The impact of technological intensity of service provision on physician expenditures: an exploratory investigation.

Advances in technology and subsequent changes in clinical practice can lead to increases in healthcare costs. Our objective is to assess the impact that changes in the technological intensity of physician-provided health services have had on the age pattern of both the volume of services provided and the average expenditures associated with them. We based our analysis on age-sex-specific patient-level administrative records of diagnoses and treatments. These records include virtually all physician services provided in the province of Ontario, Canada in a 10-year span ending in 2004 and their associated costs. An algorithm is developed to classify services and their costs into three levels of technological intensity. We find that while the overall age-standardized level and cost of services per capita have decreased, the volume and cost of high technologically intensive treatments have increased, especially among older patients.

Predicting the impact of hospital health information technology adoption on patient satisfaction.

OBJECTIVES: To develop and explore the predictability of patient perceptions of satisfaction through the hospital adoption of health information technology (HIT), leading to a better understanding of the benefits of increased HIT investment. DATA AND METHODS: The solution proposed is based on comparing the predictive capability of artificial neural networks (ANNs) with the adaptive neuro-fuzzy inference system (ANFIS). The latter integrates artificial neural networks and fuzzy logic and can handle certain complex problems that include fuzziness in human perception, and non-normal and non-linear data. Secondary data from two surveys were combined to develop the model. Hospital HIT adoption capability and use indicators in the Canadian province of Ontario were used as inputs, while patient satisfaction indicators of healthcare services in acute hospitals were used as outputs. RESULTS: Eight different types of models were trained and tested for each of four patient satisfaction dimensions. The accuracy of each predictive model was evaluated through statistical performance measures, including root mean square error (RMSE), and adjusted coefficient of determination R(2)(Adjusted). For all four patient satisfaction indicators, the performance of ANFIS was found to be more effective (R(Adjusted)(2)=0.99) when compared with the results of ANN modeling in predicting the impact of HIT adoption on patient satisfaction (R(Adjusted)(2)=0.86-0.88). CONCLUSIONS: The impact of HIT adoption on patient satisfaction was obtained for different HIT adoption scenarios using ANFIS simulations. The results through simulation scenarios revealed that full implementation of HIT in hospitals can lead to significant improvement in patient satisfaction. We conclude that the proposed ANFIS modeling technique can be used as a decision support mechanism to assist government and policy makers in predicting patient satisfaction resulting from the implementation of HIT in hospitals.