Identifying cancer patients who received palliative care using the SPICT-LIS in medical records: a rule-based algorithm and text-mining technique.

BACKGROUND: Due to limited numbers of palliative care specialists and/or resources, accessing palliative care remains limited in many low and middle-income countries. Data science methods, such as rule-based algorithms and text mining, have potential to improve palliative care by facilitating analysis of electronic healthcare records. This study aimed to develop and evaluate a rule-based algorithm for identifying cancer patients who may benefit from palliative care based on the Thai version of the Supportive and Palliative Care Indicators for a Low-Income Setting (SPICT-LIS) criteria. METHODS: The medical records of 14,363 cancer patients aged 18 years and older, diagnosed between 2016 and 2020 at Songklanagarind Hospital, were analyzed. Two rule-based algorithms, strict and relaxed, were designed to identify key SPICT-LIS indicators in the electronic medical records using tokenization and sentiment analysis. The inter-rater reliability between these two algorithms and palliative care physicians was assessed using percentage agreement and Cohen's kappa coefficient. Additionally, factors associated with patients might be given palliative care as they will benefit from it were examined. RESULTS: The strict rule-based algorithm demonstrated a high degree of accuracy, with 95% agreement and Cohen's kappa coefficient of 0.83. In contrast, the relaxed rule-based algorithm demonstrated a lower agreement (71% agreement and Cohen's kappa of 0.16). Advanced-stage cancer with symptoms such as pain, dyspnea, edema, delirium, xerostomia, and anorexia were identified as significant predictors of potentially benefiting from palliative care. CONCLUSION: The integration of rule-based algorithms with electronic medical records offers a promising method for enhancing the timely and accurate identification of patients with cancer might benefit from palliative care.

Diabetes Mellitus as Cancer Risk: A 14-year, Cross-Sectional Analysis.

Diabetes mellitus is widely thought to be a risk factors of cancers, but evidence of the association remains inconclusive, especially in Asian countries where few relevant studies have been conducted. Our study aimed to estimate overall and specific types of cancer risks among diabetes patients in Southern Thailand. Patients diagnosed with diabetes who visited the outpatient clinic of Songklanagarind Hospital during 2004 to 2018 were included. Newly diagnosed cancer patients were identified using the hospital-based cancer registry. Age-standardized incidence ratios (ASRs) and standardized incidence ratios (SIRs) were used to estimate and compare the cancer risks among diabetes patients and the general population in Southern Thailand. Of 29,314 diabetes patients identified during the study period, 1,113 patients had developed cancer. An increased risk for overall cancer was observed in both genders, with SIRs [95% CI] of 2.99 [2.65, 3.39] in men and 3.51 [3.12, 3.96] in women. Increases in the risk of several site-specific cancers including liver cancer, non-melanoma skin cancer, colon cancer and lung cancer in both sexes; prostate cancer, lymphoid leukemia, and multiple myeloma in men; and endometrial, breast, and thyroid cancer in women were observed. Our study found that diabetes generally increased the risk of both overall and site-specific cancers.

Systematic adaptation of the Thai version of the supportive and palliative care indicators tool for low-income setting (SPICT-LIS).

BACKGROUND: Identification of patients who might benefit from palliative care among countries with different socioeconomic and medical contexts is challenging. The Supportive and Palliative Care Indicators Tool for a Low-income Setting (SPICT-LIS) was designed to help physicians identify patients in low-income setting who might benefit from palliative care. We aimed to systematically adapt and refine the SPICT-LIS for Thai general palliative care providers. METHODS: We followed the WHO guidelines for translation, cross-cultural adaptation and validation of an instrument for the SPICT-LIS. Three expert panel members did the initial adaptation using forward and backward translations with pretested data. Two iterations of pretesting were conducted to test for applicability and reliability. The case vignettes which were used in the pretesting were modified hospital medical records. The pretesting was done with 30 respondents from various specialties in a community health center and 34 general palliative care providers from a regional referral hospital in the first and second iterations, respectively. To examine instrument reliability, interrater reliability and internal consistency were evaluated. Cognitive interviewing was conducted using semi-structured interviews with general practitioners (GPs) using the "think aloud strategy" and "probing questions". RESULTS: The adapted Thai SPICT-LIS had a total of 34 indicators which included 6 general and 28 clinical indicators. The assessment of the adapted Thai SPICT-LIS found that it provided consistent responses with good agreement among the GPs, with a Fleiss kappa coefficient of 0.93 (0.76-1.00). The administration time was 2.3-4.3 min per case. Most respondents were female. The 8 interviewed GPs said they felt that the SPICT-LIS was appropriate for use in a general setting in Thailand. CONCLUSION: The study found that the Thai SPICT-LIS could be an applicable, acceptable, and reliable tool for general palliative care providers in Thailand to identify patients who might benefit from palliative care.

