to develop a risk prediction model (AUS-HF model) for 30-day all-cause re-hospitalisation or death among patients admitted with acute heart failure (HF) to inform follow-up after hospitalisation. The model uses routinely collected measures at point of care.

METHODS

We analyzed pooled individual-level data from two cohort studies on acute HF patients followed for 30-days after discharge in 17 hospitals in Victoria, Australia (2014-2017). A set of 58 candidate predictors, commonly recorded in electronic medical records (EMR) including demographic, medical and social measures were considered. We used backward stepwise selection and LASSO for model development, bootstrap for internal validation, C-statistic for discrimination, and calibration slopes and plots for model calibration.

RESULTS

The analysis included 1380 patients, 42.1% female, median age 78.7 years (interquartile range = 16.2), 60.0% experienced previous hospitalisation for HF and 333 (24.1%) were re-hospitalised or died within 30 days post-discharge. The final risk model included 10 variables (admission: eGFR, and prescription of anticoagulants and thiazide diuretics; discharge: length of stay>3 days, systolic BP, heart rate, sodium level (<135 mmol/L), >10 prescribed medications, prescription of angiotensin converting enzyme inhibitors or angiotensin receptor blockers, and anticoagulants prescription. The discrimination of the model was moderate (C-statistic = 0.684, 95%CI 0.653, 0.716; optimism estimate = 0.062) with good calibration.

CONCLUSIONS

The AUS-HF model incorporating routinely collected point-of-care data from EMRs enables real-time risk estimation and can be easily implemented by clinicians. It can predict with moderate accuracy risk of 30-day hospitalisation or mortality and inform decisions around the intensity of follow-up after hospital discharge.

Discussing HPV and oropharyngeal cancer in dental settings: gender and provider-type matter

Human papillomavirus (HPV) is the most common sexually transmitted infection in the US and the leading cause of oropharyngeal cancer (OPC), an oral cancer most often identified by dental providers. Given the rise in HPV-associated OPC and recent Food and Drug Administration (FDA) approval of the HPV vaccine to prevent OPC, dental providers have a unique role in HPV prevention. This study assessed US adults' comfort levels discussing HPV and OPC with dental providers. An online survey platform was used to recruit a nationally representative sample of US adults (n = 300). The questionnaire assessed participants' knowledge, acceptability, and comfort discussing HPV-related topics with dental providers. SPSS 24 was utilized for data analyses. In general, participants reported feeling comfortable discussing HPV and OPC with dental providers. Participants reported feeling more comfortable with dentists than dental hygienists when discussing (t = 2.85, p < .01) and receiving recommendations about the HPV vaccine (t = 2.09, p < .05). Participants were less comfortable discussing HPV as a risk factor for OPC compared to non-HPV related risk factors (t = 2.94, p < .01). Female participants preferred female providers, whereas male participants had no preference. Previous research has indicated dental providers recognize their role in HPV prevention, but research is needed to understand patients' perceptions of dental providers' role in HPV prevention. Findings demonstrate that US adults are comfortable discussing HPV and OPC with dental providers, which may be key to OPC-HPV prevention. Future research is needed to facilitate HPV communication between patients and dental providers.

Estimating the health loss due to poor engagement with cardiac rehabilitation in Australia

BACKGROUND

Cardiac rehabilitation (CR) programs are effective in reducing cardiovascular mortality and readmissions. However, most patients are denied the benefits of CR due to low referral rates. Of those patients referred, commencement rates vary from 28.4% to 60%. This paper quantifies the scale of health loss in Australia due to poor engagement with the program, and estimates how much public funding can be justifiably reallocated to address the problem.

METHODS

Economic decision modelling was undertaken to estimate the expected lifetime health loss and costs to Medicare. Key parameters were derived from Australian databases, CR registries and meta-analyses. Population health gains associated with uptake rates of 60%, and 85% were calculated.

