Assessing Factors Predicting Pediatricians' Decisions to Refer Young Children for Autism Evaluatio

Assessing factors predicting pediatricians' decisions to refer young children for autism evaluation is instrumental in reducing the lack of referrals for this disorder. A demographic survey was originally placed on the researchers' Face Book and LinkedIn pages to obtain participants with no results. Next, the use of an email company (MMS, Inc.,) and the U.S. Postal Service (USPS) acquired the required number of participants from the American Medical Association (AMA). Multiple linear regression analysis allowed a determination to be made on specific demographic factors that influenced pediatricians' referral rates for children suspected of possessing autistic characteristics for further evaluation before 24 months of age. The survey collected data from a minimum of N= 85 pediatricians ranging in age from 32-75, as defined by G-Power Analysis with an effect size ( f2) of .15, alpha error probability of p < .05, and power (1-beta error probability) of .80. Statistical data analyses provided evidence that there was a statistically significant relationship between predictor variables (pediatricians' age at p < .020, and pediatricians' gender at p < .021) and the outcome variable of pediatricians' referral rates. Additionally, it was determined that there was a significant correlation between variables in combination, documented as pediatricians' gender at p < .002 and pediatricians' geographic location that approached a significant correlation at p > .051. However, the overall model for those findings were not statistically significant. Finally, a backward elimination regression showed that pediatricians' gender (female) and pediatricians' geographic location (metropolitan) were better predictor of pediatricians' referral rates. Pediatricians' gender (females) had a statistical significance of 0.029 or p < .05 and an unstandardized beta coefficient of -.577, whereas, pediatricians' geographic location (metropolitan) had a statistical significance of .090 or p > .05 and an unstandardized beta coefficient of -.233. However, both variables would be needed to accurately predict pediatricians' referral rates. Thus, the data analysis provided information necessary to develop recommendations for future research that would remediate the non-influential factors by using larger sample sizes and different outcome measures. Recommendations for future research also supported a focus on more practical research being conducted, which examines the influence of other variables on referral rates of pediatricians. Finally, it is recommended that replication of this study be completed to confirm the results found.