Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity

OBJECTIVE

To explore factors that potentially impact external validation performance while developing and validating a prognostic model for hospital admissions (HAs) in complex older general practice patients.

STUDY DESIGN AND SETTING

Using individual participant data from four cluster-randomised trials conducted in the Netherlands and Germany, we used logistic regression to develop a prognostic model to predict all-cause HAs within a 6-month follow-up period. A stratified intercept was used to account for heterogeneity in baseline risk between the studies. The model was validated both internally and by using internal-external cross-validation (IECV).

RESULTS

Prior HAs, physical components of the health-related quality of life comorbidity index, and medication-related variables were used in the final model. While achieving moderate discriminatory performance, internal bootstrap validation revealed a pronounced risk of overfitting. The results of the IECV, in which calibration was highly variable even after accounting for between-study heterogeneity, agreed with this finding. Heterogeneity was equally reflected in differing baseline risk, predictor effects and absolute risk predictions.

CONCLUSIONS

Predictor effect heterogeneity and differing baseline risk can explain the limited external performance of HA prediction models. With such drivers known, model adjustments in external validation settings (eg, intercept recalibration, complete updating) can be applied more purposefully.

TRIAL REGISTRATION NUMBER

PROSPERO id: CRD42018088129.

The Effect of a Comprehensive, Interdisciplinary Medication Review on Quality of Life and Medication Use in Community Dwelling Older People with Polypharmacy

Background: We conducted a comprehensive medication review at the patients’ home, using data from electronic patient records, and with input from relevant specialists, general practitioners and pharmacists formulated and implemented recommendations to optimize medication use in patients aged 60+ years with polypharmacy. We evaluated the effect of this medication review on quality of life (QoL) and medication use. Methods: Cluster randomized controlled trial (stepped wedge), randomly assigning general practices to one of three consecutive steps. Patients received usual care until the intervention was implemented. Primary outcome was QoL (SF-36 and EQ-5D); secondary outcomes were medication changes, medication adherence and (instrumental) activities of daily living (ADL, iADL) which were measured at baseline, and around 6- and 12-months post intervention. Results: Twenty-four general practices included 360 women and 410 men with an average age of 75 years (SD 7.5). A positive effect on SF-36 mental health (estimated mean was stable in the intervention, but decreased in the control condition with −6.1, p = 0.009,) was found with a reduced number of medications at follow-up compared to the control condition. No significant effects were found on other QoL subscales, ADL, iADL or medication adherence. Conclusion: The medication review prevented decrease of mental health (SF36), with no significant effects on other outcome measures, apart from a reduction in the number of prescribed medications.

Predicting negative health outcomes in older general practice patients with chronic illness: Rationale and development of the PROPERmed harmonized individual participant data database

The prevalence of multimorbidity and polypharmacy increases significantly with age and are associated with negative health consequences. However, most current interventions to optimize medication have failed to show significant effects on patient-relevant outcomes. This may be due to ineffectiveness of interventions themselves but may also reflect other factors: insufficient sample sizes, heterogeneity of population. To address this issue, the international PROPERmed collaboration was set up to obtain/synthesize individual participant data (IPD) from five cluster-randomized trials. The trials took place in Germany and The Netherlands and aimed to optimize medication in older general practice patients with chronic illness. PROPERmed is the first database of IPD to be drawn from multiple trials in this patient population and setting. It offers the opportunity to derive prognostic models with increased statistical power for prediction of patient-relevant outcomes resulting from the interplay of multimorbidity and polypharmacy. This may help patients from this heterogeneous group to be stratified according to risk and enable clinicians to identify patients that are likely to benefit most from resource/time-intensive interventions. The aim of this manuscript is to describe the rationale behind PROPERmed collaboration, characteristics of the included studies/participants, development of the harmonized IPD database and challenges faced during this process.

Factors associated with appropriate knowledge of the indications for prescribed drugs among community-dwelling older patients with polypharmacy

BACKGROUND

polypharmacy contributes to patients' non-adherence with physicians' prescriptions. Patients' knowledge about the indications for their medicines is one of the factors influencing adherence.