RESULTS

CR was associated with a 99.9% probability of being cost-effective, even at a cost-effectiveness threshold lower than conventionally applied. Importantly, an average of 0.52 years of life expectancy are lost due to national uptake being below 60% achieved in some best performing programs in Australia, equivalent to 0.28 quality adjusted life years. The analysis indicates that $12.9 million/year could be justifiably reallocated from public funds to achieve a national uptake rate of 60%, while maintaining cost-effectiveness of CR due to the large health gains that would be expected.

CONCLUSION

CR is a cost-effective service for patients with coronary heart disease. In Australia, less than a third of patients commence CR, potentially resulting in avoidable patient harm. Additional investment in CR is vital and should be a national priority as the health gains for patients far outweigh the costs.

Self-Care Interventions That Reduce Hospital Readmissions in Patients With Heart Failure; Towards the Identification of Change Agents

Unplanned hospital readmissions are the most important, preventable cost in heart failure (HF) health economics. Current professional guidelines recommend that patient self-care is an important means by which to reduce this burden. Patients with HF should be engaged in their care such as by detecting, monitoring, and managing their symptoms. A variety of educational and behavioural interventions have been designed and implemented by health care providers to encourage and support patient self-care. Meta-analyses support the use of self-care interventions to improve patient self-care and reduce hospital readmissions; however, efficacy is variable. The aim of this review was to explore methods to achieve greater clarity and consistency in the development and reporting of self-care interventions to enable ‘change agents’ to be identified. We conclude that advancement in this field requires more explicit integration and reporting on the behaviour change theories that inform the design of self-care interventions and the selection of behaviour change techniques. The systematic application of validated checklists, such as the Theory Coding Scheme and the CALO-RE taxonomy, will improve the systematic testing and refinement of interventions to enable ‘change agent/s’ to be identified and optimised.

The Role of Dental Providers in Preventing HPV-Related Diseases: A Systems Perspective

Successfully educating dental providers and patients about the link between human papillomavirus (HPV) and oropharyngeal cancer requires coordinated efforts to increase HPV-related prevention practices. The aim of this study was to identify, using a systems perspective, the multi-level determinants related to how dental providers can promote HPV prevention in dental practices. Data for this qualitative study were collected in 2015-16 from focus groups with dentists (four focus groups, n=33), focus groups with dental hygienists (four focus groups, n=48), and in-depth interviews with dental opinion leaders (n=13). Results were triangulated and mapped along micro, meso, and macro system levels. At the micro level, participants identified patient characteristics and low self-efficacy as influential determinants when discussing HPV prevention. At the meso level, relationships among dentists, dental hygienists, and the physical practice environment were factors affecting dental providers' HPV prevention efforts. At the macro level, professional organizations impacted how dental providers interacted with their patients on this topic. These results suggest that improving HPV prevention among dental providers requires a multi-level approach that considers the distinctive context of dental settings, dental training, and perceptions of professional roles. The findings suggested that the macro- and meso-level determinants may be challenging to modify due to the distinctive culture and practice models of dentistry. Nevertheless, the association between HPV and oral cancer requires an expansion of prevention strategies used in dental practices. Improving dental providers' self-efficacy to communicate HPV prevention through continuing education and integration of skill-guided training in dental and dental hygiene curricula could facilitate this process.

Abstract P6-13-08: Palbociclib and paclitaxel on an alternating schedule for advanced breast cancer: Results of a phase Ib trial

Background: Palbociclib (P) is an oral CDK 4/6 inhibitor (CDKi) that was recently FDA approved in combination with endocrine therapy for metastatic breast cancer. We have performed a Phase I trial of P in combination with paclitaxel (T) based on preclinical studies suggesting that P synergizes with T when given on an alternating schedule, enabling cell cycle synchronization in tumor cells. We now present the dose expansion cohort.

Methods: Patients (Pts) enrolled on the trial had Rb-expressing tumors of any estrogen/progesterone/HER2 receptor type, adequate organ function, and ≤3 prior chemotherapy regimens for metastatic breast cancer (mBC). Prior adjuvant or metastatic taxane was allowed. Dose escalation led to expansion at P100mg or 75mg, starting with 3 days of P (run-in) and reduction of P dosing from 5-day to 3-day intervals (days 2-4, 9-11, 16-18 of each 28 day cycle). T at 80mg/m2 was given weekly for 3 cycles; thereafter, T was administered days 1, 8 and 15 of 28 day cycle. Weekly toxicity assessments were performed; RECIST 1.0 response was assessed every 2 cycles as partial response (PR), stable disease (SD) or progressive disease (PD). Pts had the option to discontinue T and continue on P alone (3 on/1 off schedule) if they attained SD after cycle 6.

