Predicting health-related quality of life change using natural language processing in thyroid cancer.

Background: Patient-reported outcomes (PRO) allow clinicians to measure health-related quality of life (HRQOL) and understand patients' treatment priorities, but obtaining PRO requires surveys which are not part of routine care. We aimed to develop a preliminary natural language processing (NLP) pipeline to extract HRQOL trajectory based on deep learning models using patient language. Materials and methods: Our data consisted of transcribed interviews of 100 patients undergoing surgical intervention for low-risk thyroid cancer, paired with HRQOL assessments completed during the same visits. Our outcome measure was HRQOL trajectory measured by the SF-12 physical and mental component scores (PCS and MCS), and average THYCA-QoL score.We constructed an NLP pipeline based on BERT, a modern deep language model that captures context semantics, to predict HRQOL trajectory as measured by the above endpoints. We compared this to baseline models using logistic regression and support vector machines trained on bag-of-words representations of transcripts obtained using Linguistic Inquiry and Word Count (LIWC). Finally, given the modest dataset size, we implemented two data augmentation methods to improve performance: first by generating synthetic samples via GPT-2, and second by changing the representation of available data via sequence-by-sequence pairing, which is a novel approach. Results: A BERT-based deep learning model, with GPT-2 synthetic sample augmentation, demonstrated an area-under-curve of 76.3% in the classification of HRQOL accuracy as measured by PCS, compared to the baseline logistic regression and bag-of-words model, which had an AUC of 59.9%. The sequence-by-sequence pairing method for augmentation had an AUC of 71.2% when used with the BERT model. Conclusions: NLP methods show promise in extracting PRO from unstructured narrative data, and in the future may aid in assessing and forecasting patients' HRQOL in response to medical treatments. Our experiments with optimization methods suggest larger amounts of novel data would further improve performance of the classification model.

Patients Follow Different Financial Hardship Trajectories in the Year after Injury.

OBJECTIVE: To determine if distinct financial trajectories exist and if they are associated with quality-of-life outcomes. SUMMARY OF BACKGROUND DATA: Financial hardship after injury measurably impacts Health-Related Quality of Life outcomes. Financial hardship, encompassing material losses, financial worry, and poor coping mechanisms, is associated with lower quality of life and increased psychological distress. However, recovery is dynamic and financial hardship may change over time. METHODS: This is a secondary analysis of a cohort of 500 moderate-to-severe nonneurologic injured patients in which financial hardship and Health-related Quality of Life outcomes were measured at 1, 2, 4, and 12 months after injury using survey instruments (Short Form-36). Enrollment occurred at an urban, academic, Level 1 trauma center in Memphis, Tennessee during January 2009 to December 2011 and follow-up completed by December 2012. RESULTS: Four hundred seventy-four patients had sufficient data for Group- Based Trajectory Analysis. Four distinct financial hardship trajectories were identified: Financially Secure patients (8.6%) had consistently low hardship over time; Financially Devastated patients had a high degree of hardship immediately after injury and never recovered (51.6%); Financially Frail patients had increasing hardship over time (33.6%); and Financially Resilient patients started with a high degree of hardship but recovered by year end (6.2%). At 12-months, all trajectories had poor Short Form-36 physical component scores and the Financial Frail and Financially Devastated trajectories had poor mental health scores compared to US population norms. CONCLUSIONS AND RELEVANCE: The Financially Resilient trajectory demonstrates financial hardship after injury can be overcome. Further research into understanding why and how this occurs is needed.

What is the experience of our patients with transient hypoparathyroidism after total thyroidectomy?

BACKGROUND: We sought to better understand the experience of patients with transient hypoparathyroidism using patient interviews and quality of life surveys. METHODS: This is a prospective analysis of 62 patients after total thyroidectomy at a high-volume institution. Semistructured patient interviews and quality of life surveys were conducted preoperatively and postoperatively at 2 weeks, 6 weeks, 6 months, and 1 year and compared based on postoperative parathyroid hormone levels. RESULTS: Postoperative parathyroid hormone levels were <10 pg/mL in 32% of patients (n = 20), 10 to 20 pg/mL in 19% (n = 12), and >20 pg/mL in 48% (n = 30). Hypocalcemic symptoms at 2 weeks were reported in 28 of 55 patients (51%), but patients felt "well prepared" and reported it "wasn't a big deal." If symptoms persisted at 6 weeks, they became more bothersome. At 6 months and 1 year, patients reported calcium supplementation prevented most symptoms and did not interfere with daily activities. Quality of life as measured by the European Organization for Research and Treatment of Cancer and the 12-Item Short Form Survey demonstrated a slight improvement at 1 year postoperatively regardless of parathyroid hormone level. CONCLUSION: Early postoperative transient hypoparathyroidism is common but when appropriately managed did not have a substantial negative impact on the overall quality of life.

A Randomized Controlled Clinical Trial: No Clear Benefit to Prophylactic Central Neck Dissection in Patients With Clinically Node Negative Papillary Thyroid Cancer.

OBJECTIVE: The aim of this prospective randomized-controlled trial was to evaluate the risks/benefits of prophylactic central neck dissection (pCND) in patients with clinically node negative (cN0) papillary thyroid cancer (PTC). BACKGROUND: Microscopic lymph node involvement in patients with PTC is common, but the optimal management is unclear. METHODS: Sixty patients with cN0 PTC were randomized to a total thyroidectomy (TT) or a TT+ pCND. All patients received postoperative laryngoscopies and standardized radioiodine treatment. Thyroglobulin (Tg) levels and/or neck ultrasounds were performed at 6 weeks, 6 months, and 1 year. RESULTS: Tumors averaged 2.2 ±â€Š0.2 cm and 11.9% had extra-thyroidal extension. Thirty patients underwent a pCND and 27.6% had positive nodes (all ≤6 mm). Rates of postoperative PTH < 10 (33.3% vs 24.1%, P = 0.57) and transient nerve dysfunction (13.3% vs 10.3%, P = 1.00) were not significantly different between groups. Six weeks after surgery, both TT and TT + pCND were equally likely to achieve a Tg < 0.2 (54.5% vs 66.7%, P = 0.54) and/or a stimulated Tg (sTg) <1 (59.3% vs 64.0%, P = 0.78). At 1 year, rates of Tg < 0.2 (88.9% vs 90.0%, P = 1.00) and sTg < 1 (93.8% vs 92.3%, P = 1.00) remained similar between groups. Neck ultrasounds at 1 year were equally likely to be read as normal (85.7% in TT vs 85.1% in pCND, P = 1.00). CONCLUSIONS: cN0 PTC patients treated either with TT or TT + pCND had similar complication rates after surgery. Although microscopic nodes were discovered in 27.6% of pCND patients, oncologic outcomes were comparable at 1 year.