The comprehensive diagnostic study is suggested as a design to model the diagnostic process

Identifying Existing Evidence to Potentially Develop a Machine Learning Diagnostic Algorithm for Cough in Primary Care Settings: Scoping Review

BACKGROUND

Primary care is known to be one of the most complex health care settings because of the high number of theoretically possible diagnoses. Therefore, the process of clinical decision-making in primary care includes complex analytical and nonanalytical factors such as gut feelings and dealing with uncertainties. Artificial intelligence is also mandated to offer support in finding valid diagnoses. Nevertheless, to translate some aspects of what occurs during a consultation into a machine-based diagnostic algorithm, the probabilities for the underlying diagnoses (odds ratios) need to be determined.

OBJECTIVE

Cough is one of the most common reasons for a consultation in general practice, the core discipline in primary care. The aim of this scoping review was to identify the available data on cough as a predictor of various diagnoses encountered in general practice. In the context of an ongoing project, we reflect on this database as a possible basis for a machine-based diagnostic algorithm. Furthermore, we discuss the applicability of such an algorithm against the background of the specifics of general practice.

METHODS

The PubMed, Scopus, Web of Science, and Cochrane Library databases were searched with defined search terms, supplemented by the search for gray literature via the German Journal of Family Medicine until April 20, 2023. The inclusion criterion was the explicit analysis of cough as a predictor of any conceivable disease. Exclusion criteria were articles that did not provide original study results, articles in languages other than English or German, and articles that did not mention cough as a diagnostic predictor.

RESULTS

In total, 1458 records were identified for screening, of which 35 articles met our inclusion criteria. Most of the results (11/35, 31%) were found for chronic obstructive pulmonary disease. The others were distributed among the diagnoses of asthma or unspecified obstructive airway disease, various infectious diseases, bronchogenic carcinoma, dyspepsia or gastroesophageal reflux disease, and adverse effects of angiotensin-converting enzyme inhibitors. Positive odds ratios were found for cough as a predictor of chronic obstructive pulmonary disease, influenza, COVID-19 infections, and bronchial carcinoma, whereas the results for cough as a predictor of asthma and other nonspecified obstructive airway diseases were inconsistent.

CONCLUSIONS

Reliable data on cough as a predictor of various diagnoses encountered in general practice are scarce. The example of cough does not provide a sufficient database to contribute odds to a machine learning-based diagnostic algorithm in a meaningful way.

The value of the GP's clinical judgement in predicting dementia: a multicentre prospective cohort study among patients in general practice

BACKGROUND

Clinical judgement is intrinsic to diagnostic strategies in general practice; however, empirical evidence for its validity is sparse.

AIM

To ascertain whether a GP's global clinical judgement of future cognitive status has an added value for predicting a patient's likelihood of experiencing dementia.

DESIGN AND SETTING

Multicentre prospective cohort study among patients in German general practice that took place from January 2003 to October 2016.

METHOD

Patients without baseline dementia were assessed with neuropsychological interviews over 12 years; 138 GPs rated the future cognitive decline of their participating patients. Associations of baseline predictors with follow-up incident dementia were analysed with mixed-effects logistic and Cox regression.

RESULTS

A total of 3201 patients were analysed over the study period (mean age = 79.6 years, 65.3% females, 6.7% incident dementia in 3 years, 22.1% incident dementia in 12 years). Descriptive analyses and comparison with other cohorts identified the participants as having frequent and long-lasting doctor-patient relationships and being well known to their GPs. The GP baseline rating of future cognitive decline had significant value for 3-year dementia prediction, independent of cognitive test scores and patient's memory complaints (GP ratings of very mild (odds ratio [OR] 1.97, 95% confidence intervals [95% CI] = 1.28 to 3.04); mild (OR 3.00, 95% CI = 1.90 to 4.76); and moderate/severe decline (OR 5.66, 95% CI = 3.29 to 9.73)). GPs' baseline judgements were significantly associated with patients' 12-year dementia-free survival rates (Mantel-Cox log rank test P<0.001).

CONCLUSION

In this sample of patients in familiar doctor-patient relationships, the GP's clinical judgement holds additional value for predicting dementia, complementing test performance and patients' self-reports. Existing and emerging primary care-based dementia risk models should consider the GP's judgement as one predictor. Results underline the importance of the GP-patient relationship.

Development of practical recommendations for diagnostic accuracy studies in low-prevalence situations

OBJECTIVE

Low disease prevalence poses challenges for diagnostic accuracy studies because of the large sample sizes that are required to obtain sufficient precision. The aim is to collate and discuss designs of diagnostic accuracy studies suited for use in low-prevalence situations.

STUDY DESIGN AND SETTING

We conducted a literature search including backward citation tracking and expert consultation. Two reviewers independently selected studies on designs for estimating diagnostic accuracy in a low-prevalence situation. During a 1-day expert meeting, all designs were discussed and recommendations were formulated.

RESULTS

We identified six designs for diagnostic accuracy studies that are suitable in low-prevalence situations because they reduced the total sample size or the number of patients undergoing the index test or reference standard depending on which poses the highest burden. We described the advantages and limitations of these designs and evaluated efficiencies in sample sizes, risk of bias, and alignment with the clinical pathway for applicability in routine care.

CONCLUSION

Choosing a study design for diagnostic accuracy studies in low-prevalence situations should depend on whether the aim is to limit the number of patients undergoing the index test or reference standard, and the risk of bias associated with a particular design type.

Causes of chest pain in primary care--a systematic review and meta-analysis

Aim

To investigate the frequencies of different and relevant underlying etiologies of chest pain in general practice.

Methods

We systematically searched PubMed and EMBASE. Two reviewers independently rated the eligibility of publications and assessed the risk of bias of included studies. We extracted data to calculate the relative frequencies of different underlying conditions and investigated the variation across studies using forest plots, I², tau², and prediction intervals. With respect to unexplained heterogeneity, we provided qualitative syntheses instead of pooled estimates.

Results

We identified 11 eligible studies comprising about 6500 patients. The overall risk of bias was rated as low in 6 studies comprising about 3900 patients. The relative frequencies of different conditions as the underlying etiologies of chest pain reported by these studies ranged from 24.5 to 49.8% (chest wall syndrome), 13.8 to 16.1% (cardiovascular diseases), 6.6 to 11.2% (stable coronary heart disease), 1.5 to 3.6% (acute coronary syndrome/myocardial infarction), 10.3 to 18.2% (respiratory diseases), 9.5 to 18.2% (psychogenic etiologies), 5.6 to 9.7% (gastrointestinal disorders), and 6.0 to 7.1% (esophageal disorders).

Conclusion

This information may be of practical value for general practitioners as it provides the pre-test probabilities for a range of underlying diseases and may be suitable to guide the diagnostic process.