A model for predicting 6-minute walk test (6MWT) outcomes in patients with chronic obstructive pulmonary disease (COPD) that does not involve a physical assessment could facilitate ongoing at-home monitoring of disease progression. The investigators who created and tested this model reported their findings Computer Methods and Programs in Biomedicine.

The 6MWT is commonly used to evaluate functional exercise capacity in patients with COPD; however, some patients may have practical issues in completing this test due to multiple factors, including comorbidities and weather. Researchers therefore developed a novel model to predict key outcomes of the 6MWT in a way that did not require specific physical assessment and provided continuous monitoring information. The model used cardiopulmonary and clinical parameters as inputs of a Bayesian network to estimate relevant outputs of the 6MWT, including the 6-minute walking distance (6MWD), the maximum heart rate (HRmax) achieved when completing the 6MWT, and the heart rate recovery index evaluated after 3 minutes of recuperation following the 6MWT (ie, HRR3).

The researchers conducted a cohort study that included 46 adult patients with COPD recruited from Ziekenhuis Oost-Limburg, Genk, Belgium. Patients performed the 6MWT and relevant outcome data was collected. In addition, the researchers collected patient clinical and physiological data that included heart rate before the test, maximum heart rate after the test, decay after 3 minutes, fragmentation index, forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC) and FVC percent predicted, age and height, dyspnea score based on the mBorg scale, and peripheral capillary oxygen saturation (SpO2).  Using a Bayesian network approach, the patient data collected was used to develop a model predicting each patient’s 6MWD, HRmax after walking, and HRR3.


Continue Reading

Disease severity parameters were inferred by the Bayesian network based on actual 6MWT results, and patient status was assessed by the Bayesian network  predicting 6MWT outcomes. A strong correlation was found between the actual and estimated 6MWT measures for HRmax (R=0.84, mean absolute percentage error [MAPE]=8.10%). A moderate correlation was found between actual and estimated 6MWT measures for 6MWD (R=0.58, MAPE=15.43%) and for HRR3 (R=0.58, MAPE=32.49%). Researchers reported the classification of disease severity revealed more than 78% accuracy using 3 severity groups.

“[O]ur model provides a dual-function tool,” the researchers noted. “Firstly, the trained model allows the prediction of the 6MWT outcomes and thus, the evaluation of the functional exercise capacity of the patients. And secondly, it can assess the disease severity and progression by inferring the predefined FEV1 [percent predicted] groups, and how disease severity might progress (ie, improved or worsened) by modifying the available patient data. Both capabilities enable the proposed model to be used for more personalized monitoring of COPD patients in their home environment,” said the researchers.

Study limitations include underpowered sample size, the fact that the actual 6MWT was conducted only once, unaccounted-for comorbidities, and lack of a control group.

Researchers concluded that their model could become a “powerful tool to continuously monitor the COPD patient’s condition and disease progression at home, without physical performance measures.” Further studies are needed to “validate and refine the model, as well as to improve the overall performance, especially for extreme low or high FEV1 [percent] measures.”

Reference

Romero D, Blanco-Almazán D, Groenendaal W, et al. Predicting 6-minute walking test outcomes in patients with chronic obstructive pulmonary disease without physical performance measures. Comput Methods Programs Biomed. Published online July 11, 2022. doi:10.1016/j.cmpb.2022.107020

This article originally appeared on Pulmonology Advisor