Results: 27 pts enrolled on study (15- dose escalation, 12- dose expansion). Results are shown in the Table. 21 pts had received prior taxane; pts had received a median of 2 chemotherapy regimens for mBC. DLTs were grade 3 AST/ALT (n=1, at 125 mg) and febrile neutropenia (FN) (n=1, at 100 mg). Uncomplicated grade 3/4 NTP was common and frequently led to dose reduction or dose interruption during the first cycle of therapy. Frequency of NTP did not change with reducing the days of P. Among 24 evaluable patients, 14 (58%), had PR or SD ≥ 6 months across all dose levels. Of 14 pts who responded, 10 (71%) had received prior taxane. 20 pts are off study; 19 for PD, and 2 for toxicity (NTP in cycle 17 and FN in cycle 1); 7 pts remain on study. Prolonged tumor responses were seen.

Conclusions: P and T can be safely combined on an alternating dosing schedule; the optimal combination dose is 75 mg of P and 80mg/m2 of weekly T. The high response rate warrants a randomized trial to determine the incremental benefit over T alone. Additional mechanistic studies are in progress to understand the in vivo effects of the alternating dosing schedule on cell cycle activity and tumor proliferation.

Starting Dose Level P (mg)Number (Total 27)DLTGrade 3/4 NTP (n)Final Dose P mg (n)Dose Interruption (n)Best Response (n)5030050 (1) 50 (1) 50 (1)No (2) Yes (1)PR (1) SD (1) PD (1)7530275 (1) 50 (1) 25 (1)No (1) Yes (2)PR (2) SD (1)100605100 (2) 75 (3) 25 (1)No (1) Yes (5)PR (2) SD (1) PD (3)12531- LFT375 (1) 50 (2)No (0) Yes (3)PR (1) SD (2)75 (Run-In)60175 (5) 50 (1)No (4) Yes (2)PR (1) SD (2) PD (1) N/A (2)*100 (Run-In)61- FN5100 (1) 75 (4) 25(1)No (1) Yes (5)PR (4) SD (1) N/A (1)^*2 pts not yet evaluable. ^1 pt went off study due to FN after cycle 1.

Citation Format: Clark AS, O'Dwyer P, Troxel A, Lal P, Feldman M, Gallagher M, Driscoll A, Colameco C, Lewis D, Rosen M, Matro J, Bradbury A, Domchek S, Fox K, DeMichele A. Palbociclib and paclitaxel on an alternating schedule for advanced breast cancer: Results of a phase Ib trial. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P6-13-08.

Prevention of HPV-related oral cancer: assessing dentists' readiness

OBJECTIVES

Epidemiological research indicates an association between the Human Papillomavirus (HPV) with a subset of oral cancers (OC). Dentists may play a role in primary prevention of HPV-related OC by discussing the HPV vaccine with patients. This study assessed dentists' readiness to discuss the HPV vaccine with female patients.

STUDY DESIGN

Cross-sectional web-based survey.

METHODS

A web-based survey based on the Transtheoretical Model was administered among Florida dentists (n = 210).

RESULTS

The majority of participants (97%) fell into the precontemplation and contemplation stages of readiness to discuss the HPV vaccine with patients. Perceived role and liability were determined to be predictive of dentists in contemplation stage as opposed to those in precontemplation (P < 0.05).

CONCLUSIONS

Findings suggest liability and perceived role as processes of change necessary to guide dentists to primary prevention of HPV-related OC despite high levels of knowledge. As public awareness of HPV-related OC increases, dentists may become more involved in primary prevention. Results of the current study may assist in developing intervention strategies for engaging dentists in discussing the HPV vaccine with patients.

Evidence-based chronic heart failure management programs: reality or myth?