The trend of opioid prescriptions among cancer patients in a tertiary hospital: A multimethod quantitative study.

Introduction: Pain is a major symptom in cancer patients. World Health Organization recommends opioids as the main analgesic agent. Few studies have examined the amount of opioid uses in cancer patients in Southeast Asia, however, none of them have examined the factors associated with the amount of opioid uses which were lower than required. Objectives: To assess the trends and factors associated with opioid prescriptions for cancer patients in Songklanagarind Hospital, the largest referral center in Southern Thailand. Design: Multi-method quantitative study. Methods: We reviewed the electronic medical records of 20,192, outpatients aged ≥18 years diagnosed with cancer between 2016 and 2020 who received opiod prescriptions. Oral morphine equivalents (OME) were calculated using the standard conversion factors and the OME trend during the study period was assessed by a generalized additive model. Factors affecting the morphine equivalent daily dose (MEDD) were assessed using multiple linear regression with a generalized estimating equation. Results: The mean overall MEDD for all study patients was 27.8 ± 21.9 mg per day per patient. The bone and articular cartilage cancer patients had the highest MEDD. For every 5-year increase in the duration of cancer, the MEDD increased by 0.02 (95% confidence interval [CI]: 0.01 - 0.04). Patients with stage 4 cancer received a higher average MEDD of 4.04 (95% CI: 0.30-7.62) as compared to those with stage 1 cancer. Patients with bone metastasis received a average higher MEDD of 4.03 (95% CI: 0.82-7.19) compared to those without. Age was inversely associated with the MEDD. Patients aged 42-58, 59-75 and >76years old received MEDDs of 4.73 (95% CI: 2.31-7.15), 6.12 (95% CI: 3.66-8.59) and 8.59 (95% CI: 6.09-11.09) compared with those aged 18-42 years old. Brain metastasis was inversely associated with MEDD of 4.49 (95% CI: 0.61-8.37) compared to those without. Conclusion: Opioid use in cancer patients in this study is lower than the average global usage. Promoting opioid prescriptions for pain management through medical education can help doctors overcome opiophobia.

Drug Recommendation from Diagnosis Codes: Classification vs. Collaborative Filtering Approaches.

Over time, large amounts of clinical data have accumulated in electronic health records (EHRs), making it difficult for healthcare professionals to navigate and make patient-centered decisions. This underscores the need for healthcare recommendation systems that help medical professionals make faster and more accurate decisions. This study addresses drug recommendation systems that generate an appropriate list of drugs that match patients' diagnoses. Currently, recommendations are manually prepared by physicians, but this is difficult for patients with multiple comorbidities. We explored approaches to drug recommendations based on elderly patients with diabetes, hypertension, and cardiovascular disease who visited primary-care clinics and often had multiple conditions. We examined both collaborative filtering approaches and traditional machine-learning classifiers. The hybrid model between the two yielded a recall at 5 of 76.61%, a precision at 5 of 46.20%, a macro-averaged area under the curve of 74.52%, and an average physician agreement of 47.50%. Although collaborative filtering is widely used in recommendation systems, our results showed that it consistently underperformed traditional classification. Collaborative filtering was sensitive to class imbalances and favored the more popular classes. This study highlighted challenges that need to be addressed when developing recommendation systems in EHRs.