OBJECTIVE

to identify factors associated with appropriate knowledge about the indications for drugs prescribed to older patients with polypharmacy.

METHODS

in a primary care setting, using home interviews and postal questionnaires, patients aged 60 and over who were taking five or more prescribed drugs simultaneously were asked about their medication. Multiple logistic regression analysis was used to evaluate the association (odds ratio, OR) between medication knowledge and explanatory variables like medication use, sex, age, living situation and educational level.

RESULTS

seven hundred and fifty-four participants (mean age 73.2 years) reported an average daily intake of nine (SD 3.0) prescribed drugs. Only 15% of the patients were able to recall the indication for each of their prescribed drugs. Variables that were negatively associated with correct reporting of all indications were taking many prescribed drugs (e.g. ≥10 versus ≤5: OR 0.05), age 80 years or over (versus 60-69 years: OR 0.47) and male sex (OR 0.53). Patients living with a partner were more knowledgeable than patients living alone (OR 2.11). We did not find an association with educational level.

CONCLUSION

among older patients using five or more prescribed drugs, there was little understanding of the indications for their drugs, especially among patients taking the highest number of drugs, patients aged 80 or over, and men. Patients living independently with a partner were more knowledgeable than others.

[What do older patients on polypharmacy know about their prescribed drugs?]

OBJECTIVE

To identify factors associated with appropriate knowledge about the indications for drugs prescribed to older patients on polypharmacy.

DESIGN

Cross-sectional study.

METHOD

In a primary care setting, patients aged 60 years and over who were taking five or more prescribed drugs simultaneously were asked about their medication by conducting home interviews and postal questionnaires. Multiple logistic regression analysis was used to evaluate the association between medication knowledge and explanatory variables such as medication use, sex, age, residential status and educational level.

RESULTS

Seven hundred and fifty-four participants, mean age 73.2 years, reported an average daily intake of nine (SD 3.0) prescribed drugs. Only 15% of the patients were able to recall the indication for each of their prescribed drugs. Variables that were negatively associated with correct reporting of all indications were: taking many prescribed drugs (≥ 10 vs. ≤ 5 drugs: odds ratio (OR) 0.05), age 80 years or over (vs. 60-69 years: OR 0.47) and male sex (vs. females: OR 0.53). Patients living with a partner were more knowledgeable than patients living alone (OR 2.11). We did not find an association with educational level.

CONCLUSION

Among older patients using five or more prescribed drugs, there was little understanding of the indications for their drugs, especially among patients taking the highest number of drugs, patients aged 80 years or over, and men. Patients living independently with a partner were more knowledgeable than others.

[How much do patients and health professionals (really) know? The surplus value of a home visit to the patient with polypharmacy by the practice nurse, to support medication reviews in primary care]

BACKGROUND

Polypharmacy in older people should be addressed by an annual review of the chronic medication. In the PIL-study this was done by an integrated approach by GP, practice nurse, pharmacist, specialist and patient. All patients were first visited at home by the practice nurse.

RESEARCH QUESTIONS

What 'over the counter' (OTC) medications do polypharmacy patients use? Do they know the indications of the prescribed medication? Does medication use according to the patient match with medication use according to the records of GP and pharmacist?

METHOD

Inclusion criteria were: age 60 years or older, daily use of five or more chronic medications, mental competence, and adequate command of the Dutch language. All patients were visited at home by the practice nurse, who made an inventory of the actual drug use.

RESULTS

Five hundred fifty patients used a total of 5576 drugs, including 527 (9.4%) OTC medication. Patients knew the indication of 64% of the prescribed medication. The number of prescribed drugs that a patient actually used did not match the numbers known to GP and pharmacist. In 60.4% of all medication prescriptions there was complete agreement between GP, pharmacist and patient. On a patient level agreement was 18.7%.

CONCLUSIONS

Home visits by the nurse practitioner to make an inventory of the medication as reported by the patient seem to have an added value.