BACKGROUND

Chronic heart failure management programmes (CHF-MPs) have become part of standard care for patients with chronic heart failure (CHF).

OBJECTIVE

To investigate whether programmes had applied evidence-based expert clinical guidelines to optimise patient outcomes.

DESIGN

Prospective cross-sectional survey was used to conduct a national audit.

SETTING

Community setting of CHF-MPs for patients after discharge.

SAMPLE

All CHF-MPs operating during 2005-2006 (n = 55). 10-50 consecutive patients from 48 programmes were also recruited (n = 1157).

MAIN OUTCOME MEASURES

(1) Characteristics and interventions used within each CHF-MP and (2) characteristics of patients enrolled into these programmes.

RESULTS

Overall, there was a disproportionate distribution of CHF-MPs across Australia. Only 6.3% of hospitals nationally provided a CHF-MP. A total of 8000 post-discharge CHF patients (median, 126; IQR, 26-260) were managed via CHF-MPs representing only 20% of the potential national case load. Significantly, 16% of the case load comprised patients in functional New York Heart Association class I with no evidence of these patients having had previous echocardiography to confirm a diagnosis of CHF. Heterogeneity of CHF-MPs in applied models of care was evident with 70% of CHF-MPs offering a hybrid model (a combination of heart failure outpatient clinics and home visits), 20% conducting home visits and 16% an extended rehabilitation model of care. Less than half (44%) allowed heart failure nurses to titrate medications. The main medications that were titrated in these programmes were diuretics (n = 23, 96%), beta-blockers (n = 17, 71%), ACE inhibitors (n = 14, 58%) and spironolactone (n = 9, 38%).

CONCLUSION

CHF-MPs are being implemented rapidly throughout Australia. However, many of these programmes do not adhere to expert clinical guidelines for the management of patients with CHF. This poor translation of evidence into practice highlights the inconsistency and questions the quality of health-related outcomes for these patients.

Suboptimal management of cardiovascular risk factors in coronary heart disease patients in primary care occurs particularly in women

BACKGROUND

Patients with established coronary heart disease (CHD) are at the highest risk of further events. Despite proven therapies, secondary prevention is often suboptimal. General practitioners (GPs) are in an ideal position to improve secondary prevention.

AIM

To contrast management of cardiovascular risk factors in patients with established CHD in primary care to those in clinical guidelines and according to gender.

METHODS

GPs throughout Australia were approached to participate in a programme incorporating a disease management software (mdCare) program. Participating practitioners (1258 GPs) recruited individual patients whose cardiovascular risk factor levels were measured.

RESULTS

The mdCare programme included 12,509 patients (58% male) diagnosed with CHD. Their mean age was 71.7years (intra-quartile range 66-78) for men and 74years (intra-quartile range 68-80) for women. Low-density-lipoprotein cholesterol was above target levels in 69% (2032) of women compared with 58% (2487) in men (P < 0.0001). There was also a higher proportion of women with total cholesterol above target levels (76%, 3592) compared with men (57%, 3787) (P < 0.0001). In patients who were prescribed lipid-lowering medication, 53% (2504) of men and 72% (2285) of women continued to have a total cholesterol higher than recommended target levels (P < 0.0001). Overall, over half (52%, 6538) had at least five cardiovascular risk factors (55% (2914) in women and 50% (3624) in men, P < 0.0001).

CONCLUSION

This study found less intensive management of cardiovascular risk factors in CHD patients, particularly among women, despite equivalent cardiovascular risk. This study has shown that these patients have multiple risk factors where gender also plays a role.

How do patients with interstitial cystitis present?

PURPOSE

We determined how interstitial cystitis progresses from initial symptoms to diagnosis.

MATERIALS AND METHODS

We retrospectively analyzed the records of 45 patients to determine the dates of symptom onset and diagnosis, and sequence of urgency/frequency, nocturia and pain. We also documented alternate and previous diagnoses, and previous surgical treatments.

