Estimating Patient Satisfaction Through a Language Processing Model: Model Development and Evaluation.

BACKGROUND: Measuring patient satisfaction is a crucial aspect of medical care. Advanced natural language processing (NLP) techniques enable the extraction and analysis of high-level insights from textual data; nonetheless, data obtained from patients are often limited. OBJECTIVE: This study aimed to create a model that quantifies patient satisfaction based on diverse patient-written textual data. METHODS: We constructed a neural network-based NLP model for this cross-sectional study using the textual content from disease blogs written in Japanese on the Internet between 1994 and 2020. We extracted approximately 20 million sentences from 56,357 patient-authored disease blogs and constructed a model to predict the patient satisfaction index (PSI) using a regression approach. After evaluating the model's effectiveness, PSI was predicted before and after cancer notification to examine the emotional impact of cancer diagnoses on 48 patients with breast cancer. RESULTS: We assessed the correlation between the predicted and actual PSI values, labeled by humans, using the test set of 169 sentences. The model successfully quantified patient satisfaction by detecting nuances in sentences with excellent effectiveness (Spearman correlation coefficient [ρ]=0.832; root-mean-squared error [RMSE]=0.166; P<.001). Furthermore, the PSI was significantly lower in the cancer notification period than in the preceding control period (-0.057 and -0.012, respectively; 2-tailed t47=5.392, P<.001), indicating that the model quantifies the psychological and emotional changes associated with the cancer diagnosis notification. CONCLUSIONS: Our model demonstrates the ability to quantify patient dissatisfaction and identify significant emotional changes during the disease course. This approach may also help detect issues in routine medical practice.

Analysis of Patient Narratives in Disease Blogs on the Internet: An Exploratory Study of Social Pharmacovigilance.

BACKGROUND: Although several reports have suggested that patient-generated data from Internet sources could be used to improve drug safety and pharmacovigilance, few studies have identified such data sources in Japan. We introduce a unique Japanese data source: tobyoki, which translates literally as "an account of a struggle with disease." OBJECTIVE: The objective of this study was to evaluate the basic characteristics of the TOBYO database, a collection of tobyoki blogs on the Internet, and discuss potential applications for pharmacovigilance. METHODS: We analyzed the overall gender and age distribution of the patient-generated TOBYO database and compared this with other external databases generated by health care professionals. For detailed analysis, we prepared separate datasets for blogs written by patients with depression and blogs written by patients with rheumatoid arthritis (RA), because these conditions were expected to entail subjective patient symptoms such as discomfort, insomnia, and pain. Frequently appearing medical terms were counted, and their variations were compared with those in an external adverse drug reaction (ADR) reporting database. Frequently appearing words regarding patients with depression and patients with RA were visualized using word clouds and word cooccurrence networks. RESULTS: As of June 4, 2016, the TOBYO database comprised 54,010 blogs representing 1405 disorders. Overall, more entries were written by female bloggers (68.8%) than by male bloggers (30.8%). The most frequently observed disorders were breast cancer (4983 blogs), depression (3556), infertility (2430), RA (1118), and panic disorder (1090). Comparison of medical terms observed in tobyoki blogs with those in an external ADR reporting database showed that subjective and symptomatic events and general terms tended to be frequently observed in tobyoki blogs (eg, anxiety, headache, and pain), whereas events using more technical medical terms (eg, syndrome and abnormal laboratory test result) tended to be observed frequently in the ADR database. We also confirmed the feasibility of using visualization techniques to obtain insights from unstructured text-based tobyoki blog data. Word clouds described the characteristics of each disorder, such as "sleeping" and "anxiety" in depression and "pain" and "painful" in RA. CONCLUSIONS: Pharmacovigilance should maintain a strong focus on patients' actual experiences, concerns, and outcomes, and this approach can be expected to uncover hidden adverse event signals earlier and to help us understand adverse events in a patient-centered way. Patient-generated tobyoki blogs in the TOBYO database showed unique characteristics that were different from the data in existing sources generated by health care professionals. Analysis of tobyoki blogs would add value to the assessment of disorders with a high prevalence in women, psychiatric disorders in which subjective symptoms have important clinical meaning, refractory disorders, and other chronic disorders.

Effect of Predialysis Recombinant Human Erythropoietin on Early Survival After Hemodialysis Initiation in Patients With Chronic Kidney Disease: Co-JET Study.

Progression of anemia in patients with chronic kidney disease (CKD) is associated with an increased risk of death and hospitalization. It is not sufficiently clear whether treating renal anemia with recombinant human erythropoietin (rHuEPO) has a beneficial effect on early survival after hemodialysis (HD) initiation in patients with CKD. The study was an open-label multicenter retrospective cohort study to evaluate the relationship between rHuEPO treatment and early survival after HD initiation in patients with CKD. Predialysis patients with CKD were divided into two groups: an rHuEPO-treated group (rHuEPO group) and a non-treatment group. The primary endpoint was all-cause mortality in the year after HD initiation. A total of 3261 patients were enrolled (2275 in the rHuEPO group and 986 in the non-treatment group). One-year survival was 95.36% in the rHuEPO group and 90.36% in the non-treatment group. The survival rate was significantly higher in the rHuEPO group (P < 0.0001). The results of multivariate analysis confirmed that predialysis treatment with rHuEPO is a predictor for reduced mortality risk (hazard ratio = 0.61, 95% confidence interval: 0.42-0.87, P = 0.006). Risk for the composite event of death/hospitalization was also lower in the rHuEPO group (hazard ratio = 0.88, 95% confidence interval: 0.78-0.98, P = 0.026). The results of this study suggest that treatment with rHuEPO can decrease early mortality risk after initiation of HD in patients with CKD. A prospective study is needed to further investigate early survival after HD initiation.

