Enhancing Homeopathic Prescribing for Chronic Cough by Classifying the Reliability of Polar Symptoms

Assessment of relationship between polar symptoms (PS), such as weather, activity or food, and homeopathic medicines in chronic cough has produced some confusing outcomes. Statistical variation in the data partly explains the seemingly conflicting outcomes in prognostic factor research (PFR) of PS. Classification of statistical probability and selection of statistically reliable data can reduce conflicting outcomes.

To make an inventory of statistically reliable PS and their relationship with homeopathic medicines in chronic cough.

A prospective observational cohort study was conducted at 10 outpatient centers of the Central Council for Research in Homoeopathy, India. Patients with chronic cough were followed for 12 months. Effect of treatment was assessed using the ORIDL (Outcome in Relation to Impact on Daily Living) instrument, with symptom severity and health status evaluated respectively via the Cough Severity Index and EuroQoL 5D-5L. Prevalence and likelihood ratios (LRs) of PS were calculated for general and medicine-specific populations. Cumulative binomial probability (CBP) was used to determine whether symptom prevalence among good responders to specific medicines significantly differed from that of the general population.

Some LRs were unexpectedly low, which could be explained by asymmetry of frequency distributions, resulting in high prevalence of one symptom pole. CBP for homeopathic medicines was subsequently performed, with 131 symptoms classified as 'certain' or 'probable': 21 symptoms for Arsenicum album, 23 for Bryonia, 11 for Calcarea carbonica, 3 for Lycopodium, 15 for Natrium muriaticum, 5 for Nux vomica, 17 for Phosphorus, 13 for Pulsatilla, 10 for Silicea, and 13 for Sulphur.

Conflicting outcomes in PFR of PS can partly be explained by statistical variation. A classification of reliability was not feasible when expressed using the classical 95% confidence interval but could be achieved by CBP, resulting in certain and probable PS for 10 medicines. We still need clinical expertise and corroboration for future research. Practitioners need scientific training to prevent bias and use research instruments such as Likert scales and questionnaires.