RESULTS

Of the patients 89% presented with only 1 symptom. Median time from the initial symptom to all symptoms was 2 years (mean 5.5). The most common previous diagnoses were urinary infection in 19 cases, a gynecologic diagnosis in 14 and urethral diagnoses in 6. A previous urinary infection was documented in only 1 of 19 patients, while 11 had undergone hysterectomy and 5 diagnosed with endometriosis had no pathological documentation available.

CONCLUSIONS

Early interstitial cystitis presents variably and usually with only a single symptom of urgency/frequency, nocturia or pain. Clinicians may fail to appreciate the symptoms of early interstitial cystitis, which leads to delayed diagnosis until the patient is more symptomatic. With time multiple symptoms manifest. Bacterial cystitis, prostatitis, endometriosis and chronic pelvic pain are common initial misdiagnoses. Interstitial cystitis should be considered when laboratory documentation of alternate diagnoses is lacking or when patients fail to respond to therapy for alternate diagnoses.

Managing post-discharge care at home: an analysis of patients' and their carers' perceptions of information received during their stay in hospital

This small-scale study carried out in a Melbourne metropolitan hospital explored patients' and their carers' perceptions of information, adequacy of information, and their utilization of information concerning post-discharge care received from health professionals during their stay in hospital. The research design consisted of two stages. Stage one involved a qualitative approach using focused interviews of five pairs of patients and their carers, 2 weeks after discharge from hospital. Five main themes emerged from the content analysis of the interview transcripts: information given by health professionals to patients and carers, patients' and carers' psychological well-being, activities of daily living, caring tasks of the patients, and community linkages. A quantitative approach was used for stage two involving two sets of questionnaires, one for the patient and one for the carer, developed from the themes identified in stage one. A pilot study was conducted on three pairs of patients and their carers, 2 weeks after discharge from hospital. The main study consisted of a convenience sample of 40 pairs of patients and their carers who completed the questionnaires 2 weeks post-discharge. Data analysis of stage two of the study consisted of descriptive statistics and cross-tabulations. The main findings suggested that carers received very little information from health professionals concerning their patients' health problems and care at home. The carers' health and employment states were often not considered in their patients' discharge plan. Carers who were present with their patients when they received information concerning post-discharge care experienced a decrease in anxiety during their patients' convalescence at home, greater satisfaction with the information they received, and their patients experienced fewer medical problems post-discharge. The implications for nursing practice and research include recommendations for a more effective system of discharge planning, and further research to include a larger population with a more varied group of participants.

Young women's degree of control over first intercourse: an exploratory analysis

CONTEXT

While policymakers and researchers alike often seem to believe that young women's decision to initiate sexual intercourse is conscious and free of ambiguity, the actual degree of control that such young women exert over first intercourse has rarely been explicitly examined.

METHODS

The 1995 National Survey of Family Growth asked all women who had experienced intercourse to rate, on a 1-10 scale, the wantedness of their first intercourse; they were then asked whether the experience was voluntary. Logistic regression analysis of data for women aged 15-24 who had experienced first premarital intercourse was performed to test the effect of background factors and wantedness scores on contraceptive use at voluntary first intercourse.

RESULTS

Twenty-four percent of women aged 13 or younger at the time of their first premarital intercourse report the experience to have been nonvoluntary, compared with 10% of those aged 19-24 at first premarital intercourse. About one-quarter of respondents who reported their first intercourse as voluntary chose a low value (1-4) on the wantedness scale. Women whose first partner was seven or more years older than themselves were more than twice as likely as those whose first partner was the same age or younger to choose a low value (36% vs. 17%). Women whose partner had been seven or more years older were also less likely than other women to have used contraceptives at first intercourse. After the introduction of controls for demographic and background factors, partner age discrepancy and relationship status, wantedness of voluntary first intercourse was not independently related to the odds of contraceptive use at that intercourse.

CONCLUSION

Characterizing women's first intercourse as simply voluntary or nonvoluntary is inadequate. Measures that take into account degrees of wantedness may help elucidate relationships between sexual initiation, contraceptive use and teenage pregnancy. The fact that substantial numbers of young women voluntarily participated in a first sexual experience about which they felt ambivalent or negative deserves the attention of program planners and service providers.