Bias in spontaneous reporting of adverse drug reactions in Japan.

BACKGROUND: Attitudes of healthcare professionals regarding spontaneous reporting of adverse drug reactions (ADRs) in Japan are not well known, and Japan's unique system of surveillance, called early post-marketing phase vigilance (EPPV), may affect these reporting attitudes. Our objectives were to describe potential effects of EPPV and to test whether ADR seriousness, prominence, and frequency are related to changes in reporting over time. METHODS: A manufacturer's database of spontaneous ADR reports was used to extract data from individual case safety reports for 5 drugs subject to EPPV. The trend of reporting and the time lag between ADR onset and reporting to the manufacturer were examined. The following indices for ADRs occurring with each drug were calculated and analyzed to assess reporting trends: Serious:Non-serious ratio, High prominence:Low prominence ratio, and High frequency:Low frequency ratio. RESULTS: For all 5 drugs, the time lag between ADR onset and reporting to the manufacturer was shorter in the EPPV period than in the post-EPPV period. All drugs showed higher Serious:Non-serious ratios in the post-EPPV period. No specific patterns were observed for the High prominence:Low prominence ratio. The High frequency:Low frequency ratio for peginterferon alpha-2a and sevelamer hydrochloride decreased steadily throughout the study period. CONCLUSIONS: Healthcare professionals may be more likely to report serious ADRs than to report non-serious ADRs, but the effect of event prominence on reporting trends is still unclear. Factors associated with ADR reporting attitude in Japan might be different from those in other countries because of EPPV and the involvement of medical representatives in the spontaneous reporting process. Pharmacovigilance specialists should therefore be cautious when comparing data between different time periods or different countries. Further studies are needed to elucidate the underlying mechanism of spontaneous ADR reporting in Japan.

Impacts of recombinant human erythropoietin treatment during predialysis periods on the progression of chronic kidney disease in a large-scale cohort study (Co-JET study).

The effect of recombinant human erythropoietin (rHuEPO) treatment on the progression of chronic kidney disease (CKD) has not been fully evaluated in Japan. We therefore retrospectively evaluated this in a sub-cohort of a prospective multicenter study to investigate optimal hemoglobin (Hb) level of CKD patients on hemodialysis (HD) treated with rHuEPO; Japan Erythropoietin Treatment Study for Target Hb and Survival (JET study). Effect of rHuEPO treatment during predialysis period to delay initiation of HD was retrospectively assessed in 2434 patients from the JET study comparing groups with and without rHuEPO treatment. The assessment was done by Cox proportional hazards regression analysis and inverse probability-weighted (IPW) analysis to adjust for time-dependent confounders. The weights used in the IPW analysis were calculated using a logistic model that included baseline confounders and time-dependent variables. During the predialysis period, 71.7% (1746 patients) were treated with rHuEPO (mean Hb level of 8.7 g/dL at initiation of rHuEPO treatment). Covariates significantly associated with initiation of rHuEPO treatment were Hb level, serum creatinine level, age, diabetes, cardiac insufficiency, and hypertension. The adjusted hazard ratio for time until HD initiation under rHuEPO treatment was 0.272 (95% CI, 0.223-0.331; P < 0.001) in the Cox analysis and 0.63 (95% CI, 0.53-0.76; P < 0.0001) in the IPW analysis. This retrospective study suggests that rHuEPO treatment during the predialysis period has preventive effects on the progression of CKD although further prospective investigation on the efficacy is needed.

Low hemoglobin levels and hypo-responsiveness to erythropoiesis-stimulating agent associated with poor survival in incident Japanese hemodialysis patients.

Although erythropoiesis-stimulating agents (ESAs) are effective at treating anemia, the association between hemoglobin (Hb) levels and survival is still unclear, especially for the incident Japanese hemodialysis (HD) population. The Japan Erythropoietin Treatment (JET) Study is an open multi-center, prospective, observational study designed to evaluate the relationship between the maintenance of Hb levels and new HD patient prognosis after the first administration of epoetin beta. Landmark analyses were performed to examine the relationship between Hb levels at 6 months and survival. Among a total of 10,310 patients, 6631 completed the initial 6 months of epoetin beta treatment (induction phase) and were followed up for a further 2.5 years (maintenance phase). Three-year survival rate of patients with <9 g/dL Hb levels after 6 months was 74.1%, which was significantly lower than 89.3% for patients with Hb levels 10 to 11 g/dL; the adjusted hazard ratio (HR) was 2.08 (95% CI, 1.57-2.77; P < 0.0001). Moreover, the 3-year survival rate for poor responders defined by Hb levels <10 g/dL and weekly epoetin beta doses ≥ 9000 IU during the induction phase was 71.6%, which was significantly lower than 89.4% for the group, which had Hb levels 10 to 11 g/dL excluding poor responders and those with excursion; the HR was 1.71 (95% CI, 1.13-2.60; P = 0.0118). Adverse events related to the treatment were reported in 71 of 10,310 patients (0.69%). These findings suggest that the achieved low Hb levels and poor response to ESA therapy are significantly associated with high mortality.